Natural capital and distributive justice - UQ eSpace

Natural capital and distributive justice; a

multidisciplinary, multi-scalar assessment Glenn Althor

Bachelor of Environmental Management

A thesis submitted for the degree of Doctor of Philosophy at

The University of Queensland in 2019

Faculty of Science

2

Abstract

People depend on natural capital goods and services for their well-being. However, throughout

history the distribution of the benefits and burdens of natural capital exploitation has not been equal

across people and groups. These unequal distributions can result in matching unequal distributions

of human capability and quality of life. Within this thesis, I engage with distributive justice theories

which allow for an examination of the fairness of distributive patterns and apply these insights to

real world natural capital exploitation cases across diverse geographic scales and locations, using a

range of methodological approaches. I use empirical research methods to describe distributions and

use theoretical frameworks of distributive fairness to evaluate them.

To do this, I sought to meaningfully engage with describing and evaluating the distribution of

natural capital in a way that offers new links between theoretical perspectives and empirical

applications; a critical synthesis for advancing knowledge on distributive justice. I therefore have

pursued an overarching and primary research question: Is it possible to re-frame distributive

justice theories in ways that are more amenable to (a diversity of) empirical research

methods? To answer this research question, the thesis first presents new principles for examining

natural capital distributive justice, developed through theoretical argumentation and presented in

Chapter 1:

P1: Social and economic inequalities, resulting from the utilisation of natural capital, are to

be of the greatest benefit of the least advantaged members of society.

P2: Where harm is accepted as a tolerable by-product of natural capital use for the

betterment of society, exposure levels of harm ought to be equal among all persons.

P3: Where burdens upon any individual or exclusive group are the by-product of the use of

natural capital for the betterment of another individual or exclusive group, burdens are to be

accompanied by (at least) commensurate benefits.

These principles provide a new theoretical contribution to the study of natural capital distributive

justice in that they explain foundational principles from theories of justice in a form suitable for

empirical interrogation. This potential of the new principles, then, is examined in the remainder of

the thesis through multi-scalar cases.

First, in Chapter 2, a systematic review of the natural capital distributive justice is presented in

order to understand the breadth of literature of relevance to this thesis. I find this literature is rapidly

growing, and somewhat diverse, but is dominated by particular geographies, methods,

demographics, livelihood indicators, and topics. Overall, evidence indicates that natural capital

3

distributive issues are inequitable, with over half of the literature yielding this result, but with many

unclear determinations and only 2% of studies finding equitable distributions.

Next, in Chapter 3, I statistically analysed the global distributions of benefits and burdens resulting

from global greenhouse gas emissions. In this analysis I find an unjust inequity, which is globally

pervasive, in which nations that have the greatest responsibility for climate change are also those

that are the least vulnerable.

Then, shifting scale from global to national, in Chapter 4 I present statistical modelling that

examines Australia-wide the distributive relationships between industrial pollution and

socioeconomic indicators. Through this detailed empirical analysis, I find little evidence of unfair

distributions, a finding which does not align with existing research. This points to the importance of

policy in shaping the equitability of outcomes and methodological design in how we establish

evidence.

Again, shifting scale from national to local scale, in Chapter 5 I present a qualitative study of the

lived experiences of Cambodian subsistence fisher-people in the context of rapid anthropogenic

environmental change. This study offers a rich and in-depth complement to the prior quantitative

analyses that unpacks the ways in which natural capital distributive inequities affect some of the

world’s most vulnerable people. I find that these communities have little capacity to respond to this

change and are suffering unfair burdens as a result of economic activities from which they derive no

benefit.

Taken together, these studies demonstrate the multiple ways in which natural capital distributive

justice can be examined across scales and with a range of methodological approaches. The thesis

also contributes new understandings about pressing real-world distributive justice issues. In Chapter

6 I offer a synthesis of these findings, returning again to my principles for study of natural capital

distributive justice. This thesis highlights the importance of continued scholarship on natural capital

distributive justice, and I conclude that the four studies show that there are indeed ways in which

theory and methods can be integrated to answer questions about distributive justice in real world

settings. However, my research shows that there is no singular answer to the question of justice in

natural capital distribution; outcomes are context dependent. Critically, I also conclude that natural

capital distributive justice can be evaluated transparently and rigorously through empirical research

supported by my distributive justice principles.

4

Declaration by author

This thesis is composed of my original work, and contains no material previously published or

written by another person except where due reference has been made in the text. I have clearly

stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical

assistance, survey design, data analysis, significant technical procedures, professional editorial

advice, financial support and any other original research work used or reported in my thesis. The

content of my thesis is the result of work I have carried out since the commencement of my higher

degree by research candidature and does not include a substantial part of work that has been

submitted to qualify for the award of any other degree or diploma in any university or other tertiary

institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for

another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and,

subject to the policy and procedures of The University of Queensland, the thesis be made available

for research and study in accordance with the Copyright Act 1968 unless a period of embargo has

been approved by the Dean of the Graduate School.

I acknowledge that copyright of all material contained in my thesis resides with the copyright

holder(s) of that material. Where appropriate I have obtained copyright permission from the

copyright holder to reproduce material in this thesis and have sought permission from co-authors for

any jointly authored works included in the thesis.

5

Publications included in this thesis

Althor G and Witt B. (2019) A quantitative systematic review of distributive environmental justice

literature: a rich history and the need for an enterprising future. Journal of Environmental Studies

and Sciences online.

– incorporated as Chapter 2.

Contributor Statement of contribution

Althor G (Candidate) Conception and design (100%)

Analysis and interpretation (90%)

Drafting and production (90%)

Witt B Conception and design (0%)



Althor G, Watson JEM, Fuller RA. (2016) Global mismatch between greenhouse gas emissions and

the burden of climate change. Scientific Reports 6.

– incorporated as Chapter 3.





Watson JEM Conception and design (10%)

Analysis and interpretation (2 %)


Fuller RA Conception and design (10%)



Althor G, Mahood S, Witt B, Colvin RM, Watson JEM. 2018. Large-scale environmental

degradation results in inequitable impacts to communities: a case study from the floating villages of

Cambodia. Ambio 47 (7).

- incorporated as Chapter 5.


6




Colvin R Conception and design (5%)






Mahood S Conception and design (5%)



Watson J Conception and design (5%)



Submitted manuscripts included in this thesis

Althor G, Witt B, Colvin RM. (forthcoming) The distribution of industrial pollution across socio-

economic groups in Australia; a ten-year, a fine scale distributive justice analysis. Target Journal:

Environmental pollution

- incorporated as Chapter 4.








Other publications during candidature

Althor G, McKinnon M, Cheng SH, Klein C, Watson J. (2016) Does the social equitability of

community and incentive based conservation interventions in non-OECD countries, affect human

well-being? A systematic review protocol. Environmental Evidence 5 (1):26.

7

Brown CJ, Althor G, Halpern BS, et al. (2018) Trade-offs in triple-bottom-line outcomes when

recovering fisheries. Fish and Fisheries 19: 107-116.

Di Marco M, Chapman S, Althor G, et al. (2017) Changing trends and persisting biases in three

decades of conservation science. Global Ecology and Conservation 10: 32-42.

Lewis SC, Perkins-Kirkpatrick SE, Althor G, et al. (2019) Assessing contributions of major

emitters' Paris-era decisions to future temperature extremes. Geophysical Research Letters 46:7.

McIntosh EJ, Chapman S, Kearney SG, et al. (2018) Absence of evidence for the conservation

outcomes of systematic conservation planning around the globe: a systematic map. Environmental

Evidence 7: 22.

Contributions by others to the thesis

No contributions by others.

Statement of parts of the thesis submitted to qualify for the award of

another degree

“No works submitted towards another degree have been included in this thesis”.

Research Involving Human or Animal Subjects

Chapter 5 involved human subjects. Ethics approval granted by Dr Paul Dargusch at the University

of Queensland, GPEM Ethics Officer. Approval number: GPEM number 20150701

8

Acknowledgements

I acknowledge the Traditional Owners of the continent now called Australia. In particular, I

acknowledge that I now live and work on land of the Ngunnawal people – the Traditional Owners

of the area now known as Canberra, and I also acknowledge the Traditional Owners of the land on

which the University of Queensland is situated, the Turrbal and Jagera people. I acknowledge that

this land was invaded and colonised by my European ancestors, and that I live in privilege as a

result of stolen land and resources, historic and ongoing racism, and inhumane structural violence. I

acknowledge that the strength and resilience of Aboriginal and Torres Strait Islander people is

astounding, and that Australians like me need to do more in supporting Aboriginal and Torres Strait

Islander people to freely achieve the functionings and capabilities they deem valuable.

Interpersonally, first and foremost I want to give thanks and praise to my partner, best friend,

mentor, teacher, and all-round infinite-patience-machine, Bec Colvin. Without her guidance,

kindness, and companionship this thesis simply would not exist. There’s not a word, thought, idea,

or statistic in this thesis that hasn’t been touched by Bec in some manner. You’re the best!

I also want to thank my Mum, Anille ‘Nelly’ Ellis, who only got to see her eldest son’s first thirteen

years. Even though Fortune only allowed us to spend a shockingly short time together, my Mum

instilled in me an ethic of kindness and empathy for others, a profound love of animals and the

natural world, and a powerful devotion to scholarship and knowledge. Every single day the wisdom,

concern, and care that Mum imparted on me motivates me to work towards supporting the world’s

least represented and most vulnerable people in their fight for the lives and capabilities we all

deserve.

A million and more thanks to my wonderful friend, mentor, and supervisor, Bradd Witt. Bradd’s

disciplined patience, kindness, and support have pushed me over the line many times when I was

facing seemingly insurmountable barriers and thinking this this thesis was an impossible task.

My thanks to my ‘Outlaw’ family – the Colvins and the Hoogleys. All of whom have supported me

with a grace and kindness that I will never be able to repay.

Many thanks to the wonderful folks who gave me the opportunity to discover my love of tertiary

teaching, as a tutor, demonstrator, and lecturer in their courses. Karen McNamara, Lorrae Van

9

Kerkhoff, Bruce Doran, Matthew Brookhouse, and Sophie Lewis. Thanks also to all my amazing

friends and family who are sadly too numerous to name here, I hope you know who you are and

how much I appreciate your support!

I also want to thank all the wonderful people who in big and little ways have made this thesis a

reality with kindness and support, and those who I’ve also had the pleasure to work and write with.

It’s been an absolute privilege to work with some of the most amazing minds from around the

world, including: James Watson my co-supervisor; and the folks at WCS Cambodia for their

amazing support during and beyond my fieldwork (particularly Sarah, Simon, Visal, and Ross).

Also including my co-authors (in no particular order, you’re all amazing) Richard Fuller, Madeleine

Mckinnon, Samantha Cheng, Carissa Klein, Simon Mahood, Moreno DiMarco, Sarah Chapman,

Stephen Kearney, Joseph Maina, Oscar Venter, Christopher Brown, Ben Halpern, Sayed Iftekhar,

Elizabeth Pride, Steve Schlizzi, Becky Twohey, Hugh Possingham, Emma McIntosh, Brooke

Williams, Jessica Thorn, Sophie Lewis, Sarah Perkins-Kirkpatrick, Andrew King, Luke Kemp, and

of course all the wonderful academics at ANU and UNSW who I’m currently developing new

research horizons with! Finally, I want to thank all of the folks who have provided other valuable

support and help along the way, particularly Karen McNamara, Ruth Garside, Phil Kokic, Grace

Chou, Genna and the world’s best finance team - SEES finance, and all the amazing and

hardworking UQ librarians.

I wish there was more I could do to show my appreciation for everyone beyond a simple thank you,

but I promise that I ‘pay it forward’, and use your kindness to guide me whenever I have

opportunities to support others.

10

Financial support

Chapter 5 was part funded by an Australian Research Council Discovery Grant (DP140100733) to

JW. Additional funding during candidature by The Australian Postgraduate Award scholarship to

GA (PN: 3018293). Funding for translation work provided by WCS, Cambodia.

Keywords

Distributive justice, natural resource management, natural capital, human wellbeing, environmental

management, social justice.

Australian and New Zealand Standard Research Classifications

(ANZSRC)

ANZSRC code: 220104 Human Rights and Justice Issues 34%

ANZSRC code: 050205 Environmental Management 33%

ANZSRC code: 160403 Social and Cultural Geography 33%

Fields of Research (FoR) Classification

FoR code: 0502 Environmental Science and Management 34%

FoR code: 1604 Human Geography 33%

FoR code: 2201 Human Rights and Justice Issues 33%

11

Table of Contents

Abstract ................................................................................................................................................ 2

Declaration by author ........................................................................................................................... 4

Publications included in this thesis ...................................................................................................... 5

Submitted manuscripts included in this thesis ..................................................................................... 6

Other publications during candidature ................................................................................................. 6

Contributions by others to the thesis .................................................................................................... 7

Statement of parts of the thesis submitted to qualify for the award of another degree ........................ 7

Research Involving Human or Animal Subjects .................................................................................. 7

Acknowledgements .............................................................................................................................. 8

Financial support ................................................................................................................................ 10

Keywords ........................................................................................................................................... 10

Australian and New Zealand Standard Research Classifications (ANZSRC) ................................... 10

Fields of Research (FoR) Classification ............................................................................................ 10

Table of Contents ............................................................................................................................... 11

Chapter 1 Thesis overview ................................................................................................................. 16

Foreword ........................................................................................................................................ 16

Introduction .................................................................................................................................... 18

Natural capital and distributive justice........................................................................................... 19

Distributive justice ..................................................................................................................... 21

Thesis aim and research questions ................................................................................................. 25

Caveats ........................................................................................................................................... 29

Thesis structure .............................................................................................................................. 30

Chapter 2 A quantitative systematic review of distributive environmental justice literature; a rich

history, and the need for an enterprising future ................................................................................. 31

Overview ........................................................................................................................................ 31

Abstract .......................................................................................................................................... 33

Keywords ....................................................................................................................................... 33

Introduction .................................................................................................................................... 33

Methods .......................................................................................................................................... 35

Systematic quantitative literature review ................................................................................... 35

Search string development ......................................................................................................... 35

Database creation, inclusion criteria, and article screening ....................................................... 36

12

Typologies and article categorization ........................................................................................ 37

Data analysis and figure generation ........................................................................................... 37

Results ............................................................................................................................................ 38

The scope of EDJ article types, methods and data ..................................................................... 38

The geographic scope of EDJ articles ........................................................................................ 38

The thematic scope of EDJ articles ............................................................................................ 42

Demographics ............................................................................................................................ 43

Human well-being outcomes ..................................................................................................... 43

Natural resource exploitation issues .......................................................................................... 43

Natural hazards .......................................................................................................................... 44

Equitability of outcomes ............................................................................................................ 45

Relationships between data ........................................................................................................ 46

Discussion ...................................................................................................................................... 48

Study Limitations ....................................................................................................................... 51

Conclusion ..................................................................................................................................... 52

Chapter 3 Global mismatch between greenhouse gas emissions and the burden of climate change . 53

Overview ........................................................................................................................................ 53

Abstract .......................................................................................................................................... 55

Introduction .................................................................................................................................... 55

Methods .......................................................................................................................................... 57

Results ............................................................................................................................................ 59

Discussion ...................................................................................................................................... 62

Conclusion ..................................................................................................................................... 63

Author contributions ...................................................................................................................... 64

Competing Financial Interests statement ....................................................................................... 64

Chapter 4 The distribution of industrial pollution across socio-economic groups in Australia; a ten-

year, a fine scale distributive justice analysis .................................................................................... 65

Overview ........................................................................................................................................ 65

Abstract .......................................................................................................................................... 66

Introduction .................................................................................................................................... 66

1.1 A testable principle for distributive justice .......................................................................... 69

Methods .......................................................................................................................................... 69

2.1 Spatial data ........................................................................................................................... 70

2.2 Industrial pollutant emissions data ....................................................................................... 71

13

2.3 Total Toxicity Potential ....................................................................................................... 71

2.4 Socio-demographic Advantage and Disadvantage .............................................................. 72

Results ............................................................................................................................................ 74

3.1 Summary statistics ............................................................................................................... 76

3.2 Proportions of variance explained and statistical significance ............................................ 77

Discussion ...................................................................................................................................... 78

4.1 Study limitations .................................................................................................................. 80

Conclusions .................................................................................................................................... 81

Acknowledgements ........................................................................................................................ 81

Author Contributions ..................................................................................................................... 82

Chapter 5 Large-scale environmental degradation results in inequitable impacts to already

impoverished communities: A case study from the floating villages of Cambodia .......................... 83

Overview ........................................................................................................................................ 83

Abstract .......................................................................................................................................... 85

Keywords ....................................................................................................................................... 85

Introduction .................................................................................................................................... 85

The biophysical dimensions of the Tonle Sap Great Lake ........................................................ 88

Environmental Protection within the Tonle Sap ........................................................................ 90

Threats to the ecosystem and fishery ......................................................................................... 91

Methods .......................................................................................................................................... 92

Prek Toal Biosphere Reserve and surrounding communities .................................................... 92

Data Collection .......................................................................................................................... 94

Data Analysis ............................................................................................................................. 95

Results ............................................................................................................................................ 96

Fishing is critical to local livelihoods ........................................................................................ 96

People believe the fishery is highly vulnerable to environmental change ................................. 97

People believe the fishery is undergoing rapid declines ............................................................ 99

Discussion ...................................................................................................................................... 99

Conclusion ................................................................................................................................... 102

Chapter 6 Thesis summary and conclusion...................................................................................... 104

Modifying distributive justice principles ..................................................................................... 106

Contribution of each research chapter ......................................................................................... 107

Chapter 2: A quantitative systematic review of distributive environmental justice literature; a

rich history, and the need for an enterprising future ................................................................ 107

14

Chapter 3: Global mismatch between greenhouse gas emissions and the burden of climate

change ...................................................................................................................................... 109

Chapter 4: The distribution of industrial pollution across socio-economic groups in Australia; a

ten-year, a fine scale distributive justice analysis .................................................................... 110

Chapter 5: Large-scale environmental degradation results in inequitable impacts to already

impoverished communities: A case study from the floating villages of Cambodia ................ 111

Advantages and limitations of the distributive justice approach, and future research ................. 112

Advantages and limitations ...................................................................................................... 112

Future research ......................................................................................................................... 114

Concluding remarks ................................................................................................................. 114

Bibliography..................................................................................................................................... 116

Supplementary materials .................................................................................................................. 133

Chapter 2 Supplementary materials ............................................................................................. 133

Additional file 1 - List of key EDJ studies............................................................................... 133

Additional file 2 - Search string development log ................................................................... 135

Additional file 3 – Variable typologies .................................................................................... 141

Additional file 3 – Data tables ................................................................................................. 150

Additional file 4 Complete dataset ........................................................................................... 157


S1 Overview ............................................................................................................................. 158

S2 Summary of Lorenz curve .................................................................................................. 158

S3 A summary of per capita GHG emissions and national vulnerability ................................ 158

Supplementary Figure S3 ......................................................................................................... 159

Supplementary table S4 ........................................................................................................... 161


Supplementary file 1 ................................................................................................................ 172




Supplementary material S1 ...................................................................................................... 199




15

16

Chapter 1 Thesis overview

Foreword

First and foremost, I want to acknowledge and thank you, the reader of this work. I very much

appreciate your time spent reading this thesis.

Next, I need to discuss my philosophical positionality as a researcher. As scientists, I believe that it

is critical that we understand and acknowledge the moral and philosophical ideas and ideals that we

bring into our work. These (consciously and unconsciously) pre-conceived ideas shape our research

choices (topic, theoretical framing etc.), and are ultimately the lens through which we interpret our

data and findings – I firmly believe that pretending otherwise isn’t helpful for social or scientific

progress, nor is it somehow more objective. As such, I will clearly state how/why I came to

undertake this work, and how this has shaped my interpretation of how my studies model reality.

However, I want to push back on this interpretation (or perhaps straw person of my own making?)

and re-enforce my earlier claim that I (and other, highly accomplished people e.g. Moon et al.

(2019)) firmly believe that as scientists we can be blinded by the notion of pure objectivity, and

forget the moral positing, and inherent biases of the person(s) behind the research – ourselves. This

is eloquently stated by the authors above:

“The philosophical position of the researcher frames their theoretical perspective (i.e., the ideas,

concepts, and assumptions the researcher brings to their research), influencing the kinds of

questions they ask and how they seek to answer them. These elements inform which methodologies

will best suit the philosophy, how theory and the desired research outcome/s are integrated, and the

rationale for the chosen methods. While philosophy might not always appear to drive research, it

will always implicitly underpin the choices made.” (Moon et al., 2019).

It is imperative that we reflect on, and openly state and acknowledge our philosophical

positionality, when and where possible/practicable. This of course requires a deliberate degree of

self-reflection, which I argue is a critical scientific skill. A wonderfully executed example of this

can be seen in Carney (2016), which openly discusses the author’s positionality as an activist,

scholar, and woman of colour in a “Personal Reflexive Statement”. I strongly believe that this is a

standard all researchers should aspire to. In the absence of such transparency we risk making it easy

17

for others to misinterpret our work, which can in turn lead to erroneous conclusions, critiques, and

perhaps most importantly can feed into the crisis of scientific replication.

My research positionality is mostly as an extremely concerned outsider to the issues I have

investigated. I have the privilege to live in a very wealthy, and relatively fair and equitable country,

and consider it imperative that I use this privilege to work toward a fairer, and more equitable

world. In so far as I understand these terms, I consider myself an intersectional feminist, a

humanitarian, a strong believer in the purposefulness of a universal moral framework for the

equality of rights and fair treatment of all people, regardless of any form of socially constructed or

biologically determined difference, and I am a staunch advocate for improving the lives and

capabilities of disadvantaged people and communities. The driving force behind this thesis is my

belief that historic and ongoing injustices committed by any group against some ‘other’

(particularly in cases of socially constructed and re-enforced identity) are abominable, only serve to

further the nihilistic whims of a few ‘elites’ while compromising the needs of the majority, and

ultimately serve to hinder the just progress of our pretty wonderful species. I don’t think it’s a

stretch of the imagination to claim that in our conflict riven, and nuclear armed world, which is

facing potential ecological collapse, that the true enemies of keeping humanity extant (and happy,

and healthy) are parochialism, and socially constructed division.

By my layperson’s reading, the collapse of the Roman Empire was driven by the kind of internal

division, parochialism, elitism, environmental change, and racism that we’re seeing rise in

occurrence across the world in contemporary times. While I believe that we need to be wary about

holding up a mirror to history as if it is a one-to-one model of the contemporary world, I do think

we can gain worthy insights from our shared history, and the fate of the Romans paints a grim

picture of where unchecked elitism and social division lead.

18

Introduction

It is widely accepted that every living person has intrinsic value (Bradley, 2006). This should make

weighing the value of one person’s life and well-being against another’s impossible. Yet, this is

how our systems of natural resource management and use most often manifest. Throughout human

history, the natural environment has been, and remains, the primary source of material that we

utilise, extract, and exploit, in order to progress our societies, and ultimately improve our collective

capabilities. Be it the ancient (though by no means extinct) practice of hunting wildlife for

sustenance, or the modern utilisation of rare and highly refined minerals in complex technologies,

people have and will always depend on natural capital for personal, social, and economic progress.

However, throughout history, including in our contemporary world, these resources have not been

equally accessible or distributed among individual people, or among our global, multiscalar

communities – neither across time nor space (Gerlagh and Keyzer, 2001). Given that people depend

on natural capital for well-being, these unequal distributions can result in matching unequal

distributions of human capability and well-being. As people, groups, and nations do not exist in a

vacuum and are instead highly interactive, these unequal distributions can result in situations where

the comparatively excessive lot of one group, when contrasted against others, can be reasoned to be

unfair and immoral (Sen, 1995; Sen, 2009).

Within this thesis I empirically identify and describe present day cases of these kinds of divisions

through a distributive justice lens, in the context of natural resource use, with a broad aim of asking

if our national and global institutions are producing fair outcomes. I have not approached this work

with an a priori position that strict equality (i.e. equal distribution of resources regardless of need or

contribution) is either a sufficient or tenable position, neither philosophically nor practically.

Rather, within this thesis I have adopted the position argued by Rawls (1971), and strengthened by

Scanlon (2018) whereby distributive injustices can only be called such if there is sufficient evidence

or moral reasoning to do so. That is to say, inequalities (economic or otherwise) are only unfair, if

they cause harm, particularly if such harm is experienced by any given society’s most vulnerable

people (e.g. economically poor, or ethnically marginalised) (Scanlon, 2018). Furthermore, such

inequalities can be said to be fair (and even desirable), if they result in an improvement in the well-

being of people, again, particularly if the benefit is experienced by the most vulnerable (Crick,

1987; Rawls, 1971).

19

I aim to add to the social justice scholarship by combining theoretical distributive justice

frameworks with the tools of empirical science, in the context of environmental problem solving.

Within this framing, and to meet this aim, I have developed and authored several empirical,

descriptive, and evaluative studies (as per definitions in Robeyns (2017)), which make up the body

of this thesis. Each is descriptive in that I have designed them to provide empirical accounts of

phenomena, and evaluative as I have applied normative (largely fairness) values to evaluate my

empirical descriptions. Explicitly, I have described the distributive patterns of natural resource use

across several cases and evaluated them through a lens of distributive justice.

In this thesis, I present four publications, a systematic literature review and three cases. Each is self-

contained and includes an account of its specific background and importance, the methods I used,

my findings, and my conclusions. As such, this thesis consists of six distinct chapters – two

ancillaries (this introduction and an overall conclusion), and four publications – and is also

supported by several technical and method appendices. First, I introduce the key concepts and

theoretical frameworks that I have used, modified, and developed throughout my doctorate, and the

research questions underpinning this body of work. In following chapters, I present the four

empirical studies, organised in ‘spatial’ order, presenting the most geographically broad studies

first, followed by the smaller scale studies. Finally, I summarise each research chapter’s

contribution to the overall thesis, the advantages and limitations of the distributive justice approach,

and present proposals for future research directions and my concluding remarks.

Natural capital and distributive justice

Much of human history, including current times, can be characterised by the exploitation, trade

politics, and wars of natural capital. These purely social phenomena are generally an artefact of

natural resource scarcity and their naturally and socially unequal distributions. This is exemplified

by humanity’s defining historical moments, such as the momentous, exploitation colonialism of

Europe throughout the 15th to 19th centuries, which resulted in unprecedented levels of imperial

growth, trade, wealth creation, transnational politics, and widespread and interrelated horrors such

as genocide, cultural and sovereign extinction, human slavery, and dispossession – all of which

were largely driven by one group’s desire for another’s natural capital (Findlay and O’Rourke,

2012; Gilmartin, 2009; Piketty, 2014).

20

In our modern world, covetous nationalism is still very much a primary driver of trade, politics, and

war. However, the disastrous levels of human suffering experienced during the globally scaled wars

of the 20th century, and the substantial though less visceral grief caused by ‘softer’ conflicts such as

the Cold War, have led to a relatively more cooperative world (Copeland, 1996). Rather, it is now

relatively more common for international trade and foreign politics to further national interest,

rather than spilled blood and spent munitions. While the modern world is far from utopian, it is

undeniable that in the last century there has been critically important and meaningful progress

beyond simplistic, uncooperative national parochialism. While national self-interest is far from a

spectre of the past, today the majority of the world’s nations and their leaders subscribe (though to

widely varying degrees) to moral frameworks that are relatively much more universal in

recognising social justice, such as the rights of individuals, groups, and nations beyond their own

interests (e.g. quite literally, as signatories of international treaties such as the Universal Declaration

of Human Rights (UN General Assembly, 1948), and more recently the Vienna Declaration and

Programme of Action (OHCHR, 2019)). Such progress has not happened by accident. Rather, in

many areas of the world, the larger part of the last century has often been (and still is) marked by

the extraordinary efforts of political activists, civil rights movements/leaders, philosophers,

academics, politicians, and every-day-people, arguing and fighting for the recognition of their own

rights, or the rights of other people (Davis, 1981; Freischlag and Faria, 2018; Carney, 2016).

However, while the relatively recent progress of social justice and equality has been incredible in its

scale, it has hardly been perfect, nor is it by any means complete, or is a path to global

egalitarianism certain (not even an egalitarianism of so called ‘basic needs’). With even a cursory

glance at the state of the world any observer can see that it is typified by inequality and social

hierarchies that result in deep injustices and unfair distributions of power, wealth, and natural

capital (Piketty, 2014). Globally, income inequality is rising alarmingly, resulting in the repugnant

and absurd situation where some individuals have more wealth than entire nations (Bagchi and

Svejnar, 2015), while hundreds of millions of people are materially deprived of the basic needs

required to realise their full potential (Sumner, 2016; Sen, 1999). Armed conflict is still associated

with natural capital exploitation (Ayres, 2012), and the inequitable nature of greenhouse gas

emissions and the resulting climatic effects see industrialised economies achieving historically high

profit levels, while entire small island nations are sinking under water (Willcox, 2016). And so on

and so forth.

21

As such, it is clear that much work remains to be done if human societies are to address

recognizably harmful forms of inequality. Within this thesis I will use the context of natural capital

distributions to demonstrate how an initial stage of this process is to use empirical research methods

to describe distributions and use theoretical frameworks of distributive fairness to evaluate them. In

doing so, I hope to (humbly) contribute alongside the traditions, research, and extraordinary efforts

of political activists, civil rights movements/leaders, philosophers, academics, politicians, and

every-day-people, arguing and fighting for the recognition of their or others equal rights to a live

they deem worth living (Davis, 1981; Freischlag and Faria, 2018; Carney, 2016).

Distributive justice

It is important to note that an exhaustive discussion of the history and philosophies of justice and

fairness are beyond the scope of my thesis, which instead focuses on the distributive justice theories

that I have utilised within this thesis, which I believe (as per the forward) best suit my analytical

approach and support my broader aims of using research to further the equality of rights and fair

treatment of all people. As such, the core theoretical framework that I use within this thesis, is a

modified form of Rawls’ difference principle - arguably the most influential and impactful theory of

distributive justice (Allingham, 2014; Olsaretti, 2018). Positive economics alone cannot, without

the guidance of normative principles, recommend which policies, structures, or institutions to

pursue. As such there has been meaningful theoretical progress over the last century, for examining

and analysing the fairness of natural resource distributions, and how such analysis can be used to

guide social decision making such as policy.

Distributive justice theories allow for an examination of these frameworks and their resulting

distributive patterns (Lamont and Favor, 2017; Allingham, 2014). Through distributive justice

approaches, we are able to measure and normatively evaluate how the social, environmental, and

economic constructs and contexts of a society (e.g. laws, policy, natural resource access, taxes,

welfare) result in a diverse array of benefit and burden distributions. Furthermore, we can assess

how these distributions differ across social and demographic groups, and (also normatively)

evaluate if these distributive structures affect people’s capabilities and by extension well-being

(Sen, 1995). Additionally, and most importantly for this thesis, distributive justice frameworks

provide clear ethical frameworks through which we can make normative claims of what are just or

fair distributions (Allingham, 2014). The application of distributive justice theories (among others,

e.g. Gross (2008)) provides researchers, governments, policy makers, and activists an outline of

22

moral guidance for how social, environmental, and economic frameworks ought to be shaped, in

order to actually change distributions in a fair manner1.

Due to an ever-increasing utilisation of natural capital globally, there are likewise an ever-

increasing number of environmental justice issues surfacing. Some of these issues can be highly

complex, and require an in-depth examination of associated procedural, institutional, or social

justice. However, for this thesis I have elected to examine cases where understanding each person

(or group’s) distributive due ought to be fairly easy to describe and analyse. This is not to say that

more complex issues are more or less important, however I argue (and will use this work to support

the notion) that, particularly in environmental issues, there are cases where examining, describing

burden and benefit distributions among groups, and reaching conclusions about how fair they are,

can be a relatively simple process. I further argue that understanding and undertaking these (again,

relatively) simple analyses can be a time and resource efficient method for establishing a

distributive justice baseline or ‘snapshot’ of an issue. Such a baseline is intended to be useful for

interested parties (e.g. researchers, policy makers, NGOs, or global intergovernmental

organizations) to understand the landscape of EDJ issues and prioritise actions to better understand

or address them. As such, I have deliberately selected cases of natural resource use where I can use

relatively simplistic forms of distributive justice to understand their fairness. This is not to say I

have selected unimportant issues; I have still elected to examine cases which are critical for the

well-being of groups of people. Furthermore, I have deliberately selected cases at widely diverse

scales, from a global issue between nations, to a local issue affecting a small group of rural villages.

Furthermore, distributive justice theories allow for the explicit justification of ethical choices. Often

the ethics or morality underlying decision making, or assessments of distributions are simply

implied (e.g. It is wrong that air pollution occurs more in places where particular racial groups live).

While this is potentially sufficient, by instead making fairness claims through a lens of distributive

justice we are able to argue in a clear, logical, and rational framework, which I believe strengthens

our position.

In this light, distributive justice is not simply a series of moral theories. Rather, they are best

thought of as a highly practical set of theoretical tools that we can use to describe, examine and

resolve real world problems (Lamont and Favor, 2017). I argue that within a universal moral

1 I must stress that distributive justice outlines guidance, as while it may be possible to do so, it is best that distributive

justice is not applied in a vacuum, and instead as part of a greater equitability process, which includes other critically

important forms of justice such as procedural and relational fairness (as per Gross 2009).

23

framework of equality of rights and fair treatment of all, distributive justice theories are enormously

useful in describing and evaluating scientific data for use in real world dilemmas. Through the

application of these theoretical tools, I have been able to study physical and social phenomena

through a lens of the moral right to fairness in the distribution of benefits and burdens associated

with natural capital among different groups of people. Furthermore, I argue (and illustrate) that

distributive justice principles as applied to scientific enquiry can facilitate practical examination and

provides tools to shape fair policy pertaining to natural capital. It is important however to

acknowledge that distributive justice is neither perfect, nor infallible, or without critics (particularly

by libertarians (e.g. Nozick (1974); Hayek (1976)). However, these limitations are useful as they

highlight the normative nature of distributive justice claims and why we ought to avoid making

universal claims via distributive justice.

As such, and as argued elsewhere (Lamont and Favor, 2017), I contend that the most important

strength of distributive justice theory is its practical applicability, an idea which I will demonstrate

throughout. In particular, I have made extensive use of a modified form of John Rawls’ Difference

Principle (Rawls, 1971; Rawls, 1993; Allingham, 2014; Lamont and Favor, 2017; Rawls, 2001).

Specifically, I utilise and modify part (b) of Rawls’ second principle of justice, which (in full)

states:

“Social and economic inequalities are to satisfy two conditions: (a) They are to be attached to

positions and offices open to all under conditions of fair equality of opportunity; and (b), they are

to be to the greatest benefit of the least advantaged members of society”

Specifically, I have modified this principle to be appropriate to the examination of issues regarding

distributive justice and natural capital, and have removed condition (a) as my work does not

examine liberty or governance. As follows (and as utilised in Chapter 5):

P1: Social and economic inequalities, resulting from the utilisation of natural capital, are to be of

the greatest benefit of the least advantaged members of society.

Furthermore, I also utilise the simplest distributive justice principle - strict egalitarianism. Strict

egalitarianism states (Lamont and Favor, 2017):

Every person should have the same level of material goods (including burdens) and services.

24

In practice, this means uniformity of access to resources for all people, regardless of need or

contribution. It is important to acknowledge the severe limitations of such a rigid and strict theory

of distribution. In a complex world, strict egalitarianism is both problematic and utopian, and as

such is rarely seriously considered in the empirical literature, and I would argue is impractical or

useable in a general manner for distributing (or re-distributing) resources. However, I argue that it

does have an important role, specific to the distribution of harm. If harm is caused by economic

activities (under the assumption that such activities are accepted/tolerated (by law) as they benefit

all of society, e.g. via taxation, economic growth) such as industrial production, any unequal

distribution of this harm, is unacceptable. As such, I have modified this principle as follows (and as

utilised in Chapter 4):

P2: Where harm is accepted as a tolerable by-product of natural capital use for the betterment of

society, exposure levels of harm ought to be equal among all persons.

Furthermore, these two principles may be combined in cases where some harm (or other type of

burden) is produced, for the specific benefit of a single person, or exclusive group, as follows (and

as utilised in Chapter 3):

P3: Where burdens upon any individual or exclusive group are the by-product of the use of natural

capital for the betterment of another individual or exclusive group, burdens are to be accompanied

by (at least) commensurate benefits.

It is critical that I acknowledge that this combined principle is potentially problematic if it were to

be misunderstood or misused, as it does not acknowledge differential power or agency in the choice

of exposure to burdens or benefits of one group due to the activities by another2. The purpose of this

principle is for purely pragmatic applications, for example examining cases where the power

dynamics between groups are realistically unable to be affected (as I will illustrate and discuss in

Chapter 5). As such, it is my intention that it should only ever be applied when assessing existing

distributions, and never for the purposes of policy or project design.

2 While agency is morally paramount, its examination is outside of the scope of this thesis, though I must again stress

that distributive justice approaches ought not be applied to real-world situations in a vacuum, rather their application

should take place alongside other analytical forms examining justice.

25

As these principles make up the key theoretical foundation of this thesis, I will utilise them

throughout to justify the relevance of each research chapter and associated research questions, to the

overall thesis aims. For brevity they will be referred to henceforth as “modified distributive justice

principle(s) P1, P2, or P3”.

In summary, the main purpose of this thesis is to provide several cases (studies) that highlight how

distributive justice frameworks can be applied to empirical research in order to strengthen

normative arguments around fairness. The ever-increasing utilisation of natural capital globally is

giving rise to similarly increasing environmental justice issues. Many of these issues are highly

complex, and require an in-depth examination of associated procedural, institutional, and relational

fairness. However, for this thesis I have elected to examine cases where understanding each person

(or group’s) distributive due ought to be fairly easy to describe and analyse. This is not to say that

more complex issues are more or less important, however I argue that, particularly in environmental

issues, there are cases where examining, describing burden and benefit distributions among groups,

and reaching conclusions about how fair they are, can be a relatively simple process, and an

important first step to their resolution. By describing and evaluating cases through these relatively

simple distributive justice approaches we gain a time and resource efficient method for establishing

a fairness baseline or ‘snapshot’ of an issue. Such a baseline is likely to be highly useful for those

wanting to broadly understand the landscape of distributive justice issues, and to develop and

prioritise follow actions to better understand or address them. As such, I have deliberately selected

cases of natural resource use where I can use relatively simplistic forms of distributive justice to

understand their fairness. This is not to say I have selected unimportant issues; I have still elected to

examine cases which are critical for the well-being of groups of people. It is my hope that the

studies I present will both provide value in and of themselves, while forming part of a broader

narrative around the importance of distributive justice theories in empirical research.

