MEASURING THE SOCIAL COSTS OF COAL-BASED ELECTRICITY GENERATION IN SOUTH AFRICA by

MEASURING THE SOCIAL COSTS OF COAL-BASED ELECTRICITY GENERATION IN SOUTH AFRICA by
MEASURING THE SOCIAL COSTS OF COAL-BASED ELECTRICITY
GENERATION IN SOUTH AFRICA
by
Nonophile Promise Nkambule
A thesis submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy (PhD) in Economics
in the Faculty of Economic and Management Sciences
UNIVERSITY OF PRETORIA
SUPERVISOR: PROF. J.N. BLIGNAUT
CO-SUPERVISOR: PROF. J.H. VAN HEERDEN
FEBRUARY 2015
DECLARATION
I declare that this thesis, I hereby submit for the degree of Doctor of Philosophy in Economics at the
University of Pretoria is my own work and has not been submitted for a degree at this or any other
university.
Signed: ……………………………………………
Name: Nonophile Promise Nkambule
- ii -
DEDICATION
I dedicate this thesis to the Almighty without whom I would have not made it and to Lindiwe Millicent
Nkambule, my mother.
- iii -
ACKNOWLEDGEMENTS
I give all thanks to the Almighty God, who blessed me with strength, patience and health to carry out this
work. All exaltation to the Sovereign God for blessing me with support from all the people and institutions I
am thanking below. I am deeply thankful to my supervisor, Professor James N Blignaut, for his unceasing
guidance, support, encouragement, enthusiasm and timely feedback without which it would have been
difficult to complete this work. His prompt, intellectual ideas and suggestions are greatly appreciated. I
have grown a lot as a researcher working with him and have learnt valuable lessons, among which are
diligence, perseverance, thankfulness, collabouration and punctuality. I gratefully acknowledge Professor
Jan H van Heerden, my co-supervisor, for his constructive comments and ideas at each stage of the thesis. I
thank both my study leaders (supervisor and co-supervisor) for the countless technical discussions on the
COAL-based Power and Social Cost Assessment (COALPSCA) Model.
I would like to express my greatest gratitude to my mother, Lindiwe Millicent Nkambule for her emotional
support, love, encouragement and spiritual support throughout the duration of my studies. My gratitude
goes to the rest of my beloved family, especially my sister, Nomphilo Primrose Nkambule for assisting with
housekeeping when it was most needed.
I would like to gratefully acknowledge the enthusiastic willingness of Dr Douglas J Crookes for offering me
lectures on system dynamics modelling. His encouragement gave me the confidence to embark on
developing the COALPSCA Model. Many thanks to Mrs Gina Downes and her team in Eskom who helped in
providing the necessary data and information needed for this compilation. I would also like to convey
thanks to the National Research Fund for providing financial support. To the Centre for Environmental
Economics and Policy in Africa (CEEPA) I am grateful for the courses and periodic training in environmental
economics that reinforced my inspiration of focusing on environmental economics.
Special thanks go to my colleagues and PhD classmates at the Department of Economics, at CEEPA and
elsewhere for their encouragement, stimulating discussions and general advice, especially Dr Mulatu
Zerihon, Dr Josephine Musango, Dr Sana Abusin and Tumi. I am also grateful to Linda Vos for technical
support with proofreading the document. Special appreciation goes to Ms Marita, Ms Louis, Ms Sindi and
Ms Sonja for much needed support during the course of this work.
- iv -
MEASURING THE SOCIAL COSTS OF COAL-BASED ELECTRICITY
GENERATION IN SOUTH AFRICA
By
Nonophile Promise Nkambule
Supervisor: Prof. J.N. Blignaut
Co-supervisor: Prof. J.H. van Heerden
Department: Economics
Degree: PhD Economics
ABSTRACT
Energy technologies interact with the economic, social and environmental systems, and do so not only
directly but indirectly as well, through upstream and downstream processes. In addition, the interactions
may produce positive and negative repercussions. To make informed decisions on the selection of energy
technologies that assist a nation in reaping the socio-economic benefits of power generation technologies
with minimal effects on the natural environment, energy technologies need to be understood in the light of
the multifaceted system in which they function. But frequently, as disclosed by the literature review
conducted in this research, the evaluation of energy technologies lacks clear benchmarks of appropriate
assessments, which has resulted in difficulty to compare and to gauge the quality of various assessment
practices. The assessment methods and tools tend to be discipline specific with little to no integrations.
Parallel with the tools, the technology assessment studies offer piecemeal information that limits deeper
understanding of energy technologies and their consequent socio-economic-environmental repercussions.
Improved energy technology assessment requires the use of a holistic and integrative approach that
traverses the disciplinary nature of energy technology assessment tools, examines the long-term
implications of technologies while at the same time embracing energy technologies’ positive-and-negative
interactions with the economic, social and environmental systems and in this manner offering economic,
social and environmental indicators to assist decision makers in the decision-making process. Accordingly,
this study focuses on improving the assessment of energy technologies through the application of a holistic
and integrative approach, specifically system dynamics approach along a life-cycle viewpoint. Precisely,
-v-
focus is on coal-based electricity generation and in particular, the Kusile coal-fired power station near
eMalahleni as a case study.
A COAL-based Power and Social Cost Assessment (COALPSCA) Model was developed for: (i) understanding
coal-based power generation and its interactions with resource inputs, private costs, externalities,
externality costs and hence its consequent socio-economic, and environmental impacts over its lifetime
and fuel cycle; (ii) aiding coal-based power developers with a useful tool with a clear interface and graphical
outputs for detecting the main drivers of private and externality costs and sources of socio-environmental
burdens in the system; (iii) aiding energy decision makers with a visual tool for making informed energysupply decisions that takes into account the financial viability and the socio-environmental consequences of
power generation technologies; and for (iv) understanding the impacts of various policy scenarios on the
viability of coal-based power generation.
The validation of the COALPSCA Model was also conducted. Five structural validity tests were performed,
namely structure verification, boundary adequacy, parameter verification, dimensional consistency and
extreme condition tests. Behavioural validity was also conducted which included an analysis of the
sensitivity of the model outcomes to key parameters such as the load factor, discount rate, private cost
growth rates and damage cost growth rates using univariate and multivariate sensitivity analysis.
Finally, while attempts were made to incorporate most of the important aspects of power generation in a
coal-fired power plant, not all intrinsic aspects were incorporated due to lack of data, gaps in knowledge
and anticipated model complication. The shortcomings of the model were highlighted and
recommendations for future research were made.
Key Words: system dynamics, coal-fired power plant, externality, social cost, private cost, externality cost,
coal mine, plant construction, flue gas desulphurisation, economics
- vi -
TABLE OF CONTENTS
DECLARATION…………………………………………………………………………………….ii
DEDICATION………………………………………………………………………………………iii
ACKNOWLEDGEMENTS .................................................................................................. iv
ABSTRACT…………………………………………………………………………………………v
TABLE OF CONTENTS……......………………………………………………………………..vii
LIST OF TABLES………………………………………………………………………………...xii
LIST OF FIGURES………………………………………………………………………………xiv
ACRONYMS AND ABBREVIATIONS.............................................................................. xvi
CHAPTER 1:
INTRODUCTION ....................................................................................... 1
1.1
SOUTH AFRICA’S ECONOMIC DEVELOPMENT & ENVIRONMENTAL AND
DEVELOPMENT PLANNING PROCESS ........................................................................ 1
1.2
EIA REGULATION AND PROCESS IN SOUTH AFRICA ................................................ 2
1.3
EIA EFFECTIVENESS AND WEAKNESSES .................................................................. 4
1.4
POSSIBLE SUGGESTIONS TO SELECTED EIA ISSUES .............................................. 7
1.5
TECHNOLOGY AND TECHNOLOGY ASSESSMENT .................................................. 10
1.6
TECHNOLOGY ASSESSMENT SHORTCOMINGS AND SOLUTIONS ........................ 12
1.7
PROBLEM STATEMENT .............................................................................................. 16
1.8
RATIONALE FOR SYSTEM DYNAMICS APPROACH AND A LIFE-CYCLE
VIEWPOINT .................................................................................................................. 18
1.9
RESEARCH OBJECTIVES............................................................................................ 19
1.10
ORGANISATION OF THESIS ....................................................................................... 20
CHAPTER 2:
EXTERNALITIES AND SOUTH AFRICA’S POWER AND COAL
INDUSTRIES............................................................................................................. 21
2.1
INTRODUCTION ........................................................................................................... 21
2.1
EXTERNALITIES DEFINED .......................................................................................... 21
2.2
THE ENVIRONMENTAL AND SOCIETAL IMPACTS LINKED WITH THE COALFUEL CYCLE ................................................................................................................ 24
2.2.1.1
2.2.1.2
2.2.1.3
2.3
Coal mining and transportation impacts ......................................................... 25
Plant construction impacts ............................................................................. 26
Plant operation impacts ................................................................................. 28
SOUTH AFRICA’S POWER INDUSTRY ....................................................................... 29
2.3.1
2.3.2
2.3.3
2.3.4
Eskom’s power stations ......................................................................................... 30
Eskom’s electricity sales ........................................................................................ 31
Coal quality and emissions profile of Eskom’s coal-based plants ........................... 32
Eskom’s coal supply methods and coal supply contracts ....................................... 33
- vii -
2.4
SOUTH AFRICA’S COAL INDUSTRY ........................................................................... 34
2.4.1
2.4.2
2.4.3
2.4.4
2.4.5
2.5
Coal production and consumption .......................................................................... 34
Coal prices ............................................................................................................. 35
Domestic coal consumers ...................................................................................... 37
Export coal consumers .......................................................................................... 38
South Africa’s coal producers ................................................................................ 40
SUMMARY .................................................................................................................... 42
CHAPTER 3:
ECONOMIC
PHILOSOPHY
AND
SYSTEM
DYNAMICS
PHILOSOPHY ........................................................................................................... 43
3.1
INTRODUCTION ........................................................................................................... 43
3.2
RESEARCH PARADIGMS DEFINED ............................................................................ 43
3.3
GUBA AND LINCOLN’S SOCIAL SCIENCE RESEARCH PARADIGM
FRAMEWORK ............................................................................................................... 43
3.3.1
3.3.2
3.3.3
3.4
Positivists and post-positivists ................................................................................ 44
Constructivists ....................................................................................................... 45
Critical theorists ..................................................................................................... 45
ECONOMIC DISCIPLINES AND RESEARCH PARADIGMS......................................... 46
3.4.1
Mercantilism and Physiocracy................................................................................ 46
3.4.1.1
Criticism of the early political economy schools ............................................. 47
3.4.1.2
Appraisal........................................................................................................ 48
3.4.2
Classical economics school ................................................................................... 48
3.4.2.1
Appraisal........................................................................................................ 49
3.4.3
Neoclassical economics school ............................................................................. 49
3.4.3.1
Appraisal........................................................................................................ 50
3.4.4
Heterodox economics ............................................................................................ 51
3.4.4.1
Austrian economics and appraisal ................................................................. 51
3.4.4.2
Institutional economics and appraisal ............................................................ 51
3.4.5
Environmental economics and ecological economics............................................. 52
3.4.5.1
Appraisal........................................................................................................ 53
3.5
RESEARCH PARADIGM(S) UNDERPINNING THE CURRENT STUDY....................... 53
3.6
SYSTEM DYNAMICS ORIGINS, FEATURES, MODELLING PROCESS AND ITS
LINKS WITH SOCIAL THEORIES ................................................................................. 54
3.6.1
Origins of systems theory....................................................................................... 54
3.6.2
Origins of system dynamics and its main features ................................................. 55
3.6.3
System dynamics modelling process ..................................................................... 58
3.6.3.1
Problem formulation ....................................................................................... 58
3.6.3.2
Dynamic hypothesis formulation .................................................................... 59
3.6.3.3
Model formulation .......................................................................................... 60
3.6.3.4
Model validation ............................................................................................. 61
3.6.3.5
Policy design and evaluation.......................................................................... 62
3.6.4
Social theoretic assumptions underpinning system dynamics practice ................... 63
- viii -
3.7
ONTOLOGY, EPISTEMOLOGY AND METHODOLOGIES OF ENERGY-RELATED
SYSTEM DYNAMICS PRACTICES ............................................................................... 66
3.8
THE
ONTOLOGICAL,
EPISTEMOLOGICAL
AND
METHODOLOGICAL
PLACEMENT OF THIS CURRENT STUDY AND ITS LINKS WITH THE
ECONOMICS SCHOOLS UNDERLYING THIS STUDY ................................................ 67
3.9
SUMMARY .................................................................................................................... 68
CHAPTER 4:
A REVIEW OF POWER GENERATION ASSESSMENT TOOLS
AND THEIR APPLICATION...................................................................................... 70
4.1
INTRODUCTION ........................................................................................................... 70
4.2
POWER GENERATION TECHNOLOGIES ASSESSMENT TOOLS ............................. 70
4.2.1
Financial analysis methods .................................................................................... 71
4.2.2
Impact analysis methods ....................................................................................... 73
4.2.2.1
Valuation of externalities in theory ................................................................. 73
4.2.2.2
Valuation of externalities in practice ............................................................... 74
4.2.2.3
Controversies of valuing human life ............................................................... 82
4.2.3
Systems analysis methods..................................................................................... 84
4.3
A REVIEW OF POWER GENERATION STUDIES ........................................................ 89
4.3.1
4.3.2
4.3.3
4.3.4
4.3.5
4.3.6
4.3.7
CHAPTER 5:
International studies assessing power generation private costs ............................. 89
International studies assessing power generation externality costs........................ 92
International studies modelling power generation systems..................................... 94
Local studies assessing power generation private costs ........................................ 99
Local studies assessing power generation externality costs................................. 101
Local studies modelling power generation systems.............................................. 102
Summary ............................................................................................................. 103
RESEARH DESIGN AND METHODS ................................................... 106
5.1
INTRODUCTION ......................................................................................................... 106
5.2
RESEARCH PARADIGM/PHILOSOPHY ..................................................................... 106
5.3
DESCRIPTION OF INQUIRY STRATEGY .................................................................. 107
5.4
RESEARCH METHOD ................................................................................................ 109
5.4.1
Study site: Kusile power station and supporting collieries .................................... 109
5.4.2
Boundary of the study .......................................................................................... 112
5.4.3
Data collection process ........................................................................................ 114
Compiling an inventory of materials and resources requirements ................ 115
5.4.3.1
5.4.3.2
Compiling an inventory of private costs ........................................................ 115
5.4.3.3
Compiling an inventory of environmental and societal burdens .................... 115
5.4.3.4
Collection of economic valuation data .......................................................... 116
5.4.4
System dynamics modelling................................................................................. 119
5.5
CONCLUSION............................................................................................................. 121
- ix -
CHAPTER 6:
COAL-BASED POWER AND SOCIAL COST ASSESSMENT
(COALPSCA) MODEL ............................................................................................ 123
6.1
INTRODUCTION ......................................................................................................... 123
6.2
SOFTWARE USED IN THE MODELLING ................................................................... 123
6.3
PROBLEM FORMULATION ........................................................................................ 123
6.4
DYNAMIC HYPOTHESIS FORMULATION ................................................................. 126
6.5
MODEL BOUNDARY................................................................................................... 129
6.6
MODEL FORMULATION: STRUCTURE AND EQUATIONS ....................................... 130
6.6.1
6.6.2
6.6.3
6.6.4
6.6.5
6.6.6
6.6.7
6.6.8
6.6.9
6.7
Power generation sub-model ............................................................................... 131
Generation cost sub-model .................................................................................. 135
Morbidity and fatalities sub-model ........................................................................ 143
Water consumption sub-model ............................................................................ 150
Water pollution sub-model ................................................................................... 155
Ecosystem services loss sub-model .................................................................... 157
Air pollution sub-model ........................................................................................ 160
Global pollutants sub-model................................................................................. 164
Social cost sub-model .......................................................................................... 168
SUMMARY .................................................................................................................. 174
CHAPTER 7:
RESULTS .............................................................................................. 175
7.1
INTRODUCTION ......................................................................................................... 175
7.2
BASELINE RESULTS.................................................................................................. 175
7.2.1
7.2.2
7.2.3
7.2.4
7.2.5
7.3
Electricity generation............................................................................................ 176
Private costs ........................................................................................................ 177
Externalities inventory .......................................................................................... 181
Externality costs ................................................................................................... 185
Social cost ........................................................................................................... 190
MODEL VALIDATION.................................................................................................. 193
7.3.1
Structural validity ................................................................................................. 193
7.3.1.1
Structure verification test ............................................................................. 194
7.3.1.2
Dimensional consistency test ....................................................................... 195
7.3.1.3
Parameter verification test ........................................................................... 195
7.3.1.4
Extreme condition test ................................................................................. 196
Behaviour validity ................................................................................................. 198
7.3.2
7.3.2.1
Face validity test, reference test and modified-behaviour prediction test ...... 198
7.3.2.2
Behaviour sensitivity test/sensitivity analysis ............................................... 199
7.4
POLICY ANALYSIS ..................................................................................................... 209
7.4.1
7.4.2
7.5
Export parity coal pricing scenarios...................................................................... 210
Carbon tax scenarios ........................................................................................... 212
SUMMARY .................................................................................................................. 217
-x-
CHAPTER 8:
CONCLUSION, LIMITATIONS AND RECOMMENDATIONS ............... 221
8.1
THE RESEARCH CONDUCTED IN THIS STUDY AND MAIN FINDINGS .................. 221
8.2
COALPSCA MODEL LIMITATIONS ............................................................................ 225
8.3
WHAT COULD BE DONE TO IMPROVE THE COALPSCA MODEL AND ENERGY
RESEARCH ................................................................................................................ 226
8.4
WAY FORWARD FOR THE SOUTH AFRICAN GOVERNMENT ................................ 227
REFERENCES ................................................................................................................ 230
APPENDICES ................................................................................................................. 288
APPENDIX A: COALPSCA MODEL EQUATIONS .................................................................. 288
- xi -
LIST OF TABLES
Table 2.1: Coal-fuel cycle environmental and societal impacts .......................................... 25
Table 2.2: Eskom’s existing and future coal-fired power stations ...................................... 31
Table 2.3: Coal quality and emissions profile of Eskom’s coal-fired power plants ............. 33
Table 2.4: South Africa’s historic coal production and consumption (tonx10 6)................... 35
Table 2.5: SA’s historic coal price with and without adjustment (R/ton) ............................. 37
Table 3.1: Research paradigms ......................................................................................... 44
Table 3.2: System dynamics modelling process ................................................................ 58
Table 3.3: Pruyt’s extended paradigm table....................................................................... 63
Table 4.1: Power generation technology assessment tools and methods ......................... 70
Table 4.2: International studies assessing power generation private costs (2010 values) . 90
Table 4.3: International studies on power generation externality costs (2010 values) ....... 94
Table 4.4: Local studies assessing power generation private costs ................................ 100
Table 4.5: Local studies assessing power generation externality costs (2010 values) .... 101
Table 6.1: Endogenous, exogenous and excluded variables ........................................... 130
Table 6.2: Parameters used in the power generation sub-model..................................... 135
Table 6.3: Parameters used in the generation cost sub-model ........................................ 143
Table 6.4: Parameters used in the morbidity and fatalities sub-model ............................. 150
Table 6.5: Parameters used in the water consumption sub-model .................................. 155
Table 6.6: Parameters used in the water pollution sub-model ......................................... 157
Table 6.7: Parameters used in the ecosystem services loss sub-model .......................... 159
Table 6.8: Parameters used in the air pollution sub-model .............................................. 164
Table 6.9: Parameters used in the global pollutants sub-model ...................................... 168
Table 7.1: Baseline scenario input parameters ................................................................ 176
Table 7.2: Economic and socio-environmental indicators ................................................ 176
Table 7.3: Baseline scenario electricity production (gross and net) ................................. 177
Table 7.4: Coal consumption (Million tons) ...................................................................... 177
- xii -
Table 7.5: Baseline scenario private costs over Kusile’s lifetime ..................................... 178
Table 7.6: Coal-fuel cycle water use (Million m3) over Kusile’s lifetime............................ 182
Table 7.7: Coal-fuel cycle fatalities and morbidity over Kusile’s lifetime .......................... 182
Table 7.8: Coal-fuel cycle land use (Hectares) ................................................................ 183
Table 7.9: Coal-fuel cycle classic air pollutant loads over Kusile’s lifetime ...................... 183
Table 7.10: Annual emissions of classic air pollutants - coal combustion (Thousand t) ... 184
Table 7.11: Coal-fuel cycle greenhouse gas pollutant loads over Kusile’s lifetime .......... 185
Table 7.12: Annual emissions of greenhouse gases - coal combustion (Million t) ........... 185
Table 7.13: Coal-fuel cycle externality cost (Billion Rands) over Kusile’s lifetime ............ 186
Table 7.14: FGD system or not, costs and savings (Billion Rands) over Kusile’s lifetime 190
Table 7.15: Levelised externality and social cost of energy (R/MWh) over Kusile’s lifetime
........................................................................................................................ 191
Table 7.16: Selected present value output (Billion Rands) .............................................. 192
Table 7.17: Examples of structure test............................................................................. 195
Table 7.18: Selected parameters, values and data sources - power generation ............. 196
Table 7.19: Minimum and maximum parameter values versus baseline values .............. 200
Table 7.20: Lower and higher range damage cost estimates versus baseline values ..... 200
Table 7.21: Lower and higher range lifetime externality costs versus baseline over the
lifetime of Kusile .............................................................................................. 207
Table 7.22: Lower and higher range life-time externality costs per phase versus baseline
........................................................................................................................ 208
Table 7.23: Policy scenarios ............................................................................................ 210
- xiii -
LIST OF FIGURES
Figure 2.1: Externality costs of production ......................................................................... 23
Figure 2.2: Eskom’s electricity sales by sector 2010/11..................................................... 32
Figure 2.3: SA’s historic coal production and consumption (tx10 6) .................................... 35
Figure 2.4: Domestic and export coal prices in R/ton (not adjusted) .................................. 37
Figure 2.5: SA’s domestic coal users by sector 2011 ........................................................ 38
Figure 2.6: South Africa’s coal export sales by region 2011 .............................................. 39
Figure 2.7: South Africa’s active coal mines ...................................................................... 42
Figure 3.1: Positive and negative causality ........................................................................ 59
Figure 3.2: Positive and negative feedback loops .............................................................. 60
Figure 3.3: Stock and flow diagram ................................................................................... 61
Figure 6.1: Causal loop diagram of the modelled system ................................................ 127
Figure 6.2: Power generation sub-model stock and flow diagram ................................... 131
Figure 6.3: Lookup function for effect of profitability on desired functional capacity ........ 134
Figure 6.4: Generation cost sub-model stock and flow diagram ...................................... 140
Figure 6.5: Morbidity and fatalities sub-model stock and flow diagram ............................ 147
Figure 6.6: Water consumption sub-model stock and flow diagram................................. 154
Figure 6.7: Water pollution sub-model stock and flow diagram........................................ 156
Figure 6.8: Ecosystem services loss sub-model stock and flow diagram......................... 158
Figure 6.9: Air pollution sub-model stock and flow diagram ............................................. 160
Figure 6.10: Global pollutants sub-model stock and flow diagram ................................... 167
Figure 6.11: Social cost sub-model stock and flow diagram ............................................ 169
Figure 7.1: LCOE outputs ................................................................................................ 179
Figure 7.2: Present value costs and revenues ................................................................. 181
Figure 7.3: Social NPV..................................................................................................... 192
Figure 7.4: COALPSCA Model behaviour under extreme condition test 1 ....................... 197
Figure 7.5: COALPSCA Model behaviour under extreme condition test 2 ....................... 198
- xiv -
Figure 7.6: Confidence bounds for discount rate (range: 0.04 to 0.12) on selected model
outcomes ........................................................................................................ 201
Figure 7.7: Confidence bounds for private cost growth rates (range: -0.05 to 0.05) on
selected model outcomes ............................................................................... 202
Figure 7.8: Confidence bounds for load factor (range: 0.85 to 0.95) on selected model
outcomes ........................................................................................................ 203
Figure 7.9: Confidence bounds for damage cost growth rates (range: -0.0055 to 0.0165)
on selected model outcomes .......................................................................... 204
Figure 7.10: Confidence bounds for low and high damage costs estimates (range: Table
7.20) on selected model outcomes ................................................................. 205
Figure 7.11: Confidence bounds for of all uncertain parameters on selected model
outcomes (multivariate) ................................................................................... 206
Figure 7.12: Export parity price outcomes ....................................................................... 212
Figure 7.13: Carbon tax cost at various tax regimes and growth rates ............................ 215
Figure 7.14: Carbon tax effects on LCOE ........................................................................ 216
Figure 7.15: NPV at various tax regimes and growth rates .............................................. 217
- xv -
ACRONYMS AND ABBREVIATIONS
AAIC
Anglo American Inyosi Coal
ATSE
Australian Academy of Technological Sciences and Engineering
BDFM
Business Day and Financial Mail
B-BBEE
Broad-Based Black Economic Empowerment
BNEF
Bloomberg New Energy Finance
BP
Beyond Petroleum
c
cents
CCS
Carbon Capture and Storage
CEEPA
Centre for Environmental Economics and Policy in Africa
CF
Capacity Factor
CO2
Carbon Dioxide
CO2e
Carbon Dioxide equivalence
COALPSCA
COAL-based Power and Social Cost Assessment
CVM
Contingent Valuation Method
EIA
Environmental Impact Assessment
EPRI
Electric Power Research Institute
ESP
ElectroStatic Precipitator
ESPs
ElectroStatic Precipitators
etc.
et cetera
EU
European Union
FBC
Fluidized Bed Combustion
FGD
Flue Gas Desulphurisation
FMA
Financial Model Approach
FOB
Free On Board
FOR
Free On Rail
GDP
Gross Domestic Product
GHG
Greenhouse Gas
GWh
Gigawatt hour
IEA GHG
International Energy Agency Greenhouse Gas R&D Programme
IGCC
Integrated Gasification Combined Cycle
IOA
Input-Output Analysis
- xvi -
ACRONYMS AND ABBREVIATIONS (CONTD.)
IPCC
International Panel on Climate Change
IRP
Integrated Resource Plan
IRR
Internal Rate of Return
kcal
kilocalorie
kg
kilogram
kt
kilotons
kWh
kilowatt hour
LCA
Life Cycle Assessment
LCC
Life Cycle Cost
LCOE
Levelised Cost of Energy
LECOE
Levelised Externality Cost of Energy
LSCOE
Levelised Social Cost of Energy
LEAP
Long-range Energy and Alternatives Planning
MIRR
Modified Internal Rate of Return
MIT
Massachusetts Institute of Technology
MJ
Megajoule
Mt
Mega tons
MW
Megawatt
MWh
Megawatt hour
N2O
Nitrous Oxide
NERSA
National Energy Regulator of South Africa
NO2
Nitrogen Dioxide
NOx
Oxide of Nitrogen
NPV
Net Present Value
OECD
Organisation for Economic Co-operation and Development
O&M
Operation and Maintenance
ORNL & RfF
Oak Ridge National Labouratory and Resources for the Future
PC
Pulverised Combustion
PE
Policy Engineering
PM
Particulate Matter
PPP
Public Private Partnership
R
Rand
- xvii -
ACRONYMS AND ABBREVIATIONS (CONTD.)
RB
Richards Bay
ROM
Run-Of-Mine
SCR
Selective Catalytic Reduction
SD
System Dynamics
sLCOE
simplified Levelised Cost of Energy
SNPV
Social Net Present Value
SO2
Sulphur Dioxide
USEIA
United States Energy Information Administration
t
ton
T21
Threshold 21
UK
United Kingdom
US
United States
USAID
United States Agency International Development
VOLY
Value of a Life Year
W/
With
W/O
Without
WCA
World Coal Association
WTP
Willingess to Pay
ZAR
South African Rand
$
Dollars
%
Per cent
∂
Discount factor
- xviii -
CHAPTER 1:
1.1
INTRODUCTION
South Africa’s economic development & environmental and development planning process
In the initial stages of economic development South Africa relied greatly on its rich mineral resources.
Primary capital was mainly accumulated from the mining sector, out of which South Africa supported a
strong manufacturing sector (Blignaut & Hassan, 2001). The contribution of mining to Gross Domestic
Product (GDP) was about 13.2% in 1970 (South African Reserve Bank cited by Blignaut & Hassan, 2001) and,
nowadays, the mining sector still remain the main stimulant behind the development of the country’s
economy, contributing about 8.8% to GDP. The sectors contribution is about 18% if one factor in the mining
sector’s indirect and induced effects (Chamber of Mines, 2012). With regards to the range of minerals and
quantities produced, South Africa is one of the world’s leading mining countries (Statistics South Africa,
2012) and its mineral industry is mainly based on coal, gold and platinum group metals.
Coal is South Africa’s main source of energy, providing over 70% of its primary energy and 93% of its
electricity (Department of Energy, 2010; World Coal Association (WCA) (2010). Owing to the development
of the economy, and the fact that South Africa has not recently invested in augmenting its power
generation supply capacity, the maximum production capacity of the existing power stations has been
reached (Department of Energy, 2009). The South African government has planned energy projects, to
augment its electric power supply reflected in its Integrated Resource Plan (IRP). The IRP investigates how
South Africa’s electricity demand can be met between 2010 and 2030. The plan include investing in various
energy technologies including pulverised combustion plants, Fluidized Bed Combustion (FBC) plants,
Integrated Gasification Combined Cycle (IGCC) plants, nuclear plants and renewable energy sources like
solar and wind (IRP, 2011). On the other hand, Eskom, a utility owned by the state and that dominate the
country’s power industry, has begun constructing new two coal-fired power stations in an effort to meet
the country’s growing demand for electricity, namely the Kusile and Medupi power stations in eMalahleni
and Limpopo, respectively (Eskom, 2011; Eskom, 2012a; Eskom. 2013a).
Generally, the environmental and development planning process in the form of an Environmental Impact
Assessment (EIA) have been the main driver of project development in South Africa (Hoosen, 2010). EIA is a
project-oriented environmental assessment tool for assessing the impacts of planned activities on the
environment - the environment is broadly defined to include the economic, social and natural dimensions
(Hugo, 2004; Southern African Institute for Environmental Assessment, 2004; Department of Environmental
Affairs and Tourism (DEAT), 2008; Department of Environmental Affairs (DEA), 2010). The proper
-1-
assessment of the ecological, economic and social effects of planned development projects is therefore a
fundamental process of an EIA, which provides essential information to decision-makers on the potential
impacts of planned projects and hence place them in a better position to make informed decisions on
whether or not approval should be granted. EIA ensure that activities that are unacceptably damaging to
the environment are not authorized and that those that are authorized are carried out in a way that the
environmental impacts are minimized or mitigated to acceptable levels (DEAT, 2006a). It is therefore
appreciated as a process for minimizing/mitigating the adverse impacts of development on the
environment early in the design stages (Ministry of Environment and Tourism, 1997; Hoosen, 2010).
This study focuses on the energy sector and specifically on power generation developments, that is, on
developments that are complex, that require multi-billion Rand investments and that are associated with
diverse and long-lasting environmental and societal effects at some points in their fuel cycle (World Energy
Council, 2004; Georgakellos, 2010). The following sections outline the history of the EIA process in South
Africa, discuss EIA effectiveness and weaknesses, and reflect on possible solutions to selected EIA issues.
Later on, the project-orientated manner of EIA shifts the direction of this background section to technology
assessment owing to its broader scope. Following this route, a discussion of technology assessment
concept, various types, functional elements, shortcomings and possible solutions is conducted. This
information provides the foundation for framing the research problem and subsequent research objectives.
1.2
EIA regulation and process in South Africa
The EIA process in the country began on non-obligatory grounds in the 1970s. During this time it was
undertaken out of free will as a component of Integrated Environmental Management. In September 1997,
EIA became mandatory with the declaration of EIA regulations in the form of the Environment Conservation
Act (ECA) of 1989 (South Africa, 1989, 1997a,b). The Act provided EIA procedures, which were incorporated
into the regulation (Republic of South Africa, 1998). In addition, the 1997 EIA regulations were
supplemented by EIA guidelines which charted the application for authorization and the steps of the EIA
process (Department of Environmental Affairs and Tourism (DEAT), 1998).
The key steps of the EIA process included: the submission of application for authorization; screening;
scoping report writing (which included a plan of study for EIA coupled with extensive public participation);
Environmental Impact Report (EIR) preparation which included specialist reports, public involvement and
draft environmental management plan; review of environmental impact report by the competent
authority; decision making; and monitoring (Wood, 1999). Notable from the early regulation was the
requirement for extensive public participation and comprehensive scoping for all projects. About 80% of
-2-
the proposed developments were therefore authorized on grounds of an extended scoping report (South
Africa, 1997a), also referred to by Sandham, Siphugu and Tshivhandekano (2005) as the mini-EIA or beefedup scoping report. Furthermore, a large number of development projects were subjected to a full EIA
process (Hoosen, 2010). The compounded results of the 1997 EIA process was a lengthy and expensive
administrative process (Sandham, van Heerden, Jones, Retief & Morrison-Saunders, 2013) which created
bottlenecks in EIA authorization which were perceived to be retardation the country’s development
(Swanepoel, 2008).
In 2006, new EIA regulations were announced in terms of the National Environmental Management Act
(NEMA) and they substituted the Environmental Conservation Act EIA regulations (DEAT, 2006a; South
Africa, 2006). The 2006 regulation key changes comprised of the institution of time frames, extending
developments that required EIA (e.g. mining), consideration of alternatives to a proposed development,
provision for monitoring after authorization and separating the environmental assessment processes into
basic and full assessments. The former is suited for developments characterized by minor environmental
effects while the latter is undertaken for developments with potentially significant environmental impacts,
for example activities characterized by high levels of pollution, land generation and waste generation
(DEAT, 2006a; DEAT, 2006b; Kidd & Retief, 2009). Electricity generation and mining activities therefore fall
under the full EIA process (Department of Environmental Affairs and Development Planning, 2006).
The assessment process in the new regulation is broadly separated into a basic assessment, scoping
procedure and appeal procedure (DEAT, 2006a). Generally the main difference between the assessment
process for a basic and full assessment is that a full EIA is subjected to a more detailed scoping assessment
than a basic assessment, it requires an EIA and in addition it requires a submission of an application prior to
the competent authority (DEAT, 2006a; Republic of South Africa, 2006). For the full assessment, upon
submitting the application form and authorization of the scoping report and the EIA proposal, the
Environmental Assessment Practitioner (EAP) proceeds with the EIA proposed study. The aim of the EIA is
to address concerns raised in the course of scoping, to assess impacts, to determine impacts significance, to
frame mitigation actions and to assess in comparative manner alternatives to the proposed activity (DEAT,
2006b; DEAT, 2006c). The new EIA regulation reduced the portion of full EIAs undertaken, quickened the
completion of EIAs due to commenting period timeframes and eliminated EIA authorization backlogs
(Swanepoel, 2008). With the aim of improving EIA effectiveness and efficiency, in 2010 the third set of EIA
regulations became operative (Hildebrandt & Sandhamb, 2014).
-3-
1.3
EIA effectiveness and weaknesses
Ever since the initiation of EIA in the United States of America in 1970 and its successive adoption by
various governments and environmental agencies around the world, its effectiveness has been a subject of
interest to scholars and the EIA practice community (Christensen, Kornov & Nielsen, 2005; Retief &
Chabalala, 2009; Heinma & Poder, 2010). The shared response by authorities to perceived poor EIA system
performance is to adjust the governing legislation, for example, the appraisals of the Canadian and South
African EIA systems (DEAT, 2006a; DEAT, 2008a; Standing Committee on Environment and Sustainable
Development, 2011).
EIA effectiveness generally refers to two criteria, namely whether the EIA process achieves its objectives
and a procedural criterion which pertains to level of conformity with procedural requirements (Cashmore,
Gwilliam, Morgan, Cobb & Bond, 2004; Glasson, Therivel & Chadwick, 2005; Jay, Jones, Slinn & Wood,
2007). It is argued that though research has focused on the procedural side of EIA effectiveness, the
evaluation of EIA effectiveness with regards to its goals is a better measure of EIA effectiveness (Cashmore
et al. 2004; Jay et al., 2007). Since EIR holds project information in decision-making, it is commonly
acknowledged that EIRs of poor quality contribute to EIA ineffectiveness (Wood, 2003; Glasson et al.,
2005).
Many researchers have therefore evaluated the quality of EIRs in South Africa (Sandham et al., 2005; Van
der Vyver, 2008; Mbhele, 2009; Hildebrandt & Sandhamb, 2014; Sandham, Carroll & Retief, 2010) and
internationally (Androulidakis & Karakassis, 2005; Pinho, Maia & Monterrosa, 2007; Heinma & Poder, 2010;
Jalava, Pasanen, Saalasti & Kuitunen, 2010) in an effort to gauge the effectiveness of the EIA process.
The literature in developed, developing and transitional economies over the world disclosed that the
quality of EIAs and decisions involved had improved over the past half-century with improvement of EIA
procedures, enhancement of EIA capacity, increased use of mitigation measures and the occasional nonexecution of potentially environmental damaging undertakings that would otherwise would have been
permitted (Barker and Wood, 1999; Canelas, Almansa, Merchan & Cifuentes, 2005; Lee, 2000; Jay et al,
2007; Polonen, Hokkanen & Jalava, 2010).
Despite the improvements of EIAs worldwide, the weaknesses usually faced relate to limited capacity of
authorities, inadequate public participation, limited scope of impact assessments, poor consideration of
project alternatives, impact prediction challenges, inadequate consideration of cumulative impacts and
-4-
inadequate follow-up monitoring (Barker and Wood, 1999; Gray and Edwards-Jones, 1999; Jay et al., 2007;
Tzoumis, 2007; Kruopiene, Zidonienė & Dvorioniene, 2009; Peterson, 2010).
In South Africa researchers have highlighted more or less similar shortfalls, for instance insufficient public
participation (DEAT, 2006a; Hoosen, 2010), lack of political will, for example EIA has been blamed for
delaying construction by government officials, lack of skilled government officials to conduct EIR review
(Sandham & Pretorius, 2008; Hildebrandt & Sandhamb, 2014), inconsistency in EIR review (DEAT, 2006a;
Sandham & Pretorius, 2008), lack of a reference frame for EAP to adhere to (Sandham & Pretorius, 2008),
inadequate use of assessment methodologies (Sandham et al., 2010; Sandham & Pretorius, 2008) and poor
EIR report quality especially with regards to the provision of information pertaining to impact identification,
impact magnitude prediction, impact significance assessment, project alternatives, mitigation measures
and monitoring (Sandham et al., 2013). The study by Sandham et al. (2013) is in essence a comparative
study of the quality of EIRs conducted under the 2006 and 1997 EIA regulation. 3 of the 7 EIRs under the
1997 EIA system and 8 of the 11 EIRs under the 2006 were planned developments by Eskom (Electricity
Supply Commission). The reports were reviewed under four appraisal areas, namely: development
description; local environment and baseline conditions; impact identification and evaluation; and,
alternatives and mitigation. The more analytical tasks (i.e. impact identification and evaluation, and
alternatives and mitigation) which form the basis for decision making performed poorly, along with
monitoring.
Analogous analyses of the - South African mining industry EIRs (Sandham, Hoffmann & Retief, 2008), quality
of EIRs of various developments in the North West province of South Africa (Sandham & Pretorius, 2008),
and quality of EIRs in the context of biological pest control in the Limpopo province (Sandham et al., 2010),
also disclosed generally satisfactory grades in descriptive and presentational areas of EIRs while impact
identification, prediction and evaluation, and alternatives and mitigation measures remain weaker aspects
of EIRs. Such poorer scores in the more analytical components of the EIRs are shared also internationally
(Barker & Wood, 1999; Lee, 2000; Polonen et al., 2010; Barker & Jones, 2013).
Since the environment is defined widely in the NEMA (Aucamp, Woodborne, Perold, Bron & Aucamp, 2011;
Du Pisani and Sandham, 2006) requiring an estimation of the nature and extent of the effects of the
proposed project on the biophysical, economic, social and cultural facets of the environment (DEA, 2010;
DEAT, 2008a,b; DEAT, 2006d), some researchers have focused on evaluating the quality of specific
dimensions of the environment as opposed to the entire EIR quality discussed above. Focusing on the
quality of social impact assessment (SIA) report as a component of EIA, Hildebrandt and Sandhamb (2014)
-5-
highlights weak SIA report quality particularly with regards to defining and identifying impacts, impact
significance prediction, project alternatives and mitigation measures. The research findings by Hildebrandt
and Sandhamb (2014) concur with the findings of Du Pisani and Sandham (2006) in South Africa and with
those of Fisher (2011), Burdge (2003) and Glasson and Heaney (1993) who also found SIA or the treatment
of socio-economic impact a poor component of EIA in the UK and the US.
Aucamp et al. (2011) and Du Pisani and Sandham (2006) highlighted that the EIA is strongly weighted to
the biophysical environment while Kruger and Chapman (2005) noted that many EIRs do not consider socioeconomic impacts of planned developments. Hoosen (2014) have noted that though the EIA regulations
require an estimation of the nature and extent of the negative and positive effects of the planned project
and identified alternatives on the various dimensions of the environment (DEA, 2010; DEAT, 2008a,b; DEAT,
2006d), it does not specify the criteria that needs to be used to estimate the effects.
Perdicoúlis and Glasson (2006) in a review of causal networks (i.e. the diagrammatic illustrations of
interactions between elements and the designation of causality to those relations) use in EIA, disclosed that
causal networks though they tie well with EIA as they are specially suited for making cause and effect
relations explicit (European Commission, 1999a), soliciting where and how impacts arise (Glasson, 2001)
and hence suited to fulfill specific principles of EIA conduct, for example transparency, integration and
being systematic, their use in modern EIA practice is minimal, simplistic and dwindling. In their random
sample of environmental impact statements they found zero count of causal networks use. Among the
causal networks discussed in this study are cause-and-effect diagrams, tree diagrams, digraphs, flow
diagrams and system dynamics (a discussion of these causal networks is provided in the following section).
Perdicoúlis and Glasson (2006) findings concur with those of Wood, Glasson and Beker (2006) who also
found the use of similar methods such as decision trees and flow charts in the region of 3% or lower in
England and Wales. The scarcity of causal networks in EIA practice might explain the limitation of
identifying, predicting, assessing and evaluating impacts in EIA highlighted in the above paragraphs.
On the other hand, Burdge (2003) notes that the economic evaluation of externalities do not feature in the
assessment process while Roth and Ambs (2004), Icyk (2006) and Australian Academy of Technological
Sciences and Engineering (ATSE), 2009) emphasize the importance of considering externality costs
alongside financial costs in decision-making. In addition, since EIA is project-oriented, its narrow focus
conceals the true life-cycle impacts associated with a proposed development. For example for a proposed
coal-fired power station, a separate independent application will be submitted and additional independent
ones will be submitted for other activities upstream or downstream of the power station (e.g. for the coal
-6-
mine(s) that will supply that power station, coal transportation and electricity transmission). Not
investigating the entire fuel cycle of energy development projects conceals the true impacts and costs
associated with energy technologies/sources and may lock a nation in a costly energy path, since power
generating projects from any fuel source (e.g. coal, oil, gas, solar and hydropower) are costly activities
involving multi-billion Rand investments and are associated with diverse and long-lasting environmental
and societal effects at some points in their fuel cycle (World Energy Council, 2004; Georgakellos, 2010). The
comprehensive assessment and full cost pricing of energy technologies supports the selection of best
source of power from a perspective that accounts for environmental preservation, human health and
economic feasibility (Roth & Ambs, 2004).
1.4
Possible suggestions to selected EIA issues
The scarcity of causal networks in EIA practice might explain the limitation of identifying, predicting,
assessing and evaluating impacts in EIA highlighted in the previous section. Causal networks are specially
good for making cause and effect relations explicit (European Commission, 1999a), soliciting how and from
where impacts emanate (Glasson, 2001) and when the causal relationships convey quantitative information
(equations) they become capable of numerical simulations, making possible forecasting and hence
enhancing decision-making (Perdicoúlis & Glasson, 2006).
In their review of the typology of causal networks use in environmental impact assessment, Perdicoúlis and
Glasson (2006) disclosed that causal networks though they are suited to fulfill specific principles of EIA
practice like transparency, integration and being systematic, their use in modern EIA practice is simplistic,
minimal and dwindling. Specifically the review consisted of: non-graphical expressions of causality in
environmental impact assessment namely, (i) text and (ii) matrices; graphical expressions of causality
(causal networks) in environmental impact assessment namely, (iii) digraphs/directed-graphs, (iv) causeand-effect diagrams, (v) flow diagrams and (vi) tree diagrams; and causal networks beyond EIA, that is
those in other professional/academic fields that can enhance EIA, for example, system dynamics. In the
following paragraphs a brief discussion of the non-graphical expressions of causality in environmental
impact assessment, causal networks in EIA and system dynamics is conducted to highlight the superior
attributes that system dynamics can offer to enhance EIA practice.
(i) Text - gives considerable liberty when relating project and environmental elements and their
interactions. However, text may result in misunderstanding/omissions owing to the complexity of EIA
systems (Perdicoúlis & Glasson, 2006). (ii) Matrices – express causality by relating the effects of individual
project actions in columns (e.g. construction) to individual environmental parameters in rows (e.g. to
-7-
whether or not construction causes noise, the duration of the interaction, probability of occurrence,
reversibility, etc.). The shortcomings of impact matrices is that they permit the illustration and study of
interactions between only two sets of data so consideration of third set of data (interactions between
effects/impacts or indirect impacts) cannot easily be represented in the same matrix (Perdicoúlis &
Glasson, 2006). This limitation can be overcome by representing the matrix in a (iii) digraph/directed-graph
causal network. Digraphs are conceivably the easiest causal networks with elements represented by nodes
and causality between elements represented by unidirectional arrows. The arrows may also incorporate
the polarity (+/-) between the elements (Canter, 1996; Perdicoúlis & Glasson, 2006).
On the other hand, though (iv) cause-and-effect diagrams are digraphs their elements are specified in text
form in different designs (commonly rectangles). Like in directed-graphs causality between elements is still
marked by unidirectional arrows but without demonstrating link polarity. In addition, causal relationships
generally convey no quantitative information so the diagrams are less rich in information. In EIA they are
used for identifying and predicting impacts (Perdicoúlis & Glasson, 2006; Glasson et al., 2005; Glasson,
2001; European Commission, 1999a). In contrast to cause-and-effect graphs which trace activities and their
effects, (v) flow diagrams trace movements of materials/energy. There exists various forms of flow
diagrams but some are not causal (Perdicoúlis & Glasson, 2006; Glasson et al., 2005). Last but not least, (vi)
tree diagrams resemble trees and may or may not be causal (Perdicoúlis & Glasson, 2006). There are
various types of tree diagrams, for instance event trees which are employed towards studying development
options concurrently (United Nations Environment Programme, 2002) and decision trees which are used for
outlining actions and their effects, impact significance and delineate corresponding decision options (e.g. to
draft mitigating measures or not) (Glasson et al., 2005).
System dynamics is a causal mathematical model that represents systems (for example the natural
environment, economy, social and energy) and analyses how they behave over time (Sterman, 2000;
Forrester, 1961). There are two fundamental types of diagrams in system dynamics namely, causal loop and
stock-and-flow diagrams. The former are special digraphs (Sterman, 2000; Ford, 1999) that capture the
structure of the system in a qualitative manner. The diagrams indicate cause and effect relations between
the system variables, link polarity, feedback loops and delays – all being fundamental attributes of dynamic
systems. Causal loop diagrams contain more information than typical digraphs, for example time delays and
feedback loops (Perdicoúlis & Glasson, 2006).
Stock and flow diagrams are flow diagrams and they too reflect cause and effect relations (Perdicoúlis &
Glasson, 2006) and unlike causal loop diagrams which illustrate the system structure qualitatively, they
-8-
capture the quantitative relationships between the variables of the system. The stocks/levels are denoted
by rectangles and they show accumulations while the flow variables (i.e. inflow and outflow rates) are
denoted by valves and they regulate changes in stocks (Jeong, Kim, Park, Lim, & Lee, 2008). There are two
styles of expressing the equations in stock and flow diagrams namely, mathematics and chemistry styles
(Perdicoúlis & Glasson, 2006):
(
)………….…………………………………………...............Mathematics style
…….….…………..….…………….………Chemistry style
The stock and flow diagrams therefore show in an explicit manner the relations between elements in the
system both textually and mathematically. The diagrams are for this reason richer in information than the
corresponding causal diagrams and are capable of numerical simulations (Jeong et al., 2008; Perdicoúlis &
Glasson, 2006). The stock and flow diagrams permit simulations based on specified scenarios, for example
scenarios characterized by project activities, system states and mitigation measures (Perdicoúlis & Glasson,
2006). System dynamics is therefore an experimental approach that can permit learning about
development projects through “what if” analysis (Wolstenholme, 2003). It is also a flexible tool (Anand et
al., 2005) that can work with numerous bottom-line facets (ecological, economic, social, energy, etc.)
through its capability to model a widespread assortment of processes and relationships (Auerhahn, 2008),
through decomposing the system into smaller, interacting sub-models that can be analyzed and integrated,
keeping the mutual interactions among them. For this reason there is no restriction on what a system
dynamics model can be designed to do. It has the capability to model complex problems in terms of flows,
stocks, time delays and feedback loops at any level of aggregation, be it at company, industry, country,
regional or global level and has capability to handle not only numerous variables but also innumerable units
of measure with ease. Lastly, it permits the modeler to control the complexity or boundary of the model
and hence the data needs. For instance, a simpler model can be built in the beginning and can be easily
extended to address further questions. System dynamics can therefore offer superior attributes to enhance
EIA practice.
While the employment of casual networks and specifically system dynamics in EIA practice may rectify the
limitation of identifying, predicting, assessing and evaluating impacts, as well as permit transparency,
integration and being systematic, the project-orientated manner of EIA, however, limit the scope of impact
assessment and hence does not permit a comprehensive assessment of the life-cycle impacts and costs, a
limitation that becomes more evident in the context of energy generation development projects due to the
importance of fuel-cycle impacts and costs towards informing energy technology selection. For this reason
-9-
one could argue that EIA is not broad enough to enable sound energy technology (or power generation
projects) assessment to inform policy-making. For this reason an exploration of technology assessment is
conducted in the following section since it is broader than EIA. Berg (1994) and Brooks (1994) classify
technology assessment into various types that illustrates its broader scope than EIA. The various forms of
technology assessment are discussed later in the following section.
1.5
Technology and technology assessment
Technology is the science that deals with the construction and usage of technical artifacts and their
interconnection with social, natural and economic environments (Grubler, 1998). Technologies are
developed and shaped by social actors, social information (i.e. human skill, reason and techniques) and the
economic system (Grubler, 1998). The production and use of technology in turn shapes the social, natural
and economic environment (Berkhout & Goudson, 2003; Grubler, 1998). Technology and the social,
economic and natural environments are therefore inseparable. The relation between technology and the
social, economic and natural environment is, however, a complex one (Grubler, Nakicenovic & Nordhaus
2002). While technology development has the capability of stimulating economic growth (Berkhout &
Goudson, 2003), providing societal benefits, improving efficiency of existing activities (i.e. affecting the
production function of companies) (Andries, Janssen & Ostrom, 2004; Berkhout & Goudson, 2003),
repairing/minimizing/reversing the negative environmental impacts of existing activities (Berkhout &
Goudson, 2003), technologies are dependent on the natural environment for raw-materials/resources,
their production, use and/or disposal impose negative effects on the natural environment and they depend
on the natural environment for waste assimilation (Smith & Stirling, 2008). Technology is therefore
fundamental to the well-functioning of economies and societies but needs to be managed to minimize
negative impacts, and energy technologies1 are no exception.
The energy sector has for a long time been driven by technological development (Sagar & Holdren, 2002).
Energy technologies and the resultant energy/electricity they generate are essential to meeting basic
human needs and for the advancement and development of economies (Ghosh, 2002; Ghader, Azadeh &
Zahed, 2006; Vardar & Yumurtaci, 2010; Alter & Syed, 2011). However, aside from the beneficial
consequences, energy technologies like all technologies generate undesirable effects in their production,
use and/or disposal and in addition, electricity production from any fuel source (e.g. coal, oil, gas, solar and
hydropower) also poses undesirable environmental and social effects at some points in its fuel cycle (World
1
Energy technologies are devices that produce or transmit or use energy, for instance power plants, boilers,
automobiles, etc., and are characterized by various attributes, namely efficiencies, costs, benefits, emissions, etc.
(Energy Technology Systems Analysis Program, 2007).
- 10 -
Energy Council, 2004; Georgakellos, 2010). Since energy technologies are not self-governing, their
management is essential to enable the realization of the socio-economic benefits with minimal effects on
the natural environment.
Technology assessment an imperative discipline in technology management is a strategic designing device
for policy making regarding technologies. The concept was developed in the late 1960’s at a time when the
extensive application of technology began to visibly affect United States inhabitants (Tran, 2007). It was
designed to support public policy decision-making by providing an understanding of the likely implications
of the extensive expansion of currently in operation technologies or the introduction of new ones
(Berloznik & Van Langenhove, 1998). It is therefore a policy study designed to offer decision makers with
information regarding the implications of technologies (Coates, 2001; CEFIC, 1997). Its aim is to produce
policy alternatives for answers to societal and organizational difficulties which at the practical level apply
new technologies or modifies/alters existing technology.
Technology assessment focuses on direct and indirect effects (Coates, 2001; CEFIC, 1997) plus benefits and
downsides (CEFIC, 1997). With awareness that technology schemes are rooted within the socio-economicecological system, technology assessment uses a theoretic structure that is determined by the three facets
of sustainability namely, social, economic and ecological facets (Assefa & Frostell, 2006). The concept of
technology assessment was as a result redefined as the evaluation with respect to sustainability of an
object fashioned by social actors towards the realization of a goal (Eriksson & Frostell, 2001). Technology
assessment enables therefore the assessment of a technology with respect to its supposed defined setting
of operation and enables its comprehensive evaluation with reverence to sustainability and in contrast with
other solutions yielding similar functions.
The concept is utilized in a number of organizational settings that vary widely in scope and depth (Assefa &
Frostell, 2006) including government, industry, academia, research laboratories, power executives
(Berloznik & Van Langenhove, 1998; Tran, 2007) and businesses (Berloznik & Van Langenhove, 1998; Tran &
Daim, 2008). Through offering information aiding decision making technology assessment can be
imperative in influencing improvement in existing technologies, adoption of new technologies,
manufacture and purchase decisions and research direction (De Piante, 1997).
Various researchers/institutions distinguish among various forms of technology assessment. For instance,
the Institute for Technology Assessment and Systems Analysis classified technology assessment into three
types namely problem-induced, project-induced and technology-induced technology assessment (Berg,
- 11 -
1994) while Brooks (1994) distinguishes among five types of technology assessment namely, project, policy,
generic, problem and global focused technology assessments. Generic oriented technology assessment
examines generic technologies with no orientation to a specific project/place whereas project oriented
technology assessment focuses on a concrete project. Problem oriented technology assessment studies an
extensive problem area and explores a broad spectrum of technologies and non-technical alternatives
towards managing the problem while policy assessment technology assessment is synonymous with
problem oriented technology assessment with a greater consideration of technological measures to
achieving social goals. Global oriented technology assessment focuses on a cluster of technical, economic,
social and political problems affecting the entire globe. Technology assessment is therefore broader than
EIA which is project- centered.
Armstrong and Harman (1980) categorized technology assessment into main functional elements namely:
technology description and alternative projections – which include a description of the technology,
establishing the boundary of the assessment and projection of technology alternatives; impact assessment
– which include establishing the impact selection criteria and impact prediction, assessment, comparison
and presentation; and policy analysis which involves the implementation of technology/alternatives and
relating the assessment of impacts to the address of societies concerns (Durbin & Rapp, 1983; Armstrong
and Harman, 1980).
1.6
Technology assessment shortcomings and solutions
Similar to environmental impact assessment which focuses on the impacts of planned development
projects, technology assessment centers primarily on the impacts/consequences of a technology before the
effects are with ease identifiable (Fleischer, Decker & Fiedeler, 2005). Likewise, policy-making requires an
understanding of the potential effects of the institution of technologies before they are extensively applied.
Proper assessment of the social, economic and ecological effects of planned developments that employ
new technologies is therefore a fundamental process of technology assessment.
The embrace of economic, social and environmental indicators (i.e. sustainability indicators) can therefore
be helpful in the evaluation of various developments or developments employing new technologies (Assefa
& Frostell, 2006). But frequently, as disclosed by the literature review conducted in this research, the
assessment of energy technologies lacks clear benchmarks of appropriate assessments, which has resulted
in difficulty to compare and to gauge the quality of various assessment practices. The assessments methods
and tools tend to be discipline specific with little to no integrations (Palm & Hansson, 2006). For example,
the technology assessment tools are often categorized into financial analysis tools, externalities/impact
- 12 -
analysis tools, systems analysis tools, risk assessment and technical performance assessment (Short, Packey
& Holt, 1995). As a result the literature is characterized by energy technology studies that exclusively assess
these groupings with little/no integration and with variations in scope and depth. A few selected groupings
of technology assessment tools plus examples of related studies are described briefly below:

Financial analysis tools - financial analysis is essential to corporate decision makers as it entails
comparing cash inflows and cash outflows of power generation developments and calculating the
corresponding financial return ratios. Financial analysis is therefore an essential constituent of
technology assessment but on its own it does not provide an all-inclusive assessment. Financial
feasibility may be assessed using different kinds of metrics such as life cycle cost analysis, levelised cost
of energy, cost effectiveness analysis, return on investment, net present value and breakeven point
analysis (Short et al., 1995). Examples of local studies focusing on the private costs of various energy
technologies include those by the Electric Power Research Institute (EPRI) (2010) and Mokheseng
(2010);

Externalities/impact analysis tools - externalities have been given many definitions and names in the
literature (Sundquivist, 2000), but the implications of externalities are somewhat the same (Baumol &
Oates, 1993). Generally, an externality occurs each time the production/consumption decisions of an
agent affects the utility of another in an unintentional manner and when no compensation is made to
the affected party by the producer of the undesirable effect. This definition follows the one of Baumol
and Oates (1988), Cornes and Sandler (1986), Mishan (1969) and Perman, Ma, McGilvray and Common
(1999). In the context of technologies, externalities are the unintended, non-compensated
accompanying effects of a technology that are borne by a third party (e.g. society or the environment).
Van Horen (1997), Spalding-Fecher and Matibe (2003) and Blignaut, Koch, Riekert, Inglesi-Lotz and
Nkambule (2011) offer examples of studies focusing on externalities but with the shortcomings of
emphasising the coal combustion phase and a subset of the coal-fuel cycle externalities mainly (e.g.
climate change and human health impacts);

Systems analysis tools - systems thinking analysis is an approach that looks at problems as parts of a
whole system. It is centered on the understanding that a system can best be grasped by examining the
linkages and interactions between its components and elements. In the viewpoint of technologies,
system analysis offers a systems view to technology assessment that enables the assessment of
technologies within their domain of operation (Crepea, 1995). There are a number of energy systems
analysis models and they may be categorized into bottom-up and top-down energy approaches. Top- 13 -
down energy models, also called macroeconomic models, address the energy-economy feedback. The
models describe the economic system in detail but they typically describe the energy system in an
aggregated manner and as a subdivision of the whole economy. The technical potential of various
energy technologies is thus not represented explicitly. Top-down modelers apply general equilibrium
models or models that are demand prompted (Hodge et al., 2008). In contrast, bottom-up energy
models study the energy system extensively but they do not consider the economic system in detail as
in top-down models (Berglund & Soderholm, 2006). As emphasized by Grubler et al. (2002), bottom-up
models normally aim at finding the minimum-cost mix of energy technologies serving a specified
energy demand. For this reason the models are optimization models that minimize total discounted
system cost (or maximize the income of energy systems) conditional on technological and
environmental constraints (Kiviluoma & Meibom, 2009). Bottom-up models include the PERSEU,
Balmorel, MARKAL and HOMER models.
The top-down and bottom-up energy systems models discussed above, offer piecemeal information that
limits deeper understanding of energy technologies and their consequent economic, environmental and
societal impacts. This is so, because the top-down models present the energy system as a black-box, by
paying no attention to the processes and activities because the matrices used can only analyse a sector as a
whole, and as a result differentiation between a range of products or production methods nor technologies
is not possible (Weisser, 2007). In addition, environmental focus is on GHGs and the links between plant
type/performance and environmental/societal burdens are hidden. The bottom-up models’ shortcomings
include that they are generally static models, with no feedback loops and time delays. In addition, they
optimize for least cost in private terms not in social terms, and environmental focus is on GHGs. Local
studies that employed top-down energy models include Pauw (2007) whereas bottom-up energy models
have been used by Haw & Hughes (2007) and Winkler, Hughes, Marquard, Haw and Merven (2011).
In the light of the shortcomings of energy technology assessment studies and tools, there have been
researchers who have advocated for the use of a holistic and integrated approach, specifically system
dynamics approach for the assessment of energy technologies partly due to its capability to permit
understanding of energy technologies in the domain of the multifaceted system wherein they are rooted,
for example, it permit the study of energy technologies in relation to economic, environmental and social
systems (Wolstenholme, 2003). As a special system analysis tool also recognized as a tool for energy
technology assessment (Tran & Daim, 2008) system dynamics is special in that it has the supplementary
ability to investigate dynamic cause and effect interactions and has the capability to model an extensive
diversity of processes and relations in a dynamic manner (Auerhahn, 2008). Among other advantages of
- 14 -
such an approach that makes it suitable for energy technology assessment as stated by Wolstenholme are its experimental nature that permits learning about the technology in question and its interaction with the
sphere of its application through “what if” analysis; its support for collaboration between various
stakeholders about the technology; and its support for the examination of the advantages and side effects
of an energy technology. System dynamics is, however, not limited to energy analysis owing to its diverse
application in various settings, for example in urban planning, economics, medicine, industrial engineering
and management (Damle, 2003).
There are a number of system dynamics energy models, for example the Feedback-Rich Energy Economy
model, which is a climate-economy model focusing solely on economy-climate interactions (Fiddaman,
1997), the Energy Transition model, which is a general disequilibrium model considering energy-economy
interactions (Sterman, 1981) and the IDEAS model, which is a dynamic energy supply and demand policy
simulation model of the United States (AES Corporation, 1993) which considers energy in isolation. Most
recent applications of coal-related system dynamics models have been developed for the study of various
issues, for example comparison of power generation cost (e.g. Jeong et al., 2008), forecasting coal demand,
supply and reserves (e.g. Hou et al., 2009), modelling energy supply and demand (e.g. AES Corporation,
1993; Musango, Brent, & Tshangela in press), assessment of coal production environmental pollution loads
(e.g. Hou et al., 2009; Yu & Wei, 2012), and GHG mitigation (e.g. Jeong et al., 2008; Saysel & Hekimoglu,
2010).
Often, however, the models provide a partial-view, for example the models were not tailored to specificcoal technologies, are not comprehensive in their assessment of fuel-cycle burdens to help inform energy
technology selection (i.e. focus is on a single phase (usually power generation or coal mining) and on a
subset of the coal-fuel cycles’ undesirable side-effects (usually GHGs)), tend to be still discipline-specific and
they tend not to address social cost. Roth and Ambs (2004) for instance, advocates the improvement of
assessment practices through the employment of a full cost approach that encompasses not only the
traditional costs incurred directly by power utilities (i.e. private costs) but costs incurred in the entire fuel
cycle including the conventionally neglected externality cost (social cost = private costs + externality costs)
(Roth & Ambs, 2004). According to Roth and Ambs the focus on private costs is unacceptable since it gives
limited attention to indirect factors (e.g. environmental and societal burdens and costs) which are taken as
exterior to the energy technology. Unless the externalities of electricity generation technologies are
identified, quantified and monetized they continue to be unknown, playing no role in the selection of
energy technologies (Australian Academy of Technological Sciences and Engineering (ATSE), 2009) and
posing hindrance to efficient and sustainable allocation of resources (Icyk, 2006).
- 15 -
In the light of the problems faced with evaluating energy technologies, the consequent need for formal
comprehensive assessment methods to evaluate energy technologies, and the suggestions for improving
the assessment of energy technologies as discussed above, this study focuses on improving the assessment
of energy technologies through the application of a system dynamics approach along a life-cycle viewpoint,
and specifically focuses on the Kusile coal-fired power station near eMalahleni in the Mpumalanga province
of South Africa as a case study.
1.7
Problem statement
South Africa has a number of planned development projects, including energy projects with coal-based
investments. Generally, the environmental and development planning process, in the form of an EIA have
been the main driver of project development in the country (Hoosen, 2010). The analysis of the quality of
EIRs, however, disclosed that amongst other issues, the more analytical components of the EIRs which form
the basis for decision making are performed poorly for instance with regards to the provision of
information pertaining to impact identification and assessment of key impacts (Sandham et al., 2008;
Sandham & Pretorius, 2008; Sandham et al. 2013). Concerning the assessment of impacts various
researchers have expressed inadequate use of assessment methodologies (Sandham et al., 2010; Sandham
& Pretorius, 2008), for instance, causal networks despite their suitability to fulfill specific principles of EIA
conduct like transparency, integration and being systematic (Perdicoúlis and Glasson, 2006; Wood et al.,
2006). Other concerns pertains to: overemphasis on biophysical environment (Aucamp et al.,2011; Du
Pisani & Sandham, 2006); limited consideration of socio-economic impacts of planned developments
(Kruger & Chapman, 2005); no consideration of the economic value of externalities (Burdge, 2003) despite
the importance of considering externality costs alongside financial costs in decision-making (ATSE, 2009;
Icyk, 2006; Roth & Ambs, 2004).
While the employment of causal networks and specifically system dynamics in EIA practice may rectify the
limitation of impact identification and the limited scope of impact assessment, as well as permit
transparency, integration and being systematic, the narrow project-orientation of EIA, however, limit the
scope of impact assessment and hence it hinders a comprehensive assessment of the life-cycle impacts and
social costs of developments, a limitation that becomes more evident in the context of energy generation
projects due to the importance of fuel-cycle impacts and social costs towards informing energy technology
selection. For this reason one could argue that EIA is not broad enough to enable sound energy technology
assessment to inform energy policy formulation and therefore an exploration of technology assessment
was conducted since it is broader than EIA (Berg, 1994; Brooks, 1994).
- 16 -
The energy technology assessment tools and studies, however, are also not without weaknesses for
instance they provide a partial view and partial analysis, respectively, to making informed decisions on the
selection of energy technologies. The reason for this being that the assessment tools and methods tend to
be discipline specific with little to no integrations, with tools often grouped into financial analysis tools,
impact analysis tools, technical performance assessment and so on (Palm & Hansson, 2006), which has
consequently resulted in energy technology studies that exclusively assess these groupings with little/no
integration and with variations in scope and depth. Other concerns pertain to the none consideration of the
economic evaluation of externalities and social costs (Roth & Ambs, 2004) as well as variations in scope and
depth in the assessment of externalities (i.e. limited scope of impact assessment) which make comparing
various energy development project involving (new) technologies difficult. For instance, the studies differ in
terms of the types of externalities they consider, the fuel-cycle stage(s) they investigate, and they do not
factor in the long-standing repercussions of the technologies on the environment and social systems.
These shortcomings highlight the lack of recognized technology assessment frameworks to support energy
policy formulation in the field of environmental and development planning processes (i.e. in both
technology assessment and as well as EIA) and therefore suggests the need for comprehensive assessment
to help inform decision-making on energy developments. Wolstenholme (2003) have supported improving
energy technology assessment through the use of a holistic and integrated approach, namely system
dynamics due to its superior attributes while Roth and Ambs (2004) advocates the improvement of
assessment practices through the measurement of not only the traditional costs incurred directly by power
utilities but costs incurred in the entire fuel cycle including the conventionally neglected externality costs.
This study therefore aspires to promote proper technology assessment at the extensive project level
through improving the environmental and development planning processes by means of employing a
systems approach, namely system dynamics due to its superior attributes and embedding it within the
processes to account for the lifecycle and long-term economic, social and environmental repercussions and
social costs of energy development projects. The current study specifically focuses on coal-based electricity
generation as a case study.
In the light of this problem, this study advances the environmental and development planning practices
(technology assessment/EIA) by contributing in terms of:
- 17 -

The consideration of the economic value of externalities (Burdge, 2003) and social costs (Roth &
Ambs, 2004) owing to the importance of considering externality costs alongside financial costs in
decision-making (ATSE, 2009; Icyk, 2006; Roth & Ambs, 2004);

The use of causal networks, specifically system dynamics to assess the environmental impacts of
planned developments (specifically in this study to model the life-cycle impacts and social costs of
energy generation projects/developments) due to its suitability to fulfill specific principles of EIA
and energy technology practices such as transparency, integration and being systematic
(Perdicoúlis and Glasson, 2006; Wood et al., 2006) as well as its ability to offer the numerous
positive attributes that are suited for energy technology assessment that were highlighted earlier in
this chapter and below, and later in chapter 3; and,

Permitting a comprehensive assessment of the life-cycle and long-term impacts and social costs of
energy generation projects/developments (Roth & Ambs, 2004), due to its importance towards
informing energy technology selection.
1.8
Rationale for system dynamics approach and a life-cycle viewpoint
The system dynamics approach was found conducive to the assessment of the environmental impacts and
social costs of power-generating technologies, because:

It permits operation with numerous bottom-line facets through its capability to model an extensive
diversity of processes and relations (Auerhahn, 2008), through decomposing the system into
smaller, interacting sub-models that can be analyzed and integrated, keeping the mutual
interactions among them). For this reason there is no restriction on what a system dynamics model
can be designed to do;

It permits the understanding of energy technologies in the domain of the multifaceted system in
which they are rooted (Wolstenholme, 2003);

It supports for the examination of the advantages and side effects of an energy technology
(Wolstenholme, 2003);

It has capability of modelling complex problems in terms of flows and stocks, feedback loops and
time delays (Perdicoúlis & Glasson, 2006) at any level of aggregation, be it at company, industry,
country, regional or global level;

It is a flexible tool (Anand et al., 2005) that can accommodate socio-economic-ecological indicators
that can assist decision-makers in the appraisal of energy technologies;

It is an experimental approach that permits learning about the technology in question and its
interaction with the sphere of its application through “what if” analysis (Wolstenholme, 2003), for
- 18 -
example it permit simulations based on specified scenarios, for example scenarios characterized by
project activities, system states and mitigation measures (Perdicoúlis & Glasson, 2006);

It has capability to handle not only numerous variables but also innumerable units of measure with
ease (i.e. the measurement metric is not fixed like in general equilibrium models);

It can be used in data poor problems, for instance, one can conceptualize and formulate a system
dynamics model for an anticipated future planned system and, unlike in statistical models, one
does not need time series data to drive the model. This is important in this study because the coalfired power station that is being studied is currently under construction. It is, however, worth
mentioning that system dynamics models can be used in conjunction with statistical models or
other energy models; and,

The tool permits the modeler to control the complexity or boundary of the model and hence the
data needs. For instance, a simpler model can be built in the beginning and can be easily extended
to address further questions.
Other advantages of system dynamics that favoured its use in this study include its offering of quick model
design and simulation, in-built error checking capacities, various model outputs comprising diagrams, tables
and graphs, extensive sensitivity analysis capabilities, conceptualization of a range of scenarios and
effortless experimentation with model structure, parameters, data and outcomes.
On the other hand, a life-cycle viewpoint was deemed important owing to the environmental and social
repercussions originating from the various stages of an energy technology’s cycle. The approach is
therefore opted for because it systematically assesses over the life cycle, all flows (e.g. materials, energy
and environmental flows) that go into the investigated system from nature and those that flows out from
the system to nature (Ampofo-Anti, 2008; Varun & Ravi, 2009). A life-cycle viewpoint can therefore limit
the exclusion of important externalities. Since the approach delineates between the various phases of an
energy technology, it is better suited to detecting the transference of impacts between life-cycle stages or
between environmental media. It can also serve as a tool for identifying potential socio-economicecological improvements (Sherwani, Usmani & Varun, 2010), by this means yielding vital trade-offs
information that can be beneficial to decision makers and managers.
1.9
Research objectives
The objectives of the present study are as follows:

To understand the resource inputs, material requirements and private costs of building, operating
and maintaining a coal-fired power station.
- 19 -

To understand the coal-fuel cycle environmental and societal burdens and costs (coal fuel cycle
phases considered: coal mining, coal transportation, plant construction, plant operation, and waste
disposal).