Thesis aim and research questions

This thesis pursues the overarching research question of:

Is it possible to re-frame distributive justice theories in ways that are more amenable to (a diversity

of) empirical research methods?

In order to answer this question as systematically as a doctoral thesis permits, I have developed four

broad sub-questions. Each of these sub-questions specifically utilises one or more components of

26

my modified distributive justice principles and is answered via four empirical examinations

(presented here as Chapters 2 to 5).

This is divided into the following sub-questions, which are variously utilised in each research

chapter:

1. Sub-question 1: Can distributive justice principles be re-framed and applied in a manner

that allows for the development and application of a systematic review of the natural

resources literature?

2. Sub-question 2: Can the benefit and burden distributions between people or groups,

resulting from natural resource exploitation, be empirically examined using modified

distributive justice principles?

3. Sub-question 3: Are distributive justice principles able to be modified to allow for a

quantitative empirical examination of the distribution of harms resulting from natural

resource exploitation?

4. Sub-question 4: Are distributive justice principles able to be modified to allow for a

qualitative empirical examination of the distribution of harms resulting from natural


As my modified principles have the potential for a diverse array of applications, they demanded a

diversity of empirical responses at this initial exploratory stage, to highlight their flexibility. As

such, the research chapters herein are diverse in their geographic scale, empirical methods, and

scale of outcomes. In doing so, it was my intention to highlight, indicate, and paint as broad a

picture of the potential diversity of empirical applications for my modified distributive justice

principles (as time and resources would allow).

Chapter specific research questions

The following is a breakdown of each of the research chapters’ research questions. For clarity and

due to each chapter being written to appear as a standalone peer-reviewed manuscript, I have opted

to delineate between thesis and study specific questions. As such, this section is structured by

specifying 1) the thesis research question(s), which explicitly link each to the overall thesis question

and the relevance to my modified distributive justice principles, 2) a list of which modified

principles I have applied, 3) the study specific research questions and sub-questions, which treat

each chapter as a self-contained study such as they might (or do in some cases) appear in a

27

standalone peer-reviewed manuscript, and 4) a brief overview of how these questions are addressed

in the relevant chapter.

Chapter 2: A quantitative systematic review of distributive environmental justice literature; a

rich history, and the need for an enterprising future

Thesis sub-question 1: Can distributive justice principles be re-framed and applied in a manner

that allows for the development and application of a systematic review of the relevant literature?

Modified distributive justice principles utilised: P1, P2, and P3.

Study specific research questions: What is the depth and breadth of the distributive justice and

natural capital literature?

Study specific sub-questions:

1) Where are the research institutions of the authors of this research located;

2) What has been the geographic focus of the literature;

3) What has been the demographic, environmental, and human well-being scope of the literature;

4) What types of publications and study methods have been used; and

5) What are the gaps and areas for future focus in the literature.

These questions are answered in Chapter 2 by applying all three modified distributive justice

principles (P1, P2, and P3) as a framework to develop and undertake a quantitative systematic

review of the literature. This approach adheres to the method developed by Pickering et al. (2014),

and summarises the literature using a rigorous and robust, replicable review method.

Chapter 3: Global mismatch between greenhouse gas emissions and the burden of climate

change

Thesis sub-question 2: Can the benefit and burden distributions between people or groups,

resulting from natural resource exploitation, be empirically examined using modified distributive

justice principles?

Study specific research question: How are greenhouse gas emissions and the burdens of climate

change distributed globally?

28


1) What is the extent of global climate inequity; and

2) How is this expected to change in the near future.

In Chapter 3 these questions are answered in a study wherein I utilised my modified distributive

justice principles P1, and P3 to develop and undertake the first global level study to analyse the

distribution of burdens and benefits between nations resulting from greenhouse gas emissions

(GHGs).

Chapter 4: The distribution of industrial pollution across socio-economic groups in Australia;

a ten-year, a fine scale distributive justice analysis

Thesis sub-question 3: Are distributive justice principles able to be modified to allow for a

quantitative empirical examination of the distribution of harms resulting from natural resource

exploitation?

Study specific research question: Are there social inequities in the distribution of exposure to

harmful industrial pollution across Australia?


1) Have these distributions changed over the last 10 years; and

2) If any, which social or economic groups are more likely to be exposed to industrial pollution.

These questions were developed and answered in Chapter 4 where I used my second modified

distributive justice principle (P2), to develop and undertake a study where I quantitatively modelled

the relationships between industrial pollution exposure and socio-demographic distributions across

the nation of Australia.


impoverished communities: A case study from the floating villages of Cambodia

Thesis sub-question 4: Thesis sub-question 3: Are distributive justice principles able to be modified

to allow for a qualitative empirical examination of the distribution of harms resulting from natural


29

Study specific research question: How are the fishing communities of the Tonle Sap floating

villages experiencing major environmental changes?


1) How are floating village communities experiencing a change in fisheries; and

2) How are floating village communities responding to these changes.

Finally, in Chapter 5 I show how I used my second modified distributive justice principle (P2) to

develop and answer these research questions through a qualitative method which captures the

harmful lived experiences of Cambodian subsistence fisher people in the context of an environment

which is rapidly changing due to interventions beyond their control.

Caveats

A note on terminology

The modified principles described above, and associated terminology, were developed over the

early stages of my PhD, but not completely before publications had been finalised and published.

As such, there are some cases of varying terminology applied throughout. To make the connection

between principle and study clear, I have included a short overview before each research chapter,

reiterated the conclusion, and described how each relates to the underlying theory and overall

research questions.

A note on GDP

Throughout I have made statements regarding GDP which may be read as problematic, such as in

Chapter 2 where I state: “…economic statistics (such as GDP) can be used as a coarse analogue for

human well-being…”. Such statements are not intended to be normative claims about GDP as a

good, worthwhile, non-problematic analogue for well-being, or that there are not far more

appropriate quantitative measures of human well-being such as the Human Development Index, or

Genuine Progress Indicator (as I qualify in the text following the above quote “…, [GDP] offer[s]

only limited insight into important dimensions such as capabilities” (Sen, 2009)). Rather their use is

intended to reflect the reality that economic growth reigns supreme in the minds and policies of the

majority of the world’s decision makers, such as national governments, as such GDP is the most

commonly used statistic (to assess well-being) by those with power (Ward et al., 2016). As such, I

wholly support the evidence based research that shows how GDP is problematic, but is also

appropriate as it speaks in the “language” of policy makers (Kenny et al., 2019; Sen, 2009; Ward et

30

al., 2016). So, while I strongly believe that other quantitative metrics of well-being are important,

they are orthogonal to my purpose here.

Thesis structure

I have structured this thesis as follows:

1. Foreword material including this introductory chapter.

2. Four chapters based on the four distinct studies I have developed and undertaken as part of

this thesis. Each contains:

a. A brief overview discussing how they relate to the overall distributive justice

theories underpinning this thesis, the research questions they answer, and highlights.

b. The actual study, each presented as a peer-reviewed publication. Unless otherwise

stated they appear either exactly as published for currently published/under review

works, or how they are expected to appear for forthcoming works.

3. Concluding material.

4. A bibliography which aggregates all works cited throughout this thesis, including for each

individual study, and all ancillary texts.

5. Supplementary material for each chapter. Unless otherwise stated they appear either exactly

as published for currently published/under review works, or how they are expected to appear

for forthcoming works.

31

Chapter 2 A quantitative systematic review of distributive

environmental justice literature; a rich history, and the need for an

enterprising future

Full citation: Althor G and Witt B. (2019) A quantitative systematic review of distributive

environmental justice literature: a rich history and the need for an enterprising future. Journal of

Environmental Studies and Sciences. Online.

Overview

This study is currently under review. It was designed to use the quantitative systematic review

method developed by Pickering et al. (2014). The overall aim of this study is to capture the depth

and breadth of literature which relates to natural capital and distributive justice.

Research question: Does existing literature on distributive justice have evident biases or limitations,

and what is the opportunity for use of empirical evidence to address these?

Sub-questions:

1) What is the depth and breadth of the distributive justice and natural capital literature?






The resulting ‘map’ of the literature was used to identify key knowledge gaps, which were further

used to develop and select specific topic areas for Chapters 3, 4, and 5. All of the theoretical

principles underlying this thesis P1, P2, and P3 (as described in Chapter 1) were used as a guide to

decide on the relevance of each individual study included in this systematic review.

Highlights:

• First systematic review to specifically seek to understand the depth and breadth of the

natural capital and distributive justice literature.

32

• Rigorously applied systematic review method which reduced over 46,000 potential studies

down to a final list of 354.

• Provides a key resource for easily finding all literature under particular sub-topics and

identifying topical knowledge gaps in the literature.

33

The following text appears as largely as published, with minor amendments based on thesis

examiner feedback.

Abstract

Environmental distributive justice contextually assesses social equities in relation to natural

resources. While there is a rich literature on environmental distributive justice there have been few

assessments quantifying the biases and scopes of this literature. We conduct a systematic review of

the literature. We find several key biases and summarize the breadth of subjects that have been

studied or discussed. We find a very distinct overlap between authorship nationality and study

location, which is concerning, as some of the world’s most polluted and inequitable societies are

least represented in authorship. Additionally, we find a dominance of quantitative studies. These

results are important for understanding both where future research efforts in this area could best be

directed, and how the literature could be enriched by diversified approaches. Improving

environmental justice studies is critical and important for many people across our global society,

which is increasingly shaped by widespread natural resource depletion.

Keywords

Distributive justice, natural resource management, environmental exploitation, burdens, benefits,

climate change, pollution.

Introduction

Throughout human history the use of natural resources has been the key means by which societies

have created social and economic progress. By exploiting the natural environment, societies are

able to make the places they live more habitable for humankind, and produce goods and services to

improve their well-being. This is (and has been) the case for all societies, from the earliest of

civilizations, whose existence was only possible due to the innovation of agriculture, to

contemporary consumption-based societies, which exploit natural capital globally for the

production of primary and secondary goods (Pearce and Turner, 1990).

While societies may benefit from exploiting natural capital, these processes rarely come without a

social cost. These costs take various forms and can result in a diverse range of human well-being

burdens. Examples of these issues are diverse (Brulle and Pellow, 2006), and can include a range of

issues from the loss of amenity and recreation due to water pollution in Australian lakes (Crase and

Gillespie, 2008), the ruin of local livelihoods due to deforestation in Pakistan (Shehzad et al., 2014),

to increases in cardiovascular disease due to exposure to air pollution in the USA (Brunekreef and

34

Hoffmann, 2016). The key message from this literature is: societies exploit natural capital to

produce a diverse range of well-being benefits, however in doing so they also generate a diverse

range of social costs. While analysing and understanding this message is important for improving

human well-being, it does not elaborate on two critical and interlinked social issues: 1) How are the

costs and benefits, which result from the non-sustainable exploitation of natural capital, distributed

across social groups? and vitally 2) How fair are these distributions?

These questions are important as they highlight the potential imbalance between the benefits and

burdens of environmental exploitation among different social groups. Recognising and addressing

such imbalances is imperative, as they can lead to unjust outcomes for particular groups. Within the

broader social justice literature, issues of cost and benefit imbalances are commonly referred to as

issues of ‘distributive justice’ (Allingham, 2014). Understanding and describing distributive justice

in broad social issues has long been the preoccupation of many of the world’s preeminent political

philosophers (such as Rawls, Sen, Dworkin, and Nozick (as per Sen, 2009)). There has also been a

relatively recent increase in scientific research aiming to empirically quantify a diverse range of

distributive justice issues within and across societies (e.g. Essoka (2010); Rodriguez-Lara (2013)).

Furthermore, there has been a marked increase in research which specifically investigates the

distributive justice dimensions of the exploitation of natural capital (e.g. Dobbie and Green (2015);

Kyne and Bolin (2016)). Environmental distributive justice (EDJ) research is increasingly critical as

natural resource exploitation becomes ever more widespread, and as resultant benefits and burdens

progressively affect more people (Althor et al., 2016a). The practical benefit derived from EDJ

research is to highlight issues where particular populations may be unfairly burdened by

environmental exploitation. As such, EDJ studies can be used as a guide for decision makers. While

there have been significant and meaningful efforts to synthesize the geographic distribution of

environmental justice issues (e.g. The EJ Atlas (Temper et al., 2015)), and rigorously examined and

characterized in narrative style literature reviews (e.g. Brulle and Pellow (2006); Mohai et al.

(2009); Mohai and Saha (2015); Banzhaf et al. (2019)) the literature has not been systematically

quantified until now. As such this study significantly compliments existing works by providing a

unique, systematic quantification of the literature.

Here we utilise the systematic quantitative literature review method (Pickering et al., 2014), to

interrogate the peer-reviewed literature that regards distributive justice and the non-sustainable

exploitation of natural capital. This allows us to map the available literature both conceptually and

diagrammatically, and to answer several important questions: 1) Where are the research institutions

of the authors of this research located; 2) What has been the geographic focus of the literature; 3)

What has been the demographic, environmental, and human well-being scope of the literature; and

35

4) What types of publications and study methods have been used. We use the data and analyses

generated by answering these questions to not only quantify the existing literature, but also to

identify potential research gaps and key trends. Our study provides insights into the nature of the

EDJ literature and ensure that future research in this area targets these gaps.

Methods

Systematic quantitative literature review

We undertook a systematic quantitative literature review using a well established method (see

Griffith University (2019) and Pickering (2014) for an overview of the method and a list of

publications utilising it). This method was developed in order to systematically search for and

categorize the literature on a particular academic topic. The purpose of this process is to quantify

the scope of the literature, and identify historic strengths or biases. Such information can be used to

identify potential knowledge gaps, and inform future research (CEE, 2013).

While this method is excellent for quantitatively examining the characteristics of any given topic

area, it is worth noting that it has limitations. Most importantly and by design, the method does not

wholly synthesize the literature, neither in a narrative nor analytical manner. As such theoretical

syntheses and meta-analyses often found in traditional systematic reviews are beyond the scope of

quantitative reviews.

Search string development

The purpose of our study was to find and quantify the peer-reviewed literature specifically relating

to EDJ. As such, we did not include articles which exclusively examined other aspects of

environmental justice (such as procedural justice), though multidimensional articles including EDJ

dimensions were retained.

To conduct our search do so we developed a Boolean style search string using the online academic

database Web of Knowledge (WOS) (Thomson Reuters, 2017). Google Scholar was deliberately

avoided, as it has been found to be inappropriate as a principal search system (Gusenbauer and

Haddaway, 2019). Further, due to time and resource constraints, we decisively chose to limit the

focus and targets of our searches (both for search string development, and final searches) to the

most relevant, and high-quality peer reviewed literature. Explicitly, we limited searches to peer-

reviewed articles written in the English language, within research areas relevant to EDJ, and

accessible from in three key academic literature databases, WOS, SCOPUS, and Proquest. These

databases were selected as they cover a large proportion of the academic literature.

36

Our search string, and list of relevant research areas, were developed over an iterative five-step

process whereby we: 1) developed an initial search string based on key EDJ, distributive justice,

and environmental exploitation articles (Bournay, 2017; McKinnon et al., 2016; Shortt et al., 2012;

Heyward, 2007; Tornblom, 1977; Cook and Hegtvedt, 1983); 2) tested search string in WOS by

topic (denoted “TS” in WOS); 3) sorted results by relevance (a WOS parameter); 4) reviewed the

first 100 titles, screening for any EDJ terms missing from the initial search string; 5) updated the

search string with new terms, and repeated steps 2, 3 and 4, until the first 100 titles all included at

least one relevant EDJ term, finalizing the search string (for a log of search string development see

supplementary materials). Research areas were determined from these 100 articles, using the WOS

results analysis tools (Thomson Reuters, 2017). The final search string was run on the three key

academic databases, WOS, SCOPUS, and Proquest. We modified the search string to fit each

database’s particular Boolean parameters, and research areas.

The final search string produced a very large number of results (n = 46,860), and given the limited

resources available for this project, further refinements were required. In order to reduce the number

of results, we undertook a coarse screening of titles by only including in our search results those

papers which also included a distributive justice term in the title (denoted “TI” in WOS). While this

process reduced results to a manageable number (n = 2008), it likely resulted in an unknown

number of EDJ articles being missed. However, we attempted to address this by conducting a final

reliability check to ensure results were appropriate and accurate, by verifying that the final results

included several key EDJ studies (n =10) identified prior to any searches through preliminary

readings of the literature (for list see supplementary materials). As such, our review presents a

systematic insight into the EDJ literature.

Database creation, inclusion criteria, and article screening

We exported the final list of articles as citations to Endnote reference manager software (Thomson

Reuters, 2015), which we then exported to a Microsoft Excel database. We used this database to

conduct all article screening. Before commencing article screening we first developed an exhaustive

list of 19 inclusion and exclusion criteria which were used throughout the screening process to

identify which articles should be kept, and which discarded. The overall ‘theme’ of the inclusion

criteria was to: Identify peer-reviewed articles which explicitly relate to the distributive justice

dimensions of the non-sustainable exploitation of natural capital (for full list of inclusion and

exclusion criteria see supplementary materials). Examples of these criteria include: must relate to an

environmental distributive justice issue, must relate to the non-sustainable exploitation of natural

capital, must not be related to nature conservation/sustainable development, must not be related to

procedural justice etc. Article screening was conducted according to our inclusion criteria. We

37

carried out the literature screening over four phases: the (aforementioned) coarse (keyword) title

screening, fine title screening (screened against inclusion and exclusion criteria), abstract screening,

and full text screening. At each phase of screening we discarded any articles not relevant according

to the inclusion and exclusion criteria. Throughout the screening process we also discarded any

articles which had missing information (such as title or author), or did not have full text available

online. Each stage reduced the number of total articles (fine scale title screening, n = 1107; abstract

screening, n = 492; full text screening, n = 354).

Typologies and article categorization

We read and categorized the contents of every article passed to the full text screening phase.

Articles were categorized using a list of key variables (n = 91). Variables included bibliographic

information (e.g. author names, and first author institute location), the scope of articles (e.g. article

and data types, geography, and themes), and where mentioned, the equitability of outcomes (see

supplementary materials for complete dataset, including list of variables).

To categorize the thematic scope of each article in the data extraction phase, we developed a series

of thematic typologies. Each typology consisted of a series of variables under four overarching

categories: social demographics, human well-being outcomes, natural resource exploitation issue,

and natural hazards (for complete typologies, see supplementary materials). We developed each

typology by modifying existing typologies (Bournay, 2017; McKinnon et al., 2016; Shortt et al.,

2012; Heyward, 2007; Tornblom, 1977; Cook and Hegtvedt, 1983). Each typology was updated as

appropriate during the title and abstract screening phases. For example, we used the typology

developed by (McKinnon et al., 2016) as a starting point for our social demographics typology.

However, this did not include a variable for LGBTQI+ people, who we found referred to during the

screening phase. Therefore, we added this group as a social demographic variable.

Data analysis and figure generation

All data analysis and all figure generation was conducted using R statistical software (R Core

Development Team, 2015). Linear models were used to assess: the increase in EDJ articles

published over time; and relationships between first author institute location and variable selection

(Burkholder and Edler, 2014). A Pearson’s Chi-squared test was used to assess relationships

between first author institute location and study location (Lancaster, 1969). All other analyses are

proportions.

38

Results

The scope of EDJ article types, methods and data

Overall, we found articles relating to EDJ published between 1986-2017, and that the number of

publications per year has significantly increased over time (figure 1a). Our assessment of the EDJ

literature shows several article types (figure 1b), the use of qualitative, quantitative, and mixed data

types (figure 1c), and a diverse range of methods (figure 1d). The majority of published EDJ articles

are empirical research (n = 205, 57.91%), but there are also many essays (n = 70, 19.77%),

literature reviews (n = 54, 15.25%), commentaries (n = 14, 3.95%), and discussion articles (n = 11,

3.11%). When assessing data types, we found that a clear majority of articles used quantitative

methodologies such as desktop analyses of secondary data (n = 181, 51.13%). While we also

identified studies which utilised qualitative (n = 70, 19.77%), and mixed data methods (n = 11,

3.1%) they are less common. We discovered a high diversity of methods used in EDJ articles, but a

few dominated (for table of full methods counts see supplementary material). The most common

methods are desktop studies using secondary data (n = 73, 20.62%), GIS based spatial analysis (n =

67, 18.93%), and document analyses such as literature reviews (n = 48, 13.56%). Philosophical

argument pieces were also common (n = 54, 15.25%), which almost exclusively related to climate

change.

The geographic scope of EDJ articles

We assessed several geographical scales for EDJ literature including: the geographic scale of the

article, the location of the first author’s institute, and for research articles the primary study

location. This study is limited to English language papers only, which may have introduced some

biases.

We found a diverse use of these geographic scales used across articles (figure 2), including local (n

= 87, 24.58%), global (n = 83, 23.45%), national (n = 57, 16.1%), no scale (n = 54, 15.25%), sub-

national (n = 45, 12.7%), regional (n = 17, 4.8%), and multi-city (n = 11, 3.11%) scales. While the

majority of papers applied to the global scale, it is worth nothing that we separately classified

articles that were purely philosophical or theoretical in nature as no scale, as they could be

applicable to any human population.

Our assessment of the geographic location of the first author institutes producing EDJ literature also

revealed a relatively small number of countries (figure 3a). Furthermore, even within the countries

producing the literature, there are only a few producing the majority of the work. The United States

of America (USA), hosts a clear majority of first authors (n = 149, 42.1%), and most others come

from highly developed nations such as the United Kingdom (n = 45, 12.7%), Australia (n = 24,

39

6.78%), and Canada (n=20, 5.65%). Even when grouped, all other nations are proportionally

underrepresented (n = 92, 25.99%; for table of all nations see supplementary material). Africa,

Central and South America, and much of Asia are least represented. Furthermore, our assessment of

articles which focused on a specific geographic area of interest shows that here too the majority are

located in a small number of highly developed countries (figure 3b). This is in clear contrast to

geographic distributions shown elsewhere (such as in the EJ Atlas (Temper et al., 2015)).

Figure 1 Total counts of articles, percent of types of articles, percent of data types, and

percent of methods used in environmental distributive justice articles, published from 1986 to

2017. (a) The total papers published per year. There has been a significant increase over time (β=

1.31, p < 0.00, r2 = 0.68), which has accelerated in recent years. 2017 is not shown, as it is an

40

incomplete (year) record. (b) The total counts for each type of published article, for all years. (c)

The total counts of articles by data type. (d) The total counts for the methods used in the literature,

for all years - method counts are shown in two ways, the first (yellow) shows counts of non-climate

change articles only, and the second (purple) shows counts of all articles (including those related to

climate change - studies in yellow)).

Note that we have aggregated sub-national areas of interest to the national level (for complete data

set, including sub-national locations see supplementary material). Similarly, to the first author

institute location, we discovered that areas of interest are largely concentrated in only a few

countries. Here too the USA (n = 89, 25.14%), the United Kingdom (n = 18, 5.08%), Australia (n =

11, 3.11%), and Canada (n = 12, 3.39%) are the most represented nations. Areas of interest in all

other countries combined accounted for a relatively small number of articles, even when grouped (n

= 69, 13.45%). Again Africa, Central and South America, and much of Asia are least represented.

41

Figure 1 The geographic scale of environmental distributive justice articles from 1986 to 2017.

The geographic scale (x axis) is sorted from most to least common. The percent of total articles (y

axis) is for all articles, for all years. Note that no scale refers to (mostly theoretical) papers with no

specific geography which could apply to any human population.

42

Figure 2 World maps showing the geographic spread of environmental distributive justice

articles, from 1986 to 2017. (a) The percent of total publications aggregated at the national level,

by the location of first author institute (n = 224). Most institutes are in a small number of highly

developed nations, with a clear majority located in the USA. (b) Geographic areas of interest as

represented by the percent of total articles aggregated at the national level, by location ((n = 224),

no scale studies (n = 130) not shown). Most non-global articles regard a small number of highly

developed nations, with a majority in the USA. Scales show percent of density of articles with

highest in purple (darkest) to lowest in yellow (lightest).

The thematic scope of EDJ articles

We used theme and variable typologies to organise extracted data for four key topical areas used in

the EDJ literature: social demographics; human well-being outcomes; natural resource exploitation

issues, and; natural hazards.

43

Demographics

We found a diverse spread of demographics either discussed in articles, or used as variables in

empirical studies (figure 4a). Economic status (n = 147, 41.52%), geographic location (n = 112,

31.64%), and race or ethnicity (n = 96, 27.19%) were the most commonly used demographic

groups. Whereas, religious (n = 1, 0.28%) and LGBTQ+ (n = 1, 0.28%) groups were comparatively

underrepresented (for complete demographic data see supplementary material). Importantly, we

also recorded several articles under the ‘other’ category as they were inadequately captured by our

social demographics typologies. The majority of these articles used unique demographic groups e.g.

regional grouping of states (Morello-Frosch and Jesdale, 2006), percentage of people employed in

manufacturing (Su et al., 2010), and percentage of single parents (Padilla et al., 2014). The only

relatively common demographic variable recorded under ‘Other’, but not captured in our typology,

was population density.

Human well-being outcomes

For human well-being outcomes, we also found a diversity of topics (figure 4b). However, there

was an overwhelming majority of articles in which health or mortality were discussed or measured

(n = 214, 60.45%). Living standards were also relatively common (n = 78, 22.03%), but equally so

were articles with no human well-being outcomes specified (n = 78, 22.03%). Less commonly

explored outcomes included education (n = 4, 1.13%), and mental health (n = 5, 1.41%). For table

with complete human well-being theme counts see supplementary material. Key examples include

the examinations of toxic substance exposure (Verde et al., 2016; Almaskut et al., 2012), disease

incidence (Szabo et al., 2016), and mental health incidence (Maantay and Maroko, 2015).

Natural resource exploitation issues

We also discovered a range of natural resource issues, but here too a few topics dominated (figure

4c). The majority of articles discussed or measured some form of climate change issue (n = 152,

42.94%) (e.g. Caney (2014)), or air pollution (n = 130, 36.72%) (e.g. Almaskut et al. (2012)), but

few discussed or measured both climate change and air pollution (n = 6, 1.69%) (e.g. Perez and

Egan (2016)). Ozone depletion (n = 2, 0.56%), hunting (n = 2, 0.56%), and hydrological

modification (n = 4, 1.13) are the least commonly explored topics. For table with complete natural

resource exploitation theme counts see supplementary material.

44

Natural hazards

As with the other topical areas, our data extraction also revealed a spread of natural hazard topics,

but a majority for only a few topics (figure 4d). Here, the use of air quality as a discussion point or

measurement dominates (n = 143, 40.39%) (e.g. Almaskut et al. (2012)). Air quality aside, it is

common for articles to not specify any environmental hazards (n = 83, 23.45%). Other common

hazards include water quality (n = 59, 16.67%) (e.g. Babidge (2016)), natural disaster intensity (n =

56, 15.82%) and disease (n = 34, 9.6%) (e.g. Austin and McKinney (2016)). Least common hazards

explored are over grazing (n = 1, 0.28%) and extreme fires (n = 2, 0.56%). It is worth nothing that

most natural hazards can either be attributed to or be intensified by climate change (Lewis, 2017).

For table with complete natural hazard theme counts see supplementary material.

45

Figure 4 Percent of total articles by topical areas, divided into sub-categories, for all years.

Includes articles where the sub-category was either discussed or measured. Note that articles may

discuss or measure multiple sub-categories. (a) Percent of total articles by social demographic sub-

categories (b) Percent of total articles by environmental exploitation issue sub-categories (c) Percent

of total articles by human well-being sub-categories (d) Percent of total articles by environmental

hazard sub-categories.

Equitability of outcomes

We also assessed and recorded the equitability of human well-being outcomes for each paper

(figure 5). We found that the majority of articles stated that well-being outcomes were inequitable

46

(n = 184, 51.98%) (e.g. Sun et al. (2017)), meaning that in these cases the authors found one or

more social groups were unfairly burdened by some form of natural resource exploitation. A large

proportion of articles either did not specify an outcome or did not discuss equitability at all (n =

150, 42.37%). This was the case for all non-empirical studies, where inequities may have been

discussed, but were not measured. Very few articles stated that the outcomes were

unclear/undetermined (n = 12, 3.4%), or equitable (n = 8, 2.26%) (e.g. Rivas et al. (2017)).

Figure 5 Percent of total articles by equitability of outcomes. The outcome(s) (y axis) are sorted

from least to most common. The percent of total articles (x axis) is for all articles, for all years.

Relationships between data

In addition to quantifying the scope of the literature, we also identified several key relationships.

For instance, an assessment of the relationship between first author institute location and study

location revealed a very strong dependence (Pearson’s Chi-squared test with Monte Carlo simulated

47

p-value (2000 replicates), χ2 = 3033, p < 0.001). As such, authors are primarily writing about EDJ

issues in countries in which they reside or work, or they focus on global issues. We also found

several statistically significant relationships between first author institute location and the inclusion

of particular social demographics, however the power of most of the tests was too weak due to

small per country sample sizes (for 0.8 power n was approximately > 300 for each demographic

variable, as such only the USA had sufficient samples). As such, the location of an institute

potentially has some effect on which demographic groups are selected during study design, however

we do not have enough data to quantify the strength of this relationship (for complete results of

linear modelling see supplementary materials).

We also identified a relationship between the environmental issue of interest and whether an article

reported on an empirical study (measurement or model of data), or simply discussed the issue (e.g.

philosophical essays, discussion, or comment pieces). In particular, authors of climate change EDJ

literature generate a large amount of discussion articles (n = 106, 29.94%), as opposed to studies

which measure or model data (n = 30, 8.47%). Furthermore, few climate change articles mention

human well-being outcomes (n = 59, 16.67%), and even fewer measure or model a climate change

effect and a human well-being outcome (n = 14, 3.95%; figure 6a). As such, the majority of the

climate change EDJ literature does not involve any empirical measurement or modelling. Similarly,

almost all air pollution articles make some mention of human health outcomes (n = 121, 34.18%).

However, we identified an uneven distribution between air pollution articles which only discuss

health outcomes (n = 83, 23.45%), and those which empirically measured or modelled health data

(n = 38, 10.73%; figure 6b).

48

Figure 3 Total counts of key environmental exploitation issues and the disparity between the

measurement and discussion of issues. (a) Total counts of climate change articles divided by

climate change discussion or measurement, and the inclusion (or not) of human well-being

assessments. (b) Total counts of air pollution articles divided by the discussion or measurement of

human health outcomes.

Discussion

Overall, we found that while there is diversity in the EDJ literature, the majority is narrow in scope.

Our results show that the literature has, and continues to grow significantly over time at a rate well

ahead of the general increase in academic literature, which highlights the increasing importance of

this field globally. Methodologically, we found a diversity of research types, though a few method

categories dominate the literature. In addition, we found that for climate change literature

specifically, there are a much larger number of papers putting forth ethical frameworks or

arguments (e.g. how emissions ought to be calculated), than empirical studies. Furthermore, we

found that the majority of EDJ research applies quantitative methods, such as desktop analyses of

secondary data, and the use of spatial methods such as GIS. There are far fewer studies draw on

qualitative methods, such as interviews and focus groups. Even fewer are mixed methods studies.

While mixed methods research is more time and resource expensive to conduct, it has been

described as the most effective means of producing validated, complex, and rich data (Borkan,

2004; Creswell, 2014; Clark and Ivankova, 2015). As such, we argue that increased application of

mixed methods is a key opportunity area for developing robust EDJ studies.

49

Our results show a large amount of the literature that examines issues at either the global or local

geographic scales. While examining issues at these scales is important for addressing EDJ issues,

we argue that there is also much to be done at other scales, such as at the national and or sub-

national levels. Examining EDJ issues at these scales is becoming increasingly important as

governments and scientists alike attempt more complex and nuanced approaches to solving issues,

such as ‘bottom-up’ action for mitigating greenhouse gas emissions, or reducing air pollution (Sabel

and Victor, 2017). However, the most critical issue we found in the geographic scope in the EDJ

literature, both in terms of study location and author location, is the narrow focus on a few, highly

developed nations. Interestingly, this highly contrasts with other resources showing much broader

geographic distributions of EJ issues (Temper et al., 2015). Combined with the close association

between first author institute location and area of interest, we posit that this represents a significant

geographic bias in (English language) academic literature. There is much evidence to suggest that

nations which are relatively least developed are far more likely to be socially inequitable, and to

have poor environmental exploitation standards (Adger, 2002), which in turn can lead to human

health inequalities (Drabo, 2011).

While there are certainly good reasons for a lack of academic EDJ research in developing countries

such as poor data availability, given the significant inequities observed between nations, the need

for EDJ research in developing countries is vital. It will therefore be a challenge for the field to

ensure that researchers can undertake projects in countries other than that of their research

institution, guided by attentiveness to the greatest need for EDJ research. In such situations, it will

be critical that Western researchers focus on building the capacity of researchers in those countries

thereby developing more diversity in EDJ authorship, and contributing to the broader movement of

decolonialising academia. It is worth highlighting that issues around study location and author

diversity are by no means unique to this field. Several papers have been published highlighting

similar issues across other disciplines (e.g. development studies (Cummings and Hoebink, 2017),

and Africa-based studies (Briggs and Weathers, 2016)).

While we found a diverse spread of social demographic examination across the literature, the

majority of articles focus on the economic differences between groups. While economic statistics

(such as GDP) can be used as a coarse analogue for human well-being, they offer only limited

insight into important dimensions such as capabilities (Sen, 2009). As such, it is recommended that

economic status is used with caution and only where data allowing more complex metrics, such as

the human development index (UNDP, 2017), are absent. Furthermore, we found a lack of articles

discussing minority religious groups, and people who identify as LGBTQ+, both of which are

recognized as potentially vulnerable populations (Fineman, 2014; Grim and Finke, 2007).

50

Interestingly, we also found that primary author institute location and certain demographics

discussed in an article are potentially related, which is potentially due to the importance of cultural

context. For example, European studies tended to focus on economic status race very commonly.

However, the reasoning for demographic selection is often not clear or explicitly stated in the study

design. We suggest that (where appropriate to study design) authors utilise a more universal set of

social demographics, such as the PROGRESS framework used in the synthesis of health inequity

studies (O’Neill et al., 2014). From this universal framework authors can then tailor their study to

the appropriate cultural context.

Given the nature of EDJ, it is unsurprising that links to human well-being outcomes feature heavily

in the literature. While we did find a range of outcomes, a few areas dominated. Interestingly, while

economic status is the most commonly used social demographic in articles, livelihoods are not the

most commonly examined human well-being outcome (see results, figure 4). Rather, the majority of

articles examine health and mortality outcomes. While human health is of obvious importance,

many other well-being issues can result from environmental exploitation. Furthermore, non-health

outcomes can affect large groups of people, and themselves lead to health issues. For example,

climate change enhanced droughts and floods can destroy livelihoods, and increase disease

incidence (Gichere et al., 2013). As such, a more diverse exploration of human well-being outcomes

is a good candidate area for future EDJ research.

We found a clear tendency for EDJ articles to relate to either air pollution, or climate change which

reflects the large-scale nature of these issues and their potential for devastating impacts on human

well-being and livelihoods. However, we discovered an imbalance between the discussion, and

some form of measurement (or modelling) across these topics. This is particularly true for the

climate change literature, where only a very small proportion of papers measured some climate

change effect and a human well-being outcome. While the discussion of how we ought to decide

what might be a just distribution of climate change costs and benefits is of obvious importance,

further benefit to the field can be gained through conducting empirical studies which develop the

evidence-base for the EDJ implications of climate change. Similarly, but not to the same degree, for

the literature related to air pollution, we found many articles which discuss health outcomes from

air pollution exposure, but far fewer which attempted to measure them or record incidence.

Furthermore, health outcomes were often only discussed, rather than measured across much of the

literature. For example, statements which indicate air borne pollutants have been linked to

respiratory illness, with no accompanying respiratory illness data are common. This illustrates an

opportunity for the EDJ literature, as to maximize the applicability of EDJ research (e.g. to

influence future policy) it is not only important to continue to identify injustices in distribution of

51

environmental issues. Rather, it is equally (if not more), important to also link such injustices to

human well-being outcomes. While causative links are difficult to establish, this could potentially

be addressed by an increase in more complex, mixed methods studies. For example, spatial

quantitative methods could be used to identify correlates, controls, and counterfactual areas across a

large space for use in the experimental design of follow on, fine scale studies.

We found that almost all papers (90.2%) which assessed the equitability of human well-being

outcomes concluded that there are distributive inequities. However, the reason for this finding is

unclear. It may be that almost all natural resource exploitation is profoundly inequitable. However,

it may alternatively represent a form of confirmation bias in the EDJ literature, where authors may

be selecting cases where they have prior knowledge of an inequity. Confirmation bias is an issue of

increasing relevance to most scientific disciplines (MacCoun and Perlmutter, 2015). As the reason

for such a high statistic is unclear, we highly recommend that future EDJ studies take precautions to

prevent potential biases where possible, and utilise methods such as counterfactual thinking and

randomized trials, quasi-experimental designs, blind analysis, and the use of control variables

(MacCoun and Perlmutter, 2015; Miguel et al., 2014; Ferraro, 2009). In either case, it was not in the

scope of this study to assess the quality of methodology within the literature, so our 90.2% inequity

statistic ought to be considered with caution. To clarify this finding, we recommend that future

reviews of the EDJ literature attempt to incorporate study quality assessment.