To develop and validate a system dynamics model for understanding coal-based power generation
and its interactions with resource inputs, private costs, externalities, externality costs and hence its
consequent economic, social and environmental impacts over its lifetime and fuel cycle.
1.10 Organisation of thesis
The thesis is organized in eight chapters. This first chapter presents the research, research problem,
research objectives and organization of the research. Chapter 2 presents the concept of externalities, coalfuel cycle externalities and a review of the South African power and coal industry. Chapter 3 grounds the
research conducted in this study within the economic discipline of study within which it falls and motivates
the use of a systems approach by studying the links between system dynamics and the schools of economic
thought that underpin this study. Chapter 4 presents a comprehensive survey of the different tools used by
various researchers to evaluate power generation technologies and afterwards an analysis of the
application of the tools in the power sector with a special focus on coal-based power generation
applications. Chapter 5 discusses the strategy of inquiry and the methodological approach that was
employed to achieve the study’s objectives. Chapter 6 discusses and presents the COAL-based Power and
Social Cost Assessment (COALPSCA) Model developed for understanding coal-based power generation and
its interactions with resource inputs, material inputs, private costs, externalities, externality costs and
hence its consequent economic, social and environmental impacts. Chapter 7 presents the COALPSCA
Model outcomes, discusses the validation of the model and evaluates the model outcomes under various
policy scenarios. Chapter 8 summarises the findings of this study, highlights its limitations and makes
recommendations for future research.
- 20 -
CHAPTER 2:
EXTERNALITIES AND SOUTH AFRICA’S POWER AND
COAL INDUSTRIES
2.1
Introduction
Before proceeding with the presentation of the thesis, it is imperative to provide background information
on externalities and social cost, coal-fuel cycle externalities and the South African power and coal
industries. This chapter begins by defining the concept of externalities which is followed by a discussion of
the environmental and societal impacts linked with the coal-fuel cycle. The South African power industry is
discussed in the third section. Special focus is given to Eskom’s power stations, electricity sales, coal quality,
emissions profile, coal supply and supply contracts. A discussion of the South African coal industry is
provided in the fourth section. Focus is on the trends of coal production, consumption and prices. The
country’s main coal producers and consumers and as well as export coal consumers are also discussed. The
fifth section summarises this chapter.
2.1
Externalities defined
Marshall (1890) was the first economist who dwelled on the concept of externalities, followed by his
student Pigou (1920). Ever since these early days, economists have paid a great deal of attention to the
concept of externalities. However, in the literature there are many definitions of externalities and in
addition, externalities have been given many names, including external diseconomies/economies, external
effects, adders, third party effects, and neighbourhood effects (Sundquivist, 2000), but nevertheless, the
implications of externalities are somewhat the same though a definition that captures all the concept
ramifications is regarded by some economists as difficult to provide (Baumol & Oates, 1993).
Generally, an externality occurs each time the production/consumption decisions of an agent affects the
utility of another in an unintentional manner and when no compensation is made to the affected party by
the producer of the undesirable effect. This definition follows the one of Baumol and Oates (1988), Cornes
and Sandler (1986), Mishan (1969) and Perman, Ma, McGilvray and Common (1999). The definition thus
states that an externality can occur in the production of a good or consumption of a good and the agents
that receive the effect can be a producer, consumer, an individual or society at large, secondly, an
externality can be a cost or benefit (Baumol & Oates, 1988), thirdly, the effects of an externality falls on a
third party (Cornes & Sandler, 1986), fourthly, it is an unintentional action, so it does not include intentional
actions by an agent or else it would be handled within the existing justice system (Mishan, 1969) and lastly,
- 21 -
it involves no compensation, hence it causes inefficiencies and misallocations of resources (Baumol &
Oates, 1988).
In order to understand the last point (i.e. that externalities cause inefficiencies and misallocation of
resources) it is necessary to discuss externalities in the context of the framework of welfare economic
theory. Welfare economics focus on the study of resource allocation and income distribution in an
economy in such manner that an efficient state is achieved, a state whereby no individual can be bettered
without making others worse-off (Pareto efficiency) (Mishan, 1960; Arrow & Scitovsky, 1969). Welfare
economics is therefore concerned with testing the efficiency of economic activities in utilizing society’s
productive assets. It aims at attaining maximum social welfare.
There are a number of conditions under which social welfare maximisation becomes achievable, for
example, the existence of perfect competition, free trade, etc. In the midst of these conditions, the
necessary and sufficient condition is the presence of perfect competition. Economic efficiency in allocating
resources is attained on the equality of marginal costs and prices. This marginal argument is broadened to
incorporate the proposal that marginal social benefit must equate marginal social cost for social welfare
maximisation (Ferguson, 1972).
Although the marginal conditions pictured in the establishment of social welfare maximisation are essential
to society’s welfare improvement, in the real world the conditions are hardly ever met, causing a deviation
among social and private benefits and social and private costs. Various causal factors can contribute to the
deviation of social and private costs, among which is the existence of externalities (others are, barriers to
market entry and trade, poorly defined property rights, etc.). Because of the presence of externalities
markets fail to achieve Pareto efficiency (i.e. causing an incident whereby the First Theorem of Welfare
Economics fails to apply), hence a divergence between private and social costs.
For illustration purposes consider Firm A that operates in a competitive market. Assume that Firm A
produces q quantities of output, at a cost of c(q) (private cost) and sells its product at a market price of p .
Firm A' s profit maximization problem then becomes:  A  max pq  c(q) .  A is the profit for Firm A . Firm
q
A' s equilibrium amount of output (q  ) is yielded by the first order conditions p  c ' (q* ) showing that Firm
A should produce up to the point where price  p  equals marginal private cost (MPC ) (i.e. the point
denoted as X in Figure 4.1).
- 22 -
Now assume that the production of q units by Firm A also leads to the production of q units of pollution,
i.e. an externality cost of pollution equal to e(q) . Hence from society’s point of view, the q  units produced
by Firm A are too large, because Firm A only considered private costs in its optimization. Thereby, not
taking into account the externality cost of pollution, it imposes on society. The efficient level of production
for Firm A in society’s view is that which internalises the externality cost. Hence, in order to internalise the
externality cost of pollution, Firm A needs to incorporate the externality cost into its profit maximization
problem, which then becomes:  A  max pq  c(q)  eq  .
q
The efficient level of output for Firm A in the presence of externalities is yielded by the first order
conditions p  c ' (q e )  e ' (q e ) showing that Firm A should produce at the point (point Z ) where price equals
the sum of marginal private cost and marginal externality cost, i.e. marginal social cost (MSC ) . q e is the
Pareto efficient output (see Figure 4.1). The summation of private and externality costs makes up the social
cost of production (social cost = private costs + externality costs) (Pearce & Turner, 1990), which is also
called the total economic cost (Kim, 2007).
Marginal social cost (MSC)
Z
Marginal private cost (MPC)
X
Demand
Figure 2.1: Externality costs of production
Source: Own construction
Thus, without internalising the externality cost of pollution produced by Firm A , Firm A only faces MPC
and produces an output that is higher in the viewpoint of society (q   q e ) . Also the price faced when Firm
A does not incorporate the externality is lower than when externality costs are incorporated ( p  p e ) .
- 23 -
This therefore means that using a good without taking into account its social cost, results in resource
misallocations because of producers’ choice to produce a higher output level than is economically ideal.
This higher rate of production translates into more rapid consumption of resources and even more
pollution because as long as externality costs are not internalised, the incentive to produce less pollution
does not exist, hence more and more pollution is produced since releasing it into the environment is cheap
(Tietenberg, 1992).
Though highlighted that externalities need to be internalised, unregulated markets will not internalise
externalities themselves, some kind of government intervention is needed. One way, according to
Pigouvian teachings, is through taxing the producers of the externality an amount equivalent to the
damages caused. Some economists recommend that focus should be on clarifying property rights instead of
taxing and yet some favour other methods, such as user fees and tradable emission permit scheme (Energy
Information Administration, 1995). To conclude, externalities cause market failure, which in turn leads to
non-optimal resource allocation in society’s view. So using a good without taking into account its social cost
causes misallocation of resources. The externalities linked with the coal-fuel cycle are discussed next.
2.2
The environmental and societal impacts linked with the coal-fuel cycle
Table 2.1 below presents some of the environmental and societal impacts linked with the coal-fuel cycle.
The environmental and societal impacts in Table 2.1 are externalities on condition they are negative
unintentional consequences of an economic activity that are borne by a third party without or without-full
compensation. In the table, the impacts have been broadly categorised into three main classes, namely coal
mining and transportation impacts, plant construction impacts and plant operation impacts. Where
necessary during the discussion, special reference is made to Kusile’s impacts as discussed in the
environmental impact assessment report (i.e. NINHAM SHAND, 2007).
- 24 -
Table 2.1: Coal-fuel cycle environmental and societal impacts
Activity
Biodiversity
Air
pollution
GHG
Damage
to roads
Accidents
Noise
Water
quality
Coal mining & transportation impacts
Coal mining
Beneficiation
Coal transportation
Plant construction impacts
Site preparation
Materials production
Materials transportation
Construction
Plant operation impacts
Material inputs production
Material inputs transportation
Raw material storage: coal, fuels, etc.
Coal combustion
Flue-gas clean-up: FGD
Ash & FGD waste disposal
Source: Own construction
2.2.1.1
Coal mining and transportation impacts
Coal mining is linked with various societal and environmental hazards. It normally impacts the environment
in the course of extraction, beneficiation and during coal transportation to a power plant (Mishra, 2009).
The main impacts associated with coal mining and transportation include air pollution human health
burdens, climate change impacts owing to GHG emissions, injuries and fatalities, water pollution and landuse linked impacts. Air pollution in coal mines stems mostly from emissions of particulate matter,
underground fires, coal dust, burning discard dumps (Goldblatt et al., 2002) and methane (CH4) emissions –
a greenhouse gas emitted in the course of coal extraction at a time when coal seams are cut (National
Research Council, 2009; Singh, 2008). The main operations producing gases and dust in mines are drilling,
blasting, haulage, crushing and transportation. Opencast mines are associated with more air pollution than
underground mines, due to that opencast mines create pollution within mining premises and beyond
(Singh, 2008).
High incidences of deaths and accidents are linked with coal mining due to falling rocks, CH4 explosions,
material handling and as well as due to coal transporting accidents. Noise pollution is another problem in
coal mines which causes problems such as hearing loss and pneumoconiosis (Goldblatt et al., 2002). The
quality of water can also be affected by opencast mines by means of leachate from discard dumps, dirty
mine water releases or acid mine drainage. Extensive land surfaces may also be disrupted by surface mines.
Such mines may also displace people, erode the soil, and impact on local biodiversity through vegetation
- 25 -
cover removal which has the likelihood to negatively impact endangered plant species with subsequent
effects on the faunas that use that habitat (NINHAM SHAND, 2007; Singh, 2008). Conversely underground
mining may result in surface subsidence (Singh, 2008).
The cleaning of coal using wet cleaning methods may decrease the content of sulphur in coal but it also
produces coal slurry which is discarded in slurry dams (Wassung, 2010). The dams are major water
contaminant due to their susceptibility to collapse during heavy rains. A number of the coal processing
chemicals are also acknowledged to be carcinogenic and to cause lung and heart damages (Epstein et al.,
2011).
Coal transportation is also associated with various negative externalities, mainly in the forms of accidents,
damage to roadways, air pollution, noise, global warming and congestion (Jorgensen, 2010). Among the
classic air pollutants released during the transportation of coal are carbon monoxide (CO), non-methane
volatile organic compounds (NMVOC), nitrogen oxides (NOx), particulate matter (PM2.5), hydrocarbons
(HC), sulphur dioxide (SO2) and lead (Pb). The ailments linked with these air pollutants comprise of lung
cancer, bronchitis, lower respiratory illnesses, chronic respiratory disease and eye irritation. The GHGs
linked with coal transportation comprise of carbon dioxide (CO2), nitrous oxide (N2O) and CH4. Noise is also
linked with coal transportation due to car alarms, road contact and engine noise. Related to accidents are
occupational and non-occupational injuries, deaths, material damage and lost productivity (Gaffen et al.,
2000). Coal transportation may aslo lead to roadways damage plus congestion (Jorgensen, 2010) while the
construction of new roadways may impact on local biodiversity (NINHAM SHAND, 2007).
2.2.1.2
Plant construction impacts
Plant construction can be a very destructive operation resulting in a large number of negative externalities.
Among the main impacts are biodiversity impacts and increased sediment loads on rivers and streams from
the removal of vegetation cover to construct the plant and its ancillary infrastructure (NINHAM SHAND,
2007; US. Department of Energy, 2009). In the case of biodiversity, the removal of vegetation cover can
alter the diversity of plants and animals in the study site and/or even impact endangered plant species,
with negative impacts on the faunas that use that habitat. In the site in which Kusile is being constructed,
there are a range of protected species amongst which are six endangered plant species plus a red data bird
species. The overall significance of the impact of the construction phase on terrestrial and aquatic flora and
fauna without and with mitigation measures in place has been deemed to be medium (negative) and low
(negative), respectively due to the presence of protected species and little natural vegetation cover, owing
to extensive agricultural activities in the site (NINHAM SHAND, 2007).
- 26 -
Specifically, though the proposed position of Kusile interconnects with a seasonal watercourse that runs
into a river coupled with the occurrence of protected species in wetland communities, the effect of
establishing the plant and its related infrastructure on terrestrial flora and fauna was deemed to be low
(negative) without mitigation and very low (negative) with mitigation measures implemented (the
proposed conveyor belt, pipeline and road placements crossed land largely dominated by agricultural
undertakings so the impact was also deemed minimal). Concerning the effects of the power plant and its
linked infrastructure on aquatic flora and fauna, the impact assessment results disclosed that the plants
infrastructure (consisting of coal stockyard, dams, water and wastewater treatment structures) would
directly impact the aquatic environment owing to being right on segments of the wetland and indirectly
through loss of wetland services. Though the coal stockyard proposed location could not directly impact
any wetlands the seepage from it could impact surrounding aquatic flora and fauna. The proposed layout of
the surface infrastructure was deemed to reflect a low (negative) significance impact (NINHAM SHAND,
2007).
On the other hand, the establishment of roads, conveyors, railway line and pipelines were found to affect
various wetlands systems with the overall significance of the impact deemed high (negative) owing to its
high magnitude, long-term duration and local extent. Due to the proposed location of the above ash dump
being in the middle of a high integrity wetland, it would have a direct effect on the aquatic environment
coupled with an indirect effect through increased sediment levels owing to dust blown away from it. The
ash dump is anticipated to have high (negative) significance impact without mitigation due to its long-term
duration and high magnitude, and a very low (negative) significance impact with mitigation measures
implemented. Overall the plants and its associated infrastructure layout could effect a low (negative) to
very low (negative) impact on the aquatic environment without and with mitigation measures, respectively
owing to that the site is generally characterized by low biodiversity, poor and degraded biotic integrity and
no endangered aquatic species (NINHAM SHAND, 2007).
Air pollution can also become an issue during the construction phase due to fuel use in heavy machinery,
which can, depending on the locality, have a negative effect on the water quality of open water bodies
through the deposition of particulates and chemicals in the water (U.S. Department of Energy, 2009).
Construction noise can become an issue contingent on the position of the plant, the operations being
performed, the construction period and the size of the site (Bohlweki Environmental, 2006; NINHAM
SHAND, 2007; Tech Environmental, 2009).
- 27 -
Other negative impacts include: visual effects from construction activities resulting from dust generation,
construction equipment, vehicles and presence of workers (Bohlweki Environmental, 2006); negative
impacts of construction activities on soils such as soil compaction which decrease aeration, porousness and
water holding ability of soils which can increase surface overflow and possibly soil erosion (NINHAM
SHAND, 2007); negative impacts linked with the transportation of material inputs necessary for the
establishment of the power station - i.e. an activity that requires a number of trips per day leading to an
increase in daily traffic volumes on road networks thereby resulting in negative externalities mainly in the
forms of air pollution, global climate change, congestion, accidents, damage to roadways and noise
(Jorgensen, 2010); and, indirect negative impacts linked with the production of material inputs (such as
iron, steel, aluminium, cement, concrete and glass) and manufactured products such as boilers and
turbines, which generates a range of negative impacts such as their contribution to global climate change,
transportation related impacts and biodiversity impacts (Russell, 2008; InEnergy, 2010).
2.2.1.3
Plant operation impacts
The operation of a coal-based plant can also impact the biophysical and social environments in numerous
ways (NINHAM SHAND, 2007). The impacts include the emissions of air pollutants such as particulates, SO2,
NOX, CO2, N2O (nitrous oxide) and various trace metals from flue stacks. SO2 and NOX contribute to acid
deposition, eventually leading to a wide range of environmental impacts, including damage to vegetation,
soils, human health, animals and materials (Ma & Jin, 2010). Particulate matter coupled with SO 2, NOx and
heavy metals are associated with harmful effect on the health of communities in vicinity to the power
station. These air pollutants can also impacts negatively on the quality of water of open water bodies and
subsequently aquatic flora and fauna through being deposited in water (U.S. Department of Energy, 2009).
CO2 and N2O are GHGs released from coal-fired power stations and they contribute to the greenhouse
effect as they trap long-wave radiation exiting the surface of the earth, leading to heating-up of the lower
atmosphere of the earth, with variations in global/regional climates (Georgakellos, 2010), extended
desertification and rising sea levels.
Also associated with the plant operation phase are occupational and non-occupational injuries and
fatalities (Department of Minerals and Energy (DME), 2010; Eskom, 2011). Visibility can also be reduced
due to particulates/dust generation from materials handling facility, ash-disposal facility and from flue
stacks (Ma & Jin, 2010). Visual impairment can also emanate from power station infrastructure such as coal
stockyard, ash dump and the power plant. Impact on ambient noise quality can become an issue contingent
on the position of the plant, the operations being performed, and the size of the site (Bohlweki
Environmental, 2006; NINHAM SHAND, 2007; Tech Environmental, 2009). Above ground ash dumps can
- 28 -
directly and indirectly impact aquatic environment depending on the location of the ash dump. For
example, directly through being sited in a middle of a high integrity wetland or near a river and indirectly
through dust carried-away from the dump leading to increased sediment loads in aquatic systems, causing
loss of habitat and decreased photosynthesis and physiological stress on organisms (NINHAM SHAND,
2007).
Groundwater can be contaminated through the use of process chemicals, acidic leachate emanating from
the coal stockyard, run-off from the coal stockyard, permeation and run-off from dirty water dams, leakage
and infiltration of liquid fuel and through runoff and seepage from the ash dumps. At the same time,
groundwater levels can also increase owing to artificial renewal from dirty and clean water dams and
through runoff from ash dumps and coal stockpiles (NINHAM SHAND, 2007). A range of impacts are aslo
associated with electricity usage in the operation phase, for example electricity consumed by the plant
itself and electricity used in conveyor belt to transport coal to the plant from the stockpile.
Link with power generation are also a range of indirect impacts connected with the production and
transportation of material inputs that are necessary for the operation of the plant, such as limestone for
SO2 abatement in the FGD system which generates a range of upstream impacts such as emissions of GHGs,
air pollution, biodiversity impacts, water quality impacts and transport-related impacts such accidents and
damage to roadways (Singleton, 2010). On plants fitted with FGD devices, the plant operation phase will
also generate a range of impacts linked with the operation of the FGD system. Generally, the impacts are as
a result of increased effluent discharge, increased solid waste, increased water use, traffic and transport
impacts, visual impacts, increased land-use and increased GHGs especially the principal gas CO2 owing to a
decrease in plant efficiency and also due to FGD chemistry (Singleton, 2010). The wet slurry from the FGD
could have negative impacts on the aquatic environment should there be spillage, however, due to waste
disposal minimum requirements in the case of Kusile, the significance of the effect is expected to be very
low (negative) (NINHAM SHAND, 2007). The South African power industry is discussed next.
2.3
South Africa’s power industry
The power industry in South Africa is dominated by Eskom, a utility owned by the state. The utility produces
over 90% of the electric power needs of the country (Department of Energy, 2010). It was established in
1923 as the Electricity Supply Commission and by 2002 it was entirely owned by government. The National
Energy Regulator of South Africa controls the utility’s power prices. In 2011 Eskom burnt 124.7 million tons
of coal. In the same period, 237 430 net Gigawatt hours (GWh) of electricity was produced by Eskom of
which about 93% was produced from coal-fired power plants (220 219 net GWh) (Eskom, 2011). Total
- 29 -
Eskom electricity sales in 2011 amounted to 224 446 GWh, earning the utility a revenue of R90 375 million.
Approximately 6% of the electricity produced was exported to neighbouring countries (Eskom, 2012a).
2.3.1
Eskom’s power stations
The existing and future Eskom power stations are presented in Table 2.2. The utility runs 10 base load
power stations which are mainly found in the Mpumalanga province. In addition, three power stations have
been or are being returned to service in Mpumalanga. All 13 power stations use conventional pulverisedcoal technology and are fitted with electrostatic precipitators in order to reduce particulate emissions. On
average, the utility's power stations have a generation capacity of 3 400 Megawatt (MW) with a wet recirculating cooling process and are fitted with precipitators to control dust (Wassung, 2010). Two new
power stations are currently under construction, namely Kusile and Medupi power stations in the Witbank
and Waterberg coalfields, respectively.
Converse to the conventional pulverised-coal technology used in Eskom’s coal plants, the new plants will
use supercritical technology (Eberhard, 2011). In a pulverised-coal power plant the coal is first crushed into
a smooth-textured powder and then fed into a boiler where the coal is burned to create heat. The heat
produces steam which is used to spin turbine(s) to generate electricity. Supercritical plants on the other
hand, form part of the pulverised-coal system but use higher pressure and temperatures to boost the
efficiency of the plant to about 40% or more (Bohlweki Environmental, 2006). The average thermal
efficiencies of Eskom’s pulverised-coal power plants were 32.6% in 2011 (Eskom, 2011) and 33.1% in 2010
(Eskom, 2010a). In addition, the new plants will use dry-cooling systems and have capacities above 4
700MW. The Kusile power station will be fitted upfront with a Flue Gas Desulphurisation (FGD) system to
remove SO2 from flue gas while the Medupi power station will initially not be fitted with an FGD system
(Njobeni, 2010).
- 30 -
Table 2.2: Eskom’s existing and future coal-fired power stations
New Return
to
Build
service
Base load
Plant name
Arnot
Duvha
Hendrina
Kendal
Kriel
Lethabo
Majuba
Matimba
Matla
Tutuka
Camden
Grootvlei
Komati
Medupi
Kusile
Province
Mpumalanga
Mpumalanga
Mpumalanga
Mpumalanga
Mpumalanga
Free State
Mpumalanga
Limpopo
Mpumalanga
Mpumalanga
Mpumalanga
Mpumalanga
Mpumalanga
Limpopo
Mpumalanga
Capacity
2100 MW
3600 MW
2000 MW
4116 MW
3000MW
3708 MW
4110 MW
3990 MW
3600 MW
3654 MW
1600 MW
1200 MW
1000 MW
4788 MW
4800 MW
Cooling system
Wet re-circulating
Wet re-circulating
Wet re-circulating
Indirect dry
Wet re-circulating
Wet re-circulating
Wet re-circulating & dry
Direct dry
Wet re-circulating
Wet re-circulating
Wet re-circulating
Wet re-circulating & dry
Wet re-circulating
Direct dry
Direct dry
Pollution
control
technology
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP
ESP, FGD
Year
1975
1984
1976
1993
1979
1990
2001
1994
1983
1990
1973
1976
1968
2015
2017
ESP = ElectroStatic Precipitator for controlling dust; FGD = Flue Gas Desulphurisation for controlling sulphur dioxide
Source: Adapted from Wassung (2010); Eskom (2013a)
2.3.2
Eskom’s electricity sales
The state-owned utility distributes electricity to customers in the commercial, industrial, agricultural,
mining and residential sectors, and to redistributors (municipalities). Direct electricity sales by Eskom
during the 2010/2011 period are presented in Figure 2.2. Most of Eskom’s electricity sales are to
municipalities (about 41%), industry (about 27%) and the mining sector (about 14%). Electricity sales to
commercial and agricultural, foreign, residential and rail sectors are low at 6%, 6%, 5% and 1%,
respectively. About 84 000 agricultural customers, 4 million residential customers, 3 000 industrial
customers, 49 000 commercial customers and about 1 000 mining customers are served directly by Eskom
(Eskom, 2011).
- 31 -
Eskom's electricity sales by sector
5.9%
4.7%
1.3%
Municipality
6.2%
40.8%
Industry
Mining
14.5%
Commercial and agriculture
Foreign
Residential
Rail
26.6%
Figure 2.2: Eskom’s electricity sales by sector 2010/11
Source: Adapted from Eskom (2011)
2.3.3
Coal quality and emissions profile of Eskom’s coal-based plants
The thermal coals used by Eskom in the generation of domestic electrical energy are generally poor coals
with a lower calorific value and high ash content. Table 2.3 below shows the quality of coal used and
pollutants emitted by Eskom’s power plants in the financial years 2006/07 to 2010/11. The table discloses
that, in general, Eskom burns coals with an average calorific value of about 19 Megajoules/kilogram
(MJ/kg), ash content of about 29% and sulphur content of about 0.80%.
In 2011, coal with an average ash content of 29.03% and an average calorific value of 19.45MJ/kg was
burned in Eskom’s plants while CO2, SO2 and N2O emissions stood at 230.3 million tons, 1 810 kilotons and
2 906 tons, respectively. Particulates emissions were about 75.8 kilotons (Eskom, 2011). The coal supplied
for power generation is mostly from screened run-of-mine (ROM) production while a third is sourced from
coal middling from coal washing (run-of-the-mill). The coal quality received by the utility has, however,
been worsening in recent years as higher grades are reserved for export market (Eberhard, 2011).
- 32 -
Table 2.3: Coal quality and emissions profile of Eskom’s coal-fired power plants
Variable
Units
2006/2007
2007/08
2008/09
2009/2010
2010/11
Average calorific value
Average ash content
Average sulphur content
Average thermal efficiency
Particulates
SO2
CO2
N2O
NOx as NO2
MJ/kg
%
%
%
kt
kt
Mt
t
kt
19.06
29.70
0.86
33.9
46.08
1 876
208.9
2 730
930
18.51
29.09
0.87
33.4
50.84
1 950
223.6
2 872
984
19.10
29.70
0.83
33.4
55.64
1 874
221.7
2 801
957
19.22
29.56
0.81
33.1
88.27
1 856
224.7
2 825
959
19.45
29.03
0.78
32.6
75.84
1810
230.3
2 906
977
Source: Adapted from Eskom (2011)
2.3.4
Eskom’s coal supply methods and coal supply contracts
All of Eskom’s mining is undertaken by private companies and the plants are mainly mine-mouth fed
through conveyor belts, though road transportation is also becoming common. For instance, of the 126.23
million tons of coal purchased in 2011, 30.5 million tons were transported by road (Eskom, 2011), signifying
that about a quarter of Eskom’s coal is supplied through road transportation. The utility is thus exposed to
short-term contracts which, in part, are spurred by the fact that the plants run at higher capacities than
originally planned and also owing to some collieries failing to meet production expectations for power
stations such as Tutuka and Majuba. The increasing reliance of Eskom on short-term contracts has upsurged the utility’s average coal price but it is still much lower than international prices (Eberhard, 2011).
Coal road supply has, however, impacted road infrastructure to such an extent that Eskom began financing
the upkeep of roads. In addition, the increase in the number of road accidents and fatalities has prompted
Eskom to invest in public safety awareness initiatives (Eskom, 2011).
Most of the collieries supply coal to Eskom plants through long-term coal contracts. Nine of the power
plants are served through long-standing coal contracts. Three of these are fixed-price contracts while six
are cost-plus contracts. Coal supply for the fixed-price contract is at a base price which is augmented using
an assented formula. For the cost-plus contracts, the power utility and the coal contractor share the capital
cost of establishing the mine, with Eskom additionally paying for the operation cost. The coal contractor
then receives a net income from Eskom, based on the return on money invested by it. The return on capital
invested consists of two components, namely a fixed component that is not based on coal production and a
variable component that is based on coal supplied to Eskom. Future coal supply contracts, however, are
envisaged to be fixed-price (indexed-priced) contracts since cost-plus arrangements discourage cost
minimisation (Eberhard, 2011).
- 33 -
2.4
2.4.1
South Africa’s coal industry
Coal production and consumption
Coal has been the backbone of the development of South Africa since 1870, when coal was first used in a
Kimberley diamond mine. Since then, coal production rose to approximately 30 million tons in the 1950s
and 115 million tons in 1980, following the oil crisis, rising international coal demand and soaring domestic
electricity demand (Department of Energy, 2010; Statistics South Africa, 2010). It is estimated that coal
production in 2011 stood at about 253 million tons, rendering the country the seventh major producer of
coal globally. In the same year the country produced 3.3% of total global coal production (Beyond
Petroleum (BP), 2012). If ROM production is considered, instead of saleable coal production, the country
produced a total ROM of about 316.2 million tons in 2011, which is a decrease of 0.43% from the tonnes
produced in 2010 (Department of Mineral Resources, 2012).
South Africa’s saleable coal production and consumption between 2000 and 2011 is presented in Table 2.4
and Figure 2.3. Between 2000 and 2011, South Africa produced a total of about 2 901 million tons of coal.
The country’s coal production for 2011 (252.8 million tons) decreased by 1.7% from 257.8 million tons in
2010. Compared to approximately 115 million tons of coal produced in 1980, the country’s 2011 coal
production is roughly 120% higher. Figure 2.3 shows the general steady increase of the country’s coal
production since 2000.
A closer look at Table 2.4, and more explicitly Figure 2.3, discloses that approximately two thirds of the coal
produced in the country is consumed locally while about one third is exported. In 2011, approximately
178 million tons of coal was consumed locally while about 69 million tons were exported. Local coal
consumption tonnage in 2011 decreased by 4.7% from 2010, while coal exports in 2011 increased by about
3% from 2010. The domestic consumption of about 178 million tons makes South Africa the fifth major
domestic consumer of coal in the globe after China, US (United States), India and Russia (BP, 2012).
- 34 -
Table 2.4: South Africa’s historic coal production and consumption (tonx106)
Year
Total production
Local consumption
2000
224.1
154.6
2001
223.5
152.2
2002
220.2
157.6
2003
239.3
168
2004
242.8
178.3
2005
245
173.4
2006
244.8
177
2007
247.7
182.8
2008
252.7
197
2009
250.6
184.7
2010
257.2
186.4
2011
252.8
177.7
Total
2900.7
2089.7
Source: Adapted from Department of Mineral Resources (2012)
Exports
69.9
69.2
69.2
71.5
67.9
71.4
68.7
67.7
60.6
60.5
66.8
68.8
812.2
Coal production and consumption
300
250
tx106
200
Total production
150
Local consumption
Exports
100
50
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 2.3: SA’s historic coal production and consumption (tx106)
Source: Adapted from Department of Mineral Resources (2012)
2.4.2
Coal prices
Table 2.5 and Figure 2.4 show the trajectory of local and export sale prices in the past decade. The
domestic coal sale price (Free on rail – FOR) increased by 16% from R181 per ton (2010) to R210 per ton
(2011), yielding a higher domestic sale revenue of R37.3 billion in 2011 despite the 4.7% decrease in local
coal sales between 2010 and 2011. A higher export sale tonnage of approximately 3% between 2010 and
2011, coupled with an export sale price (Free on board – FOB) surge of 31% between 2010 and 2011 (i.e.
- 35 -
from R561 to R735 per ton), explains the higher export sale revenue of R50.5 billion in 2011. Figure 2.4
further shows the price of exported coal to be very unstable compared to the domestic price of coal.
However, a correction was made to the coal prices for differentials in calorific values. Eskom, for instance,
burns coal with a calorific value ranging between 17 and 22MJ/kg (Chamber of Mines, 2011) while export
coal, on the other hand, is generally washed and is characterized by high heating values ranging between
24.7 and 26MJ/kg (Eberhard, 2011). The average domestic calorific value was calculated to be 19.5MJ/kg
(17–22MJ/kg) while the export calorific value was calculated to be 25.35MJ/kg (24.7–26MJ/kg). This yielded
an adjustment factor of 0.769231 (i.e. 19.5MJ/kg ÷ 25.35MJ/kg) which was used to adjust the export coal
price. Column C in Table 2.5 shows the adjusted export coal price while Column D shows the price ratio (i.e.
adjusted export coal price ÷ domestic coal price). Comparing the domestic coal price (Column A) to the
adjusted export coal price (Column C), it becomes evident that the domestic price of coal is lower (by a
great margin) than the export coal price with a similar calorific value.
For 2011, for example, the domestic coal price was R210/t while export coal with the same calorific value as
domestic coal, would have been sold for R565/t. Domestic miners receive lower returns per ton of coal
from serving the domestic market than serving the export market. The price ratio in Column D shows that
on average between 2000 and 2011, the price of domestic coal with a calorific value of 19.5MJ/kg was
three times lower than that of export coal with a similar calorific value. In 2011, domestic miners supplying
the export market received 207% higher returns per ton of coal than domestic miners supplying the
domestic market. Coal investors and producers thus have an incentive to supply the export coal market
rather than to feed the domestic coal market. Nonetheless, the country’s old and inefficient rail
infrastructure presents a major hurdle to supplying the export market.
- 36 -
Table 2.5: SA’s historic coal price with and without adjustment (R/ton)
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Average
Local price – FOR
(A)
57
63
75
79
76
86
92
108
153
187
181
210
114
Export price – FOB
(B)
Adjusted export price
(C)
160
245
280
189
213
296
316
361
737
512
561
735
384
123
188
215
145
164
228
243
278
567
394
432
565
295
1
Price ratio
(D = C/A)
2.2
3.0
2.9
1.8
2.2
2.6
2.6
2.6
3.7
2.1
2.4
2.7
3
1
Export coal price adjusted for differences in calorific values between domestic and export coal
Source: Adapted from Department of Mineral Resources (2012)
Coal price: domestic and export
800
700
600
R/ton
500
Domestic price - FOR
400
Export price - FOB
300
200
100
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 2.4: Domestic and export coal prices in R/ton (not adjusted)
Source: Adapted from Department of Mineral Resources (2012)
2.4.3
Domestic coal consumers
Domestic coal is used by various users, amongst others, Eskom’s use of coal for electricity generation, coal
use by Sasol in its coal-to-liquid-fuel plants and the use of coal by small merchants (supplying residential
users and small businesses), metallurgical industries and other industries. Of the 177.7 million ton of coal
consumed in South Africa in 2011, electricity generation consumed almost two thirds of domestic coal sales
- 37 -
(65.5%), followed by the synthetic fuel sector at 22.6%, industries at 5.2%, metallurgical at 3.1%, merchants
and domestic sector at 3%, mining at 0.2% and others at 0.4% (Department of Mineral Resources, 2012).
See Figure 2.5 below.
Domestic coal users by sector 2011
5.2%
3.1% 3%
0.2%
0.4%
Electricity
Synthetic fuel sector
22.6%
Industries
Metallurgical
65.5%
Merchants & domestic sector
Mining
Others
Figure 2.5: SA’s domestic coal users by sector 2011
Source: Adapted from Department of Mineral Resources (2012)
2.4.4
Export coal consumers
South Africa exports coal to a number of regions, including Asia, Europe, the Middle East, Latin America,
Islands and Africa. Among the Asian countries, the main customers are India and China, while European
exports go to countries like Germany, Belgium, Switzerland, France and The Netherlands. The Middle
Eastern customers include Israel, United Arab Emirates and Turkey, while in Africa coal is exported to
Mozambique and Morocco, among other countries. Island customers include Mauritius and Latin American
customers include Brazil, Mexico and Chile (Eberhard, 2011; Department of Mineral Resources, 2012).
Figure 2.6 presents the country’s export coal sales by region in 2011.
- 38 -
Export coal sales by region 2011
2% 3.8%
0.3%
6.3%
Asia/Far East
11.1%
Europe
Middlle East
58.1%
18.4%
Africa
Islands
America
Australasia
Figure 2.6: South Africa’s coal export sales by region 2011
Source: Adapted from Eberhard (2011); Department of Mineral Resources (2012)
Of the 68.8 million tons of coal exported in 2011, the Asian market consumed the most (58.1%), followed
by the European market (18.4%), the Middle East (11.1%), Africa (6.3%) and the Islands (2%). The Asian
market is South Africa’s largest coal importer. India was the leading customer in 2011, importing 25.2% of
South Africa’s coal, followed by China with 17.8%. The Middle East increased its coal imports in 2011,
doubling its 2010 volume. Mozambique became the main importer of South Africa’s coal in Africa,
importing 5.3% in 2011 (Department of Mineral Resources, 2012). Coal exports to Europe have been
decreasing since 2005, from a high of three-quarters of the country’s exports in 2005 to below half in 2009.
Colombia and Russia are South Africa’s competitors in the European markets (Eberhard, 2011).
While there are various export coal products in South Africa, export coal is broadly classified into RB1 and
RB2 (RB = Richards Bay) coal specifications. These specifications generally refer to an A grade product with
an ash content of 15% and a calorific value of 6 000kcal/kg. RB1 and RB2 differ mainly with regard to their
volatile matter. For an RB1 specification, volatile matter has to be a minimum of 22% on an as-received
basis, while for RB2, volatile matter has to be a minimum of 25%. In South Africa, higher coal grades are
generally reserved for the export market. A higher calorific value and lower ash content constitutes a
higher grade of coal (Steyn & Minnitt, 2010). The export coal is washed, an activity that has guaranteed a
homogeneous product and has earned South Africa a good reputation in the international coal market
(Eberhard, 2011).
- 39 -
While South Africa’s coal sales to Europe have been decreasing, the country’s coal sales to Asia have been
increasing. The country’s competitors in the Asian markets are Indonesia and Australia. Although export
coal in South Africa is generally characterized by high heating values ranging between 24.7MJ/kg and
26MJ/kg (Eberhard, 2011) with a maximum ash content of no more than 20% (Chamber of Mines, 2011),
China imports low grade coal from South Africa. Although China is proposing to impose an import ban of
low grade coal in order to curb pollution and favour local coal mines, it has faced fierce protest from power
utilities (Business Day and Financial Mail (BDFM) Publishers, 2013a). In 2012, South African coal exports to
China surpassed those of India and it is forecasted that South Africa and Australia are likely to surpass
Indonesia as the leading primary coal provider to China (Mineweb, 2013).
In South Africa in the past, low grade coal (i.e. coal with a calorific value of 17–22MJ/kg) was used only by
Eskom to fuel its power stations. Recently though, low grade coal has become a contested commodity. No
longer is it for the exclusive use of Eskom to fuel its power stations, but it is also exported to China and sold
at export parity prices (Uninterruptible Power Supplies Direct, 2012). The emergence of export markets for
Eskom-grade coal, coupled with other issues such as underinvestment in new capacity in the coal industry,
have caused a high level of uncertainty on future domestic coal prices (Creamer Media, 2013). The
domestic coal market faces migration of domestic prices to export parity price levels. Although Eskom
purchases most of its coal through long-term contracts (cost-plus and fixed-cost contracts) (Eberhard,
2011), which basically means that most of its coal requirements are secured, the amount of coal acquired
by Eskom through short-term contracts has been rising over the years due to underperformance of its costplus mines (Creamer Media, 2013). In addition, securing long-term coal supplies has been reported as a
problem for Eskom (Uninterruptible Power Supplies Direct, 2012).
Coal acquired through short-term contracts was 17% in 2007, and rose to 30% in 2011. Currently, the utility
acquires approximately 30 million tons of coal through short-term contracts (Creamer Media, 2013) while
burning about 124.7 million tons of coal (Eskom, 2011). It is further forecasted that coal shortages of about
40 million tons per annum will be forthcoming after 2018 (Creamer Media, 2013), an incident that is likely
going to cause coal prices to soar tremendously.
2.4.5
South Africa’s coal producers
In South Africa only private companies conduct coal mining. The coal is mainly positioned in thick level
seams at low depths, making its extraction easier and relatively cheaper. It is, however, for the most part
low-quality coal with high ash content (Department of Energy, 2010). Almost half the coal harvested in
South Africa is mined from opencast mines while the rest is harvested through underground mining
- 40 -
methods. Of the country’s ROM production of about 316.2 million tons in 2011, opencast mining
contributed the highest (at 61.9%), followed by board-and-pillar mining (at 33.9%) while long-wall and
stopping mining each accounted for 2.1% (Department of Mineral Resources, 2012).
The active coal mines in South Africa are shown in Figure 2.7 and most of them are located in the
Mpumalanga province (Statistics South Africa, 2010). The Mpumalanga Central basin – a basin consisting of
three coalfields, namely Witbank, Highveld and Ermelo coalfields – accounted for 83.3% of the country’s
total production in 2011. The Witbank, Highveld and Sasol-Vereeniging coalfields together accounted for
89.4% of total production. The Witbank coalfield produced the highest tonnage (52.3% of total production),
followed by the Highveld coalfield (29.5%) and Sasol-Vereeniging coalfield (7.6%) (Department of Mineral
Resources, 2012).
The coal mining companies in South Africa can, in general, be categorized into three groups, namely major
coal miners, junior coal miners and Broad-Based Black Economic Empowerment (B-BBEE) companies. This,
however, is not a precise classification because B-BBEE companies fit into more than one category. The six
major producers that produced 80.7% of the country’s total production in 2011 are Exxaro Resources, BHP
Billiton Coal South Africa, Sasol Mining, Anglo Coal, Xstrata Coal South Africa and Optimum Coal Holdings.
Exxaro Resources and Optimum Coal Holdings are B-BBEE companies. The remaining 19.3% was produced
by junior coal producers. Junior coal miners and B-BBEE companies jointly accounted for 41% of total coal
production. Three major B-BBEE companies, namely Exxaro Resources, Optimum Coal Holdings and
Umcebo Mining produced 26% of the country’s total production (Department of Mineral Resources, 2012).
- 41 -
Figure 2.7: South Africa’s active coal mines
Source: Statistics South Africa (2010)
2.5
Summary
The concept of externalities, coal-fuel cycle externalities and the South African power and coal industries
were
reviewed
in
this chapter.
An
externality
was
discussed to
occur each
time
the
production/consumption decisions of an agent affects the utility of another in an unintentional manner and
when no/full compensation is made to the affected party by the producer of the undesirable effect.
Externalities cause market failure, which in turn leads to non-optimal resource allocation in society’s view.
A number of environmental and societal impacts associated with the coal-fuel cycle were discussed. The
impacts were categorised into three main classes, namely coal mining and transportation impacts, plant
construction impacts and plant operation impacts. In the discussion of the South African power industry,
special focus was given to Eskom’s power stations, electricity sales, coal quality, emissions profile, coal
supply and coal supply contracts. The discussion of the county’s coal industry on the other hand, focused
on the trends of coal production, consumption and prices. The country’s main coal producers and
consumers and as well as export coal consumers were also discussed. The review also highlighted the
problems faced by the power and coal industries. The power and coal industries were found to be of major
importance to the development of the South African economy.
- 42 -
CHAPTER 3:
ECONOMIC PHILOSOPHY AND SYSTEM DYNAMICS
PHILOSOPHY
3.1
Introduction
This chapter seeks to ground the research conducted in this study within the economic discipline of study
within which it falls and to motivate the use of a systems approach to model the life-cycle burdens and
social costs of coal-based electricity generation by studying the links between system dynamics and the
schools of economic thought that underpin this study. In pursuit of these aims, this chapter begins by
defining the concept of research paradigms which is followed by a discussion of Guba and Lincoln’s social
science research paradigm framework for deliberating main matters of research methodology in social
science. Section 3.4 reviews the history of economic thought, discussing developments in the economics
research field since the 16th century and using Guba and Lincoln’s conceptual framework to evaluate the
developments. Based on section 3.4, section 3.5 briefly discusses the research paradigms that provide the
theoretical basis for this study. Section 3.6 reviews system dynamics origins, main features, and the
modelling process and its links with social theories. Based on section 3.6, the section 3.7 and 3.8 try to
place the system dynamics practices of the energy literature and that of this current study on Pruyt’s
extended paradigmatic table. Section 9 summarizes this chapter.
3.2
Research paradigms defined
Generally “paradigms” are beliefs that guide actions (Guba in Creswell, 2008) and have also been termed
“worldviews” (Patton, 1990). According to Kuhn (1970), a research paradigm represents the whole
collection of beliefs, values, and methods that are mutually embraced by associates of a research
community. A research paradigm is therefore a set of beliefs or assumptions that regulate inquiry in a
discipline by providing a philosophical and conceptual framework through which organized investigations in
that discipline are accomplished (Filstead in Ponterotto, 2005; Schnelker, 2006; Weaver & Olson, 2006).
3.3
Guba and Lincoln’s social science research paradigm framework
The literature unveils various research paradigms that could potentially guide a research effort but also
various ways in which paradigms have been categorized and labeled (Ponterotto, 2005; Creswell, 2008).
One influential typology of research paradigms was that developed by Guba and Lincoln (1994; 2005).
These researchers distinguish between four research paradigms, namely positivism, post-positivism (critical
realism), critical theory, and constructivism (interpretivism). These research paradigms are presented in
Table 3.1 and discussed below, in terms of the philosophical beliefs/assumptions researchers place on
- 43 -
reality (ontology), manner of knowing and construction of knowledge (epistemology), the values
underpinning ethics, aesthetics and religion (axiology) and the procedures and techniques the researchers
use to investigate what can be known (methodology).
Table 3.1: Research paradigms
Research paradigms
Research
inquiry
Positivism
Ontology
Naïve realist - real
but apprehendable
Epistemology
-Objectivist/dualist
-Causal determinism
-Finding true
Axiology
Methodology
Post-positivism
Critical realism – reality
but only imperfectly
and probabilistically
apprehendable
-Modified
objectivist/dualist
-Causal determinism
-Critical community
-Findings probably true
Propositional knowing about the world is an
end in itself, is intrinsically valuable
-Experimental
-Chiefly quantitative
-Verification of
hypothesis
-Modified
experimental
-Critical multiplism
-May include
qualitative
-Falsification of
hypothesis
Critical theory
Historical realism -virtual
reality shaped by social,
cultural, political, ethnic,
economic and gender
values
-Formed over time
-Subjectivist
-Value-mediated findings
Constructivism
Relativism - local
and specific
constructed
realities
-Subjectivist
-Created findings
Propositional, transactional knowing is
instrumentally valuable as a means to social
emancipation, which as an end in itself, is
intrinsically valuable.
Dialogic/dialectical
Hermeneutical/
dialectical
Source: Adapted from Guba and Lincoln (1994)
Traditionally, before 1930, research was modelled after the physical and natural sciences (i.e. hard
sciences). During this period scientific investigation and empiricism were the yardstick for research and
were represented by positivism. Following the physical and natural sciences, new research fields such as
social sciences imitated this successful paradigm but later other paradigms such as post-positivism emerged
as researchers started to question the applicability of the positivist approach to human behaviour and
society (Plack, 2005).
3.3.1
Positivists and post-positivists
Ontologically, positivists believe in an apprehendable, identifiable and measureable reality, while postpositivists - though they acknowledge an objective reality they believe in one that is only
partially/imperfectly apprehendable (Denzin & Lincoln, 1994). In terms of the nature of knowing and
construction of knowledge (i.e. epistemology), they both believe in the existence of laws/theories which
- 44 -
control/regulate/direct planet earth, which must be tested/verified and refined to enable humans to
understand the world. The knowledge that develops through both these research paradigms is bound by
causality.
Therefore
both
paradigms
hold
causal
determinism
in
that
causes
determine
outcomes/effects/events. The problems studied by positivists and post-positivists thus display the necessity
of identifying and evaluating the sources that shape results through cautious examination plus
measurement of the objective reality in the actual world. The development of numeric measures of
examination is therefore paramount to both research paradigms (Plack, 2005; Creswell, 2008). Positivists
use quantitative research to verify truth while post-positivists primarily use quantitative methods but also
some qualitative approaches as a method of falsifying a priori propositions (Guba & Lincoln, 1994; Crotty,
1998). Axiologically, they both look at propositional knowledge. Positivists argue that investigations must
be value and bias-free replicable while post-positivists recognize the existence of human-being interactivity
and try hard to minimize such bias (Guba & Lincoln, 1994; Ponterotto & Grieger, 2007).
3.3.2
Constructivists
On the other hand, constructivists ontologically believe in multiple socially constructed realities that can
only be imperfectly grasped (i.e. subjective reality). Epistemologically, constructivists create knowledge
through interactions and knowledge is accepted through relative consensus. Unlike positivist/post-positivist
researchers, constructivist researchers have no interest in forecasting the future or making gross
generalizations but within a specific context of human action, the focus is on understanding the subtle and
distinctive differences in human behaviour through trying to understand the way in which meanings are
fashioned, negotiated and customized (Plack, 2005). The researcher is the primary research tool and is
intimately involved with the inquiry (Merriam, 2002). Axiologically, constructivists focus on both clear and
linguistic-centered propositional information and implicit plus tacit information (Guba & Lincoln, 1981).
Methodologically, constructivists’ inquiry differs from that of positivists/post-positivists, in that they
commence with an enquiry/concern instead of an a priori proposition from theory and the inquiry takes
place in the real-world setting (Creswell, 2008). The inquiry is informal, interactive and takes an explanatory
and/or descriptive stance. The constructivist researcher uncovers embedded meaning through words and
text and therefore employs only qualitative methods (Ponterotto & Grieger, 2007).
3.3.3
Critical theorists
The critical theorist, on the other hand, believes that describing and understanding human behaviour as
done by the constructivist researcher is not sufficient. Thus they aim at advancing the welfare of humans,
especially marginalized individuals in society through fighting oppression and inquiring the status quo. The
- 45 -
critical theorist researcher’s goal is therefore to empower members to alter the status quo and liberate
own-selves from on-going domination (Plack, 2005). The critical theorist researcher is therefore not simply
concerned with generating new knowledge but facilitating social change (Kim, 2003). In essence, critical
theory is a collection of various paradigms, among which are feminism, cultural studies, neo-Marxism,
social theorists, materialism and racialized discourses (Kincheloe & McLaren, 1994; Denzin & Lincoln, 2000).
Ontologically, they accept as truth that all that can be known is fundamentally a historical realism formed
by a number of factors, including political, social, racial, cultural and economic factors. Epistemologically, as
the manner of investigation is wholly value based, the critical inquirer’s values are central to the inquiry
coupled with those of participants. Knowledge is shaped by the relations between the inquirer and
participants (Guba & Lincoln, 1994). Axiologically, this research paradigm is dialectic dialogue that reveals
the unrevealed suppositions through which day-to-day happenings are interpreted (Kincheloe & McLaren,
2000). In the following section the history of economic thought is reviewed.
3.4
Economic disciplines and research paradigms
Economics, like other social sciences, is characterised by the existence of diverse schools of thought. The
history of economic thought begins with a discussion of two concepts, namely mercantilism and
physiocracy, which denotes a system of early economic policy and the development of economic doctrines.
This is then followed by a discussion of several concepts, for instance classical, neoclassical, heterodox and
environmental economics. At the end of each discussion an attempt is made to classify the economic
disciplines according to the research paradigms of Guba and Lincoln discussed earlier.
3.4.1
Mercantilism and Physiocracy
Among the most primitive economic ideas in history is mercantilism. Mercantilism theorists hold that
wealth consists in gold and silver (Butler, 2011) and that trade is a zero sum game, with no mutual benefits
from trade and that if there was a nation that gains from trade the other nations were losers. At the core of
mercantilism is the view that national prosperity can best be attained through maximizing net exports.
Mercantilism, in essence, is based on bullionism - an economic theory that considers a country’s wealth and
success by the quantity of valuable metals (gold or silver) the country owns (Pojer, n.d.). A country with
more precious metals than another was therefore considered a rich one (Butler, 2011). This notion had
fundamental repercussions for economic policy. A trade surplus had to be maintained by means of export
surplus, so a country had to export more than it imports (Magnusson, 2003) and tariffs were used to
encourage exports and discourage imports (Rothbath, 2010). Agriculture and manufacturing were
promoted in order to increase exports and restrict imports, and sea power was essential to control foreign
markets (Pojer, n.d.). Maintenance of a positive trade balance was time and again backed by military might
- 46 -
(Rommelse, 2010). The mercantilism doctrine overshadowed European business activities’ course of action
in the 16th to late 18th century. During this time mercantilism promoted colonial expansion and was the
reason behind persistent European wars (LaHaye, 2008). In spite of its prevalence, however, it only
appeared in print in 1763 by Marquis de Mirabeau and was popularised by Adam Smith, a classical
economist, who was strongly against its ideas (Magnusson, 2003). Amongst other scholars against
mercantilism were John Locke and David Ricardo.
The mercantilism doctrine coexisted with the physiocracy doctrine - the first school that rejected
mercantilism. Physiocracy is a new science that saw the wealth of the nation originating from nature, in
particular agriculture. It is also called the government of nature. Physiocrats, like mercantilists, studied the
economy with the goal of developing economic policies, but unlike mercantilism, physiocracy was led by an
intellectual leader, Francois Quesnay. Quesnay was analytical, designed conceptual models that gave the
stance of science to the study of economy (Lluch & Argemi, 1994) and supported perfect liberty (Butler,
2011). Quesnay’s economic theories included the idea that the source of national wealth was the
productive sector, in particular agriculture, and that taxation should be solely imposed on the landowning
class (Lluch & Argemi, 1994; Pojer, n.d.). Quesnay developed a tableau economique, which represented the
economic system in three networking classes, namely property, productive and sterile classes which
represent landowners, agricultural labourers and artisans and merchants, respectively. In the tableau
shown is the regeneration of income, a landowner receives rent and spends it on products of artisans and
agriculture, who also in turn purchase other products (Phillips, 1995). The tableau portrays the intersectoral
flow of money and commodities in an economy (Bilginsoy, 1994). Quesnay’s tableau has captivated
numerous students of economic doctrines and is regarded as one of the greatest discovery and an initial
fruitful attempt to examining a nation’s wealth on a macro-economic basis (Marx, 1952; Newman, 1962;
Giancarlo, 2011). It marked the beginning of general equilibrium theories and was a basis for welfare
analysis (Bilginsoy, 1994).
3.4.1.1
Criticism of the early political economy schools
Hume is regarded as a precursor of Adam Smith in that some of his ideas, even those against mercantilism,
were reflected in Smith’s book “Wealth of Nations” (McGee, 1989), among which is the fallacious goal of
continuous positive balance of trade (Ekelund & Hebert, 1975; McGee, 1989) and mercantilists’
misconception of money and wealth (McGee, 1989). According to Smith, wealth consists of not only gold
and silver but also land, houses and various consumable goods. Silver and gold do not therefore describe
the wealth of a country nor is it the sole benefit of foreign trade (Butler, 2011). On the other hand, John
Locke pointed out that human labour generated the wealth of the globe and that it is not unchanging as
- 47 -
mercantilists believed. David Ricardo indicated the failure of mercantilists to comprehend the concepts of
absolute and comparative advantage and the benefits of trade. Smith was also in opposition to the
mercantilist acceptance as truth that trade was a zero-sum game but recognized instead that trade was a
positive-sum game, as foreign trade can encourage growth of production capability of a country and
increase a country’s real worth (Rosenberg, 1960). Using the theory of laissez faire (“leave things alone”, a
belief than an economy is self-regulating) in analysing the trade problem, Smith explained that an invisible
hand would guide trade in a similar manner as done in domestic economic performances, hence he
critiqued the mercantilist trade policy of intervention and monopolising trade. Smith considered the wealth
ideas and trade theory of mercantilists as nonsensical and untenable (Rosenberg, 1960; Manis, 2005). On
the other hand, though Smith did not fully approve of all the ideas of physiocracy, he preferred it over
mercantilism because it recognized wealth to consist of not only gold and silver but in a nation’s
production, and embraced perfect liberty as the finest manner to maximize a nation’s wealth. Smith found
the main error with physiocracy to be the view of artisans and merchants as a sterile or unproductive class
(Butler, 2011).
3.4.1.2
Appraisal
As stated earlier, the purpose of describing the mercantilism and physiocracy theories was to offer
background on the development of economic doctrines, so no attempt is made to classify the early
doctrines into the economic research paradigms of Guba and Lincoln (1994). It is also doubtful that
nowadays there is any economist that considers himself/herself mercantalist or physiocrat, however, the
theories developed by mercantilists and physiocracts are the foundation of what became modern
economics.
3.4.2
Classical economics school
Influenced by mercantilism and physiocracy theories, the classical economic school is often called classical
liberalism, as it is based on the liberal doctrine. Beginning with the wealth of nations work in 1776, classical
economists supported free market economy. They believed that free markets regulate themselves. Smith,
using the concept of the invisible hand, explained how resources would be allocated and how the market
would move towards equilibrium without intervention. Classical economists defended that free trade
would promote efficient use of resources and boost welfare. Out of the explanations of foreign trade and
its mutual benefits came concepts such as specialization and theories of comparative and absolute
advantage. Classical economists developed the labour theory of value. A manufactured good’s value was
believed to be contingent on the costs of creating it, i.e. rent for the landlord, wages for workers and profits
for capitalists. The real price of anything was therefore the trouble of getting it. The market price
- 48 -
fluctuations varied according to market forces and altered factor prices (wages and profits). From these
came the concepts of perfect market competition and the law of one price (Kucukaksoy, 2011). Later some
classical economist began stressing the value of a good to the consumer. Classical economists focused on
analysing the causes of economic growth, allocation of surplus output and promoted policies that
enhanced the wealth of nations (De Vroey, 1975). Emphasis was therefore on production and what
influenced the supply of goods. The goal of classical economists was to assist policy makers to increase the
wealth of nations (Meek, 1973). Classical economists introduced numerous ideas that are used in presentday economies, especially in international economics and microeconomics doctrines. Among the classical
economists were Adam Smith, William Petty, David Ricardo, John Stuart Mill, and the unorthodox Robert
Malthus.
3.4.2.1
Appraisal
The problems studied by classical economists reflected the necessity of identifying and evaluating the
sources that shape results through cautious examination plus measurement of the objective reality in the
actual world. The development of numeric measures of examination was therefore paramount and the
knowledge that developed was bound by causality. Classical economists, in analysing social phenomena,
stressed the concept of class (i.e. landlords, capitalists and workers) instead of individuals and historical
analysis was the tradition. For instance, using historical analysis, they made efforts to explain capitalist
mode of production. The ontology, epistemology and methodology of classical economists imply a positivist
paradigm.
3.4.3
Neoclassical economics school
Neoclassical economics was first coined by Thorstein Veblen over a century ago while referring to a school
of economic thought (Lawson, 2013). The literature, however, discloses diverse interpretations of the term
neoclassical economics (Hahn, 1982; Weintraub, 2002), perhaps due to the fact that the school has evolved
since its introduction in the 1870s (Dequech, 2007). Broadly, neoclassical economics is characterized by
rationality, economic self-interest (Weintraub, 2002), methodological individualism, equilibrium analysis
(Hahn, 1984), and mathematical techniques (Brennan & Moehler, 2010; Lawson, 2013). Some scholars,
however, note that present-day economics is transitioning from some of the substantive categories, namely
economic selfishness, rationality and equilibrium to well-informed self-interest, purposeful behaviour and
sustainability (Colander, Holt & Rosser, 2004; Davis, 2005).
Unlike classical economists whose focus was on economic growth and hence assisting policy makers,
neoclassical economists focused on efficiency, i.e. optimal allocation of scarce resources among alternative
- 49 -
uses (De Vroey, 1975). Also in contrast to classical economists whose core theoretical structure was
centered on capital, the central concept for neoclassical economists is focused on prices – i.e. economic
analysis focuses on the determination of equilibrium prices in factor and products markets (Eagly, 1974).
Prices in this school are determined by subjective preferences (desires and beliefs) of consumers. Agents
are assumed to have rational preferences. While in principle the preferences of agents could be driven by
group interests, in practice it is argued there is predominance towards self-interested motives with agents
maximizing their own well-being. Economic behaviour is thus conceived as a complex exchange between
rational individuals (Brennan & Moehler, 2010) - individuals maximize utility while firms maximize profits
(Weintraub, 2002). The analysis of the interplay among rival interests has, however, tended to focus on
equilibrium analysis. Mathematical techniques formalize the complex interactions of agents (Brennan &
Moehler, 2010).
Neoclassical economics has dominated the twentieth century and introduced a number of theories in
connection with economic activity, among which are theory of production, theory of consumption,
marginal (productivity and utility) theory, theory of diminishing returns and theories of general
equilibrium and Pareto efficiency. The school thus dominates microeconomics. It has, however, been
criticized: for reliance on methodological individualism as its unit of analysis; on the assumption of rational
choice of individuals and optimization, which has been viewed as overlooking essential aspects of human
behaviour (i.e. real people often do not resemble the “economic man”, they lack the ability to maximise
benefits from their choices (Boldeman, 2007); for normative bias in that instead of explaining actual
economies as observed empirically it focuses on utopia (Pareto-optimality and welfare) (Eichner & Kregel,
1975); and on the suitability of its general equilibrium theory in explaining evolving economies (Boldeman,
2007). Some of these criticisms have been merged into latest forms of neoclassical theory as cognizance of
economic benchmarks’ evolution while most of it has manifested itself in heterodox economics.
3.4.3.1
Appraisal
Neoclassical economics acknowledges a reality that is controlled by absolute laws of nature, and views
economics as an objective science that is value free. The school of thought focuses on rational explanation
of social matters while relying on mathematical techniques. The concept of rationality is an end in itself
with no queries raised concerning the source/value of preferences. The ontology, epistemology and
methodology of neoclassical economics thus imply a positivist paradigm.
- 50 -
3.4.4
Heterodox economics
Heterodox economics designates various schools of economic thought which oppose the neoclassical
approach to understanding socio-economic performance. The reasons behind the rejection of the
neoclassical orthodoxy vary among the heterodox schools and there is no single heterodox theory (Gabriel,
2003). The heterodox schools are a growing movement that has been challenging neoclassical economics
since the 1870s. Heterodox schools of the time included historical schools and various supporters of
mercantilism. After 1945 Keynesian economics became absorbed into the mainstream, forming neoclassical
synthesis - partitioned into microeconomics and macroeconomics. The heterodox schools that opposed this
synthesis were post-Keynesians, Austrians, Marxist and institutional schools. The mainstream after 1980
became challenged by various research programmes which can also be adapted to heterodox economics,
namely evolutionary economics, behavioural economics, experimental economics, complexity economics
and neuroeconomics (Davis, 2006). Among the heterodox schools, two influential heterodox schools are
discussed further and classified into the research paradigms of Guba and Lincoln namely, Austrian and
institutional economics.
3.4.4.1
Austrian economics and appraisal
Austrian economics rejected the neoclassical economics approach of explaining market phenomena by way
of exact and universal laws. While Austrian school embraces methodological individualism in explaining
economic phenomena, it rejects the neoclassical economics’ “economic man” and considers a “perceiving
man”, particularly a “man who grasps the future” (Selgin, 1988). In addition, unlike neoclassical economists
the Austrian economists embrace methodological subjectivism, embrace a subjective theory of value and
reject empirical modelling, mathematical and statistical methods as they consider individuals too complex
(Fritz, 2004) and embrace instead historical description and understanding of social occurrences (Selgin,
1988). Austrian economics also advocates for complete elimination of government control, an extreme
case of laissez faire approach (Raico, 1995). Among Austrian economists are Carl Menger, Friedrich von
Wiese and Eugen von Bohm-Bawerk. From this discussion the ontology, epistemology and methodology of
the Austrian school imply a critical theorist research paradigm.
3.4.4.2
Institutional economics and appraisal
Original institutional economics opposes the rational “economic man” model of neoclassical economics,
and instead stresses the habitual and routinized personality of human conduct (Veblen, 1919; Stanfield,
1999). The institutions as opposed to individuals are the center of analysis. Institutions are seen as having a
cognitive dimension which provides a frame for processing data into meaningful knowledge (Hodgson,
1988). The approach to economics is holistic, systematic plus evolutionary (Wilber & Harrison, 1978) in
- 51 -
pursuit of understanding the dynamics of the socio-economic system (Veblen, 1919; Ayres, 1962). Focus is
therefore not on rational, static and equilibrium processes. Theory development is focused on explanation
instead of prediction (Arvanitidis, 2006). Among the original institutional economists are Thorstein Veblen
and John Commons (Mirowski, 1987). Based on the research paradigms of Guba and Lincoln discussed
earlier, the ontology, epistemology and methodology of the original institutional economics suggest a postpositivist research paradigm.
3.4.5
Environmental economics and ecological economics
The importance of nature/environment was noted by classical and neoclassical economists but the
comments they made were not reflected in their exposition of theories. The Malthusian scarcity (1798),
Ricardian scarcity (1817) and Jevon’s coal question (1865) represent such earlier works (Cracker & Rogers,
1971; Common, 1988). It was, however, not until the 1960s and 1970s that the deteriorating quality of the
natural environment prompted scholars to apply economic tools to environmental science. The increasing
scarcity of none market reflected resources, for example clean air, soil and water became viewed as a
consequence of market failure but disagreement arose on how the environmental crisis should be studied,
mainly stemming from differences in scientific views. Environmental economics developed and used the
theories and methods of neoclassical regime, however, continued environmental deterioration and
numerous oppositions to the environmental economics’ approach of treating the natural environment
within the neoclassical economics framework, led to the formulation of ecological economics (Boyce,
2011).
Environmental economics is the study of economy and environmental association with specific focus on
regulation/control with the ultimate goal of sustainable development, while ecological economics is the
study of economy and ecosystem interrelationships in the light of biophysical limits with a specific focus on
stewardship and has the same goal of ensuring sustainable development (Sahu & Nayak, 1994). Ecological
and environmental economics are thus two different sub-disciplines of economics that address
environmental issues that now operate as two dissimilar disciplines of economics and environmental
science (Sahu & Nayak, 1994). Following in the practice of neoclassical economics, environmental
economists assume a reality that is controlled by absolute laws of nature (Tacconi, 1998), and maintain that
economics is a positive science that is value-free, that is, they maintained that economics is intended to
describe facts devoid of personal value or subjectiveness permitted to affect the facts (Sahu & Nayak, 1994;
Tacconi, 1998). Ecological economists on the contrary, recognise a subjective reality, that is, not devoid of
personal value (Tacconi, 1998; Boyce, 2011).
- 52 -
The schools’ treatment of resources when solving environmental resource-related problems also differed.
In the light of a globalized network of resources resulting from globalization, environmental economists
employed a model of relative scarcity (prices that reflect scarcity) but in contrast to neoclassical
economists, they internalised the dreadful environmental consequences from production and
consumption. In contrast, ecological economists used the concept of absolute scarcity, with bounds on the
thermodynamics of resources (Sahu & Nayak, 1994), and in line with the biophysical approach to resources,
irrespective of the utility they provide, they are deemed to have value (Venkatachalam, 2007).
The schools also differ in terms of the valuation of environmental resources. Environmental economists
quantify environmental services using measures such as contingent valuation (willingness to pay and accept
cost surveys), travel cost method, hedonic pricing and cost benefit analysis while ecological economists use
evaluative methods such as environmental impact assessment, systems analysis and including other more
qualitative methods (Batabyal, Kahn & O’Neill, 2003). While ecological economists would make use of such
methods too, they would not entirely be responsible for the resource value (Panagopoulos, 2009).
3.4.5.1
Appraisal
As elucidated above, economic and ecological economics vary in terms of the underlying philosophy, views
on resource scarcity, valuation and methods. The assumption of a reality that is controlled by absolute laws
of nature, and the maintenance of that economics is an objective science that is value free, coupled with
the use of quantitative methods to value resources and to analyse environmental issues, renders
environmental economics a positivist philosophy like neoclassical economics. The appreciation of a
subjective reality, and perhaps multiple realities that can only be imperfectly grasped (Tacconi, 1998)
renders the ontology of ecological economics to that of a post-positivist and constructivist. The
constructivism ontology is useful for explaining matters such as sustainability, in the event that individuals
hold different views/explanations of limits (Boyce, 2011). The use of both quantitative and qualitative
methods, however, renders ecological economics a post-positivist philosophy.
3.5
Research paradigm(s) underpinning the current study
Among other negative consequences, the environmental degradation as a spin-off of coal-based electric
power production is viewed as a consequence of market failure. Among other factors, the presence of
externalities causes markets to fail to achieve Pareto efficiency, thereby causing a divergence between
private and social costs. Therefore an understanding of the intricate relations between the environment
and the power production system is needed to redress market failures in the electric power sector. The
review of economic thought discloses that the main concepts in this study, namely production, externalities
- 53 -
and social cost are rooted in neoclassical and environmental economics, particularly, in welfare economic
theory, theory of production and Pareto efficiency (more on these in the following Chapter). Neoclassical
and environmental economics are therefore the main economic disciplines that provide the theoretical
basis for this study. The ontology, epistemology and methodology of both neoclassical and environmental
economics fall within the positivist research paradigm of Guba and Lincoln’s classification. In the following
section, a review is conducted of the origins of system dynamics and its main features, its modelling
process, and its links with social theories.
3.6
System dynamics origins, features, modelling process and its links with social theories
3.6.1
Origins of systems theory
Ludwig von Bertalanffy, while conducting a biological study on living organisms, recognized the need to
study the living organisms not only in isolation but to consider their relations when studied as a whole. He
used the term “organism-as-a-whole” and suggested that this approach be employed in other fields of
study too (Von Bertalanffy, 1973). This approach is a systems thinking approach. Systems thinking analysis
looks at problems as parts of a whole system. It is premised on the understanding that a system can best be
known by examining the linkages and interactions between its elements. Ludwig von Bertalanffy
characterizes systems inquest into three key spheres of influence, namely systems technology, systems
science and systems philosophy. Systems technology developed from technological and organizational
challenges that necessitated integration of skills and knowledge from various domains of study in the
second half of the 20th century. Systems technology was, however, constrained by its chiefly instrumental
focus and mechanistic world conception (Von Bertalanffy, 1952).
In the early 20th century, while reacting to the growing disintegration and replication of scientific and
technological research, Von Bertalanffy proposed the development of a general science of organized
complexity (that is general theory of systems or, as commonly known, general systems theory) (Laszlo &
Krippner, 1998) as a way of reviving the unity of science (Von Bertalanffy, 1968). General systems theory is
whole, integrative and emphasises a structured world and is hence a drastic departure from the
mechanistic, linear causality and analytic paradigm of classical science (Von Bertalanffy 1955; Hommand,
2005).
But similar to other pioneering frameworks of thought, general system theory suffered ridicule and
abandonment (Laszlo & Krippner, 1998). For Von Bertalanffy the whole and humanistic methodology to
knowledge and practice is general systems theory’s major contribution (Hommand, 2005). Notwithstanding
the criticism general systems theory faced, it profited from analogous developments and prominence of
- 54 -
cybernetics plus information theory, coupled with their extensive application in many fields. In 1954
Kenneth Boulding and his colleagues realized that the systems approach was not restricted to the hard
sciences and they applied it to the social sciences (Laszlo & Krippner, 1998). The systems thinking approach
was also seen relevant to industrial engineering and management. In a study of management problems in
corporate settings, Jay W. Forrester (1958) proposed industrial dynamics (Damle, 2003). Later, various
disciplines employed industrial dynamics to address various problems, for example, it was used in urban
planning, economics and medicine. Owing to its diverse application in various settings, it became
transformed into a more general term called system dynamics (Damle, 2003). By the 1960s, on a transdisciplinary plane, systems thinking started being acknowledged, as an archetypal attempt at scientific
unification and theory construction (Laszlo & Krippner, 1998).
Systems science is therefore an important development that demonstrated the diffusion of Von Bertalanffy
thought (i.e. systems approach or system thinking) in all sciences (for example biology, physics, behavioural
and social sciences), accentuating interactions between parts and studying any system in association with
its environment (Hommand, 2005). Systems theory therefore centres on the arrangements of and
associations between parts which link them into a whole (Heylighen & Joslyn, 1992), and as a general frame
of inquiry concerned with the study of phenomena and events in a holistic and interactive manner, it is
connected to both epistemological and ontological views (Laszlo & Krippner, 1998). Stemming from a
systems (science) viewpoint of a relational approach to understanding reality, systems philosophy mirrors a
similar reality of worldview, that is, that of organization and interdependence, emphasizing relational
patterns between systems parts and the whole (Hommand, 2005). The ontological view thus suggests a
nature of reality that consists of systems. The epistemological view suggests a holistic approach highlighting
the relationship between systems and their parts/elements (Hjorland & Nicolaisen, 2005).
3.6.2
Origins of system dynamics and its main features
The origins of system dynamics were highlighted briefly while discussing the systems approach above. In
this section more background on system dynamics is provided in order to address its philosophical
background and its main features. The discussion firstly explains the work of the creator of system
dynamics, in order to establish the initial drives, assumptions and aims behind Jay Forrester’s development
of system dynamics.
What is presently known as system dynamics began in the 1950s when Jay Forrester started searching for
links between engineering and management. Equipped with knowledge in feedback control systems, in a
faculty seminar in 1956, Forrester criticized economic models on a number of accounts, for example, their
- 55 -
failure to reflect in a satisfactory manner the loop structures that characterize economic systems which
consequently led to exclusion of closed loops properties such as accumulations and delays, for not being
holistic and integrative - for example they failed to incorporate the flows of money, labour, goods and
information in one unified model, for describing systems using linear equations, for not incorporating
changing mental attitudes, for overconfidence in regression analysis in defining economic behaviour, and
lack of discussions of assumptions underlying economic models (Forrester, 2003).
Forrester visualized new firm-economy models that embrace characteristics such as dynamic structure,
incremental changes in variables, information flows, non-linear systems, non-linear differential equations,
model complexity, empirical solutions, symbolism (flow diagrams) and correspondence with real
counterparts and lastly, models that emphasize structure over coefficient accuracy (Forrester, 1975a).
Based on these initial thoughts, Forrester published a paper “Industrial dynamics: a major breakthrough for
decision makers” which emphasized that management should embrace unified systems given their
profession which necessitates relating the flows of materials, information, capital equipment, money, and
labour. The relations among these factors form a foundation for understanding the structure of the system
and for anticipating the consequence of decisions and policies (Forrester, 1975b).
Shortly after this paper, Forrester published a book in 1961 entitled “Industrial Dynamics” whereof the
major intention was to develop a science for designing effective industrial and economic systems. In the
book Forrester explains the industrial dynamics approach for devising effective systems (Damle, 2003). To
be included in the model are main factors that will help address the questions to be answered, cause-effect
information feedback loops are to be traced and focus must be on closed loop information feedback
structures, pictorial representations of the model through flow diagrams are to be conducted and the
model variables are to resemble those in the system being represented, formal decision policies are to be
formulated, interactions among the system components are to be described via mathematical
formalization, the model behaviour can then be generated and its outcomes compared against available
knowledge of the real system, model revision is to be conducted until the model resembles actual system in
a satisfactory manner, last but not least the model structure and policies are to be redesigned as can be in
actual system such that the modifications that produce better or worse behaviour can be established,
lastly, the real system can be altered in the manner that improves performance. Forrester also highlighted
that model validation/significance rests on the suitability of the model for the purpose it was designed for.
The focus therefore is not on predictions but on understanding the structure of the system and our
assumptions about it (Forrester, 1961; Damle, 2003).
- 56 -
In these early days, Forrester’s focus was on corporate operations, but in a follow-up book “Principles of
Systems”, he outlined the broader view of systems. He describes a system as an organization of parts that
function collectively for a common purpose, highlights the importance of structure in organizing knowledge
in any field of study, stressing the notion of mental models and importance of levels and rates as key
variables (Forrester, 1971; Forrester, 1975a). The science of designing systems that began in the 1950s as
industrial dynamics was rephrased into a broad-spectrum phrase called system dynamics, owing to its
diverse application in various settings, for example, in urban planning, economics and medicine (Damle,
2003).
System dynamics, though labelled by many as a method (Sterman, 2000; Lane, 2001), methodology
Roberts, 1978), theory (Jackson, 2003) and field of study (Coyle, 2000) has twofold intentions, firstly to
understand the behaviour of systems through detecting the factors driving the behaviour of the system.
Secondly to observe how the system responds to alterations of the said factors and then to make policy
recommendations that improve system performance (Damle, 2003). System dynamics models embrace
such characteristics as complexity, information flows, feedback behaviour, dynamic structures, causal loop
diagrams and their correspondence with real counterparts, stock and flow diagrams, difference equations,
non-linearity, confidence based on model structure, experimental approach and the construction of formal
models using computers (Forrester, 1975a; Lane & Oliva 1998; Damle, 2003; Pruyt, 2006).
As a tool that has been characterized as most suited for the framing and understanding of complex
problems, it is important to clarify such complex problems (Richardson & Pugh, 1981), which basically
refers to problems in feedback rich environments. The complexity stems from the system consisting of
various parts that interrelate and generate feedbacks. Feedbacks occur, for example, when a system factor
affects another system factor, which in turn affects the first factor, thus forming a loop. The feedbacks can
be self-reinforcing (positive) or self-correcting (negative). Self-reinforcing loops amplify change in the
system while self-correcting loops oppose change in the system and attempt to bring the system into
equilibrium. The feedbacks, together with delays, create dynamics in the system (Damle, 2003).
There are a number of basic modes of system dynamics behaviours, namely exponential growth which
occurs because of self-reinforcing feedback (i.e. the rate of growth rises rapidly over time)), exponential
decay where overtime the growth of decay rises rapidly, goal-seeking behaviour which occurs because of
self-balancing loop (i.e. the structure/corrective-action that attempts to move the state of the system
towards a desired state) and oscillation which occurs due to a delay in the self-balancing loop (Damle, 2003;
Sterman, 2000). The interactions among the basic behaviours result in derived modes of behavior, among
- 57 -
which are S-shaped growth which occurs because of interactions of a self-reinforcing and a balancing
feedback loop, S-shaped growth with overshoot which occurs because of the presence of delays in the
balancing loop of the S-shaped growth, and overshoot and collapse which is also displayed by the S-shaped
curve when the system’s capacity is variable and is diminished and destroyed by the system state. Some
other types of system dynamics behaviours include equilibrium behaviour where the system’s state
remains fairly constant and chaotic behaviour where the system oscillates irregularly (Damle, 2003). The
features of system dynamics mentioned in this section will be discussed more elaborately, while presenting
the system dynamics modelling process in the following section.
3.6.3
System dynamics modelling process
Various system dynamics modelling arrangements have been suggested by numerous researchers in the
literature, for example, Richardson and Anderson (1980), Richardson and Pugh (1981), Roberts, Anderson,
Deal, Grant and Shaffer (1983), Ford (1999), Forrester (2000) and Sterman (2000). Some of the suggested
modelling steps are presented in Table 3.2. The table suggests that there is substantial consistency in the
modelling steps to be followed when constructing system dynamics models.
Table 3.2: System dynamics modelling process
Richardson & Anderson
(1980)
Problem recognition
System conceptualisation
Roberts et al. (1983)
Ford (1999)
Problem definition
System conceptualization
Understanding the system
Dynamic problem
Model representation
Model representation
Model behaviour &
evaluation
Model use
Model behaviour &
evaluation
Policy analysis and model
use
Stock & flow diagrams
Causal loop diagram
Reference mode
Sensitivity analysis
Policy evaluation
Sterman (2000)
Problem articulation
Dynamic hypothesis
formulation
Simulation model
formulation
Testing
Policy formulation and
evaluation
Source: Adapted from Richardson & Anderson (1980), Roberts et al. (1983), Ford (1999), Sterman (2000)
3.6.3.1
Problem formulation
Problem formulation is the first and most important step in the model building process. It embraces a
number of activities, amongst others, proper description of the problem, identification of key variables that
need to be considered, determining the system boundary and establishing the time horizon for the model
(Sterman, 2000). A good understanding of the system by the modeller is therefore necessary for proper
formulation of the problem.
- 58 -
3.6.3.2
Dynamic hypothesis formulation
The dynamic hypothesis formulation stage involves constructing a working theory that explains the
problem. This theory explains/describes the dynamic behaviour of the system centered on the feedbacks
and causal structure of the system (Sterman, 2000). Causal loop diagrams are part of the dynamic
hypothesis formulation step and are an essential feature of system dynamics models that capture the
structure of the system in a qualitative manner. They indicate the cause and effect relations amongst the
variables in the system and feedback loops of the system. The relationships between the variables are
either positive or negative (see Figure 3.1). Positive polarity designates that, all else being equal, an
increase (decrease) in the “cause” element will increase (decrease) the “effect” element. So the cause and
effect elements travel in a similar direction. Negative polarity indicates that, all else being equal, an
increase (decrease) in the “cause” element will decrease (increase) the “effect” element. So the cause and
effect elements travel in opposite directions (Sterman, 2000).
Caus
Effect 0
Connector
+
Cause
Effect
Positive polarity
Cause
Effect
Negative polarity
Figure 3.1: Positive and negative causality
Source: Own construction
The interactions between the variables generate feedback loops which determine the dynamics of the
system. Feedbacks occur, for example, when one variable in the system affects another variable, which in
turn affects the first variable, thereby forming a loop. Feedback loops are either positive or negative (see
Figure 3.2) and one having an even number of “-” signs (or only “+” signs) is a positive loop (also called a
reinforcing loop), while one having an uneven number of “-” signs is a negative loop (also called a selfbalancing loop). Self-reinforcing loops amplify change in the system while self-correcting loops oppose
change in the system and attempt to bring the system into equilibrium (Coyle, 1996; Sterman, 2000). The
- 59 -
causal loop diagram, as a tool that illustrates in a qualitative manner the linkages and feedback loops of the
system, serves as a quick tool for capturing the hypothesis relating to the basis of dynamics. Model
construction tests this hypothesis and it must be adjusted if evidence from the model or from the real
system refutes it (Lane, 2000).
+ .. 10
. 10
Birth
Population
Positive or self-reinforcing loop
+
- .. 10 0
. 10 0
Deaths
Population
Negative or self-balancing loop
+
Figure 3.2: Positive and negative feedback loops
Source: Own construction
3.6.3.3
Model formulation
Model formulation includes developing maps of causal structure (i.e. constructing stock and flow diagrams)
and estimating the parameters of the model (Ford, 1999). Stock and flow diagrams, unlike causal loop
diagrams which illustrate the system structure qualitatively, capture the quantitative relationships between
the variables of the system by adding stock and flow variables. The stocks or levels denoted by rectangles
show accumulations in the complex-whole formed from related parts, while the flow variables (i.e. inflow
and outflow rates) denoted by valves, regulate changes in stocks (i.e. by means of filling or draining the
stocks), see Figure 3.3. The flow rates are given by various factors, for instance, stock levels or exogenous
variables (Jeong et al., 2008) and they can go to other stocks or into infinite sinks and sources, denoted by
clouds. Also incorporated into stock and flow diagrams are auxiliary variables. They either represent
constants or are calculated from other auxiliary variables or from stocks (Ford, 1999). Shadow variables
show variables that relate with other variables in other views of the model.
- 60 -
Stock
Inflow rate
Outflow rate
Auxiliary 2
Auxiliary 3
Auxiliary 1
<Variable x>
Valve (flow regulation)
Source/sink (stock outside
model boundary)
Shadow variable
Figure 3.3: Stock and flow diagram
Source: Own construction
Concerning parameter estimation, numerous techniques can be employed to estimate model parameters,
such as use of actual data if data is available, conducting surveys if data is unavailable, use of expert input,
basing parameter values on modellers’ own observation, use of secondary data from literature sources and
use of lookup tables which are functions that relate a variable and its causes by sketching a graph of the
relationship. The lookup tables can be based on actual data, expert opinion, experiments or artificial data.
3.6.3.4
Model validation
A fundamental element of all models, and especially system dynamics, is model validation. It is a
continuous series of actions of testing and establishing confidence in the model’s usefulness (Forrester &
Senge, 1980; Sterman, Richardson & Davidsen, 1988). Building confidence is a gradual process that runs
throughout the whole course of model building, beginning with model conceptualisation up until
implementation of policy recommendations (Forrester & Senge, 1980; Sterman et al, 1988). Forrester
(1961) further emphasises that model validation ought to be judged with reference to a particular purpose,
that is, detached from purpose, model validity is worthless. This is considered important for system
dynamics models because they are constructed to accomplish a purpose (Holling, 1978; Barlas & Carpenter,
1990).
The purpose of the system dynamics model informs both the conceptual/qualitative-model (i.e. causal loop
diagram) and the quantitative/simulation model. During the conceptual modelling phase, focus is on proper
problem conceptualisation and on causal relationships identification. If the causal relationships conflict
with a known causality or if the problem is misrepresented, then the model outcomes or recommendations
would be misleading. In addition, the system dynamics model would be refuted even though the model
- 61 -
behaviour matches observed behaviour. It is for these reasons that in system dynamics, validation of the
internal structure of the model is priority, followed by behaviour validity. The accuracy of the behaviour of
the model is only meaningful once adequate confidence on model structure was established prior (Barlas,
1989; Barlas, 1994).
Though lack of formal validation tools is regularly the critic of system dynamics methodology (Barlas, 1994),
the literature discloses a number of validation tests (Forrester & Senge 1980; Richardson & Pugh, 1981;
Sterman, 2000). Structural validity concerns establishing validity with regards to the internal structure of
the model. These tests include comparing model structure versus knowledge of the real system or versus
general knowledge about the system as evidenced by literature (Barlas, 1994). Five direct structure
validation tests were introduced by Forrester and Senge (1980) for system dynamics, namely structure
verification, dimensional consistency, boundary adequacy, extreme condition and parameter verification
tests. Behaviour validity on the other hand, seeks to establish the extent to which the model’s behaviour
matches the behaviour of the real system (Barlas, 1996). The focus is on patterns. Among the behaviour
validation tools are the behaviour sensitivity test, reference test, modified-behaviour prediction and a face
validity test. Finally, based on models being simple representations of actual-world situations, they can
never be fully validated (Sterman, 2000) and in addition, no particular test can completely verify a model,
but the confidence in a model is improved as the model passes a range of tests (Forrester & Senge, 1980).
3.6.3.5
Policy design and evaluation
As highlighted earlier, the intentions of system dynamics are to understand the behaviour of systems by
detecting the factors driving the behaviour of the system and observing how the system responds to
alterations of the said factors and then making policy recommendations that improve system performance.
Policy design and evaluation aimed at alleviating existing problems in the system is therefore central to the
development of system dynamics models. Policy scenarios are crafted based on the model
outcomes/learning from the model and from anticipations/expectations in the real world (Sterman, 2000).
Key model outcomes are then examined and recommendations made that improve the performance of the
system (Grant, Pederson & Marin, 1997).
Having discussed the origins of system dynamics, its main features and the modelling process, sufficient
background has therefore been provided on system dynamics, the following section discusses the links
between system dynamics and social theory by exploring the social theoretic assumptions underpinning
system dynamics practice. Specific focus is on the system dynamics paradigms of Pruyt (2006).
- 62 -
3.6.4
Social theoretic assumptions underpinning system dynamics practice
A number of researchers have attempted to position systems sciences within a pragmatic framework for
social theories. For instance, Checkland (1981) and Lane (1994) have used the Burrell-Morgan framework of
social sciences to position systems sciences and operational research methodologies, while Lane (2001)
used the Burrell-Morgan Framework to map system dynamics. On the other hand, Pruyt (2006) considers a
different framework founded and extended from the frameworks of Mertens (2002) and Tashakkori and
Teddlie (1998), which he uses to discuss various strands of system dynamics practice. To avoid duplication,
extensive discussions of these frameworks and attempts at placing system dynamics within a social theory
can be found in the said studies. A summary, however, of Pruyt’s extended paradigmatic table is provided
and discussed in this section, mainly in order to inform the ontological, epistemological and methodological
placement of this current study and its links with the schools of economic thought that underpin it (section
3.8), and partly in order to help position the energy literature utilizing system dynamics approach (3.7).
Pruyt’s extended paradigmatic table consists of six paradigms, namely positivist, post-positivist, critical
pluralism, pragmatism, constructivism and transformative-emancipatory-critical (see Table 3.3).
Table 3.3: Pruyt’s extended paradigm table
Transformativeemancipatorycritical
Constructivism
(Pragmatism)
Realism
Relativism
Relativism
Subjective (and
objective)
Non-neutral valueladenness
Subjective
Concerned by
value-ladenness
Objective and
subjective
Unconcerned by
value-ladenness
Primarily
quantitative
Quantitative &
qualitative
Quantitative &
qualitative
Qualitative,
quantitative,
mixed
Reasonably stable
causal relationships
(not necessarily
used)
Primarily deductive
Causality is key to
understanding of
real world
Maybe causal
relationships but not
exactly knowable
Deductive &
Inductive
Do models lead
to real insight &
understanding
Potential to
structural
transformation?
Deductive &
inductive
Closest to goal or
own value system?
Positivist
Post-positivist
Critical
pluralism
Pragmatism
(Naive)
Realism
(Transcendental)
Realism
(Critical)
Realism
Epistemology
Objective
(Probably) objective
Subjective
Axiology
Value-free
Controllable valueladenness
Method(ologie)s
Purely
quantitative
Causality
Knowable
real causes
Logic
Deductive
Ontology
Appropriateness
of model
Refutable but
not refuted
Appropriateness
of strategies
Optimal
strategy
Validated models,
results closest to
the real world
Probably optimal
or most appropriate
strategy
Close to goal or own
value system
Value-bound
Qualitative
Indistinguishable
causes & effects
Deductive &
inductive
Advancing justice,
democracy &
oppressed?
Advancing justice,
democracy and
oppressed?
Inductive
Confidence in
constructed
model
Any strategy
(if agreed to)
Source: Pruyt (2006)
Positivist system dynamics & Post-positivist system dynamics: The ontological position of positivist
(functionalist or objectivist) system dynamics practices is that the modelled systems resemble real-world
systems (i.e. realist) and that of post-positivist is also realist. The epistemological position of positivist
- 63 -
system dynamics practices is that the causal loop and stock and flow diagrams are good objective
representations of reality and that the manner to replicate the dynamics of the real-world systems is
through quantitative system dynamics simulation (i.e. objective). On the other hand, the epistemological
position of post-positivist system dynamics practices is also objective but to a lesser extent contains
subjective elements. Axiologically positivist system dynamicists assume value-free investigations which are
achieved, among other factors, through modelling the physical flows whereas axiologically post-positivist
system dynamicists though they acknowledge that the researcher’s theories and values influence
knowledge and that modelling and interpretation are value-laden, the employment of the scientific method
controls for such influences.
The practice of system dynamics by positivist assumes that real causes may be pinned down and
measurement and interpretation of results is quantitative and objective and if models do not resemble
reality they should be refuted which suggests that model validation is done by comparing simulation results
to real-world facts. On the other hand, the practice of system dynamics by post-positivist assumes lawful,
reasonably stable causal relations which could be probabilistically known and which only change slightly
over time. The method of research is primarily quantitative with qualitative models (causal loop diagrams)
used with the purpose of developing quantitative models. Model theories are tested, validated or refuted
and the logic is principally deductive and the best model is one that mostly resembles the actual-world
system. Typical representations of positivist system dynamics practice are neoclassical economics
modelling, marginal practices, optimization, forecasting and policy engineering while post-positivist system
dynamics practice is to a lesser extent represented in contemporary system dynamics.
Critical pluralist system dynamics: The ontological view of this system dynamics practice is realist in that
the actual world exists. The epistemology is, however, subjective in that the actual world is accessible
sorely through subjective mental models. The axiology is value-laden in terms of research choice,
methodologies, assumptions, etc. The research method is both quantitative and qualitative and modelling
is a repetitive series of actions of building, simulation and interpretation and is therefore inductive and
deductive. Causality is fundamental to understanding the real world as it generates model behaviour. The
models are constructed in cooperation with stakeholders, making them ideographic. The models are
centred on generating understanding between the underlying structures and ensuing dynamics and are
deemed proper if they are helpful in altering mental models and actual-world structures.
Examples of critical pluralist system dynamics practice include mainstream system dynamics, interactive
system dynamics focused on increasing understanding, and broad-system dynamics practice. The specific
- 64 -
ontological and epistemological positions of mainstream system dynamics seem to be indeed (moderately)
realist and (moderately) subjective (Pruyt, 2006). Mainstream system dynamicists often start with
qualitative system dynamics, then turn to quantitative and then qualitative. The modelling process is
therefore qualitative-quantitative-qualitative in that it begins with qualitative information and diagramming
(e.g. during problem definition, system conceptualization) then quantitative simulation models are
developed and used (i.e. model building/formulation, simulation) but the results of simulations and
analyses are interpreted and communicated qualitatively (Pruyt, 2013). Most system dynamics modelling is
also highly interactive/participative (a high degree of participation of stakeholders and decision makers is
desirable and often necessary) which allows for the: exchange of knowledge and information on existing
systems and desired systems; gradual development of understanding, insight, confidence and commitment,
and enables the address of factors omitted from the actual models (Forrester 1971; Lane 2000).
Pragmatist system dynamics: The ontological and epistemological positions for this practice in the
simulation stage are primarily realist and objective respectively, while often nominalist and subjective in
the modelling and explanation stages. The axiology is one of unconcern by value-ladenness of the research
choice, theory used, modelling, and interpretation. The methodology is ideographic and the logic is
inductive and deductive (i.e. from assumptions and perceptions the model is induced with the simulation
results being deduced from simulation). Concerning the issue of causality, pragmatist system dynamicists
assume that actual causality in social-economic systems cannot be exactly known as institutions, cultures,
and societies evolve, altering existing causality. In addition they assume the model that is closest to reality
cannot be known. The focus of pragmatist system dynamicists is not on understanding structural causality
that generates observable behavior, but on models that correspond to values or work towards reaching a
goal. Measurement is both qualitative and quantitative.
Constructivist system dynamics: The ontological position of this practice is relativist in that in reality
systems are nonexistent but only concepts/holons associated with the knower can be described. The
epistemological position is subjective in that models describe concepts from a specific viewpoint. The
axiology is certainly value bound while a voluntarist human nature is supposed. The methodology is mainly
qualitative and it is ideographic. This practice assumes subjective causal interpretations of real world. The
models are likely used for learning about other points of view; for gaining insight into potential evolutions;
for building shared interpretation; and for finding compromises amongst different views. Examples of this
practice include holon dynamics and modelling as radical learning. With regards to transformativeemancipatory-critical system dynamics: The ontological position for this system dynamics practice is
relativist while the epistemological position is subjective. Its aim is to assist the oppressed and
- 65 -
disadvantaged through using system dynamics tools to promote democracy and Justice. Examples of this
practice include the strands of modelling as radical learning to enhance group debates and to deal with
power, oppression and beliefs. An attempt at placing the system dynamics practices of the energy-related
literature on Pruyt’s extended paradigmatic table is attempted in the following section.
3.7
Ontology, epistemology and methodologies of energy-related system dynamics practices
Having reviewed the literature on social theoretic beliefs underlying system dynamics practice and
particularly the system dynamics paradigms of Pruyt, an attempt at placing the system dynamics practices
of the energy literature on Pruyt’s extended paradigmatic table is attempted in this section. While it is not
that easy to position the bulk of the energy literature utilizing system dynamics to study an assortment of
energy issues wholly on Pruyt’s extended paradigmatic table categorised in terms of ontology,
epistemology, axiology, methodologies, causality, logic and appropriateness of model and strategy, due to
that no detailed information is given in the articles to enable full placement, it is very clear from the model
built (reviewed fully in chapter 4) that the ontological/epistemology positions and methodologies of most
of the present-day energy literature corresponds with the critical pluralist paradigm and to a lesser extent
the post-positivists paradigm.
Examples of these type of modelling in the energy literature are the models of: Pruyt (2007) who built a
system dynamics model and used it to investigate the transition of EU-25 electricity generation system,
towards a more sustainable energy system characterised by lower CO2 emissions; Bassi, Shilling and Herren
(2007) who constructed a system dynamics model designed to analyse the main energy challenges and
choices faced by the United States of America in the wider context of their relation to society, environment
and the economy, and with associations with rest of the world; Saysel and Hekimoglu (2010) who
developed a dynamic simulation model of the electric power industry in Turkey and used it to study options
for CO2 mitigation; Ford, Vogstad and Hilary (2007) and Vogstad (2005) who modeled green electricity
certificates; Jeong et al. (2008) who designed a system dynamics model for power generation costs
comparison in a liquefied natural gas combined cycle and coal-based power plants while also taking into
account control costs of CO2 and NO2; Vogstad, Botterud, Maribu and Jensen (2002) who built a system
dynamics model for the Nordic electricity market and used it to investigate the short-term and long-term
energy planning trade-offs. The aim was to find efficient policies to aid the transition from fossil-fueled
based power supply to renewables; and Musango, Brent, Amigun, Pretorius, & Müller (2012) who used a
system dynamics approach to develop a model for assessing the sustainability of bioenergy and used it to
assess the effects of the development of a biodiesel industry on a number of sustainability indicators in the
Eastern Cape province of South Africa.
- 66 -
Other examples of these system dynamics practices include those that have combined system dynamics
models with other methods, for example, Pereira & Saraiva (2010, 2011) who combined system dynamics
with generic algorithms, Sanchez, Barquin, Centeno and Lopez-Pea (2008) who combined system dynamics
with game theoretical approaches, Tan, Anderson, Dyer and Parker (2010) who combined system dynamics
with decision trees, Pasaoglu (2006) who integrated system dynamics with analytical hierarchy processes
and Dyner, Ochoa and Franco (2011) who built a system dynamics model linked to an iterative algorithm.
System dynamics approach has thus been used widely for modelling various energy issues. The ontological,
epistemological and methodological placement of this current study and its links with the economics
schools underlying this study are discussed in the following section.
3.8
The ontological, epistemological and methodological placement of this current study and
its links with the economics schools underlying this study
The placement of the system dynamics research conducted in this study based on Pruyt’s extended
paradigmatic table is undertaken. The ontological and epistemological positions for the system dynamics
that is taken in this study is realism and (moderately) objective with subjective elements. The view taken is
thus that an external real-world exists (or the modeled system resembles a real-world system) and the
causal loop and stock-and-flow diagrams are interesting formulations to structure, describe and understand
real-world issues such as the social cost assessment issue investigated in this study. Though no primary
valuation of externalities is conducted in this study, the manner of knowing and construction of this
knowledge (externality costs), can only be grasped mainly through subjective views of the participants,
hence the subjective stance. The methodology is mainly quantitative with qualitative models (causal loop
diagrams) used for developing quantitative models. The model developed in this current study was also
validated in keeping with mainstream system dynamics and due to concerns of value-ladeness. Based on
Pruyt’s (2006) system dynamics paradigms, the system dynamics investigation conducted in this study can
therefore be categorized within the critical pluralist and post-positivist paradigms.
In section 3.5 it was determined that neoclassical and environmental economics provided the theoretical
base for this study. The ontology, epistemology and methodology of both neoclassical and environmental
economics were discussed to be realist, objective and quantitative, respectively, and hence to fall within
the positivist research paradigm of Guba and Lincoln’s classification. The proposed modelling approach
(system dynamics) thus shares many elements that are consistent with the two economic disciplines that
underpin this study, for instance, ontology and epistemology elements and the usage of quantitative
techniques. In addition though, the modelling approach proposed in this study offers more features than
- 67 -
the two economic disciplines, such as non-linear structures, dynamic structures, experimental approach
(Forrester, 1975a; Robertshaw, Mecca & Rerick, 1978), transdisciplinarity methods, disequilibrium
approach and case study approach instead of using abstractions to develop models (Beed & Beed, 2006).
Other additional attributes include that it offers a complex unitary approach with the ability to deal with
large number of elements and many interactions between elements, a problem-orientated approach,
empirical solutions (Forrester, 1975a; Flood & Jackson, 1991) and confidence based on model structure
over coefficient accuracy, focus on closed loop information feedback structures and focus not on
predictions but on understanding the structure of the system and our assumptions about it (Forrester,
1961).
3.9
Summary
A historical review of the schools of economic thought and system dynamics was provided in this chapter,
with the ultimate aims of determining the schools of economic thought that underpin this study and its
links with system dynamics. In pursuit of these aims, the history of economic thought was reviewed by
discussing developments in the economics research field since the 16th century and through using Guba
and Lincoln’s social science research paradigm framework to evaluate the developments into distinct social
science research paradigms. From this review it became clear that neoclassical and environmental
economics provided the theoretical basis for this study. The ontology, epistemology and methodology of
both neoclassical and environmental economics were discussed to be realist, objective and quantitative,
respectively, and hence to fall within the positivist research paradigm of Guba and Lincoln’s classification.
A review of system dynamics origins, main features, modelling process, its links with social theories and its
links with the schools of economic thought that underpin this study was then conducted. From this review
it became clear that while the proposed modelling approach (system dynamics) shares many elements that
are consistent with neoclassical and environmental economics (for instance, through sharing the same
ontological position, epistemological position (to a certain extent) and the use of quantitative techniques),
the proposed modelling approach also offers more features such as non-linear structures, dynamic
structures, experimental approach, transdisciplinarity methods2, disequilibrium approach, case study
approach instead of using abstractions to develop models, problem-orientated approach, empirical
solutions, complex unitary approach with the ability to deal with large number of elements and many
interactions between elements and confidence based on model structure over coefficient accuracy, focus
on closed loop information feedback structures and focus not on predictions but on understanding the
structure of the system and our assumptions about it.
2
Transdisciplinary methods are none discipline-specific approaches used in transdisciplinary research.
- 68 -
- 69 -
CHAPTER 4:
A REVIEW OF POWER GENERATION ASSESSMENT TOOLS
AND THEIR APPLICATION
4.1
Introduction
A number of approaches have been used by various researchers to evaluate power generation technologies
contingent on the goals plus scopes of the applications. The application of the tools is commonly performed
from a financial or environmental viewpoint. In this chapter, an overview of the various tools used by
various researchers to estimate the private and/or externality costs of power generation technologies is
conducted, followed by a review of the application of the assessment tools in the power sector with a
special focus on coal-based power generation applications.
4.2
Power generation technologies assessment tools
The literature discloses various tools and methods that have been used by researchers to evaluate power
generation technologies, which can at least be categorised into three broad categories of methods, namely
financial analysis methods, impact analysis methods and systems analysis methods (see Table 4.1). The
grouping shown is, however, not a precise classification due to that some of the tools fit into more than
one category.
Table 4.1: Power generation technology assessment tools and methods
Financial analysis







Life cycle cost analysis
Levelised cost of energy
Simple payback period
Discounted payback period
Internal rate of return
Modified internal rate of return
Net present value
Impact analysis
Systems analysis
 Damage cost approach
 Abatement cost approach
 Benefit transfer technique
o Simple unit transfer
o Unit transfer with income
adjustment
o Benefit function transfer
o Meta-analysis
 System dynamics
 System optimization techniques
o Linear programming
o Integer programming
o Quadratic programming
o Dynamic programming
o Nonlinear programming
o Stochastic programming
 Life cycle assessment (LCA)
 Hybrid LCA
o Tiered hybrid LCA
o Input-output hybrid LCA
o Integrated hybrid LCA
 Energy systems analysis models
o PERSUE model
o Balmorel model
o EnergyPLAN model
o MARKAL model
o HOMER model
o RETScreen software
o MESSAGE model
o POLES model
 Environmental impact assessment
o Ecological impact assessment
o Health impact assessment
o Social impact assessment
Source: Adapted from Short et al. (1995), Berglund and Soderholm, 2006; Tran (2007)
- 70 -
4.2.1
Financial analysis methods
Financial analysis is essential to corporate decision makers as it entails comparing revenue (cash inflows)
and expenses (cash outflows, e.g. capital/investment cost, maintenance and operation costs) of power
generation project alternatives and calculating the corresponding financial return ratios. The financial
feasibility of an energy generation project may be assessed using different kinds of metrics such as life cycle
cost analysis, levelised cost of energy, cost effectiveness analysis, return on investment, net present value
and breakeven point analysis.
Life cycle cost (LCC) analysis is a method for assessing the total costs of constructing/developing,
operating/owning and disposing/retiring a product/facility/project. The analysis therefore evaluates costs
over a power system’s /product’s lifetime and is particularly useful for comparing project alternatives that
fulfil similar performance requirements but vary in terms of investment/initial and operational costs (Fuller
& Petersen, 1996). LCC estimating techniques may generally be categorised into parametric, analogous and
detailed models. Cost estimation using a parametric model involves predicting a process’s /product’s cost
by means of regression analysis founded on technical information and historical cost (Dean, 1995; Asiedu &
Gu, 2010). Such models correlate technical information and costs with parameters such as design
complexity, weight, and performance, which describe the system. They are top-down estimations and are
deemed not very accurate, especially for the approximation of product costs that use new technologies
(Asiedu & Gu, 2010).
Analogous models on the other hand, estimate cost by analogy/comparison through identification of a
comparable product and correcting its cost analogously to the new target product (Shields & Young, 1991).
The models’ chief shortcoming is that they are highly judgmental (Asiedu & Gu, 2010). Detailed models are
bottom-up estimation techniques that estimate direct costs of a product/activity through the use of
estimates of material quantities, material prices, labour and labour rates, coupled with an allocation rate
for overheads (Shields & Young, 1991; Greves & Schreiber 1993). They are data intensive as they require
detailed knowledge of the product and processes but this drawback is counteracted in that bottom-up
approaches can achieve the most accurate cost estimates (Asiedu & Gu, 2010). Finally, LCC analyses have
been criticised for not considering environmental costs, revenues and returns.
Levelised cost of energy (LCOE) is the cost of electricity per kWh (kilowatt hour) that over the lifespan of
the power generating plant fully recovers capital, fuel, operating, financial and decommissioning costs
(Davis & Owens, 2003; Denholma & Margolis, 2007; Sovacool, 2008; Paltsev et al., 2011). It is thus the cost
per kWh over the lifespan of the investment technology that equals the total life cycle cost when
- 71 -
discounted back to the base year (Short, Packey & Holt, 1995) and is hence quite synonymous with LCC
analysis. Like LCC analysis the LCOE can be used to evaluate the cost-effectiveness of various power
generation technologies (Park et al., 2011), but is conversely said to provide the fairest/best comparison
between energy supply technologies since it takes into account not only the lifetime cost but also the
lifetime energy production associated with an energy system (Bandyopadhyay, Groo, Hartley, LeBrun, &
Moazed, 2008; Darling, You, Veselka, & Velosa, 2011)
There are multiple calculation methods for the LCOE depending on the level of financial detail. The most
common approaches are the simplified LCOE (sLCOE) and the Financial Model Approach (FMA). The sLCOE
is the minimum price at which energy must be sold over the life of the energy development to break even,
i.e. the LCOE is calculated such that the project’s Net Present Value (NPV) is zero (Kornbluth, Greenwood,
Jordan, McCaffrey, & Erickson, 2012; Darling et al., 2011). The FMA, on the other hand, captures more
complex financial considerations such as revenue requirements, taxes, subsidies and depreciation and
calculates the required revenue to achieve a certain internal rate of return (Black and Veatch, 2011). Both
methods (i.e. sLCOE and FMA) can be computed in real or nominal terms - that is as real LCOE or nominal
LCOE. A nominal LCOE accounts for the effect of inflation over the lifetime of the energy project whereas
real LCOE excludes inflation associated with fuel, operation and maintenance costs (Wang, Kurdgelashvili,
Byrne, & Barnett, 2011). The choice between real and nominal LCOE is contingent on the purpose of the
assessment, with the former mainly preferred by policy makers and the latter by project developers.
Despite the form of LCOE selected, the most cost-effective energy technology will not change as long as all
energy supply technologies are evaluated using the same method (Short et al., 1995). Finally on the
downside, researchers employing LCOE have been criticised for not considering correct plant lifetimes, real
load factors of the technologies, the full costs of the plant, for instance decommissioning and
environmental costs, and lastly, poor treatment of the uncertainties associated with input parameters
(Darling et al., 2011).
The simple payback period, internal rate of return (IRR), discounted payback period, Modified IRR (MIRR)
and NPV are other financial indicators that are employed by building economists, cost engineers,
operations researchers and others to evaluate (energy) projects. They are commonly computed as
secondary/supplementary measures of economic evaluation and are therefore discussed briefly.
The simple payback period refers to the number of years that are necessary to recover the development’s
cost of an investment. This financial indicator provides a quick and simple way of comparing alternative
energy projects but it does not consider the time value of money and returns after payback. The
- 72 -
discounted payback period is an upgrade of the simple payback period in that it considers the time value of
money but it continues to disregard the returns after payback (Short et al., 1995). The internal rate of
return (IRR) refers to the discount rate at which the cash inflow of a development equals its cash outflow.
That is, a discount rate which makes the NPV of all cash inflows of an investment project zero. This financial
indicator can be used to compare investment projects. Generally, the higher the IRR of a project, the more
desirable the project. The fallacy with the IRR is that it assumes that the cash flows or interim proceeds
from the projects are reinvested at the internal rate of return. In this manner it may overstate profitability
(Short et al., 1995). The modified internal rate of return (MIRR), also called adjusted IRR, is a form of IRR
that assumes that all proceeds from the investment are reinvested at a company's capital cost. The MIRR
reflects the profitability of an investment project more accurately. It can be used to evaluate projects
having different lifespans or scales since it accounts for varying investment rates. This is an exceptional
advantage of the MIRR, especially since project ranking criteria such as IRR and NPV may produce
conflicting results under such circumstances due to dissimilar reinvestment assumptions. The net present
value (NPV) method assumes reinvestment at the discount rate while, as stated earlier, the IRR method
assumes reinvestment at the IRR (Short et al., 1995).
In summary, the review of financial measures in this section discloses that different financial measures are
suitable for different computations. Generally though, cost-effective energy projects are those with lowest
LCOE, LCC, simple payback period and discounted payback period, plus those with high IRR, MIRR and NPV.
A combination of these methods is usually used in practice when comparing investments, though different
measures may provide dissimilar outcomes.
4.2.2
Impact analysis methods
Explained earlier in the initial chapters of this thesis was that all power generation technologies are
accompanied by undesirable side effects at some point in their fuel cycles. They inflict costs on third parties
by means of negative influences on human health, climate change, crops, biodiversity, structures, etc.
(ATSE, 2009). In this section the various techniques that researchers have used to quantify and value
externalities are discussed. The first section discusses how externalities are valued in theory, while the
valuation of externalities in practice is covered in the second section. Section three provides an in-depth
discussion of the controversies surrounding the placement of monetary values on human life. VSL
methodologies are used as an example during this discussion.
4.2.2.1
Valuation of externalities in theory
- 73 -
Externalities have a real, direct effect on the utility of consumers because an economic activity that
produces an externality (for example, reduce environmental quality), directly causes a change in the utility
of individuals. However, externalities are generally not revealed in market transactions and thus not in
market prices. So in theory, to achieve a socially optimal level of production in the presence of externalities,
externalities need to be valued through monetising individuals’ preferences (Sundqvist, 2000). For this
reason, the valuation of externalities is theoretically based on welfare economics, which identifies the
economic value of a resource as a function of individuals’ preferences (i.e. is based on measuring peoples’
preferences) (Kim, 2007). The valuation process is anthropocentric and preference revelation involves
investigating how much people are willing to pay or accept as reparation for the environmental
improvement taking place or not, respectively (Sundqvist, 2000; Kim, 2007). By so doing, economists obtain
direct welfare measures associated with specific effects. The compensation principle is therefore the
theoretical model of valuing externalities (Kim, 2007).
4.2.2.2
Valuation of externalities in practice
In practice, the literature reveals two broad classes of methods employed by scholars to value externalities,
namely the damage cost approach and the abatement cost approach. The abatement cost approach
employs the cost of mitigating or controlling damage as a proxy for the damage caused by an externality.
This approach involves analysing existing/proposed regulation with the aim of identifying the marginal cost
reduction strategy as required in legislation, which is then taken as an estimate of the value that regulators
(and society) implicitly place on specific impacts.
The abatement cost foundation in economic theory is, however, dubious because it depends on the strong
belief that regulators make optimal decisions (an assumption that regulators know the damage and
abatement costs when designing regulations) (Venema & Barg, 2003; Thopil & Pouris, 2010). However, as
the European Commission (1999b) explains, regulators are not aware of these costs and the manner in
which they make policy decisions is not reflective of that they set abatement costs to be equivalent to
social damages. Secondly, the strong assumption renders the approach irrational since it assumes precisely
what it should be trying to evaluate (Office of Technology Assessment, 1994). Thirdly, an additional
shortcoming of the abatement cost approach stems from that, say, at a given point in time, regulations
were set such that they produce an optimal level of pollution, such a state cannot last for long since
society’s preferences evolve with time because of changes in information and values (Joskow, 1992).
Therefore past preferences may not be reflective of actual effects and their worth to society today. Lastly,
according to Joskow (1992), the condition under which abatement cost will bear a resemblance to damage
cost is solely when the pollution reduction strategy used in the abatement cost approach as a basis for
- 74 -
externality cost estimation is based on the least cost of controlling emissions. Otherwise, the externality
cost estimates will overstate the true damages.
The abatement cost approach is easier to implement as it is not as data intensive as the damage cost
approach, but the drawback is that it does not offer an equivalent level of precision (Owen, 2004), as there
is no relation to actual damage (Faaij et al., 1998). The abatement cost approach has been employed in
various electricity externality studies, for example, the study conducted by Bernow, Biewald and Marron,
(1991) and Roth and Ambs (2004). The literature in South Africa on electric sector externality studies,
however, does not reflect much use of this approach.
In contrast to the abatement cost approach which estimates the cost of actions/technology that would limit
or control the externality, the damage cost approach approximates the real externality impacts and
allocates a cost to the effects by means of valuation techniques. The approach can be performed in a
bottom-up or top-down fashion. The top-down approach estimates externality cost of environmental
burdens based on national/regional level studies approximating quantities of (prevailing) pollutants and
damage caused by pollutants (Sundqvist, 2000). The two main critics of the top-down approach are that it
does not allow for the consideration of site specificity of impacts and the various fuel cycle stages. The
approach has also been criticised for being derivative due to its dependence on previous estimates (Clarke,
1996). On the positive side, the approach is less data intensive than the bottom-up approach. Various
electric sector externality studies have used the top-down approach, for example, the study by Hohmeyer
(1988), Hohmeyer (1992), Friedrich and Voss (1993) and Pearce (1995). Most of the studies, however, that
employed the top-down approach were conducted in the 80’s and 90’s. Recent studies made use of the
bottom-up approach or used the benefit transfer technique that adjusts monetary estimates from earlier
studies and transfers them to new settings.
The bottom-up approach, also identified as the impact pathway approach, traces contaminants and other
burdens from their original source, quantifies effects and monetises effects by means of valuation
techniques. For example, for pollutants the assessment originates with the determination of emissions
loads from a distinct source and the dispersion of these pollutants. This is followed by the determination of
marginal damages resulting from emissions using dose-response functions and finally marginal externality
costs are obtained as a product of the marginal damages multiplied by their estimated monetary values.
In this approach externalities are therefore quantified in a logical manner. The approach is more in line with
economic theory. However, since the approach is location specific, in principle the obtained costs are not
- 75 -
transferable (Kim, 2007). As evidenced by electric sector externalities studies, the bottom-up approach is
the most favoured method. On the downside, the bottom-up approach is data intensive in relation to other
methods and has been criticised for focusing on impact pathways that are easier to establish, so the
approach omits some externality impacts due to lack of data or lack of monetisation ability (Owen, 2004).
Most of the bottom-up studies were carried out in developed economies. Examples include the electric
sector externalities studies undertaken by Oak Ridge National Labouratory and Resources for the Future
(ORNL & RfF) (1994), Rowe et al. (1995), European Commission (1999b) and European Commission (2005).
While most recent studies used the bottom-up approach, some studies in both developing and developed
nations made use of the benefit transfer technique. The valuation methods used for monetising externality
impacts are discussed next.
Two types of valuation methods can be used to monetise externality impacts, namely the direct valuation
method and the indirect valuation method. Linked with these two methods are a number of costing
techniques. The direct valuation method aims to derive values using direct methods that simulate markets.
They are direct in that they are directly designed to elicit willingness to pay or willingness to accept. The
stated preference method, also known as the contingent ranking method and the Contingent Valuation
Method (CVM), are well-known direct valuation methods. The CVM elicits preferences through asking
individuals direct questions through questionnaires (i.e. individuals are asked the amount they are willing
to accept or pay for the damage imposed on them (compensation) or for the avoidance of a damage)
(Sundqvist, 2002; Icyk, 2006). The stated preference method, on the other hand, is based on questionnaires
that are designed to elicit ranking of preferences.
The indirect valuation method, in contrast to the direct valuation method is based on actual behaviour of
individuals. The various techniques aim to derive value from market observations. The damage is valued
indirectly using a relationship between a marketed good and the externality. Example of indirect valuation
methods include change in productivity technique, replacement cost technique, hedonic pricing method
and change in income technique. The change in productivity technique is suited for measuring externality
impacts that directly affect the production process, for example, those that affect the quantity and quality
of output. The observable change in price is then used as a measure of the externality cost. The change in
income technique is most suited to measure externality costs that are health related, for example,
externalities that cause health effects are measured through individual income changes.
- 76 -
In the replacement cost technique, quantified expenditure necessary to replace a resource/good/service
that has been affected by an externality, is used as a measure of the externality cost. The hedonic pricing
method is suited to measure externalities that affect characteristics of products, for instance, for measuring
noise pollution from wind turbines, house prices in both a quiet and a noisy area can be used to infer
persons’ willingness to pay to avoid the noise, thereby obtaining an estimate for noise pollution (Icyk,
2006).
As evidenced by the damage cost approach, especially the bottom-up valuation approach, and by the direct
and indirect methods of monetising externality impacts above, the study of externalities (identifying,
quantifying and monetising externalities) is a time-consuming exercise. In addition, conducting primary
valuation assessments in certain contexts, especially in developing countries, might prove difficult as
respondents might lack the knowledge of fully understanding what is being valued. Researchers in more
recent electric sector externality studies did not necessarily conduct primary valuation studies but rather
used the benefit transfer and dose-response techniques discussed below.
The benefit transfer technique does not derive monetary estimates for externalities but rather adjusts and
transfers monetary approximations of externalities from earlier studies to new settings. Since extensive
work on the valuation of electricity sector impacts (on the environment, on humans, etc.) has been
conducted in developed countries (especially in Europe and the US), most recent works have adjusted and
transferred the monetary estimates from these studies to present contents. Hence some of the electric
sector studies in South Africa and elsewhere have adjusted and used these estimates.
The Benefit transfer technique, also called the value transfer technique, is another technique for assessing
externality effects of energy technologies, adopted when there are not enough resources and time to
perform primary valuation investigations (Navrud, 2004; Navrud & Ready, 2007; New Energy Externalities
Developments for Sustainability (NEEDS), 2009). The unit value transfer and function transfer approaches
are two key methods to benefit transfer. With the unit value transfer approach, the unit value or damage
cost at the study site is taken as a proxy for the new site and is either - (i) taken simply as it is (i.e. simple
unit value transfer); or (ii) adjusted for income differences between the study site and new sites using GDP
per capita and/or adjusted for differences in the costs of living using purchase power parity indices (i.e. unit
transfer with income adjustment). Though the simple unit value transfer method provides the easiest
means of transferring estimates between sites, people between the two sites may be dissimilar in such
factors as education, income and other socio-economic characteristics, which might affect the values
yielded. The approach therefore ought not to be employed to transfer estimates among nations with
- 77 -
diverse income levels and costs of living. In such cases the unit transfer with income adjustment works best
as it makes adjustments for such differences, though it also does not correct factors such as differences in
individual preferences, institutional and cultural conditions and initial environmental quality between
countries/states/provinces (NEEDS, 2009).
With the benefit function approach either a benefit function - (i) at the study site is estimated and
transported to the new site (i.e. benefit function transfer); or (ii) it is approximated from several research
sites by means of meta-analysis (i.e. meta-analysis). The benefit function can be written as:
, where
household
and
denotes household
characteristics at site ,
are parameters while
willingness to pay at the site ,
is a set of
denotes the environmental good set of attributes at site ,
,
is the error term. The bid/WTP-values can be estimated using stated
preference or revealed preference methods. The benefit function transfer approach is then implemented
by finding a study in the body of written works with estimates for the parameters including the constant. At
the new site, the researcher collects data on household and environmental good characteristics and inserts
them in the benefit function and then calculates households’ WTP. Though transferring the whole benefit
function is theoretically more attractive compared to transferring just unit values, for the reason that
extensive information is captured by such a transfer, the main drawback of this approach is omission of
pertinent variables in the WTP function (NEEDS, 2009).
Meta-analysis on the other hand, combines results from several original valuation studies into one common
benefit function. In the regression analysis the outcome of each of the studies is taken as a distinct
observation. But in the event of multiple outcomes from one study, various meta-regression specifications
are specified. Such equations clarifying differences in unit values may subsequently be used in conjunction
with data on
and
(explanatory variables) collected at the new site to construct an adjusted unit value.
Meta-analysis thus permits a broader evaluation of the environmental good characteristics, population
characteristics and modelling assumptions (NEEDS, 2009).
The dose-response technique does not derive individual preferences but uses the links between pollution
and impacts and values the final impact at a market shadow price. Intensive research has been conducted
in Europe and in North America on dose-response functions that connect human health to the quality of
the environment. So researchers elsewhere have modified the dose–response relationships from previous
studies for country-specific socio-economic conditions and estimated the health impacts using the
opportunity costs of the health effects. Examples of electric sector externality studies utilising this
- 78 -
technique in South Africa include the study conducted by Van Horen (1997) and that conducted by
Spalding-Fecher and Matibe (2003).
Life cycle assessment (LCA) is another method for assessing the environmental influences of a product or
project (e.g. wind turbine). It assesses such throughout the product/project life cycle (SANS 14040, 2006),
that is from the procurement of raw materials, processing, manufacturing, use and finally disposal. This
analytic tool systematically defines and measures over the life cycle, all flows (e.g. materials, energy and
environmental flows) that go into the investigated system from nature and those that flows out from the
system to nature (Ampofo-Anti, 2008; Varun & Ravi, 2009). A LCA study consists of four components,
namely the goal and scope (which describes the aim of the assessment, the system and its borders and the
functional unit), life-cycle inventory stage (which involves the collection of all the materials, resources and
environmental inflows and outflows), life-cycle impact assessment (which conventionally involves
classifying the inventory flows into specific impact classes (e.g. global warming) then normalization and
weighting of the impacts) and lastly, interpretation of the study results (Tan & Culaba, 2003; Scientific
Applications International Corporation, 2006).
As an environmental assessment tool, LCA is favoured because it systematically captures the environmental
performance of products over their whole life cycle while embracing all processes, material, energy and
environmental flows (Varun & Ravi, 2009). This strength is, however, conditional on the comprehensiveness
of the LCA. LCA can also be used as a technique for detecting the transfer of environmental impacts
between life-cycle stages or between environmental media, thereby serving as an instrument for detecting
possibilities for improvements with the intention of reducing negative impacts on the environment, human
health and resource depletion (Sherwani, Usmani & Varun, 2010). This yields vital environmental trade-offs
information that can be beneficial to decision makers and managers. On the downside, LCA will not
determine the most cost-effective products or processes owing to cost data being missing (Kannan, Leong,
Osman & Ho, 2007). Lastly, conducting an LCA may be time consuming and resource demanding,
contingent on the comprehensiveness of the LCA (Scientific Applications International Corporation, 2006).
Hybrid LCA refers to any method that combines LCA and input-output analysis (IOA). It is an environmentaleconomic tool that has been employed by researchers to study environmental and economic issues.
Basically IOA is a quantitative method that describes the monetary or physical flows amongst various
sectors in the national economy. The economic system’s sectoral structure is described by I-O tables while
sectoral changes within the system are analysed by I-O models. The tool takes a top-down linear approach
to describe industrial structure. It can analyse the entire world economy, national economy, regional area
- 79 -
or even an enterprise. The basic unit of analysis is either a sector, industry or a product group (Leontief,
1986). Economists began applying IOA to environmental issues and problems since the late 1960’s, through
extending the I-O accounting framework with environmental data.
Three kinds of hybrid LCAs exist namely, input-output hybrid LCA, integrated hybrid LCA and tiered hybrid
LCA (Huppes & Suh, 2002). Tiered hybrid LCA is a tool that develops and analyses separately the I-O system
and process-based system (i.e. LCA). Direct, downstream and lower-order upstream requirements are
covered in a thorough process LCA, while IOA is used to examine higher-order requirements (Lenzen, 2001).
The tool combines the advantages of site specificity and completeness. Integrated hybrid LCA develops
independently and merges systematically process-based system and IO-based system (Huppes & Suh,
2002). The two systems become intricately looped. Total production in the I-O system is used to normalise
monetary flows whereas the operation time in the process-specific part of life-cycle inventory technology
matrix is used to normalise physical product flows (Huppes & Suh, 2002). Input-output hybrid LCA (IOALCA) starts off from a conventional IOA by disaggregating part of the I-O table in the event of availability of
comprehensive sector-based monetary data. Furthermore, substitution of sectoral I-O data with detailed
process data or its augmentation with sectoral physical unit data can be conducted. The disaggregation,
substitution and augmentation of the direct requirements matrix with process data may result in
undesirable flow-on effects (Lenzen, 2001). IOA-LCA has generally been regarded as a quick data collection
strategy whose results in comparing products should be interpreted as relative performance indicators
rather than absolute indicators (Joshi, 2000).
The hybrid LCA models are linear models whose results represent economic-environmental impacts
through industrial sectors’ production in line with increased demand. Accordingly, downstream phases such
as use and decommissioning phases are not explicitly incorporated in the results (Lenzen,
2001). Moreover, an industry sector embodies an assortment of industry types – this form of aggregation
therefore ignores the diversity of industries, products and production methods (Weisser, 2007). Some
researchers such as Gronow (2001) find it illogical that price levels affect the estimation of emissions and
materials use. The models are incomplete due to considerations of a limited number of environmental
effects. Many assumptions also go into crafting the impact vectors (i.e. the values for the environmental
impacts and materials consumption) but even so, the I-O data are more unreliable than process LCA data
(Treloar & Love, 2000) owing to uncertainty inherent in original data (source data uncertainty), estimation
uncertainty of capital flow, allocation uncertainty, imports assumption uncertainty and gate-to-grave
truncation error (Lenzen, 2001). Lastly, though the models combine economic and natural systems, the
- 80 -
study of ecological processes interconnections is mostly unsuccessful partly because they are most often
too complex to fit into the rigid I-O framework (Fankhauser & McCoy, 1995).
Environmental Impact Assessment (EIA), also termed environmental auditing, environmental impact
analysis, environmental appraisal or environmental assessment, is a site-specific environmental
management tool for assessing the effects of a planned activity on the environment - social, economic and
biophysical dimensions (Ministry of Environment and Tourism, 1997; Hugo, 2004; Southern African Institute
for Environmental Assessment, 2004). The holistic definition of EIA thus given encompasses social, health
and ecological impact assessments and EIA practitioner teams should therefore encompass expertise from
these fields.
The literature presents various EIA models (Weaver, 2003; Hugo, 2004), but common to the models are the
main phases in the EIA process. These are project pre-feasibility/screening phase (which is the key planning
stage though with minimal information on the design of the proposed development, in this phase a
decision is arrived at on whether or not to subject a project to a full EIA, based on evaluating the project
mainly against simple checklists for the type of activity), scoping phase (which determines the proposed
development’s nature of impacts, extent of impacts, their significance and whether they are direct or
indirect, or reversible or not – this phase is therefore about identifying significant issues and eliminating
insignificant
ones),
preparation
of
the
draft
EIA
statement/report,
draft
EIA
statement
review/environmental management plan (which assesses the quality of the draft EIA report in terms of data
gathered, models used for impact prediction, findings obtained, stakeholders’ views on findings and
ensures strong commitment to the implementation of the environment management plan), monitoring
(which assesses the occurrence of the predicted environmental impacts and checks the effectiveness of
mitigation measures), environmental auditing (which audits the performance of the development in line
with the final EIA statement) and decommission (which ensures rehabilitation of the environment once
operations cease) (Hugo, 2004; Nhamo & Inyang, 2011). The EIA tool thus unveils to decision makers the
likely implications of a development project. It is an imperative guide to decision making because it
integrates into the planning process of development projects, social and biophysical considerations at the
same time that financial and technical factors are considered. It enables the mitigation of adverse
environmental effects and enhancement of positive impacts early in the design stages (Ministry of
Environment and Tourism, 1997.
In summary, the review of impact analysis tools in this section discloses that various tools are suited for
identifying, quantifying and monetising externalities. The advantages and limitations of the reviewed tools
- 81 -
were also discussed. Depending on the aims of the investigations and a number of issues surrounding the
research such as availability/unavailability of previous primary valuation studies, various researchers (e.g.
environmental economists, environmental practitioners, operation researchers, etc.) have therefore
employed various impact analysis tools. The controversies of valuing human life are discussed next.
4.2.2.3
Controversies of valuing human life
Among the highest disputed areas of public policy are those concerning dangers to the safety and health of
humans. At the heart of these disputes is the valuation of human life. The placement of an economic value
on human life is likely to stimulate ethical, philosophical and religious questions. It therefore not unusual
for people to object the placement of a monetary value to human life, arbitrating such an exercise to be
heartless and belittling the value of existence. Even though some considers a human life priceless or are
rather against the expression of the economic value of human life, conflicting demands on limited public
funds means it is impossible to save all life, making tradeoffs necessary on programs that saves lives. The
refutation to attach explicitly a value on human life simply forces implicit valuations that are reflected in
decisions on either to fund or not public projects along with decisions to enforce other regulatory activities
(Brannon, 2005; Landefeld & Seskin, 1982). Assigning a value to life is an effort towards making rational
decisions about these tradeoffs. Controversy, however, still remains on the correct method for generating
approximations for valuing risks to life. Though a standard concept for placing value on human life does not
exist, when observing health risk/reward trade-offs made by people, economist frequently consider the
value of a statistical life (VSL) (Brannon, 2005). In the following paragraphs VSL methodologies and the
criticisms surrounding them are reviewed.
The VSL is the amount of money a person/society is willing to spend to save a human life or rather the
value placed on a change in the risk of death. Owing to lack of a formal market for lives, the VSL is only
measured through indirect methods, for example through surveys or by means of observing the behavior of
humans in risky environments (Brannon, 2005). One manner by which economists estimate the VSL is by
observing a person’s actual choice, through observing the risks humans are voluntarily willing to take and
how much they must be paid for taking them. This is the revealed preferences method (Mankiw, 2012). The
implied value of life is inferred from labor-market choices. Much of the revealed preference research uses a
wage hedonic approach, which observes the changes in wages as job characteristics changes. Economists
estimate the VSL by studying the differences in pay between jobs by controlling for many job characteristics
and hence establish the share of the wage compensating for the risk of injury or death. This number (i.e.
the risk premium) is then multiplied by the inverse of the risk difference and is said to be the VSL (Leeth &
Ruser, 2003; Viscusi, 2003; Viscusi & Aldy, 2003).
- 82 -
Shogren and Stamland (2001) contend that almost all revealed preference assessments are biased
upwards. The average VSL is less than the marginal VSL due to that the pay at a specific job is just adequate
to lure the marginal employee. Another issue with this method is the appropriate time period that should
be used to measure fatality rates. Such is imperative because death rates fluctuate yearly and the choice
can affect the VSL. Another problem is that of which form of death rate to use - workers’ perceived chances
of death or actual death rates. In any case, wage premiums are most probably based on perceived risk than
actual risk, so the two can differ. Yet another issue concerns whether it is okay to calculate the VSL by
simply multiply the risk premium by the inverse of the risk assumed. Researchers such as Krupnick,
Cropper, Alberini, Simon, O’Brien and Goeree, 2002) have found that the risk premium does not necessarily
double if the risk doubles (nonlinearity in valuing risk reduction). Lastly, another issue with this approach is
the wide variation in VSL estimates by various researchers (Brannon, 2005).
Yet another method used by economists to estimate the value placed by people on their lives is by
conducting surveys and asking each person how much money he/she would accept for a marginally higher
chance of dying (or how much money each person is willing to pay for mortality risk reduction). In essence
each person is asked a series of questions up until the person refuse the money for the higher risk or refuse
to pay any amount for risk reduction. This is a contingent valuation method. After the survey the researcher
imputes the implied value of a life by each respondent and multiplies it by the inverse of the extra risk
taken and averages the valuations (Krupnick et al., 2002). Problems of this approach include the subjectivity
of this approach. All questions are imaginary, so why must the respondents responses truly reflect the
trade-offs they are willing to make? Another problem is the “protest” vote in which a respondent insists
he/she cannot be enticed by any amount of money to accept a higher risk. Should such respondent’s value
of life (an outlier) be considered in the final average or not or should researchers use a median or truncate
the sample? There is no consensus in this issue except that such an outlier should not form part of the
estimate. Another critics concerns whether respondents accurately perceive small changes in the
probability of death (Murphy, Allen, Stevens, & Weatherhead, 2004; Brannon, 2005), for example a 4 in 10
000 risk from a 6 in 10,000 risk.
The consumer market behavior method is another approach in which the implicit value of life is inferred
from product choices, for example purchasing safety improvements like purchasing a car with antilock
brakes (this device reduces the occurrence of crashes and death). The implied value of life is then inferred
from the cost of this device. Criticisms of this approach include the difficulty of interpreting VSL values from
different devices, the issue of whether a distinction is made between safety features from other product
- 83 -
attributes (i.e. the difficulty of determining the extent to which the decision of the buyer was influenced by
safety consideration (e.g. purchase decisions might be influenced by proximity of the buyer to a store) and
the issue of whether buyers understand the safety improvements inborn in a purchase (Brannon, 2005).
The meta-analysis method is another method. Based on existing studies of the VSL, a meta-analysis seeks to
find a representative VSL through attempting to control for exogenous elements that may possibly affect
the estimated VSL. Though meta-analyses vary in complexity, a complex one can consider the different risks
in various studies and neutralize such differences in the final VSL. A meta-analysis though an upright tool, it
is difficult to execute well due to that it can solely be performed on analogous studies using the similar
statistical approximation technique. Contingent valuation studies cannot be in a meta-analysis study with
revealed preference studies (Mozrek & Taylor, 2002). Another VSL estimation method is the forensic
economics method often used by economists when estimating the VSL not for regulatory purposes. The
value of a life is placed after death through computing the value today of the future stream of income lost
by the household due to death. The contested assumptions of this approach include the issue of the growth
rate of income and the deceased retirement age, and whether population averages should be used in such
computations? (Brannon, 2005).
Lastly, a common criticism of VSL in regulatory examination is failure to differentiate between saving a life
of a young person versus saving a life of a person close to end of life. The value of a statistical life year
(VSLY) and the quality adjusted life year (QALY) are variants of VSL that make such modifications. Both
approaches make an effort to compute the value of one additional year of a saved life, with the earlier
correcting the value for a saved life by means of discounting future life years saved and the later correcting
for the amount of life saved plus the quality of life saved. Both approaches differ from the VSL in that VSL is
computed from human decisions made either directly in surveys or indirectly in market choices whereas
these two approaches require no human behavior observation. The two approaches appear more
appealing to policy makers than VSL computations (Brannon, 2005).
4.2.3
Systems analysis methods
Systems thinking analysis is an approach that looks at problems as parts of a whole system. It is centred on
the understanding that a system can best be grasped by examining the linkages and interactions between
its elements. Any kind of system, for example natural, engineered, human, scientific or conceptual can be
studied by systems thinking techniques. System thinking techniques such as system dynamics, systems
optimization techniques (e.g. linear, nonlinear and dynamic programming) and energy systems analysis
models are discussed in this section.
- 84 -
System dynamics is an approach to understanding how systems behave over time. It is a causal
mathematical model (Barlas, 1996) centred on understanding the structure of a system under
consideration which results in observable and predictable behaviour (Forrester, 1987). It has gained
recognition because of its emphasis on the structure of a system and because of its flexibility (Anand, Vrat
& Dahiya, 2005). It combines theory and methods required to study the behaviour of systems. It has a
capability to model an extensive diversity of processes and relations in a dynamic fashion (Auerhahn, 2008).
Though system dynamics, in an integrated sense, focuses on a system, the system is disintegrated into
various interrelating subsystems. It therefore facilitates the understanding of complex systems. It can also
capture in an intuitive manner the complex actual-world behaviour of uncertainties which stem from
nonlinear feedback constructions, in this manner providing clearer understandings of the sources of the
effects of strategic action (Sterman, 2000; Johnson, Taylor & Ford, 2006).
The interactions in a system are fed via interactive feedback loops. The feedback structure of the system
that is investigated is what the approach centres on. It is normally represented by means of causal loop
diagrams, which provide a qualitative expression of the interactions in the system (Chi, Reiner & Nuttall,
2009), for example, interactions between burning coal for electricity and the related environmental
releases, e.g. CO2. Stock and flow diagrams are then constructed premised on the causal loop diagram and
coupled with the addition of equations for all variables in the model (Anand et al., 2005). The stocks control
the inertia of the investigated system and can either increase or decrease regularly. The stock in-flows and
out-flows regulate the rate of change in stocks. The flow rates are characterised by such factors as the
stocks level, exogenous variables and can be taken to represent the output/input of policies (Jeong et al.,
2008). Lastly, system dynamics models could be readily constructed and could be used to test various
alternative model specifications (Jeong et al., 2008; Chi et al., 2009).
System optimization techniques are methods that are developed to offer “best values” of system design
and policy elements - values yielding highest ranks of the performance of systems (Loucks, van Beek,
Stedinger, Dijkman & Villars, 2005). The methods therefore select the best element with regard to some
criteria (e.g. cost or benefit) from a set of available alternatives. In general, optimization problems in
general concern minimizing or maximizing a real function by way of selecting, in a systematic manner, input
values
from
an
allowable
set
and
calculating
the
function’s
value.
Mathematical-
programming/optimization/constrained-optimization techniques include, linear, nonlinear, integer,
quadratic, dynamic, stochastic and geometric programming. Linear programming is a mathematical
method that determines the best outcome (e.g. lowest cost or maximum profit) in a mathematical model
- 85 -
with an objective function that is linear and is subject to linear equality and inequality constraints. It has
been applied to solve problems in a number of fields of study, including economics, business and
engineering. It has shown worth in modelling various problems in planning, design and scheduling,
allocation/assignment. Quadratic programming, on the other hand, studies the case where the objective
function has quadratic terms, with linear equality and inequality constraints.
Integer programming, also called integer linear programming, centres on linear programmes characterised
by all/some of the decision variables taking integer values. A mixed integer programming problem is
characterised by some (not all) unknown variables taking integer values. A special case is binary integer
programming where variables take values 0 or 1. Dynamic programming is a mathematical method for
solving complex problems by splitting the problem into smaller sub-problems (optimization strategy). It is
therefore appropriate to problems with overlapping sub-problems. The sub-problems are solved
independently, their solutions stored and when needed later they are simply looked up and incorporated to
reach an overall solution. The approach is useful in the event of repeating sub-problems that grow
exponentially as a function of the input size. The word “dynamic” pertains to the time varying aspect of the
problems. Stochastic programming, on the other hand, models optimization problems where a certain
number of the constraints are contingent on variables that are random. The approach therefore deals with
problems that involve uncertainty. Most problems in the real world almost always contain some unknown
parameters. Stochastic programming has been broadly applied, for instance in investment portfolio
optimization overtime and energy optimization (Wallace & Ziemba, 2005). Nonlinear programming models
the optimization problem where either/both the objective function and constraints are characterised by
nonlinear elements.
Energy systems analysis models concern the use of energy systems models for the study of the
connections between various elements of energy technologies and the consequences of different decisions
on the environment, physical, technical, economic and political systems (Vattenfall Research and
Development Magazine, 2011). Energy technologies are devices that produce or transmit or use energy, for
instance power plants, boilers, automobiles, etc., and are characterized by various attributes, namely
efficiencies, costs, benefits, emissions, etc., (Energy Technology Systems Analysis Program, 2007).
Various energy systems models have been developed and may be categorised into top-down and bottomup energy methods (Hodge, Aydogan-Cremaschi, Blau, Pekny & Reklaitis, 2008). Top-down energy models,
also called macroeconomic models, address the energy-economy feedback. The models describe the
economic system in detail but they typically describe in an aggregated manner the energy system and as a
- 86 -
subdivision of the whole economy. The technical potential of various energy technologies is thus not
represented explicitly. The outcomes of the model are induced primarily by relative price changes. Topdown modellers apply general equilibrium models or models that are demand prompted (Berglund &
Soderholm, 2006; Hodge et al., 2008). Top-down models, however, present the energy system as a blackbox, by paying no attention to the processes and activities because the matrices used can only analyse a
sector as a whole, and as a result differentiation between a range of products or production methods nor
technologies is not possible (Weisser, 2007). Top-down models have been used to study a variety of issues,
comprising of the role of energy or specific energy technologies in a national economic system
(Papatheodorou, 1990; Galinis & van Leeuwen, 2000), and impact of greenhouse gas reduction policies
(Zhang & Baranzini, 2004).
In contrast, bottom-up energy models study the energy system extensively but they do not consider the
economic system in detail as in top-down models (Loschel, 2002; Berglund & Soderholm, 2006). As
emphasized by Grubler et al. (2002), bottom-up models normally aim at finding the minimum-cost mix of
energy technologies serving a specified energy demand. For this reason the models are optimization
models that minimize total discounted system cost (or maximise the income of energy systems) conditional
on technological and environmental constraints (Berglund & Soderholm, 2006; Jensen & Meibom, 2008;
Karlsson & Meibom, 2008; Kiviluoma & Meibom, 2009). Bottom-up models include PERSEU, Balmorel,
MARKAL, HOMER and RETScreen software.
PERSEU model is a family of material and energy flow models that apply a multi-periodic linear
programming method. It aims at minimising expenditures (e.g. investment, fuel supply, variable and fixed
costs) in the context of the whole energy supply system. The energy supply system techno-economic
characteristics are reflected by means of applying equations that take into account technical, political and
ecological restrictions. Environmental/sustainability issues can also be considered through the extension of
the objective function with electricity production externality cost (Fleury, Fichtner & Rentz, 2003). Balmorel
model is a deterministic linear optimization model which centres on optimizing investments in electricity
production, heat production, storage and transmission units while meeting electricity and heat demands in
each area and time period. It focuses on a number of countries. Costs minimised in the energy system
include new units annualised investment costs, fuel costs, existing and new units operation and
maintenance costs plus carbon dioxide quota costs. The model can also be programmed to optimise in
other criteria including minimising carbon dioxide emissions (Munster, 2009). A number of countries have
applied this model to analyse various technologies and market conditions (Ball, Wietschel & Rentz, 2007;
Karlsson & Meibom, 2008).
- 87 -
EnergyPLAN model is an analytical deterministic programming model that has the ability to model different
regulation strategies and to integrate fluctuating energy sources (Østergaard, 2009). It optimises energy
production while meeting heat and electricity demands in each area and time period. The model minimises
marginal production cost and/or fuel consumption and is designed for national analysis. Optimization is
carried out manually through iterations. Among the outputs generated by the model are energy
production, excess electricity production, electricity import/export, fuel consumption and carbon dioxide
emissions (Lund, Duic, Krajacic & Carvalho, 2007; Munster, 2009). A number of energy systems analysis
activities have been analysed using this model, including waste-to-energy technologies and renewable
energy systems (Lund, 2007; Lund & Mathiesen, 2009; Munster, 2009). MARKAL is a linear programming
model that is used for energy systems analysis (Loulou, Goldstein & Noble, 2004) that encompasses both
energy demand and supply (Fishbone& Abilock, 1981; Mallah & Bansal, 2010). The MARKAL family of
models minimizes cost through investment and operating decisions and has been used widely for various
purposes, including comparing electricity generation technologies (Naughten, 2003), development of
carbon mitigation strategies (Jegarl, Baek, Jang & Ryu, 2009), internalisation of power production
externality cost (Rafaj & Kypreos, 2007) and waste management modelling (Cosmi et al. 2000; Salvia,
Cosmi, Macchiato & Mangiamele, 2002). The models are intended for national, regional or global analyses.
HOMER model is an optimization model particularly designed for small, remote power systems but also
makes provision for grid connection. Various energy technologies costs and the availability of energy
resources are considered in this model. It further permits the assessment of economic and technical
viability of the various technologies (National Renewable Energy Labouratory, 2008). RETScreen software is
a decision-support tool designed for evaluating various energy technologies’ financial feasibility, energy
efficiency, cogeneration projects, benefits from clean energy production, and savings through energy
efficiency projects. The model accounts for project costs, financial risk and emission reductions (RESTScreen
International, 2012). Other bottom-up energy systems analysis models include POLES (Kouvaritakis, Soria &
Isoard, 2000) and MESSAGE (e.g. Messner, 1997). The bottom-up models’ shortcomings include that they
are generally static models, with no feedback loops and time delays.
Finally, the review of the systems analysis tools in this section discloses that various models are designed
for different purposes (e.g. modelling energy system, and/or economic system, and/or ecological system),
different technologies (e.g. renewable energy, non-renewable energy or both), different scales of analyses
(e.g. national, regional or global) and different sizes of energy systems.
- 88 -
4.3
A review of power generation studies
In this section electric sector studies are reviewed, with a special focus on those assessing coal-based
private and/or externality costs. Sections 4.3.1 to 4.3.3 review international literature assessing power
generation private costs, externality costs and studies modelling power generation systems, respectively.
Sections 4.3.4 to 4.3.6 review the same for local studies. Several of such studies were conducted in the past
three decades in both developing and developed nations. The studies/models differ in terms of the energyrelated issues they focus on, scales of analyses (e.g. national or regional), power generation technologies
they study, types of externalities they investigate, valuation methods they employ and in terms of the fuel
cycle stage(s) they investigate. For these reasons the reviewed studies highlight these issues.
4.3.1
International studies assessing power generation private costs
The literature discloses several international studies that have assessed the private costs of constructing
and operating coal-fired power generation technologies, for example Booras and Holt (2004), Davison,
(2007), Hoffmann and Szklo (2011) and Cormos (2012). The studies mainly consider three primary air-blown
coal generation technologies, namely subcritical, supercritical and ultra-supercritical pulverised coal plants.
Subcritical units are the traditional Pulverised Combustion (PC) plants with steam temperatures around
538°C (1,000°F) while supercritical and ultra-supercritical units are newer, higher efficiency cycles with
steam temperatures around 565°C (1,050°F) and above (e.g. 593°C (1,100°F)), respectively. Coal based
Integrated Gasification Combined Cycle (IGCC) technology is also another type of coal plant that is
extensively investigated in the literature, mainly developed to decrease the environmental impact of coalbased plants (Booras & Holt, 2004). Table 4.2 presents a summary of the operating performance and
private costs of these coal generation technologies. The cost estimates were adjusted to 2010 US dollars ($)
for comparison purposes.
As explained earlier, the LCOE is the cost of electricity per kWh or per MWh (Megawatt hour) over the
lifespan of the investment technology (Short et al., 1995). Though quite synonymous with LCC analysis, it is
conversely said to provide the fairest/best comparison between energy supply technologies since it takes
into account not only the lifetime costs but also the lifetime electricity production of an energy system
(Bandyopadhyay et al., 2008; Darling et al., 2011). It is therefore the focus of this review. In addition, due to
the somewhat similarities of the two approaches, with the LCOE extending the LCC analysis a step further
through factoring energy production and discounting costs, the review focuses on the LCOE. The LCOE
consists of three main cost components, namely capital cost, Operation and Maintenance (O&M) costs, and
fuel cost, hence the cost breakdown in Table 4.2.
- 89 -
Table 4.2: International studies assessing power generation private costs (2010 values)
Study
Country
Capital
cost
O&M
cost
Fuel
Cost
LCOE
Type of plant
$/MWh
Booras & Holt,
2004
DNR & PSCOW,
2007
Karmis, 2005
BNEF, 2011
Rubin, Rao &
Chen, 2005
IEA GHG, 2004
Hoffmann &
Szklo, 2011
MIT, 2007
Cormos, 2012
USEIA, 2012
US
US
29.64
8.89
16.60
55.14
30.95
8.89
15.42
55.26
33.32
10.91
15.30
59.53
30.83
10.20
15.30
56.33
31.38
5.41
19.47
56.26
38.95
7.57
17.31
63.83
Virginia
73.72
74.84
90.48
73-155+1
56.79
101.15
59.51
85.75
Chile
US
The
Netherlands
US
US
IGCC, W/ capture, escalation
Conventional coal plant, Transmission cost
Supercritical PC plant, W/O capture
Supercritical PC plant, W/ CCS, FGD, SCR, ESP
IGCC plant, W/O capture, FGD, SCR, ESP
IGCC plant, W/ CCS
8.11
16.12
52.37
29.21
8.11
14.39
51.72
29.86
8.11
12.77
50.74
48.90
17.31
22.07
88.28
46.96
17.31
18.93
83.20
45.87
17.31
16.23
79.41
31.38
9.74
14.39
55.50
IGCC plant, W/O capture, 500MW net
41.44
11.36
17.74
70.54
IGCC plant, W/ capture, 500MW net, plus CO2 transport and
storage costs
51.29
27.93
79.21
69.60
44.86
114.46
64.9
74.1
91.8
71.68
97.03
96.4+1.2
109.8+1.2
137.5+1.2
31.5
35.7
45.7
43.91
41.90
China
42.08
71.85
63.98
UK
IGCC, W/O capture, escalation
28.13
72.06
NZEC, 2009
Davison, 2007
Scrubbed coal, CF 70%, ∂7%, escalation
Ultra-supercritical PC plant, W/O capture, FGD, SCR, 758MW
net, bituminous coal
Ultra-supercritical PC plant, W/ capture, FGD, SCR, 666MW
net, bituminous coal
Subcritical PC, W/O capture, FGD, NOx & TSP control, 500MW
gross, Illinois#6 bituminous
IGCC plant with GE gasifier, W/ CCS,596MW gross, Illinois #6
bituminous coal
Subcritical PC, W/O capture, 500MW net,Illinois#6bituminous
coal
Supercritical PC, W/O capture, 500MWnet, Illinois #6
bituminous coal
Ultra-supercritical PC, W/O capture, 500MW net, Illinois #6
bituminous coal
Subcritical PC, W/ capture, 500MW net, Illinois #6 bituminous
coal
Supercritical PC, W/ capture, 500MW net,Illinois #6
bituminous coal
Ultra-supercritical PC, W/ capture, 500MW net,Illinois #6
bituminous coal
50.70
Romania
US
Subcritical PC plant, W/O capture, 500MW net, capacity
factor 80%, Pittsburgh #8 bituminous coal
Supercritical PC plant, W/O capture, 500MW net, Pittsburgh
#8 bituminous coal
IGCC plant with spare gasifier, W/O capture, 500MW net,
Pittsburgh #8 bituminous coal
IGCC plant with no spare gasifier, W/O capture, 500MW net,
Pittsburgh #8 bituminous coal
Supercritical PC plant, W/O capture, 600MW,
bituminous coal
IGCC plant with assumed spare gasifier, W/O capture,
600MW,bituminous coal
27.92
11.17
22.34
61.43
42.45
22.34
23.46
88.24
39.09
24.57
23.46
87.12
- 90 -
IGCC plant, W/O CCS, Shell gasifier, 485.19MW net
IGCC plant, W/ CCS, Shell gasifier, 433.18MW net
Conventional coal, W/O CCS, CF 85%, transmission cost
Advanced coal, W/O CCS, CF 85%, transmission cost
Advanced coal with CSS, CF 85%, add transmission cost
Subcritical PC plant, W/O capture, 295.1MW net, ∂ 10%,
capacity factor 85%, bituminous coal
Supercritical PC plant, W/O capture, 574.1MW net, ∂ 10%,
capacity factor 85%, bituminous coal
Ultra-supercritical PC plant, W/O capture, 824MW net, ∂10%,
capacity factor 85%, bituminous coal
Ultra-supercritical PC plant, W/ capture using MEA solvent,
824MW net, ∂ 10%, capacity factor 85%, bituminous coal
IGCC plant, W/ capture using Selexol solvent, 661.7MW net, ∂
10%, bituminous coal
Ultra-supercritical PC plant, W/O capture, FGD, SCR,758MW
net, ∂10%, capacity factor 85%, bituminous coal
Ultra-supercritical PC plant, W/ capture using oxy, FGD, SCR,
532MW net, ∂ 10%, capacity factor 85%, bituminous
Ultra-supercritical PC plant, W/ capture using Flour, FGD, SCR,
666MW net, ∂10%, capacity factor 85%, bituminous
32.39
12.29
21.22
65.90
42.45
21.78
23.46
87.68
IGCC Shell plant, W/O capture, 776MW net, ∂ 10%,
bituminous coal
IGCC plant, W/ capture using Selexol solvent, 676MW net,
10%, bituminous coal
PC – Pulverised Combustion; IGCC – Integrated Gasification Combined Cycle; SCR – Selective Catalytic Reduction; ESP – ElectroStatic Precipitators;
W/O – without; W/ – with; CCS – Carbon Capture and Storage; ∂ – discount factor; CF – Capacity Factor.
Table 4.2 discloses that regardless of the coal technology, capital cost is the largest component of the LCOE,
followed by the fuel cost and in the lower end are O&M costs. Higher LCOE estimates are also associated
with coal technologies that have more pollution control technology, due to their higher capital cost, for
instance, plants fitted with FGD, plants with Carbon Capture and Storage (CCS) and plants with only carbon
capture. Coal technologies with only carbon capture report lower LCOE than those with CCS in spite of the
coal technology in question.
Furthermore disclosed by Table 4.2 is that pulverised combustion using subcritical, supercritical and ultrasupercritical steam cycles without CCS/capture report lower LCOE than similar IGCC plants without
CSS/capture, demonstrating that IGCC technology without CCS/capture is a more expensive technology
compared to any of the pulverised coal boilers (such is evident in the studies by Booras & Holt, 2004;
Karmis, 2005; Rubin et al., 2005; Davison, 2007; Department of Natural Resources & Public Service
Commission of Wisconsin (DNR & PSCOW), 2007; Massachusetts Institute of Technology (MIT), 2007).
Conversely, IGCC plants with CCS/capture are associated with lower LCOE than pulverised combustion
steam cycles with CCS/capture, demonstrating the benefits of avoided CO2 in IGCC plants over pulverised
steam cycles (this is evident in the studies conducted by Rubin et al. (2005), MIT (2007) and NZEC (2009)).
In addition, Table 4.2 further shows no distinct difference between LCOE outcomes of studies that
investigate subcritical, supercritical and ultra-supercritical steam cycles. For example, MIT (2007) reports
$52.37, $51.72 and $50.74/MWh for subcritical, supercritical and ultra-supercritical units without carbon
capture, respectively. MIT (2007) and NZEC (2009) report highest LCOE for subcritical units while Booras &
Holt (2004) and the United States Energy Information Administration (USEIA) (2012) report lowest LCOE for
subcritical units and high LCOE estimates for supercritical/ultra-supercritical steam cycles.
The higher efficiency units as shown in Table 4.2 are all associated with slightly higher capital cost than the
subcritical units. For example, MIT (2007) reports a levelised capital cost of $29.86, $29.21 and
$28.13/MWh for ultra-supercritical, supercritical and subcritical units, respectively while Booras & Holt
(2004) report $55.26 and $55.14/MWh for supercritical and subcritical units, respectively. The increased
efficiency offered by supercritical and ultra-supercritical boilers is one way of reducing GHG emissions per
unit of power produced for the reason that less coal is burned in such units. The reduction in pollutants in
- 91 -
such units is not only limited to GHGs but also applies to other pollutants such as SO2 and NOX due to less
coal being burned. Booras & Holt (2004) estimate that an ultra-supercritical unit with an efficiency of 4648% would emit about 18-22% less CO2/MWh of power generated than an equivalent-sized subcritical unit.
A closer look at Table 4.2 also supports the low fuel consumption of the higher efficiency boilers in that
higher efficiency boilers are evidently associated with lower fuel cost in the studies by Booras & Holt (2004)
and MIT (2007). These units will therefore offer lower emissions and favourable LCOE comparisons over
subcritical units in coal-importing countries, that is, in countries where the fuel cost constitutes a higher
fraction of the LCOE.
The LCOE of subcritical plants without capture, generally range from a low value of $44/MWh to a higher
value of $96/MWh. In the upper-end are subcritical plants with more pollution control technology and/or
plants that consider electricity transmission costs. The study by Bloomberg New Energy Finance (BNEF)
(2011) which reports a high value of $155/MWh for such plants was considered an outlier and is therefore
not included in the given range. The LCOE for higher efficiency plants (i.e. supercritical and ultrasupercritical) without capture, with capture and with CCS range between $51 - $110/MWh, $72 - $88/MWh
and $101 - $138/MWh, respectively. IGCC LCOE for plants without capture, with capture and with CCS
range between $56 - $75/MWh, $64 - $90/MWh and $86 - $114/MWh. The estimates are sensitive to a
number of factors, among which are escalation, discount rate, fuel prices and capacity factor, so most of
the studies reviewed here conduct sensitivity analyses. Lastly, while LCOE is a beneficial initial step for
approximating the costs of generating electricity, the tool does not consider externality costs associated
with the power technologies. Studies that address externality costs of power generation technologies are
discussed in the following section.
4.3.2
International studies assessing power generation externality costs
Several international studies have made attempts to quantify the externality costs of coal-based power
using various valuation methods. The inflation adjusted damages (2010 values) of the reviewed studies are
contained in Table 4.3. The table shows that several of these studies were conducted in Europe and in the
US, with estimates varying according to the country in question, fuel cycle stage(s) studied and the range of
impacts investigated.
The abatement cost approach was the earliest approach used by researchers to derive damage cost
estimates of power generating units from various fuel sources. Some of the early work includes that of
Schuman and Cavangh (1982), Chernick and Caverhill (1989) and Bernow et al. (1991). All three studies
- 92 -
focused on air emissions from fossil fuel combustion in the US, so other fuel cycle stages and their
associated impacts were not investigated. Table 4.3 shows the estimates from Schuman and Cavangh to be
highly variable with estimates from later studies utilising the same approach being relatively stable.
Hohmeyer (1988) was among the early researchers who used the top-down damage cost approach to
derive externality estimates for a number of fuel sources including coal. Like other earlier studies, the focus
was on air pollution impacts arising from the fuel combustion stage, however, Hohmeyer’s own estimates
are lower than that of other researchers who used the same approach (i.e. Ottinger, Wooley, Robinsson,
Hodas & Babb, 1991; Pearce, Bann & Georgiou, 1992; Faaij et al., 1998) because global warming impacts
were not assessed in his study. The bottom-up approach employed in these studies, however, did not allow
for site specific type of impacts since it utilizes highly aggregated emissions data to approximate costs of
specific pollutants.
With the development of the bottom-up approach, new studies considered site specificity and made
attempts to consider the entire coal cycle, for example, European Commission (1999b) and European
Commission (2005). Also due to the comprehensiveness of these studies, higher damages were realised
than similar studies employing the same approach but focusing on a narrow range of impacts (for example,
ORNL & RfF (1994) and European Commission (1995) who report lower damages due to the exclusion of
CO2 damages) and/or a subset of the fuel cycle stages (for example, ORNL & RfF (1994). In addition, like all
studies reported in Table 4.3, the bottom-up damage cost estimates vary according to the country in which
the assessments were conducted, thus making country specific assessments fundamental.
The benefit transfer technique has also been used by a number of researchers through transferring and
adjusting damage cost estimates estimated using the bottom-up approach from other studies to the new
contexts, for example Epstein et al. (2011), International Panel on Climate Change (IPCC) (2007a),
Sevenster, Croezen, Van Valkengoed, Markowska and Donszelmann (2008) and Bjureby et al. (2008). As
expected, the first two studies report damage cost estimates that are within the range of estimates
reported by studies employing the bottom-up approach. The latter two studies’ damage cost values are,
however, not normalised to per kWh and are therefore not reported in Table 4.3.
Lastly, it is important to highlight that few studies (as evidenced by Table 4.3) focused on plant construction
and that for those investigating coal mining and transportation the focus was on mainly three impacts,
namely climate change impacts, human health burdens due to air pollution, and fatalities due to coal
transportation. Finally, though the studies reviewed investigate generation phase externalities, only direct
- 93 -
emissions from combusting coal were quantified and monetised. Thereby a number of operation phase
impacts, including indirect impacts linked to the material requirements of operating the plant, were not
investigated.
Table 4.3: International studies on power generation externality costs (2010 values)
Study
Country
Method
Externality
1
costs
Phases & impacts considered
US cents/kWh
Schuman & Cavangh, 1982
Chernick & Caverhill, 1989
Bernow et al., 1991
Hohmeyer, 1988
Ottinger et al., 1991
Pearce et al., 1992
ORNL & RfF, 1994
European Commission,
1995
European Commission,
1999b
Epstein et al., 2011
IPCC, 2007a
US
US
US
Abatement
Abatement
Abatement
0.14—99.67
7.69—13.62
6.61—14.78
Combustion phase (only CO2 effects)
Germany
US
UK
Top-down
Top-down
Top-down
0.15—7.82
5.80—14.19
4.15—22.44
Combustion phase (air pollution effects, not GHGs)
US
Bottom-up
0.16—0.71
Mining, transport and combustion phases (air pollution
effects, not CO2)
UK
Bottom-up
Bottom-up
Bottom-up
Bottom-up
Bottom-up
1.40
3.42
0.60—20.59
2.55—25.53
1.81—26.40
Entire fuel chain - including decommissioning (air pollution
effects, not CO2)
Benefit
transfer
9.48(low)
18.07(best)
27.24(high)
7.7
Germany
Finland
Germany
The
Netherlands
US
US
Benefit
transfer
Combustion phase (air pollution effects, plus GHGs)
Combustion phase (air pollution effects, plus GHGs)
Combustion phase (air pollution effects, plus GHGs)
Combustion phase (air pollution effects, plus GHGs)
Entire fuel chain - including decommissioning (air pollution
effects, plus GHGs)
Mining, transport and combustion phases (air pollution
effects, plus GHG, coal transportation accidents)
Mining and combustion phases (air pollution effects, plus
GHG)
1
Inflation adjusted values to 2010.
4.3.3
International studies modelling power generation systems
As evidenced by the systems modelling tools in Table 4.1 there are a variety of computer models that have
been designed for energy analysis/optimization/planning. As indicated below, the various models are
designed for different purposes, for example assessing and comparing energy technologies, cost
minimization and reducing greenhouse gases. The models are additionally designed for different
technologies (e.g. renewable energy, non-renewable energy or both) and scales of analyses (e.g. national,
regional or global).
For instance, the EnergyPLAN model in the literature has mainly been used to simulate renewablepenetrations. It was used by Lund and Mathiesen (2009) and Cosic, Krajacic and Duic (2012) to design 100%
renewable energy systems for Denmark and Macedonia, respectively. Liu, Lund and Mathiesen (2011) used
it to study the influences and barriers of integrating wind power into China’s present energy system while
Lund (2006) used it to study the integration of photovoltaics - wind plus wave power - into the electricity
supply system of Denmark. The model is designed for national or regional analysis. The MARKAL family of
- 94 -
models on the other hand, are energy/economic/environmental tools that have been used widely for
various purposes, including comparing electricity generation technologies (Naughten, 2003), development
of carbon mitigation strategies (Jegarl et al., 2009), internalisation of power production externality cost
(Rafaj & Kypreos, 2007), analysing the market effects of CO2 emission markets and the effects on electricity
of green certificate market (Unger & Ahlgren, 2005) and waste management modelling (Cosmi et al., 2000;
Salvia et al., 2002). The models have been designed for national, regional or global analyses.
The Balmorel model, like the MARKAL tools, has been used to analyse a number of issues, including security
of electricity supply (Morthorst, Jensen & Meibom, 2005; Jensen & Meibom, 2008), expansion of electricity
transmission (Heggedal, 2006), development of international electricity markets (Ea Energy Analyses,
Hagman Energy, COWI. 2008), wind power development (Ea Energy Analyses, 2007; 2008), international
green certificates markets and the trade of emissions (Lindboe, Werling, Kofoed-Wiuff & Bregnbaek, 2007).
HOMER, a micro-power design tool, has also been used to determine and compare a green (solar and wind)
and a diesel-based energy system in Malaysia with respect to net present cost and pollutant gas emission
(Ashourian et al., 2013), to simulate a 100% renewable energy system (Lambert, Gilman & Lilienthal, 2006)
and to study wind energy potential in Ethiopia (Bekele & Palm, 2009).
The energy systems analysis models thusfar reviewed show diverse application and according to Bassi et al.
(2007) although being detailed tools, they do not efficiently simulate the interaction between the energy
system and the main factors in the entire economy, environment and society, as does an innovative
Threshold 21 (T21) model. T21 is a dynamic simulation framework built to aid extensive, integrated, longstanding, nation-wide planning with severe devotion to causality (Barney, Eberlein & Sharma, 1995). T21
can be built into a system dynamics platform (system dynamics based T21 model) (Bassi et al., 2007). The
model has been adapted for developed nations like the US and Italy and developing nations such as Malawi
and Mozambique. For example, in Italy it was used by the national environmental agency to study how the
Italian government could meet the terms of the Kyoto Protocol GHG commitments without hindering the
economy, while in Malawi it was used by the National Economic Council to analyse strategies for reaching
Malawi’s Vision 2020. In Mozambique it was used by the Ministry of Planning and Development to support
the national visioning process, Agenda 2025 and national development planning (Millennium Institute,
2010) while in the US it was used to analyse the main energy challenges and choices faced by the state in
the wider context of their relation to society, environment and the economy, and with links to the rest of
the world (Bassi et al., 2007).
- 95 -
Other system dynamics energy models that fall into the shortcoming noted by Bassi et al. (2007) include,
Energy Transition Model which is a general disequilibrium model considering only energy-economy
interactions (Sterman, 1981), Feedback-Rich Energy Economy model which is a climate-economy model
focusing solely on economy-climate interactions (Fiddaman, 1997), Petroleum Life Cycle Model (Sterman,
Richardson & Davidsen, 1988; Davidsen, Sterman & Richardson, 1990), FOSSIL model (Backus, Green &
Masevice, 1979) and IDEAS model (AES Corporation, 1993) which consider energy in isolation. The
Petroleum Life Cycle Model depicts the development of the petroleum resource and ancillary industry of
the US, beginning in 1870. The IDEAS model is a dynamic energy supply and demand policy simulation
model of the US (i.e. an improved version of the FOSSIL model).
Most recent applications of system dynamics modelling to energy-related issues analysis include those that
have focused on fossil fuels, for example, Jeong et al. (2008) designed a system dynamics model for power
generation costs comparison in a coal-based power plant and a liquefied natural gas combined cycle plant
while also taking into account control costs of CO2 and NO2. Hoffmann, Hafele and Karl (2013) analysed
climate change effects on efficiency and power generation in selected German thermal power plants with
once-through and closed-circuit cooling systems through a dynamic simulation model based on a system
dynamics methodology. Hou, Xia, Zhang, Lou, Zhang and Xin (2009) developed a system dynamics model to
forecast growth trends in coal demand, supply, reserves and pollution under several economic growth
scenarios while Fan, Yang and Wei (2007) designed a system dynamics model taking into account coal
industry investment, mine construction and reserves and used it to optimise coal investment size.
On the other hand, Robalino-Lopez, Mena-Nieto and Garcia-Ramos (2014) developed a system dynamics
model to study the effects of improving the efficiency of fossil energy use and that of reducing fossil energy
on CO2 emissions of Ecuador while Shih and Tseng (2014) focusing on coal-fired power generation as the
marginal supplier of electricity, built a system dynamics model to study the social benefits of an energy
policy promoting sustainable energy. The model was used to simulate energy saving under energy
efficiency improvements and renewable energy promotion. Life-cycle co-reductions of GHGs and classic air
pollutants were estimated. Li, Dong, Li, Li, Li, & Wan (2012) designed a system dynamics model for a
traditional industrial region in China characterized by high CO2 emissions (Liaoning Province) and used it to
simulate CO2 emission trends under various scenarios while Feng, Chen and Zhang (2013) modeled energy
consumption and CO2 emission trends for the City of Beijing. Carbon rich fuel (coal) and low carbon fuel
such as natural gas were considered. Mao, Dai, Wang, Guo, Cheng, Fang et al. (2013) concentrated on all
industries in China and simulated carbon emissions and GDP growth under three scenarios using a system
dynamics model. Energy consumption focused on nine energy sources including coal, natural gas and coke.
- 96 -
Qudrat-Ullah and Davidsen (2001) built a dynamic simulation model based on system dynamics to study the
dynamics of the electricity system in Pakistan. Focus was on assessing the effects of government policy
incentives to private sector investments (coal, oil and gas power plants) on resource import dependency,
electricity supply and CO2 emissions. Dastkhan and Owlia (2014) developed a regional dynamic integrated
electricity model to explore the right policies for electricity generation in the Middle East. Among the
technologies considered were coal, gas and solar power plants. A number of scenarios and policies were
studied. Qudrat-Ullah (2013) developed a system dynamics model to understand electricity supply and
demand in Canada. Coal, uranium, crude oil, natural gas, wind and hydro are among the major electricity
sources considered.
Other researchers have used system dynamics to model renewable energy technologies or to study the
transition towards more sustainable energy systems, for example, Saysel and Hekimoglu (2010) who
developed a dynamic simulation model of electric power industry in Turkey (i.e. a model that represents
the investment, production, pricing and financing structures of a number of energy technologies including
coal) and used it to study options for CO2 mitigation through fossil fuel based power early retirements and
replacements with clean energy resources. Pruyt (2007) used a system dynamics model to study the
transition of EU-25 electricity generation system, towards a more sustainable energy system characterised
by lower CO2 emissions while Ford et al. (2007) used system dynamics to simulate price dynamics in a
market for tradable green certificates to encourage wind electricity. Vogstad et al. (2002) built a system
dynamics model for the Nordic electricity market and used it to investigate the short-term and long-term
energy planning trade-offs. The aim was to find efficient policies to aid the transition from fossil-fueled
based power supply to renewables. Focusing on the Swiss electricity market, Ochoa and Van Ackere (2009)
built a system dynamics model to study the dynamics of capacity expansion and the effects of various
policies such as phasing out nuclear and imports and export electricity policies. Among the power plants
considered were hydro, nuclear, solar panels and wind turbines.
Additional research in this category include that of Aslani, Helo and Naaranoja (2014) who constructed a
system dynamics model to evaluate the role of renewable energy promotion policies on Finland’s energy
dependency, and Cepeda and Finon (2013) who built a system dynamics model for simulating electricity
investment decisions in the case of either market driven or subsidized large-scale wind power
development. Qudrat-Ullah (2014) developed a dynamic simulation model based on a system dynamics
methodology to investigate the dynamics of electricity generation capacity in Canada. Focus was on
identifying a sustainable and balanced electricity generation capacity scenario for the country. Among the
energy sources considered were hydro, thermal and nuclear. Focusing on biofuel production Rendon- 97 -
Sagardi, Sanchez-Ramirez, Cortes-Robles, Alor-Hernandez and Cedillo-Campos (2014) developed a system
dynamics model for assessing the viability in ethanol supply chain for biofuel generation in Mexico. The
availability of cropping area, capacity of ethanol and fuel, reduction of CO2 emissions, as well as five
scenarios were evaluated. The feedstocks considered were grain sorghum and sugarcane. Also focusing on
biofuels, Barisa, Romagnoli, Blumberga and Blumberga (2014) developed a system dynamics model to gain
understanding into the longstanding dynamic behavior of Latvia’s biodiesel market. A number of policy
instruments in support of biofuel production were explored including state subsidies and increasing taxes
on fossil fuels.
Other examples of energy-related system dynamics models include those that have combined system
dynamics models with other methods, for example, Yu and Wei (2012) who developed a hybrid model
centred on system dynamics and generic algorithm for analysis of coal production and environmental
pollution load (specifically three kinds of waste - waste gas, water plus solids) in China. Dyner et al. (2011)
built a system dynamics model linked to an iterative algorithm to evaluate the effects of integration of
electricity markets on system expansion and security of supply while Pereira and Saraiva (2011) combined
system dynamics with generic algorithms to help market agents to develop long-term generation
investment plans. System dynamics was used to simulate the evolution of capacity factors, electricity
demand and prices while the generic algorithm was utilized towards maximizing the profits of each
generation agent (Pereira & Saraiva, 2011). Focusing on wind turbines Tan et al. (2010) combined system
dynamics with decision trees to analyze investment alternatives in the face of multiple uncertainities and
high managerial flexibility. The combination allowed for the consideration of dynamic complexity (system
dynamics) and managerial flexibility (decision tree method).
On the other hand, Sanchez et al. (2008) combined system dynamics with game theoretical methods to
study long-term investment dynamics in electricity generation while Pasaoglu (2006) built the model
Liberalized Electricity Market Microworld (LEMW) which integrates system dynamics and analytical
hierarchy processes. Short-term and long-term dynamics of electricty demand and supply were considered.
Also considered were socio-economic and political issues like environmental impact, environmental costs
and resource availability. The model permits the evaluation of various business strategies for utilities along
with regulatory authorities’ programs. Lastly, Vogstad (2005) combined system dynamics and experimental
economics to evaluate the effect of emissions trading on the Swedish electricity market. Experiments were
used to identify various trading plans for renewable energy certificates.
- 98 -
Systems modelling tools have therefore been developed for several countries and for addressing various
energy-related issues, for example, to model fossil fuels and renewable energy, to study the transition
towards more sustainable energy systems and furthermore researchers have combined energy-related
system dynamics models with other methods such as game theoretical approaches, analytical hierarchy
processes, generic algorithms, iterative algorithm and decision trees. However, as evidenced by the review
not all facets/features of coal-based power production had been studied by the researchers, for example
the models were not tailored to specific coal-based power generation technologies, did not address social
cost nor permit deeper (comprehensive) and explicit understanding of coal-based power generation and its
interactions with resource inputs, private costs, externalities, externality costs and hence its consequent
economic, social and environmental impacts over its lifetime and fuel cycle. Also evidently environmental
focus was on quantifying direct GHG emissions from the coal combustion phase thus numerous combustion
phase and upstream environmental impacts can still be incorporated and monetised to advance coal
energy analysis.
4.3.4
Local studies assessing power generation private costs
The literature discloses a number of local studies that have studied or rather-modified, private costs of
power generation technologies estimated in international studies to the South African context, for example
EPRI (2010) and IRP (2011). Mokheseng (2010) estimates the NPV of solar photovoltaics and compares
these to coal-based power through adjustments of cost data from the literature. EPRI (2010) provides cost
and performance data on a number of power generation technologies, for example fossil fuel based
technologies such as pulverised coal, IGCC, Fluidized Bed Combustion (FBC) and renewable resource based
technologies such as wind, biomass and solar photovoltaics. The construction and O&M costs of the various
power generation technologies were presented as overnight costs, which assumes that the plant is built
overnight and for this reason the costs do not include interest and financing costs. Used as a baseline for
the cost estimates were recent EPRI studies on US-based plants. Adjustment factors for materials, labour
productivity and labour rates were used to convert construction costs for the US to construction costs in
South Africa. Assumptions of the fraction of equipment imported and supplied locally were also made.
Water consumption and CO2 emissions were also estimated and reported in the EPRI report. The reported
LCOE for various power technologies is shown in Table 4.4.
The outcomes of the EPRI report have been used to facilitate the IRP process of South Africa. The IRP (2011)
investigates how South Africa’s electricity demand can be met between 2010 and 2030. Various
technologies’ private costs under various scenarios were reported for a representative pulverised
combustion plant, IGCC plant, FBC plant and other fossil–based power plants such as nuclear and
- 99 -
renewable energy sources like wind and solar. Emissions in the form of CO2, NOx, SO2 and particulates were
also estimated for the various scenarios. Pulverised coal-based power LCOE from the IRP report are also
reported in Table 4.4. Other economic analysis studies in the country address renewable energy sources for
example, Pouris (1987) and Du Plessis (2011).
Table 4.4 discloses the LCOE to be highly composed of capital cost, followed by fuel and O&M costs. The
fuel costs are generally similar irrespective of the coal technology and plant size, though slightly higher for
PC plants than IGCC plants. Plants without FGD show slightly lower LCOE than plants with FGD. The table
furthermore shows pulverised units to have lower LCOE than IGCC plants. The limitations of the cost and
performance estimates as stated in the EPRI report are that they are conceptualised for the South African
context. South Africa’s ground-up estimates as stated in the EPRI report were not feasible due to time
constraints. Site and company specificity conditions are therefore not reflected by the estimates. Lastly,
though the LCOE is a beneficial initial step for approximating the costs of generating electricity, the tool
does not consider the externality costs linked with the power technologies. Local studies that address
externalities of power generation technologies are discussed in the following section.
Table 4.4: Local studies assessing power generation private costs
Study
EPRI, 2010
IRP, 2011
EPRI, 2010
Capital cost
295
305.5
321.5
338.8
351.4
373.9
115.3
119.1
125.9
212
424
468.1
536
629.3
666.6
583
1859.8
1348.7
O&M cost
Fuel cost
ZAR/MWh
83
144.6
84.6
144.6
87.2
144.6
105.5
146.5
107.8
146.5
111.9
146.5
424
146.4
439.8
146.4
468.1
146.4
95
147
155.2
146.4
125.9
146.4
95.2
67.3
118.1
64.1
87.3
74.9
166
168
LCOE
522.6
534.7
553.4
590.8
605.7
632.4
685.6
705.2
740.4
464
685.6
740.4
698.5
811.5
754.3
657.8
2025.8
1516.7
Type of plant
PC plant, without FGD, 4500MW net, capacity factor 85%
PC plant, without FGD, 3000MW net, capacity factor 85%
PC plant, without FGD, 1500MW net, capacity factor 85%
PC plant, with FGD, 4500MW net, capacity factor 85%
PC plant, with FGD, 3000MW net, capacity factor 85%
PC plant, with FGD, 1500MW net, capacity factor 85%
Shell IGCC, 3,865MW net
Shell IGCC, 2,577 MW net
Shell IGCC, 1288MW net
Pulverised fuel, capacity factor 85%
IGCC – six 2x2x1 Shell IGCC
IGCC – two 2x2x1 Shell IGCC
Nuclear Areva EPR – 6 units
Nuclear AP1000 – 6 units
Wind - 10x2MW farm – wind class 6
Wind - 100x2MW farm – wind class 6
Soar - Parabolic trough, storage 9hrs, net power 125MW
Solar - Central receiver, storage 14hrs, net power 125MW
PC – pulverised combustion; IGCC – Integrated Gasification Combined-Cycle Generation; ZAR – South African Rand
- 100 -
4.3.5
Local studies assessing power generation externality costs
Several local studies have made an effort to quantify the externality costs of coal-based power, for
example, Dutkiewicz and De Villiers (1993), Van Horen (1997), Van Zyl, Raimondo and Leiman (2002),
Spalding-Fecher and Matibe (2003) and Pretorius (2009). The studies by Van Zyl et al. (2002) and Pretorius
(2009) focus strictly on the coal mining phase with Pretorius (2009) estimating the water pollution
externality cost for Eskom’s coal requirements to be R0.38/kWh while Van Zyl et al. (2002) estimate the
impact of coal mining on the quality of water in the eMalahleni catchment to be between R8.56 million and
R17.13 million (R0.12–R0.23/t). The Van Zyl study further estimates climate change impact of CH4 emissions
emitted during coal mining to range between R180 million and R1.260 billion (R0.98–R6.83/t).
The rest of the studies focus mainly on the operation phase with Dutkiewicz and De Villiers making use of
the top-down approach to value externalities while the other local studies used the bottom-up approach or
rather transferred damage cost estimates estimated using the bottom-up approach from international
studies to the South African context. The inflation adjusted externality costs (2010 values) of the reviewed
studies is shown in Table 4.5. The externality cost estimates produced by Dutkiewicz and De Villiers fall in
the lower range of the estimates produced by international studies using a similar approach (see Tables 4.3
and 4.5). Those produced by Van Horen are higher than those of Spalding-Fecher and Matibe due to a
broader range of impacts being considered. Nonetheless, the estimates from both studies are lower than
the damage cost estimates from similar studies conducted abroad (see Tables 4.3 and 4.5), partly because
of focusing on a subset of the fuel cycle stages.
Table 4.5: Local studies assessing power generation externality costs (2010 values)
Top-down
Externality
1
costs
US cents/kWh
0.51
van Horen, 1997
Benefit transfer
0.76—4.27
Mainly combustion phase (air pollution effects, GHG, water
consumption & mining accidents)
Spalding-Fecher & Matibe, 2003
Benefit transfer
0.34—2.24
Combustion phase (air pollution effects, GHG)
Nkambule & Blignaut, 2012
Benefit transfer
Inglesi-Lotz & Blignaut, 2012
Riekert & Koch, 2012
Blignaut et al (2012)
Statistical
Benefit transfer
Benefit transfer
Study
Dutkiewicz & de Villiers, 1993
Method
Phases & impacts considered
Coal mining and transportation (air pollution effects, GHGs,
mortality, morbidity, water use & pollution, etc.)
4.23—25.66
Combustion phase (water use externality)
Combustion phase (air pollution effects)
Combustion phase (CO2)
1
Own calculations based on values reported in the studies. Inflation adjusted values (ZAR) and converted to 2010 US dollars ($).
The rest of the studies in Table 4.5 are independent studies, but that were executed in a single project for a
specific plant (Kusile), so their externality costs were summed. Nkambule & Blignaut (2012) focused on the
externalities of mining coal and transporting it to Kusile. This study is a product of this thesis and it
concentrated on climate change effects, air pollution-related health effects, mortality, morbidity, water
pollution, water use externality and the loss of ecosystem services. Riekert and Koch (2012), Inglesi-Lotz
- 101 -
and Blignaut (2012), and Blignaut (2012) focused on the coal combustion phase in Kusile, and studied air
pollution-related health effects, water consumption externality and climate change effects, respectively.
The externality costs of Kusile were approximated to range between 4c/kWh – 26c/kWh, values that are
comparable to those produced by similar studies conducted abroad.
The outcomes of these studies are an improvement over the earlier black-box national level studies as they
focus on a specific plant and somehow disclose the links between plant type/performance and
environmental and societal burdens. Nonetheless, these studies can also be improved upon by making the
cause-effect relationships explicit (through a system dynamics model), by widening the breadth and width
of the measurable externality costs within the combustion phase (e.g. through assessing public and
occupational health impacts (fatalities and mortalities), non-CO2 GHG impacts, etc.) and through assessing
indirect burdens linked with the production and transportation of material requirements for operating
Kusile, construction phase burdens, FGD system burdens (as Kusile will be fitted with this technology) and
by embracing the long-term repercussions of the coal-fuel chain on the environment and social systems.
4.3.6
Local studies modelling power generation systems
Locally, there are studies that have employed computer models to analyse energy-related issues. For
instance, Taviv et al. (2008) used the Long-range Energy and Alternatives Planning (LEAP) energy modelling
tool to model energy demand and supply from 2005 to 2030 under alternative assumptions on energy
drivers in South Africa. Haw and Hughes (2007) used two energy models, namely the LEAP system to
generate South Africa’s future energy demand based on GDP and population growth coupled with supplyside options for meeting demand, and the MARKAL model to optimize for least cost. Musango et al. (2009)
used a partial T21 model to study energy supply and demand in South Africa, and how energy efficiency
measures and nuclear energy production expansion could help meet the country’s future energy
requirements. Winkler et al. (2011) used the MARKAL model to project GHG emissions under business as
usual in South Africa, while Hughes et al. (2007) explored a number of scenarios including final energy
demand reduction by 15% lower than the forecasted 2015 levels and the 2013 renewable energy target of
10 000 GWh. Pauw (2007) used the Standard General Equilibrium model to study the probable impact that
different climate change mitigation alternatives may possibly have on the South African economy with
regards to the wellbeing of households, employment and GDP.
In a different application, Davis, Cohen, Hughes, Durbach & Nyatsanza, (2010) used three models, namely
system dynamics modelling, statistical regression analysis and the LEAP model to study the rebound effects
of energy efficiency interventions in the residential sector in South Africa and the effectiveness of measures
- 102 -
aimed at reducing the rebound effects. System dynamics was used to model household energy
consumption behaviour, regression models were used, among other issues, to test the hypotheses
generated by the system dynamics model while the LEAP model was used to assess the rebound effect and
its mitigation. Other energy models have been applied to solely address renewable energy issues. For
example, Dekker, Nthontho, Chowdhury and Chowdhury (2012) used HOMER software to study the
economic viability of introducing photovoltaics/diesel hybrid power systems in each of the six climatic
zones of South Africa, while Musango et al. (2012) used a system dynamics approach to develop a model
for assessing the sustainability of bioenergy and used it to assess the effects of the development of a
biodiesel industry on a number of sustainability indicators in the Eastern Cape province of South Africa.
Based on the T21 framework, South African Green Economy Modelling (SAGEM) was developed to study
the transition of South Africa to a green economy. The effects of green economy investments in selected
circumstances and sectors including the energy sector were assessed (Department of Environmental Affairs
and United Nations Environment Programme, 2013). Though the study include coal-based electricity
generation the coal electricity module and coal-related modules (e.g. water demand electricity generation
module and air emissions module) are to a great extent black-boxes (e.g. in terms of coal cost, CO2
emissions and water use) and the study pay no attention to coal technologies and life cycle analysis. In
addition, the focus of the study is at a national level and it does not address externality costs and social
costs.
The local literature thus discloses that various researchers have employed energy modelling tools to study
an assortment of energy issues, among which is modelling South Africa’s energy demand and supply (Taviv
et al., 2008; Haw & Hughes 2007), modelling bioenergy supply (Musango et al., 2012); projecting the
country’s GHG emissions under various scenarios (Winkler et al. 2011; Hughes et al. 2007); and studying
impacts of various climate change mitigation options on employment, household welfare and GDP (Pauw,
2007). None of the energy models have been employed to study coal-based power generation and its
interactions with resource inputs, private costs, externalities, externality costs and hence its consequent
economic, social and environmental impacts over its lifetime and fuel cycle.
4.3.7
Summary
In this chapter, an overview of the various tools employed by various researchers to estimate the private
and/or externality costs of power generation technologies was conducted, followed by reviewing the
application of the assessment tools in the power sector with special emphasis on coal-based power
generation. The review discloses that an assortment of methods and tools have been adopted by
- 103 -
researchers to evaluate power generation technologies contingent on the specific goals and scopes of the
applications. The tools were grouped into three broad categories of methods, namely financial analysis
methods, impact analysis methods and systems analysis methods. The review of financial measures
discloses that different financial measures are suited for different computations. Generally though, cost
effective energy projects are those with lowest LCOE, LCC, simple payback period and discounted payback
period plus those with high IRR, MIRR and NPV. The review of impact analysis tools discloses that various
tools are suited for identifying, quantifying and monetising externalities. Depending on the aims of the
investigations and a number of issues surrounding the research (such as time and financial constraints and
availability/unavailability of previous primary valuation studies), various researchers employ various impact
analysis tools. The review of the systems analysis tools discloses that various systems models are designed
for different purposes (e.g. modelling energy system, and/or economic system, and/or ecological system),
different technologies (e.g. renewable energy, non-renewable energy or both) and different scales of
analyses (e.g. national, regional or global).
Concerning the application of the tools in the power sector, the review discloses that in the past three
decades a variety of studies were conducted on electric sector private and externality costs in both
developed and developing countries. Earlier externality studies used the abatement cost and bottom-up
approaches to derive externality costs estimates while recent studies used the bottom-up approach and/or
benefit transfer technique to estimate externality costs of power generation. The studies differ in terms of
the types of externalities they focus on, the fuel-cycle stage(s) they investigate, and they do not factor in
the long-standing repercussions of the technologies on the environment and social systems. The most
investigated externalities internationally and locally are climate change and human health impacts
associated with airborne pollution from coal combustion. More attention is still paid to the power
generation phase even in more recent studies. These differences in scope affect the outcomes of the
studies, make comparing them difficult, and highlight the need for comprehensive externality investigations
that widen the range of externalities studied, that consider the various fuel-cycle stages and that embrace
the long-term repercussions of the technologies. A systematic investigation of burdens in a life-cycle
manner, can limit the exclusion of important externalities in the coal fuel chain, for example, coal mining
and processing externalities, coal transportation-related externalities, plant construction-related impacts
and can ensure that externality assessments are reflective of how the plant and its associate upstream and
downstream processes are operated.
Finally, the literature discloses that systems modelling tools have been developed for several countries and
for addressing various energy-related issues (locally the models are mainly used for modelling energy
- 104 -
supply and demand, projecting GHG emissions and studying climate change mitigation options), however,
as evidenced by the review not all facets/features of coal-based power production had been studied by the
researchers, for example, the models were not tailored to specific coal-based power generation
technologies, did not address social cost nor permit deeper (comprehensive) and explicit understanding of
coal-based power generation and its interactions with resource inputs, private costs, externalities,
externality costs and hence its consequent economic, social and environmental impacts over its lifetime
and fuel cycle.
Specifically, the top-down and bottom-up energy systems models were found to offer piecemeal
information that limits deeper understanding of energy technologies and their consequent economic,
environmental and societal impacts. This was so because the top-down models presented the energy
system as a black-box, by paying no attention to the processes and activities because the matrices used can
only analyse a sector as a whole, and as a result differentiation between a range of products or production
methods nor technologies was not possible. In addition, environmental focus was on GHGs and the links
between plant type/performance and environmental/societal burdens were hidden. The bottom-up
models’ shortcomings included that they are generally static models, with no feedback loops and time
delays. In addition, they optimized for least cost in private terms not in social terms, and environmental
focus was on GHGs, especially direct GHG emissions from the coal combustion phase. As a result numerous
combustion phase and upstream burdens can still be incorporated and monetised to advance coal energy
analysis.
Argued in this study is that a comprehensive assessment of social costs is highly necessary to aid decisionmaking on least social cost energy options for future energy supply. Advocated for such an assessment is a
systems thinking approach namely, system dynamics along a life-cycle viewpoint. The current study thus
develops a system dynamics model for understanding coal-based power generation and its interactions
with resource inputs, private costs, externalities, externality costs and hence its consequent economic,
social and environmental impacts over its lifetime and fuel cycle.
- 105 -
CHAPTER 5:
5.1
RESEARH DESIGN AND METHODS
Introduction
Research designs are detailed plans and procedures outlining how a research project will be conducted. The
designs explain and motivate decisions taken by the researcher regarding the philosophical beliefs that
underpins the study (research philosophy/paradigm), the strategy of inquiry for the study (research
approach) and the specific methods of data collection, analysis and interpretation (research methods)
(Creswell, 2008). This chapter therefore discusses in detail the three main components of the study’s
research design, namely the research philosophy/paradigm that underpins the study, the strategy of
inquiry and the specific research methods that were used to collect and analyse data so as to realize the
specific objectives of the study.
5.2
Research paradigm/philosophy
The primary aim of this section is to discuss the research paradigm that underpins this study. In chapter 3
the research conducted in this study was grounded within the economic discipline of study in which it falls
and the use of a systems approach to model the life-cycle burdens and social costs of coal-based electricity
generation was motivated through studying the links between system dynamics and the schools of
economic thought that underpin this study. In pursuit of these aims: the concept of research paradigms
was defined; a discussion of Guba and Lincoln’s social science research paradigm framework for
deliberating main matters of research methodology in social science was conducted; a review of the history
of economic thought and the classification of the economic disciplines according to the research paradigms
of Guba and Lincoln were conducted; and a review of the literature on social theoretic beliefs underlying
system dynamics practice, particularly the system dynamics paradigms of Pruyt was conducted. Based on
this information an attempt was made to: determine the schools of economic thought that underpin this
study and to classify them according to the research paradigms of Guba and Lincoln; to place the system
dynamics practice of this current study on Pruyt’s extended paradigmatic table; and to study the links
between system dynamics and the schools of economic thought that underpin this study.
The review of economic thought disclosed that the main concepts in this study, namely production,
externalities and social cost are rooted in neoclassical and environmental economics, particularly, in
welfare economic theory, theory of production and Pareto efficiency. Neoclassical and environmental
economics are therefore the main economic disciplines that provide the theoretical basis for this study. The
ontology (i.e. the philosophical beliefs/assumptions researchers place on the nature of reality),
- 106 -
epistemology (i.e. the nature of knowing and construction of knowledge) and methodology (i.e. the
procedures and techniques the researcher use to investigate what can be known) of both neoclassical and
environmental economics were discussed to be realist, objective and quantitative, respectively, and hence
to fall within the positivist research paradigm of Guba and Lincoln’s classification (i.e. both schools
acknowledges a reality that is controlled by absolute laws of nature, views economics as an objective
science that is value free and use quantitative methods or mathematical techniques).
The ontological and epistemological positions for the system dynamics that is taken in this study is realism
and (moderately) objective with subjective elements. The view taken is thus that an external real-world
exists (or the modeled system resembles a real-world system) and the causal loop and stock and flow
diagrams are interesting formulations to structure, describe and understand real-world issues such as the
life-cycle burdens and social costs assessment issue investigated in this study. Though no primary valuation
of externalities is conducted in this study, the manner of knowing and construction of this knowledge
(externality costs), can only be grasped mainly through subjective views of the participants, hence the
subjective stance. The methodology is mainly quantitative with qualitative models (causal loop diagrams)
used for developing quantitative models. The model developed in this current study was also validated in
keeping with mainstream system dynamics and due to concerns of value-ladeness. Based on Pruyt’s (2006)
system dynamics paradigms, the system dynamics investigation conducted in this study can therefore be
categorized within the critical pluralist and post-positivist paradigms.
The modelling approach (i.e. system dynamics) thus shares many elements that are consistent with the two
economic schools that underpin this study, for instance, through sharing the same ontological position,
epistemological position (to a certain extent) and the use of quantitative techniques. In addition though,
the proposed modelling approach also offers more features such as non-linear structures, dynamic
structures, experimental approach, transdisciplinarity methods, disequilibrium approach, case study
approach, problem-orientated approach, empirical solutions, complex unitary approach with the ability to
deal with large number of elements and many interactions between elements and confidence based on
model structure over coefficient accuracy, focus on closed loop information feedback structures and focus
not on predictions but on understanding the structure of the system and our assumptions about it.
5.3
Description of inquiry strategy
An inquiry/research strategy refers to the broad approach that a researcher will employ to address the
research problem. In essence, a research strategy provides important links between the research paradigm
and data collection and analysis methods. The choice of a specific inquiry strategy will therefore determine
- 107 -
the data collection methods and indirectly affect the decisions concerning suitable data analysis
techniques. Research strategies are therefore not in themselves methods for collecting/analysing data
(Hiles, 1999).
In the previous section while discussing the research paradigms underpinning this study, it was highlighted
in passing that the goal of the study necessitates a quantitative approach, in this section quantitative
research together with other core characteristics of the current study are discussed in more detail.
Quantitative research have been seen to be a more scientific and objective form of research (Blaxter,
Hughes & Tight, 2006) that is intended to scientifically elucidate phenomenon and issues linked with
phenomenon using numerical data. The research approach attempts exact/specific measurement of
phenomenon by soliciting answers to such questions as how much, how many, how often, who and when
Cooper & Schindler, 2006; Fox & Bayat, 2007). These questions are descriptive as they aim at describing the
phenomena under investigation (descriptive research). Looking at the first two objectives of the study, a
descriptive quantitative strategy of inquiry is well suited to understanding the resource inputs, material
requirements and private costs of building, operating and maintaining a coal-fired power station (objective
1) and to understanding the coal-fuel cycle environmental and societal burdens and costs (objective 2).
Quantitative research may also answer the why and how questions if the aim is to provide explanations for
phenomena (i.e., if the aim is to establish cause(s)-effect(s) of phenomena) (Plack, 2005). The proposed
study is thus also explanatory in nature in that it seeks to develop and validate a system dynamics model
for understanding coal-based power generation and its interactions with resource inputs, private costs,
externalities, externality costs and hence its consequent economic, social and environmental impacts over
its lifetime and fuel cycle (objective 3). The current study is therefore classified as using a descriptive and
explanatory quantitative strategy of inquiry.
The current study can further be classified as pure research (as opposed to applied research). This is
attributable to that the study does not directly focus on solving a specific business/managerial problem
(applied research), it is therefore classified as pure research as it is carried out for the sake of advancing
human knowledge (Saunders, Lewis & Thornhill, 2009) through understanding coal-based power
generation fuel-cycle processes, burdens and social costs. It can also be classified as an empirical study, for
the reason that the researcher re-analyzes existing data (Babbie & Mouton, 2001). More information on
data collection is given in section 5.4.3.
- 108 -
5.4
Research method
The ultimate purpose of this study is to develop a COAL-based Power and Social Cost Assessment
(COALPSCA) Model for understanding coal-based power generation and its interactions with resource
inputs, private costs, externalities, externality costs and hence its consequent economic, social and
environmental impacts over its lifetime and fuel cycle. The research methodology chosen to attain this
objective is system dynamics. In chapter 3, a number of modelling steps to building system dynamics
models were discussed. In this study the system dynamics modelling steps followed when developing the
COALPSCA Model were those suggested by Roberts et al. (1983), Ford (1999) and Sterman (2000) but prior
to commencing with system dynamics modelling two more modelling steps were incorporated. The
methodological framework that was followed in this study is presented in Figure 5.1 and is discussed in the
following sub-sections.
Study scope