Study Limitations

While our overall aim with this study was to be as exhaustive in our investigation of the EDJ

literature as possible, available resources constrained us. As such, there are several limitations to

our study which need to be highlighted for disclosure. Most importantly was our lack of ability to

conduct a full, manual screening of the literature at the article title level. We did not have the time

and resources to manually screen over 40,000 articles. Relatedly, as the academic databases alone

yielded over 40,000 articles, we did not have the resources available to screen all potential sources

of literature, including very literature diverse databases such as Google Scholar, the grey literature,

and book chapters. Consequently, there is the potential that we missed some EDJ articles at each

subsequent screening level, and our results largely capture the academic peer-reviewed literature,

which may have biased our sample and results. However, given the relatively large number of full

text screened articles we were able to screen, our study can be thought of as a non-exhaustive, yet

representative sample of the EDJ literature. Furthermore, our study was limited to English language,

peer reviewed literature which no doubt introduced some biases. These potential biases may affect

the findings around geographic bias in the EDJ literature, e.g. if non-English language or non-peer

52

reviewed studies are available in greater numbers in developing countries. It is recommended,

where possible, future reviews on this subject area address these limitations.

Conclusion

This study makes a meaningful contribution to the broader literature reviewing EDJ. Consistent

with previous reviews, we found overall, that there is rich and somewhat diverse EDJ literature

(Banzhaf et al., 2019; Brulle and Pellow, 2006). However, there are key trends apparent across the

majority of articles. This is particularly true for the geographic scope of the literature, which is

clearly misaligned to the real world scope of EDJ problems. While a few most developed nations

are very well represented, many of the world’s least developed and/or least socially equitable

societies are not. This is not to state that work which examines the very real and often harmful

environmental inequities found in developed nations. Rather, we argue this gap highlights a key

area for expanded focus by EDJ researchers, which would both advance the knowledge on EDJ and

contribute to achieving more equity in the investment of research efforts internationally.

Consequently, we encourage EDJ researchers from the world’s most developed nations to look

further afield when developing future projects, via channels such as capacity building, funding and

to develop supportive relationships with researchers in those countries of interest that are most in

need of EDJ research. In addition, there is an opportunity to direct efforts to linking EDJ

observations with human well-being outcomes in order to develop the evidence base for decision-

making, such as environmental and planning policies, informed by EDJ literature. More generally,

we strongly encourage EDJ researchers to use our results when developing future research projects

to contribute to a diverse, robust and insightful literature.

53

Chapter 3 Global mismatch between greenhouse gas emissions and the

burden of climate change

Online publication: http://dx.doi.org/10.1038/srep20281

Full citation: Althor G, Watson JEM and Fuller RA. (2016) Global mismatch between greenhouse

gas emissions and the burden of climate change. Scientific Reports 6.

Overview

This study was published in Nature Scientific Reports in 2016. It is an innovative and novel

approach to examining the fairness dimensions of climate change. I produced the first global level

study to use data for all greenhouse gas emissions (GHGs), against the vulnerability of nations to

climate change. I argue that the economic benefits some few countries are gaining from emitting

GHGs are far outweighed by the burdens they are placing on other lower emitting nations. I use

quantitative Gini and Robin Hood coefficients to present this idea.

Research question: How are greenhouse gas emissions and the burdens of climate change

distributed globally?

Sub-questions:



This study utilises my combined principles which regard cases where some harm (or other type of

burden) is produced, for the specific benefit of a single person, or exclusive group, as follows:

P3: Where burdens upon any individual or exclusive group are the byproduct of the use of natural



Highlights:

• First global study to use data for all greenhouse gas emissions to assess distributive justice

of climate change.

• Prime example of how natural capital distributive justice principles can be used to design

and evaluate empirical research.

http://dx.doi.org/10.1038/srep20281

54

• Uses quantitative measures of inequality to support claims, and provides geographical maps

making findings easily understandable.

55

The text appears exactly as published, with the exception of change from Nature referencing

style to Sage Harvard referencing style and placement of Methods section before Results

section.

Abstract

Countries export much of the harm created by their greenhouse gas (GHG) emissions because the

Earth’s atmosphere intermixes globally. Yet, the extent to which this leads to inequity between

GHG emitters and those impacted by the resulting climate change depends on the distribution of

climate vulnerability. Here, we determine empirically the relationship between countries’ GHG

emissions and their vulnerability to negative effects of climate change. In line with the results of

other studies, we find an enormous global inequality where 20 of the 36 highest emitting countries

are among the least vulnerable to negative impacts of future climate change. Conversely, 11 of the

17 countries with low or moderate GHG emissions, are acutely vulnerable to negative impacts of

climate change. In 2010, only 28 (16%) countries had an equitable balance between emissions and

vulnerability. Moreover, future emissions scenarios show that this inequality will significantly

worsen by 2030. Many countries are manifestly free riders causing others to bear a climate change

burden, which acts as a disincentive for them to mitigate their emissions. It is time that this

persistent and worsening climate inequity is resolved, and for the largest emitting countries to act on

their commitment of common but differentiated responsibilities.

Introduction

The current generation is the first to feel the effects of anthropogenic climate change (IPCC, 2014a;

Ipcc, 2014b). Despite their well-known harmful impacts to the world’s climate system (IPCC,

2014a; United Nations Framework Convention on Climate Change, 1992), greenhouse gases

(GHG) are deliberately emitted by countries to drive economic growth and enhance human

wellbeing (Jorgenson, 2014). Spatially localised environmental issues, such as city air pollution

(Sheehan et al., 2014), may result from high GHG emissions, but the most damaging and long

lasting consequence, that of global climate change (Montzka et al., 2011), is not constrained within

the border of the emitting country (IPCC, 2014a). Rather, by polluting the Earth’s atmosphere with

GHG emissions through fossil fuel combustion, deforestation and agricultural activities, emitting

countries are degrading the world’s climate system, a common resource shared by all biodiversity,

including people (Betts, 2008).

56

Because the impacts of GHG emissions can be felt beyond a country’s border, and the impacts of

climate change on countries are highly variable, there is potential for some emitters to contribute

more or less to the causes of climate change than is proportionate to their vulnerability to its effects

(Cazorla and Toman, 2001; Kjellstrom et al., 2009; Trenberth, 2011). This inequity has not gone

unnoticed in international climate negotiations or global reporting (IPCC, 2014a; United Nations

Framework Convention on Climate Change, 1992). As far back as 1992, the United Nations

Framework Convention on Climate Change (UNFCCC) committed to the principle of “common but

differentiated responsibilities”, in which countries have a common responsibility in reducing GHG

emissions, but historic emissions and differences in current development levels mean that countries

have different levels of emissions reduction obligations (Cazorla and Toman, 2001). Both of the

previous IPCC Assessment Reports have acknowledged the inequity in the causes and effects of

climate change (IPCC, 2014a; IPCC, 2007) although operationalising the principle has proved

difficult (Ostrom, 2014). This is primarily because developing and developed countries continue to

disagree over the extent of each other’s responsibilities (Ostrom, 2014; Cole, 2015). One major

impediment to resolving such debates is a poor quantitative understanding of the magnitude of the

global inequity in emissions and impacts. ‘Free rider’ countries contribute disproportionately to

global GHG emissions with only limited vulnerability to the effects of the resulting climate change,

while ‘forced rider’ countries are most vulnerable to climate change but have contributed little to its

genesis (Rao, 2014; Füssel, 2010). This is an issue of environmental equity on a truly global scale

(World Resources Institute, 2014).

Here, we measure the current pattern of global climate change equity, and assess whether the

situation will improve or worsen by 2030, using data on GHG emissions (World Resources

Institute, 2014) and newly available national climate change vulnerability assessments (DARA,

2012). We address the lack of a contemporary, qualitative assessment of global climate equity that

incorporates key variables. Previous studies have been limited to CO2 emissions datasets, omitting

the most potent and long lasting GHGs (IPCC, 2014a; Montzka et al., 2011; Füssel, 2010), and

used vulnerability variables that do not capture the complexity of climate change threats, and cannot

be forecasted. Here, we use the most recently available datasets based on comprehensive national

vulnerability assessments and comprehensive GHG emissions data to produce an easily replicable

snapshot of the relationship between countries’ GHG emissions and their vulnerability to the

negative effects of climate change (World Resources Institute, 2014; DARA, 2012), and forecast

this to 2030. We employ economic metrics, the Gini and Robin Hood coefficients (Coulter, 1989b),

to quantify the present level of equity in GHG emissions. Only through a proper empirical

57

understanding of the pattern of climate equity now, and how it will change in the near future, can

signatories of the UNFCCC make meaningful progress toward resolving the inequity in the burden

of climate change impacts.

Methods

We quantified climate change equity, defined as the distribution of climate change benefits and

burdens, using data from two publicly available datasets and national GDP data. National level data

sets suffer from some weaknesses such as a lack of accounting for sub-national variability and

scaling. Nonetheless, they are still highly useful as global metrics as they provide aggregated

assessments at the national level, which is the most meaningful for international policy negotiations.

We extracted data on national vulnerability to the negative impacts of climate change from DARA’s

Climate Vulnerability Monitor (CVM) (DARA, 2012). The CVM uses 22 climate vulnerability

indicators across four impact areas (Environmental Disasters, Habitat Change, Health Impact, and

Industry Stress) to evaluate the vulnerability of 184 countries to climate change impacts for the

years 2010 and 2030. Each of the 22 indicators is individually aggregated from various data sources

and models and then combined to determine a country’s overall climate vulnerability, measured by

impact to share of GDP and mortality (as these impacts are comparable across the wide range of

countries). The CVM calculates vulnerability projections for 2030 using human population growth,

mortality and GDP predictions. The CVM uses five vulnerability categories (low, medium, high,

severe and acute) which are determined using a mean absolute standard deviation method18. The

CVM categories do not of course capture the full complexity of national climate vulnerability, as

capturing this would require an impractical degree of data. However, we consider the 22 indicators

used by the CVM as capturing a high enough level of complexity to provide a meaningful

approximation of national vulnerability.

Data on GHG emissions (by countries) were exported from the World Resource Institute’s (WRI)

Climate Analysis Indicators Tool (CAIT) (World Resources Institute, 2014), a database of national

and international GHG emissions derived from multiple sources. The CAIT data set compiles data

for the six main GHG gasses (carbon dioxide (CO2), methane (CH4), nitrous oxide (N2O),

hydrofluorocarbons (HFCs), perfluorocarbons (PFCs) and sulfur hexafluoride (SF6)) from 185

countries over the period from 1990-2012. We used the 2010 data for this study to match the CVM

vulnerability data. The WRI compiled GHG data from UNFCCC reports and complemented with

data from several NGO sources (World Resources Institute, 2014), including emissions data from

58

six major sectors (Land Use Change & Forestry, Energy, Industrial Processes, Agriculture, Waste,

and International Bunkers) and several subsectors. The CAIT data set reports at the national level,

however we extrapolated per capita emissions results by dividing data by 2010 and 2030 population

data from the World Bank (World Bank Group, 2015) (see Supplementary Table S4 online).

We excluded the ten countries (Cook Islands, Federated States of Micronesia, Marshall Islands,

Montenegro, Nauru, Niue, Saint Kitts & Nevis, Serbia, Somalia and Taiwan) with data missing in

any dataset, and 179 remained for analysis. In addition, there were also insufficient data available

for many of the world’s island and archipelagic countries. Given the negligible GHG emissions and

high climate change vulnerability of such countries, the majority are highly likely to qualify as

climate forced riders (Nurse et al., 2014; Widlansky et al., 2013) and as such, we expect that climate

forced rider countries are likely underrepresented in our results. National GDP (measured in Current

US$) was extracted from the World Bank Group (World Bank Group, 2015), who measure GDP as

the gross value of all resident producers in an economy plus taxes.

We created a Lorenz curve to represent the variation of GHG emissions among countries using the

CAIT dataset, and calculated the Gini index to measure inequity in GHG emissions among

countries, and the Robin hood index to measure how much of the total global emissions would have

to be redistributed to achieve equity among countries (see Supplementary Fig. S2 online).

We compared the CAIT GHG data and the CVM vulnerability data both in 2010 and 2030 to assess

whether the most heavily polluting countries were also those least vulnerable to the negative effects

of climate change. We divided the CAIT GHG emissions into quintiles, matching the CVM data, to

enhance comparability between the datasets and enable visualisation of climate equity in the recent

past (2010) and near future (2030). We placed the emissions quintiles on a scale between the

highest (acute emissions) and the lowest (low emissions) emitting countries. We also tested the

correlations between GHG emissions and GDP against vulnerability to climate change by treating

vulnerability categories as ordinal data and undertaking spearman’s rho tests using R statistical

software (R Core Development Team, 2015). R has a computational limitation for p-values lower

than 2.2e-16, as such, where values this small were reported we wrote “p = 0”. Additionally, we

counted countries in each CVM category and compared them between each time period.

In common with other studies of inequity in climate change (Otto et al., 2015), we used terminology

from the economics literature to define ‘free riders’ and ‘forced riders’ (Ahnlid, 1992), recognising

59

that a strict definition of these terms often applies only to situations where one agent’s use of a

resource does not directly incur a cost to another agent. We define climate free riders as those

countries in the ‘acute’ GHG emissions quintile and the ‘low’ vulnerability category, as they

disproportionately receive benefits from climate change (via the national wellbeing generated by

GHG emissions) but pay few costs in the sense they are the least vulnerable to negative climate

change effects (Keohane, 2015; Kennedy and Basu, 2014). Conversely, we define climate forced

riders as those countries that fall within the ‘acute’ vulnerability category and the ‘low’ GHG

emissions quintile, as they are the most susceptible to the negative consequences of climate change

but receive the least benefits. Those countries that we define as equitable, fall in the same emissions

quintile and vulnerability category (for example, low emissions quintile, low vulnerability

category), as their emissions benefits are concomitant with their climate change burden.

Results

Greenhouse gas emissions are spread highly unevenly across the world’s countries (Fig. 1), with the

top ten GHG emitting countries generating > 60% of total emissions, and three countries, China

(21.1%), the United States of America (14.1%) and India (5.2%) being by far the largest

contributors. A Gini coefficient of 80.9 indicated extreme inequality in the distribution of emissions

among countries, given that the index can only vary between 0 (perfectly even responsibility) and

100 (one country responsible for all emissions) (Coulter, 1989b). A Robin Hood index of 64

indicated that 64% of GHG emissions would need to be redistributed to achieve an even distribution

among countries (Coulter, 1989b). Vulnerability to the impacts of climate change was also

unevenly spread among countries, with 17 countries acutely vulnerable to climate change impacts in

2010 (Fig. 2). The majority of these were island countries located in the Atlantic, Pacific and Indian

oceans (n=7, 35.3%) and African countries (n=8, 47%). By 2030 the number of acutely vulnerable

countries is predicted to rise dramatically (n=62; Fig. 2), and the majority of these will again be

island (n=20, 32.8%) and African (n=27, 44.2%) countries.

60

Figure 4. Global inequity in the responsibility for climate change and the burden of its

impacts. (a) Climate change equity for 2010. (b) Climate change equity for 2030. Countries with

emissions in the highest quintile and vulnerability in the lowest quintile are shown in dark red (the

climate free riders), and those countries with emissions in the lowest quintile and vulnerability in

the highest quintile are shown in dark green (the climate forced riders). Intermediate levels of equity

are shown in graduating colours, with countries in yellow producing GHG emissions concomitant

with their vulnerability to the resulting climate change. Data deficient countries are shown as grey.

Maps generated using ESRI ArcGIS (ESRI ArcGIS, 2011).

61

Figure 5. Vulnerability to climate change, mean GHG emissions, and mean GDP. (a) Number

of countries in each climate change vulnerability category, derived from DARA vulnerability data

(DARA, 2012), for 2010 (blue bars) and 2030 (green bars). (b) Mean GHG emissions for 2010,

derived from CAIT GHG emissions data (World Resources Institute, 2014), shown in CO2

equivalent units and climate vulnerability categories for 2010 (blue bars, with standard error) and

2030 (green bars, with standard error). (c) GDP shown in current US$ (in billions), derived from the

World Bank GDP 2010 data (World Bank Group, 2015), and 2010 GHG emissions. (d) Mean GDP

for 2010 shown in current US$ (in billions) and climate change vulnerability for 2010 (blue bars,

with standard error) and 2030 (green bars, with standard error).

Countries least vulnerable to the impacts of climate change were generally the highest GHG

emitters, and conversely those most vulnerable to climate change were the least responsible for its

genesis. This inequity held true for both 2010 and 2030, with a negative relationship between

emissions and climate vulnerability in both years (2010: ρ = -0.4, n = 175, p = 0; 2030: ρ = -0.37, n

= 175, p = 0). The only exception is in 2030, where countries acutely vulnerable to climate change

will have slightly higher average emissions than those in the severe category (2030: severe = 48.83

mtCO2e, acute = 103.13 mtCO2e).

62

In 2010, of the 179 countries assessed, 28 (15.6%) were in the same quintile for GHG emissions

and vulnerability to the negative impacts of climate change. This indicates that their vulnerability to

climate change approximately matched their relative contribution to its genesis (Fig. 1). Ninety

countries (50.3%) had GHG emissions in a higher quintile than their 2010 climate vulnerability, and

20 (11.2%) countries were free riders, with GHG emissions in the highest quintile and climate

vulnerability in the lowest quintile (Fig. 1; see Supplementary Table S4 online). Sixty-one (34%)

countries had GHG emissions in a lower quintile than their climate vulnerability, and six (3.4%)

countries were forced riders, with GHG emissions in the lowest quintile and climate vulnerability in

the highest quintile (Comoros, Gambia, Guinea-Bissau, São Tomé and Príncipe, Solomon Islands

and Vanuatu; see Supplementary Table S4 online).

By 2030, climate change inequity will rise further, with an increase in the proportion of countries

that are forced riders (n=20; 11.2%), but fewer free riders (n=16; 8.9%) and equitable countries

(n=23; 12.8%; see Supplementary Table S4 online). Free riders are typically located in the world’s

sub-tropical and temperate regions, while forced riders are frequently located in tropical regions

(Fig. 1).

Greenhouse gas emissions were positively correlated with GDP (2010: ρ = 0.84, n=175, p = 0; Fig

2c), while climate vulnerability declined with increasing GDP (2010: ρ = -0.69, n = 175, p = 0;

2030: ρ = -0.65, n = 175, p = 0; Fig. 2d). Our analysis considers the absolute contribution of each

country to climate change, but we also examined climate change equity in per capita terms to

provide a more complete picture of emissions responsibilities. The patterns were broadly similar,

with, for example, Australia, Russia and the United States of America remaining free riders (see

Supplementary Fig. S3 online). However, several populous major emitters (e.g. United Kingdom,

China, and Brazil) were no longer categorised as free riders.

Discussion

Climate change inequity is globally pervasive, and correlated with economic output. Some

countries, such as China and the United States of America, are in a win-win position of achieving

economic growth through fossil fuel use with few consequences from the resulting climate change,

while many other, mostly Island and African, countries suffer low economic growth and severe,

negative climate change impacts (see Supplementary Table S4 online). The beneficiaries of this

climate inequity have few incentives to meaningfully reduce or halt their GHG emissions. Despite

many of the broad issues around climate equity being well known (IPCC, 2014a), well-funded

63

global mechanisms that are being implemented still do not exist. This has serious consequences for

our ability to slow the rate of climate change, and reduce the wellbeing implications for forced rider

countries.

There are several global policy frameworks currently being debated that could address elements of

the problem. The Paris Agreement (UNFCCC, 2015a), secured at the 21st UNFCCC Conference of

the Parties (COP21), for example, sets an ambitious target of limiting global warming to 1.5C

above preindustrial levels. However, the 160 indicative nationally determined contributions

(INDCs) pledges submitted by signatories to the UNFCCC prior to COP21 (UNFCCC, 2015b),

indicate that current targets for GHG emissions are unlikely to limit warming to below 2C

(Jackson et al., 2015) With no binding agreement established at COP21 for INDCs, there is no clear

indication of how successful the Paris Agreement will be (UNFCCC, 2015a). Addressing GHG

emissions is clearly an important first step in ensuring the burden of climate change is not amplified

in the future. However, the historic commitment to GHG emissions reduction by key free riders has

been slow. Only 50 countries ratified the previous Doha Amendment to the Kyoto protocol, which

did not include key free riders such as the United States and Russia (Fekete et al., 2013).

Furthermore, some countries have actually backtracked on their commitments to emissions

reductions (e.g. Canada and Australia) (Pizer and Yates, 2015; Harrison, 2012).

Likewise, the Paris Agreement calls for urgent and adequate financing of US$100 billion per year

by 2020 for climate mitigation and adaptation through the Financial Mechanism of the Convention

(FMC) (UNFCCC, 2015a). However, there is no legally binding mechanism under which parties are

responsible for providing this funding. History suggests such funding goals are not always met. For

example, the Green Climate Fund (GCF) was established in 2010 under the UNFCCC to mobilise

funding support for the least developed countries that are most vulnerable to climate change, yet it

remains poorly funded, with only US$10.2 billion received in pledges by November 2015 (Green

Climate Fund, 2015). Addressing these issues around climate funding will play a critical role in

addressing climate inequity (Pickering et al., 2015).

Conclusion

It is clear climate change inequity must be addressed. If the commitment to the principle of

common but differentiated responsibilities that was widely accepted early on in the UNFCCC is to

be acted upon, member states now need to do much more to hold climate free riders to account. To

ensure equitable outcomes from climate negotiations, there needs to be a meaningful mobilization

64

of policies, such as the Paris Agreement, that achieve national level emissions reductions, and to

ensure the vulnerable forced-rider countries are able to adapt rapidly to climate change. The

provisioning of these policy mechanisms will require a distribution of resources and responsibilities

and we believe our results provide one way to understand where these responsibilities lie. The Paris

Agreement may be a significant step forward in global climate negotiations. However, as the

Agreement’s key policies are yet to be realized, member states have both an exceptional

opportunity and a moral impetus to use these results to address climate change equity in a

meaningful manner.

Author contributions

G.A., J.E.M.W. and R.A.F. designed the analysis. G.A. performed the analysis and analysed the

results. G.A., J.E.M.W. and R.A.F. wrote the paper.

Competing Financial Interests statement

The authors declare no competing financial interests.

65

Chapter 4 The distribution of industrial pollution across socio-

economic groups in Australia; a ten-year, a fine scale distributive

justice analysis

Full citation: Althor G, Witt B and Colvin R. (forthcoming) The distribution of industrial pollution

across socio-economic groups in Australia; a ten-year, a fine scale distributive justice analysis.

Target journal: Environmental pollution.

Overview

This study is forthcoming, but as presented, is appropriate for submission to a peer-reviewed

journal. With this study I build on existing work which assessed the distribution of harmful

industrial pollution across Australian socio-economic groups. It is the finest scale study of its kind

undertaken in Australia, and the first to add a temporal dimension. Furthermore, I address the

limitations of previous work in weighting emissions by toxicity, which is critical for understanding

the potential impacts on people’s well-being.

Research question: Are there social inequities in the distribution of exposure to harmful industrial

pollution across Australia?

Sub-questions:



This study specifically utilises my principle of strict equality, as follows:



Highlights:

• Highest resolution study of its kind undertaken for Australia.

• First study of its kind to introduce a temporal, as well as geographic element.

• Provides the most robust method to date for weighting emissions by toxicity, specifically for

use with data provided by the Australian Government.

66

The following text appears exactly as manuscript prepared for peer-review submission.

Abstract

Environmental distributive justice assesses the cost and benefit trade-offs among affected social and

economic groups in situations of natural resource use. Industrial pollution is a wide spread

phenomenon, and the largely unavoidable byproduct of many economic activities. Industrial

pollution can have negative effects on human health when people are exposed to emissions. As such

understanding the distribution of, and human exposure to, industrial pollutants has been

increasingly studied across the world, and within Australia. In particular there is a focus on

understanding if specific social groups (such as racial or economic groups) are more or less likely to

be situated near industrial pollution sources, and as such more or less likely to suffer negative health

effects. However, an up to date, comprehensive and nationally representative study assessing the

distribution of industrial pollutants and nearby social groups has not been undertaken in Australia.

Here we present a study using data gathered at three intervals, over 10-years for all of Australia

where we assessed emissions for all 93 industrial pollution substances that are currently monitored

in Australia and weighted each substance according to its potential to cause harm to people. We

show that there are no significant correlations between industrial pollutant sources and social

groups within Australia. This result is in contrast to previous Australian studies which have shown

distributive injustices between industrial pollution and social groups.

Introduction

In Australia industrial pollution occurs as a side effect of a highly diverse range of economic

activities such as power stations, mining and chemical manufacturing (Keywood et al., 2016).

These activities broadly improve human well-being in Australia through their contribution to Gross

Domestic Product (GDP). This is best illustrated by the approximately $1.69 Trillion (AUD) value

of Australian industry (ABS, 2017a). However, via pollution emissions into the environment,

industry can also have serious negative well-being effects on people, such as causing disease and

premature death (Moore and Hotchkiss, 2016; Kampa and Castanas, 2008). For example, a recent

global commission on pollution and health states that pollution caused an estimated nine million

mortality related diseases in 2015, which accounted for some 16 percent of global deaths

(Landrigan et al., 2017). As such, the costs and benefits of industry can be described as a complex

trade-off between improving well-being via economic growth and degrading it via toxic emissions.

Furthermore, extant evidence indicates these trade-offs are rarely evenly distributed among social

and economic groups within populations.

67

In many of the world’s nations, industrial pollution disproportionately causes death and disease

among least advantaged groups, such as low-income countries globally, or among vulnerable socio-

economic groups at various sub-national scales (Padilla et al., 2014; Charafeddine and Boden, 2008;

Germani et al., 2014). These fairness implications of such distributions make this an issue of

environmental distributive justice (EDJ). Distributive justice can broadly be described as principles

which argue that the distribution of social benefits and burdens ought to be guided by moral

principles of fairness (Allingham, 2014). EDJ applies distributive justice principles to

environmental issues. When choosing a pathway to increased GDP through the utilization of, and

impact on, the natural environment, a society (or its decision-makers) should seek to ensure an

equitable burden of any hazards or impacts on human wellbeing. This argument holds true at any

scale from the fine (i.e. between individuals in a village) up to the global (i.e. between nations). As

such, EDJ principles can be applied to industrial pollution in order to: 1) understand the resultant

trade-offs between positive and negative well-being outcomes; and 2) identifying any unjust, socio-

economic distributive patterns of pollution.

Since the 1980s, increasing public awareness of human and environmental impacts has resulted in a

distinct increase in scholarship and activism around EDJ and industrial pollution, particularly in the

United States and Europe (Padilla et al., 2014; Holnicki et al., 2017; Charafeddine and Boden,

2008). However, in countries such as Australia there is much less research dedicated to the

investigation of EDJ and industrial pollution (Chakraborty and Green, 2014). Pollution (industrial

and otherwise) is a serious issue in Australia with around 3,000 deaths per year attributed to urban

pollution alone (Keywood et al., 2017). Furthermore, it has been estimated that the health costs of

pollution-related premature deaths are approximately $2.5 billion AUD annually (Broome et al.,

2015). While the overall disease and mortality burdens related to pollution in Australia are

relatively well understood, far less is known about the distribution of this pollution, and if it is

correlated with socio-economic factors such as ethnicity or income.

We found only three studies based in Australia that have assessed the EDJ implications of pollution

nationally (Chakraborty and Green, 2014; Knibbs and Barnett, 2015; Dobbie and Green, 2015).

Chakraborty & Green (2014) aggregated a sub-set of air pollution data from the Australian

Government’s National Pollution Inventory (NPI), and four aggregates of demographic data from

several data sets from the Australian Bureau of Statistics (ABS). The authors state that their study

shows socio-economically disadvantaged populations are more likely to be distributed near

concentrated sources of toxic air pollutants. However, in this study the emissions data were

68

“toxicity-weighted volume of air pollution”, based on an out of date toxicity rating which was not

developed for use in this manner (NEPC 1999, appendix III) (Chakraborty and Green, 2014: pg. 2).

Furthermore, details of the actual weighting method used are unclear making replication of the

study difficult. As such, the validity of weighting the data in such a way may be problematic. In

addition, the study used the relatively large Australian standard Statistical Area Level 2 geographic

units. Recent research has shown that spatial analysis which utilises such large units can produce

unreliable results (Hanigan et al., 2017), such as aggregation errors and other measurement errors

(Anderton et al., 1994).

Knibbs & Barnett (2015), also assessed the distribution of exposure to air pollutant NO2 by social

stratification, including variables for socio-economic advantage and disadvantage, and percent of

people identifying as Aboriginal or Torres Strait Islander. This robust study utilises high quality

complex pollution and socio-economic data at the smallest available spatial scale. However, the

authors only assessed exposure to a single pollutant (nitrogen dioxide (NO2)) and did not assess the

distribution of sources of the pollutant (i.e. whether or not it was generated by industry). Their

results show evidence of some distributive inequalities in NO2 exposure across Australia, with

people identifying as Aboriginal or Torres Strait Islander more likely to be exposed to NO2.

Dobbie & Green (2015), focused on the regulation of air pollution in Australia and the geographic

distribution of breaches of pollution licence conditions. In particular, they assessed two case studies

to illustrate the nature of these distributions. Dobbie & Green (2015) suggest that results of their

study indicate disproportionate human health impacts present in the case study communities, and

argue that this is largely the result of failures in existing regulatory systems.

These previous studies suggest that there are pollution related cases of distributive injustice in

Australia, in particular they suggest least economically advantaged groups and Indigenous

Australians are more likely to be exposed to pollution. However, in an earlier meta-analysis of EDJ

studies (mostly from the USA), Bowen (2002) found, contrary to the dominant opinion, that there

was little statistically significant evidence to suggest patterns of distributive injustices in the

location of pollutant emitting sites. Given Bowen’s analysis and the limitations with the existing

Australian literature, we present here a quantitative spatial analysis of industrial emissions in

relation to EDJ principles using Australian nationwide, fine-scale, and the most recently available

data. As such, this study aims to understand how industrial emissions are distributed among social

and economic groups in Australia.

69

The results of this study will highlight any current industrial pollution injustices, and the methods

used will provide future researchers with a replicable framework to re-assess EDJ and industrial

pollution in the future.

1.1 A testable principle for distributive justice

In this study we apply a relatively simple strict equality distributive justice principle drawn from the

theoretical argumentation in broader distributive justice literature (Allingham, 2014), as such we

posit:

1) exposure to harmful pollution (i.e. pollution beyond a safe threshold) will always reduce human

well-being by causing disease and/or death;

2) all people residing in Australia of any socio-economic group are morally and legally equal, and

lastly;

3) given all people residing in Australia are equal, the rates of exposure to industrial pollution ought

to be the same for all socio-economic groups.

Using these premises our research offers an analysis determine whether our strict equality

distributive justice principle is violated for industrial pollution in Australia. As such, our research

question is:

Is there evidence of social inequities in exposure to harmful industrial pollution in Australia?

In order to answer this research question, and in light of the findings of previous research in

Australia, we will test the following Hypothesis:

H1 = The prevalence of exposure to industrial pollution is related to socio-economic status.

H0 = The prevalence of exposure to industrial pollution is not related to socio-economic status.

It is important to note that with this study we are not aiming to prove a causative effect between

pollution exposure and associated health burdens, rather we are seeking to identify whether a

relationship between these variables is present.

Methods

We use spatial statistics to explore the relationship between industrial pollution and socio-economic

status, Australia wide. Our analyses required the aggregation and adjustment of several large, data

rich, and fine scale databases (figure 1). For each data set we use annual data taken at three time-

points: 2006, 2011, and 2016, to analyse any overall changes over the 2006-2016 decade. The script

for all data aggregation and analysis methods undertaken using R statistical software and associated

packages can be found in Supplementary file 1.

70

Figure 1 Data analysis process. For each year (2006, 2011, and 2016) three large, Australia wide

data sets were used to create total toxicity potential ratings and analysed using regression models.

2.1 Spatial data

To best understand the distribution of industrial pollution within Australia, we have used fine scale

spatial data produced by the Australian Bureau of Statistics (ABS). For 2006, we use Collection

Districts (CD). There are 38,200 CDs Australia wide, with each representing an average of 225

dwellings (ABS, 2011). In 2011 the ABS structure changed to the Australian Statistical Geography

Standard, which resulted in CDs changing to finer scale Statistical Areas (SA), as such we use SA1

units (the smallest SA unit) for years 2011 and 2016. The SA1 units have approximately 400 people

each, and as such the number of SA1 units will change as the Australian population changes. At the

time of writing, there are 57,523 SA1 units (ABS, 2017b). Research has shown that using fine scale

data such as CD and SA1 produces the most reliable results for data analysis (Hanigan et al., 2017).

71

2.2 Industrial pollutant emissions data

To examine the distribution of industrial pollution, we use data from the Australian National

Pollutant Inventory (NPI) (NPI, 2015). The NPI is a sub-division of the Australian Department of

the Environment and Energy and provides a self-reporting and tracking system for industrial

pollutants in Australia. It primarily acts as a publicly accessible, online database of emissions data.

Australian law requires that all facilities which exceed particular (substance dependent) emissions

thresholds for 93 defined substances must self-report emissions to the NPI (NPI, 2014). Emissions

records include both fugitive and point source. Here we utilise all point source emissions data from

the NPI (from the appropriate annual reports), which includes emissions data for every polluting

facility in the country, in order to estimate a potential path of pollutant exposure to surrounding

communities. The 93 substances are those that potentially harm people or the environment. The NPI

annually collects, records, and reports data for emissions to air, water and land. It is critical to note a

key limitation to the NPI data is that the reported data are available only above the threshold at

which reporting is required (NPI, 2015), meaning very low level emissions from individual

pollution sources are not recorded.

2.3 Total Toxicity Potential

The NPI records the mass of emissions from each source but does not give any indication of their

human toxicity potential. As such, adjusting the NPI emissions data for toxicity is critical for any

meaningful use in impact analyses, as a small volume of a highly toxic pollutant is potentially much

more harmful than a large amount of a less toxic substance (Wu et al., 2009). To address this we

have utilised the Human Toxicity Potentials (HTP) defined by Hertwich et al. (2001) to weight NPI

emissions data. The HTP is an index that was specifically designed for use with emissions

inventories such as the NPI. Its primary use is to understand the potential harm caused by pollutant

release. HTP scores reflect the potential harm of a unit of chemical released, as each HTP score is

based on the intrinsic toxicity of a substance, and accounts for a number of exposure routes. The

HTP classifies chemicals into two different vectors (air and water) and two types of health effects

(cancer causing and non-cancer-causing). While it is very extensive, the HTP is not exhaustive.

There are many substances (several of which are known to be harmful to human health) which are

not included in the HTP, as such we have excluded from our analysis any pollutants present in the

NPI which do not have a HTP score. We have weighted the NPI point source emissions data by the

HTP scores to produce a Total Toxicity Potential (TTP), per each emissions point source ( ).

72

For simplification, we have combined the air, water, and cancer and non-cancer-causing scores into

a single toxicity potential value , which is described by:

Total toxicity potential for each point source = Sum of (air emissions*cancer effects) + (water

emissions*cancer effects) + (air emissions*non-cancer effects) + (water emissions*non-cancer

effects)

NPI emissions are represented by E, Cancer effects by C and Non-cancer by N, and all emissions

and health effects per point source i are separated into air borne a, and water borne w emissions. It

is important to note that as each chemical can have both, either, or neither C and N health effects,

emissions can potentially contribute nil, once, or twice to the TTP. Any potential interactions

between C and N are not accounted for.

2.3.1 Mapping the distribution of Total Toxicity Potentials ( ) across Australia

Using ArcGIS Desktop (ESRI, 2011), for each year included in the study, we assigned TTP values

to each CD (for 2006) or SA1 unit (for 2011 and 2016) based on proximity to each emissions point

source. For each emissions point source (provided with NPI data, and represented as point

coordinates with six decimal places of accuracy) we added a one kilometre buffer to compensate for

the ‘edge effect problem’ as per Chakraborty and Green (2014), which is methodologically in line

with an analytical review of environmental justice studies (Bowen, 2002). This provides a realistic

area of effect of pollution around the point source. Any SA1 unit which fell (even partially) within

this one kilometre buffer was assigned the of the point. There were many spatial units which

intersected with multiple buffered points, in these cases, all were summed.

2.4 Socio-demographic Advantage and Disadvantage

To examine social and economic groups within each spatial unit, we have used four indicators

which make up the Socio-Economic Indexes for Areas (SEIFA) in line with previous studies

(Knibbs and Barnett, 2015; Chakraborty and Green, 2014). SEIFA is an index developed by the

Australian Bureau of Statistics (ABS) to score areas (e.g. SA1 units) according to their relative

socio-economic disadvantage and/or advantage (ABS, 2018a). The index consists of four sub-

indices created from weighted social and economic data, gathered from the five-yearly Australian

Census (ABS, 2018a). The four indices in SEIFA are: Index of Relative Socio-Economic

73

Disadvantage (IRSD); Index of Relative Socio-Economic Advantage and Disadvantage (IRSAD);

Index of Economic Resources (IER); and Index of Education and Occupation (IEO). SEIFA is

particularly useful for Australian studies as it was specifically designed to capture those variables

which are most representative of Australian culture and diversity. These indices consist of values

which are standardised to a mean of 1000 and a standard deviation of 100 (ABS, 2018a).

In addition to SEIFA data, we have included include percent Indigenous population per spatial unit,

using data from the Australian Census (ABS, 2018b). Previous work on environmental justice in

Australia highlighted a correlation between Indigenous Australian populations and pollution (e.g.

Knibbs and Barnett, 2015). It is worth noting that the SEIFA 2006 IRSD sub-index included percent

of people identifying as Indigenous. However, this was removed for later iterations (SEIFA 2011,

2016). In sum, we analysed both socio-economic advantage/disadvantage via the SEIFA indices and

the percent Indigenous per spatial unit against the pollution data.

For each regression model we used TTP as a dependent variable and the socio-demographic

advantage and disadvantage indices plus percent Indigenous population as separate, independent

variables. We constructed each model to understand if any particular indicators of social advantage

or disadvantage can predict the levels of TPP, or have more or less explanatory power for variance

within the TTP.