Study site
Coal-fuel cycle stages
Societal and environmental burdens/impacts
Data colletion process



Compiling an inventory of materials and resource requirements
Compiling an inventory of private costs
Compiling an inventory of environmental and health burdens
System dynamics modelling





Problem formulation
Dynamic hypothesis formulation
Model formulation (structure and equations)
Model validation
Policy design and evaluation
Figure 5.1: Procedural framework
5.4.1
Study site: Kusile power station and supporting collieries
In chapter 1 it was highlighted that this study will focus on the energy sector and particularly on coal-based
power generation developments. In chapter 2 background information on the South African power industry
was provided. Focusing on the dominant power utility (i.e. Eskom) the existing and future Eskom power
- 109 -
stations were presented in Table 2.2. The utility runs 10 base load power stations with three additional
power stations which have been or are being returned to service. All 13 power stations use conventional
pulverised-coal technology and are fitted with electrostatic precipitators in order to reduce particulate
emissions. On average, the utility's power stations was said to have a generation capacity of 3 400
Megawatt (MW) with a wet re-circulating cooling process and are fitted with precipitators to control dust
(Wassung, 2010). Two new power stations are, however, currently under construction, namely Kusile and
Medupi power stations in the Witbank and Waterberg coalfields, respectively. These two new power
stations are considerably larger than the average power station described above (i.e. with capacities above
4 700MW) and will use a variety of new technologies (e.g. combustion technology, cooling system and
pollution abatement).
Originally based on these information the plan was to select two plants as representative of South Africa’s
power plants – an old plant representing old technology and capacity as described above and one of the
new power stations in particular Kusile power station as it was reckoned that such a plant better represent
the cost structure and societal and environmental impacts of future coal-fired power stations in South
Africa. The unwillingness in the end of Eskom to share data on an old existing plant resulted in the study
focusing on only Kusile power station as a case study.
The Kusile power station plant is situated in the Mpumalanga province, south of the N4 highway between
eMalahleni and Bronkhorstspruit (see Figure 5.2). It is currently under construction and is located on the
Hartbeesfontein and Klipfontein farms. The site covers approximately 5 200 hectares and was previously
used for maize farming and cattle grazing (NINHAM SHAND, 2007; Eskom, 2010b). The plant will consist of
six units, each having 800MW generating capacity, yielding a maximum installed capacity of 4 800MW. The
power station is expected to be fully operational in 2018/19 with the first unit coming into operation
towards the end of 2014 (Eskom, 2012a). It has a projected lifespan of 50 years (Zitholele Consulting, 2011).
Unlike conventional plants, Kusile will use a variety of new technologies with regards to its combustion
technology, cooling system and pollution abatement. As opposed to conventional pulverised-coal
technology that is used in all of Eskom’s plants, Kusile will use supercritical technology. In a pulverised-coal
power plant the coal is firstly fine-crushed into a powder and then fed into a boiler where it is burnt to
create heat. The heat produces steam, which is used to spin turbine(s) to generate electricity. Supercritical
plants on the other hand, form part of the pulverised-coal system but use higher pressures and
temperatures to boost the efficiency of the plant to about 40% or more (Bohlweki Environmental, 2006).
Kusile will also be a dry-cooled station (African Development Bank, 2009; Wassung, 2010) that will be fitted
- 110 -
with FGD technology for removal of SO2 from exhaust flue gases (Eberhard, 2011). Limestone will be used
as feedstock in the FGD system (Eskom communication, 2012).
AAIC proposed New Largo mining
right area (area enclosed by the
“black bold line”)
Figure 5.2: Location of Kusile, Phola coal-processing plant and AAIC mining area
Source: Wolmarans and Medallie (2011)
It is estimated that at full capacity, Kusile will require approximately 17 million tons of coal per annum
(Synergistics Environmental Services & Zitholele Consulting, 2011). The coal will be sourced from the New
Largo coal reserve located east of the Kusile power station and 30km west of eMalahleni. Anglo American
Inyosi Coal (AAIC), a subsidiary of Anglo American, through its proposed coal mine (i.e. New Largo colliery),
will extract coal from the New Largo coal reserve and supply it to Kusile (see Figure 5.1). The New Largo
colliery will be an open-cast coal mine with a minimum raw-coal processing capacity of 12.7 million tons
per annum (Wolmarans & Medallie, 2011). While waiting for the completion of the new colliery, Kusile will
use coal from the Phola coal-processing plant along with supplementary coal from other collieries, for
instance Vlakfontein colliery (Synergistics Environmental Services & Zitholele Consulting, 2011).
The Phola coal-processing plant is located approximately 20km south-east of the Kusile power station (see
Figure 5.1) and is owned by Anglo American and BHP Billiton. It has the capacity to beneficiate 16 million
tons of coal per annum which is mainly exported. The middlings coal (secondary product) will be supplied
- 111 -
to Kusile via a proposed Phola-Kusile coal conveyor of approximately 21km long, depending on the route
chosen and will be designed to transport about 10.4 million tons of coal per annum, over the life of Kusile
(Synergistics Environmental Services & Zitholele Consulting, 2010).
5.4.2
Boundary of the study
This study seeks to provide insight into the life-cycle burdens and social costs of investing in a coal-fired
power station. The main activities/processes/stages in the coal fuel chain are shown in Figure 5.3. These
consist of coal mining, coal processing, coal transportation, plant construction, plant operation, waste
disposal and electricity transmission and use. The manufacture and transportation of material inputs for
these main activities are also important. The main coal fuel cycle phases considered in this current study
are coal mining, coal transportation, manufacture and transportation of main material inputs for
constructing the power plant, plant construction, production and transportation of material inputs for
operating the plant, power plant operation and waste disposal. The study focused therefore on a broader
project scope (life-cyle project wide scope) (see section 5.4.4 under problem formulation for reasons
behind the chosen scope). Power transmission and use are, however, not considered in this study as these
are generic/standard activities for all sources of electricity.
Material inputs
manufacturing &
transportation
Coal mining
Coal processing
Coal transport
Plant
construction
Plant operation
Power
transmission
Waste disposal
Power use
Figure 5.3: Coal-fuel cycle
Source: Own construction
The launching of Kusile power station and its ancillary activities is a source of a number of concerns. For
instance, the area that will house Kusile and the mines has already been declared a priority area for air
quality management (pollution hotspot) and there are salinity problems in the area (Munnik, Hochmann &
Hlabane, 2009). The operation of the coal mines and the power plant will therefore add to the air pollution
health crises in the area. Furthermore, the coal mines will increase water pollution, and disrupt large land
- 112 -
surface areas in addition to causing injuries and fatalities (Mishra, 2009). In chapter 2, a detailed discussion
was provided of the environmental and societal impacts linked with the coal-fuel cycle as reflected by the
literature and Kusile’s EIA report (i.e. NINHAM SHAND, 2007). A summary of some of the environmental
and societal impacts associated with the coal-fuel cycle were presented in Table 2.1.
Informed by the literature, Kusile’s EIA, eMalahleni specific environmental issues, data availability and ease
of quantification the externalities considered in this research are: climate change impacts due to GHG
emissions, human health impacts due to classic air-pollutants emissions, injuries and fatalities, water
consumption, water pollution, and loss of ecosystem services. A detailed block diagram of the life-cyle
stages/processes and externalities of interest in this study are shown in Figure 5.4.

Climate change

Human health

Mortality

Morbidity

Water pollution

Water
Material inputs
transportation
Coal processing
consumption

Material inputs
manufacturing
Coal mining
Coal transportion
Plant construction
Ecosystem

Climate change

Human health
Plant operation
Waste disposal

Climate change

Human health

Mortality

Morbidity

Water use

Climate change

Human health

Mortality

Morbidity

Ecosystem

Water
consumption
Figure 5.4: Coal-fuel cycle stages and externalities studied
Source: Own construction
As shown by Figure 5.4, the considered externalities vary with the life-cyle stages/processes. For instance
concerning transportation, transport-related externalities in a broader sense include such externalities as
human health effects due to emissions of classic air pollutants, global warming due to GHG emissions,
damage to roadways, noise, accidents and congestion (Jorgensen, 2010). The transport externalities that
were instead considered in this current study were those related to fuel use (emissions of classic air
pollutants and GHGs), injuries and deaths. Though impact on roadways by coal haulage trucks is a major
- 113 -
issue of concern on Mpumalanga roads and also though a fraction of the coal requirements of Kusile is
going to be transported by road, damage to roadways do not form part of the transportation externalities
that were considered in this study owing to lack of data.
In order to estimate the damage to roadways linked to Kusile, needed is to know/estimate such cost
upfront (say in cents/ton-km) based on the coal haulage by road that is already happening in Mpumalanga
since Kusile is not operational. Doing so will require among other things truck load data, truck
characteristics, road characteristics and road repair/maintenance cost data, data most of which is unknown
and therefore will necessitate an in-depth survey that will take a while to conduct. It is, however, important
to mention that for the coal hauling roads in Mpumalanga Eskom has began a road upkeep and repair
program. It is reported that Eskom has spent R548 million on the roads in Mpumalanga between 2007 and
2010 plus an extra R100 million on the repairs of potholes (Generation Communication CO 0001 Revision 4,
2011). This of course partly internalizes some of the damage to roadways externality cost.
5.4.3
Data collection process
A wide spectrum of data was collected to address the specific objectives of this study. The data gathering
process followed in this study is summarised below while the specific forms of data linked to each activity
in the data gathering process are discussed in the following sub-sections.
The data gathering process is as follows:

Compiling an inventory of the materials and resources used in the construction, operation and
maintenance of a coal-fired power station (activity 1);

Collection of data regarding the Rand costs of building, operating and maintaining a coal-fired power
station (i.e. private costs - capital cost, fuel cost, maintenance and operating costs) (activity 2);

Compiling an inventory of the environmental and societal burdens associated with the generation of
electricity from coal. Upstream burdens linked with coal mining, material manufacture, plant
construction and waste disposal were also solicited (activity 3); and

Collection of economic valuation data that will assist in the computation of externality costs for
example, monetary values for morbidity and mortality and climate change damage cost (activity 4).
The specific features and sources of the data that were collected are discussed below. During this
discussion where appropriate the specific data requirements are linked to the specific objectives of the
study.
- 114 -
5.4.3.1
Compiling an inventory of materials and resources requirements
The production of electricity from coal requires material and resource inputs from a number of
manufacturers and upstream processes. Collection of data on the resources and material requirements
necessary to build, operate and maintain Kusile coal-fired power station was therefore necessary in order
to partly address objective 1. This entailed conducting an inventory of: (i) materials and resource
requirements for constructing the power plant, for example steel, concrete, aluminium, cooling technology
and pollution control technologies; (ii) materials and resource requirements for operating the power plant,
for example coal requirements, limestone requirements and water use; and (iii) plant operation data that
will assist in the building of the power generation sub-model, for example plant capacity (MW), load factor
(%), coal composition (e.g. ash, carbon and sulphur), coal energy content (MJ/kg), power plant life span
(years) and plant operating hours.
The data requirements for activity one were sourced from secondary and primary sources. The former
included Eskom annual reports, information from Eskom website, published studies, media reports and
project appraisal reports while primary data was also sought from consultation with Eskom personnel
through arranged meetings and by email. Full details of the exact data requirements and corresponding
sources of data are reported in chapter 6 while presenting the sub-models.
5.4.3.2
Compiling an inventory of private costs
The Rand data on the costs of building, operating and maintaining Kusile power station was collected in
order to estimate the private costs of producing coal-based electricity in such a plant (i.e., second activity in
the data gathering process). Among other costs these data included coal cost, limestone cost, capital cost
and water cost. In the end the data was categorised into capital cost, fuel cost and operating and
maintenance costs. Also sourced was data for basing fuel price escalation rates, escalation rates for
operation and maintenance costs and interest rate. Data for this activity was sourced from various sources
including Eskom communication (2012), Eskom reports, and published studies. Full details of the exact data
requirements and corresponding sources of data are reported in chapter 6 while presenting the submodels.
5.4.3.3
Compiling an inventory of environmental and societal burdens
The third activity in the data gathering process concerns the collection of data that concerns the
environmental and societal burdens associated with coal mining and transportation, plant construction,
plant operation and waste disposal (data associated with objective 2). As discussed in chapter 2,
environmental and societal burdens arise at most stages in the coal-to-electricity fuel cycle (materials
- 115 -
manufacturing, mining, transportation, construction, combustion and decommissioning) generating various
hazards that affect the health of human and the environment (Epstein et al., 2011). For example,
generating electricity from coal is indirectly accountable for air pollution in coal mines, occupational
fatalities and injuries in coal mines, air pollution associated with manufacturing material requirements of
the power stations and transportation-related pollution too. Hence an inventory of such stage-wise
burdens was conducted.
An inventory of information and data to enable the estimation of coal-fuel cyle societal and environmental
impacts comprising of climate change impacts due to GHG emissions, human health impacts due to classic
air-pollutants emissions, injuries and fatalities, water consumption, water pollution and loss of ecosystem
services was solicited from various sources including published studies, EIA reports and Eskom annual
reports. The specific data requirements and corresponding sources of data to enable the estimation of coalfuel cyle burdens are reported in chapter 6 while presenting the sub-models.
5.4.3.4
Collection of economic valuation data
The fourth activity in the data gathering process concerned the collection of valuation data that assisted in
the computation of externality costs of the studied burdens (objective 2). Briefly discussed therefore in this
section are the valuation approaches and the associated sources of data that permitted the valuation of the
various burdens studied.
To (i) value morbidity (injuries), two methods can be used, one is based on individual preferences (i.e.
willingness to pay and accept compensation studies) and the other is based on opportunity costs namely
the cost of illness approach (Guh, Xingbao, Poulos, Qi, Jianwen, von Seidlein et al, 2008; Kochi, Donovan,
Champ & Loomis, 2010). Owing to the lack of valuation data in South Africa on individual preference
approaches coupled with that for a developing country like South Africa the individual preference
approaches are complex and contentious (Van Horen, 1997), estimates based on the cost of illness
approach were used. The approach requires the collection of data on actual expenditure on medical
treatment, transportation cost and the opportunity cost of not working (i.e., foregone income because of
lost time at work). Morbidity value estimates were adapted from a study by Van Horen (1997) who valued
injuries using the cost-of-illness approach in South Africa. The values were adjusted to cater for some form
of internalisation and inflation. Detailed explanations of these adjustments are provided in chapter 6 when
discussing the morbidity and fatalities sub-model.
- 116 -
To (ii) value pre-mature deaths3 (mortality) two approaches exist, the human capital approach and
individual preference approach. In the human capital approach lost life is valued by discounting an
individual’s future income stream, usually average GDP is used as a substitute for the individual’s earnings.
The approach is, however, sensitive to the discount rate used. The individual preference approach is the
most preferred approach in the literature. A number of studies have valued pre-mature deaths in
developed countries but primary research is lacking in many developing countries including South Africa.
For this reason, researchers in South Africa such as Van Horen (1997), Turpie, Winkler, Spalding-Flectcher
and Midgley (2002) and Turpie, Winkler and Midgley (2004) adopted values from international studies
derived through revealed preference and adjusted them for GDP per capita and exchange rates, while
Spalding-Fletcher and Matibe (2003) inflated the estimates. The economic value for premature mortality in
this study was adapted from the NEEDS (2007) and NewExt (2004) studies. Adjustments were made to the
values to reflect the disparity of income levels between the European Union (EU) and South Africa and to
cater for inflation and some form of internalisation. Detailed explanations of these adjustments are
provided in chapter 6 when discussing the morbidity and fatalities sub-model.
To (iii) estimate damage cost of climate change, generally two approaches can be used, first a bottom-up
approach which involves conducting a sectoral analysis which determines the economy-wide impact of
climate change and second an approach based on global/national impacts of climate change and its
associated damage costs, also called the social damage cost of carbon (Blignaut, 2011). The bottom up
approach is, however, plagued by difficulty so a number of researchers approximate the social damage cost
of climate change, for example, IPCC (1995), IPCC (2000), Nordhaus, (1993) and Stern (2007), Tol (2005) and
Tol (2009). The social damage cost of climate change on national economies is, among other factors,
influenced by the choice of the discount rate, countries income levels and the distribution of income
amongst and within countries. The aforementioned studies therefore yielded varied estimates.
Locally, a number of studies estimate the climate change damage cost of burning coal for electricity, for
example, Blignaut (2012), Blignaut and King (2002), Spalding-Fetcher and Matibe (2003) and Van Horen
(1997). Unit damage costs of CO2 estimated by Blignaut (2012) were used in this current study to estimate
the climate change damage costs related to coal mining, coal transportation, plant construction and coal
3
There are, however, controversies with this valuation, for example, the ethical problem arising from assigning a fixed
monetary value to human life.
- 117 -
combustion. More detailed explanations of the estimates used are provided in chapter 6 when discussing
the global pollutants sub-model.
To (iv) value human health impacts of classic air pollutants (i.e. damage cost of classic air pollutants)
released during coal mining/transportation/combustion, one can use the impact pathway methodology
developed in the ExternE project – which begins with estimating emissions of pollutants, tracking pollutants
dispersion in the atmosphere (using dispersion modelling), evaluating the exposure of people, crops and
materials to pollutants (quantifying impacts) (AEA Technology Environment, 2005) and then estimating the
damage cost of the classic air pollutants by using the individual preference approach, for example, basing
the valuations of air pollution mortality on the change of life expectancy (i.e. establishing individual’s WTP
for gain in life expectancy to estimate the Value of a Life Year (VOLY) lost by air pollution mortality (NewExt,
2004; AEA Technology Environment (2005); NEEDS, 2007; 2008; 2009) or establishing valuations based on
accidental death or a small change in the probability of dying (mortality risk) (i.e. an individual’s willingness
to pay to reduce/avoid the risk of death (Van Horen, 1997). Basing the valuations of air pollution mortality
on the change of life expectancy, as opposed to a valuations based on accidental death or a small change in
the probability of dying is more advantageous because the approach automatically factor in the constraint
that humans die only once regardless of pollution, it offers a unified framework for time series, cohort and
intervention studies plus directly yields the life expectancy change as a time integral of the observed
mortality rate (Rabl, 2006). In addition, change in life expectancy is further favourable because respondents
during surveys show too much difficulty understanding small probability variations while a change in life
expectancy is well understood (NewExt, 2003).
Another approach is the benefit transfer technique which too can be based on damage costs calculated
based on VOLY or a change in the probability of death. This later approach it involves transferring damage
cost of classic air pollutants from previous studies and adjusting the values for income differences between
countries. It is normally used if local values of health costs are not available. This approach has been used
by the AEA Technology Environment (2005), NEEDS (2007; 2008; 2009) and by Sevenster et al. (2008). All of
the damage costs used by these studies were calculated based on VOLY (explained more in chapter 6 when
discussing the air pollution sub-model). The procedure adopted by the above studies was followed to
approximate the unit damage cost of exposure to classic air pollutants.
To (v) Estimate the economic value of water use in coal mines, during plant construction and during plant
operation, needed is to establish the opportunity cost for water to society when engaging in each of these
activities. Computing such, if time and resources allow for it, is essential because water is highly
- 118 -
underpriced in South Africa so water users rarely pay the full cost of this resource (Inglesi-Lotz & Blignaut,
2012). This issue is further intensified by the scarcity of water in the country (Turton, 2008). The
opportunity cost of water to society when engaging in Kusile coal-fired electricity generation was estimated
by Inglesi-Lotz and Blignaut (2012). The opportunity cost values computed in this study were used to base
the values used in this current study for plant operation and construction and as well as for coal mining as
the coal produced by the proposed coal mine will be 100% dedicated to coal-fired power generation. Some
adjustments were, however, made to the estimates by Inglesi-Lotz and Blignaut (2012) and these are
explained in chapter 6 when discussing the water consumption sub-model.
To (vi) value water pollution damages by coal mines on other water users, the benefit transfer technique
was used. Estimates from a previous local study (i.e. Van Zyl et al., 2002) were used. Van Zyl et al. (2002)
estimated the cost imposed on other water users in the eMalahleni catchment due to water pollution
emanating from various individual industries. This study and its drawbacks are discussed fully in chapter 6
when discussing the water pollution sub-model. For the power generation phase, Eskom claim to operate
under a zero liquid effluent discharge policy, however, as of to date, no formal evaluation of this policy has
been conducted and published (Inglesi-Lotz & Blignaut, 2012). For this reason water pollution linked to the
power station was not be considered.
To (vii) estimate loss of ecosystem services due to coal mining and plant construction, needed was to
establish the opportunity cost of using the land areas occupied by the coal mine and the power station for
these uses. Since the mined area and the power station sites are mainly used for maize cultivation and
grazing (Eskom, 2010b; Ninham, 2007; Wolmarans & Medallie, 2011) the opportunity cost of these uses is
therefore the forgone benefits derived from agricultural production and ecosystem services generated by
grasslands (i.e. carbon sequestration potential and carbon storage of the vegetation cover and soils).
Estimates of the value of maize and that of ecosystem goods and services generated by grasslands
computed in a study by Blignaut et al. (2010) were adapted to this study. Full details of this study and the
modifications to the estimates are discussed fully in chapter 6 when discussing the ecosystem services loss
sub-model. The following section discusses the main research approach that was chosen to attain the last
objective of this study (objective 3).
5.4.4
System dynamics modelling
In assessing the fuel-cyle burdens and social costs of coal-based electricity generation over the lifetime of a
coal-based power plant system dynamics modelling was employed. While various modelling steps to
building system dynamics models exists as disclosed by the literature in chapter 3, the modelling process
- 119 -
followed in this study was informed by those of Roberts et al. (1983), Ford (1999) and Sterman (2000) and it
consisted of problem formulation, dynamic hypothesis formulation, model formulation (structure and
equations), model validation and policy design and evaluation. The Vensim software was used to
conceptualize, construct, simulate and analyze the COALPSCA Model. Causal loop diagrams, stock and flow
diagrams plus simulation modelling are with simplicity and flexibility provided by the Vensim software
(Ventana Systems, 2003). In this section a brief description of the system dynamics modelling process is
provided with extensive details in the following chapters (chapter 6 and 7).
Problem formulation: Problem formulation is the first and most important step in the model building
process. This step embraces a number of activities, among which are defining the problem, identifying key
variables, determining the boundary of the system and establishing the time horizon for the model
(Sterman, 2000). Informed by the literature review conducted in this study, the research problem
addressed in this study was framed and the key variables that needed to be considered were identified.
Based on the purpose of the model and the literature review the boundary and time horizon of the model
were determined. The study focused on a broader project scope (life-cyle project wide scope) while the
time frame was selected such that it was long enough to address the key fuel-cyle burdens and social costs
issues of power generation (i.e. a period of 50 years was selected – more explanations in chapter 6).
An inception meeting was also held with a knowledgeable Eskom worker (Unit head) to introduce the then
proposed work, to request views about the project, the company’s participation and to request the
company’s willingness to provide data that will foresee the attainment of the research goals. Following this
inception meeting and the requests raised to improve the research work, the proposed work was modified
to incorporate the need to study the externality costs of one of Eskom’s existing plants and to explore the
costs and benefits of retrofitting such a plant with new pollution abatement technology as that of Kusile. A
follow up meeting and countless data requests, however, did not yield the data on an existing plant, so the
study reverted to focusing on Kusile power station. It is also important to mention that the original plan for
this research study incorporated a comparative study of the life-cyle burdens and social costs of coal-based
power versus that of wind and solar power generation technologies, but time restraints could not permit
such investigations.
Dynamic hypothesis formulation: This step involves creating a working theory that explains the system’s
dynamic behaviour premised on feedbacks and causal structure of the system (Sterman, 2000). Causal loop
diagrams (i.e. diagrams that capture the structure of the system in a qualitative manner) were formulated
- 120 -
and they displayed the associations between the main variables in the system and feedback loops. An
extensive explanation of this step is provided in chapter 6.
Model formulation: In this step the stock and flow diagrams of the modeled system were constructed and
they provided the quantitative relationships between the variables of the system. A number of sub-models
were yielded by this step and are presented and discussed in chapter 6.
Model validation: This step involves repeated actions of testing and establishing confidence in the model’s
usefulness (Forrester & Senge, 1980; Sterman et al., 1988). Validation of the internal structure of the model
was conducted first followed by behaviour validity because the accuracy of the model behaviour is only
meaningful once adequate confidence on model structure was established prior (Barlas, 1989; Barlas,
1994). Five direct structure validation tests that were introduced by Forrester and Senge (1980) for system
dynamics were performed in this study, namely structure verification, dimensional consistency, boundary
adequacy, extreme condition and parameter verification tests. Behaviour validity on the other hand, seeks
to establish the extent to which the model’s behaviour matches the behaviour of the real system (Barlas,
1996). The behaviour sensitivity test was conducted in this study. Detailed explanations of these tests are
provided in chapter 7.
Policy design and evaluation aimed at alleviating existing problems in the system is central to the
development of system dynamics models. Policy scenarios are crafted based on model results/learning
from the model and from anticipations/expectations in the actual world (Sterman, 2000). A number of
policy scenarios were defined and evaluated with reference to the baseline scenario. A detailed discussion
of the policy design and evaluation step is provided in chapter 7.
5.5
Conclusion
In this chapter explained and motivated were the decisions taken by the researcher regarding the
philosophical beliefs that underpins the study, the strategy of inquiry and the methodological approach
that was employed to achieve the ultimate objective of constructing and validating a system dynamics
model for understanding coal-based power generation and its interactions with resource inputs, private
costs, externalities, externality costs and hence its consequent economic, social and environmental impacts
over its lifetime and fuel cycle. Model validation and policy design and evaluation as part of the system
dynamics modelling process were also addressed. A more thorough discussion of the modelling process is
provided in chapter 6 and 7.
- 121 -
- 122 -
CHAPTER 6:
COAL-BASED POWER AND SOCIAL COST ASSESSMENT
(COALPSCA) MODEL
6.1
Introduction
This chapter discusses and presents the COAL-based Power and Social Cost Assessment (COALPSCA) Model
developed for understanding coal-based power generation and its interactions with resource inputs,
private costs, externalities, externality costs and hence its consequent economic, social and environmental
impacts over its lifetime and fuel cycle. In chapter 3, a number of modelling steps to building system
dynamics models were discussed. In this study the modelling steps followed when developing the
COALPSCA Model were those suggested by Roberts et al. (1983), Ford (1999) and Sterman (2000). These
modelling steps include problem formulation, dynamic hypothesis formulation, model formulation
(structure and equations), model validation and policy design and evaluation. The first three modelling
steps are discussed in this chapter while the remaining two are discussed in the following chapter. Before
the discussion of the modelling steps, the modelling software employed in this research is discussed. This is
the followed by a discussion of the problem formulation step, dynamic hypothesis formulation, model
boundary and model formulation (structure and equations).
6.2
Software used in the modelling
Vensim software was used to conceptualize, construct, simulate and analyze the COALPSCA Model. The
software was specifically developed by Ventana Systems, Inc. for building system dynamics models. Causal
loop diagrams, stock and flow diagrams plus simulation modelling are with simplicity and flexibility
provided by the Vensim software (Ventana Systems, 2003). There are a number of Vensim software
packages, namely Vensim PLE (Personal Learning Edition), PLE Plus, Standard, Professional and DSS in
ascending order of increasing functionality. In this study Vensim PLE Plus software was used.
6.3
Problem formulation
Problem formulation is the first and most fundamental step in the model building process. This step
embraces a number of activities, among which are defining the problem, identifying key variables and
establishing the time horizon for the model (Sterman, 2000). Having a clear purpose of the model is
essential to keeping all those participating in the modelling focused on a single problem and keeping the
modelling process on course. In addition, models are simple representations of real complex systems, so
modellers must refrain from designing a model of the whole system to prevent the model from being as
- 123 -
complex as the system one aims to model. The focus must instead be on a small problem or on models that
address a few issues.
The problem addressed in this study can be framed as follows: South Africa has a number of planned
development projects, including energy projects with coal-based investments. Generally, the
environmental and development planning process, in the form of an EIA have been the main driver of
project development in the country (Hoosen, 2010). The analysis of the quality of EIRs, however, disclosed
that amongst other issues, the more analytical components of the EIRs which form the basis for decision
making are performed poorly for instance with regards to the provision of information pertaining to impact
identification and assessment of key impacts (Sandham et al, 2008; Sandham & Pretorius, 2008; Sandham
et al. 2013). Concerning the assessment of impacts various researchers have expressed inadequate use of
assessment methodologies (Sandham et al., 2010; Sandham & Pretorius, 2008), for instance, causal
networks despite their suitability to fulfill specific principles of EIA practice such as transparency,
integration and being systematic (Perdicoúlis and Glasson, 2006; Wood et al., 2006). Other concerns
pertains to: overemphasis on biophysical environment (Aucamp et al.,2011; Du Pisani & Sandham, 2006);
limited consideration of socio-economic impacts of planned developments (Kruger & Chapman, 2005); no
consideration of the economic value of externalities (Burdge, 2003) despite the importance of considering
externality costs alongside financial costs in decision-making (ATSE, 2009; Icyk, 2006; Roth & Ambs, 2004).
While the employment of causal networks and specifically system dynamics in EIA practice may rectify the
limitation of impact identification and the limited scope of impact assessment, as well as permit
transparency, integration and being systematic, the narrow project-orientation of EIA, however, limit the
scope of impact assessment and hence it hinders a comprehensive assessment of the life-cycle impacts and
social costs of developments, a limitation that becomes more evident in the context of energy generation
projects due to the importance of fuel-cycle impacts and social costs towards informing energy technology
selection. For this reason one could argue that EIA is not broad enough to enable sound energy technology
assessment to inform energy policy formulation and therefore an exploration of technology assessment
was conducted since it is broader than EIA (Berg, 1994; Brooks, 1994).
The energy technology assessment tools and studies, however, are also not without weaknesses for
instance they provide a partial view and partial analysis, respectively, to making informed decisions on the
selection of energy technologies. The reason for this being that the assessment tools and methods tend to
be discipline specific with little to no integrations, with tools often grouped into financial analysis tools,
impact analysis tools, technical performance assessment and so on (Palm & Hansson, 2006), which has
- 124 -
consequently resulted in energy technology studies that exclusively assess these groupings with little/no
integration and with variations in scope and depth. Other concerns pertain to the none consideration of the
economic evaluation of externalities and social costs (Roth & Ambs, 2004) as well as variations in scope and
depth in the assessment of externalities (i.e. limited scope of impact assessment) which make comparing
various energy development project involving (new) technologies difficult. For instance, the studies differ in
terms of the types of externalities they consider, the fuel-cycle stage(s) they investigate, and they do not
factor in the long-standing repercussions of the technologies on the environment and social systems.
These shortcomings highlight the lack of recognized technology assessment frameworks to support energy
policy formulation in the field of environmental and development planning processes (i.e. in both
technology assessment and as well as EIA) and therefore suggests the need for comprehensive assessment
to help inform decision-making on energy developments. Wolstenholme (2003) have supported improving
energy technology assessment through the use of a holistic and integrated approach due to its superior
attributes while Roth and Ambs (2004) advocates the improvement of assessment practices through the
measurement of not only the traditional costs incurred directly by power utilities but costs incurred in the
entire fuel cycle including the conventionally neglected externality costs. This study therefore aspires to
promote proper technology assessment at the extensive project level through improving the environmental
and development planning processes by means of employing a systems approach, namely system dynamics
due to its superior attributes and embedding it within the processes to account for the lifecycle and longterm economic, social and environmental repercussions and social costs of energy development projects.
The current study specifically focuses on coal-based electricity generation as a case study. The primary aim
of this study is therefore to design and validate a system dynamics model for understanding coal-based
power generation and its interactions with resource inputs, private costs, externalities, externality costs
and hence its consequent economic, social and environmental impacts over its lifetime and fuel cycle.
The purpose of developing the model is twofold, firstly is to aid energy decision makers with a tool for
making informed energy supply decisions that consider not only the financial feasibility of power
generation technologies, but also the socio-environmental consequences of the technologies. Secondly, the
model is to aid coal-based power developers4 with a useful tool for detecting the main drivers of the
burdens and costs in the system which should yield vital socio-economic-environmental tradeoff
information that can be beneficial to them.
4
Coal-based power developers refers to the coal-based power plant project developers or companies that run or plan
to develop coal-based power plants (e.g. Eskom).
- 125 -
Another important aspect to be considered is establishing the time horizon for the problem at hand, which
should be long enough to address the key issues. The Kusile coal-fired power station will run for a period of
50 years, therefore this time frame was considered sufficient to allow for most of the activities involved.
While it is true that some of the activities associated with producing power from a coal-fired power plant
are likely going to produce lasting effects that exceed the 50 years, for example water pollution from coal
mining, a balance needed to be trucked and 50 years was considered a reasonable time frame.
6.4
Dynamic Hypothesis formulation
The dynamic hypothesis formulation step involves constructing a working theory that explains the problem.
This theory explains/describes the dynamic behaviour of the system premised on the feedbacks and causal
structure of the system (Sterman, 2000). The causal loop diagram is therefore a diagram that illustrates in a
qualitative manner the linkages and feedback loops of the system and serves as a quick tool for capturing
the hypothesis relating to the basis of dynamics. Model construction tests this hypothesis and it must be
adjusted if evidence from the model or from the real system refutes it (Lane, 2000). The causal loop
diagram displaying the interactions between the key elements and the feedback loops of the modelled
system are shown in Figure 6.1. The interactions associated with coal-based power generation, generation
cost and externality costs are qualitatively expressed in the causal loop diagram.
Each arrow in the diagram shows the influence of one variable on another. The relationships between the
variables may be either positive or negative. Positive polarity designates that an increase (decrease) in the
“cause” variable will increase (decrease) the “effect” variable while negative polarity shows that an
increase (decrease) in the “cause” variable will decrease (increase) the “effect” variable (Sterman, 2000).
The polarity of the feedback loops is also shown in the causal loop diagram and it can be positive or
negative. Self-reinforcing/positive loops (i.e. loops having an even number of “-” signs (or only “+” signs))
amplify change in the system while self-correcting/negative loops (i.e. loops having an uneven number of “” signs) oppose change in the system and attempt to bring the system into equilibrium (Coyle, 1996;
Sterman, 2000).
- 126 -
+
Plant, FGD & waste
disposal GHGs
Plant, FGD & waste
+
disposal classic air
+
pollutants..
+
Loss of ecosystem
services due to plant
Emission
operation
+
factors..
Morbidity and
fatalities..
+
+
+
+
Electricity
production
Plant capacity
+
+
Plant capacity
during construction
+
Desired capacity
after construction
+
Planned investment
Steel, concrete, in plant capacity
+
+
Diesel
use
Distance
travelled
+
+
+
GHGs
Classic air +
pollutants +
Profits
+
Embodied
water
+
Fixed cost
Unit O&M
costs
+
Change in
O&M costs
+
Change in
fuel cost
Fatality & injury rates
per ton of material input
+
Loss of ecosystem
services due to plant
construction
+
Grand externality
costs
+
+
Loss of ecosystem
Classic air
services due to coal
pollutants.
mining
+
Emission
factors
Morbidity and
fatalities
Fatality & injury
rates coal mining
+
Plant & FGD water
consumption
+
+
Fatality &
injury rates
per MWh
Plant & FGD
water consumption
per MWh
+
+
+
Coal mining
+
externality costs
+
+
+
+
+
+
+
+ Plant construction
+ +externality costs
+
+
GHGs.
+
Surface mine water
requirements per
ton of coal
Unit fuel
cost
Social costs
Emission
factors.
Water
consumption
+
+
+
Sulphate
pollution
+
+
+
-
+
+
Private costs
+
Morbidity and
fatalities.
+
+
Plant construction
water requirements
+
Fuel cost
+
Water
consumption.
Sulphate
pollution.
+
O&M costs
-
+
+
+
+
Revenue
+
aluminium
+
+
Coal consumption
Electricity
price
Methane
Diesel use &
electricity use
+
+
+
Plant operation
externality costs
+
+
+
+
+
Change in
damage cost
Unit damage
cost
+
Figure 6.1: Causal loop diagram of the modelled system
There are five main loops (red, green, blue, pink and purple) shown by the diagram. The red reinforcing
feedback loop shows plant capacity to be increased by plant capacity during construction period and
desired functional capacity after construction period, which are in essence in turn positively influenced by
planned investment in plant capacity and profits, respectively. In turn plant capacity stimulates electricity
generation. An increase in electricity generation in turn generates revenues and profits which stimulates
the desired functional capacity after construction and hence the plant capacity after consideration of plant
capacity during the construction phase.
While it is generally true that both the expectations of capacity needs and profitability play an important
role in the decision-making process to invest in electricity generation, however, due to the scope/boundary
of this model (i.e. a life-cycle project wide scope discussed in chapter 5/7) the effect of the forces of
electricity supply and demand on investment decisions was not modelled explicitly as refleted in the above
discussion but the investment in plant capacity was based on exogenously planned investment in plant
capacity by the developer (i.e. “Planned investment in plant capacity” was taken as a proxy for all factors
that affect investment decisions). The final maximum capacity of Kusile at the end of the construction
phase is therefore largely a fixed value (e.g. a plant size of 4 800 MW) that is determined by the size of the
plant the developer planned to construct in the beginning. The amount that the plant manager wishes to
run/operate at a specific point in time after construction (i.e. desired functional capacity after construction)
was modeled as a function of expected profitability, coupled with other factors such as plant operating
hours and the load factor.
- 127 -
The green balancing loops can be called the “private cost loops” and they basically depict the interactions
between electricity production and the private costs of generating electricity. The first green balancing loop
shows electricity production to cause a rise in operation and maintenance (O&M) costs which in turn
increases the private costs of generating power, which then decreases profits, desired functional capacity
after construction, plant capacity and electricity production. The second green balancing loop shows that
electricity generation leads to an increase in coal consumption, which increases the fuel cost, which in turn
increases the private costs of generating power, which then reduces profits. A decrease in profits reduces
the incentive to finance functional capacity after construction, which in turn lowers plant capacity, which
consequently reduces electricity production.
The blue, pink and purple collection of loops can be called the “externality cost loops”. The purple loops
show the interactions between plant capacity construction, plant operation externality costs and profits.
Plant capacity (precisely the construction phase component of plant capacity) is shown by the collection of
purple balancing loops to increase plant construction water requirements, loss of ecosystem services due
to plant construction and a number of burdens (i.e. GHGs, classic air pollutants, sulphate pollution,
morbidity and fatalities and water consumption) linked with the main material input requirements for
constructing the plant (i.e. steel, concrete and aluminium). These burdens together with the likely damage
cost they impose on humans and on the environment, increase the plant construction externality costs
which in turn amplify the grand externality costs and social costs, which then reduce profits, desired
functional capacity after construction, plant capacity and electricity production.
Electricity production is shown by the collection of blue balancing loops to increase plant, FGD and waste
disposal GHG emissions, classic air pollutants, morbidity and fatalities, loss of ecosystem services and water
consumption burdens, which in turn coupled with the likely damage cost imposed by these externalities,
increase plant operation externality costs which intensify the grand externality costs. The grand externality
costs together with the private costs of generating power, in turn raises the social costs, which then reduce
profits, desired functional capacity after construction, plant capacity and electricity production.
The pink loops show the interactions between electricity production, coal mining externality costs and
profits. Electricity production is shown by the collection of pink balancing loops to increase coal
consumption, which in turn increases sulphate pollution, GHG emissions, classic air pollutants, morbidity
and fatalities, water consumption and loss of ecosystem services. These externalities coupled with the likely
damage cost imposed by them on third parties, in turn augment the coal mining externality costs which in
- 128 -
turn amplify the grand externality costs and social costs, which then reduce profits, desired functional
capacity after construction, plant capacity and electricity production. The other remaining reinforcing loops
in the diagram show the dynamics of unit fuel cost, O&M costs and damage cost. The model boundary is
discussed next.
6.5
Model boundary
System dynamics focuses on understanding the structure of the system so as to provide insight into the
behaviour of the system. Accordingly, system dynamics models should include all the important variables
that influence a system’s behaviour. The aim of the model or the problem addressed by the model, would
determine the variables that are to be treated as endogenous, exogenous or excluded. The COALPSCA
Model is a model for understanding the resource requirements, power generation, externalities, private
costs and externality costs of a coal-fired power plant in South Africa, namely Kusile power station. The
model thus seeks to provide insight into the coal-fuel cycle social cost of investing in a coal-fired power
station.
The causal loop diagram presented the interactions between certain important variables of the COALPSCA
Model. Table 6.1 summarizes some of the main endogenous, exogenous and excluded variables. The table
does not provide the whole list of the variables which are reported fully in section 6.6, where the model
equations are discussed. The table indicates that many of the key variables were endogenously generated
while some exogenous variables also drove the model. Some variables were excluded due to lack of data
(e.g. fatalities and injuries linked to plant construction) and the anticipated complication of including such
variables in the model (e.g. ecosystem services lost upstream of the power plant excluding those linked to
the coal mine).
- 129 -
Table 6.1: Endogenous, exogenous and excluded variables
Endogenous variables
Exogenous variables
Gross electricity production
Net electricity production
Operational plant capacity
Coal consumption
Material inputs inventory (coal, steel, water, diesel, etc.)
Pollutant loads (CO2, SO2, CH4, N2O, etc.)
Dry waste
Levelised cost of energy
Levelised externality cost
Levelised social cost
Levelised capital cost
NPV before tax and after tax
Social NPV before tax and after tax
Coal-fuel cycle externality cost of water use
Coal-fuel cycle fatalities and morbidity costs
6.6
Unit water cost
Unit coal cost
Unit limestone cost
Other variable O&M costs
Other FGD O&M costs
Growth rate of the various private costs
Escalation of damage costs
Planned plant capacity
Excluded variables
Ecosystem services loss upstream of plant & coal mine
Plant construction fatalities & injuries
Plant construction water pollution
Plant operation water pollution
Electricity demand
Model formulation: structure and equations
The causal loop diagram presented in section 6.4 displayed the qualitative description of the system, in this
section the stock and flow diagrams of the modelled system are constructed and they provide the
quantitative relationships between the variables of the system. The stocks/levels are denoted by rectangles
and they show accumulations in the system while the flow variables (i.e. inflow and outflow rates) are
denoted by valves and they regulate changes in stocks. Stocks are differential equations and are
mathematically denoted as follows:
t
Stock (t )  Stock (t0 )   Inflow(t )  Outflow(t )dt …………….……….……………………..………………Integral equation
t0
d ( Stock )
 Net change in stock   Inflow(t )  Outflow (t )dt ……………………………….…Differential equation
dt
....
....
....
The integral equation shows that at time t the value of the stock is given by the summation of the stock
value at time t0 and the integral from t0 to t of the change between inflow and outflow rates. The
differential equation shows that at time t the rate at which the stock changes is given by the change
between inflow and outflow rates. Also incorporated into stock and flow diagrams are auxiliary variables
and shadow variables.
The system dynamics model designed in this study for the assessment of coal-based power and its
associated life-cycle private and externality costs is composed of nine sub-models, namely power
- 130 -
generation, generation cost, water consumption, water pollution, morbidity and fatalities, ecosystem
services loss, air pollution, global pollutants and social cost sub-models. The sub-models and associated
equations are presented in the following sections, together with a presentation of the parameters used.
6.6.1
Power generation sub-model
The power generation sub-model, models the generation of electricity at Kusile power station. A plant, that
is currently under construction and will be fully operational in 2018/19. It will run for a period of 50 years
with the first unit becoming operational towards the end of 2014 (Eskom, 2012a). In this study the base
year of the model is 2010 so the first unit becomes operational in 2010 and the plant is fully operational in
2015. The model therefore runs for a period of 50 years from 2010 up until 2060. The structure of
electricity production in such a plant is represented in Figure 6.2. This sub-model consists of four stock
variables, namely plant capacity construction, plant capacity during and after construction as planned,
cumulative gross electricity production and cumulative net electricity production.
Plant construction
time
Plant capacity during and
after construction as planned
Plant capacity construction
New capacity
Capacity
construction start
<Time>
<Unit capital cost>
Capacity
investment
<Time>
Effect of profitability on
desired functional capacity
<Time>
Functional capacity
during construction
Desired functional
capacity after
construction
<Expected
profitability>
Function for effect of
profitability on desired
functional capacity
Gross electricity
production
Coal
consumption
Days per year
Hours per day
Plant operating
hours
<kg to ton>
Planned investment in
plant capacity table
Energy availability
factor
Load factor
Fraction of electricity
consumed internally
1-Internal
consumption rate
MWh/MW*h
Coal energy
content
Heat rate
MWh to kWh
Cumulative gross
electricity production
Net electricity
production
Conversion factor
Cumulative net
electricity production
Figure 6.2: Power generation sub-model stock and flow diagram
Plant capacity construction (PCC, MW) is increased by capacity construction start (CC, MW/Year) and
reduced by new capacity (NC, MW/Year) upon the completion of construction. Mathematically, the
dynamics of plant capacity construction is represented as follows:
PCC(t )  PCC (800)  CC  NC dt ..............................................................................................................(1)
- 131 -
The first component on the right-hand side represents the initial value of PCC, which is 800 MW. Capacity
construction start (CC, MW/Year) is determined by capital investment (CINV, R/Year) divided by the unit
capital cost (KC, R/MW). This is represented as:
CC  CINV / KC .............................................................................................................................................(2)
The capital investment (CINV, R/Year) is a product of exogenously planned investment in plant capacity
(PIPC, MW/Year) and unit capital cost (KC, R/MW). This is denoted as:
CINV  PIPC * KC ……..................................................................................................................................(3)
The new capacity (NC, MW/Year) on the other hand, is determined by plant capacity construction (PCC,
MW) divided by plant construction time (PCt, Year), as follows:
NC  PCC / PCt .............................................................................................................................................(4)
In turn the new capacity (NC, MW/Year) determines plant capacity during and after construction as
planned (PCDAC, MW), as follows:
PCDAC(t )  PCDAC (800)  NC dt ............................................................................................................(5)
The first component on the right-hand side represents the initial value of PCDAC, which is 800 MW. Given
the plant capacity during and after construction as planned (PCDAC, MW), functional capacity during
construction (FCC, MW) and desired functional capacity after construction (DFCA, MW) were computed.
Functional capacity during construction was taken as it was from PCDAC over the construction period. It is
given by the following equation:
FCC(t )  IF THEN ELSE(Time <= 2015, PCDAC,0) ....................................................................................(6)
Which states that FCC is to take values of PCDAC if the time is less or equal to 2015 and otherwise values of
zero (i.e. if the time is different from the specified one). Regarding desired functional capacity after
construction (DFCA, MW), it was modeled as a function of PCDAC (MW) and the effect of profitability on
desired functional capacity (EPC, Dmnl), as follows:
- 132 -
DFCA(t )  IF THEN ELSE(Time <= 2015, 0, PCDAC * EPC) ......................................................................(7)
Which states that FCA is to take values of zero if the time is less or equal to 2015 and otherwise values that
are determined by the product of PCDAC and EPC (i.e. if the time is greater than 2015). The effect of
profitability on desired functional capacity (EPC, Dmnl) on the other hand, was modeled as a function of
expected profitability. A lookup table was used. Lookups or lookup tables/functions permit the modeller to
customize relationships between a variable and its causes. They are useful in the absence of simple
arithmetic equations that describe the relationship between input and output variables. In a lookup table
the input variable alters the output variable through the lookup function, which is normally a non-linear
function (Ventana Systems, 2002). The lookup tables may be informed by experimental data or may be
artificially generated. In this study, the lookup function for effect of profitability on desired functional
capacity was informed by expected hypothetical behavior.
Figure 6.3 presents the lookup function for effect of profitability on desired functional capacity. The X-axis
denotes expected profitability while the Y-axis represents the effect on desired functional capacity.
Expected profitability is a shadow variable in the power generation sub-model so it is elaborated on in the
social cost sub-model in section 6.6.9. The input variable (i.e. expected profitability) was normalized in
order to make certain that both the input and output variables were independent of the units of measure
of other variables in the model (i.e. dimensionless)). As an illustration, the function states that when the
price of electricity is equal to the unit cost of production, the expected profit is zero and hence the effect
on desired functional capacity is 0.75.
- 133 -
Baseline
Function for effect of profitability on desired functional capacity
2
1.5
1
.5
0
-1
-0.50
0
-X-
0.50
1
Figure 6.3: Lookup function for effect of profitability on desired functional capacity
Given the functional capacity during construction (FCC, MW), desired functional capacity after construction
(DFCA, MW) coupled with plant operating hours (POH, h/Year) and the load factor (LF, Dmnl), gross
electricity production (EP, MWh/Year) is estimated as follows:
GEP  ( FCC * POH )  ( DFCA * POH ) * LF  ................................................................................................(8)
Where, POH (in h/Year) is given by the product of the number of days per year (DPY, Day/Year), hours per
day (HPD, h/Day) and energy availability factor (Dmnl). The energy availability factor is the amount of time
that the power plant is able to generate energy over some time period, divided by the amount of the time
in the period or is simply the percentage of the time that the power plant is able to provide energy to the
grid. The plants energy availability factor is mainly a factor of its reliability and the periodic maintain it
requires. All else being equal, power plants that are operated less regularly have higher energy availability
factors for the reason that they require less maintenance. The load factor on the other hand, refers to the
ratio of power produced by a power plant over the theoretical maximum it could produce at full capacity
over a time period (e.g. hours, days or weeks or yearly). It is a key variable here as it is important for
predicting the amount of power a plant can produce. The load factor is also a key concept for generation
cost estimates. The higher the load factor the lower the generation cost per MWh (Lopez, 2006).
Gross electricity production is in turn an inflow to cumulative gross electricity production, which is a third
stock of the power generation sub-model. Gross electricity production gives rise to net electricity
- 134 -
production (NEP, MWh/Year), once the fraction of electricity consumed internally by the plant is subtracted
(i.e. 1 – Internal consumption rate (Dmnl)). Net electricity production is thus represented as:
NEP (t )  GEP * (1  Internal consumption rate) ..............................................................................................(9)
Where 1 - Internal consumption rate is 1 minus the fraction of electricity consumed internally by the plant.
Cumulative net electricity production, which is the fourth stock, is therefore an accumulation of net
electricity production.
Finally, gross electricity production also determines the amount of coal consumption (CConsump, ton/Year)
coupled with data on coal energy content (CEC, MJ/kg) and heat rate (HR, MJ/kWh). Coal consumption is
given by:
CConsump  ((GEP * MWhtoKWh ) * HR) / CEC  / kgtoton ........................................................................(10)
The parameters used in the power generation sub-model are presented in Table 6.2. Input variables taken
from other sub-models (red variables) are not shown in this table. The complete respective equations of
the power generation sub-model are presented in Appendix A.
Table 6.2: Parameters used in the power generation sub-model
Parameter
Coal energy content
Days per year
Energy availability factor
Fraction of electricity consumed internally
Heat rate
Hours per day
Load factor
Planned plant construction table
6.6.2
Units
Baseline value
MJ/kg
Day/Year
Dmnl
Dmnl/Year
MJ/kWh
h/Day
Dmnl/Year
MW/Year
19.22
365
0.94
0.075
9.769
24
0.9
Time series
Data source
Eskom, 2010a.
Eskom communication, 2012.
Eskom communication, 2012; Eskom, 2012b.
Eskom communication, 2012.
EPRI, 2010.
Eskom communication, 2012.
NINHAM SHAND. 2007
Calculated based on Eskom (2012b).
Generation cost sub-model
The generation cost sub-model focuses on the private costs of electricity generation at Kusile power
station, specifically the cost per MWh incurred by the electricity producing entity (i.e. Eskom) to produce
electricity over the lifespan of the investment technology (i.e. LCOE). Computing the LCOE thus requires
both the cost of energy and power generated by an energy generating system to be assessed over the
lifetime of the energy generating system (Bandyopadhyay et al., 2008; Zweibel, Mason & Fthenakis, 2008).
- 135 -
The approach, though somehow synonymous with life cycle cost analysis, is said to provide the best
comparison between energy technologies because it takes into account not only the lifetime cost but also
the lifetime energy production associated with an energy system (Bandyopadhyay et al., 2008; Darling et
al., 2011).
Owing to the LCOE’s focus on the lifetime of the power generating system when assessing costs and power
(Zweibel et al., 2008), the future time series of expenditures and revenues have to be discounted to their
present values, by applying a discount rate (Hearps, McConnell, Sandiford & Dargaville, 2011). Accordingly,
the LCOE (R/kWh) is the ratio of total lifespan expenses to total anticipated output (i.e. electricity),
expressed in present value. Equations (11), (12a), (12b) and (12c) show the general calculation method for
the LCOE. Equation (11) shows the equivalence of the present value of the summation of discounted
revenues and costs. The calculation begins at t  0 so as to incorporate the initial cost at the start of the
first year, or alternatively the initial cost can be placed outside of the summation and then t begins at 1
t  1 .
 E  p
T
t
t 0
elect
  C  1  r  ..................................................................................................(11)
 1  r  t 
T
t
t
t 0
Where: Et is the energy generated in year t ; pelect is the price of electricity; Ct is the cost in time t ; 1  r t
is the discount factor in year t .
The sum total of the present values of the cash flows is zero, hence the NPV of the project is zero (Hearps
et al., 2011), meaning an investor breaks-even on the project. One approach therefore to calculating the
LCOE is to assume a discount rate and then to solve for the sale price of power that yields a zero NPV for
the project. Equations (12a) and (12b) therefore rearrange equation (11) and show the LCOE to be equal to
the price of electricity that equates the two discounted cash flows. The equivalence of the LCOE and the
electricity price is based on the assumptions of a stable and non-varying discount rate (r ) and electricity
price over the lifetime of the energy generating system (International Energy Agency, 2010). Equation (12c)
shows the LCOE as a ratio of the sum of the present value costs divided by the total amount of electricity
adjusted for its economic time value.
The division of each year’s physical output by the time preference factor in equations (12a), (12b), and
(12c), does not, however, seem to make intuitive sense, for the reason that physical units neither change
magnitude over time, nor pay interest. While it is true that a unit of electricity does not pay interest, it
- 136 -
indeed produces a revenue stream that does pay interest, and also a unit of electricity (MWh) generated
this year does not have the same economic value as a unit of electricity produced in the following year,
because it can be invested into projects that grow our wealth. What is discounted, in essence, is the value
of output which is the amount of electricity generated multiplied by its price. It is only after the
rearrangement of equation (11) to equations (12a), (12b) and (12c) that it seems as if physical production is
being discounted. The necessary discounting of the electricity price in equation (11), leads to the apparent
discount of physical output in equation (12a), (12b) and (12c). The substitution of physical production for its
price (i.e. economic value) in equation (12a) (from equation (11)) is possible because the nominal and
undiscounted price does not change over the lifetime of the energy generating system. The correct time
value of the revenue stream is now therefore obtained by adjusting physical production as opposed to
price with the correct discount factor. It is in effect, not physical production as such that is discounted but
its economic value (International Energy Agency, 2010). Equation 12c has been used by a number of
researchers to compute LCOE including Zweibel et al. (2008), IRP (2010), Branker, Pathak and Pearce (2011)
and Hernandez-Moro and Martnez-Duart (2013).
 C  1  r  
T
t
t
pelect 
t 0