Due to the truncation of the NPI pollution data (i.e. no values below the threshold for reporting to

the NPI) we have analysed the data using two separate regression methods. Initially we analysed the

data using a series of truncated regression models, using the truncreg R package (Croissant and

Zeileis, 2017). We created models for each of the five socio-demographic advantage and

disadvantage variables per year (four SEIFA sub-variables plus percent identification as

Indigenous), and estimated R2 values by squaring the correlations between TTP values and

predicted values (Croissant and Zeileis, 2017). Truncated regression models are useful to

accommodate for variables being truncated (in this case left truncation of TTPi < min(TTP)), and

address the bias introduced when using ordinary least squares regression, however they do have

limitations. For example, for truncated models the variance of outcome variables is reduced

compared to the non-truncated distribution. Also, the mean of the truncated variables will differ

from non-truncated variables.

74

Due to the limitations of truncated regression, we also constructed standard linear regression

models, using ad hoc imputation for TTPi < min(TTP) observations. This imputation method

simulates a complete, non-truncated dataset. Due to the nature of how the NPI pollution data are

only collected once they go over particular thresholds, the missing data are not random. We used

the R Multiple Imputation by Chained Equations (MICE) package to impute values (van Buuren

and Groothuis-Oudshoorn, 2011), using the predictive mean matching (PPM) method. For each

missing value PPM randomly fills in a value from the observed values using an observation which

has similar regression predicted values based on a simulated regression model (Heitjan and

Roderick, 1991). As such, data imputed with PPM closely maintain the distribution of the observed

values.

For each instance where TTP values were missing (i.e. TTP = 0 due to the reporting thresholds), we

generated multiple imputations (n = 50) (see supplementary file 2 for counts of ‘missing’ data per

year). We used the five socio-demographic advantage and disadvantage variables (i.e. four SEIFA

sub-variables plus percent identification as Indigenous) to generate imputed values. In effect, this

analysed the relationship between existing TTP data and socio-demographic data and used this to

provide an estimate of a likely value for the missing TTP data. The validity of the 50 imputed data

sets was assessed using means, confidence intervals (.95), and by visually comparing the

distribution of imputed and non-imputed data (see supplementary file 3 for validity figures).

Following confirmation of validity of the imputed data, analysis for a relationship between TTP and

socio-demographic variables was conducted. The 50 imputed data sets were pooled and analysed using

simple linear regression models, using log(TTP) and each socio-demographic advantage and

disadvantage variable. Note that the regression model for 2006 differs from the 2011 and 2016

models due to later removal of percept Indigenous from the IRSD sub-index of the SEIFA (see

section 2.4). As a result, 2011 and 2016 were analysed via multiple regression as they also included

the percent Indigenous population variable.

Results

We analysed total toxicity potentials (TTP) against socio-demographic measures of advantage and

disadvantage across Australia’s smallest available statistical unit (SA1). A spatial representation of

the distribution of TTP in Australia’s SA1 units is presented in Figure 2.

75

a)

b)

76

c)

Figure 2 Relative TTP 2006 CDs (a), 2011 SA1s (b), and 2016 SA1s (c). Each year is displayed

with each spatial unit’s TTP rating (normalised between 0 to 1, and using 10 natural breaks, for

readability). Melbourne and Sydney are inset to demonstrate the very small spatial units in high

population density areas.

3.1 Summary statistics

Table 1 summarises the descriptive statistics for all data used in the models. Note that distribution

of the TTP values are heavily left skewed in each year.

Table 1 Descriptive summary statistics for all variables. TTP = total toxicity potential; IRSAD =

Index of Relative Socio-Economic Advantage and Disadvantage; IRSD = Index of Relative Socio-

Economic Disadvantage; IER = Index of Economic Resources; IEO = Index of Education and

Occupation (IEO); Pct. Indig. = Percent Indigenous in the population.

Variable n Mean Std. dev. Min Max Std. err.

2006 TTP 37,457 134,341,185.28 6,417,649,383.35 0.00 498,070,637,541.00 33,159,615.88

IRSAD 37,457 1,000.01 100 474.00 1,306.00 0.52

IRSD 37,457 1,000.01 100 205.00 1,199.00 0.52

IER 37,457 1,000 100 272.00 1,276.00 0.52

IEO 37,457 1,000 100 597.00 1,365.00 0.52

Pct. Indig. 37,457 2.53 7.98 0.00 104.30 0.04

77

2011 TTP 46,743 60,463,819.19 2,443,508,383.30 0.00 194,048,594,161.00 11,302,006.23

IRSAD 46,743 999.66 99.92 298.00 1,246.00 0.46

IRSD 46,743 1,000.12 99.70 121.00 1,184.00 0.46

IER 46,743 999.68 98.52 283.00 1,290.00 0.46

IEO 46,743 1,000.58 101.04 528.00 1,289.00 0.47

Pct. Indig. 46,743 2.56 6.88 0.00 102.47 0.03

2016 TTP 55,014 23,309,284.45 841,280,692.70 0.00 73,741,562,651.50 3,586,776.46

IRSAD 55,014 999.99 100.01 188.00 1,186.00 0.43

IRSD 55,014 999.99 100.00 400.00 1,239.00 0.43

IER 55,014 1,000.04 99.95 245.00 1,281.00 0.43

IEO 55,014 999.99 99.96 523.00 1,283.00 0.43

Pct. Indig. 55,014 2.91 7.09 0.00 101.98 0.03

3.2 Proportions of variance explained and statistical significance

The statistical analysis we conducted to interrogate whether there is a relationship between TTP and

socio-demographic variables produced a number of models that we use to interpret the results. One

model was produced for each year with each socio-demographic variable and for both the truncated

and imputed analysis (Table 2). As such, thirty models were produced in this analysis.

Several models report statistically significant results (α = 0.05), however the estimated R2 sizes are

very small (R2 < 0.01) for all models (Sawilowsky, 2009). Socio-demographic advantage and

disadvantage categories accounted for a maximum of 0.11 percent of TTP variance under the

truncated models, and a maximum of 0.06 percent under the pooled imputed models. This indicates

that less than one percent of variance in the exposure of CD and SA1 units to industrial pollution

can be explained by differences in the socio-demographic variables we included in this analysis.

Based on this and following Sullivan and Feinn (2012) and Nakagawa and Cuthill (2007), although

some models yielded results that were by definition statistically significant, given the very small R2

values for those significant models we fail to reject the null hypothesis, which states the prevalence

of exposure to industrial pollution is not related to socio-economic status (H0).

Table 2 Proportions of TTP variance explained by socio-demographic advantage and

disadvantage for years 2006, 2011, and 2016 (α = 0.05). Note the 2006 SEIFA included percent

indigenous people as a variable within the IRSD index.

78

Truncated regression Pooled imputed data

regression

Year Variables Estimated R2

Truncated p-value

Estimated R2

Pooled p-value

2006 IRSAD 1.80E-06 0.94 7.20E-05 0.8

IRSD* 6.70E-06 0.88 4.70E-05 0.72

IER 1.20E-04 0.53 5.08E-05 0.95

IEO 4.50E-06 0.9 8.63E-05 0.66

Pct. Indig. 1.10E-03 0.04* 3.21E-04 0.06

2011 IRSAD 6.70E-05 0.63 2.40E-05 0.71

IRSD 1.05E-03 0.7 5.60E-04 0.88

IER 3.90E-05 0.71 5.80E-05 0.32

IEO 3.30E-07 0.97 2.29E-05 0.88

Pct. Indig. 6.90E-04 0.1 4.26E-04 0.03*

2016 IRSAD 2.00E-04 0.38 3.43E-05 0.34 IRSD 6.13E-05 0.62 1.31E-04 0.45 IER 7.60E-04 0.09 3.53E-05 0.4 IEO 9.80E-07 0.95 1.36E-05 0.89 Pct. Indig. 1.20E-03 0.02* 9.52E-05 0.07

*denotes statistically significant result (α = 0.05).

Discussion

With this study we aimed to understand if there are implications for environmental distributive

justice from industrial pollution in Australia. Based on previous studies we hypothesized that the

prevalence of exposure to industrial pollution would have a relationship with socio-economic

variables. However, our results show there is little evidence to support this relationship.

Only models which used percent of people identifying as Indigenous yielded any statistically

significant results. While any type of distributive justice issue that disproportionately burdens

Indigenous people is a critical issue, the analysis yielded very small estimated R2 values even at an

overly generous type I error threshold of 0.05. The largest of these estimated R2 values was 0.0011

(2011 IRSD), meaning that the proportion of people identifying as Indigenous explained only 0.1%

to the variability in the Total Toxicity Potential (TTP) of industrial pollution. Given the very small

effect size of this result, we do not consider this to be a finding meaningful to the question of

environmental distributive justice concerning industrial pollution in Australia. However, we note

that this finding may indicate that Indigenous Australians may be most prone to risk should the

spatial distribution of industrial pollution change in a way that concentrates pollution in higher-TPP

areas as reduces pollution in lower-TPP areas.

79

This suggests that indicators of socio-economic demographics (included in this study as the SEIFA

sub-indices and percent of people identifying as Indigenous) do not have a meaningful relationship

with the distribution of harmful industrial emissions. This finding held true for both analytical

approaches (truncation or imputation). Using the best available data, at the finest scale available, we

did not find sufficient evidence to suggest that the prevalence of exposure to industrial pollution is

related to socio-economic advantage and disadvantage. As such, and although we did find some

statically significant relationships, due to the very small effect sizes, we fail reject the null

hypothesis.

This finding is somewhat contrary to previous studies on the distributive patterns of pollution in

Australia (in particular, we did not find a large effect size on Indigenous Australians cf. Knibbs &

Barnett (2015)). However, this does not suggest that there are no pollution based inequities in

Australia, (e.g. our results are not incompatible with Knibbs and Barnett (2015), as they specifically

assessed ambient and poorly regulated air emissions of a specific substance (NO2). Furthermore,

our results do not suggest that there are no inequities in exposure to pollution from industrial

emissions sources. Case study based work in Australia (e.g. Higginbotham et al.(2010)) highlights

many situations where individual communities suffer a highly disproportional and unfair burden

from local industry. What our results do show is that in aggregate across Australia there is no

violation of the strict equality distributive justice principle for industrial pollution when examined

via quantitative spatial analysis.

These findings indicate that Australia is equitably governing the distribution of industrial pollution

when it comes to broad, national-level trends (e.g. Keywood et al., 2016). Either by regulatory

design or by chance of the history of industrial development and activity (or perhaps a combination

of both), there is no evidence of a pattern of disproportionate impacts on the most socio-

demographically disadvantaged groups; a key goal for environmental distributive justice efforts.

This is contrary to findings of a similarly-designed quantitative spatial study of pollution and socio-

demographic advantage and disadvantage, where analyses have found a clear trend of greater

impacts on the socially disadvantaged across different countries (e.g. Li et al., 2018; Cushing et al.,

2015; Stewart et al., 2014; Delpla et al., 2015; Fecht et al., 2015). Nevertheless, the regulatory

framework for industrial pollution in Australia is not without its limitations. The National Pollutant

Inventory (NPI) sits within a portfolio of policy instruments, nested beneath the newly developed

National Clean Air Agreement – a framework for cooperation between the Australian

80

Commonwealth and State Governments that is in phase 2 of development (2018-2020) (Australian

Government, 2015). As a key instrument for the regulation of industrial pollution, the NPI has been

critiqued for being an industry-friendly mechanism lacking stringency in order to encourage

industry buy-in and for failing to fulfil its legislated goals (Howes, 2001; Cooper et al., 2017). This

is important as special and business interests have affected the design of air pollution policy in the

past (Crepaz, 1995). Dobbie and Green (2015) also found that when the network of national and

state systems for regulating industrial pollution have succeeded in identifying breaches, few

prosecutions follow, suggesting deficiencies with enforcement. These perspectives highlight the

importance of interpreting our results correctly positioned within their policy setting and in the

context of qualitative case study examination of clear instances of inequitable distribution of the

impacts of industrial pollution (e.g. Dobbie and Green (2015), Cooper et al. (2018), and

Higginbotham et al. (2010). From this positioning, we report the undoubtedly positive finding that

in aggregate and at the national level there is no evidence of a violation of the strict equality

distributive justice principle for industrial pollution when examined via qualitative spatial analysis.

But this does not mean there are no instances of unjust distribution of environmental burdens in

specific cases, or in instances that are not defined by socio-demographic categories that describe

social disadvantage. Similarly, this does not suggest the Australian policy framework cannot be

improved, and we encourage ongoing and critical development of the National Clean Air

Agreement.

4.1 Study limitations

The aim of this study was to provide a thorough and replicable method with which to understand

the relationship between industrial pollution and socio-demographic advantage and disadvantage

across Australia. However, it does have several limitations which must be considered before using

our findings in a decision-making context. Most importantly the recognition that this study pertains

to risks to human health only (e.g. TTP is total toxicity potential), and does not make any

correlations with actual burdens of disease (such as work by Holniki et al. (2017)). In addition,

while we aimed to use the highest resolution, and most detailed data available, the data are not

without limitations. The list of 93 substances includes most harmful pollutants emitted within

Australia, the list is by no means exhaustive. Some substances reported to the NPI are not rated in

the HPI including highly toxic substances and as such cannot be weighted. Furthermore, the HPI

treats unknown values for cancer and noncancer impacts of a substance as 0, due to a lack of

knowledge on the specific human health impacts. However, some of these substances may actually

81

be carcinogenic or disease causing, as such our data likely underestimates the threat to human

health caused by industrial emissions in Australia.

The available data, too, may be insufficient. Dobbie and Green (2015) argue that records of

pollution developed via collection of time point data may obscure major spikes in pollution between

the points of reporting, and there are risks that self-reporting by industry bodies may undermine the

quality of available data (Howes, 2001). In lieu of agencies such as the ABS and the NPI producing

more detailed data, or independent data collection of emissions sources of every harmful substance

from a nationally representative sample, combined with a comprehensive update of the HPI, our

study (at time of completion) is likely the highest resolution and most detailed possible for

investigating this issue. Future research should examine methodological issues, such as the use of

different buffer sizes, spatial units and pollution dispersion pathways.

Conclusions

Our study aimed to determine prevalence of, and identify, any associations between exposure to

industrial pollution and socio-demographic advantage and disadvantage. We find no compelling

evidence of association. This result indicates that Australian industrial emissions conform to strict

equality distributive justice principles. Understanding whether this conformity is deliberate and by

design, or a positive accident is outside of the scope of this study, though we have discussed some

of the interactions with Australian pollution policy and highlighted contrasting and supporting

insights from the extant literature. Further research in this area should address this knowledge gap

and examine why our finding differs from similar studies in similar countries (e.g. other Western

liberal democracies) and what other nations can (and can not) learn from Australian industrial

pollution and planning regulations. Furthermore, if higher resolution or more detailed data become

available, this study ought to be replicated, addressing the data limitations.

Acknowledgements

We would like to extend thanks to the following people whose contributions made this project

possible. Staff from the Australian Bureau of Statistics for their patient and thorough advice. Staff

at the National Pollutant Inventory for their advice and knowledge around toxicity weighting data.

Dr. Phil Kokic, and Dr. Grace Chiu from the Australian National University for their kindness and

advice for statistical modelling of the data.

82

Author Contributions

G.A., designed and performed the study, analysed results. G.A. and R.C. wrote the manuscript. B.W

contributed to editing the manuscript.

83

Chapter 5 Large-scale environmental degradation results in

inequitable impacts to already impoverished communities: A case

study from the floating villages of Cambodia

Online publication: http://dx.doi.org/10.1007/s13280-018-1022-2

Full citation: Althor G, Mahood S, Witt B, Colvin R, Watson J. (2018) Large-scale environmental

degradation results in inequitable impacts to communities: a case study from the floating villages of

Cambodia. Ambio 47(7): 747-759.

Overview

This study was published in Ambio in 2018. This study is somewhat unique in this thesis as it is

based on a purely qualitative method. This work is a result of time that I spend in the in remote

Cambodian floating villages. During my time in Cambodia I was able to meet with local fishing

community members and representatives to understand their lived experiences in the context of a

rapidly changing environment, which they are highly dependant on for subsistence needs. This

study fits within a broader literature which has examined the bio-physical and social impacts on the

Tonle Sap Great Lake as a result of climate change, land degradation, and damming along the

Mekong.

Research question: How are the fishing communities of the Tonle Sap floating villages experiencing

major environmental changes?

Sub-questions:



This study specifically utilises my modified (from Rawl’s second principle of justice (Rawls,

1971)) distributive justice principle, as follows:



Highlights:

• Adds an updated and lived experience narrative to a well-established body of literature on

this topic.

84

• Demonstrates the injustice of how over 100,000 rural Cambodians are being impacted by the

degradation of natural capital from which they receive no benefit.

• Further highlights the applicability for distributive justice principles in investigating issues

of natural capital utilisation.

85

The following text appears exactly as published.

Abstract

Cambodian subsistence communities within the Tonle Sap Great Lake area rely on resource extraction from

the lake to meet livelihood needs. These fishing communities – many of which consist of dwellings floating

on the lake – face potentially profound livelihood challenges because of climate change and changing

hydrology due to dam construction for hydroelectricity within the Mekong Basin. We conducted interviews

across five village communities, with local subsistence fisher people in the Tonle Sap in 2015, and

used thematic analysis methods to reveal a fishery system that is undergoing rapid ecological

decline, with local fishing communities increasingly experiencing reductions in available fish

stocks. As a result, over 100,000 people living in these communities are experiencing a direct loss

of well-being and livelihood. We discuss these losses and consider their implications for the future

viability of Cambodian floating village communities.

Keywords

Climate change, distributive justice, human well-being, Mekong basin, subsistence livelihoods,

Tonle Sap Great Lake.

Introduction

Globally, significant changes to ecosystems and the services they provide to people are occurring.

The diffuse effects of climate change are altering entire natural systems such as hydrological

regimes, which is compounded by the impacts of land use change and development. A range of

diverse and complex factors drives these changes, including economic growth in the developed

world (Leal Filho et al., 2017), to gazettement of protected areas for nature conservation (Watson et

al., 2014), and natural resource development for poverty alleviation in the developing world

(Manorom et al., 2017).

Common to these drivers of major global change are distributive justice implications. Viewed from

an environmental distributive justice perspective, such changes commonly produce inequitable

outcomes for the world’s poorest people, particularly in least developed countries. Rarely are the

distributions of well-being benefits derived from ecosystem degradation proportionate to the

resultant burdens that the poorest people are forced to bear (Pauline Dube and Sivakumar, 2015).

Furthermore, there is strong evidence which suggests that by contributing to human well-being,

ecosystem services (that is those natural, biophysical processes people use to support their well-

86

being) can reduce poverty (MEA, 2005), as examined by Fisher et al. (2014) in their framework for

analysing ecosystem services and poverty alleviation (ESPA) . As such, not only are benefits and

burdens resulting from ecosystem exploitation inequitably distributed to poor communities, their

degradation also commonly results in a double inequity by reinforcing local poverty (Berbés-

Blázquez et al., 2016).

While the environmental justice literature has a distinct focus on the justice implications of global

climate change, there are far fewer studies examining the impacts of large-scale environmental

changes on small, subsistence communities. Here we aim to partially address this gap by examining

the state of environmental distributive justice within the Tonle Sap Great Lake (hereafter ‘Tonle

Sap’). To do so, we use Rawls’ (Rawls, 1971: , pp. 5–6) second principle of distributive justice

(Rawls, 1971: , pp. 5–6), applied through an environmental justice lens, that is: Social and

economic inequalities, resulting from the degradation of natural resources, are to be to the greatest

benefit of the least advantaged members of society.

Located in Cambodia, the Tonle Sap is a prime example of a highly productive, valuable and

vulnerable wetland. Globally, wetlands are a key ecosystem type that provide many essential

ecosystem services, such as supporting and regulating biodiversity habitat, provisioning food and

other material goods, and cultural services (UNESCO, 1994). Wetlands such as the Tonle Sap are

generally highly productive, and are among the most valuable ecosystems on the planet (Al-Obaid

et al., 2017). Due to being highly productive, wetlands around the world commonly support the

well-being of large groups of people via ecosystem services, and as such, the well-being of these

people are reliant on a healthy ecosystem (Fisher et al., 2014). However, wetlands are also

characterised by their high level of vulnerability to systemic changes.

The Tonle Sap is located in rural Cambodia, where people often live subsistence lifestyles, meaning

they are highly dependent on the direct provisioning of ecosystem services to meet their livelihood

needs (Keskinen, 2006). Over one million Cambodians directly depend on the Tonle Sap system,

where they fish and farm food resources for day to day consumption and livelihood generation

(Sokhem and Sunada, 2006; FAO, 2017). The lake is particularly critical for the livelihoods of over

100,000 people who live in ‘floating villages’; communities which consist of dwellings that

permanently float on the water of the Tonle Sap (Figure 1; (Keskinen, 2006)). The livelihoods of

people living within these communities are based predominantly on fishing activities (Keskinen,

2006). As such, fish extracted from the lake are the main source of protein and livelihood

87

generation for residents (CDRI, 2013; Ratner et al., 2017). Given this dependence on the lake’s

natural resources, floating village communities are particularly vulnerable to any environmental

changes that affect the fishery.

Unfortunately for these communities, environmental change in the Tonle Sap has been rapidly

increasing over recent decades. Some of these changes are driven by the activities of the large

human population living within the system (Keskinen, 2006), and the high sensitivity of fresh

wetlands to disturbances (Middleton and Souter, 2016). However, the most significant

environmental pressures are large-scale and external to the system. Specifically, climate change and

the modification of the Mekong River are the most substantial drivers of declines in the health of

the lake’s natural systems (Arias et al., 2014; Arias et al., 2012). While the activities which generate

these pressures (such as manufacturing and damming) are undertaken to produce economic growth

and enhance social well-being for some, it is likely that the environmental declines they also

generate result in inequitable outcomes for floating villages. While these communities feel the

effects of environmental changes most directly and profoundly, they receive few benefits from the

development driving such change. As such, our present research examines the environmental

distributive justice implications of major environmental changes on people living in the floating

villages of the Tonle Sap. We aim to contribute to an understanding of how environmental inequity

may affect subsistence communities.

88

Figure 1 Prek Toal, is one of many floating villages located within the Tonle Sap. Most float

on the water using barrels, or bundles of bamboo attached beneath the dwellings.

The biophysical dimensions of the Tonle Sap Great Lake

In order to understand the livelihoods of people living within the Tonle Sap Great Lake, it is

important to understand its unique and complex biophysical dimensions. The Tonle Sap is the

largest lake in South East Asia and is part of a complex hydrological system that includes the Tonle

river and the Mekong River which is the lake’s primary tributary (Kummu and Sarkkula, 2008).

The most distinguishing feature of the lake is its exceptional annual flood cycle. Each year, the

Tonle Sap floods and expands from ~2,500-3,000 km2 in area and an average ~1.5m in depth, up to

~10,000-16,000 km2 in area and an average ~10m in depth (Figure 2). This flood occurs each year,

between July and October, and is driven by the South East Asian monsoonal regime. During this

period the monsoon generates over 60% of the Mekong Basin’s total annual rainfall (Arias et al.,

2014). The rains drive a vast flow of water into the Mekong river, which in turn reverses the flow of

the Tonle river. During the flooding large amounts of nutrients are pumped into the lake, which is

the foundation of its highly diverse set of freshwater ecosystems and fish species (FAO, 2001;

Kummu et al., 2008b; Ratner et al., 2017; FAO, 2017). As such, the lake is characterised by its

fisheries, which are among the most productive in the world (Keskinen, 2006). The most recent

estimates of total annual catch are between 180,000 – 250,000 tonnes (Chea et al., 2016). Catch

from the Tonle Sap contributes around 50 percent of Cambodia’s total catch, and around 16 percent

of the national GDP (van Zalinge et al., 2000). Therefore the annual flood pulse is critical for the

health of the livelihoods of people living within it (Keskinen, 2006), and for the lake’s natural areas

(Arias et al., 2014).

89

Figure 2 The location and seasonal extents of the Tonle Sap Great Lake, tributaries, and key

conservation areas. Map created using ESRI ArcGIS Pro with data provided by Wildlife

Conservation Society, Cambodia (ESRI ArcGIS Pro, 2017).

The Tonle Sap’s fisheries are dependent on a sustainable ecological system (van Zalinge et al.,

2000). As such, the livelihoods of villagers living within the Tonle Sap are completely reliant on the

fitness of the entire system’s functionality, particularly its regular hydrological flood pulse system

(Arias et al., 2014). Furthermore, much of the lake’s biodiversity have life cycles that are

synchronised to the flood pulse (Arias et al., 2014). Within these different ecosystems that range

from floodplains to seasonally flooded grasslands, shrub-lands, and gallery forests (Middleton and

Souter, 2016), hundreds of fish, plant, bird, reptile, invertebrate, and mammal species occur, many

of which have high nutritional or conservation value (Roos et al., 2007; Arias et al., 2014; Campbell

et al., 2006). For a significant majority of Tonle Sap species the seasonally flooded ecosystems

function as key habitat, which makes them critical for the ecological health of the lake (Arias et al.,

2012).

90

Environmental Protection within the Tonle Sap

In addition to the complex biophysical dimensions of the Tonle Sap, and due to its important social

and ecological values the lake is also host to a diverse and intricate network of protected areas

(figure 2). Understanding these protected areas is important as the lake’s floating village

communities often occur within and around them. Among the largest protected areas is the Tonle

Sap Biosphere Reserve. Designated in 2001 by the United Nations Educational, Scientific and

Cultural Organization (UNESCO) the Tonle Sap Biosphere Reserve officially recognizes the lake’s

ecological significance (UNESCO, 2007). The biosphere reserve aims to promote the conservation

and sustainable use of the Tonle Sap’s natural resources, with a particular focus on the lake’s

flooded ecosystems (UNESCO, 2015). It hosts three of the lake’s most ecologically diverse areas,

which UNESCO designated as ‘core zones’: Prek Toal, Boeng Tonle Chhmar, and Stoeng Sen

(UNESCO, 2007). Core zones act as ‘strict protected areas’ where any exploitation of natural

resources is prohibited under Cambodian law. Surrounding the core zones are buffer and transition

zones, which allow increasing degrees of human activity. The transition zones are of particular

importance to many floating village communities as this is where people are permitted to engage in

economic activities, such as fishing. Overlapping the core zones at Prek Toal, and Boeng Chhmar

are two Ramsar designated sites (Ramsar, 2017). Ramsar designation highlights these sites as

wetlands of international ecological significance, but confers no legal protection. In addition to the

globally recognized conservation sites are a network of over 400 locally managed freshwater

community fisheries and fish sanctuaries (Sok et al., 2012). Finally, there are three concentric

ecological zones around the lake: the terrestrial zone where development activities are permitted;

the floodplain zone where only specified development activities are permitted, and; the aquatic zone

where habitat is strictly protected (Mak, 2015). Adding to the complexity of the protected areas is

the overlapping jurisdictional and governance structures of the Cambodian authorities responsible

for their management. For example, the biosphere reserve and Ramsar sites are managed by the

Ministry of Environment, whereas the Fisheries Administration of the Ministry of Agriculture,

Forestry and Fisheries is responsible for the management of the community fisheries and fish

sanctuaries. Furthermore, the Tonle Sap Authority is responsible for ensuring coordinated natural

resource management across the entire Tonle Sap within the three ecological zones. All together the

various protected areas and management authorities operating within the Tonle Sap make

understanding its management and protection challenging for local people (Gillespie, 2016).

91

Threats to the ecosystem and fishery

Regardless of the many layers of environmental protection within the Tonle Sap, it is still a system

under significant threat. Since the 1980s the Tonle Sap has undergone significant environmental

change due to clearing of vegetation, declines in water quality, and increased extraction of resources

from people (Cooperman et al., 2012; Arias et al., 2014; Ratner et al., 2017). Furthermore, the lake

is affected by major ongoing changes to its hydrology, flow regime, and flood pulse, which are due

to rapid development of the upper reaches of the Mekong River (Lamberts, 2008; Arias et al., 2012;

Ziv et al., 2012).

Countries in the upper region of the Mekong River are increasingly constructing hydropower dams

and reservoirs (Arias et al., 2012). For instance, there are more than 40 mainstream and tributary

dams, irrigation, and water supply projects planned for implementation across Lao PDR, Thailand,

China, Vietnam and Cambodia in the next 20 years (Mekong River Commission, 2016). The

Mekong infrastructure projects put stress on the entire socio-ecological Mekong system, and in

particular the Tonle Sap (Lauri et al., 2012). Of specific concern for the health of the Tonle Sap is

the alteration of flow regimes (Arias et al., 2014). Damming of the Mekong dampens the annual

flood pulse, which is vital for the health of the lake’s ecosystems (Arias et al., 2012). Predictive

modelling forecasts that dampening the flood pulse will create a twofold effect: in the dry season

the average water levels will undergo a permanent increase, while in in the wet season average

levels will be permanently decreased (Piman et al., 2013). This will result in the average,

permanently flooded area increasing between 17% and 40% (Keskinen et al., 2011). This is a

serious issue as previous research suggests that even small changes in seasonal water extent are

likely to compromise the lake’s productivity (Arias et al., 2012; Baron et al., 2002; Arias et al.,

2014). This is largely a result of the increase in permanently flooded areas dramatically reducing

seasonally flooded habitats, which are the most significant for species of high economic and

conservation values (Kummu et al., 2008a; Middleton and Souter, 2016; Cunningham et al., 2011).

Reductions in flooded habitats are likely to be rapid, as wetlands, in general, degrade much faster

than other ecosystems (MEA, 2000).

While modification of the Mekong is expected to be the primary driver of the Tonle Sap’s declining

health, climate change is also predicted to affect the system (Middleton and Souter, 2016; Arias et

al., 2014). Within the Tonle Sap, climate change is increasing both atmospheric and water

temperatures, the frequency of extreme weather events, and the variability of rainfall (Nuorteva et

al., 2010). In fresh water systems, climate change modifies hydrological patterns such as

92

precipitation, evaporation, and flooding, (Erwin, 2009; Lemieux et al., 2014). These modifications

result in a scarcity of fresh water, and hydrological systems become strained.

Previous research has modelled many scenarios for both the effects that alteration of the Mekong

system and climate change will have on the Tonle Sap ecosystems (Arias et al., 2014; Arias et al.,

2012; Keskinen et al., 2011; Kummu and Sarkkula, 2008; Lamberts, 2008). Although there is a

degree of variability between studies, all agree that large tracts of the lake’s seasonally flooded

habitats will be lost. This will result in widespread reduction in the lake’s productivity (Lamberts,

2008). For floating village communities, reductions in the lake’s productivity is a serious cause for

concern. A decrease in the productivity of the lake will result in a decline in the fisheries, which is

the only source of livelihood generation for most households (CDRI, 2013; van Zalinge et al.,

2000). While there is a body of research highlighting how ongoing anthropogenic change will

adversely affect the natural systems of the Tonle Sap (Arias et al., 2014; Arias et al., 2013; Arias et

al., 2012; Kummu et al., 2008b; Kummu and Sarkkula, 2008; Orr et al., 2012), comparatively little

research has been conducted to understand how these changes will affect the lives of the

communities living within the lake’s floating villages.

As such, our research question is: How are the fishing people of the Tonle Sap floating villages

experiencing major environmental changes? To achieve this we interviewed people in five Tonle

Sap floating village communities using thematic analysis methods to record: (1) people’s

experiences living in this complex natural system; (2) how they perceive ongoing environmental

changes in the system, and; (3) if and how they respond to these changes. By collecting and

analysing these experiences and perceptions our aim was to understand the observations of, and

responses by, communities to ongoing changes in the lake’s ecosystems.

Methods

Prek Toal Biosphere Reserve and surrounding communities

We undertook an interview-based study in a selection of communities to understand and record the

experiences and perceptions of people living in floating villages within the Tonle Sap. We elected

to visit the five communities surrounding the Prek Toal Core Area of the Tonle Sap Biosphere

Reserve (Figure 3), as these communities were recommended to us by Cambodia-based

conservation practitioners as representative of the demographic makeup and livelihood model of the

Tonle Sap (fishing people), and due their proximity to other areas of interest. These villages

(Thvang, Kampong Prahok, Anlong Ta Uor, Prek Toal, and Kbal Taol - henceforth Prek Toal

93

communities) are subsistence communities permanently floating within the Tonle Sap floodplain,

and are adjacent to several protected area and fishery management zones, and as such, are of

interest in understanding how environmental change affects people in the Tonle Sap. The Prek Toal

communities are adjacent to various Community Fisheries on one side, and to the former Fishing

Lot No. 2, now a Fish Sanctuary (no-take zone) that overlaps and surrounds the Prek Toal Core

Area of the Tonle Sap Biosphere Reserve (itself also the Prek Toal Ramsar site) within which there

is a Community Protected Area. Over 10,000 people make up these communities, and by

international standards many are living below the poverty line (Varis et al., 2006). This makes the

Prek Toal communities highly vulnerable to changes to the resources that provide for their

livelihoods. The five communities represent a relevant sample for understanding the complex trade-

offs and interactions between livelihoods, nature, resource management, and conservation in the

Tonle Sap.

94

Figure 3. Map of floating villages, and other features surrounding the Prek Toal Biosphere

Reserve. Note that the background blue area in the main map indicates that the entire area is

flooded in the wet season. Map created using ESRI ArcGIS Pro with data provided by Wildlife

Conservation Society, Cambodia (ESRI ArcGIS Pro, 2017).

Data Collection

We conducted semi-structured interviews with subsistence fisher people (n=19) using qualitative

interview methods (Guest et al., 2006). Before any recorded interviews took place, we met with the

chief of each of the villages, and the commune chief. Village chiefs are responsible for governance

and community leadership within their village. Communes are a multi-village Cambodian

95

administrative division. The commune chief we spoke to governs the Kaoh Chiveang commune,

which includes all five villages. There were several purposes for these meetings. First, we explained

the nature of our research, and importantly, asked for their permission and endorsement to conduct

research within their communities. Once permission was granted we asked the chiefs to assist us in

understanding the communities, and the issues people face, relevant to our research. As such, the

village chief interviews were vital for participant recruitment, and to design our interview questions

and interview guide (supplementary material 1).

With the assistance of the chiefs we were able to use purposive sampling, targeting those people

with fishing as their primary means of livelihood generation, who had long-term residence within

the community. This was decided as fishing is by far the most common and important livelihood

activity in the communities, and long-term residence would avoid some of the bias that occurs with

shifting baselines (Sáenz-Arroyo et al., 2005). Further participants were engaged using snowball

sampling (Lewis-Beck et al., 2004), by asking participants for recommendations following

interviews. Participant sample size (n=19) was established based on data saturation techniques

(Bernard and Gravlee, 2014; Bryman, 2012; Francis et al., 2010; Guest et al., 2006). We then

conducted interviews with all participants in their villages using participatory tools including

seasonal fishing charts (see supplementary materials 2), and participant-generated maps of the local

area and features (see supplementary materials 3). The creation and discussion of these charts were

critical as discussion tools during the interviews, however the sample size is too small for any

quantitative analyses. As some of the issues discussed in our interviews were potentially

contentious within the community or could place participants at risk, we collected no identifiable

information from participants. All discussions with chiefs and interviews with participants were

subject to ethics approval, granted by the University of Queensland (number 20150701). We

conducted all interviews in the Cambodian language, Khmer, with the assistance of a translator. All

interviews were recorded for later translation and transcription. It is important to note that as

interviews were translated, some meaning from either interviewer or participants may have been

unclear. In order to reduce this issue, we ensured interview prompts and probing questions were

simple and easy to communicate.

Data Analysis

The lead author undertook all analyses, including designing the strategy, developing the code book,

coding the data, and performing the thematic analyses. The data corpus consisted of (1) the

transcribed interviews, and participant-generated (2) fishing charts and (3) participant-generated

96

maps. Professional Cambodian translators were contracted to translate and transcribe audio data

from Khmer into English. The field translator conducted translations of fish charts and maps. Data

were not quantified due to the small sample size suited to qualitative analysis, and the semi-open

ended structure of the interviews. The purpose of the analysis was to explore participant responses,

in the context of our research aims. The interview data largely informed the analytical strategy, and

we used inductive analysis to explore and code participant responses.

Thematic identification for creation of our code book was used (see supplementary materials 4).

Codes were developed using the literature, and key words, trends, themes, and ideas within the data

corpus (as per Guest et al., 2012). Throughout the screening processes only data considered relevant

to the research aims were analysed. All data coding was undertaken using R v. 3.3.3 (R Core

Development Team, 2015), with qualitative data analysis package RQDA v. 0.2-8 (Huang, 2014).

Coding required three data screening phases (as per Frith and Gleeson, 2004):

1. An initial screening of data for the purposes of identifying key and sub-themes, and

organisation of themes into a draft code book.

2. A second screening of data to test the precision of, and to refine, the code book.

3. A final screening of data, where we re-coded all data using the final, refined code book.

Thematic analysis methods were used to identify, analyse, and describe thematic patterns within the

coded data sets (Guest et al., 2006). As the aim was to capture community perceptions and

experiences, we determined thematic prevalence based on the number of interviewees who

discussed any particular topic. Data were analysed and described using a narrative approach (Braun

and Clarke, 2006; Kitzinger and Willmott, 2002), whereby we extracted an understanding of

participants’ realities from the data.

Results

Our analysis revealed several key themes from the data: (1) fishing is increasingly critical to local

livelihoods, (2) people believe that the fishery is highly vulnerable to environmental change, and;

(3) people believe the fishery is undergoing rapid declines.

Fishing is critical to local livelihoods

All participants explained that the majority of floating village families catch fish as their primary

source of food and income. Some villagers sometimes buy fish, but most villagers generally catch

97

fish to meet their needs. For those who buy fish, buying is mainly restricted to the dry season, when

the catch is reduced.

There was little relevance of the fish species caught. Participants described how families will both

eat and sell any fish species that they catch. However, they tend to only eat the smaller fish, which

have less market value, whereas larger fish are sold. Some participants described how the majority

of the income derived from selling fish is spent on basic food, and in particular rice which is eaten

at each meal, year-round. Participants explained that the fishery changes from month to month, and

that different fish species occur at different times of the year, with the dry season being lengthened

and increasingly hot. As different species have different monetary values, this has the effect of

making family incomes seasonal.