 T

Et  1  r  t

 t 0





.......................................................................................................................(12a)
 C  1  r  
T
t
t
t 0
LCOE  pelect 

 T

Et  1  r  t

 t 0





.............................................................................................................(12b)
 C  1  r  
T
t
t
LCOE 
t 0

 T

Et  1  r  t

 t 0





.........................................................................................................................(12c)
Equation 9c’s approach to computing the LCOE was used in this study. But before presenting the
generation cost sub-model structure, it is import that one address one of the important parameters in the
LCOE formula, namely the discount rate. A discount rate is used in the computation of present values of
future cash flows. It is fundamental whenever the cash flows accrue at different time frames and especially
over long periods. Discounting stems from that a dollar that is received now is worth more than a dollar
- 137 -
that is received in the future. Therefore choosing a discount rate is synonymous with choosing future dollar
values or with putting relative values on cash flow estimates occurring in various time periods (Harrison,
2010). A higher discount rate makes future cash flows count for less, so the present value of future cash
flows becomes smaller, while by not discounting, one assumes a zero discount rate and basically implies
that a dollar received in the future (no matter how distant the future) carries the same value as a dollar
received today.
In spite of the importance of discounting there is, however, little consensus over the suitable discount rate
to employ for computing present values. The literature recommends various estimates of discount rates in
various countries mainly ranging between estimates of 1% to 15%, with developed countries and
environmental projects using estimates lying in the lower end, while developing countries generally use
higher estimates. Two main schools of thought to discount rate selection have largely influenced the rates
used, namely the descriptive and prescriptive approaches. The descriptive approach offers a discount rate
that is centered on the opportunity cost of capital invested in the project; it centers on efficiency criterion.
On the other hand, the prescriptive approach or normative approach offers a discount rate that is swayed
by ethical views about intergenerational equity. It mixes equity and efficiency factors and is encouraged
whenever projects have an influence on future generations (Harrison, 2010).
In line with the main schools of thought there are two focal groups of approaches to deriving the discount
rate, namely market and non-market rates. The market rates are based on market interest rates and they
include the (i) marginal rate of return on private investment, also called investment rate, private sector
rate, before-tax rate of return or producer rate, (ii) social marginal rate of time preference, also called
consumption rate, after-tax rate of return or consumer rate, (iii) weighted average rate of the investment
and consumption rates, and (iv) government borrowing rate or government bond rate which is a risk-free
rate (Boardman, Greenberg, Vining & Weimer, 2006).
The social time preference rate (STPR) is an example of a non-market based rate and it accounts for the
value that society attaches to current consumption versus future consumption (Boardman et al., 2006;
European Commission, 2008). It is generally estimated as STPR  p  g , where: p denotes the rate at
which future consumption is discounted over present consumption by individuals;  denotes the elasticity
of marginal utility of consumption while g denotes the long-run rate of growth of per capita consumption
(HM Treasury, 2003). Proponents of STPR, argue against basing the discount rate on market variables due
to the market being imperfect, consumers being irrational, and the prevalence of information asymmetry
and other distortions (Rozylow, 2013). The STPR is, however, in turn criticized - for difficulty in estimating
- 138 -
the parameter g , including that there might be a flaw in the estimation of the growth rate because
national income might not precisely measure consumption. Another criticism stems from the judgments
about intergenerational equality on the parameters  and g which might be wrong (Boardman et al.,
2006). A number of researchers or organizations have suggested using the STPR, including Evans (2005),
European Commission (2008) and Rozylow, (2013).
In South Africa, the National Treasury does not stipulate a discount rate for Public Private Partnership (PPP)
projects. Various institutions have therefore used various estimates, with some regarding an appropriate
discount rate to be the same as the government bond yield (which is considered a risk-free rate) with a
maturity matching the PPP project length (Kelman, 2008), the risk-adjusted cost of capital to government
and the nominal government bond yield rate over the project term (National Treasury, 2004). The use of
the government bond yield has been supported for the main reason that it reflects at any time period,
government cost of funds. A discount rate of 8% was recommended by the Department of Environmental
Affairs and Tourism (2004) to discount costs and benefits that accrue in the future in cost-benefit studies,
with sensitivity analysis carried out at 3% and 10%. On the other hand, the EPRI (2010), in its assessment of
power generation technologies in South Africa, used an 8.6% real before tax weighted average cost of
capital (after tax was at 7.4%), while sensitivity analysis was carried out at 4% (after tax at 3.2%). The IRP
(2011), in its integrated energy plan for electricity in the country, used a real discount rate of 8% which was
signed off by the National Treasury as per its use by the National Energy Regulator of South Africa (NERSA)
in the utility price application. In line with the above review on power generation technologies, the current
study adopts an 8% discount rate in the baseline model and in its response to the uncertainty regarding the
appropriate discount rate, sensitivity analysis is conducted at 4%, 6%, 10%, 12% and 15%.
Now, looking at the sub-model, the structure of the generation cost sub-model is represented in Figure 6.4.
This sub-model consists of fourteen stock variables, six of which signify the main components of the
generation cost, namely cumulative present value (CPV) fuel cost, CPV variable O&M costs, CPV fixed O&M
costs, CPV FGD operation cost, CPV net electricity production and cumulative capital cost escalated (capital
cost escalated, though included as a stock variable, was not discounted (Branker et al., 2011)). These six
main stocks are key inputs into the computation of the LCOE. The other eight stocks in gold text, namely
unit capital cost, unit water cost, unit limestone cost, unit coal cost, unit transport cost, other variable O&M
costs, fixed O&M costs and other FGD O&M costs, denote the unit cost components behind the key stocks
and together with some other variables give rise to the six key stocks.
- 139 -
Fixed O&M cost
escalation
Plant water cost
Fixed O&M
Plant water
Unit water cost
Plant water
costs
consumption
Change in water cost
PV fixed O&M
consumption per MWh
Other variable
cost
Water cost
<Year>
<Gross electricity
Other O&M
O&M costs year
escalation
Fixed O&M
Year
production>
costs escalation
Unit coal cost
costs year
Variable O&M
Change in coal
Cumulative PV fixed
Other
variable
costs
cost
Change
in
other
O&M
costs
<Present value
O&M costs
<Cumulative PV net
PV fixed O&M
Coal cost
variable O&M costs
factor>
<Coal
electrity production>
costs
escalation
consumption>
Cumulative PV variable
O&M costs
Levelised fixed
PV variable
Levelised variable
Coal cost
O&M costs
O&M costs
<Present value
O&M costs
Unit transport
factor>
Cumulative PV fuel
cost
Change in
Year of cost
cost
Levelised
fuel
transport cost
Levelised
PV fuel cost
cost
<Time>
O&M costs
Limestone
Transport cost
Cumulative PV net
Present value
transportation distance
escalation
electrity production
<Conversion
PV net electricity
factor
factor>
production
Limestone
Limestone
Levelised cost of energy
<Net electricity
transport cost
consumption per hour
production>
Discount rate
Limestone cost
Limestone
consumption
<Plant operating
hours>
Change in
limestone cost
Cumulative PV FGD
operation cost
PV FGD
operation cost
Limestone
consumption cost
Unit limestone cost
Limestone cost
escalation
FGD operation
cost
FGD water cost
FGD water consumption
per MWh
<Capacity
investment>
Other FGD O&M
costs year
<Unit water
cost>
FGD water
consumption
Levelised FGD
operaton cost
Other FGD
O&M costs
<Year>
<Gross electricity
production>
Capital
investment rate
Change in other FGD
O&M cost
<Other O&M costs
escalation>
Capital cost
escalation table
Cumulative capital
cost escalated
Unit capital cost
Change in capital
cost
Levelised
capital cost
<Cumulative PV net
electrity production>
Capital cost
Overnight cost
<Time>
Plant size
Figure 6.4: Generation cost sub-model stock and flow diagram
The eight subsidiary stocks, namely unit coal cost (UCC, R/ton), unit water cost (UWC, R/m3), unit transport
cost (UTC, R/ton/km), unit limestone cost (ULC, R/ton), other variable O&M costs (OVO & MC,R), fixed
O&M costs (FO & MC,R), other FGD O&M costs (OFGDO & MC,R) and unit capital cost (UKC, R/MW), have a
relatively similar structure and are influenced by exogenous fractional rate. For example, the unit coal cost
(UCC, R/ton) is influenced by the change in coal cost (∆CC, R/ton/Year), which is in turn determined by coal
cost escalation. The equation for unit coal cost is given by:
UCC (t )  UCC (210)  UCC dt ...............................................................................................................(13)
In a similar manner, the other subsidiary stocks are estimated as follows:
UWC(t )  UWC(0.7)  UWC dt ..............................................................................................................(14)
UTC(t )  UTC(1.22)  UTCdt ....................................................................................................................(15)
ULC(t )  ULC(335)  ULCdt ..................................................................................................................(16)
OVO & MC (t )  OVO & MC (7.26e + 008)  OVO & MC dt .............................................................................(17)
FO &MC (t ) OO&MC (8.93e +008)  FO &MC dt .......................................................................................(18)
OFGDO & MC (t )  OFGDO & MC (1.705e + 008)  FGDO & MC dt ……………..............................................(19)
UKC (t )  UKC (Capital cos t / Plantsize )  UKC dt .................................................................................(20)
- 140 -
In turn the unit coal cost (UCC, R/ton) together with the amount of coal consumption (ton/Year),
determines the coal cost (CC, R/Year). The coal cost together with the present value factor (PVF, Dmnl),
determines the present value fuel cost (PVFC, R/Year), which is an inflow to the cumulative PV fuel cost
(CPVFC, R) which is the ninth stock, given by:
CPVFC (t )  CPVFC (0)  PVFC dt ............................................................................................................(21)
The cumulative PV fuel cost (R) coupled with the cumulative PV net electricity production (CPVNEP, MWh)
determines the levelised fuel cost (LFC, R/MWh), as follows:
LFC  CPVFC / CPVNEP ............................................................................................................................(22)
The plant water cost (PWC, R/Year) and other variable O&M costs year (OVO & MCY, R/Year) determine the
variable O&M costs (VO & MC, R/Year), which together with the present value factor, determine the
present value variable O&M costs (PVVO & MC, R/Year) which is an inflow to the cumulative PV variable
O&M costs (CPVVO & MC, R) which is the tenth stock, given by:
CPVVO & MC (t )  CPVVO & MC (0)  PVVO & MC dt ..................................................................................(23)
In turn the cumulative PV variable O&M costs (R) together with the cumulative PV net electricity
production (CPVNEP, MWh) determine the levelised variable O&M costs (LVO&MC, R/MWh), as follows:
LVO & MC  CPVO & MC / CPVNEP ....................................................................................................(24)
The fixed O&M costs year (FO&MCY, in R/Year), together with the present value factor, determine the
present value fixed O&M costs (PVFO&MC, R/Year) which is an inflow to the cumulative PV fixed O&M
costs (CPVFO&MC, R) which is the eleventh stock, given by:
CPVFO & MC (t )  CPVFO & MC (0)  PVFO & MC dt ......................................................................(25)
In turn the CPVFO&MC (R) together with the cumulative PV net electricity production (CPVNEP, MWh)
determine the levelised fixed O&M costs (LFO&MC, R/MWh), as follows:
LFO & MC  CPVFO & MC / CPVNEP .....................................................................................................(26)
- 141 -
The levelised variable O&M costs and levelised fixed O&M costs represent the direct O&M costs linked to
plant operation (excluding the FGD system). These levelised costs are summed to arrive at the levelised
O&M costs (LO&MC, R/MWh).
The FGD operation cost (FGDOC, R/Year) which is composed of limestone cost, FGD water cost and other
FGD O&M costs year, coupled with the present value factor, determines the present value FGD operation
cost (PVFGDOC, R/Year), which is an inflow to the cumulative PV FGD operation cost (CPVFGDOC, R) which
is the twelfth stock, given by:
CPVFGDOC (t )  CPVFGDOC (0)  PVFGDOC dt …….........................................................................(27)
In turn the CPVFGDOC (R), coupled with cumulative PV net electricity production (CPVNEP, MWh),
determines the levelised FGD operation cost (LFGDOC, R/MWh), as follows:
LFGDOC  CPVFGDOC / CPVNEP .....................................................................................................(28)
The thirteenth stock is cumulative PV net electricity production (CPVNEP, MWh) and is determined by the
present value net electricity production (PVNEP, MWh/Year) which is a function of net electricity
production (MWh/Year) and the present value factor (Dmnl). Cumulative PV net electricity production
(CPVNEP, MWh) is given by:
CPVNET (t )  CPVNET (1)  PVNEP dt .................................................................................................(29)
The last stock (14th) is cumulative capital cost escalated (CKCE, R) and is determined by capital investment
rate (KIR, R/Year) which in turn is a function of capital investment. Cumulative capital cost escalated (CKCE,
R) is estimated as follows:
CKCE (t )  CKCE (0)  KIRdt .................................................................................................................(30)
Now, concerning the levelised capital cost (LKC, R/MWh), it is determined by cumulative capital cost
escalated (KCE, R) and cumulative PV net electricity production (CPVNEP, MWh), as follows:
LKC  CKCE / CPVNEP ...........................................................................................................................(31)
- 142 -
Finally, the LCOE (R/MWh) is mathematically represented as a summation of the levelised capital cost (LKC,
R/MWh), levelised fuel cost (LFC, R/MWh), levelised O&M cost (LO&MC, R/MWh) and levelised FGD
operation cost (LFGDOC, R/MWh), as follows:
LCOE  LKC  LFC  O & MC  LFGDOC ...............................................................................................(32)
The parameters used in the generation cost sub-model are presented in Table 6.3. The complete respective
equations of the generation cost sub-model are presented in Appendix A.
Table 6.3: Parameters used in the generation cost sub-model
Parameters
Capital cost
Capital cost escalation table
Coal cost escalation
FGD water consumption per MWh
Limestone transportation distance
Limestone consumption per hour
Limestone cost escalation
Plant O&M costs (both fixed & variable)
Plant size
Transport cost escalation
Unit capital cost
Unit coal cost
Unit limestone cost
Unit transport cost
Unit water cost
Water consumption per MWh
Water cost escalation
6.6.3
Units
Baseline value
R
Dmnl/Year
Dmnl/Year
3
m /MWh
Km
ton/h
Dmnl/Year
%
MW
Dmnl/Year
R/MW
R/ton
R/ton
R/ton/km
R/m3
m3/MWh
Dmnl/Year
R118.5b
Time series
0.001
0.145
100
70
0.001
46 of fuel cost
4800
0.001
R118.5b ÷ 4800MW
210
335
1.22
0.7
0.2
0.001
Data source
Eskom, 2012a.
Assumption.
Assumption.
NINHAM SHAND, 2007.
Assumption (round trip).
Eskom communication, 2012.
Assumption.
BDFM Publishers, 2013b.
Integrated report, 2011 & 2012.
Assumption.
Eskom, 2012a.
Eskom communication, 2012.
Calculated based on Souza et al (2002).
Calculated based on Botes (2006).
Assumption.
Eskom communication, 2012.
Assumption.
Morbidity and fatalities sub-model
The morbidity and fatalities sub-model focuses on human injuries and deaths that arise in the coal-fuel
cycle, namely, during coal mining, construction and power generation. Before discussing the sub-model,
the address of accidents in externality analysis is discussed first. Since accidents are a complicated topic in
externality analysis, care needs to be taken to ensure that what is measured are externality costs. If
workers are receiving an occupational risk premium in their wage rate and are voluntarily choosing to bear
the risk, there is no externality. So workers are fully compensated for the risk of accidents they are exposed
to if such cost is fully internalized through the wage rate. The high frequency of wage-related strikes in the
mining/energy sector in South Africa, however, indicates that workers are not happy with their wages and
therefore that they are barely receiving an occupational risk premium in their wage rate. In addition, the
wage-related strikes coupled with the high level of unemployment rate in the country signify that it is very
- 143 -
unlikely that workers are voluntarily choosing to bear the occupational risk but instead that they are rather
forced to bear it as they need to provide for themselves and their families.
On another note, accidents that are suffered by employees involved in the coal-fuel cycle may also be
internalized by way of ex post compensation to relatives of the victim. In this regard, in the South African
case there are two legislations that govern mining health compensation with different benefits, namely the
compensation for occupational injuries and diseases act (COIDA) and occupational diseases in mines and
works Act (ODIMWA). The ODIMWA is only applicable to lung diseases in the mining industry (i.e. covers
permanent incurable conditions) while COIDA applies to all injuries in and beyond the mining industry plus
diseases not covered by ODIMWA (i.e. covers incurables and curable conditions). Only a once-off payment
is provided by the ODIMWA (i.e. a lump sum payment based on a statutory formula is paid to an individual
who becomes disabled) whereas COIDA has a lifetime pension (i.e. the dependents of a worker are entitled
to 75% of the worker’s salary depending on the level of disability (31% - 100%) in the event that the worker
dies due to occupational-related injuries, for instance the widow is paid till the widow dies whereas
children benefit until they become self-supporting (United States Agency International Development
(USAID), 2008)
On the other hand, medical expenses are only covered for two years by the COIDA whereas ODIMWA
prolongs the responsibility over the worker’s life. In addition, a number of serious problems have been
raised with regards to the acts especially the ODIMWA compared to the COIDA, including poor service
delivery (an insignificant proportion of certified disabled miners receive successful compensation), delays in
compensation payment, virtually no revisions of compensation figures (not even inflationary alterations).
There have also been calls to harmonize the two acts into an integrated compensation system (USAID,
2008). Inadequate and inequitable compensation arrangement therefore characterizes the compensation
legislation.
In addition to occupational accidents there are coal-fuel chain accidents that affect the general public. Nonoccupational accidents in the fuel chain are mostly involuntarily suffered by the general public though to
certain degrees the costs/losses from accidents maybe reduced by individuals through two different
protective measures, namely mitigation measures and through buying insurance. Economic theory
recommends internalizing the cost of accidents through liability insurance (Kopp & Prud’homme, 2007).
Liability insurers pay a combination of annuities and once-off payments related to wage losses and medical
costs for injuries and a combination of annuities and once-off payments for fatalities to the family
(European Commission, 2005).
- 144 -
In the light of this discussion on occupational and public accidents, it is evident that some degree of
internalization is to be expected but the absence of hard data in South Africa with which to approximate
and validate the percentage of internalization rendered the researcher to base the internalization risk on
the study by the European Commission (2005). In the European Commission externality study, occupational
and non-occupational accidents in the fuel cycle were estimated for both Organisation for Economic Cooperation and Development (OECD) countries and non-OECD countries. The internalization estimates used
in the study indicated that occupational risk is recognized as largely internalized in industrialized economies
while a lower degree of internalization is expected in non-OECD countries. For occupational accidentrelated mortality, 70% (low), 80% (central) and 100% (high) ranges of internalization were assumed for
OECD countries while 0% (low), 50% (central) and 100% (high) were assumed for non-OECD countries. On
the other hand, for non-occupational accident-related mortality, 30% (low), 50% (central) and 70% (high)
ranges of internalization were assumed for OECD countries while 0% (low), 20% (central) and 50% (high)
were assumed for non-OECD countries. The internalization estimates used in the European Commission
study further disclose that non-occupational accident-related mortality impacts are recognized as
substantially externalized than occupational accident-related mortality. No form of risk internalization was
made with regards to injuries in the European Commission study but instead injuries were taken to be 1%
to 13% of mortality values (i.e. value(s) of prevented fatality).
In this current study the unit morbidity and mortality values used (i.e. the values of treating injuries
suffered by occupational personnel and the general public, and the economic values for premature
mortality, respectively) were based on the study by van Horen (1997), and NEEDS (2007) and NewExt
(2004) studies, respectively. Van Horen (1997) valued injuries using the cost-of-illness approach. Estimates
of medical treatment costs and the opportunity costs of not working were obtained through discussions
with public health practitioners. Low, high and central estimates were computed in the Van Horen study
and were adjusted for inflation in the current study. It was, however, not possible to gather whether or not
any form of internalization of costs for injuries were incorporated in the study as the book providing indepth explanation to the estimates reported in Van Horen (1997) is out of print. Concerning mortality, the
values for mortality were obtained through the adjustment of valuations of changed life expectancy,
obtained from the NEEDS (2007) and NewExt (2004) studies. The adjustments were conducted to reflect
the disparity of income levels between the European Union (EU) and South Africa (through multiplying the
unit cost determined in the EU by the ratio of purchasing power parity gross domestic product (PPP GDP)
between the two nations), and to cater for inflation.
- 145 -
In the absence of internalization data in the South African case, morbidity and mortality values in this study
were adjusted with an average of 0% (low), 35% (central) and 50% (high) ranges of internalization in line
with the average assumed internalization of occupational and non-occupational accident for non-OECD
countries reported in the European Commission (2005) study. The internalization estimates used in this
current study therefore imply that 50%, 65% and 100% of our low, central and high estimates for mortality
and morbidity were assumed to be externalized. Accordingly, the low and central values for mortality and
morbidity were both adjusted to reflect 50% and 65% externality while the high estimates were not altered.
Having discussed the above, attention is now reverted to the morbidity and fatalities sub-model. Figure 6.5
represents the structure of this sub-model which consists of two stock variables, namely unit morbidity
value and unit mortality value. The unit morbidity value (UMV, R/person) refers to the value of treating
injuries suffered by occupational personnel and the general public. As explained earlier, the values for
morbidity (low, high and central estimates) were adapted from a study by Van Horen (1997) and were
adjusted for inflation and some form of internalization as explained above. The baseline value used in the
modelling conducted in this study is the central estimate adjusted for inflation and internalization. The unit
value for morbidity is determined by the change in morbidity value (∆UMV, R/person/Year), which is in turn
altered by escalation of damage cost (Dmnl/Year), which is estimated at the growth rate of population. The
unit morbidity value (UMV, R/person) is denoted as follows:
UMV (t )  UMV (25434)  UMV dt .........................................................................................................(33)
- 146 -
<Coal
consumption>
Fatalities per million
tons of coal mined
<Escalation of
damage cost>
Unit mortality value
Change in
mortality value
Fatalities per
MWh
Deaths power
generation
<Unit morbidity
value>
<Gross electricity
production>
Injuries power
generation
Injury rate per
MWh
Fatality cost due to
limestone production
Fatalities & mortality costs
limestone production (FGD)
Morbidity cost due to
limestone production
Deaths from
coal mining
<Escalation of
damage cost>
Fatalities & morbidity
costs (coal mining)
Fatalities cost
(coal mining)
Morbidity cost
(coal mining)
Injuries per million
tons of limestone
<Unit morbidity
value>
Limestone in
million tons
<Tons to million
tons>
<Limestone
consumption>
<Steel>
Steel in
million tons
Morbidity cost due to
material inputs production
Fatalities & morbidity
costs (construction)
Injuries limestone
production
Fatalities per million
tons of limestone
Injuries material
inputs production
<Unit morbidity
value>
Coal-fuel cycle
fatalities & morbidity
costs
Injuries per million
tons of Al
Injuries per million
tons of concrete
Unit morbidity value
Change in
morbidity value
Fatalities & morbidity
costs (power generation)
Deaths limestone
production
<Unit mortality
value>
Injuries from
coal mining
Coal consumption in
million tons
Fatality cost power
generation
Morbidity cost
power generation
Injuries per million
tons of coal mined
Tons to million
tons
<Al>
Concrete in
million tons
Al in million
tons
<Tons to <Concrete
>
million tons>
Fatalities per million
tons of Al
Fatality cost due to
material inputs production
<Unit mortality
value>
Deaths material
inputs production
Fatalities per million
tons of steel
<Tons to
million tons>
Injuries per million
tons of steel
Fatalities per million
tons of concrete
Figure 6.5: Morbidity and fatalities sub-model stock and flow diagram
Similarly, the unit mortality value (UMtV, R/person) refers to the economic value for premature mortality
(fatalities or deaths). As explained earlier, the values for mortality were adapted from the NEEDS (2007)
and NewExt (2004) studies. In transferring estimates from the EU to the South African context benefit
transfer with income adjustment approach was used. The unit transfer with income adjustment approach is
usually used when transferring estimates between countries with different income levels and costs of
living. This is usually done using purchasing power parity (PPP) and income elasticity (Navrud, 2004;
Hainoun et al., 2009). The following formula was used in this study to adjust the estimates:

 PPP 
SA 
UV  UV 
SA
Rc  PPP 
Rc 

……………………………………………………………………………….............Income adjustment formula
Where, UVSA refers to unitary value in South Africa, UVRc refers to unitary value in reference country, PPP is
the GDP per capita adjusted for purchasing power parity and γ represents the income elasticity. An income
elasticity of WTP of < 1 would imply that WTP for the improvement in environmental quality drops with
increase in income, that is, as noted by Krupnick et al. (1996) premature mortality risk is an inferior good,
meaning an income elasticity of 1 will understate the WTP of lower income countries. In their transfer of
mortality values from the United States of America to Central and Eastern Europe Krupnick et al. (1996)
used an income elasticity of 0.35 with sensitivity analysis at 1. The income elasticity of WTP was also found
by Desaigues et al. (2011) to be less than one in nine countries of the EU25, normally in the range 0.4 - 0.7.
While reviewing the literature on the elasticity of WTP Pearce (2003) concluded that the income elasticity
- 147 -
of WTP for environmental change is less than one in most of the studies and that the range 0.3 - 0.7 seem
about right. On the other hand, an income elasticity of 1 would imply that WTP for environmental quality
varies equivalently with income while an income elasticity that is > 1 would mean environmental quality is a
luxury good (McFadden, 1994).
There is therefore disagreement in the literature concerning environmental quality, it is being viewed to be
an inferior good by some, a luxury good by some and the elasticity of WTP of 1 is not supported by
everyone. In this current study income elasticity’s of 1, 0.7 and 0.4 were used, for our low, central and high
estimates, respectively. A value of 0.7 was used in the baseline scenario. If the income elasticity is lower in
South Africa than assumed in the baseline scenario, the outcome would be an underestimation of the WTP
and the externality costs while if it is higher than assumed in this study the WTP values and externality
costs would be overestimated. In essence in order to evaluate whether the South African elasticity is
underestimated or overestimated, detailed information on the preferences of individuals in South Africa
and in the EU would be needed, as would a thorough analysis of the market structures of the various
nations. Since individual preferences are not easily measured, it becomes difficult to calculate where the
elasticity of South Africa lies in relation to the elasticity in the EU countries. In the absence of data income
elasticity’s of 1, 0.7 and 0.4 were used.
Overall, the unit mortality values were adjusted to reflect the disparity of income levels between the EU
and South Africa and to cater for inflation and some form of internalization. After all the adjustments the
central estimate which is used in the baseline model became R245 438/person, with the high estimate at
R771 700/person. The unit value for mortality is determined by the change in mortality value (∆UMtV,
R/person/Year), which is in turn altered by escalation of damage cost which is estimated at the growth rate
of population. The unit mortality value (UMV, R/person) is given by:
UMtV (t )  UMtV (245438)  UMtV dt ....................................................................................................(34)
The unit mortality and morbidity values play a central role in the computation of the coal-fuel cycle
fatalities and morbidity costs (CCFMC, R/Year). CCFMC is composed of fatalities and morbidity costs
streaming from three phases in the coal-fuel cycle, namely fatalities and morbidity costs (coal mining)
(FMCM, R/Year), fatalities and morbidity costs (construction) (FMC, R/Year) and fatalities and morbidity
costs (power generation) (FMCPG, R/Year) as follows:
CCFMC  FMCM  FMC  FMCPG ……………….........................................................................................(35)
- 148 -
The fatalities and morbidity costs (coal mining) (FMCM, R/Year) are determined by fatality cost (coal
mining) plus morbidity cost (coal mining). The fatality cost (coal mining) is in turn determined by the deaths
from coal mining together with the unit mortality value. The deaths from coal mining in turn are
determined by fatalities per million tons of coal mined, coupled with coal consumption in million tons. The
morbidity cost (coal mining) on the other hand, is determined by the injuries from coal mining together
with unit morbidity value. The injuries from coal mining are a function of the injuries per million tons of
coal mined, coupled with coal consumption in million tons. The injuries and fatalities per million tons of
coal mined were calculated as averages based on estimates of the deaths, injuries and coal mined in South
Africa, reported by the Department of Minerals and Energy (2008; 2010) and WCA (2006-09) for the years
2006 to 2009.
The fatalities and morbidity costs (construction) (FMC, R/Year) are determined as a sum of fatality cost due
to material inputs production and morbidity cost due to material inputs production. Three main material
inputs, namely aluminium, steel and concrete were considered in this study. The fatality and injury rates
per million tons of the main material inputs, coupled with the quantities of the main material inputs,
determine the deaths and injuries from the production of the main material inputs, which then together
with the unit values for mortality and morbidity, determine the fatality and morbidity costs due to material
inputs production. The fatality rate, injury rate and quantities of material inputs were computed as
averages based on estimates of deaths, injuries and quantities of material inputs reported by a number of
sources including estimates by the Department of Minerals and Energy (2007-2010).
Lastly, fatalities and morbidity costs (power generation) (FMCPG, R) are the sum of four main costs, namely
fatality cost from power generation, morbidity cost from power generation and fatalities and morbidity
costs from limestone production (FGD). The fatalities and injury rates per MWh coupled with gross
electricity production determine the deaths and injuries from power generation, which then together with
the unit values for mortality and morbidity determine the fatality and morbidity costs from power
generation. The fatalities and injury rates per MWh were computed based on estimates of deaths, injuries
and power production in the years 2006 to 2009 (Eskom, 2007; 2009b; 2010a).
The fatalities and morbidity costs from limestone production (FGD) on the other hand, reflect the fatality
and morbidity costs that are linked with the FGD process. These costs were, however, only limited to the
fatality and morbidity costs linked with the production of limestone (i.e. limestone is used in the flue gas
desulphurisation (FGD) system to curb SO2), so fatality and morbidity costs linked with the direct operation
- 149 -
of the FGD system were not included, owing to lack of deaths and injuries data in the literature. The fatality
and morbidity costs due to limestone production are determined by the deaths and injuries from limestone
production coupled with the unit mortality and morbidity values. The deaths and injuries are in turn a
function of limestone requirements and the fatality and injury rates. Data to compute the fatality and injury
rates was sourced from the Department of Minerals and Energy (2005; 2007). The parameters used in the
morbidity and fatalities sub-model are presented in Table 6.4. The complete respective equations of the
morbidity and fatalities sub-model are presented in Appendix A.
Table 6.4: Parameters used in the morbidity and fatalities sub-model
Parameters
Units
Baseline value
Dmnl/Year
persons/million
tons
0.011
Statistics South Africa, 2011.
0.056
Calculated based on DME (2008; 2010) and WCA
(2006; 2007; 2008; 2009).
persons/MWh
0.00000026
Calculated based on Eskom (2007; 2009) and
Eskom (2010a).
3.174283457
Calculated based on IndexMundi (2012b) and
DME (2007-2010).
0.159
Calculated based on DME (2007 - 2010), Lafarge
(2011), IndexMundi (2012a) and Palladian
Publications (2013).
0.2906977
Calculated based on DME (2005; 2007) and DMR
(2003).
2.0158923
Calculated based on DME (2007 - 2010), South
African Iron & Steel Institute (2013).
0.823
Calculated based on DME (2008; 2010) and WCA
(2006; 2007; 2008; 2009).
19.91141441
Calculated based on IndexMundi (2012b) and
DME (2007-2010).
0.995
Calculated based on DME (2007-2010), Lafarge
(2011),
IndexMundi
(2012a),
Palladian
Publications (2013).
1.3372093
Calculated based on DME (2005; 2007) and DMR
(2003).
0.3213741
Calculated based on DME (2007-2010), South
African Iron & Steel Institute (2013).
Persons/MWh
0.00000010
Calculated based on Eskom (2007; 2009) and
Eskom (2010a).
Unit morbidity value
R/person
25 434
Unit mortality value
R/person
245 438
Escalation of damage cost
Fatalities per million tons of coal mined
Fatalities per MWh
Fatalities per million tons of Al
Fatalities per million tons of concrete
Fatalities per million tons of limestone
Fatalities per million tons of steel
Injuries per million tons of coal mined
Injuries per million tons of Al
Injuries per million tons of concrete
Injuries per million tons of limestone
Injuries per million tons of steel
Injuries per MWh
6.6.4
Persons/million
tons
persons/million
tons
persons/million
tons
persons/million
tons
persons/million
tons
persons/million
tons
persons/million
tons
persons/million
tons
persons/million
tons
Data source
Calculated based on Van Horen (1997).
Calculated based on NEEDS (2007) and AEA
Technology Environment (2005).
Water consumption sub-model
Water is utilized in various activities in the coal-fuel cycle, for example, during the coal mining phase, water
is primarily used for dust control, extraction, coal washing and is also lost through evaporation (Wassung,
2010). Water is also used during the building and operation of the plant, operation of the FGD system and
in disposing of waste. The water consumption sub-model focuses on estimating the coal-fuel cycle
externality cost of water use. Before discussing the sub-model, the study will focus on the importance of
- 150 -
estimating the opportunity cost of water, the water issues in the Olifants river catchment, the opportunity
cost of water use in Kusile power station and how and why it was adjusted.
Estimating the opportunity cost of water use is imperative for a number of reasons. Among which are that:
water is a scarce resource in South Africa (Turton, 2008) that is not traded in the market; the administered
price of water does not reflect the scarcity of water; the price of water seldom reflects the full cost of water
delivery (Inglesi-Lotz & Blignaut, 2012), meaning water is under-priced, and lastly, the price of water does
not reflect the actual loss of welfare to society attributable to misallocation of water to suboptimal
applications, i.e. the administered water prices do not capture society’s welfare impact owing to the
presence of externalities (Spalding-Fecher & Matibe, 2003).
The Kusile power station, which is the focus of this study, is located in the Olifants River catchment, in
particular, upper Olifants together with other power stations, namely Arnot, Kriel, Hendrina, Matla, Kendal,
Duvha and Komati (DWA, 2011). The Olifants River is situated in the north-east of South Africa and
originates in Gauteng province (Wester, Merrey & De Lange, 2003). Water is a contested resource in the
Olifants River catchment which is perceived as one of South Africa’s severely stressed catchments in the
context of water quantity and quality. Over the years, the water requirements in the catchment have
increased extensively owing to rising water demand in various sectors. The water issues in the catchment
have led the Department of Water Affairs to undertake a reconciliation strategy for the basin and its users
to alleviate existing water deficits and as well as to ensure sustainable supply of water for the future. The
conflicting requirements of the various water users, however, present a major challenge in the reconciling
process (DWA, 2011).
In the mid portion of the catchment, water is mainly used for irrigation purposes while the Kruger National
Park found in the lower end of the catchment, necessitates sufficient river flow for the maintenance of the
system’s ecological integrity. In the upper portion of the catchment, apart from being used at thermal
power plants, water is used mainly for mining and for urban purposes. Most of the thermal power stations
in the upper catchment have large water requirements owing to their wet-cooling processes. Owing to the
water crisis in the Olifants catchment, all of the previously mentioned power stations are supplied with
water either from the upper Komati or the Vaal systems. About 228 million m3 of water per year are
transferred into the Olifants catchment to meet the water requirements of the power stations. The Kusile
power station will also receive water from the Vaal system (DWA, 2011).
- 151 -
The opportunity cost of water use to society when engaging in coal-fired electricity generation, was
adapted from Inglesi-Lotz and Blignaut (2012). Since the administered prices of water in South Africa do not
reflect the actual loss of welfare to society attributable to misallocation of water to suboptimal
applications, that is to say the administered water prices do not capture society’s welfare impact because
externalities are not incorporated into those prices (Spalding-Fecher & Matibe, 2003), Inglesi-Lotz and
Blignaut (2012) measured the externality cost of water use through estimating the shadow price of water
which served as an indicator of the opportunity cost to society of using water in coal-fired power
generation. Shadow prices are commonly relevant in the event that real prices do not represent the actual
loss of welfare to society (Moolman et al., 2006).
When estimating the opportunity cost to society of water use in coal-fired electricity generation, the
authors firstly estimated the shadow price of water when putting water use into coal-fired power
generation. Secondly, they estimated the shadow prices of water when the water was put into various
alternative technologies, for example, electricity production from renewable energy technologies like wind
and solar. It is important to note that Eskom is a strategic water user and receives its water at about 99.5%
level of assurance (Eskom, 2009a). Thus the water that will be used in Kusile will be strictly reserved for the
power sector, which is why the authors kept the water within the power sector, by evaluating alternative
energy technologies. The opportunity cost of water therefore computed focuses on water earmarked for
the power sector.
The shadow prices were computed such that they disclose the net marginal revenue (NMR) of water (i.e.
the additional revenue that will be generated by increasing water use by a cubic meter (Moore, 1999)). The
higher the NMR of water, the more efficiently water is used. The methodology of determining the true
scarcity value of water through estimating its shadow price and thereby comparing shadow prices of water
utilising technologies (NMRs), is an approach that has been applied successfully within the agriculture
sector (Moore, 1999; Moolman et al., 2006). The marginal revenue function of water determines the unit
cost as the opportunity cost.
Thirdly, the opportunity cost of using one technology over another was represented by the difference
between the NMRs (i.e. the forgone value of using water on coal-fired power generation rather than in a
wind-plant is given by the difference in NMRs between the two water using technologies). The estimated
NMR of water for all technologies was in R/m3 and in order to arrive at the opportunity cost values, a three
step process was undertaken: (1) the NMRs of water from the various alternative technologies were
subtracted from the NMR of water from the baseline model (i.e. coal-fired power generation); (2) the
- 152 -
differences in NMRs (in R/m3) were then multiplied with the water volume of the various technologies (m3)
yielding the society wide loss or gain (in R), and then lastly, (3) the opportunity cost (R/kWh) was calculated
as the societal loss/gain (R) divided by the net generation output of the baseline model (MWh), times 1000.
Low and high estimates of the opportunity cost of water use in Kusile power station were computed in the
Inglesi-Lotz and Blignaut study, i.e. the society-wide loss (opportunity cost) of water use in Kusile coal-fired
power station was computed to range between R21 305 million and R42 357million. Dividing these values
by the amount of water requirements for Kusile power station (26.166million m3 (Inglesi-Lotz & Blignaut,
2012)) yields opportunity cost of water per m3 amounting to a low and high value of R814/m3 and R 1
619/m3, respectively. The average of the low and the high estimates is used as the baseline value in this
study, but firstly all the opportunity cost values need some adjustment because the power purchased by
the water when put into renewables is in essence not real, owing to the fact that the technologies are not
yet into such large scales, so renewables will not be able to uptake the water. The following formula was
used to adjust the opportunity cost values:
  PS SW
1  
  PS K

  OCi ......................................................................................................................................(36)