Participants also discussed the local practice of aquaculture (fish and, very rarely, crocodile

farming). They explained that aquaculture is usually restricted to those in the community who are

relatively wealthy. This is due to the capital required to set up the fish farms, which require timber

and other resources which are scarce in the area. However, many participants also explained that eel

farming is a very common practice, as it does not need any capital. To raise caught eel people

simply use the hollows of bamboo stalks, submerged in water. The majority of fish produced in the

farms is sold to traders, with few farmed fish being eaten by farm owners. Many participants

explained that fish farmers will still catch some fish for consumption, but also for use as feed in the

farms. The majority of fish farmed in the region are Snakehead species (Channidae), meaning this

farming is technically illegal. However, participants spoke of how these species are of the highest

value for resale and as such, very desirable. There was no mention or evidence that these

restrictions are in any way enforced.

People believe the fishery is highly vulnerable to environmental change

When people where asked about when and where they go to fish, participants talked about how

villagers fish in all open aquatic and flooded habitats (grasslands, shrub lands and gallery forests) in

the area. However, access to different areas is seasonal, and can be restricted by Cambodian

authorities. Participants described differences in fishing between the wet and dry seasons. During

the dry season, villagers fish within and along the shore of the Tonle Sap, within and along the

shore of local streams and channels, and along the border of the protected areas. In the wet season,

people will fish anywhere that the water has risen. However, communities reduce fishing efforts in

the deepest areas of the lake during the wet season due to the danger of large waves and storms.

98

While most people told us that they can never fish in the protected areas in either season, some

stated that “a little poaching” occurs within the protected areas as enforcement is either lax or prone

to minor corruption.

We also spoke with participants about when and where fish are the most abundant, and where the

greatest catch volume occurs. Participants explained that the greatest catch occurs during the

transition between the end of the dry season and the beginning of the wet season, as the water levels

rise. Participants told us that fish are most abundant in and around the nearby protected areas, and

that the best catch occurs along the borders of the protected areas. Participants also told us that the

catch is better along the shores of the lake and local streams than in other areas. When asked why

they thought the protected areas have the most fish, people told us that it is due to government

prohibitions, and because armed guards patrol these areas. Furthermore, they spoke of how these

areas are most abundant as this is where fish migrate to during the dry season. Many participants

told us that the protected areas are important for the local fish, however some participants said that

they are dissatisfied with prohibitions and that they would prefer complete fishing access. We asked

participants about the effects of extreme weather on the communities, including drought, very high

floods, and storms. Participants explained how drought (such as in 2002 and 2015) has terrible

effects on communities. The majority of people described that 2015 was the hottest and driest year

in living memory, and how the water level increased slowly during the start of the wet season. They

also explained how the lake water can become polluted and ‘smell bad’ during drought. Many

people described how during a drought the fish catch is much smaller than the average year. They

attributed this to many of the dry season fishing areas, such as local streams, completely drying up

or becoming stagnant, which kills the fish. Furthermore, participants spoke of how high water

temperature in local streams causes a high rate of fish mortality. Fish that die in this manner are

rendered inedible. We asked participants about the effect that unusually high annual floods (such as

in 2011 and 2013) have on families in the villages. All participants stated that during these floods

the fish catch is greater, and many also told us that the size of fish is larger. A consistent comment

from many participants was ‘big flood, big fish’. Additionally, we asked participants about the

effect of the severe storms which occur in the area. Most participants told us that storms are not

very disruptive, but can result in a short-term suspension of fishing activities, and damage fishing

gear.

99

People believe the fishery is undergoing rapid declines

The participants spoke of perceptions about changes to the fishery in recent history (the last 10-20

years). There was a very strong consensus that the fishery has rapidly declined, both in terms of

catch size and individual fish size. Participants explained how every fish species they catch has

declined in abundance by anywhere between 30% and 90%, and that some species had disappeared

altogether. Many participants showed us that the size of some species had gone from the girth of

their forearm down to the girth of a few fingers. When prompted about why they thought that the

fishery had declined, participants gave a range of reasons although virtually all participants told us

that the main causes for declines in the fishery were due to an (unspecified) increase in the local

human population, resulting in overfishing, and that many villagers were now using modern fishing

gear, such as electric nets. Other less common reasons given for the declines were the increasing hot

weather, water pollution, and illegal fishing in restricted areas.

We prompted the villagers to discuss how they respond to the declines in the fishery, and how they

adapt to extreme weather such as drought and flood. Most people told us that the communities do

not respond or adapt at all. Due to the much-reduced catch, many families now need to supplement

their catch by buying fish to eat from local traders. However, often families cannot afford to buy

fish. A few participants reported that during drought they can go through periods where “people

almost have nothing to eat” as they have no alternate source of food or income. We were told that

during extreme conditions such as drought, even though people know there is little catch, most will

still try to undertake fishing activities. Few fisher people will change to alternative livelihood

activities. This was attributed to villagers not having any choice for alternative livelihood

generation in the floating villages. Some participants explained that families who can afford the

expense of buying food will leave the village for factory work in Thailand.

Discussion

The ongoing and projected changes in the Tonle Sap’s ecosystems are expected to increasingly

compromise the lake’s ecosystems (Arias et al., 2014). As such, it is very likely that the Tonle Sap

fisheries will become increasingly less productive. As a result, floating village communities will

suffer from significantly reduced livelihoods. This outcome clearly fails the requirements of

environmental distributive justice as the well-being outcomes of the developments driving these

changes are not equitably distributed to those most in need. Rather, our findings demonstrate that

floating village communities are the most likely people to suffer from ecosystem service declines

100

in the Tonle Sap. We demonstrate that people believe there are rapid changes and declines in the

ecological system. Furthermore, we show how floating villagers believe that the entire fishery has

declined rapidly over the past two decades, a phenomenon which is likely true given the consistent

reports in interviews, and supporting empirical studies (Arias et al., 2014; Ratner et al., 2017).

Interestingly, our results show how the effects of climate change and hydrological alteration, such

as reductions in seasonally flooded habitats (Arias et al., 2012), are to some extent invisible to

floating villagers. Rather, they largely attribute fishery declines to overpopulation and the increased

use of illegal fishing gear.

Importantly, our research also demonstrates how people living in floating villages have very little

capacity, opportunity or means to adapt to these events and changes in situ. While some families

can seek international factory work, or trade fish, the majority of people describe themselves as too

poor to afford the capital required for such activities. We also found that in many cases, people may

not attempt to adapt at all, rather they simply continue fishing in much declined waters. This finding

is similar to previous survey based research in Cambodia, where over 40% of participants stated

they responded to severe drought by either organising religious ceremonies, or planting crops as

usual (MoE, 2005). This lack of adaptive capacity in floating villages is very likely due to economic

poverty (Keskinen, 2006). Poverty has been shown in general to make societies highly vulnerable to

changes in livelihoods (Shepherd and Brunt, 2013). For floating village communities, a distinct lack

of alternative livelihood options further exacerbates this vulnerability. The vulnerability of floating

villages is a serious concern as communities with little capacity to adapt to threats can fall into

poverty traps, which if left unresolved can cause catastrophic social breakdown (Carpenter and

Brock, 2008). As such, if the Tonle Sap’s ecosystems continue to degrade alongside the conditions

of poverty suffered by floating villagers, it is unclear if and how these communities will be able to

survive into the future.

While the overall adaptive capacity of floating villages is low, the protected areas near these

villages may play a role in buffering their vulnerability to change. While the role of the protected

areas in conserving biodiversity is obvious, the co-benefits to communities can be less so (Adams et

al., 2004). By maintaining healthy habitat the protected areas also provide important ecosystem

services to nearby communities (Middleton and Souter, 2016). In the Tonle Sap, protected areas

offer sanctuary, and spawning grounds for the fish on which the communities depend. Furthermore,

participants described an effect very similar to the fishery spill over seen around marine protected

areas (Russ and Alcala, 2011). In well managed marine reserves, species thrive in the absence of

101

human pressure and as a result, fish can ‘spill over’ into the surrounding waters (Goñi et al., 2010).

However, the extent of such an effect in the Tonle Sap is poorly understood and should be the

subject of future research in this region. Regardless of the potential spill over, previous research

(and one participant who we interviewed) describe the protected areas as a burden on communities

(Gillespie, 2016). For local people, the complex nature of the over-lapping layers of protection and

regulation for the environment can be confusing. Gillespie (2016) reported that while many local

people value the services protected areas provide, their management was described as ‘clunky’ and

‘confusing’. This confusion is a concern, as local communities are those most affected (positively or

negatively) by the protected areas. This is also of concern from a biodiversity conservation

perspective, as a lack of buy-in from local communities have been shown to potentially undermine

the effectiveness of protected areas (Beger et al., 2004). As such, there is a need for an improved

understanding and communication of human and biodiversity conservation relationships in the

Tonle Sap. This presents an important research need which may identify opportunities for attaining

benefits for both people and nature in Cambodia.

Future studies assessing protected area spill over, or the human-nature relationships in the Tonle

Sap would benefit from improved data collection methods. In our study, the use of translation from

Khmer to English to conduct interviews is a time consuming and exhausting task which introduced

the potential for issues of miscommunication and misunderstanding. We therefore recommend that

native Khmer speakers undertake future interview based research with floating village communities.

While the communities and the biophysical nature of the Tonle Sap are highly unique, the lack of

distributive justice is not. Globally, these wetland ecosystems provide the majority of resources

required by local people to support their livelihoods and well-being. However, many (if not all)

major wetland systems are in a state of decline due to large scale anthropogenic changes, which

result in the unfair distribution of burdens to local communities. Recent examples include

subsistence or poor communities impacted by the Belo Monte dam in the Amazon Basin (Abers et

al., 2017), and the Grand Ethiopian Renaissance Dam on the Blue Nile River (Nalepa et al., 2017).

Even within the Mekong the situation in the Tonle Sap is not unique, with agrarian communities

throughout the Mekong suffering similar injustices (Tilt and Gerkey, 2016). As such, understanding

and addressing issues of distributive justice in wetland ecosystems is a problem of global concern.

The drivers of large scale environmental change in the world’s wetlands are happening at (regional

and global) scales which are beyond the ability of local communities to affect. As such, the

environmental degradation of wetlands is also an issue of social equity and distributive justice. The

people, businesses, and nations who are receiving the benefits from degrading these systems (i.e.

102

gross domestic product growth from producing greenhouse gas emissions, or the sale of hydro-

power) are forcing others, such as the Tonle Sap floating village communities, to bear the costs.

However, there are efforts being made to address these inequities. The Mekong River Commission

(the inter-governmental organisation which manages the development of the Mekong River) state

they will make up for the losses and spread the benefits expected from hydropower development to

small communities via benefit sharing mechanisms (MRCS, 2014). Likewise, climate adaptation

funds, such as the Adaptation Fund, the Green Climate Fund, and the Special Climate Change Fund

aim to commit extensive resources to helping developing countries adapt to climate change.

However, there is no clear indication if (or how) any such funds may or may not be distributed to

specific communities (such as floating villages within the Tonle Sap). The Mekong River

Commission documentation related to benefit sharing offer little to no detail regarding beneficiaries

or actual distribution mechanisms, and this was reflected in our interviews with participants making

no mention of any such arrangements. Similarly, the climate adaptation funds offer little detail on

actual fund distribution mechanisms and are struggling to gain financial commitment from wealthy

nations (Kumar, 2015). As such, there is a critical need for clear and detailed mechanisms for

benefit sharing to Tonle Sap communities. These mechanisms urgently need to be formed and

implemented before livelihoods in the lake become further compromised. The parties driving

environmental degradation in the Tonle Sap have a clear moral obligation to share the benefits of

their activities with those bearing the costs. These issues also have implications at a broader scale

as these issues are hardly unique to the Tonle Sap. Previous research shows that inequitable impacts

of large-scale environmental degradation can be found the world over (Althor et al., 2016b), and

much like the case of the Tonle Sap, the main beneficiaries of environmental exploitation elsewhere

pay relatively few costs (O'Faircheallaigh, 2015; Ehrlich et al., 2012).

Conclusion

Globally, large scale environmental changes threaten the viability of subsistence communities.

Development along the Mekong basin will permanently change the ecological structure of the Tonle

Sap. Global climate change will increase temperatures within the lake, and increase the incidence of

extreme weather events. The development of dams on the Mekong River will affect the Tonle Sap’s

hydrology, leading to poor consequences for sustainability of the fishery. Furthermore, the effects

of Mekong basin development and climate change on the Tonle Sap are likely to intensify each

other. As a result, the lake’s ecosystems will become increasingly stressed, and some habitats such

as the lake’s seasonally flooded habitats will potentially disappear altogether. Permanent loss of

103

habitat in the Tonle Sap could lead to a collapse of the lake’s (already) stressed fisheries. If the

lake’s fisheries fail, life in floating villages will become even more difficult, if not impossible. As

such, climate change and the development of the Mekong basin will dramatically, and potentially

catastrophically, affect these people. Furthermore, given the low adaptive capacity of these

communities, it is unclear how they might respond or adapt to a loss of their main livelihood source.

Efforts to secure a sustainable future are needed to correct the inequities of the manifold pressures –

climate change, Mekong basin development, fishery decline, population growth – affecting the

people of the Cambodian Tonle Sap Great Lake.

104

Chapter 6 Thesis summary and conclusion

In this chapter I summarise how each research chapter of this thesis has overall contributed to how

distributive justice theories can be used to evaluate fairness for issues of natural capital distribution

while also producing rigorous scientific analyses of contemporary issues. In addition, I outline the

advantages and limitations, both broad and specific, of this approach, and provide suggestions for

how this work may be extended in the future.

Overall, throughout this thesis I have first argued then empirically demonstrated how distributive

justice principles can be adapted and modified to inform, develop, and understand empirical

research. More specifically, I have shown how research of this nature can be undertaken to examine

and understand natural capital distribution case studies through a justice lens.

A key foundation for understanding the quality and quantity of evidence regarding any particular

academic topic is the systematic review. I firmly believe that (high quality and rigorous) systematic

reviews are vital to gain an understanding of the whats, whens, wheres, and whos of particular topic

areas. Through these wonderful tools, we can systematically assess our fields of interest and

understand key unanswered questions, and how existing evidence paints the known truth of the

world. In chapter 2 I have demonstrated how a theoretical framework can be used alongside

systematic review methods for study design and to analyse peer reviewed data specific to

distributive justice and the environment. Moreover, I have shown how the results of this kind of

review can be used to identify knowledge gaps, which can in turn be used to guide the direction the

researcher takes. In my case, this review guided me toward the topics and cases investigated with

empirical methods in chapters 3, 4, and 5.

Issues of just or fair natural capital distribution can be empirically examined at the widest possible

scales, that is between nation states at the global level. While studies at this level do not allow for

the nuance of on-ground and inter-personal methods, they are nonetheless vital for the well-being of

many of the world’s people. Global issues which can and are affecting all of Earth’s human and

non-human inhabitants, generally require a global response. This response manifests in global

treaties, conventions, and other such political mechanisms. In chapter 3 I have shown how climate

change, perhaps the most concerning issue our species has faced in written history, is demonstrably

unjust. Essentially only a few of the world’s nations are driving the climate change phenomenon in

order to grow their economies, and in doing so are forcing unprecedented environmental change on

105

the majority. This is clearly a repugnant situation, which must be quantified through studies such as

that which I undertook in order to inarguably bring such injustices to light. From this perspective, I

would argue that such studies can be seen as providing essential empirical tools for the burdened to

call out and argue against the unfair cost and benefit distributions of natural capital exploitation

during global negotiations such as the UN Climate Change Conferences.

Logically, distributive justice issues also occur at the national scale, and can be examined as such.

In the absence of any truly egalitarian society, most if not all nations are characterised by some

degree of unfair natural capital distribution. This is exemplified in chapter 2, the systematic review,

where I have shown that the vast majority of studies in this area have concluded that social

characteristics such as race and ethnicity often have a deterministic relationship with poor

environmental quality. Such work is critical, as it can give voice to issues faced by groups that often

also tend to have the least power and self-determination within a nation. Where there is data

available, there is a way to use distributive justice theories to critically examine these issues. In

chapter 4 I demonstrated how to examine the potential issue of unjust distributions of industrial

pollution across an entire nation, even within serious data quality constraints. As such, this method

can be replicated elsewhere to examine similar issues. The development and application of such

methods is critical to produce an evidence base upon which policy design and analysis can be

conducted, particularly where distributive justice is of concern. Furthermore, this chapter

demonstrates how important the ecological fallacy can be in studies based on spatial methods,

where disaggregating spatial units used in previous studies can lead to very different conclusions.

Additionally, cases of unjust or unfair natural capital distribution occur and can be examined at the

scale of impacts to individuals and their immediate communities. Such examinations are critical as

they convey the richness and nuance of real-world lived experiences that are missing from broad or

large scale quantitative analyses. While large scale analyses can quantify injustices, they are absent

the humanity of how real people experience, regard, and reflect on injustices. In the scientific realm,

it is only through these qualitative methods that we can gain some degree of understanding how, for

example, a person experiences living next to heavily polluting industry. Or, as I have examined in

chapter 5, how subsistence fisher folk in rural Cambodia are experiencing the literal evaporation of

their primary source of livelihood so that others, elsewhere in the world, might power their homes,

or contribute to climate change through unchecked consumerism and economic growth.

106

I believe that this thesis adds an important, and at times absent, moral dimension to rigorous,

empirical studies, which in turn adds weight to their importance. It is my intention that this thesis

might be used by other researchers to inform the theory underlying similar studies, and to also

provide examples for how distributive justice theories can be utilised in empirical application. I

argue that this gives researchers a clear framework for discussing the results of such cases and

provides a baseline for drawing conclusions. As demonstrated, this process of theory adaptation and

study design can be applied to a wide range of problems, data types, and research methods.

Also importantly, throughout this thesis I have shown that it is not only appropriate for academic

research to be transparently normative in its execution, but also that this is unavoidable and as such

ought to be done in a transparent manner, which makes the researcher’s positionality obvious to the

reader. I believe that this is a critical for the interpretability of empirical research, and ought to be

undertaken when and where possible.

Modifying distributive justice principles

A key idea driving this thesis was my desire to transparently utilise philosophical theories in

combination with empirical research to evaluate real-world natural capital distribution issues. Very

early on, I came across the work on distributive justice by John Rawls (Rawls, 1971). His principles

fit very comfortably within the natural capital distribution context in which I was working. In

particular, I saw a lot of value in utilising part b of Rawls’ second principle of distributive justice:

“Social and economic inequalities are to satisfy two conditions: (a)… ; and (b), they are to be to

the greatest benefit of the least advantaged members of society”

As demonstrated in this thesis, and as argued elsewhere, the pragmatism and practicality of this

principle is well-suited to evaluating the way that natural capital goods and services are distributed

across different groups of people. Additionally, I have demonstrated the potential for the principle

of strict egalitarianism to be useful in such settings:

Every person should have the same level of material goods (including burdens) and services.

This principle, though, is highly unpragmatic when applied to general concepts of natural capital

distributions. As such, I have modified the principle to capture the idea that every person is of equal

107

moral worth and that harm to persons is universally immoral. Therefore, if we are to use natural

capital to benefit society in a manner which generates harm as a tolerable by-product of that benefit,

then harm ought strictly to be distributed equally across all beneficiaries.

To apply the work of Rawls and incorporate the principle of strict egalitarian in the context of harm,

I elected to modify these principles into distinct forms that are explicit in their relationship to the

distribution of natural capital goods and services, as follows:





P1 and P2 can be utilised separately, or combined, as follows:

P3: Where burdens upon any individual or exclusive group are the byproduct of the use of natural



These three principles make up the underlying theoretical approaches used throughout the thesis.

They were applied separately and in combination across research chapters.

Contribution of each research chapter

The following is a chapter by chapter breakdown of each study’s key research question and sub-

questions, a summary of how they utilised distributive justice principles, and finally the key

outcomes of each.

Chapter 2: A quantitative systematic review of distributive environmental justice literature; a rich

history, and the need for an enterprising future

108

Research question: What is the depth and breadth of the distributive justice and natural capital

literature?

Sub-questions:






In this Chapter I utilised all three modified principles (P1, P2, and P3) within a systematic review of

the distributive justice and natural capital literature. I used the principles to develop the study aims,

and methods such as the search strings, and inclusion/exclusion criteria. Specifically, P1, P2, and P3

were all used to develop the scope of the literature review, and were critical for the structure of the

introduction and arguing for the purpose and importance of the study. P1 and P2 were used to

develop the initial search string by using keywords such as “social inequality”, and “natural capital”

in the initial testing phase. Furthermore, I utilised P1 and P2 to develop the typologies used in the

review to ensure that all data I was capturing were related to the thesis principles.

Key outcome 1

With this study I provided a rigorously developed database of the peer-reviewed literature which

relates to distributive justice and natural capital. This database can be used to identify gaps in the

literature and to quickly and easily identify literature using cross tabulated sub-categories.

Key outcome 2

This literature is rapidly growing, and somewhat diverse, but is dominated by particular

geographies, methods, demographics, livelihood indicators, and topics. I recommend that a more

comprehensive approach is taken in the future, and suggest that the tabulated results of this review

can be used to identify these kinds of gaps.

Key outcome 3

Studies in this area are dominated by findings of distributive inequities (90 percent plus). However,

it is unclear if this is the true nature of natural capital distributions, or a result of biases such as

researcher and design bias, or the general bias in all sciences against the publication of negative

109

results. Based on this finding I recommend that future studies in this area apply the highest degree

of rigour possible, particularly in terms of controls.

Chapter 3: Global mismatch between greenhouse gas emissions and the burden of climate change

Research question: How are greenhouse gas emissions and the burdens of climate change

distributed globally?

Sub-questions:



In this Chapter I utilised my third modified distributive justice principle (P3) to empirically describe

and evaluate how the benefits and burdens associated with climate change are distributed.

Specifically, I utilised this principle in the design and results interpretation of this study. P3 posits

that cases where groups or individuals might be burdened by the use of natural capital by other

groups, these burdens ought to be accompanied by commensurate benefits (at a minimum). This is

clearly applicable to global climate change, where some few nations are improving their nation

economies through greenhouse gas emitting industries, while others are suffering from resulting

environmental burdens with few to no benefits.

Key outcome 1

Climate change inequity is globally pervasive, and correlated with economic output. The world’s

highest GHG emitting nations are driving climate change in order to grow their national economies,

but are (relatively) the least vulnerable (partially due to their large economies) to the negative

impacts of climate change. This is inversely true for low emitting nations, who are benefiting very

little from the processes driving climate change, but are inordinately suffering from their effects.

Key outcome 2

Climate change inequity is expected to worsen in the future. I project that by 2030 the distributions

of benefits and burdens will continue along a trajectory of inequity.

110

Key outcome 3

Global policy frameworks have been developed to address these kinds of inequities, but are grossly

inadequate, underfunded, and lack enforceability. As such, there are few incentives for the highest

emitting nations to curtail activities.

Chapter 4: The distribution of industrial pollution across socio-economic groups in Australia; a ten-

year, a fine scale distributive justice analysis

Research question: Are there social inequities in the distribution of exposure to harmful industrial

pollution across Australia?

Sub-questions:



In this Chapter I utilised my second modified distributive justice principle (P2) to empirically

describe and evaluate how harm from industrial pollution is distributed across social demographics

Australia wide. Specifically, I used P2 in the design and results interpretation of this study. P2

argues that cases where harm is accepted as a by-product of using natural capital to improve

society, then harms ought to be distributed equally. This principle is less generalisable than P1, but

does have a specific purpose for use in cases of harm distribution, such as the distribution of

industrial pollution. Industrial pollution (within regulated thresholds) is accepted as a by-product of

economic activities that is known to be harmful. As such, it provides a prime model for how P2 can

be used to examine the distribution of harm across social and demographic groups.

Key outcome 1

My results show there is little evidence aligned with previous studies which claimed that exposure

to industrial pollution has a clear relationship with socio-economic variables in aggregate. I found

some statistically significant relationships, however the effect sizes were too small to attribute any

causal relationship. I argue that this may be a result of overly large spatial units being used in

previous studies, and improperly applied toxicity weightings.

111

Key outcome 2

The findings may indicate that Indigenous Australians are the most at risk of exposure to industrial

pollution, but further research needs to clarify this relationship.


impoverished communities: A case study from the floating villages of Cambodia

Research question: How are the fishing communities of the Tonle Sap floating villages experiencing

major environmental changes?

Sub-questions:



In this chapter I utilised my first modified distributive justice principle (P1) using qualitative

research methods to empirically describe and evaluate the lived experiences of Cambodian

subsistence fisher people, in regards to how upstream environmental degradation is affecting their

well-being. Specifically, I used P1 in the design and results interpretation of this study. P1 argues

that social and economic inequalities, resulting from the utilisation of natural capital, are to be of

the greatest benefit of the least advantaged members of society. In this case there are gross

inequalities experienced by Tonle Sap fishing communities, resulting from Mekong damming and

climate change, however these communities are receiving little to no benefit from these activities.

As such, this chapter is exemplary of how to examine distributions of costs and benefits resulting

from natural capital exploitation.

Key outcome 1

I demonstrate that people living within the Tonle Sap Great Lake believe there are rapid changes

and declines in the ecological system, aligning with existing ecological and environmental science

research.

Key outcome 2

However, my results also show how the effects of climate change and hydrological alteration, such

as reductions in seasonally flooded habitats, are to some extent invisible to floating villagers.

112

Instead, they largely attribute fishery declines to overpopulation and the increased use of illegal

fishing gear.

Key outcome 3

I also demonstrate how people living in floating villages have very little capacity, opportunity or

means to adapt to a rapidly changing environment. For floating village communities, a distinct lack

of alternative livelihood options exacerbates their vulnerability to these changes. In most cases,

people are unlikely to attempt to adapt, but rather they will simply continue fishing in much

declined waters.

Advantages and limitations of the distributive justice approach, and future

research

This section contains the conclusions I have reached throughout my PhD about the advantages and

limitations of the distributive justice approach I have used (in its modified form), and future

research needs. As each research chapter discusses its own specific limitations and future research

needs, I have not replicated these here. Rather, this section pertains to the overall advantages,

limitations, and future research.

Advantages and limitations

The distributive justice approach provides a clear, practical, and easy to understand philosophical

framework with which to evaluate empirical research findings related to the fairness in how natural

capital is distributed. This is important for researchers interested in distributive justice issues, as it

not only provides a useable framework, it also allows them to be transparent in acknowledging their

positionality, and how this may affect their decisions and interpretations of empirical findings.

Furthermore, distributive justice studies have the advantage of simplicity, although this is also a

limitation (as outlined below). As I have demonstrated throughout this thesis, distributive justice

evaluations and their underlying moral arguments are intuitive, easy to understand, and

communicate. For example, practically anyone intuitively understands that it is unfair for some

nations to suffer from a global climate crisis they contributed very little to creating. The value of

this is manifold, not only does it provide a formalised and rational argument for intuitive ideas, by

reinforcing intuitive arguments it can be used to produce outputs including graphical figures that are

113

easily interpretable by a wide range of audiences, importantly including policy makers and the

general public.

Another key advantage is its scalability across geographies. With Chapters 3, 4, and 5. I have

demonstrated how these principles can be applied at a range of vastly different scales. Between

these chapters, I show how distributive justice principles can be applied anywhere from and

between the global scale (in Chapter 3), to the national scale (in Chapter 4), and all the way down to

evaluating the lived experiences of individuals in a local community setting (Chapter 5). Moreover,

this approach is highly useful as a guide in selecting and developing research methods. The

dimensions of distributive justice, such as costs and benefits, can be used by researchers and

practitioners in their selection of appropriate research methods. Throughout this thesis, I have

demonstrated the use of distributive justice principles in methodological design across a wide range

of method types. In Chapter 2 I used these principles to develop screening criteria for a systematic

literature review, in Chapters 3 and 4 I used them to develop statistical methods and modelling to

interpret quantitative data, and in Chapter 5 I demonstrated how they can be used to guide the

design and interpretation of qualitative research.

The most critical limitation of singularly distributive justice studies is their simplicity (although also

an advantage as discussed above), and lack of multidimensional justice perspectives. As discussed

throughout, due to this simplicity I caution against uses of distributive justice that treat it as an all-

encompassing evaluation of justice. There are many complex forms of justice which interact and

overlap to form a complete picture of what might universally be considered just or unjust.

Distributive justice is of most value and use if it is combined with other approaches to justice,

including but not limited to procedural, relational, restorative, or legal. While these approaches

were beyond the scope of the work contained in this thesis, it is critical that I acknowledge their

value alongside distributive approaches. More specific to this thesis, this work is also limited in the

scope of distributive justice forms examined. By basing this thesis almost solely on the principles

created by John Rawls, I have not rigorously examined all forms of distributive justice itself. While

Rawls is certainly the most prolific writer in this space, he is by no means the only writer. There are

several other forms of distributive justice which would likely have been just as legitimate for

developing my evaluations – e.g. the luck egalitarian approaches of Dworkin, Arneson, and Cohen,

or Nozick’s entitlement theory (Allingham, 2014).

114

Future research

As is likely obvious, this thesis is limited in its ability to holistically capture the ideas that it has

touched on. As discussed, the forms of justice examined herein are very limited, and are by no

means comprehensive. For future work undertaken in this in this area, rather than examine many

cases at different scales and using different methods, there would be great benefit in critically

examine a single case from a multi-dimensional and holistic justice approach.

As it stands, I have made many claims for what is fair, unfair, just, or unjust throughout, but they

are singularly limited to a distributive perspective, and as such cannot be called universally just or

not. Furthermore, and based on the findings of the systematic review outlined in Chapter 2,

distributive justice approaches clearly need to be more diverse, and more inclusive. For example, it

would be highly desirable to see more research dedicated to Indigenous people around the world,

LGBTQI+ people, and other under-represented minorities. Lastly, while I have argued that

distributive justice lends itself to an ease of understanding and real-world applicability (via its

simplicity), I have not empirically demonstrated this idea. It would be excellent for future work in

this area to address this issue by directly engaging with policy makers or science communicators

about problems they are facing, and work with them to see if this approach is truly as useful as I

have hypothesized.

Concluding remarks

I opened this thesis noting my own positionality as a researcher and reasserting the claim that every

human being has equal moral worth and that we must attach value to the well-being of each and

every person. My lived experiences, disciplinary background and training in environmental

management, and belief in equality guided my PhD research toward the questions of distributive

justice and natural capital examined in this thesis.

I modified principles of distributive justice and examined these across a range of studies and scales

via a diversity of methods. Through this I have come to learn, and my thesis has shown, that

distributive justice is flexible and practical for use in analysing empirical research related to the way

in which natural capital is distributed. I did not set out to capture a holistic perspective on justice as

related to natural capital, but instead sought a pragmatic and practical way to look at how benefits

and burdens are shared, and critically, the impact this has on the well-being of people.

Based on these studies, I cannot assert that the world is fair, nor is it unfair. Each study explored a

specific case of natural capital distributions and their social impacts. The nuances of each of these

115

studies are shaped by their own social and geographical contexts. Rather, and unsurprisingly, the

world is a complex place that cannot be universally essentialised. However, through this thesis I

have demonstrated that there is a practical value in applying singular dimensions of justice when

examining real world cases. Furthermore, I have met my personal goal of ensuring my work

transparently and honestly reflects my positionality as a researcher – a person who believes in a

universal moral framework for the equality of rights and fair treatment of all people.

116

Bibliography

Abers RN, Oliveira MSd and Pereira AK. (2017) Inclusive Development and the Asymmetric State:

Big Projects and Local Communities in the Brazilian Amazon. The Journal of Development

Studies 53: 857-872.

ABS. (2011) 1259.0.30.002 - Statistical Geography - Australian Standard Geographical

Classification (ASGC), Digital Boundaries, 2006. Available at:

http://www.abs.gov.au/AUSSTATS/[email protected]/Lookup/1259.0.30.002Main+Features12006?

OpenDocument.

ABS. (2017a) 1345.0 - Key Economic Indicators, 2018 Available at:

http://www.abs.gov.au/AUSSTATS/[email protected]/mf/1345.0.

ABS. (2017b) Australian Statistical Geography Standard (ASGS): Volume 1 - Main Structure and

Greater Capital City Statistical Areas. Available at:

http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/1270.0.55.001July%202016?Ope

nDocument.

ABS. (2018a) Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA),

Australia, 2011. Available at:

http://www.abs.gov.au/ausstats/[email protected]/DetailsPage/2033.0.55.0012011?OpenDocument.

ABS. (2018b) TableBuilder. Available at:

http://www.abs.gov.au/websitedbs/D3310114.nsf/Home/2016%20TableBuilder.

Adams WM, Aveling R, Brockington D, et al. (2004) Biodiversity conservation and the eradication

of poverty. Science 306: 1146-1149.

Adger WN. (2002) Inequality, Environment, and Planning. Environment and Planning A 34: 1716-

1719.

Ahnlid A. (1992) Free or Forced Riders?: Small States in the International Political Economy: The

Example of Sweden. Cooperation and Conflict 27: 241-276.

Al-Obaid S, Samraoui B, Thomas J, et al. (2017) An overview of wetlands of Saudi Arabia: Values,

threats, and perspectives. Ambio 46: 98-108.

Allingham M. (2014) Distributive Justice, Florence: Taylor and Francis.

Almaskut A, Farrell PJ and Krewski D. (2012) Statistical methods for estimating the environmental

burden of disease in Canada, with applications to mortality from fine particulate matter.

Environmetrics 23: 329-344.

Althor G, Watson J and Fuller R. (2016a) Global mismatch between greenhouse gas emissions and

the burden of climate change. Scientific Reports 6.

http://www.abs.gov.au/AUSSTATS/[email protected]/Lookup/1259.0.30.002Main+Features12006?OpenDocument

http://www.abs.gov.au/AUSSTATS/[email protected]/Lookup/1259.0.30.002Main+Features12006?OpenDocument

http://www.abs.gov.au/AUSSTATS/[email protected]/mf/1345.0

http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/1270.0.55.001July%202016?OpenDocument

http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/1270.0.55.001July%202016?OpenDocument

http://www.abs.gov.au/ausstats/[email protected]/DetailsPage/2033.0.55.0012011?OpenDocument

http://www.abs.gov.au/websitedbs/D3310114.nsf/Home/2016%20TableBuilder

117

Althor G, Watson JEM and Fuller RA. (2016b) Global mismatch between greenhouse gas

emissions and the burden of climate change. Scientific Reports 6: 20281.

Anderton DL, Anderson AB, Rossi PH, et al. (1994) Hazardous Waste Facilities:"Environmental

Equity" Issues in Metropolitan Areas. Evaluation Review 18: 123-140.

Arias ME, Cochrane TA and Elliott V. (2014) Modelling future changes of habitat and fauna in the

Tonle Sap wetland of the Mekong. Environmental Conservation 41: 165-175.

Arias ME, Cochrane TA, Norton D, et al. (2013) The flood pulse as the underlying driver of

vegetation in the largest wetland and fishery of the Mekong Basin. Ambio 42: 864-876.

Arias ME, Cochrane TA, Piman T, et al. (2012) Quantifying changes in flooding and habitats in the

Tonle Sap Lake (Cambodia) caused by water infrastructure development and climate change

in the Mekong Basin. J Environ Manage 112: 53-66.

Austin KF and McKinney LA. (2016) Disaster Devastation in Poor Nations: The Direct and Indirect

Effects of Gender Equality, Ecological Losses, and Development. Social Forces 95: 355-

380.

Australian Government. (2015) National Clean Air Agreement. In: Energy DoEa (ed). online:

Department of Environment and Energy.

Ayres CJ. (2012) The international trade in conflict minerals: coltan. Critical perspectives on

international business 8: 178-193.

Babidge S. (2016) Contested value and an ethics of resources: Water, mining and indigenous people

in the Atacama Desert, Chile. Australian Journal of Anthropology 27: 84-103.

Bagchi S and Svejnar J. (2015) Does wealth inequality matter for growth? The effect of billionaire

wealth, income distribution, and poverty. Journal of Comparative Economics 43: 505-530.

Banzhaf S, Ma L and Timmins C. (2019) Environmental Justice: The Economics of Race, Place,

and Pollution. The Journal of Economic Perspectives 33: 185-208.

Baron JS, Poff NL, Angermeier PL, et al. (2002) Meeting ecological and societal needs for

freshwater. Ecological Applications 12: 1247-1260.

Beger M, Harborne AR, Dacles TP, et al. (2004) A framework of lessons learned from community-

based marine reserves and its effectiveness in guiding a new coastal management initiative

in the Philippines. Environ Manage 34: 786-801.

Berbés-Blázquez M, González JA and Pascual U. (2016) Towards an ecosystem services approach

that addresses social power relations. Curr Opin Environ Sustain 19.

Bernard HR and Gravlee CC. (2014) Handbook of methods in cultural anthropology: Rowman &

Littlefield.

Betts R. (2008) Comparing apples with oranges. 7-8.

118

Borkan JM. (2004) Mixed Methods Studies: A Foundation for Primary Care Research. Annals of

Family Medicine 2: 4-6.

Bournay E. (2017) Typology of Hazards. Available at: http://www.grida.no/resources/7805.

Bowen W. (2002) An Analytical Review of Environmental Justice Research: What Do We Really

Know? Environmental Management 29: 3-15.

Bradley B. (2006) Two Concepts of Intrinsic Value. Ethical Theory and Moral Practice 9: 111-130.

Braun V and Clarke V. (2006) Using thematic analysis in psychology. Qualitative Research in

Psychology 3: 77-101.

Briggs RC and Weathers S. (2016) Gender and location in African politics scholarship: The other

white man's burden? African Affairs 115: 466-489.

Broome RA, Fann N, Cristina TJN, et al. (2015) The health benefits of reducing air pollution in

Sydney, Australia. Environmental Research 143: 19-25.

Brulle RJ and Pellow DN. (2006) Environmental justice: human health and environmental

inequalities. Annu. Rev. Public Health 27: 103-124.

Brunekreef B and Hoffmann B. (2016) Air pollution and heart disease. The Lancet 388: 640-642.

Bryman A. (2012) Social research methods, Oxford: Oxford university press.

Burkholder I and Edler L. (2014) Linear Model: Overview. Wiley StatsRef: Statistics Reference

Online: John Wiley & Sons, Ltd.

Campbell IC, Poole C, Giesen W, et al. (2006) Species diversity and ecology of Tonle Sap Great

Lake, Cambodia. Aquatic Sciences 68: 355-373.

Caney S. (2014) Climate change, intergenerational equity and the social discount rate. Politics

Philosophy & Economics 13: 320-342.

Carney N. (2016) All Lives Matter, but so Does Race: Black Lives Matter and the Evolving Role of

Social Media. Humanity & Society 40: 180-199.

Carpenter SR and Brock WA. (2008) Adaptive capacity and traps. Ecology and Society 13: online.

Cazorla M and Toman M. (2001) International Equity and Climate Change Policy. Climate Change

Economics and Policy: An RFF Anthology. Washington DC: RFF Press, 235.

CDRI. (2013) Climate Change Adaptation and Livelihoods in Inclusive Growth: A Review of

Climate Change Impacts and Adaptive Capacity in Cambodia. In: CDRI (ed). Phnom Penh:

CDRI.

CEE. (2013) Guidelines for Systematic Reviews in Environmental Management, Bangor, UK:

Collaboration for Environmental Evidence.

Chakraborty J and Green D. (2014) Australia’s first national level quantitative environmental

justice assessment of industrial air pollution. Environmental Research Letters 9: 044010.

http://www.grida.no/resources/7805

119

Charafeddine R and Boden LI. (2008) Does income inequality modify the association between air

pollution and health? Environmental Research 106: 81-88.

Chea R, Guo C, Grenouillet G, et al. (2016) Toward an ecological understanding of a flood-pulse

system lake in a tropical ecosystem: Food web structure and ecosystem health. Ecological

Modelling 323: 1-11.

Clark VLP and Ivankova NV. (2015) Mixed methods research: A guide to the field: Sage

Publications.

Cole DH. (2015) Advantages of a polycentric approach to climate change policy. Nature Clim.

Change 5: 114-118.

Cook KS and Hegtvedt KA. (1983) Distributive justice, equity, and equality. Ann Rev Sociol 9.

Cooper N, Green D and Meissner JK. (2017) The Australian National Pollutant Inventory Fails to

Fulfil Its Legislated Goals. International Journal of Environmental Research and Public

Health 14.

Cooper N, Green D, Sullivan M, et al. (2018) Environmental justice analyses may hide inequalities

in Indigenous people’s exposure to lead in Mount Isa, Queensland. Environmental Research

Letters 13: 084004.

Cooperman MS, So N, Arias M, et al. (2012) A watershed moment for the Mekong: newly

announced community use and conservation areas for the Tonle Sap Lake may boost

sustainability of the world’s largest inland fishery. Cambodian Journal of Natural History

2012: 101-106.

Copeland DC. (1996) Economic Interdependence and War: A Theory of Trade Expectations.

International Security 20: 5-41.

Coulter PB. (1989a) Measuring inequality: a methodological handbook, Boulder: Westview Press.

Coulter PB. (1989b) Measuring inequality: a methodological handbook, Boulder: Westview Press.

Crase L and Gillespie R. (2008) The impact of water quality and water level on the recreation

values of Lake Hume. Australasian Journal of Environmental Management 15: 21-29.

Crepaz MML. (1995) Explaining national variations of air pollution levels: Political institutions and

their impact on environmental policy‐making. Environmental Politics 4: 391-414.

Creswell JW. (2014) A concise introduction to mixed methods research: SAGE publications.

Crick B. (1987) Socialism, Milton Keynes: Open University Press.

Croissant Y and Zeileis A. (2017) truncreg: Truncated Gaussian Regression Models. Available at:

http://r-forge.r-project.org/projects/truncreg/, Version = 0.2-4.

http://r-forge.r-project.org/projects/truncreg/

120

Cummings S and Hoebink P. (2017) Representation of Academics from Developing Countries as

Authors and Editorial Board Members in Scientific Journals: Does this Matter to the Field

of Development Studies? The European Journal of Development Research 29: 369-383.

Cunningham SC, Thomson JR, Mac Nally R, et al. (2011) Groundwater change forecasts

widespread forest dieback across an extensive floodplain system. Freshwater Biology 56:

1494-1508.

Cushing L, Faust J, August LM, et al. (2015) Racial/Ethnic Disparities in Cumulative

Environmental Health Impacts in California: Evidence From a Statewide Environmental

Justice Screening Tool (CalEnviroScreen 1.1). Am J Public Health 105: 2341-2348.

DARA. (2012) Methodological Documentation for the Climate Vulnerability Monitor, Madrid:

DARA.

Davis AY. (1981) Women, race & class: New York : Vintage eBooks.

Delpla I, Benmarhnia T, Lebel A, et al. (2015) Investigating social inequalities in exposure to

drinking water contaminants in rural areas. Environmental Pollution 207: 88-96.

Dobbie B and Green D. (2015) Australians are not equally protected from industrial air pollution.

Environmental Research Letters 10.

Drabo A. (2011) Impact of Income Inequality on Health: Does Environment Quality Matter?

Environment and Planning A 43: 146-165.

Ehrlich PR, Kareiva PM and Daily GC. (2012) Securing natural capital and expanding equity to

rescale civilization. Nature 486: 68-73.

Erwin KL. (2009) Wetlands and global climate change: the role of wetland restoration in a changing

world. Wetlands Ecology and management 17: 71.

ESRI. (2011) ArcGIS Desktop: Release 10.5. Redlands, CA: Environmental Systems Research

Institute.

ESRI ArcGIS. (2011) Release 10. Redlands, CA: Environmental Systems Research Institute.

ESRI ArcGIS Pro. (2017) 1.4.1. Redlands, CA: Environmental Systems Research Institute.

Essoka JD. (2010) The Gentrifying Effects of Brownfields Redevelopment. Western Journal of

Black Studies 34: 299-315.

European Environment Agency. (2007) Europe's environment - The fourth assessment.

Copenhagen: European Environment Agency.

FAO. (2001) Tonle Sap Fisheries: A Case Study on Floodplain Gillnet Fisheries. In: Lamberts D

(ed). online: FAO Regional Office for Asia and the Pacific.

FAO. (2017) Fishery and Aquaculture Country Profiles: The Kingdom of Cambodia. Available at:

http://www.fao.org/fishery/facp/KHM/en.

http://www.fao.org/fishery/facp/KHM/en

121

Fecht D, Fischer P, Fortunato L, et al. (2015) Associations between air pollution and socioeconomic

characteristics, ethnicity and age profile of neighbourhoods in England and the Netherlands.

Environmental Pollution 198: 201-210.

Fekete H, Vieweg M, Rocha M, et al. (2013) Analysis of Current Greenhouse Gas Emission

Trends, Berlin: Climate Analytics.

Ferraro PJ. (2009) Counterfactual thinking and impact evaluation in environmental policy. New

Directions for Evaluation 2009: 75-84.

Findlay R and O’Rourke K. (2012) War, Trade, and Natural Resources: A Historical Perspective.

In: Garfinkel M and Skaperdas S (eds) The Oxford Handbook of the Economics of Peace

and Conflict. Oxford: Oxford University Press.

Fineman MA. (2014) Vulnerability, Resilience, and LGBT Youth. Temple Political & Civil Rights

Law Review 23: 307-330.

Fisher JA, Patenaude G, Giri K, et al. (2014) Understanding the relationships between ecosystem

services and poverty alleviation: A conceptual framework. Ecosystem Services 7: 34-45.

Francis JJ, Johnston M, Robertson C, et al. (2010) What is an adequate sample size?

Operationalising data saturation for theory-based interview studies. Psychology and Health

25: 1229-1245.

Freischlag JA and Faria P. (2018) It Is Time for Women (and Men) to Be Brave: A Consequence of

the #MeToo MovementPutting an End to Sexual Harassment in Health CarePutting an End

to Sexual Harassment in Health Care. JAMA 319: 1761-1762.

Frith H and Gleeson K. (2004) Clothing and embodiment: Men managing body image and

appearance. Psychology of men and masculinity 5: 40-48.

Füssel H-M. (2010) How inequitable is the global distribution of responsibility, capability, and

vulnerability to climate change: A comprehensive indicator-based assessment. Global

Environmental Change 20: 597-611.

Gerlagh R and Keyzer MA. (2001) Sustainability and the intergenerational distribution of natural

resource entitlements. Journal of Public Economics 79: 315-341.

Germani AR, Morone P and Testa G. (2014) Environmental justice and air pollution: A case study

on Italian provinces. Ecological Economics 106: 69-82.

Gichere SK, Olado G, Anyona DN, et al. (2013) Effects of drought and floods on crop and animal

losses and socio-economic status of households in the Lake Victoria Basin of Kenya.

Journal of Emerging Trends in Economics and Management Sciences 4: 31.

Gillespie J. (2016) Catch 22: wetlands protection and fishing for survival. Geographical Research

54: 336-347.

122

Gilmartin M. (2009) Colonialism/Imperialism Key Concepts in Political Geography. London:

SAGE Publications Ltd.

Goñi R, Hilborn R, Díaz D, et al. (2010) Net contribution of spillover from a marine reserve to

fishery catches. Marine Ecology Progress Series 400: 233-243.

Green Climate Fund. (2015) Background.

GRID Arendal. (2017) Typology of Hazards. Available at: www.grida.no/resources/7805.

Griffith University. (2019) Systematic Quantitative Literature Review. Available at:

https://www.griffith.edu.au/griffith-sciences/school-environment-

science/research/systematic-quantitative-literature-review.

Grim BJ and Finke R. (2007) Religious Persecution in Cross-National Context: Clashing

Civilizations or Regulated Religious Economies? American Sociological Review 72: 633-

658.

Gross C. (2008) A measure of fairness: An investigative framework to explore perceptions of

fairness and justice in real-life social conflict. Human Ecology Review 15: 130-140.

Guest G, Bunce A and Johnson L. (2006) How Many Interviews Are Enough?: An Experiment with

Data Saturation and Variability. Field Methods 18: 59-82.

Guest G, MacQueen K and Namey E. (2012) Themes and Codes. Applied Thematic Analysis.,

Thousand Oaks, CA: SAGE Publications, Inc.

Gusenbauer M and Haddaway N. (2019) Which Academic Search Systems are Suitable for

Systematic Reviews or Meta-Analyses? Evaluating Retrieval Qualities of Google Scholar,

PubMed and 26 other Resources. Research Synthesis Methods.

Hanigan IC, Cochrane T and Davey R. (2017) Impact of scale of aggregation on associations of

cardiovascular hospitalization and socio-economic disadvantage. PLoS ONE 12: e0188161.

Harrison K. (2012) A Tale of Two Taxes: The Fate of Environmental Tax Reform in Canada.

Review of Policy Research 29: 383-407.

Hayek F. (1976) Law, Legislation and Liberty, Volume 2: The Mirage of Social Justice, Chicago:

The University of Chicago Press.

Heitjan DF and Roderick JAL. (1991) Multiple Imputation for the Fatal Accident Reporting

System. Journal of the Royal Statistical Society. Series C (Applied Statistics) 40: 13-29.

Hertwich EG, Mateles SF, Pease WS, et al. (2001) Human toxicity potentials for life-cycle

assessment and toxics release inventory risk screening. Environ Toxicol Chem 20: 928-939.

Heyward M. (2007) Equity and international climate change negotiations: a matter of perspective.

Climate Policy 7: 518-534.

www.grida.no/resources/7805

https://www.griffith.edu.au/griffith-sciences/school-environment-science/research/systematic-quantitative-literature-review

https://www.griffith.edu.au/griffith-sciences/school-environment-science/research/systematic-quantitative-literature-review

123

Higginbotham N, Freeman S, Connor L, et al. (2010) Environmental injustice and air pollution in

coal affected communities, Hunter Valley, Australia. Health & Place 16: 259-266.

Holnicki P, Tainio M, Kałuszko A, et al. (2017) Burden of Mortality and Disease Attributable to

Multiple Air Pollutants in Warsaw, Poland. International Journal of Environmental

Research and Public Health 14: 1359.

Howes M. (2001) What's Your Poison? The Australian National Pollutant Inventory versus the US

Toxics Release Inventory. Australian Journal of Political Science 36: 529-552.

Huang R. (2014) RQDA: R-based Qualitative Data Analysis. 0.2-7 ed.

IPCC. (2007) Climate Change 2007: Impacts, Adaptation and Vulnerability : Working Group ii

Contribution to the Fourth Assessment Report of the Ipcc [M.L. Parry, O.F. Canziani, J.P.

Palutikof, P.J. Van Der Linden And C.E. Hanson (Eds)], Cambridge: Cambridge University

Press.

IPCC. (2014a) Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part A: Global and

Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the

Intergovernmental Panel on Climate Change [Field, C.B., V.R. Barros, D.J. Dokken, K.J.

Mach, M.D. Mastrandrea, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Genova,

B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L. White (eds.)],

Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press.

Ipcc. (2014b) Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional

Aspects. Contribution of Working Group II to the Fifth Assessment Report of the

Intergovernmental Panel on Climate Change [Barros, V.R., C.B. Field, D.J. Dokken, M.D.

Mastrandrea, K.J. Mach, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B.

Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L. White (eds.)],

Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press.

Jackson RB, Canadell JG, Le Quere C, et al. (2015) Reaching peak emissions. Available at:

http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate2892.html.

Jorgenson AK. (2014) Economic development and the carbon intensity of human well-being.

Nature Clim. Change 4: 186-189.

Kampa M and Castanas E. (2008) Human health effects of air pollution. Environmental Pollution

151: 362-367.

Kennedy M and Basu B. (2014) An analysis of the climate change architecture. Renewable and

Sustainable Energy Reviews 34: 185-193.

Kenny DC, Costanza R, Dowsley T, et al. (2019) Australia's Genuine Progress Indicator Revisited

(1962–2013). Ecological Economics 158: 1-10.

http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate2892.html

124

Keohane RO. (2015) The Global Politics of Climate Change: Challenge for Political Science. PS:

Political Science & Politics 48: 19-26.

Keskinen M. (2006) The Lake with Floating Villages: Socio-economic Analysis of the Tonle Sap

Lake. International Journal of Water Resources Development 22: 463-480.

Keskinen M, Kummu M, Salmivaara A, et al. (2011) Baseline results from hydrological and

livelihood analyses, Exploring Tonle Sap Futures. online: Aalto University.

Keywood M, Emmerson K and Hibberd M. (2016) Atmosphere: Australia state of the environment

2016. In: Energy AGDotEa (ed). Canberra: Australian Government Department of the

Environment and Energy.

Keywood M, Hibberd M and Emmerson K. (2017) Australia state of the environment 2016:

atmosphere, independent report to the Australian Government Minister for the Environment

and Energy. online: Commonwealth of Australia.

Kitzinger C and Willmott J. (2002) 'The thief of womanhood': women's experience of polycystic

ovarian syndrome. Soc Sci Med 54: 349-361.

Kjellstrom T, Kovats RS, Lloyd SJ, et al. (2009) The Direct Impact of Climate Change on Regional

Labor Productivity. Archives of Environmental & Occupational Health 64: 217-227.

Knibbs LD and Barnett AG. (2015) Assessing environmental inequalities in ambient air pollution

across urban Australia. Spat Spatiotemporal Epidemiol 13: 1-6.

Kumar S. (2015) Green Climate Fund faces slew of criticism. Nature 527: 419-420.

Kummu M, Keskinen M and Varis O. (2008a) Modern Myths of the Mekong: A critical review of

water and development concepts, principles and policies. In: Kummu M, Keskinen M and

Varis O (eds) Modern myths of the Mekong. Online: Water & Development Publications -

Helsinki University of Technology, 206.

Kummu M, Penny D, Sarkkula J, et al. (2008b) Sediment: curse or blessing for Tonle Sap Lake?

Ambio 37: 158-163.

Kummu M and Sarkkula J. (2008) Impact of the Mekong River flow alteration on the Tonle Sap

flood pulse. Ambio 37: 185-192.

Kyne D and Bolin B. (2016) Emerging Environmental Justice Issues in Nuclear Power and

Radioactive Contamination. International Journal of Environmental Research and Public

Health 13: 700.

Lamberts D. (2008) Little impact, much damage: The consequences of Mekong River flow

alterations for the tonle sap ecosystem. In: Kummu M (ed) Modern myths of the Mekong.

online: Water & Development Publications - Helsinki University of Technology, 3-18.

125

Lamont J and Favor C. (2017) Distributive Justice. In: Zalta EN (ed) The Stanford Encyclopedia of

Philosophy.

Lancaster HO. (1969) The chi-squared distribution, New York: Wiley.

Landrigan PJ, Fuller R, Acosta NJR, et al. (2017) The Lancet Commission on pollution and health.

The Lancet.

Lauri H, de Moel H, Ward PJ, et al. (2012) Future changes in Mekong River hydrology: impact of

climate change and reservoir operation on discharge. Hydrology and Earth System Sciences

16: 4603-4619.

Leal Filho W, Pociovalisteanu D-M and Al-Amin AQ. (2017) Sustainable Economic Development

Green Economy and Green Growth, Online: Springer International Publishing.

Lemieux CJ, Gray PA, Douglas AG, et al. (2014) From science to policy: The making of a

watershed-scale climate change adaptation strategy. Environmental Science & Policy 42:

123-137.

Lewis-Beck M, Bryman A and Futing Liao T. (2004) The SAGE Encyclopedia of Social Science

Research Methods, online: SAGE Publications Ltd.

Lewis S. (2017) Extreme Climate Change: Damage and Responsibility. AQ - Australian Quarterly

88: 3-8.

Li VOK, Han Y, Lam JCK, et al. (2018) Air pollution and environmental injustice: Are the socially

deprived exposed to more PM2.5 pollution in Hong Kong? Environmental Science & Policy

80: 53-61.

Maantay J and Maroko A. (2015) 'At-risk' places: inequities in the distribution of environmental

stressors and prescription rates of mental health medications in Glasgow, Scotland.

Environmental Research Letters 10.

MacCoun R and Perlmutter S. (2015) Blind analysis: Hide results to seek the truth. Nature 526:

187-189.

Mak S. (2015) The Governance of Wetlands in the Tonle Sap Lake, Cambodia. Journal of

Environmental Science and Engineering B 4: 331-346.

Manorom K, Baird IG and Shoemaker B. (2017) The World Bank, Hydropower-based Poverty

Alleviation and Indigenous Peoples: On-the-Ground Realities in the Xe Bang Fai River

Basin of Laos. Forum for Development Studies 44: 275-300.

McKinnon MC, Cheng SH, Dupre S, et al. (2016) What are the effects of nature conservation on

human well-being? A systematic map of empirical evidence from developing countries.

Environ Evid 5.

126

MEA. (2000) Ecosystems and human well-being wetlands and water synthesis. Online: Millennium

Ecosystem Assessment

MEA. (2005) Millennium Ecosystem Assessment, Ecosystems and human well-being, Washington,

D.C: Island Press.

Mekong River Commission. (2016) Integrated water resources management-based basin

development strategy for the Lower Mekong Basin 2016-2020. online: Mekong River

Commission.

Middleton BA and Souter NJ. (2016) Functional integrity of freshwater forested wetlands,

hydrologic alteration, and climate change. Ecosystem Health and Sustainability 2: e01200-

n/a.

Miguel E, Camerer C, Casey K, et al. (2014) Promoting Transparency in Social Science Research.

Science 343: 30-31.

MoE C. (2005) Vulnerability and Adaptation to Climate Hazards and to Climate Change: A Survey

of Rural Cambodian Households, Phnom Penh: Ministry of Environment, Cambodia.

Mohai P, Pellow D and Roberts JT. (2009) Environmental Justice. 34: 405-430.

Mohai P and Saha R. (2015) Which came first, people or pollution? A review of theory and

evidence from longitudinal environmental justice studies. Environmental Research Letters

10: 125011.

Montzka SA, Dlugokencky EJ and Butler JH. (2011) Non-CO2 greenhouse gases and climate

change. Nature 476: 43-50.

Moon K, Blackman DA, Adams VM, et al. (2019) Expanding the role of social science in

conservation through an engagement with philosophy, methodology, and methods. Methods

in Ecology and Evolution 0.

Moore RJH and Hotchkiss JL. (2016) The importance of toxicity in determining the impact of

hazardous air pollutants on the respiratory health of children in Tennessee. Environmental

Pollution 216: 616-623.

Morello-Frosch R and Jesdale BM. (2006) Separate and Unequal: Residential Segregation and

Estimated Cancer Risks Associated with Ambient Air Toxics in U.S. Metropolitan Areas.

Environmental Health Perspectives 114: 386-393.

MRCS. (2014) National-to-Local benefit sharing options for hydropower on Mekong tributaries

evaluated and reported by 2014. onine: Mekong River Commission Secretariat.

Nakagawa S and Cuthill IC. (2007) Effect size, confidence interval and statistical significance: a

practical guide for biologists. Biological Reviews 82: 591-605.

127

Nalepa RA, Short Gianotti AG and Bauer DM. (2017) Marginal land and the global land rush: A

spatial exploration of contested lands and state-directed development in contemporary

Ethiopia. Geoforum 82: 237-251.

Nozick R. (1974) Anarchy, State and Utopia, New York: Basic Books.

NPI. (2014) Substance list and thresholds. Available at:

http://www.npi.gov.au/substances/substance-list-and-thresholds.

NPI. (2015) National Pollutant Inventory Guide. online: National Pollutant Inventory.

Nuorteva P, Keskinen M and Varis O. (2010) Water, livelihoods and climate change adaptation in

the Tonle Sap Lake area, Cambodia: learning from the past to understand the future. Journal

of Water and Climate Change 1: 87-101.

Nurse LA, McLean RF, Agard J, et al. (2014) Small Islands. In: Barros VR, Field CB, Dokken DJ,

et al. (eds) Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional

Aspects. Contribution of Working Group II to the Fifth Assessment Report of the

Intergovernmental Panel of Climate Change. Cambridge, United Kingdom and New York,

NY, USA: Cambridge University Press, 1613-1654.

O'Faircheallaigh C. (2015) Social Equity and Large Mining Projects: Voluntary Industry Initiatives,

Public Regulation and Community Development Agreements. Journal of Business Ethics

132: 91-103.

O'Neill J, Tabish H, Welch V, et al. (2014) Applying an equity lens to interventions: using

PROGRESS ensures consideration of socially stratifying factors to illuminate inequities in

health. Journal of Clinical Epidemiology 67: 56-64.

O’Neill J, Tabish H, Welch V, et al. (2014) Applying an equity lens to interventions: using

PROGRESS ensures consideration of socially stratifying factors to illuminate inequities in

health. J Clin Epidemiol 67.

OHCHR. (2019) World Conference on Human Rights, 14-25 June 1993, Vienna, Austria. Available

at: https://www.ohchr.org/EN/ABOUTUS/Pages/ViennaWC.aspx.

Olsaretti S. (2018) Introduction: The Idea of Distributive Justice. In: Olsaretti S (ed) The Oxford

Handbook of Distributive Justice. Oxford: Oxford University Press.

Orr S, Pittock J, Chapagain A, et al. (2012) Dams on the Mekong River: Lost fish protein and the

implications for land and water resources. Global Environmental Change 22: 925-932.

Ostrom E. (2014) A polycentric approach for coping with climate change. Ann. Econ. Finance 15:

71-108.

Otto FEL, Frame DJ, Otto A, et al. (2015) Embracing uncertainty in climate change policy. Nature

Clim. Change 5: 917-920.

http://www.npi.gov.au/substances/substance-list-and-thresholds

https://www.ohchr.org/EN/ABOUTUS/Pages/ViennaWC.aspx

128

Padilla CM, Kihal-Talantikite W, Vieira VM, et al. (2014) Air quality and social deprivation in four

French metropolitan areas—A localized spatio-temporal environmental inequality analysis.

Environmental Research 134: 315-324.

Pauline Dube O and Sivakumar M. (2015) Global environmental change and vulnerability of Least

Developed Countries to extreme events: Editorial on the special issue. Weather and Climate

Extremes 7: 2-7.

Pearce DW and Turner RK. (1990) Economics of natural resources and the environment, Hemel

Hempstead: Harvester Wheatsheaf.

Pearce JR, Richardson EA, Mitchell RJ, et al. (2010) Environmental justice and health: the

implications of the socio-spatial distribution of multiple environmental deprivation for

health inequalities in the United Kingdom. Transactions of the Institute of British

Geographers 35: 522-539.

Perez VW and Egan J. (2016) Knowledge and Concern for Sea-Level Rise in an Urban

Environmental Justice Community. Sociological Forum 31: 885-907.

Pickering C, Grignon J, Steven R, et al. (2014) Publishing not perishing: how research students

transition from novice to knowledgeable using systematic quantitative literature reviews.

Studies in Higher Education: 1-14.

Pickering J, Jotzo F and Wood PJ. (2015) Splitting the difference: can limited coordination achieve

a fair distribution of the global climate financing effort? Environmental Politics 15: 4.

Piketty T. (2014) Capital in the twenty-first century, Cambridge Massachusetts: The Belknap Press

of Harvard University Press.

Piman T, Lennaerts T and Southalack P. (2013) Assessment of hydrological changes in the lower

Mekong Basin from Basin-Wide development scenarios. Hydrological Processes 27: 2115-

2125.

Pizer WA and Yates AJ. (2015) Terminating links between emission trading programs. Journal of

Environmental Economics and Management 71: 142-159.

R Core Development Team. (2015) R: A language and environment for statistical computing.

Vienna, Austria: R Foundation for Statistical Computing.

Ramsar. (2017) Cambodia Country Profile. Available at: http://www.ramsar.org/wetland/cambodia.

Rao ND. (2014) International and intranational equity in sharing climate change mitigation burdens.

International Environmental Agreements-Politics Law and Economics 14: 129-146.

Ratner BD, So S, Mam K, et al. (2017) Conflict and collective action in Tonle Sap fisheries:

adapting governance to support community livelihoods. Natural Resources Forum 41: 71-

82.

http://www.ramsar.org/wetland/cambodia

129

Rawls J. (1971) A theory of justice, Cambridge, Mass: Cambridge, Mass., Belknap Press of Harvard

University Press.

Rawls J. (1993) Political liberalism: New York : Columbia University Press.

Rawls J. (2001) Justice as fairness : a restatement, Cambridge: Harvard University Press.

Rivas I, Kumar P and Hagen-Zanker A. (2017) Exposure to air pollutants during commuting in

London: Are there inequalities among different socio-economic groups? Environment

International 101: 143-157.

Robeyns I. (2017) Wellbeing, freedom and social justice the capability approach re-examined,

Cambridge, UK: Open Book Publishers.

Rodriguez-Lara I. (2013) An Experimental Study of Gender Differences in Distributive Justice.

IDEAS Working Paper Series from RePEc.

Roos N, Chamnan C, Loeung D, et al. (2007) Freshwater fish as a dietary source of vitamin A in

Cambodia. Food Chemistry 103: 1104-1111.

Russ GR and Alcala AC. (2011) Enhanced biodiversity beyond marine reserve boundaries: the cup

spillith over. Ecol Appl 21: 241-250.

Sabel CF and Victor DG. (2017) Governing global problems under uncertainty: making bottom-up

climate policy work. Climatic Change 144: 15-27.

Sáenz-Arroyo A, Roberts C, Torre, J., , Cariño-Olvera M, et al. (2005) Rapidly Shifting

Environmental Baselines among Fishers of the Gulf of California. Proceedings: Biological

Sciences 272: 1957-1962.

Sawilowsky SS. (2009) New effect size rules of thumb. Journal of Modern Applied Statistical

Methods 8.

Scanlon T. (2018) Why Does Inequality Matter?: Oxford University Press.

Sen A. (1995) Inequality reexamined, Cambridge, Mass: Harvard University Press.

Sen A. (1999) Development as Freedom, London: Zed Books.

Sen A. (2009) The idea of justice, London ; New York: Allen Lane/Penguin Books.

Sheehan P, Cheng E, English A, et al. (2014) China's response to the air pollution shock. Nature

Clim. Change 4: 306-309.

Shehzad K, Qamer FM, Murthy MSR, et al. (2014) Deforestation trends and spatial modelling of its

drivers in the dry temperate forests of northern Pakistan — A case study of Chitral. Journal

of Mountain Science 11: 1192-1207.

Shepherd A and Brunt J. (2013) Chronic Poverty: Concepts, Causes and Policy, online: Palgrave

Macmillan UK.

130

Shortt NK, Richardson EA, Pearce J, et al. (2012) Mortality inequalities by environment type in

New Zealand. Health Place 18: 1132-1136.

Sok S, Yu X and Wong KK. (2012) Impediments to community fisheries management: Some

findings in Kompong Pou commune, Krakor District in Cambodia's Tonle Sap. Singapore

Journal of Tropical Geography 33: 398-413.

Sokhem P and Sunada K. (2006) The Governance of the Tonle Sap Lake, Cambodia: Integration of

Local, National and International Levels. International Journal of Water Resources

Development 22: 399-416.

Stewart IT, Bacon CM and Burke WD. (2014) The uneven distribution of environmental burdens

and benefits in Silicon Valley's backyard. Applied Geography 55: 266-277.

Su JG, Larson T, Gould T, et al. (2010) Transboundary air pollution and environmental justice:

Vancouver and Seattle compared. GeoJournal 75: 595-608.

Sullivan GM and Feinn R. (2012) Using Effect Size—or Why the P Value Is Not Enough. Journal

of Graduate Medical Education 4: 279-282.

Sumner A. (2016) The Geography of Poverty: How has global poverty changed since the end of the

Cold War? Global Poverty. Oxford: Oxford University Press.

Sun C, Kahn ME and Zheng SQ. (2017) Self-protection investment exacerbates air pollution

exposure inequality in urban China. Ecological Economics 131: 468-474.

Szabo S, Hajra R, Baschieri A, et al. (2016) Inequalities in Human Well-Being in the Urban Ganges

Brahmaputra Meghna Delta. Sustainability 8.

Temper L, Bene D and Martinez-Alier J. (2015) Mapping the frontiers and front lines of global

environmental justice: The EJAtlas. Journal of Political Ecology 22: 254-278.

Thomson Reuters. (2015) EndNote X7. New York, NY: Thomson Reuters.

Thomson Reuters. (2017) Web of Science. Available at: http://wokinfo.com/.

Tilt B and Gerkey D. (2016) Dams and population displacement on China’s Upper Mekong River:

Implications for social capital and social–ecological resilience. Global Environmental

Change 36: 153-162.

Tornblom KY. (1977) Distributive Justice: Typology and Propositions. Human Relations 30: 1-24.

Trenberth K. (2011) Changes in precipitation with climate change. Climate Research 47: 123-138.

UN General Assembly. (1948) Universal Declaration of Human Rights.

UNDP. (2017) Human Development Index. Available at: http://hdr.undp.org/en/content/human-

development-index-hdi.

http://wokinfo.com/

http://hdr.undp.org/en/content/human-development-index-hdi

http://hdr.undp.org/en/content/human-development-index-hdi

131

UNESCO. (1994) Ramsar, Convention on Wetlands of International Importance especially as

Waterfowl Habitat. Online: United Nations Educational, Scientific and Cultural

Organization.

UNESCO. (2007) Biosphere Reserve Information, Cambodia, Tonle Sap. Available at:

http://www.unesco.org/mabdb/br/brdir/directory/biores.asp?code=CAM+01&mode=all.

UNESCO. (2015) Mab Strategy 2015-2025. online: United Nations Organization for Education,

Science and Culture.

UNFCCC. (2015a) Adoption of the Paris Agreement. Conference of the Parties to the United

Nations Framework Convention on Climate Change. Paris: UNFCCC.

UNFCCC. (2015b) INDCs as communicated by Parties. Available at:

http://www4.unfccc.int/submissions/indc/Submission%20Pages/submissions.aspx.

United Nations. (2017) The Sustainable Development Goals Report 2017. online: United Nations

Publications.

United Nations Framework Convention on Climate Change. (1992) Text of the Convention, Rio:

United Nations Framework Convention on Climate Change.

van Buuren S and Groothuis-Oudshoorn K. (2011) mice: Multivariate Imputation by Chained

Equations in R. 45: 1-67.

van Zalinge N, Loeung D, Pengbun N, et al. (2000) Mekong flood levels and Tonle Sap fish

catches. Meeting of the Department of Fisheries. Phnom Penh: Thai National Mekong

Committee Secretariat, 28.

Varis O, Kummu M, Keskinen M, et al. (2006) Tonle Sap Lake, Cambodia: Nature’s affluence

meets human poverty. In: Report UNHD (ed) 1 ed. online: United Nations Human

Development Report.

Verde M, Martinez-Carrion JM and Martinez-Soto AP. (2016) Biological Welfare and Inequality

During the Mining Boom: Rio Tinto, 1832-1935. Revista De Historia Industrial: 149-181.

Ward JD, Sutton PC, Werner AD, et al. (2016) Is Decoupling GDP Growth from Environmental

Impact Possible? PLoS ONE 11: e0164733.

Watson JEM, Dudley N, Segan DB, et al. (2014) The performance and potential of protected areas.

Nature 515: 67-73.

Widlansky MJ, Timmermann A, Stein K, et al. (2013) Changes in South Pacific rainfall bands in a

warming climate. Nature Clim. Change 3: 417-423.

Willcox S. (2016) Climate change inundation, self-determination, and atoll island states. Human

Rights Quarterly 38: 1022-1037.

World Bank Group. (2015) World DataBank.

http://www.unesco.org/mabdb/br/brdir/directory/biores.asp?code=CAM+01&mode=all

http://www4.unfccc.int/submissions/indc/Submission%20Pages/submissions.aspx

132

World Resources Institute. (2014) Climate Analysis Indicators Tool: WRI’s Climate Data Explorer.

Wu C-f, Wu S-y, Wu Y-H, et al. (2009) Cancer risk assessment of selected hazardous air pollutants

in Seattle. Environment International 35: 516-522.

Ziv G, Baran E, Nam S, et al. (2012) Trading-off fish biodiversity, food security, and hydropower

in the Mekong River Basin. Proceedings of the National Academy of Sciences 109: 5609-

5614.

133

Supplementary materials

Materials in this section are separated by chapter. Unless otherwise stated, they appear and are titled

either exactly as they appear in published works, or as they are expected to appear for forthcoming

publications. For example, Nature uses a style of Sx (where x equals material number) to represent

supplementary materials.

Chapter 2 Supplementary materials

Additional file 1 - List of key EDJ studies

This list of papers was identified prior to undertaking literature searches as key types of EDJ

studies. After each iteration of the search process was completed, we ensured that results included

these papers (along with all other search modifications outlined in the main manuscript Methods

section).

1. Althor, G, Watson, JEM & Fuller, RA 2016, 'Global mismatch between greenhouse gas

emissions and the burden of climate change', Scientific Reports, vol. 6.

2. Babidge, S. (2016), Contested value and an ethics of resources: Water, mining and

indigenous people in the Atacama Desert, Chile. Aust J Anthropol, 27: 84–103.

doi:10.1111/taja.12139

3. Cassidy, E., Judge, R., & Sommers, P. (2000). The Distribution of Environmental Justice: A

Comment. Social Science Quarterly, 81(3), 877-878.

4. Chakraborty, J & Green, D 2014, 'Australia’s first national level quantitative environmental

justice assessment of industrial air pollution', Environmental Research Letters, vol. 9, no. 4,

p. 044010.

5. Füssel, H-M 2010, 'How inequitable is the global distribution of responsibility, capability,

and vulnerability to climate change: A comprehensive indicator-based assessment', Global

Environmental Change, vol. 20, no. 4, pp. 597-611.

6. Ghosh, R, Lurmann, F, Perez, L, Penfold, B, Brandt, S, Wilson, J, Milet, M, Kunzli, N &

McConnell, R 2016, 'Near-Roadway Air Pollution and Coronary Heart Disease: Burden of

Disease and Potential Impact of a Greenhouse Gas Reduction Strategy in Southern

California', Environ Health Perspect, vol. 124, no. 2, pp. 193-200.

7. Gochfeld, M & Burger, J 2011, 'Disproportionate exposures in environmental justice and

other populations: the importance of outliers', Am J Public Health, vol. 101 Suppl 1, pp.

S53-63.

134

8. Knibbs, LD & Barnett, AG 2015, 'Assessing environmental inequalities in ambient air

pollution across urban Australia', Spat Spatiotemporal Epidemiol, vol. 13, pp. 1-6.

9. Lukasiewicz, A & Dare, M 2016, 'When private water rights become a public asset:

Stakeholder perspectives on the fairness of environmental water management', Journal of

Hydrology, vol. 536, no. Supplement C, pp. 183-91.

10. Nkomo, S & van der Zaag, P 2004, 'Equitable water allocation in a heavily committed

international catchment area: the case of the Komati Catchment', Physics and Chemistry of

the Earth, Parts A/B/C, vol. 29, no. 15, pp. 1309-17.

135

Additional file 2 - Search string development log

Each search was undertaken and subsequently modified as per the Methods section of the main

manuscript.

Web of science

Search #1

You searched for: (TS= (natur*) AND TS= (environment*) AND TS= ("distributive justice" OR

equity) NOT TS=(health)) AND LANGUAGE: (English)

Refined by: DOCUMENT TYPES: ( ARTICLE OR BOOK CHAPTER ) AND WEB OF

SCIENCE CATEGORIES: ( ENVIRONMENTAL STUDIES OR ENVIRONMENTAL

SCIENCES OR MULTIDISCIPLINARY SCIENCES OR GEOGRAPHY OR SOCIAL

SCIENCES INTERDISCIPLINARY OR WATER RESOURCES OR POLITICAL SCIENCE OR

AGRONOMY OR FORESTRY OR AGRICULTURE MULTIDISCIPLINARY OR FISHERIES

OR AGRICULTURAL ECONOMICS POLICY OR ETHICS OR SOCIAL SCIENCES

MATHEMATICAL METHODS )

Timespan: 1900-2017. Indexes: SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, BKCI-S,

BKCI-SSH, ESCI, CCR-EXPANDED, IC.

Search #2

Results: 258

You searched for: (TS= (natur*) AND TS= (environment*) AND TS= ("distributive justice" OR

equity) NOT TS=(health)) AND LANGUAGE: (English)

Refined by: DOCUMENT TYPES: ( ARTICLE OR BOOK CHAPTER OR REVIEW ) AND

[excluding] PUBLICATION YEARS: ( 1973 ) AND WEB OF SCIENCE CATEGORIES: (

ENVIRONMENTAL STUDIES OR ENVIRONMENTAL SCIENCES OR ECONOMICS OR

MATHEMATICS INTERDISCIPLINARY APPLICATIONS OR PLANNING DEVELOPMENT

OR GEOGRAPHY OR SOCIAL SCIENCES INTERDISCIPLINARY OR PSYCHOLOGY

SOCIAL OR PHILOSOPHY OR STATISTICS PROBABILITY OR SOCIOLOGY OR WATER

RESOURCES OR URBAN STUDIES OR DEMOGRAPHY OR ENERGY FUELS OR

POLITICAL SCIENCE OR FORESTRY OR AGRICULTURE MULTIDISCIPLINARY OR

FISHERIES OR ETHICS OR ETHNIC STUDIES OR SOCIAL ISSUES OR

MULTIDISCIPLINARY SCIENCES OR AREA STUDIES OR AGRICULTURAL ECONOMICS

POLICY OR SOCIAL SCIENCES MATHEMATICAL METHODS OR GEOGRAPHY

PHYSICAL ) AND DOCUMENT TYPES: ( ARTICLE OR BOOK CHAPTER OR REVIEW )

136

AND RESEARCH AREAS: ( BUSINESS ECONOMICS OR PUBLIC ADMINISTRATION OR

GEOGRAPHY OR MATHEMATICS OR MATHEMATICAL METHODS IN SOCIAL

SCIENCES OR SOCIOLOGY OR GOVERNMENT LAW OR WATER RESOURCES OR

URBAN STUDIES OR INTERNATIONAL RELATIONS OR AGRICULTURE OR ENERGY

FUELS OR FISHERIES OR SOCIAL SCIENCES OTHER TOPICS OR FORESTRY OR SOCIAL

ISSUES )

Timespan: All years. Indexes: SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, BKCI-S,

BKCI-SSH, ESCI, CCR-EXPANDED, IC

Search #3

Results: 606

You searched for: (TS= ("distributive justice" OR *equity) AND TS = ("climate change" OR

"ozone depletion" OR "ecosystem service*" OR Urbanization OR biodiversity OR desertification

OR "land degradation" OR "water pollution")) AND LANGUAGE: (English)

Refined by: WEB OF SCIENCE CATEGORIES: ( ENVIRONMENTAL STUDIES OR

ENVIRONMENTAL SCIENCES OR ECONOMICS OR GEOGRAPHY OR ENERGY FUELS

OR WATER RESOURCES OR FISHERIES OR URBAN STUDIES OR POLITICAL SCIENCE

OR SOCIAL ISSUES OR MULTIDISCIPLINARY SCIENCES OR FORESTRY OR ETHICS OR

INTERNATIONAL RELATIONS OR AGRICULTURE MULTIDISCIPLINARY OR SOCIAL

SCIENCES MATHEMATICAL METHODS OR AGRICULTURAL ECONOMICS POLICY OR

SOCIAL SCIENCES INTERDISCIPLINARY ) AND RESEARCH AREAS: ( BUSINESS

ECONOMICS OR SCIENCE TECHNOLOGY OTHER TOPICS OR PUBLIC

ADMINISTRATION OR GEOGRAPHY OR FISHERIES OR SOCIAL ISSUES OR WATER

RESOURCES OR GOVERNMENT LAW OR URBAN STUDIES OR SOCIAL SCIENCES

OTHER TOPICS OR AGRICULTURE OR FORESTRY OR INTERNATIONAL RELATIONS )

AND RESEARCH AREAS: ( BUSINESS ECONOMICS OR FISHERIES OR PUBLIC

ADMINISTRATION OR GEOGRAPHY OR WATER RESOURCES OR SOCIAL ISSUES OR

URBAN STUDIES OR SOCIAL SCIENCES OTHER TOPICS OR AGRICULTURE OR

FORESTRY OR INTERNATIONAL RELATIONS OR ENERGY FUELS )



Search #4

Results: 436

137

You searched for: (TI= ("distributive justice" OR equit* OR inequit* OR fair* OR justice) AND TS

= ("Climate change" OR "deforestation" OR "ozone depletion" OR "land degradation" OR

"Reclamation" OR "Mining" OR "air pollution" OR "water pollution" OR "land pollution" OR

"Fishing" OR "Water degradation") NOT TS=(conservation OR ecology OR biology OR species))

AND LANGUAGE: (English)

Refined by: WEB OF SCIENCE CATEGORIES: ( ENVIRONMENTAL STUDIES OR

AGRICULTURE MULTIDISCIPLINARY OR ENVIRONMENTAL SCIENCES OR POLITICAL

SCIENCE OR ECONOMICS OR METEOROLOGY ATMOSPHERIC SCIENCES OR ENERGY

FUELS OR SOCIAL SCIENCES MATHEMATICAL METHODS OR INTERNATIONAL

RELATIONS OR GEOSCIENCES MULTIDISCIPLINARY OR URBAN STUDIES OR

FISHERIES OR AGRICULTURAL ECONOMICS POLICY OR WATER RESOURCES OR

SOCIAL ISSUES OR MINING MINERAL PROCESSING ) AND DOCUMENT TYPES: (

ARTICLE OR REVIEW ) AND DOCUMENT TYPES: ( ARTICLE OR REVIEW ) AND

RESEARCH AREAS: ( ENVIRONMENTAL SCIENCES ECOLOGY OR SOCIAL ISSUES OR

GOVERNMENT LAW OR SOCIOLOGY OR DEMOGRAPHY OR BUSINESS ECONOMICS

OR METEOROLOGY ATMOSPHERIC SCIENCES OR AGRICULTURE OR GEOGRAPHY

OR SOCIAL SCIENCES OTHER TOPICS OR MATHEMATICAL METHODS IN SOCIAL

SCIENCES OR ENERGY FUELS OR INTERNATIONAL RELATIONS OR FISHERIES OR

URBAN STUDIES OR WATER RESOURCES OR MINING MINERAL PROCESSING )



Search #5

Results: 1,175

(TI= ( "distributive justice" OR equit* OR inequit* OR *fair* OR justice OR burden OR just OR

*equal* OR ethic*) AND TS = ( "Climate change" OR "deforestation" OR "ozone depletion" OR

"land degradation" OR "Reclamation" OR "Mining" OR “hunting” OR "air pollution" OR "water

pollution" OR "land pollution" OR "Fishing" OR "Water degradation") NOT TS=(conservation OR

ecology OR biology OR species)) AND LANGUAGE: (English)

Refined by: DOCUMENT TYPES: ( ARTICLE ) AND WEB OF SCIENCE CATEGORIES: (

ENVIRONMENTAL STUDIES OR AGRICULTURE MULTIDISCIPLINARY OR

ENVIRONMENTAL SCIENCES OR AREA STUDIES OR ECONOMICS OR ETHICS OR

GEOGRAPHY OR POLITICAL SCIENCE OR METEOROLOGY ATMOSPHERIC SCIENCES

OR LAW OR FORESTRY OR SOCIOLOGY OR ETHNIC STUDIES OR INTERNATIONAL

138

RELATIONS OR ENERGY FUELS OR SOCIAL SCIENCES MATHEMATICAL METHODS

OR MULTIDISCIPLINARY SCIENCES OR WATER RESOURCES OR WOMEN S STUDIES

OR SOCIAL SCIENCES INTERDISCIPLINARY OR FISHERIES OR DEMOGRAPHY OR

AGRICULTURAL ECONOMICS POLICY OR STATISTICS PROBABILITY OR

MATHEMATICS APPLIED OR SOCIAL ISSUES OR URBAN STUDIES ) AND RESEARCH

AREAS: ( ENVIRONMENTAL SCIENCES ECOLOGY OR AREA STUDIES OR BUSINESS

ECONOMICS OR GOVERNMENT LAW OR MINING MINERAL PROCESSING OR SOCIAL

SCIENCES OTHER TOPICS OR GEOGRAPHY OR METEOROLOGY ATMOSPHERIC

SCIENCES OR MATHEMATICAL METHODS IN SOCIAL SCIENCES OR SCIENCE

TECHNOLOGY OTHER TOPICS OR WOMEN S STUDIES OR SOCIOLOGY OR

INTERNATIONAL RELATIONS OR FISHERIES OR ENERGY FUELS OR DEMOGRAPHY

OR WATER RESOURCES OR FORESTRY OR AGRICULTURE OR ETHNIC STUDIES OR

SOCIAL ISSUES OR MATHEMATICS OR URBAN STUDIES )



Scopus

Search #1

Results: 691

TITLE ( "distributive justice" OR *equity OR fair* OR justice ) AND TITLE-ABS-KEY (

"Climate change" OR "deforestation" OR "ozone depletion" OR "land degradation" OR

"Reclamation" OR "Mining" OR "air pollution" OR "water pollution" OR "land pollution" OR

"Fishing" OR "Water degradation" ) AND NOT TITLE-ABS-KEY (conservation OR ecology

OR biology OR species ) AND PUBYEAR > 1980 AND ( LIMIT-TO ( DOCTYPE , "ar " )

OR LIMIT-TO ( DOCTYPE , "re " ) ) AND ( LIMIT-TO ( SUBJAREA , "SOCI " ) OR

LIMIT-TO ( SUBJAREA , "ENVI " ) OR LIMIT-TO ( SUBJAREA , "EART " ) OR LIMIT-

TO ( SUBJAREA , "ECON " ) OR LIMIT-TO ( SUBJAREA , "ENER " ) OR LIMIT-TO (

SUBJAREA , "BUSI " ) OR LIMIT-TO ( SUBJAREA , "AGRI " ) OR LIMIT-TO (

SUBJAREA , "MULT " ) OR EXCLUDE ( SUBJAREA , "BUSI " ) ) AND ( LIMIT-TO (

LANGUAGE , "English " ) )

Search #2

Results: 1,519

TITLE ("distributive justice" OR equit* OR inequit* OR *fair* OR justice OR burden OR just OR

*equal* OR ethic*) AND TITLE-ABS-KEY ("Climate change" OR "deforestation" OR "ozone

139

depletion" OR "land degradation" OR "Reclamation" OR "Mining" OR "air pollution" OR "water

pollution" OR "land pollution" OR "Fishing" OR "Water degradation") AND NOT TITLE-ABS-

KEY (conservation OR ecology OR biology OR species) AND PUBYEAR > 1980 AND ( LIMIT-

TO ( SUBJAREA,"ENVI" ) OR LIMIT-TO ( SUBJAREA,"SOCI" ) OR LIMIT-TO (

SUBJAREA,"EART" ) OR LIMIT-TO ( SUBJAREA,"ARTS" ) OR LIMIT-TO (

SUBJAREA,"ECON" ) OR LIMIT-TO ( SUBJAREA,"AGRI" ) OR LIMIT-TO (

SUBJAREA,"ENER" ) OR LIMIT-TO ( SUBJAREA,"MATH" ) OR LIMIT-TO (

SUBJAREA,"MULT" ) OR LIMIT-TO ( SUBJAREA,"DECI" ) OR LIMIT-TO (

SUBJAREA,"Undefined" ) ) AND ( LIMIT-TO ( DOCTYPE,"ar" ) ) AND ( LIMIT-TO (

LANGUAGE,"English" ) )

Proquest

Search #1

217 results

ti("distributive justice" OR equity OR fair???? OR justice OR inequity) AND all("Climate change"

OR "deforestation" OR "ozone depletion" OR "land degradation" OR "Reclamation" OR "Mining"

OR "air pollution" OR "water pollution" OR "land pollution" OR "Fishing" OR "Water

degradation") NOT all(conservation OR ecology OR biology OR species)

Applied filters

Publication date: 1980 - 2017

Document type:

Article OR Review

Subject:

climate change OR studies OR environmental justice OR environmental policy OR air pollution OR

social justice OR emissions OR environmental protection OR global warming OR humans OR

socioeconomic factors OR greenhouse effect OR greenhouse gases OR justice OR politics OR

environmental economics OR sustainable development OR energy policy OR industrialized nations

OR developing countries--ldcs OR international law OR environmental health OR environmental

impact OR ethics OR human rights OR community OR economic models OR emissions trading OR

environment OR international agreements OR outdoor air quality OR data collection OR

environmental law OR international relations OR low income groups OR negotiations OR public

policy OR race OR research OR water pollution OR climate OR decision making

Search #2

Results: 763

140

ti("distributive justice" OR equit???? OR inequit???? OR fair???? OR unfair???? OR justice OR

burden OR just OR equal????? OR unequal???? OR inequal???? OR ethic????) AND all ("Climate

change" OR "deforestation" OR "ozone depletion" OR "land degradation" OR "Reclamation" OR

"Mining" OR "air pollution" OR "water pollution" OR "land pollution" OR "Fishing" OR "Water

degradation") NOT all (conservation OR ecology OR biology OR species)

Applied filters

Publication date: 1980-2019

Document type: Article

Language: English

Subject: No filter –subjects too granular on Proquest to match WOS or Scopus.

141

Additional file 3 – Variable typologies

These typologies were developed as per Methods in main manuscript. Note that we took variable

descriptions in studies at face value. For example, if an article included a variable named ‘Gender’

as a study demographic, we did not further assess if the authors may have defined gender in a

different manner to other articles.

Social demographic variables

This typology was developed using the social stratifications outlined by O'Neill et al. (2014) as a

baseline. We updated this list during the title and abstract screening phases to include any

demographic variables used in screened publications, which were not included in the baseline.

Social demographic variable Description Notes

Geographic

Article relates to, or studies

people who live within a

specific geographic area. E.g.

a particular neighborhood,

city, or nation.

Children


children.

No age range for definition of

children across articles was

assessed.

Age


people of specific age range.

May also include children.

Education


level of education.

Different levels of education

used in articles were not

assessed.

Employment


type of, or lack of

employment.

Different types of employment

were not assessed.

Farmers\Fishers\Hunters


Farmers\Fishers\Hunters.

Groups distinct but aggregated,

due to very low rates of

occurrence.

Future generations


any form of future

generations.

Can also include future

generations of other

demographic groups, such as

142

children.

Immigrants


immigrants from one nation to

another.

Can also include immigrants

from other demographic groups,

such as children.

Indigenous


people identified by authors as

Indigenous to the place of

study.

Marital status


the marital status of groups.

Multiple unspecified


aggregated social groups, but

does not explicitly describe

the groups within the

aggregations.

Other

Article relates to, or studies a

social demographic not

captured within this typology.

Large number of articles fit

under this category due to the

very diverse nature of studies.

Rural/Urban


groups which are specifically

rural, or rural/urban

differences.

Does not include solely urban

studies.

LGBTQ+


any LGBTQ+ groups or sub-

groups.

Any sub-LGBTQ+ groups (e.g.

gay or lesbian) aggregated due

to very low rates of occurrence.

Economic status


economic status of groups.

Can include income, relative or

absolute poverty etc.

Racial/ethnic


racial or ethnic groups.

Some studies may define

particular religious groups as

belonging to an ethnic group.

These delineations were not

examined.

Religious Article relates to, or studies Some studies may define

143

particular religious groups. particular religious groups as

belonging to an ethnic group. If

so, these studies would be

captured under the Racial/ethnic

category.

Health/Disability


groups based on physical

health, or groups with

disability.

Gender\sex


particular genders or sexes.

Human well-being outcome variables

This typology was developed using the domains and definitions of human well-being outcomes,

outlined by McKinnon et al. (2016) as a baseline. We updated and modified this list during the title

and abstract screening phases to include any well-being outcome variables used in screened

publications, which were not included in the baseline.

Human well-being

outcome variable

Description Notes

Education


outcomes that affect education

level, or access to education

institutions, of groups.

Governance


outcomes that affect access to

structures and processes for

decision making.

Health/Mortality


outcomes that affect physical

health, mortality, or access to

144

health institutions.

Economic livelihood


outcomes that affect

livelihoods, such as income and

employment.

Living standards


outcomes that affect living

standards, such as poverty.

Mental health


outcomes that affect mental

health.

Multiple HWB

outcomes


multiple outcomes.

Captured separately to identify

studies which may be assessing

the complex nature of multiple

outcomes and their interactions.

Not specified

Article does not relate to, study,

or specify any HWB outcomes.

Natural resource

rights/access


outcomes that affect the rights

of groups to, or to access,

particular natural resources.

Security/Safety


outcomes that affect physical

security, or safety.

Social wellness


outcomes that affect social

145

interactions, social capital, or

other social relationships.

Spiritual/Cultural

fulfilment


outcomes that affect spiritual or

cultural fulfilment.

Environmental exploitation activity variables

This typology was developed using the exploitation activities described in United Nations (2017),

Pearce et al. (2010), and European Environment Agency (2007) as a baseline. We updated this list

during the title and abstract screening phases to include any exploitation variables used in screened

publications, which were not included in the baseline.

Environmental

exploitation activity

variable

Description Notes

Air pollution Article relates to, or studies

air pollution.

Can include a high variety of

sources, such as traffic and

industry.

Climate change Article relates to, or studies

the caused or effects of

climate change.

Climate change is a highly

complex topic and as such its

various causes and effects are

aggregated under the single

term.

Hydrological

modification


hydrological change, such as

damming or riverine changes.

Vegetation clearing Article relates to, or studies

vegetation clearing, such as

deforestation, riparian

clearing, or the clearing of

remnant vegetation.

Fishing Article relates to, or studies

fisheries.

146

Hunting Article relates to, or studies

hunting as an industry or

community activity.

Land degradation other Article relates to, or studies

the degradation of land which

is not related to soil pollution,

such as erosion, or salinity.

Land reclamation Article relates to, or studies

the reclamation of land, such

as creating new land from

ocean, riverbeds, or lake beds

Mining Article relates to, or studies

mining.

Ozone depletion Article relates to, or studies

ozone depletion.

Soil pollution Article relates to, or studies

soil pollution.

Fresh water degradation

other


non-hydrological

modification impacts to fresh

water systems.

Sea water

degradation/pollution


the degradation of sea water,

such as acidification, mining,

or pollution.

Fresh water pollution Article relates to, or studies

the pollution of fresh water

systems.

Environmental hazard variables

This typology was developed using hazard variables described by GRID Arendal (2017) as a

baseline. We updated this list and added descriptions during the title and abstract screening phases

to include any hazard variables used in screened publications, which were not included in the

baseline.

147

Environmental hazard

variable

Description Notes

Air quality Article relates to, or studies

exploitation activity which

results in a change to air

quality.

Amenity degradation Article relates to, or studies


results in the loss of

environmental amenity.

Fire Article relates to, or studies


results in/contributes to

unnatural fire event(s).

Often associated with

climate change.

Intensification of natural

hazards



results in natural hazards

(such as extreme weather)

intensifying.

Used where particular

hazards (e.g. fire events) are

not specified, but generic

language such as

‘intensification of natural

disasters’ etc. used.

Extreme storms Article relates to, or studies


contributes to extreme

storm events, such as

cyclones, and hurricanes.


climate change.

Deforestation Article relates to, or studies


results in loss of forest.

Desertification Article relates to, or studies


results in desertification.


climate change.

Disease Article relates to, or studies


148

results in increased

occurrence or incidence of

disease.

Drought Article relates to, or studies


results in or contributes to

drought events.


climate change.

Erosion Article relates to, or studies


results in soil erosion.

Flooding Article relates to, or studies



flood events.


climate change,

deforestation and erosion.

Food quality\supply Article relates to, or studies


results in impacts to either

the quality or availability of

food.

Heat stress/wave Article relates to, or studies



heat wave or heat stress

events.


climate change.

Landslide Article relates to, or studies



landslides.

Not specified Article does not specify any

natural hazards.

Oil spill Article relates to, or studies


results in oil entering the

149

environment uncontrolled.

Over grazing Article relates to, or studies


results in overgrazing of

land.

Usually also related to soil

erosion/pollution.

Over fishing Article relates to, or studies


results in unsustainable,

over exploitation of a

fishery.

Pest invasion Article relates to, or studies



vectors used by exotic, pest

species.

Soil pollution Article relates to, or studies



soil pollution.

Sea level rise Article relates to, or studies



rising sea levels.

Almost exclusively

associated with climate

change.

Water quality\supply Article relates to, or studies


results in or contributes to a

loss of water supply or

quality.

150

Additional file 3 – Data tables

Tables which contain all data used in figures within the main manuscript.

Table 1 – Methods used. Counts and percent (rounded to two figures) of methods used.

Method Studies Percent

Case studies 11 3.107345

Census analysis 1 0.282486

Desktop study 73 20.62147

Discussion 25 7.062147

Document analysis 48 13.55932

Ethnographic 3 0.847458

Field research 1 0.282486

Focus groups 2 0.564972

Framework 7 1.977401

Interviews 11 3.107345

Mixed methods 12 3.389831

Modelling 25 7.062147

Philosophical argument 54 15.25424

Spatial analysis 67 18.92655

Surveys 8 2.259887

Informed opinion 6 1.694915

Table 2 – First author institute location. Counts and percent (rounded to two figures) of total

authors shown. NA denotes papers where author location was indeterminable.

Country Count Percent

United states 149 42.09

United Kingdom 45 12.71

Australia 24 6.78

NA 24 6.78

Canada 20 5.65

China 8 2.26

Germany 8 2.26

Spain 8 2.26

151

France 6 1.69

India 6 1.69

Italy 6 1.69

Netherlands 5 1.41

Norway 5 1.41

Denmark 4 1.13

Austria 3 0.85

New Zealand 3 0.85

Thailand 3 0.85

Brazil 2 0.56

Chile 2 0.56

Finland 2 0.56

Indonesia 2 0.56

Ireland 2 0.56

Malaysia 2 0.56

South Africa 2 0.56

Sweden 2 0.56

Switzerland 2 0.56

Taiwan 2 0.56

Czech republic 1 0.28

Jamaica 1 0.28

Japan 1 0.28

Pakistan 1 0.28

Singapore 1 0.28

Swaziland 1 0.28

Tanzania 1 0.28

Table 3 – Study location. Counts and percent (rounded to two figures) of study location shown.

NA denotes papers where study location was indeterminable or global.

Country Count Percent

NA 155 43.79

United states 89 25.14

United Kingdom 18 5.08

152

Canada 12 3.39

Australia 11 3.11

India 10 2.82

China 8 2.26

New Zealand 7 1.98

France 6 1.69

Chile 4 1.13

South Africa 4 1.13

Brazil 3 0.85

Peru 3 0.85

Spain 3 0.85

Ghana 2 0.56

Indonesia 2 0.56

Bangladesh 1 0.28

Czech republic 1 0.28

Guatemala 1 0.28

Italy 1 0.28

Jamaica 1 0.28

Malaysia 1 0.28

Nepal 1 0.28

Netherlands 1 0.28

Panama 1 0.28

Puerto Rico 1 0.28

Russia 1 0.28

Saint Vincent and the grenadines 1 0.28

Somalia 1 0.28

Sudan 1 0.28

Sweden 1 0.28

Taiwan 1 0.28

Thailand 1 0.28

153

Table 4 Demographic themes/variables explored. Counts and percent of total for each paper

shown. Multiple unspecified was recorded where a paper used multiple demographic

themes/variable but did not explicitly state which. Other was recorded to capture the very diverse

range of variables used which were not included in our demographic typology. Note that a single

article may explore several themes and as such counts here are higher than total article counts.

Theme Count Percent

Economic Status 147 16.66667

Geographic 112 12.69841

Racial/ethnic 96 10.88435

Other 84 9.52381

Education 64 7.256236

Age 58 6.575964

Employment 41 4.648526

Gender 41 4.648526

Indigenous 40 4.535147

Multiple unspec. 37 4.195011

Children 30 3.401361

Farmers/Fishers/Hunters 28 3.174603

Health/Disability 26 2.947846

Future generations 25 2.834467

Immigrants 24 2.721088

Rural 20 2.267574

Marital status 7 0.793651

LGBTQ+ 1 0.113379

Religious 1 0.113379

154

Table 5 Human well-being themes explored. Counts and percent of total for each paper shown.

Note that a single article may explore several themes and as such counts here are higher than total

article counts.

Theme Count Percent

Health/Mortality 214 38.07829

Living standards 78 13.879

Not specified 78 13.879

Livelihood 63 11.20996

Resource access 42 7.47331

Security/Safety 33 5.871886

Social wellness 23 4.092527

Spiritual/Cultural 12 2.135231

Governance 10 1.779359

Mental health 5 0.88968

Education 4 0.711744

155

Table 6 Natural resource exploitation themes explored. Counts and percent of total for each

paper shown. Note that a single article may explore several themes and as such counts here are

higher than total article counts.

Theme Count Percent

Climate change 152 35.43124

Air pollution 130 30.30303

Mining 46 10.72261

Water pollution 30 6.993007

Soil pollution 16 3.729604

Veg. clearing 15 3.496503

Land deg. other 14 3.263403

Fishing 11 2.564103

Fresh water deg. other 7 1.631702

Hydro. modification 4 0.932401

Hunting 2 0.4662

Ozone depletion 2 0.4662

156

Table 7 Natural hazard themes explored. Counts and percent of total for each paper shown. Note

that a single article may explore several themes and as such counts here are higher than total article

counts.

Theme Count Percent

Air quality 143 24.23729

Not specified 83 14.0678

Water quality/supply 59 10

N.disaster intensity 56 9.491525

Disease 34 5.762712

Flooding 28 4.745763

Food quality/supply 28 4.745763

Heatwave/stress 27 4.576271

Sea level rise 25 4.237288

Soil pollution 24 4.067797

Drought 18 3.050847

Deforestation 17 2.881356

Intense Storm 14 2.372881

Amenity degradation 12 2.033898

Over fishing 6 1.016949

Erosion 5 0.847458

Land slide 4 0.677966

Desertification 4 0.677966

Extreme Fire 2 0.338983

Over grazing 1 0.169492

157

Additional file 4 Complete dataset

Complete dataset is hosted at url: https://data.mendeley.com/datasets/g6frzjmnzt/1

https://data.mendeley.com/datasets/g6frzjmnzt/1

158


The text appears exactly as published, with the exception of change from Nature referencing

style to Sage Harvard referencing style and placement of Methods section before Results

section.

S1 Overview

This document provides supplementary information not provided in the main text of the article

Global mismatch between greenhouse gas emissions and the burden of climate change. S2 shows

the climate equity Lorenz curve with Gini and Robin hood indices. S3 shows the results of

performing the same analyses provided in the Methods of the main text, but uses per-capita

emissions. S4 presents a table of data used for analyses.

S2 Summary of Lorenz curve

As detailed in the Methods of the main text, equity indices were created to measure climate

inequity. We created a Lorenz curve (Supplementary Fig. S2) which shows the equity of the

distribution of GHG emissions data for the year 2010(Coulter, 1989a). The blue line represents the

hypothetical line of equality where countries would reside if GHG emissions were divided with

perfect equity between all countries. The red line represents the Lorenz curve and denotes actual

GHG emissions equity for the year 2010. The distance between the blue and red lines represents

GHG emissions inequity and is quantified by the Gini (80.9) and Robin Hood indices (64). As such,

the curve and both indices show that the current distribution of GHG emissions is highly

inequitable.

S3 A summary of per capita GHG emissions and national vulnerability

We conducted separate analyses to those reported in the Methods section of the main text, using per

capita GHG emissions. We created a similar set of maps to those in the main text for per capita

emissions by dividing emissions by national populations (World Bank Group, 2015)

(Supplementary Fig. S3). Per capita emissions are shown for years 2010 (Supplementary Fig. S3a)

and 2030 (Supplementary Fig. S3b). Maps generated using ESRI ArcGIS (ESRI ArcGIS,

2011).The patterns we found here are broadly similar to those in the main text (Supplementary Fig.

S3c). For example, Australia, Russia and the United States of America remain as free riders.

However, several populous major emitters (e.g. United Kingdom, China, Brazil) were no longer

categorised as free riders.

S4 Summary of data

A table of all countries, their level of vulnerability at 2010 and 2030, GHG emissions, GDP at 2010

and population at 2010 and 2030 is presented (Supplementary Table S4).

159

Supplementary Figure S2

Supplementary Figure S3

a

b

160

c

161

Supplementary table S4

Country

Climate

Change

Vulnerabili

ty 2010

Climate

change

vulnerabili

ty 2030

GHG

Emissions

(MtCO2e)

2010 GDP 2010

Population

2010

Population

2030

Afghanistan SEVERE SEVERE 26.5382

159367844

37 28397812 43499632

Albania LOW

MODERAT

E 6.8977

119269572

55 3150143 3310564

Algeria

MODERAT

E

MODERAT

E 166.7126

161207304

960 37062820 48561408

Angola SEVERE ACUTE 262.3005

824708948

68 19549124 34783312

Antigua and

Barbuda SEVERE ACUTE 1.0505

113553903

7 87233 104982

Argentina LOW LOW 426.0262

462703793

707 40374224 46859381

Armenia

MODERAT

E

MODERAT

E 7.8514

926028741

6 2963496 2969807

Australia LOW LOW 592.4785

114126776

0188 22404488 28335501

Austria LOW LOW 89.7307

389656071

767 8401924 9005424

Azerbaijan LOW LOW 58.5329

529027033

76 9094718 10474377

Bahamas SEVERE ACUTE 2.7347

791000000

0 360498 447410

Bahrain LOW LOW 27.9484

257135448

25 1251513 1641988

Bangladesh SEVERE ACUTE 157.8669

115279077

465 151125475 185063630

Barbados

MODERAT

E HIGH 3.5481

443370000

0 280396 305709

Belarus MODERAT MODERAT 110.5189 552209326 9491070 8488334

162

E E 14

Belgium LOW LOW 129.1643

484404271

608 10941288 11664194

Belize ACUTE ACUTE 13.9414

139711345

0 308595 461277

Benin SEVERE ACUTE 27.9443

656178231

3 9509798 15506762

Bhutan SEVERE ACUTE -7.4517

158539625

6 716939 897761

Bolivia HIGH ACUTE 149.3575

196496313

08 10156601 13665316

Bosnia and

Herzegovin

a

MODERAT

E

MODERAT

E 27.3532

168474930

59 3845929 3700255

Botswana

MODERAT

E

MODERAT

E 20.1171

137467127

11 1969341 2347860

Brazil LOW

MODERAT

E 1392.6399

214306787

1760 195210154 222748294

Brunei

MODERAT

E

MODERAT

E 23.009

123697088

59 400569 499424

Bulgaria LOW

MODERAT

E 47.5596

486690605

12 7389175 6213179

Burkina

Faso SEVERE ACUTE 32.0538

920928838

3 15540284 26564341

Burundi SEVERE ACUTE 41.1171

202686441

4 9232753 16392403

Cambodia SEVERE ACUTE 48.7258

112422663

34 14364931 19143612

Cameroon SEVERE ACUTE 199.5086

236224829

55 20624343 33074215

Canada LOW LOW 841.5982

161407209

3764 34126240 40616997

Cape Verde HIGH SEVERE 0.6499 na 487601 576734

Central ACUTE ACUTE 103.4364 198601475 4349921 6318381

163

African

Republic

9

Chad SEVERE ACUTE 52.7923

106577050

72 11720781 20877527

Chile

MODERAT

E

MODERAT

E 86.8362

217501911

334 17150760 19814578

China LOW

MODERAT

E 9387.0095

593050227

0313

135982146

5

145329730

4

Colombia

MODERAT

E

MODERAT

E 216.3056

287018184

638 46444798 57219408

Comoros ACUTE ACUTE 0.5174 516962949 683081 1057197

Costa Rica

MODERAT

E HIGH 7.584

362983276

70 4669685 5759573

Côte

d'Ivoire SEVERE ACUTE 62.1688

248845039

51 18976588 29227188

Croatia

MODERAT

E

MODERAT

E 15.7967

596438181

82 4338027 4015138

Cuba

MODERAT

E SEVERE 45.2579

643282200

00 11281768 10847333

Cyprus LOW LOW 8.7575

215619413

69 1103685 1306312

Czech

Republic LOW LOW 124.4638

231324503

31 10553701 11053125

Democratic

Republic of

the Congo SEVERE ACUTE 317.3795

207016402

026 62191161 103743184

Denmark LOW LOW 60.1913

319812413

597 5550959 6009458

Djibouti HIGH SEVERE 1.1688

112861170

0 834036 1075146

Dominica SEVERE ACUTE 0.2797 492962963 71167 76952

Dominican

Republic HIGH ACUTE 31.442

530429437

31 10016797 12218615

Ecuador MODERAT HIGH 135.5452 695553670 15001072 19648546

164

E 00

Egypt LOW LOW 276.3722

218887812

550 78075705 102552797

El Salvador SEVERE ACUTE 14.1367

214183000

00 6218195 6874758

Equatorial

Guinea HIGH SEVERE 25.5787

115829177

90 696167 1138788

Eritrea HIGH SEVERE 6.1925

211703951

1 5741159 9782455

Estonia LOW LOW 29.0265

194790124

23 1298533 1212150

Ethiopia HIGH SEVERE 148.3437

299337903

34 87095281 137669707

Fiji SEVERE ACUTE 1.3405

314050883

6 860559 939469

Finland LOW LOW 83.3728

247799815

768 5367693 5649744

France LOW LOW 488.4332

264683711

1795 63230866 69286370

Gabon SEVERE ACUTE 6.661

145695271

25 1556222 2382369

Gambia ACUTE ACUTE 6.944 951827285 1680640 3056357

Georgia

MODERAT

E HIGH 9.4743

116385368

34 4388674 3953077

Germany LOW LOW 827.0112

341200877

2737 83017404 79551501

Ghana HIGH ACUTE 59.4589

321748397

13 24262901 35264291

Greece LOW LOW 107.6441

299598056

253 11109999 10975530

Grenada SEVERE ACUTE 1.7461 771015875 104677 107433

Guatemala

MODERAT

E HIGH 50.5749

413379582

52 14341576 22566243

Guinea ACUTE ACUTE 32.8177 473595647 10876033 17322136

165

6

Guinea-

Bissau ACUTE ACUTE 4.3053 847491367 1586624 2472642

Guyana SEVERE ACUTE 7.4287

225928839

6 786126 852670

Haiti ACUTE ACUTE 7.9624

662254152

9 9896400 12536811

Honduras SEVERE ACUTE 47.4738

158393445

92 7621204 10811004

Hungary LOW LOW 70.5821

129585601

616 10014633 9525243

Iceland

MODERAT

E

MODERAT

E 2.9422

132610355

17 318042 383558

India SEVERE ACUTE 2304.0432

170845887

6830

120562464

8

147637790

3

Indonesia

MODERAT

E HIGH 2033.0459

709190823

320 240676485 293482460

Iran

MODERAT

E

MODERAT

E 695.2992

422567967

405 74462314 91336270

Iraq

MODERAT

E

MODERAT

E 227.1567

138516722

650 30962380 50966609

Ireland LOW LOW 60.163

218435251

789 4467561 5346841

Israel LOW LOW 86.9162

232907996

791 7420368 9632030

Italy LOW LOW 463.1778

212662040

2889 60508978 61211831

Jamaica ACUTE ACUTE 12.0012

132308440

40 2741485 2949838

Japan LOW LOW 1119.975

549538718

2996 127352833 120624738

Jordan LOW LOW 25.8197

264253794

37 6454554 9355173

Kazakhstan LOW LOW 300.8368 148047348 15921127 18572745

166

241

Kenya

MODERAT

E HIGH 56.8606

400001388

31 40909194 66306063

Kiribati SEVERE ACUTE 0.0774 150431114 97743 130715

Kuwait LOW LOW 191.8787

115428557

470 2991580 4832793

Kyrgyzstan

MODERAT

E HIGH -4.6703

479435779

5 5334223 6871058

Laos HIGH ACUTE 45.2108

718144115

2 6395713 8806260

Latvia LOW LOW -2.2515

240096804

60 2090519 1855822

Lebanon LOW

MODERAT

E 24.3469

380099502

49 4341092 5171981

Lesotho

MODERAT

E

MODERAT

E 2.6993

217568568

1 2008921 2419217

Liberia ACUTE ACUTE 16.8827

129269647

6 3957990 6395182

Libya LOW

MODERAT

E 145.7536

747552889

17 6040612 7459411

Lithuania LOW LOW 26.7093

367095115

68 3068457 2816749

Luxembour

g LOW LOW 12.1485

521436503

83 507885 636826

Macedonia LOW

MODERAT

E 12.4073

933867407

8 2102216 2068730

Madagascar ACUTE ACUTE 78.5476

872993613

6 21079532 36000163

Malawi SEVERE ACUTE 23.3628

539861698

5 15013694 25959551

Malaysia

MODERAT

E HIGH 428.9105

247533525

518 28275835 36845517

Maldives HIGH SEVERE 1.1599

213410488

4 325694 435873

167

Mali ACUTE ACUTE 32.2724

942226726

0 13985961 26034111

Malta LOW LOW 3.012

816384106

0 424738 436792

Mauritania SEVERE ACUTE 9.9108

352694760

9 3609420 5640323

Mauritius HIGH SEVERE 5.9566

971823391

1 1230659 1287944

Mexico

MODERAT

E HIGH 706.0062

105162794

9327 117886404 143662574

Moldova

MODERAT

E HIGH 11.8209

581160405

2 3573024 3066205

Mongolia HIGH HIGH 52.5383

620035707

0 2712738 3387631

Morocco

MODERAT

E HIGH 92.0352

907706714

32 31642360 39190274

Mozambiqu

e ACUTE ACUTE 56.6871

101653535

91 23967265 38875906

Myanmar ACUTE ACUTE 238.0118 na 51931231 58697747

Namibia

MODERAT

E SEVERE 23.5119

112819964

26 2178967 3042197

Nepal HIGH SEVERE 42.8585

159940946

07 26846016 32853228

Netherlands LOW LOW 220.1469

836389937

229 16615243 17268589

New

Zealand LOW LOW 53.0408

143466743

661 4368136 5208035

Nicaragua SEVERE ACUTE 46.5508

893820965

1 5822209 7390914

Niger SEVERE ACUTE 20.259

571858955

0 15893746 34512751

Nigeria SEVERE ACUTE 491.669

369062403

182 159707780 273120384

North SEVERE ACUTE 108.6613 na 24500520 26718625

168

Korea

Norway LOW LOW 27.2304

420945705

225 4891251 5837893

Oman

MODERAT

E HIGH 99.0907

586413528

78 2802768 4920265

Pakistan HIGH ACUTE 326.6593

177165635

077 173149306 231743898

Palau SEVERE ACUTE 0.2839 197558700 20470 24836

Panama

MODERAT

E SEVERE 21.1526

288141000

00 3678128 4882047

Papua New

Guinea SEVERE ACUTE 97.6247

948004795

9 6858945 10044486

Paraguay

MODERAT

E HIGH 112.8127

200305297

33 6459721 8693133

Peru

MODERAT

E HIGH 150.7154

148522810

972 29262830 36513996

Philippines

MODERAT

E SEVERE 145.7789

199589447

424 93444322 127797234

Poland LOW LOW 333.2904

476687891

752 38198754 37447642

Portugal LOW LOW 74.1196

238303443

425 10589792 10432816

Qatar LOW LOW 75.2752

125122249

141 1749713 2760329

Republic of

Congo HIGH SEVERE 21.9944

120078800

67 4111715 6753771

Romania LOW LOW 117.5375

164792252

746 21861476 20232088

Russia LOW LOW 2133.8469

152491611

2079 143617913 133556108

Rwanda HIGH SEVERE 3.9692

569854892

3 10836732 17771249

Saint Lucia SEVERE ACUTE 1.1142

124429740

7 177397 201817

169

Saint

Vincent and

the

Grenadines

MODERAT

E SEVERE 0.2239 681225963 109316 110012

Samoa HIGH ACUTE 0.3459 643059403 186029 211105

Sao Tome

and

Principe ACUTE ACUTE 0.1718 201037917 178228 278192

Saudi

Arabia LOW LOW 510.1386

526811466

667 27258387 35634201

Senegal SEVERE ACUTE 28.3144

129324277

24 12950564 21855703

Seychelles SEVERE ACUTE 0.7709 973355738 91208 98416

Sierra

Leone ACUTE ACUTE 10.7209

257815949

6 5751976 8057580

Singapore LOW LOW 70.055

236420337

243 5078969 6577884

Slovakia LOW LOW 37.0475

890119192

05 5433437 5395535

Slovenia LOW

MODERAT

E 5.5184

479729887

42 2054232 2086066

Solomon

Islands ACUTE ACUTE 2.2297 681587105 526447 764146

South

Africa

MODERAT

E

MODERAT

E 458.765

365208432

989 51452352 58095501

South

Korea LOW

MODERAT

E 629.5671

109449935

0178 48453931 52190069

Spain LOW LOW 334.3103

143158761

2302 46182038 48235492

Sri Lanka

MODERAT

E SEVERE 48.0389

495675216

70 20758779 23271183

Sudan HIGH ACUTE 186.007

656320812

40 35652002 55077835

Suriname HIGH SEVERE 9.1899 436839804 524960 603805

170

8

Swaziland

MODERAT

E HIGH 2.8898

389156347

8 1193148 1515527

Sweden LOW LOW 39.5656

488377689

565 9382297 10690986

Switzerland LOW LOW 51.274

581208562

423 7830534 9477452

Syria LOW LOW 93.866 na 21532647 29933865

Tajikistan HIGH HIGH 10.8228

564217858

0 7627326 11407028

Tanzania HIGH ACUTE 172.5317

229150045

89 44973330 79354326

Thailand

MODERAT

E SEVERE 344.5517

318907930

076 66402316 67554088

Togo SEVERE ACUTE 16.7057

317294550

6 6306014 10014965

Tonga

MODERAT

E ACUTE 0.3755 369212477 104098 120995

Trinidad

and Tobago

MODERAT

E

MODERAT

E 46.049

207581918

58 1328095 1307826

Tunisia LOW

MODERAT

E 34.2352

444260164

87 10631830 12561225

Turkey LOW

MODERAT

E 353.3523

731168051

903 72137546 86825345

Turkmenist

an

MODERAT

E

MODERAT

E 101.005

221480701

75 5041995 6159875

Uganda HIGH SEVERE 45.6373

160309961

79 33987213 63387713

Ukraine

MODERAT

E

MODERAT

E 356.9985

136419300

368 46050220 39841900

United

Arab

Emirates LOW

MODERAT

E 202.3437

286049293

399 8441537 12330367

United LOW LOW 579.3852 240793376 62066350 68630898

171

Kingdom 7805

United

States LOW LOW 6253.7159

149644000

00000 312247116 362628830

Uruguay

MODERAT

E

MODERAT

E 13.1671

388811021

00 3371982 3581432

Uzbekistan LOW

MODERAT

E 214.0483

393327709

29 27769270 34146873

Vanuatu ACUTE ACUTE 0.4817 700804286 236299 352225

Venezuela

MODERAT

E HIGH 399.0262

393801556

872 29043283 37172167

Vietnam HIGH ACUTE 250.0867

115931749

905 89047397 101830324

Yemen

MODERAT

E HIGH 34.1512

317437511

69 22763008 33991041

Zambia SEVERE ACUTE 128.6095

202653963

26 13216985 24956509

Zimbabwe HIGH SEVERE 57.8627

945680820

0 13076978 20292380

172


The text appears exactly as intended for publication.

Supplementary file 1

#AUS_EMISSIONS_DJ.R

#Analysis of Australian air pollution, social demographics and health data.

#Written by Glenn Althor 2018

#### load packages and write functions ####

library(truncreg) # Truncated regression

library(ggplot2) # plotting

library(reshape2) # melt

library(geoR) # boxcoxfit

library(psych) # describe

library(dplyr) # data wrangling

library(mice) # ad hoc imputation

library(censReg) # censored/tobit modelling

library(viridis) # colour blind friendly colour pallete

library(Publish) # ci.mean

library(Cairo) # print figures

#### create colour palettes ####

two_col <- viridis(2)

#### Prepare emissions data ####

rm(list = ls())

setwd("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental equity in

Australia/Data/R/AUS_EMISSIONS_DJ")

npi_emissions <- read.table(header=TRUE,

sep=",","C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental equity in

Australia/Data/NPI/@Working/npi_emissions.csv", stringsAsFactors = F)

colnames(npi_emissions)[1] <- "report_year"

173

tox_matched_substances <- read.table(header=TRUE,sep =

',',"C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental equity in

Australia/Data/NPI/@Working/tox_matched_substances.csv", stringsAsFactors = F)

colnames(tox_matched_substances)[1]<- "substance_name"

# merge datasets, keeping matching toxicity to npi emissions (all.x = T) = join keeping x.

dat <- merge(npi_emissions, tox_matched_substances, by = "substance_name")

rm(npi_emissions)

rm(tox_matched_substances)

# remove entries for non-air/water emissions and 0 values for emissions

dat <- subset(dat, air_total_emission_kg != "NA" |

water_emission_kg != "NA")

dat <- subset(dat, air_total_emission_kg != 0 |

water_emission_kg != 0)

# Convert emissions NA values to 0

dat[c(3:6,10:13)][is.na(dat[c(3:6,10:13)])] <- 0

# remove entries for chemicals with no HTP (may reconsider later)

dat <- subset(dat, HTP_Chemical_name != "NA")

# Create total risk - emission*HTP

dat$weight_cancer_risk_air <- dat$air_total_emission_kg*dat$HTP_Cancer_Air

dat$weight_cancer_risk_water <- dat$water_emission_kg*dat$HTP_Cancer_Water

dat$tot_cancer_risk <- dat$weight_cancer_risk_air+dat$weight_cancer_risk_water

dat$weight_noncancer_risk_air <- dat$air_total_emission_kg*dat$HTP_NonCancer_Air

dat$weight_noncancer_risk_water <- dat$water_emission_kg*dat$HTP_NonCancer_Water

dat$tot_noncancer_risk <- dat$weight_noncancer_risk_air+dat$weight_noncancer_risk_water

dat$tot_risk <- dat$tot_cancer_risk+dat$tot_noncancer_risk

# filterout all non-necessary columns

dat <- dat[c(2,7,8,20)]

dat.2016 <- subset(dat, report_year == "2015/2016")



174

# # Export csv files

write.csv(dat.2016,file = "./outputs/2016_weighted_emissions.csv", row.names = F)



#### ARC GIS spatial join process ####

# 1. Convert csv to xls

# 2. Import xls

# 3. Right click - Display XY data

# 4. Export as shape

# 5. Spatial join shape to SA1 data - Join one to one, and sum total risk

# 6. Export new data set as text, and convert to csv "2016_emissions_join.csv"

# Add SA2 and SA3 Codes to dataset

rm(list = ls())

setwd("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental equity in

Australia/Data/R/AUS_EMISSIONS_DJ")

SA1_2016_emissions <- read.csv2("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3.

Environmental equity in Australia/Data/R/AUS_EMISSIONS_DJ/2016_emissions_join.csv", sep =

',')

colnames(SA1_2016_emissions)[5] <- "SA1_7DIGITCODE_2016"

SA_All <- read.csv2("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental

equity in Australia/Data/R/AUS_EMISSIONS_DJ/SA1_2016_AUST.csv", sep = ',')

dat <- merge(SA1_2016_emissions, SA_All)

write.csv(dat,file = "./outputs/2016_emissions_join_SA.csv")

#### Impute 2016 ####

setwd('C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental equity in

Australia/Data/Census')

## TTP 2016 data ##

# TTP 2016 import #

175



TTP_2016 <- read.table("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental

equity in Australia/Data/outputs/relative_TTP_2016.txt", sep = ',', header = T)

# Clean data up & Filter out unnecessary columns

dat_2016 <- TTP_2016[c(5,11)]

colnames(dat_2016)[1] <- "SA1_7DIGITCODE_2016"

colnames(dat_2016)[2] <- "TTP_2016"

rm(TTP_2016)

## SEIFA 2016 data ##

SEIFA_2016 <- read.csv2("./SEIFA_2016.csv", sep = ',', header = T, stringsAsFactors = F)

colnames(SEIFA_2016)[1] <- "SA1_7DIGITCODE_2016"

SEIFA_2016 <- SEIFA_2016[c(1:2,4,6,8)]

SEIFA_2016[,] <- as.numeric(as.character(unlist(SEIFA_2016[,])))

dat_2016 <- merge(dat_2016, SEIFA_2016, all.x = T)

colnames(dat_2016) <- c("SA1_7DIGITCODE_2016",

"TTP_2016",

"IRSD_Score",

"IRSAD_Score",

"IEO_Score",

"IER_Score")

dat_2016[is.na(dat_2016)] <- NA

dat_2016 <- na.omit(dat_2016)

rm(SEIFA_2016)

## Indigenous % data 2016##

IND_pct_2016 <- read.csv2("./INDI_SA1_2016.csv", sep = ',', header = T, stringsAsFactors = F)

IND_pct_2016[,] <- as.numeric(as.character(unlist(IND_pct_2016[,])))

dat_2016 <- merge(dat_2016, IND_pct_2016, all.x = T)

dat_2016$pct_ind[is.na(dat_2016$pct_ind)] <- NA

dat_2016$pct_ind <- na.omit(dat_2016$pct_ind)

rm(IND_pct_2016)

176

# Summary statistics #

# describe(dat_2016[c(2:7)])

## Adhoc imputation ##

# convert 0 obs to NA #

dat_2016$TTP_2016[dat_2016$TTP_2016 == 0] <- NA

# View missing data #

md_2016 <- md.pattern(dat_2016)

md_2016

# Mulitple imputations - Predictive mean matching #

imp_2016 <- mice(dat_2016[c(2:7)], m = 50, method = "pmm", seed = 1)

# xy plot to check imputed values

TTP_2016_imp_p <- xyplot(imp_2016, TTP_2016 ~ IRSAD_Score | .imp, pch = 20, cex = 1.4,

col=two_col, alpha = 0.5)

## Regressions ##

# Percent Ind regression

IND_2016 <- with(imp_2016, lm(log(TTP_2016) ~ pct_ind))

# summary(IND_2016)

# pool(IND_2016)

# summary(pool(IND_2016))

# pool.r.squared(IND_2016)

# IRSAD regression #

IRSAD_2016 <- with(imp_2016, lm(log(TTP_2016) ~ IRSAD_Score))

# summary(IRSAD_2016)

# pool(IRSAD_2016)

# summary(pool(IRSAD_2016))

# pool.r.squared(IRSAD_2016)

# IRSD regression #

177

IRSD_2016 <- with(imp_2016, lm(log(TTP_2016) ~ IRSD_Score + pct_ind))

# summary(IRSD_2016)

# pool(IRSD_2016)

# summary(pool(IRSD_2016))

# pool.r.squared(IRSD_2016)

# IER regression #

IER_2016 <- with(imp_2016, lm(log(TTP_2016) ~ IER_Score))

# summary(IER_2016)

# pool(IER_2016)

# summary(pool(IER_2016))

# pool.r.squared(IER_2016)

# IEO regression #

IEO_2016 <- with(imp_2016, lm(log(TTP_2016) ~ IEO_Score))

# summary(IEO_2016)

# pool(IEO_2016)

# summary(pool(IEO_2016))

# pool.r.squared(IEO_2016)

# Combine all R^2 values #

r_sq_all_2016_imp <- as.data.frame(cbind(c(pool.r.squared(IRSAD_2016)[1,1],

pool.r.squared(IRSD_2016)[1,1],

pool.r.squared(IER_2016)[1,1],

pool.r.squared(IEO_2016)[1,1],

pool.r.squared(IND_2016)[1,1])))

r_sq_all_2016_imp$y <- round(r_sq_all_2016_imp$V1*100, digits = 3)

colnames(r_sq_all_2016_imp) <- c("Estimated r^2","Estimated % variance")

rownames(r_sq_all_2016_imp) <- c("IRSAD","IRSD","IER","IEO","% Indig. Aus.")

#### Impute 2011 ####

# TTP Import #

178



dat_2011 <- read.table("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental



dat_2011 <- dat_2011[c(5,11)]

colnames(dat_2011)[1] <- "SA1_7DIGITCODE_2016"


# SEIFA 2011 DATA #



SEIFA_2011 <- SEIFA_2011[c(1:2,4,6,8)]



dat_2011[is.na(dat_2011)] <- NA


rm(SEIFA_2011)

## Indigenous % Data 2011##



IND_pct_2011 <- IND_pct_2011[c(1,4)]

colnames(IND_pct_2011)[c(1,2)] <- c("SA1_7DIGITCODE_2016","pct_ind")




rm(IND_pct_2011)


describe(dat_2011[c(2:7)])



179

dat_2011$TTP_2011[dat_2011$TTP_2011 == 0] <- NA

# Check for missing data


md_2011

# adhoc imputation prediction


# Check imputed data



## Regressions ##



# summary(IND_2011)

# pool(IND_2011)






# pool(IRSAD_2011)



# IRSD regression #

IRSD_2011 <- with(imp_2011, lm(log(TTP_2011) ~ IRSD_Score + pct_ind))

# summary(IRSD_2011)

# pool(IRSD_2011)



180

# IER regression #


# summary(IER_2011)

# pool(IER_2011)



# IEO regression #


# summary(IEO_2011)

# pool(IEO_2011)




# 2011 #









#### Impute 2006 ####

# TTP Import #



dat_2006 <- read.table("C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental



dat_2006 <- dat_2006[c(4,5)]

181

colnames(dat_2006)[1] <- "CD_CODE06"


# SEIFA 2006 DATA #


SEIFA_2006 <- SEIFA_2006[c(1:2,4,6,8)]



dat_2006[is.na(dat_2006)] <- NA


rm(SEIFA_2006)


IND_pct_2006 <- read.csv2("./INDI_CD_2006.csv", sep = ',', header = T, stringsAsFactors = F)


IND_pct_2006 <- IND_pct_2006[c(1,4)]

colnames(IND_pct_2006)[c(1,2)] <- c("CD_CODE06","pct_ind")




rm(IND_pct_2006)


describe(dat_2006[c(2:7)])



dat_2006$TTP_2006[dat_2006$TTP_2006 == 0] <- NA

# Check for missing data


md_2006

# adhoc imputation prediction


182

# Check imputed data



## Regressions ##



# summary(IND_2006)

# pool(IND_2006)






# pool(IRSAD_2006)



# IRSD regression #

IRSD_2006 <- with(imp_2006, lm(log(TTP_2006) ~ IRSD_Score))

summary(IRSD_2006)

# pool(IRSD_2006)



# IER regression #


# summary(IER_2006)

# pool(IER_2006)



# IEO regression #

183


# summary(IEO_2006)

# pool(IEO_2006)












#### Data and plot exports ####

## Data outputs ##

dir.create(file.path(getwd(), "outputs/imputed/"))

# Missing Data #

write.csv(md_2016,file = "./outputs/imputed/missing_data_2016.csv")



## Summary statistics ##

# Change NA back to 0 #

dat_2006[c(2)][is.na(dat_2006[c(2)])] <- 0

dat_2011[c(2)][is.na(dat_2011[c(2)])] <- 0

dat_2016[c(2)][is.na(dat_2016[c(2)])] <- 0

write.csv(describe(dat_2016[c(2:7)]),file = "./outputs/2016_summary_stats.csv")



184

# Pooled regressions #

write.csv(summary(pool(IND_2016)), file = "./outputs/imputed//IND_2016.csv")

write.csv(summary(pool(IRSAD_2016)), file = "./outputs/imputed//IRSAD_2016.csv")

write.csv(summary(pool(IRSD_2016)), file = "./outputs/imputed//IRSD_2016.csv")

write.csv(summary(pool(IER_2016)), file = "./outputs/imputed//IER_2016.csv")

write.csv(summary(pool(IEO_2016)), file = "./outputs/imputed//IEO_2016.csv")











# pooled R^2 estimates #

# Output #

write.csv(r_sq_all_2006_imp, file = "./outputs/imputed/r_sq_all_2006_imp.csv")



## Plots ##

# Imputed data distribution plots #

Cairo(3800, 3800, file="./outputs/imputed//plots/TTP_2016_imp_p.png",

type = "png", dpi = 300, bg="white")

TTP_2016_imp_p

dev.off()

185



TTP_2011_imp_p

dev.off()



TTP_2006_imp_p

dev.off()

#### truncated regressions ####

#### trunc 2016 ####

# TTP 2016 Import #






dat_2016_t <- TTP_2016[c(5,11)]

colnames(dat_2016_t)[1] <- "SA1_7DIGITCODE_2016"

colnames(dat_2016_t)[2] <- "TTP_2016"

rm(TTP_2016)

dat_2016_t <- subset(dat_2016_t, TTP_2016 > 0) # 0 values

# SEIFA 2016 DATA #



SEIFA_2016 <- SEIFA_2016[c(1:2,4,6,8)]


dat_2016_t <- merge(dat_2016_t, SEIFA_2016, all.x = T)

dat_2016_t[is.na(dat_2016_t)] <- NA

dat_2016_t <- na.omit(dat_2016_t)

colnames(dat_2016_t) <- c("SA1_7DIGITCODE_2016",

"TTP_2016",

186

"IRSD_Score",

"IRSAD_Score",

"IEO_Score",

"IER_Score")

rm(SEIFA_2016)

## Indigenous % data 2016 ##



dat_2016_t <- merge(dat_2016_t, IND_pct_2016, all.x = T)

dat_2016_t$pct_ind[is.na(dat_2016_t$pct_ind)] <- NA

dat_2016_t$pct_ind <- na.omit(dat_2016_t$pct_ind)

rm(IND_pct_2016)

# IND Truncated regression #

IND_2016_t <- truncreg(TTP_2016~pct_ind, data = dat_2016_t, point = 0, direction = "left")

# summary(IRSAD_2016_t)

# IND 2016 Generate r^2 estimates

dat_2016_t$IND_2016_t_yhat <- fitted(IND_2016_t)

r_IND_2016_t <- with(dat_2016_t, cor(TTP_2016, IND_2016_t_yhat))

r_sq_IND_2016_t <- r_IND_2016_t^2

rm(r_IND_2016_t)

# IRSAD Truncated regression #

IRSAD_2016_t <- truncreg(TTP_2016~IRSAD_Score, data = dat_2016_t, point = 0, direction =

"left")


# Generate r^2 estimates

dat_2016_t$IRSAD_2016_t_yhat <- fitted(IRSAD_2016_t)

r_IRSAD_2016_t <- with(dat_2016_t, cor(TTP_2016, IRSAD_2016_t_yhat))

r_sq_IRSAD_2016_t <- r_IRSAD_2016_t^2

rm(r_IRSAD_2016_t)

# IRSD Truncated regression #

187

IRSD_2016_t <- truncreg(TTP_2016~IRSD_Score, data = dat_2016_t, point = 0, direction = "left")

# summary(IRSD_2016_t)


dat_2016_t$IRSD_2016_t_yhat <- fitted(IRSD_2016_t)

r_IRSD_2016_t <- with(dat_2016_t, cor(TTP_2016, IRSD_2016_t_yhat))

r_sq_IRSD_2016_t <- r_IRSD_2016_t^2

rm(r_IRSD_2016_t)

# IER Truncated regression #

IER_2016_t <- truncreg(TTP_2016~IER_Score, data = dat_2016_t, point = 0, direction = "left")

# summary(IER_2016_t)


dat_2016_t$IER_2016_t_yhat <- fitted(IER_2016_t)

r_IER_2016_t <- with(dat_2016_t, cor(TTP_2016, IER_2016_t_yhat))

r_sq_IER_2016_t <- r_IER_2016_t^2

rm(r_IER_2016_t)

# IEO Truncated regression #

IEO_2016_t <- truncreg(TTP_2016~IEO_Score, data = dat_2016_t, point = 0, direction = "left")

# summary(IEO_2016_t)


dat_2016_t$IEO_2016_t_yhat <- fitted(IEO_2016_t)

r_IEO_2016_t <- with(dat_2016_t, cor(TTP_2016, IEO_2016_t_yhat))

r_sq_IEO_2016_t <- r_IEO_2016_t^2

rm(r_IEO_2016_t)

# Export all r^2 values #

r_sq_all_2016 <- c(r_sq_IRSAD_2016_t,r_sq_IRSD_2016_t,r_sq_IER_2016_t,r_sq_IEO_2016_t,

r_sq_IND_2016_t)

r_sq_all_2016 <- melt(data.frame(r_sq_all_2016))

r_sq_all_2016 <- r_sq_all_2016[c(2)]

r_sq_all_2016$y <- round(r_sq_all_2016$value/1*100, digits = 3)

188

colnames(r_sq_all_2016) <- c("Estimated r^2","Estimated % variance")

rownames(r_sq_all_2016) <- c("IRSAD","IRSD","IER","IEO","% Indig. Aus.")

#### trunc 2011 ####

# TTP 2011 Import #






dat_2011_t <- TTP_2011[c(5,11)]

colnames(dat_2011_t)[1] <- "SA1_7DIGITCODE_2011"


rm(TTP_2011)


# SEIFA 2011 DATA #



SEIFA_2011 <- SEIFA_2011[c(1:2,4,6,8)]





colnames(dat_2011_t) <- c("SA1_7DIGITCODE_2011",

"TTP_2011",

"IRSD_Score",

"IRSAD_Score",

"IEO_Score",

"IER_Score")

rm(SEIFA_2011)


189



IND_pct_2011 <- IND_pct_2011[c(1,4)]

colnames(IND_pct_2011)[c(1,2)] <- c("SA1_7DIGITCODE_2011","pct_ind")




rm(IND_pct_2011)







r_sq_IND_2011_t <- r_IND_2011_t^2



"left")







IRSD_2011_t <- truncreg(TTP_2011~IRSD_Score + pct_ind, data = dat_2011_t, point = 0,

direction = "left")






190

rm(r_IRSD_2011_t)







r_sq_IER_2011_t <- r_IER_2011_t^2

rm(r_IER_2011_t)







r_sq_IEO_2011_t <- r_IEO_2011_t^2

rm(r_IEO_2011_t)



r_sq_IND_2011_t)


r_sq_all_2011 <- r_sq_all_2011[c(2)]

r_sq_all_2011$y <- round(r_sq_all_2011$value/1*100, digits = 3)



#### trunc 2006 ####

# TTP 2006 Import #

191






dat_2006_t <- TTP_2006[c(4:5)]

colnames(dat_2006_t)[1] <- "CD_CODE06"


rm(TTP_2006)


# SEIFA 2006 DATA #


colnames(SEIFA_2006)[1] <- "CD_CODE06"

SEIFA_2006 <- SEIFA_2006[c(1:2,4,6,8)]





colnames(dat_2006_t) <- c("CD_CODE06",

"TTP_2006",

"IRSD_Score",

"IRSAD_Score",

"IEO_Score",

"IER_Score")

rm(SEIFA_2006)


IND_pct_2006 <- read.csv2("./INDI_CD_2006.csv", sep = ',', header = T, stringsAsFactors = F)


IND_pct_2006 <- IND_pct_2006[c(1,4)]

colnames(IND_pct_2006)[c(1,2)] <- c("CD_CODE06","pct_ind")



192


rm(IND_pct_2006)







r_sq_IND_2006_t <- r_IND_2006_t^2



"left")







IRSD_2006_t <- truncreg(TTP_2006~IRSD_Score, data = dat_2006_t, point = 0, direction = "left")






rm(r_IRSD_2006_t)





193



r_sq_IER_2006_t <- r_IER_2006_t^2

rm(r_IER_2006_t)







r_sq_IEO_2006_t <- r_IEO_2006_t^2

rm(r_IEO_2006_t)



r_sq_IND_2006_t)


r_sq_all_2006 <- r_sq_all_2006[c(2)]

r_sq_all_2006$y <- round(r_sq_all_2006$value*100, digits = 3)



#### Data and plot exports ####

## Data outputs ##

dir.create(file.path(getwd(), "outputs/truncated/"))

# truncated regression coefficients #

write.csv(summary(IND_2016_t)$coefficients, file = "./outputs/truncated/IND_2016_t.csv")

write.csv(summary(IRSAD_2016_t)$coefficients, file = "./outputs/truncated/IRSAD_2016_t.csv")

write.csv(summary(IRSD_2016_t)$coefficients, file = "./outputs/truncated/IRSD_2016_t.csv")

write.csv(summary(IER_2016_t)$coefficients, file = "./outputs/truncated/IER_2016_t.csv")

write.csv(summary(IEO_2016_t)$coefficients, file = "./outputs/truncated/IEO_2016_t.csv")

194











# R^2 estimates #

write.csv(r_sq_all_2016, file = "./outputs/truncated/r_sq_all_2016_t.csv")



#### Plots ####

ggplot(dat_2016, aes(TTP_2016)) +

geom_histogram()

# #### Unpublished code ####

#

# ## RA Data 2016 ##

# setwd('C:/Users/glenn/Dropbox/Uni/phd/@Research/Paper - Ch3. Environmental equity in


# RA_2016 <- read.table("./RA_2016_AUST.csv", sep = ',', header = T)

# # Clean data up & Filter out unnecessary columns

# RA_2016 <- RA_2016[c(3,5)]

# colnames(RA_2016 )[1] <- "SA1_7DIGITCODE_2016"

# colnames(RA_2016 )[2] <- "RA_2016"

195

# # rename values to remove state names

# RA_2016$RA_2016 <- gsub("^\\Migratory - Offshore - Shipping.*", "Migratory - Offshore -

Shipping", RA_2016$RA_2016)

# RA_2016$RA_2016 <- gsub("^\\No usual address.*", "No usual address", RA_2016$RA_2016)

# RA_2016$RA_2016 <- as.factor(RA_2016$RA_2016)

# RA_2016[,] <- as.numeric(as.character(unlist(RA_2016[,])))

# RA_2016 <- RA_2016[!duplicated(RA_2016), ]

# dat_2016 <- merge(dat_2016, RA_2016, all.x = T)

# rm(RA_2016)

#

#

# # # Merge

# # colnames(imp_complete)[2] <- "TTP_2016_imp"

# # dat_2016 <- merge(dat_2016, imp_complete, all.x = T)

#

# # ## Testing model - Adhoc imputation - and mice - vignette:

https://gerkovink.github.io/miceVignettes/Ad_hoc_and_mice/Ad_hoc_methods.html

# # fit2 <- with(data = dat_2016, expr = lm(log(TTP_2016_imp) ~ IRSD_Score + RA_2016))

# # mean(log(dat_2016$TTP_2016), na.rm = TRUE)

# # mean(fit$fitted.values)

# # confint(fit1, level = 0.9)

# # confint(fit2, level = 0.9)


Counts of ‘missing’ data per year

CD_CODE06 2006_IRSAD_Score

2006_IRSD_Score

2006_IER_Score

2006_IEO_Score

2006_pct_ind

TTP_2006

1 1 1 1 1 1 1

1 1 1 1 1 1 0

0 0 0 0 0 0 28851 SA1_7DIGITCODE_2016

2011_IRSAD_Score

2011_IRSD_Score

2011_IER_Score

2011_IEO_Score

2011_pct_ind

TTP_2011

1 1 1 1 1 1 1

1 1 1 1 1 1 0

0 0 0 0 0 0 37479 SA1_7DIGITCODE_2016

2016_IRSD_Score

2016_IRSAD_Score

2016_IEO_Score

2016_IER_Score

2016_pct_ind

TTP_2016

1 1 1 1 1 1 1

196

1 1 1 1 1 1 0

0 0 0 0 0 0 44952


Imputed data validity example

197

198

199


The text appears exactly as published.

Supplementary material S1

Food Security and conservation in the Tonle Sap: Interview guide

Glenn Althor, The University of Queensland, 2015.

This interview guide is written for the exclusive purpose to assist the interviewer while undertaking

interviews. It is broken up into sections which correspond to the project’s research questions. Only

the questions in bold will be directly asked of the participants. Each question is followed by a

checklist of actions and/or potential follow up questions. Follow up questions will only be asked if

the participants answers to open ended questions are insufficient or off topic.

Participant eligibility

How long have you lived in this village?

Must be at least 10 years, if not exit interview.

What is your main occupation?

Must be fisher, if not exit interview.

Read out participant information sheet

Offer participant to ask questions

Introductory comments before recording

Anonymity – data storage and management

Purpose of research, aims, expected outcomes.

Interested in learning from feelings, opinions, and experiences, no right or wrong answers.

Welcome to interrupt, ask for clarification, criticize line of questioning, refuse to answer, end

interview at any point.

Expected duration of 60-90 minutes, so at some points we may move along to adhere to time

requirements. But will take time to answer any questions for the participant.

Consent form.

Start recording.

Fish profile

o What are the main fish types that people in your village catch?

o Write down list of fish on seasonal calendar – try to get at least 6.

o What are the most common fish species eaten? – be insistent only require a common

name (Khmer or English)

200

What are the most common fish species that people in your village eat every day?

o Write on the seasonal calendar

What are the most common fish species that people in your village sell?

o Write on seasonal calendar

o Are different fish sold to middle merchants from other villagers?

Seasonal calendar by month

o Fill out fish section of the seasonal calendar

How do people obtain fish?

o Do people catch fish?

o How many people in the village catch fish?

o Do people catch all of the fish they eat or do people also buy fish?

o How many people in the village buy fish?

o Do people farm fish, if yes which species?

o How many people in the village farm fish?

Can you please draw a map of the areas where people in your community fish?

o Give participant pen and paper to draw a map.

o Ask them to draw wet season and dry season fishing areas (dry red, wet blue).

Over the last 10-20 years, do you think fish catch has increased, declined or stayed the same?

o By how much do you think fish catches changed?

o Have any particular fish species abundance changed? - just need a common name (Khmer

or English)

o Have fish changed in size, getting bigger or smaller? Which species?

o If they answer ‘declined’, also ask following question

Why do you think fish have declined?

o Is there overfishing?

o Has the population in the village changed in the last 10 years?

o What fishing gear are people commonly using?

o Has illegal fishing increased?

o Are there less flooded forests (fish habitat loss)?

o Are there weed problems (habitat degradation/eutrophication)?

o Are you aware of any efforts to reverse fish declines?

Impacts of Environmental variability

Is the fish catch between the annual wet and dry seasons, the same or different?

o Catch size in dry seasons, per year for last 5 years?

201

o Catch size in wet seasons, per year for last 5 years?

Is this year unusually, very dry (as dry as 2002)?

o Yes/No

Is fish catch different when it is very dry (drought years)?

o Difference in catch during drought years?

o Did the village undertake any new activities to relieve pressures in very dry years?

Do you remember when the flood was big flood in years, 2013, 2011 and 2000?

o Yes/No

Was the fish catch different in these big flood years?

o Difference in catch during flood years?

o Did the village undertake any new activities to relieve pressures in big flood years?

Does your community experience severe storms?

o Yes/No

Is the fish catch impacted by severe storms?

o Are the fishing conditions to poor to fish in?

o Does your fishing gear get damaged?

o Did the village undertake any new activities to relieve pressures in very dry years?

Conservation Area

Are you familiar with the nearby protected areas?

o Yes/No

Do people ever enter the protected areas to gather resources like fish or wood?

o Do people do other activities in the Prek Toal Core Zone such as bird egg harvest?

Some people think the bird sanctuary increases the fish catch in the area surrounding it, do

you think this happens here?

o Is fish catch better near the bird sanctuary?

o Where on the border of the sanctuary is catch better?

o Use participant’s map of area

Is access to the fisheries near the protected areas equal between the commune villages?

o Do other villages get more benefits and fewer costs from the protected areas?

Closing comments

o Are there any additional points the particpant wishes to make?

o Are there any remarks the participant wishes to clarify?

o Does the participant have any questions for the interviewer about interivew, project, etc?

202

203


204


205


Code index

Code title Definition

0 Trans Translation unclear - data not used.

1.0 Fish Not used.

1.1 Fish - catch Non-specific discussion of people or community catching fish

1.1.1 Fish - catch - freq Not used.

1.1.1.1 Fish - catch - freq - rare Species catch is rare

1.1.1.2 Fish - catch - freq - common Species catch is common

1.1.2 Fish - catch - use Catch utility

1.1.2.1 Fish - catch - use - eat Catch is for use as subsistence food

1.1.2.2 Fish - catch - use - sell Catch is for use as source of income

1.1.2.3 Fish - catch - use - farm feed Catch is for use as food in fish/crocodile farms

1.1.3 Fish - catch - seasonal Catch differences between seasons

1.1.3.1 Fish - catch - seasonal - dry Description of catch in dry season

1.1.3.2 Fish - catch - seasonal - wet Description of catch in wet season

1.1.3.3 Fish - catch - seasonal - drought Description of catch during drought periods

1.1.3.4 Fish - catch - seasonal - flood Description of catch during (extreme) floods

1.1.3.5 Fish - catch - seasonal - storm Description of catch during storms

1.1.3.6 Fish - catch - seasonal - month Monthly seasonality of catch

1.1.4 Fish - catch - area Areas where fishing takes place

1.1.4.1 Fish - catch - area - lake (within) Fishing within Tonle Sap lake

1.1.4.2 Fish - catch - area - lake (shore) Fishing on Tonle Sap shore

1.1.4.3 Fish - catch - area - stream Fishing within rivers/steams (note that lake is often translated as river)

1.1.4.4 Fish - catch - area - protected area (within)

Fishing within protected area

1.1.4.5 Fish - catch - area - protected area (border)

Fishing along protected area border

1.1.4.6 Fish - catch - area - community fishery Fishing in the community fishery (lot 3)

1.1.5 Fish - catch - species names Names of species caught by participant

1.1.4.7 Fish - catch - area - greatest Description of area where catch is greatest

1.1.4.8 Fish - catch - area - poor Description of area where catch is poor

1.1.4.9 Fish - catch - area - average Description of area where catch is average

1.2 Fish - description descriptions of fish species

1.3 Fish - size descriptions of fish sizes

1.3.1 Fish - size - decrease fish size decreases

1.3.1.1 Fish - size - decrease - reason reasons why the fish size has decreased

1.3.3 Fish - size - no change no change in fish sizes

1.3.2 Fish - size - increase fish size increases

1.3.2.1 Fish - size - increase - reason reasons why the fish size has decreased

1.4 Fish - Buy Non-specific discussion of people or community buying fish

1.4.1 Fish - Buy - Eat Fish are bought to eat

1.4.2 Fish - Buy - Farm Fish are bought to farm

1.5 Fish - farm Fish are farmed

1.5.1.1 Fish - farm - use - eat Farmed to eat

1.5.1.2 Fish - farm - use - sell Farmed to sell

206

1.5.2 Fish - farm - common Villagers commonly farm fish

1.5.3 Fish - farm - rare Villagers rarely farm fish

1.5.4 Fish - farm - species Descriptions of species that are farmed

1.5.5 Fish - farm - feed Descriptions of feed - not local catch (see 1.1.2.3)

2 Fishery Codes for fishery

2.1 Fishery - change Changes in fishery (over last 20 years)

2.1.1 Fishery - change - decline declines in fishery output

2.1.1.1.1 Fishery - change - decline - reason - other

reasons for declines - not in list below

2.1.1.1.2 Fishery - change - decline - reason - gear

declines due to fishing gear

2.1.1.1.3 Fishery - change - decline - reason - overfish

declines due to over harvesting of fish

2.1.1.1.4 Fishery - change - decline - reason - no jobs

declines due to no other employment opportunities

2.1.1.1.5 Fishery - change - decline - reason - over pop

declines due to too many people needing food

2.1.1.1.6 Fishery - change - decline - reason - conservation

declines due to the restrictions on people from conservation

2.1.1.1.7 Fishery - change - decline - reason - vietnamese

declines due to activities by Vietnamese people

2.1.2 Fishery - change - improve improvements in fishery output

2.1.2.1 Fishery - change - improve - reason reasons for improvements

2.1.1.2 Fishery - change - decline - human response

How have people respond to declines in the fishery

2.2 Fishery - access access to fisheries information

3 Climate change Mention of changing climate not related to fish catch

Natural capital and distributive justice - UQ eSpace

Documents