Where: PS SW is the maximum plant size in MW for solar and wind; PS K is the maximum plant size in MW
of Kusile power station and OCi is the opportunity cost of water with i denoting either low, baseline or
high opportunity cost estimate (i.e. R814/m3, R1 217/m3 and R 1 619/m3, respectively). The IRP (2011), in
its policy-adjusted IRP, plans an investment of about 17.8GW in renewables (wind and solar), and on an
annual basis mainly an investment of a total of about 800MW in wind-based and solar-based power. This
capacity size for wind and solar (i.e. annual value) was therefore used to adjust the opportunity cost values
of water in accordance with the above adjustment formula. In accordance with Table 2.3 of the IRP
(2011:14) the capacity values reported for the various technologies seems to be load factor adjusted
(evidently evidenced by the reported capacity for Kusile i.e. 4 338MW), so also the relevant load factor
adjusted capacity generated by COALPSCA for Kusile power station was used.
Having discussed the above, attention is now reverted to the water consumption sub-model. Figure 6.6
presents the structure of the water consumption sub-model. The sub-model has one stock variable, namely
the unit opportunity cost of water which plays an essential role in the computation of the coal-fuel cycle
opportunity cost of water use. The unit opportunity cost of water use (UOCWU, R/m3) refers to the forgone
benefit to society of water use in the coal-fuel cycle (i.e. the externality cost of water under-pricing). As
- 153 -
explained above, the value for the opportunity cost of water use was adapted from Inglesi-Lotz & Blignaut
(2012) who approximated the opportunity cost to society of water use in Kusile power station. The unit
opportunity cost of water use (UOCWU, R/m3) is determined by the change in the opportunity cost of water
use (∆OCW, R/m3/Year), which is altered by escalation of damage cost (Dmnl/Year), which is estimated at
the rate of population growth (Note: the unit damage cost estimates for all the externalities studied in this
thesis including that of the opportunity cost of water were escalated at the growth rate of population
because the effects of the externalities and hence the externality costs associated with them will be borne
by the South African residents as a whole, so the costs will therefore likely grow at the growth rate of the
population). The unit opportunity cost of water use is given by:
UOCWU (t )  UOCWU (1217)  OCW dt ..................................................................................................(37)
<Escalation of
damage cost>
Change in the opportunity
cost of water use
Factor curbing
construction
Water requirements of
constructing Kusile
Al embodied
water
Water requirements of
constructing Kusile
(curbed)
<Unit opportunity cost
of water use>
Water usage Al
<Al>
Steel embodied
water
Coal-fuel cycle externality
cost of water use
Opportunity cost of
water use (construction)
Opportunity cost of water use in
producing material inputs of
constructing Kusile
Water requirements of
construction materials
Water usage
steel
Water requirements of a
surface mine (in m3/ton)
Opportunity cost of water
use in constructing Kusile
<Time>
Opportunity cost of
water use in FGD
<Plant water
consumption>
<FGD water
consumption>
Water usage
concrete
<Concrete>
<Coal
consumption>
Water requirements of a
surface mine (in litres/ton)
Litres to m3
Opportunity cost of water
use in disposing Kusile's
waste
Opportunity cost of water use
(power generation)
Concrete
embodied water
<Steel>
Water requirements of
a surface mine
Opportunity cost of water
use in the New Largo
colliery (coal mining)
Unit opportunity cost
of water use
Water usage in
disposing waste
Dry waste Kusile
Water usage per ton of
solid waste disposed
<Unit opportunity cost
of water use>
Ash produced per ton
<Coal
of coal burnt
consumption>
Figure 6.6: Water consumption sub-model stock and flow diagram
The coal-fuel cycle externality cost of water use (CCExtWU, R/Year) is composed of three main costs,
namely the opportunity cost of water use in the New Largo colliery (coal mining) (OPWCM, R/Year), the
opportunity cost of water use (construction) (OPWC, R/Year) and the opportunity cost of water use (power
generation) (OPWPG, R/Year) as follows:
CCExtWU  OPWCM  OPWC  OPWPG .................................................................................................(38)
The opportunity cost of water use in the New Largo colliery (coal mining) (OPWCM) is a function of the
water requirements of a surface mine and the unit opportunity cost of water use. The opportunity cost of
water use (construction) (OPWC) consists of the opportunity cost of water use in constructing Kusile and in
- 154 -
producing material inputs of constructing Kusile. The opportunity cost of water use in producing material
inputs of constructing Kusile, is in essence a product of the embodied water in the main material
requirements (i.e. aluminium, concrete and steel) and the unit opportunity cost of water use.
Lastly, the opportunity cost of water use (power generation) (OPWPG) is composed of three costs, namely
the opportunity cost of water use in the FGD system, the opportunity cost of water use to operate Kusile
and the opportunity cost of water use in disposing of Kusile's waste, which are functions of the water
requirements for these activities and the unit opportunity cost of water use. For example, the opportunity
cost of water use in disposing of Kusile's waste is determined by the unit opportunity cost of water use and
the water usage in disposing of waste. The water usage in disposing of waste is a product of the amount of
dry waste and the water usage per ton of solid waste disposed. The parameters used in the water
consumption sub-model are presented in Table 6.5. The complete respective equations of the water
consumption sub-model are presented in Appendix A.
Table 6.5: Parameters used in the water consumption sub-model
Parameters
Al embodied water
Ash produced per ton of coal burnt
Concrete embodied water
Escalation of damage cost
Steel embodied water
Unit opportunity of water
Water requirements of a surface mine (in litres/ton)
Water requirements of constructing Kusile
Water usage per ton of solid waste disposed
6.6.5
Units
Baseline
value
m3/ton
Dmnl
m3/ton
Dmnl/Year
m3/ton
R/m3
l/ton
m3
m3/ton
0.000088
0.293
26.352
0.011
225
1001
469
4 123 917
0.076
Data Source
Bardhan, 2012.
Eskom, 2010a.
Crawford, 2011.
Statistics South Africa, 2011.
Bardhan, 2012.
Inglesi-Lotz and Blignaut (2012)
Wassung, 2010; Pulles et al (2001).
Assumption
Spath et al. (1999).
Water pollution sub-model
Water pollution has been characterized as an environmental issue of concern in the eMalahleni area (EO
Miners, 2011). It is a costly environmental problem (Naicker et al., 2003; Council for Geoscience, 2010) that
imposes costs on various water users. The water pollution sub-model centers on estimating the coal-fuel
cycle water pollution damage cost. Figure 6.7 represents the structure of the water pollution sub-model.
The sub-model consists of three stocks, namely the unit damage cost of sulphate pollution from coal mining
(UDSCM, R/ton), steel production (UDSS, R/ton) and Al & concrete production (UDSAC, R/ton).
The unit damage costs by these industries represent the damages caused by them to other water users in
the eMalahleni catchment. The damages were adapted mainly from Van Zyl et al. (2002) who estimated the
cost imposed on other water users in the eMalahleni catchment, due to water pollution emanating from
- 155 -
various individual industries. Sulphate was chosen by the researchers as a best available indicator of overall
salinity and a major concern in the area. Damages to the industrial and domestic sectors were estimated
using preventative expenditures while those to the agricultural sector were estimated using preventative
expenditures necessary to maintain yield and lower yields due to pollution. The drawbacks of the Van Zyl
study are its focus on sulphate and not on all pollutants, its focus on impacts in the catchment and not
downstream, and its lack of addressing natural/environmental uses. Low and high estimates representing
the dry and wet seasons, respectively were computed in the Van Zyl study. The averages of the low and the
high estimates for coal mining, steel production, and aluminium and cement production were inflated and
used as baseline values in this study.
Change in damage cost
of sulphate pollution
(coal mining)
Escalation of
damage cost
Change in damage cost
of sulphate pollution
(steel production)
Unit damage cost of
sulphate pollution from
coal mining
Damage cost of sulphate
pollution from coal mining
<Coal
consumption>
Unit damage cost of
sulphate pollution from
steel production
<Steel>
Coal-fuel cycle water
pollution externality cost
Damage cost of sulphate
pollution from steel
production
<Concrete>
Change in damage cost
of sulphate pollution (Al
& concrete production)
Unit damage cost of
sulphate pollution from
Al & concrete
production
<Al>
Damage cost of sulphate
pollution from Kusiles' raw
material requirements
Damage cost of sulphate
pollution from Al & cement
production
Figure 6.7: Water pollution sub-model stock and flow diagram
The unit damage cost of sulphate pollution from coal mining (UDSCM, R/ton), steel production (UDSS,
R/ton) and Al & concrete production (UDSAC, R/ton) is determined by changes in the damage cost of
sulphate pollution from coal mining (∆DSCM, R/ton/Year), steel production (∆DSS, R/ton) and Al & concrete
production (∆DSAC, R/ton) which are altered by escalation of damage cost, as follows:
UDSCM (t )  UDSCM (0.27)  DSCM dt ....................................................................................................(39)
UDSS (t )  UDSS (0.79)  DSS dt ..............................................................................................................(40)
UDSAC (t )  UDSAC (0.31)  DSAC dt .....................................................................................................(41)
The coal-fuel cycle water pollution damage cost (CCWPDC, R/Year) is composed of two main costs, namely
the damage cost of sulphate pollution from coal mining (DCSCM, R/Year) and damage cost of sulphate
- 156 -
pollution from Kusile’s raw material requirements (DCSMR, R/Year). Water pollution damages from the
plant operation phase were not considered in the modelling, since Eskom plans to operate the Kusile plant
under a zero liquid effluent discharge policy once it is fully operational (NINHAM SHAND, 2007). In addition,
no major effluents are said to arise from limestone mining and processing (BCS-Incorporated, 2002),
therefore water pollution emanating from such activities was also not quantified. The coal-fuel cycle water
pollution damage cost (CCWPDC) is represented as follows:
CCWPDC  DCSCM  DCSMR ..................................................................................................................(42)
The parameters used in the water pollution sub-model are presented in Table 6.6. The complete equations
of the water pollution sub-model are presented in Appendix A.
Table 6.6: Parameters used in the water pollution sub-model
Parameters
Escalation of damage cost
Unit damage cost of sulphate pollution from
coal mining
Unit damage cost of sulphate pollution from
steel production
Unit damage cost of sulphate pollution from
Al & concrete production
6.6.6
Units
Baseline value
Data Source
Dmnl/Year
0.011
R/ton
0.27
Van Zyl et al. (2002)
R/ton
0.79
Computed based on EVRAZ (2009) and
Van Zyl et al. (2002).
R/ton
0.31
Calculated based on Van Zyl et al. (2002).
Statistics South Africa, 2011.
Ecosystem services loss sub-model
The ecosystem services loss sub-model is concerned with estimating the coal-fuel cycle cost of lost
ecosystem services. The generation of power from Kusile power station will necessitate the construction of
a new open-cast mine plus the construction of the power station. The open-cast mine will be located in the
New Largo coal reserve, which signifies the extent of the area that could be mined, which covers an area of
6 817 hectares that was mainly used for maize cultivation with an extensive part falling into grasslands
(Wolmarans & Medallie, 2011). The Kusile power station, on the other hand, is located in the
Hartbeesfontein and Klipfontein farms in eMalahleni, in a site measuring approximately 5 200 hectares,
which was previously used for maize farming and cattle grazing (NINHAM SHAND, 2007; Eskom, 2010b).
The extraction of the coal resource in the New Largo reserve and the construction of the power plant will
therefore lead to loss of both farmlands and grasslands. The opportunity cost of coal mining and plant
construction in the said areas is therefore the forgone benefits derived from agricultural production and
ecosystem services generated by grasslands. Figure 6.8 presents the structure of the ecosystem services
- 157 -
loss sub-model. The sub-model has two stocks, namely the unit maize price and unit value of ecosystem
services generated by grasslands.
Mining area under
maize production
Unit maize price
<Escalation of
damage cost>
Maize yield
per hectare
Forgone benefit from
grasslands due to coal
mining
Maize production
Change in maize
price
Maize production
(dry land)
Forgone benefit from
Forgone benefit from maize maize cultivation due to
coal mining
cultivation due to building and
operating plant
Power plant area under
Maize yield per
hectare (dry land) dry land maize production
Maize yield per
hectare (irrigated land)
Ecosystem services
lost due to coal mining
Coal-fuel cycle cost of
lost ecosystem services
Unit value of ecosystem
goods & services
generated by grasslands
Maize production
(irrigated land)
Ecosystem services lost
due to plant construction &
operation
Forgone benefit from
grasslands due to building and
operating plant
Power plant area under
irrigated maize production
Mining area under
grazing/grasslands
<Escalation of
damage cost>
Change in the value of
ecosystem goods & services
Power plant area
under grazing
Figure 6.8: Ecosystem services loss sub-model stock and flow diagram
The unit maize price (UMP, R/ton) is an input in the computation of the forgone benefits from maize
cultivation. Its initial value was adapted from Blignaut et al. (2010) and is determined by the change in
maize price (∆MP, R/ton/Year), which is altered by escalation of damage cost. The unit maize price is given
by:
UMP(t )  UMP(1600)  MPdt ................................................................................................................(43)
The unit value of ecosystem services generated by grasslands (UVEG, R/ha) is an important input in the
computation of the forgone benefit from ecosystem services generated by grasslands. Its initial value was
adapted from Blignaut et al. (2010). While there are numerous services provided by grasslands, including
carbon storage, drought and flood mitigation, sediment reduction, biodiversity maintenance, wildlife
habitat provision, aesthetic beauty provision, protection of watersheds, stream and river channels, nutrient
cycling and movement, waste detoxification and decomposition, and control of agricultural pests (USDA,
2010), only three of these, namely carbon storage, drought mitigation and sediment reduction, were
valued in the study by Blignaut et al. (2010), for a fire-prone grassland ecosystem in the Maloti–
Drakensberg mountain range in South Africa. These three ecosystem services were considered immediately
viable and marketable, thus the others were excluded to avoid selling services with no immediate market.
In the current study, however, only the carbon storage value could be adapted from Blignaut et al. (2010),
not drought mitigation or sediment reduction. The reason for this is that the water values are for a highrainfall mountain catchment and cannot be equated to highlands low productive grasslands. The carbon
- 158 -
sequestration estimate adapted from the study by Blignaut et al. (2010) is thus considered conservative.
The unit value of ecosystem services generated by grasslands (UVEG, R/ha) is determined by the change in
the value of ecosystem goods & services (∆VEG, R/ha/Year), which is in turn altered by escalation of
damage cost. The exact equation is given by:
UVEG (t )  UVEG (510)  VEG dt ............................................................................................................(44)
The coal-fuel cycle cost of lost ecosystem services (CCCLES, R/Year) consists of ecosystem services lost due
to coal mining (ESLCM, R/Year) and ecosystem services lost due to plant construction & operation (ESLPCO,
R/Year) and is represented as follows:
CCCLES  ESLCM  ESLPCO .................................................................................................................(45)
The ecosystem services lost due to coal mining (ESLCM) are determined by the foregone benefits from
maize cultivation and from grasslands due to coal mining while the ecosystem services lost due to plant
construction & operation (ESLPCO) are determined by the foregone benefits from maize cultivation and
from grasslands due to building and operating the plant. The foregone benefits from maize cultivation and
grasslands are a product of the areas under maize production and the unit maize price, and the areas under
grasslands and value of ecosystem services lost due to grasslands, respectively. The parameters used in the
ecosystem services loss sub-model are presented in Table 6.7. The complete equations of the ecosystem
services loss sub-model are presented in Appendix A.
Table 6.7: Parameters used in the ecosystem services loss sub-model
Variables
Escalation of damage cost
Maize yield per hectare
Maize yield per hectare (dry land)
Maize yield per hectare (irrigated land)
Mining area under grazing/grasslands
Mining area under maize production
Power plant area under dry land maize
production
Power plant area under grazing
Power plant area under irrigated maize
production
Units
Baseline value
Dmnl/Year
ton/ha
ton/ha
ton/ha
ha/Year
ha/Year
0.011
10
4.25
10
2045.1
4771.9
ha/Year
1404
Estimated based on NINHAM SHAND (2007) and
Eskom (2010b).
ha/Year
3744
Estimated based on NINHAM SHAND (2007) and Eskom
(2010b).
ha/Year
52
Estimated based on NINHAM SHAND (2007) and Eskom
(2010b).
- 159 -
Data source
Statistics South Africa, 2011.
Blignaut et al. (2010).
Calculated based on NINHAM SHAND (2007).
NINHAM SHAND, 2007.
Estimated based on Wolmarans and Madallie (2011).
Estimated based on Wolmarans and Madallie (2011).
6.6.7
Air pollution sub-model
Classic air pollutants arise throughout the coal-fuel chain, for instance when coal is mined and transported,
during the construction of the plant (e.g. from fuel use on site when ground-works are performed) and
during the coal combustion phase. The air pollution sub-model is concerned with estimating the coal-fuel
cycle air pollution human health cost. The air pollution sub-model structure is presented in Figure 6.9. The
sub-model has seven stocks representing the damage cost of the various classic air pollutants studied,
namely SO2, NOX, PM, nickel, lead, arsenic and chromium.
<Escalation of
damage cost>
Exponent
nickel
Constant
nickel
Particulate collection
efficiency
ton to kg
<Conversion
factor>
<kg to ton>
Coal consumption
in PJ
<Gross electricity
production>
Emission factor
NOx (coal)
Emission factor
PM (coal)
Coal combustion
NOx emissions
Change in SO2
damage cost
<Escalation of
damage cost>
<Coal road
transportation distance>
Coal road transport
NOx emissions
<g to ton>
Change in NOx
damage cost
<PM emissions per
MWh>
SO2 emissions
per MWh
Emission factor PM
(transportation)
Waste disposal
NOx damages
<Unit damage
cost PM>
Waste disposal SO2
emissions
<Waste disposal
electricity use>
<NOx emissions
per MWh>
<PM emissions
per MWh>
<Emission factor NOx
Waste disposal
<Unit damage
<g to ton>
(transportation)>
NOx emissions
cost NOx>
Material transportation
Material transportation
Plant construction raw material
NOx damages
NOx emissions
transportation damages
Material transportation
SO2 damages
Material transportation
PM damages
Coal road transport
PM damages
Coal road transport
PM emissions
<SO2 emissions
per MWh>
Limestone transportation
SO2 emissions
Conveyor coal transport
air pollution damages
Unit damage
cost NOx
<Unit damage
cost PM>
Conveyor coal transport
SO2 emissions
Emissions PM
in grams
<Unit damage
cost PM>
Plant construction air
pollution health cost
Coal road transport
NOx damages
Emissions NOx
in grams
Emission factor NOx
(transportation)
<Escalation of
damage cost>
Chromium
damages
Limestone transportation
SO2 damages
Coal transportation air
pollution health cost
<g to ton>
Constant
chromium
Unit damage cost
chromium
Change in damage cost
<Escalation of
per ton of chromium
<Unit damage cost
damage cost>
NOx>
Limestone transportation
Limestone transportation
NOx emissions
<Limestone
<NOx emissions
NOx damages
transportation electricity
per MWh>
use>
Limestone transportation
Limestone transportation
FGD system air
PM damages
PM emissions
pollution health cost
Plant operation air
pollution health cost
Unit damage
cost SO2
<PM emitted per GJ
heat input>
<mg to kg>
Exponent
chromium
<Coal consumption
in PJ>
<Unit damage
cost SO2>
<Unit damage
Waste disposal air
Waste disposal
cost PM>
pollution health cost
SO2 damages
Coal road transport
SO2 damages
Coal-fuel cycle air pollution
Waste disposal PM
Waste disposal PM
human health cost
damages
emissions
Coal road transport
SO2 emissions
Emission factor SO2
(transportation)
<GJ to PJ>
Chromium emission
factor in kg/PJ
Coal combustion chromium
& compounds emissions
Coal combustion
NOx damages
Coal combustion
PM damages
Lead content
in coal
Lead emission
factor in kg/PJ
<kg to ton>
Coal combustion heavy
metals damages
Coal combustion air
pollution health damages
<Unit damage
cost SO2>
<Unit damage
cost NOx>
Coal combustion
PM emissions
g to ton
Emissions SO2
in grams
Arsenic damages
Exponent lead
Constant lead
Chromium
content in coal
Nickel damages
<kg to ton>
Coal combustion lead &
compounds emissions
<Weight fraction
of ash in coal>
Unit damage
cost nickel
<Coal consumption
in PJ>
Change in damage
cost per ton of lead
Lead damages
<GJ to PJ>
Coal combustion
SO2 damages
Coal combustion
SO2 emissions
Emission factor
SO2 (coal)
mg to kg
Constant arsenic
Unit damage
cost arsenic
Change in damage
cost of nickel
Coal combustion nickel
& compounds emissions
Arsenic content
in coal
Coal combustion arsenic
<kg to ton>
& compounds emissions
Exponent
arsenic
Change in damage cost
per ton of arsenic
<Escalation of
damage cost>
damage cost>
Arsenic emission
factor in kg/PJ
<Coal
consumption>
Coal energy
content in GJ/ton
MJ to GJ
GJ to PJ
<Coal energy
content>
Weight fraction
of ash in coal
PM emitted per GJ
heat input
Fraction of fly
ash emitted
<kg to ton>
Nickel emission
factor in kg/PJ
Nickel content
in coal
Fly ash fraction
of total ash
Unit damage
cost lead
<Coal consumption
in PJ> <Escalation of
Conveyor coal transport
NOx damages
Conveyor coal
transport SO2 damages
<Unit damage
cost NOx>
<Unit damage cost SO2>
<Electricity use by
conveyor>
Conveyor coal
transport NOx
emissions
Unit damage
cost PM
<Escalation of
Conveyor coal damage cost>
transport PM damages
NOx emissions
per MWh
<Unit damage
cost SO2>
Change in PM
damage cost
<Unit damage
cost PM>
Conveyor coal
transport PM emissions
Material transportation
SO2 emissions
<g to ton>
<Distanced travelled
(construction materials)>
<Emission factor SO2
(transportation)>
Material transportation
PM emissions
<Electricity use by
conveyor>
<Emission factor PM
(transportation)>
PM emissions per
MWh
Figure 6.9: Air pollution sub-model stock and flow diagram
The coal-fuel cycle air pollution human health cost (CCAPC, R/Year) is composed of four main costs, namely
coal transportation air pollution health cost (CTAC, R/Year), plant construction air pollution health cost
(PCAC, R/Year), plant operation air pollution health cost (POAC, R/Year), and waste disposal air pollution
health cost (WDAC, R/Year) as follows:
- 160 -
CCAPC  CTAC  PCAC  POAC  WDAC ……….......................................................................................(46)
The coal transportation air pollution health cost (CTAC) reflects the air pollution health costs emanating
from the transportation of coal by road and by the conveyor, as planned for Kusile power station. The coal
road transportation damages are a function of the transportation distances, emission factors of SO2, NOx
and PM and the unit damage cost of these gases (a discussion of the unit damage cost of the gases is
provided below). Similarly, the conveyor transportation damages are a function of the electricity use in the
conveyor, conveyor emission factors of SO2, NOx and PM and the unit damage cost of these gases. On the
other hand, the plant construction air pollution health cost (PCAC) is determined by plant construction raw
material transportation damages, which is in fact a function of the transportation distances of the raw
material requirements, emission factors and the unit damage cost of SO2, NOx and PM. No data were,
however, found on fuel use onsite during the construction of the plant, so damages that could have been
realized from such were excluded. The plant construction air pollution health cost becomes zero after the
construction period.
The plant operation air pollution health cost (POAC) consists of two main damages, namely coal
combustion air pollution health damages and coal combustion heavy metals damages. The coal combustion
air pollution health damages are determined by the coal combustion SO2, NOx and PM damages, which are
in turn a function of power production, emission factors and the unit damage cost of the gases. Concerning
the unit damage cost of the classic air pollutants (i.e. SO2, NOx and PM) in the coal-fuel chain, these were
adapted from NEEDS (2007; 2008; 2009), Sevenster et al. (2008) and from AEA Technology Environment
(2005).
AEA Technology Environment followed the impact pathway approach established in the ExternE project
when estimating the damage cost of classic air pollutants (e.g. NOX, SO2, PM2.5) – beginning with estimating
emissions of pollutants in various European countries, tracking pollutants dispersion in the atmosphere
(using dispersion modelling based on the new EMEP model), evaluating the exposure of people and crops
to pollutants and quantifying impacts (using CAFE CBA methodology) and then estimating damages of the
classic air pollutants by using estimates of VOLY (Value of a Life Year) from the NewExt (2004) study for
mortality impacts. Utilizing data from various sources Sevenster et al. (2008) estimated emissions of
pollutants (e.g. SO2, NOX, PM2.5, etc) for large coal power producing countries and for estimating the
damage costs of classic air pollutants, the authors based their estimates on damage costs per ton of
emission from the EU-based NEEDS project (ExternE series of projects). In the NEEDS project the damage
- 161 -
costs per ton of a specific local air pollutant were calculated based on value of life year (VOLY EU). Sevenster
et al. (2008) adjusted the values for purchasing power parity and population for the various countries they
investigated. The NEEDS project estimates the VOLY lost by air pollution mortality based on the results of a
new contingent valuation survey conducted in European countries. WTP questions for a 6 and 3 months
gain in life expectancy were used to estimate the VOLY (NEEDS, 2007; 2008; 2009).
Basing the valuations of air pollution mortality on the change of life expectancy, as opposed to a valuations
based on accidental death or a small change in the probability of dying or mortality risk is more
advantageous because the approach automatically factor in the constraint that humans die only once
irrespective of pollution, it offers a unified framework for time series, cohort and intervention studies plus
directly yields the life expectancy change as a time integral of the observed mortality rate (Rabl, 2006). In
addition, change in life expectancy is further favourable because respondents during surveys show too
much difficulty understanding small probability variations while a change in life expectancy is well
understood (NewExt, 2003). In this current study the estimates of the VOLY for the EU (VOLYEU) were
transferred into this study and adjusted for different levels of income between the EU and South Africa. An
adjustment factor was then obtained and was used to adjust the original unit damage costs. Income
elasticity’s of 1, 0.7 and 0.4 were used in this study, with a value of 0.7 used in the baseline scenario.
Overall, the values were adjusted to reflect the disparity of income levels between the EU and South Africa
and to cater for inflation and some form of internalization as explained in section 6.5.3 (i.e. Morbidity and
fatalities sub-model). The baseline estimates used in this study are found in Table 5.8.
The coal combustion heavy metals damages are determined by arsenic, nickel, lead and chromium
damages, which are in turn functions of coal consumption, emission factors and unit damage cost of the
said four heavy metals. The emission factors for the four heavy metals were derived using engineering
equations for black coal combustion, equations considered more accurate by the National Pollutant
Inventory (NPI) (2012) as they consider fuel type and operational settings in the plant. The generic
equations used for estimating the heavy metals emission factors are as follows:
e
Cj

EF  K *  * PM j  kg / PJ
j
j
 A


PM  A * F * ER * 1000
j
SE

- 162 -

ER  1  CE

100
Where: EFj is the emission factor for a specific type of heavy metal denoted as j (kg/PJ); K is a constant
of a specific trace metal type; C is the concentration of trace metal in the coal (mg/kg); A is the weight
fraction of ash in the coal; e is an exponent specific to a trace metal type; PM is the power plant emission
factor for total particulate matter (kg/GJ), F is the flyash fraction of total ash, ER is the fraction of flyash
emitted; SE is the specific energy as received (i.e. heating value in GJ/t) and CE is the particulate
collection efficiency.
Concerning the unit damage cost of the toxic metals, these were adapted from European Commission
(2004) and ExternE-Pol (2005). Specifically, the impact pathway approach or bottom-up damage cost
approach was followed to establish the damages of toxic metals, through determining the quantities of
metal pollutants emitted by coal-fired power plants in European countries, tracking their dispersion and
ultimate deposition in various multimedia, evaluating the human health response to various doses of the
pollutants (through dose-response functions), and then valuation of increased morbidity and mortality
through surveys assessing individual’s preference for avoiding or reducing the risk of death or illness.
Lastly, the waste disposal air pollution health cost (WDAC) is determined by the waste disposal SO2, NOx
and PM damages, which are in turn functions of electricity use during waste disposal, emission factors and
the unit damage cost of the gases. The parameters used in the air pollution sub-model are presented in
Table 6.8. The complete equations of the air pollution sub-model are presented in Appendix A.
- 163 -
Table 6.8: Parameters used in the air pollution sub-model
Parameters
Arsenic content in coal
Chromium content in coal
Constant arsenic
Constant chromium
Constant lead
Constant nickel
Emission factor NOx (coal)
Emission factor NOx (transportation)
Emission factor PM (coal)
Emission factor PM (transportation)
Emission factor SO2 (coal)
Emission factor SO2 (transportation)
Escalation of damage cost
Exponent arsenic
Exponent chromium
Exponent lead
Exponent nickel
Flyash fraction of total ash
Lead content in coal
Nickel content in coal
NOx emissions per MWh
Particulate collection efficiency
PM emissions per MWh
SO2 emissions per MWh
Unit damage cost arsenic
Unit damage cost chromium
Unit damage cost lead
Unit damage cost nickel
Unit damage cost NOx
Unit damage cost PM
Unit damage cost SO2
6.6.8
Units
Baseline
value
mg/kg
mg/kg
Dmnl
Dmnl
Dmnl
Dmnl
ton/MWh
g/km
ton/MWh
g/km
ton/MWh
g/km
Dmnl/Year
Dmnl
Dmnl
Dmnl
Dmnl
Dmnl
mg/kg
mg/kg
ton/MWh
Dmnl
ton/MWh
ton/MWh
R/ton
R/ton
R/ton
R/ton
R/ton
R/ton
R/ton
2.95
57.02
2.73
2.47
2.87
2.84
2.4
13.04
0.22
0.68
10.02
1.66
0.011
0.85
0.58
0.8
0.48
0.2
20.38
25.69
0.00389
99.8
0.000358
0.00753
339976
133866
6799530
16149
41 952
227 175
51 619
Data source
Airshed Planning Professionals (2006).
Airshed Planning Professionals (2006).
NPI (2012).
NPI (2012).
NPI (2012).
NPI (2012).
Calculated based on Riekert and Koch (2011).
Stone and Bennett (n.d.).
Calculated based on Riekert and Koch (2011).
Stone and Bennett (n.d.).
Calculated based on Riekert and Koch (2011).
Stone and Bennett (n.d.).
Statistics South Africa, 2011.
NPI (2012).
NPI (2012).
NPI (2012).
NPI (2012).
Airshed Planning Professionals (2006).
Airshed Planning Professionals (2006).
Airshed Planning Professionals (2006).
Calculated based on Eskom (2010a).
NINHAM SHAND (2007).
Calculated based on Eskom (2010a).
Calculated based on Eskom (2010a).
ExternE-Pol (2005) and European Commission (2004).
ExternE-Pol (2005) and European Commission (2004).
ExternE-Pol (2005) and European Commission (2004).
ExternE-Pol (2005) and European Commission (2004).
Calculated based on NEEDS (2007; 2008; 2009),
Sevenster et al. (2008) and from AEA Technology
Environment (2005).
Global pollutants sub-model
Scientists concur that greenhouse gases such as CO2, N2O, CH4, tropospheric ozone (O3) and water vapour
are the principal gases responsible for global warming (Gaffen et al., 2000). Greenhouse gases arise
throughout the coal-fuel chain, for example CH4 is the principal GHG associated with coal mining, released
when coal seams are cut (National Research Council, 2009; Singh, 2008), CO2 is the key GHG linked with the
transport sector with CH4 emitted in small quantities (Gaffen et al., 2000) and CO2, CH4 and N2O are
released when coal is combusted. The global pollutants sub-model is concerned with estimating the coalfuel cycle global warming damage cost. The sub-model focuses mainly on three GHGs, namely CH4, CO2, and
N2O linked with coal mining and transportation, plant construction, plant operation and waste disposal. All
the studied GHGs and their damages were expressed in their CO2 equivalence (CO2e). The structure of this
- 164 -
sub-model is presented in Figure 6.10 and it contains two stocks, namely the unit damage cost of CO2 and
the unit train emission damage cost.
The coal-fuel cycle global warming damage cost (CCGWC, R/Year) is composed of four main costs, namely
coal mining & transportation global warming damages (CMTGWD,R/Year), plant construction global
warming damages (PCGWD, R/Year), plant operation global warming damages (POGWD, R/Year) and waste
disposal global warming damages (WDGWD, R), as follows:
CCGWC  CMTGWD  PCGWD  POGWD  WSGWD ………..................................................................(47)
The coal mining & transportation global warming damages (CMTGWD) are determined by three main
damages, namely coal mining CO2e damages emanating from CH4 emissions during mining, conveyor coal
transport damages which are composed of CO2e damages of N2O, CH4 and CO2 emanating from electricity
use in the conveyor, and coal road transport global warming damages, which are essentially composed of
coal road transport CO2e damages of N2O, CH4 and CO2. These latter damages are in turn determined by the
pollutant loads, unit damage cost of CO2 and global warming potentials of the gases. The pollutant loads of
the three gases are a function of various variables, including the quantity of coal transported by road, truck
capacity, transportation distance and truck fuel consumption.
Concerning the unit damage cost of CO2 (i.e. unit value of CO2), the values that were used in this study were
based on a study by Blignaut (2012) which developed a range of unit values centred on published (peer
reviews) studies while studying the social damage cost linked with climate change in a South African coalfired power plant (i.e. Kusile). The following range of values were developed in the Blignaut study,
R5.83/tCO2 (low), R104.93/tCO2 (market), R109.80/tCO2 (median), R177.79/tCO2 (high), R600.42/tCO2 (very
high), and R819.91/tCO2 (Stern 2007; 2008). The average market rate was computed after considering
carbon prices within the EU ETS programme, prices in the voluntary carbon market and CER prices. The
median, market and high unit value estimate were arguably selected as the most likely range in the
Blignaut study and were therefore used in this study. The baseline estimate in this current study is
therefore R109.80/tCO2.
The plant construction global warming damages (PCGWD) are determined by two damages, namely firstly
construction material CO2e damages, which is in essence determined by the quantities of the main material
inputs (steel, concrete and aluminium), their embodied GHGs, the global warming potentials of the GHGs
and the unit damage cost of CO2. The capacity construction rate influences the quantities of the main
- 165 -
material inputs used during the construction period. Secondly, material transportation CO2e damages
which in real meaning is determined by a number of variables, including fuel consumption, emission
factors, unit damage cost of CO2 and global warming potential of the GHGs.
On the other hand, the plant operation global warming damages (POGWD) are determined by three
damages, namely firstly coal combustion damages, which are determined in essence by a number of
variables including electricity production, emission factors of N2O and CO2, the global warming potentials of
GHGs, and the unit damage cost of CO2. Secondly, limestone transportation damages which are determined
by limestone consumption, limestone transportation distance and the unit train emission damage cost.
Thirdly, limestone use damages which are essentially a function of limestone consumption, limestone use
CO2 emission factor and the unit damage cost of CO2. Lastly, the waste disposal global warming damages
(WDGWD) are a function of various variables, including the amount of dry waste, transportation distance,
electricity use by the conveyor, global warming potentials of the GHGs and the unit damage cost of CO2.
- 166 -
<Gravimetric factor
converting C to CO2>
Carbon oxidation
factor
<Gross electricity
production>
N2O emissions
per MWh
Density of
bituminous coal
CH4 emission
m3/ton
Coal combustion CO2e
emissions (N2O)
CH4 emission
per ton of coal
MJ to TJ
Energy density
of diesel
Diesel
consumption in TJ
Coal road
transportation distance
<Global warming
potential CH4>
Diesel consumption in
litres (coal transport)
Gravimetric factor
converting C to CO2
Coal road transpoirt
CO2e damages
Coal road transport
CO2e emissions (CH4)
CH4 emission factor for
heavy duty diesel vehicles
<Diesel consumption in
litres (coal transport)>
Fraction of coal
transported by road
Unit damage
cost CO2
<Conversion
factor>
Coal road transport
CO2e damages
Conveyor coal transport
CO2e damages
<N2O emissions
per MWh>
<Global warming
potential N2O>
<N2O emission factor for
heavy duty diesel vehicle>
<Gravimetric factor
converting C to CO2>
- 167 -
<Global warming
potential CH4>
CF4 Al production
Steel N2O
embodiment
Al CF4
embodiment
Concrete
C2F4 Al
production
Al C2F6
embodiment
Al
Al CO2
embodiment
<Capacity
construction start>
Al per MW
Concrete per MW
<Unit damage cost
Materials transportation
CO2>
CO2e emissions (CH4)
Figure 6.10: Global pollutants sub-model stock and flow diagram
<Steel>
<Al>
Materials transportation
CO2 emissions
Diesel
consumption (TJ)
<Global warming
potential CH4>
<Diesel consumption
in litres
(construction)>
<CH4 emission factor for
heavy duty diesel vehicles>
CO2 steel
production
N2O steel
production
<kg to ton>
Concrete production
CO2e emissions
Materials transportation
CO2e emissions (N2O)
Steel CO2
embodiment
Steel per MW
<Global warming
potential N2O>
CO2 Al production
Conveyor coal transport
CO2 damages
<Dry waste
Kusile>
<Capacity
construction start>
Steel production
CO2e emissions
<kg to ton>
Plant construction global
warming damages
<N2O emissions
per MWh>
Al production CO2e
emissions
Global warming
potential CF4
Distance travelled
Kusile waste
Waste disposal
electricity use
Global warming
potential C2F6
Concrete CO2e
embodiment
Material transportation
CO2e damages
<Unit damage cost
CO2>
Electricity use by conveyor
Conveyor length
Conveyor coal
transport damages
Conveyor electricity
use per ton-km
<Global warming
potential N2O>
CH4 steel
production
Steel
<Unit damage cost
Steel CH4
CO2>
embodiment
Coal mining & transportation
global warming damages
<Global warming
potential N2O>
Waste disposal CO2e
emissions (N2O)
Waste disposal CO2e
damages (N2O)
Construction materials
CO2e damages
Coal road transport
global warming
damages
Conveyor coal transport
CO2e emissions (N2O)
<Conveyor
electricity use per
ton-km>
Coal-fuel cycle global
warming damage cost
C emission factor
for diesel
<Unit damage cost
CO2>
Global warming
potential N2O
Conveyor coal transport
CO2 emissions
<CO2 emission
per MWh>
Waste disposal
CO2 emissions
Waste disposal global
warming damages
Coal road transport
CO2 damages
<Coal
consumption>
<CO2 emission
per MWh>
Conveyor
transported coal
Waste disposal
CO2 damages
N2O emission factor for
heavy duty diesel vehicle
Coal road transport
CO2e emissions (N2O)
Limestone transportation
CO2e emissions (N2O)
Limestone transportation
CO2e damages (N2O)
<Unit damage cost
CO2>
Coal mining CO2e
damages
Coal road transport
CO2 emissions
Diesel oxidation
factor
Plant operation global
warming damages
kg to ton
Change in CO2
damage cost
<N2O emissions
per MWh>
<Unit damage cost
Limestone
CO2>
transportation CO2
emissions
FGD system global
warming damages
Coal mining CO2e
emissions (CH4)
<Limestone
consumption>
Limestone transportation
electricity use
Limestone transportation
CO2 damages
<Global warming
potential N2O>
<Limestone
transportation distance>
<CO2 emission per
MWh>
<kg to ton>
Limestone use CO2
damages
Limestone use
damages (FGD)
<Unit damage cost
CO2>
<Conveyor electricity
use per ton-km>
<Limestone
consumption>
Limestone use CO2
emissions
<Coal
consumption>
Global warming
potential CH4
<Coal
consumption>
Limestone use CO2
emission factor
<Gross electricity
production>
Coal combustion
CO2e damages (N2O)
<Escalation of
damage cost>
Truck fuel
consumption in l/km
Coal combustion
CO2 damages
Coal consumption
in kJ
<Coal energy
content>
Coal energy content
in kJ/ton
CO2 emission per
MWh
Coal combustion
CO2 emissions
CO2 emission
factor
MJ to kJ
<kg to ton>
Carbon %
<Diesel oxidation
factor>
<C emission
factor for diesel>
<Energy density
of diesel>
<MJ to TJ>
Diesel consumption in
litres (construction)
Transportation
distance (round trip)
Number of
road trips
<Al>
Main material
inputs
<Truck fuel
consumption in l/km>
Distanced travelled
(construction materials)
Truck
capacity
<Steel>
<Concrete>
The parameters used in the global pollutants sub-model are presented in Table 6.9. The complete
equations of the climate change sub-model are presented in Appendix A.
Table 6.9: Parameters used in the global pollutants sub-model
Parameters
Al C2F6 embodiment
Al CF4 embodiment
Al CO2 embodiment
AL per MW
C emission factor for diesel
Carbon %
Carbon oxidation factor
3
CH4 emission m /ton
Coal road transportation distance
Concrete CO2e embodiment
Concrete per MW
Construction material use schedule
Conveyor electricity use per ton-km
Conveyor length
Density of bituminous coal
Diesel oxidation factor
Distance travelled Kusile waste
Energy density of diesel
Escalation of damage cost
Fraction of coal transported by road
Global warming potential C2F6
Global warming potential CF4
Global warming potential CH4
Global warming potential N2O
Gravimetric factor converting C to CO2
Limestone use CO2 emission factor
N2O emissions per MWh
Steel CH4 embodiment
Steel CO2 embodiment
Steel N2O embodiment
Steel per MW
Truck fuel consumption in l/km
Unit damage cost CO2
Unit train emissions damage cost
6.6.9
Units
Baseline value
kg/ton
kg/ton
kg/ton
ton/MW
ton C/TJ
Dmnl
Dmnl
m3/ton
0.04
0.8
5301
0.419
20.2
0.425
0.99
0.014
Km
2.21484e+007
kg/ton
ton/MW
Dmnl/Year
MWh/ton/km
119.72
158.758
Fraction
0.0002
Km
42
kg/m3
Dmnl
Km
MJ/l
Dmnl/Year
732
0.99
30
38.46
0.011
Dmnl
0.27
Dmnl
Dmnl
Dmnl
Dmnl
Dmnl
kg/ton
ton/MWh
kg/ton
kg/ton
kg/ton
ton/MW
l/km
R/ton
R/ton/km
12200
7390
23
310
3.667
439.71
0.0000115
0.003
2710
0.040
50.721
0.35
109.89
0.018
Data source
IPCC (2007b).
IPCC (2007b).
Bosch & Kuenen (2009).
Spath et al (1999).
IPCC (1996).
Pinheiro et al (1997).
Blignaut et al (2005).
Cook (2005) and Lloyd and Cook (2005).
Calculated based on Synergistics Environmental Services
and Zitholele Consulting (2011) and Coaltech (2009).
InEnergy (2010).
Spath et al (1999).
Timeseries
ContiTech AG (2013).
Calculated based on Synergistics Environmental
Services and Zitholele Consulting (2011).
Cook (2005).
IPCC (1996).
Zitholele Consulting (2011).
Downs et al (1998) and Lammers (2009).
Statistics South Africa, 2011.
Calculated based on Synergistics Environmental Services
and Zitholele Consulting (2011) and Coaltech (2009).
IPCC (2007c).
IPCC (2007c).
IPCC (2001).
IPCC (2001).
Blignaut et al (2005).
IPCC (2007a).
Calculated based on Eskom (2010a).
IPCC, 1996.
Bosch & Kuenen (2009)
IPCC, 1996
Spath et al (1999).
Odeh and Cockerill (2008).
Blignaut (2012).
Jorgensen (2010).
Social cost sub-model
The social cost sub-model is concerned with estimating nine economic indicators in addition to the LCOE
discussed in the generation cost sub-model, namely levelised externality cost of energy (LECOE), levelised
social cost of energy (LSCOE), cumulative PV revenue, cumulative PV cost, NPV before tax, NPV after tax,
- 168 -
cumulative PV externality cost, social NPV before tax, and social NPV after tax. Expected profitability is also
discussed in this sub-model as explained in section 6.6.1. The structure of the social cost sub-model is
presented in Figure 6.11, mainly characterised by the nine economic indicators.
<Present value
factor>
<Coal-fuel cycle externality
cost of water use>
<Present value
factor>
<Cumulative PV net
electrity production>
Cumulative PV water
pollution externality
PV fatalities &
morbidity cost
<Coal-fuel cycle water
pollution externality cost>
<Present value
factor>
Levelised water
pollution externality
Levelised externality
cost of energy
Levelised social
cost of energy
<Coal-fuel cycle
fatalities & morbidity
costs>
Coal-fuel cycle
externality costs
<Coal-fuel cycle global
warming damage cost>
Levelised fatalities &
morbidity cost
<Levelised cost of
energy>
<Cumulative PV net
electrity production>
<Plant private
costs>
<Gross electricity
production>
Unit cost of
prodution
<Time>
Revenue
<Net electricity
production>
PV revenue
PV ecosystem
services loss
<Cumulative PV
ecosystem services loss>
Externality cost
switch
<Cumulative PV
fuel cost>
<Cumulative PV net
electrity production>
<Cumulative PV <Cumulative PV externality
cost of water use>
Social NPV
fatalities & morbidity
(before tax)
cost>
Cumulative PV
externality cost
<Cumulative PV global
warming damages>
<Fixed O&M
costs year>
<Capacity
investment>
NPV (before tax)
Levelised ecosystem
services loss
Cumulative PV fatalities &
morbidity cost
<Variable O&M
costs>
Profits (after tax)
Cumulative PV revenue
Cumulative PV ecosystem
services loss
Private cost rate
<FGD
operation cost>
Plant private costs
<Coal-fuel cycle cost of
lost ecosystem services>
<Present value
factor>
Levelised global
warming damages
Levelised water use
externality
<Coal-fuel cycle
fatalities & morbidity
costs>
<Coal-fuel cycle air
pollution human health
cost>
Levelised air
pollution cost
Profits (before tax)
Electricity price
table
PV global warming
damages
<Cumulative PV net
electrity production>
Cumulative PV externality
cost of water use
Cumulative private
costs
Cumulative revenue
Revenue rate
<Tax rate factor>
Cumulative PV global
warming damages
Cumulative PV air
pollution cost
PV air pollution
cost
PV external cost of
water use
PV water pollution
externality
<Coal-fuel cycle global
warming damage cost>
<Coal-fuel cycle air pollution
human health cost>
<Cumulative PV water
pollution externality>
Cumulative PV
costs
<Coal cost>
<Present value
factor>
<Cumulative capital
cost escalated>
<Cumulative PV
fixed O&M costs>
NPV (after tax)
<Cumulative PV
variable O&M costs>
Social NPV
(after tax)
<Cumulative PV
FGD operation cost>
Tax rate factor
<Cumulative PV air
pollution cost>
Time to adjust
profit
Unit profitability
Expected profitability
Change in expected
profitability
Unit coal-fuel cycle
externality cost
<Electricity price
table>
<Tax rate factor>
<Time>
<Coal-fuel cycle cost of
lost ecosystem services>
Figure 6.11: Social cost sub-model stock and flow diagram
The levelised externality cost of energy (LECOE) and the levelised social cost of energy (LSCOE) are
computed in a similar manner as done for the LCOE. The levelised externality cost of energy is composed of
six stocks which reflect the six externalities studied in the coal-fuel cycle. The coal-fuel cycle externality cost
of water use together with the present value factor determines the PV externality cost of water use
(PVExWU, R/Year), which is an inflow to the cumulative PV externality cost of water use (CPVExWU, R)
which is the first stock given by:
CPVExWU (t )  CPVExWU (0)  PVExWU dt ........................................................................................(48)
The cumulative PV externality cost of water use, coupled with cumulative PV net electricity production
(PVNEP, MWh), determines the levelised water use externality (LWUEx, R/MWh), as follows:
LWUEx  CPVExWU / CPVNEP …………......................................................................................................(49)
On the other hand, the coal-fuel cycle water pollution externality, together with the present value factor,
determines the PV water pollution externality (PVWPEx, R/Year), which is an inflow to the cumulative PV
water pollution externality (CPVWPEx, R) which is the second stock, given by:
- 169 -
CPVWPEx (t )  CPVWPEx (0)  PVWPEx dt .........................................................................................(50)
The cumulative PV water pollution externality, coupled with cumulative PV net electricity production
(CPVNEP, MWh), determines the levelised water pollution externality (LWPEx, R/MWh), as follows:
LWPEx  CPVWPEx / CPVNEP ………….......................................................................................................(51)
In turn, the coal-fuel cycle fatalities & morbidity costs, together with the present value factor, determine
the PV fatalities & morbidity costs (PVFMC, R/Year), which is an inflow to the cumulative PV fatalities &
morbidity costs (CPVFMC, R) which is the third stock, given by:
CPVFMC (t )  CPVFMC (0)  PVFMC dt ..................................................................................................(52)
The cumulative PV fatalities & morbidity cost, coupled with cumulative PV net electricity production
(CPVNEP, MWh), determines the levelised fatalities & morbidity cost (LFMC, R/MWh), as follows:
LFMC  CPVFMC / CPVNEP .....................................................................................................................(53)
The coal-fuel cycle cost of lost ecosystem services, together with the present value factor, determines the
PV ecosystem services loss (PVESSL, R/Year), which is an inflow to the cumulative PV ecosystem services
loss (CPVESSL, R) which is the fourth stock, given by:
CPVESSL (t )  CPVESSL (0)  PVESSL dt ..................................................................................................(54)
The cumulative PV ecosystem services loss, coupled with cumulative PV net electricity production (CPVNEP,
MWh), determines the levelised ecosystem services loss (LESSL, R/MWh), as follows:
LESSL  CPVESSL / CPVNEP .....................................................................................................................(55)
In turn, the coal-fuel cycle air pollution human health cost, together with the present value factor,
determines the PV air pollution cost (PVAPC, R/Year), which is an inflow to the cumulative PV air pollution
cost (CPVAPC, R) which is the fifth stock, given by:
- 170 -
CPVAPC (t )  CPVAPC (0)  PVAPC dt .................................................................................................(56)
The cumulative PV air pollution cost, coupled with cumulative PV net electricity production (CPVNEP,
MWh), determines the levelised air pollution cost (LAPC, R/MWh), as follows:
LAPC  CPVAPC / CPVNEP .......................................................................................................................(57)
Finally, the coal-fuel cycle global warming damage cost, together with the present value factor, determine
the PV global warming damages (PVGWD, R/Year), which is an inflow to the cumulative PV global warming
damages (CPVGWD, R) which is the sixth stock, given by:
CPVGWD (t )  CPVGWD (0)  PVGWDC dt ...............................................................................................(58)
The cumulative PV global warming damages, coupled with cumulative PV net electricity production
(CPVNEP, MWh), determine the levelised global warming damages (LGWD, R/MWh), as follows:
LGWD  CPVGWD / CPVNEP ....................................................................................................................(59)
The six levelised externalities discussed above (i.e. levelised water use externality, water pollution
externality, fatalities and morbidity costs, ecosystem services loss, air pollution cost and global warming
damages) are summed to yield the levelised externality cost of energy (LECOE, R/MWh), which is the first
economic indicator, represented as follows:
LECOE  LWUEx  LWPEx  LFMC  LESSL  LAPC  LGWD ……………………....................................(60)
The levelised social cost of energy (LSCOE, R/MWh), which is the second economic indicator, is specified by
the levelised externality cost of energy (LECOE, R/MWh) and the earlier computed LCOE (R/MWh), as
follows:
LSCOE  LCOE  LECOE ..........................................................................................................................(61)
The third economic indicator is the cumulative PV revenue (CPVR, R). The revenue and the present value
factor determines the PV revenue (PVR, R/Year), which is an inflow to the CPVR. The revenue is the
function of the energy price and net electricity production. The energy price was entered as a table,
- 171 -
depicting the time series of the wholesale energy price, computed based on a number of sources including
Lana (2010), Eskom (2012c), Eskom (2013b), NERSA (2013) and BDFM Publishers (2013c). The cumulative
present value revenue (CPVR, R) is given by:
CPVR (t )  CPVR (0)  PVR dt ....................................................................................................................(62)
The fourth economic indicator is the cumulative PV cost (CPVC, R), which is a summation of the cumulative
private costs of generating coal in Kusile, namely cumulative capital cost escalated (CKCE), cumulative PV
fuel cost (CPVFC), cumulative PV variable O&M costs (CPVVO&MC), cumulative PV fixed O&M costs
(CPVFO&MC) and cumulative PV FGD operation cost (CPVFGDOC), as follows:
CPVC  CKCE  CPVFC  CPVVO & MC  CPVFO & MC  CPVFGDOC …………….…….........................(63)
The fifth and sixth economic indicators concern the net present value, which is an economic measure for
examining cash outflows (costs) and cash inflows (revenues) of investing in the Kusile project. Before and
after tax NPVs were computed. The NPV before tax (NPVbt, R) is given by the difference between the
cumulative PV revenue (CPVR, R) and cumulative PV cost (CPVC, R) whereas the NPV after tax (NPVat, R)
corrects the NPV before tax with a tax rate factor (TaxR, Dmnl) which is based on NERSA (2013). The NPVs
before tax and after tax are given by the following equations, respectively:
NPV
bt
 CPVR  CPVC .................................................................................................................................(64)
NPV at  NPV *TaxR ..................................................................................................................................(65)
bt
The seventh economic indicator is the cumulative PV externality cost (CPVExC, R), which is a summation of
the cumulative PV externality costs computed earlier, namely cumulative PV externality cost of water use
(CPVExWU), cumulative PV water pollution externality (CPVWPEx), cumulative PV fatalities & morbidity
costs (CPVFMC), cumulative PV ecosystem services loss (CPVESSL), cumulative PV air pollution cost
(CPVAPC) and cumulative PV global warming damages (CPVGWD). The cumulative present value externality
cost is represented as follows:
CPVExC  CPVExWU  CPVWPEx  CPVFMC  CPVESSL  CPVAPC  CPVGWD ..................................(66)
- 172 -
The eighth and ninth economic indicators concern the social NPV, which is an indicator that examines the
private and externality costs and benefits of the Kusile project. Before tax and after tax social NPVs were
computed. The social NPV before tax (SNPVbt, R) is yielded by the difference between the NPV before tax
(NPVbt, R) and the cumulative PV externality cost (CPVExC, R) whereas the social NPV after tax (SNPVat, R) is
yielded by the difference between the NPV after tax (NPVaf, R) and the cumulative PV externality cost
(CPVExC, R). The before tax and after tax social NPVs are given by the following equations, respectively:
SNPV
bt
 NPV
bt
 PVExC ............................................................................................................................(67)
SNPVat  NPVat  PVExC .........................................................................................................................(68)
Lastly, one more stock presented in the social cost sub-model is expected profitability (EP, Dmnl) which
basically represents the balance between the electricity price and the unit cost of production. Expected
profitability (EP, Dmnl) is driven by the change in expected profitability (CEP, Year) and is represented as
follows:
EP(t )  EP(0)  CEP dt ..............................................................................................................................(69)
The change in expected profitability is represented by the following equation:
CEP  (UP  EP) / Tap ....................................................................................................................................(70)
Where, UP is the unit profitability (Dmnl) and Tap is the time to adjust profitability (Year). The unit
profitability (UP, Dmnl) in turn is determined by the electricity price (EP, R/MWh) and the unit cost of
production (UCP, R/MWh) and is represented as follows:
UP  ( EP  UCP) / EP .....................................................................................................................................(71)
The unit cost of production (UCP, R/MWh) represents such costs such as fuel cost (i.e. coal cost), variable
O&M costs, fixed O&M costs and FGD operation cost. The effect of externality costs on the profitability of
the plant was not accounted for in the initial analysis but instead a switch was used that tested for its
incorporation. A value of 1 is taken by the switch if externality costs are included in the computation of the
unit profitability, otherwise zero if not.
- 173 -
The complete respective equations and the rest of the equations of the social cost sub-model are contained
in Appendix A.
6.7
Summary
In this chapter, the modelling process followed in developing the COALPSCA Model was discussed. The first
and essential step in the model building process is problem formulation which saw the framing of the
problem addressed in this study, the purpose of the model and a discussion of the time horizon for the
model. The main concern of this research was stated as to understand the design and performance of a
coal-fired power plant and its interconnections with resource inputs, private costs, externalities, externality
costs and hence its consequent economic, social and environmental impacts over its lifetime and fuel cycle.
The purpose of developing the model was stated as twofold, firstly to aid energy decision makers with a
tool for making informed energy supply decisions that consider not only the financial feasibility of power
generation technologies, but in addition the socio-environmental consequence of the technologies.
Secondly, the model is to aid coal-based power developers with a useful tool for detecting the main drivers
of the burdens and costs in the system which should yield vital socio-economic-environmental tradeoff
information that can be beneficial to them. A period of 50 years was considered a reasonable time frame to
address the key issues in this study.
The next important step involved the formulation of the dynamic hypothesis which saw the construction of
a working theory that explains the problem. The behaviour of the power generation system with its
upstream processes and to a certain extent its downstream processes was qualitatively expressed in the
causal loop diagram based on the causal structure and feedbacks of the system. The model boundary was
then discussed. Lastly, the stock and flow structures of the modeled system were constructed and they
provided the quantitative relationships between the variables of the system by adding stock and flow
variables. The COALPSCA Model was separated into nine sub-models, namely power generation, generation
cost, water consumption, water pollution, morbidity and fatalities, ecosystem services loss, air pollution,
global pollutants and social cost sub-models. In the next section the outcomes of the COALPSCA Model are
presented together with validation tests and policy design and evaluation.
- 174 -
CHAPTER 7:
7.1
RESULTS
Introduction
In the previous section, a life-cycle power generation and social cost assessment model for coal-based
power (COALPSCA) was constructed based on the system dynamics approach. The results of COALPSCA are
reported in this section. The model results are by no means to be viewed as predictions, but rather as likely
evolutions of coal-based power generation from which, understanding might be derived to making
informed decisions. Presented firstly are the baseline results, followed by the validation and verification of
COALPSCA, which include an analysis and discussion of the sensitivity of the model outcomes to key
parameters such as the load factor, discount rate, cost growth rates of all private costs in the model (e.g.
coal, limestone, water, O&M and capital costs), cost growth rates of all damage cost estimates and the
sensitivity of the model outcomes to lower and higher range estimates. This is then followed by an
evaluation of the model outcomes under various policy scenarios that could be faced by coal-based power
utilities, namely carbon taxation and the sale of coal domestically at export parity prices. A summary of the
results is then presented lastly.
7.2
Baseline results
The COALPSCA Model aims to demonstrate the assessment of the social cost of coal-based power
generation in the Kusile power station. An analysis of the outcomes of the model, on selected economic
and environmental indicators under various scenarios, was conducted. The focus in this section is on the
baseline scenario. The baseline scenario represents power production in the Kusile power station over a
period of 50 years as planned by Eskom. The key input parameters used in the baseline scenario are
contained in Table 7.1. As anticipated by Eskom, the baseline scenario assumes a 90% load factor while an
energy content of 19.22MJ/kg is used, as per the typical coal consumed by the entity (Eskom, 2010a). The
scenario further assumes a 0.1% growth rate of all private costs (e.g. limestone, O&M, water and coal costs)
while the studied externalities’ damage costs were escalated at the growth rate of population, i.e. 1.1%..
The damage costs were escalated at the growth rate of population growth because the effects of the
externalities and hence the externality costs associated with them will be borne by the South African
residents as a whole, so the costs will therefore likely grow at the growth rate of the population. Average
values of low and high estimates were used as starting values to value the studied externalities and based
on Eskom communication (2012) a unit cost of coal of R210 per ton was used in the baseline model. A
discount rate of 8% was used in the baseline scenario as discussed earlier in section.
- 175 -
Table 7.1: Baseline scenario input parameters
Parameter
Load factor
Coal energy content
Private cost growth rates (i.e. coal,
limestone, water, O&M & capital cost)
Escalation of damage cost
Externality damages
Discount rate
Unit coal cost
Units
Baseline value
Dmnl (%)
MJ/kg
0.9 (90%)
19.22
Dmnl/Year (%)
0.1 (0.1%)
Dmnl/Year (%)
Various units
Dmnl
R/ton
1.1%
Average estimates
0.08 (8%)
210
The economic and environmental/societal indicators used for the analysis of the COALPSCA Model
outcomes are contained in Table 7.2. In general, ten economic indicators representing the performance of
the plant, the cost incurred by plant developers and the community at large were considered. In addition,
six environmental indicators reflecting the six coal-fuel cycle externalities quantified and monetized in this
study were also considered.
Table 7.2: Economic and socio-environmental indicators
Economic indicator










7.2.1
Environmental/societal indicator






Electricity production
Generation cost (capital, fuel & O&M)
Levelised cost of energy (LCOE)
Levelised externality cost of energy (LECOE)
Levelised social cost of energy (LSCOE)
Cumulative PV revenue
Cumulative PV cost
NPV
Cumulative PV externality cost
Social NPV
Water use
Water pollution
Fatalities & morbidity
Ecosystem services loss
Air emissions
GHG emissions
Electricity generation
There are various factors that influence the amount of electricity generation, including plant capacity, load
factor, operating hours, idle capacity, profits and plant’s own electricity consumption. An increase in plant
capacity, load factor, operating hours and profits positively affect power generation while an increase in
idle capacity and plant’s own electricity consumption negatively affects generation. The plant was modelled
to produce 90% of the total amount of electricity it could theoretically produce over its lifetime (i.e. load
factor of 90%). For this reason, 10% of the plant capacity was held idle. 7.5% of the electricity produced by
the plant was modelled to be consumed internally by Kusile according to Eskom communication (2012).
Plant operating hours were estimated at about 8234 hours annually after correcting for energy availability
factor of about 90% (based on Eskom communication, 2012). The energy availability factor as explained
- 176 -
earlier is mainly a factor of its reliability and the periodic maintain it requires (i.e. it corrects for the fact
that the power plant does not run continuously over its lifetime but needs to be shut down for service at
times).
The baseline scenario electricity production outcomes are presented in Table 7.3. The model estimates an
annual net electricity production of 32.8 million MWh (gross 35.5 million MWh) once Kusile is fully
operational. Eskom on the other hand, using a 90% load factor and an energy availability factor of 84%,
estimates Kusile’s annual net electricity production at about 32.7 million MWh (Eskom, 2010c). The
COALPSCA Model estimate is therefore analogous to the estimate by Eskom with about 0.3% variation.
Total net electricity production over Kusile’s lifetime is estimated by COALPSCA at about 1.6 billion MWh
(gross 1.7 MWh). About 18 million tons of coal is estimated by COALPSCA to be consumed annually once
Kusile is fully operational. An estimate that is comparable with NINHAM SHAND (2007) estimates of 21.1
million tons for a 5400MW plant or 18.8 million tons if one corrects the estimate to Kusile’s actual size,
yielding about 4% variation (Table 6.7). In addition, Synergistics Environmental Services and Zitholele
Consulting (2011), as well as Wolmarans and Medallie (2011), estimate annual coal requirements of about
17 million tons for Kusile, estimates yielding about 5.6% variation to the estimate by the COALPSCA Model
(Table 7.4).
Table 7.3: Baseline scenario electricity production (gross and net)
Output variable
Net electricity production
Gross electricity production
Cumulative net electricity production
Cumulative gross electricity production
Units
Model
outcomes
Eskom
projection
% variation
Million MWh/Year
Million MWh/Year
Billion MWh
Billion MWh
32.8
35.5
1.6
1.7
32.7
0.3
Table 7.4: Coal consumption (Million tons)
Classic air pollutant
Coal consumption (annual)
Coal consumption (lifetime)
Model
outcomes
NINHAM SHAND
(corrected)
%
variation
Wolmarans &
a
Medallie
%
variation
18
870.3
18.8
4.3
17
5.6
a
Aslo Synergistics Environmental Services and Zitholele Consulting (2011)
7.2.2
Private costs
The economic indicators linked to the private costs of generating electricity include - (i) the generation cost
which consists of capital, fuel, FGD, fixed and variable O&M costs; (ii) the LCOE which is categorized into
capital, fuel, FGD, fixed and variable O&M costs; and (iii) the NPV which consists of cumulative PV revenues
- 177 -
and costs. The private costs of producing coal-based electricity are presented in Table 7.5. The overall
escalated lifetime generation cost of power in Kusile is estimated by the COALPSCA Model to be about
410.5 billion Rands. The main generating cost components determining the generation cost are fuel and
capital costs, which individually constitute about 46% and 29%, respectively. Fixed and variable O&M costs
make up about 11% and 9% of the generation cost, respectively, while FGD operation cost contributes the
least at about 5%.
Table 7.5: Baseline scenario private costs over Kusile’s lifetime
Output variable
Capital cost
Fuel cost
Fixed O&M costs
Variable O&M costs
FGD operation cost
Total generation cost
Levelised capital cost
Levelised fuel cost
Levelised fixed O&M costs
Levelised variable O&M costs
Levelised FGD operation cost
LCOE
Cumulative PV cost
Cumulative PV revenue
NPV (before tax)
NPV (after tax)
Units
Model outcomes
R billion
R billion
R billion
R billion
R billion
R billion
R/MWh
R/MWh
R/MWh
R/MWh
R/MWh
R/MWh
R billion
R billion
R billion
R billion
118.8 (28.9%)
a
187.7 (45.7%)
a
46.7 (11.4%)
a
38.2 (9.3%)
a
19.2 (4.7%)
410.5
362.2
117
33.7
27.5
13.8
554.2
181.8
274
92.2
66.4
Calc. Eskom (2011)
EPRI (2010)
a
338.8
146.5
105.5
b
c
540.2 - 584.6
590.8
a
Proportions (%) of electricity generation cost; b Calculated based on Eskom’s reported benchmarked value to the same base year and exchange
rate as that of EPRI (2010) (i.e. USD73/MWh); c Calculated based on Eskom’s reported non-benchmarked value (i.e. USD73/MWh).
For the LCOE computations, the time series of expenses were discounted to present values (2010 base
year) by the use of a discount rate. A discount rate of 8% was used for the baseline scenario. So the LCOE
(R/KWh) is the ratio of total lifetime expenses to total expected output (i.e. electricity), expressed in
present value. The LCOE for generating electricity in Kusile is estimated at R554.2/MWh. Levelised capital
and fuel costs are about R362.2/MWh and R117/MWh, respectively, while levelised fixed and variable
O&M costs are estimated at about R33.7/MWh and R27.5/MWh, respectively. Constituting the least to the
lifetime LCOE is the levelised FGD operation cost of about R13.8/MWh (see Table 7.5). Figure 7.1 below
shows the LCOE simulation outputs as estimated by the COALPSCA Model. The LCOE computations
conducted in this study are in nominal (i.e. current) Rands, meaning the effects of inflation are taken into
account when looking at future costs, however, it must be noted that for the baseline scenario escalation
was assumed to be 0.1% (a value that is less than the current inflation rate in South Africa).
- 178 -
LCOE outputs
5000
R/MWh
3750
2500
1250
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised cost of energy : Baseline
Levelised capital cost : Baseline
Levelised fuel cost : Baseline
"Levelised fixed O&M costs" : Baseline
"Levelised variable O&M costs" : Baseline
Levelised FGD operaton cost : Baseline
Figure 7.1: LCOE outputs
While it is quite problematic to compare the LCOE from various studies due to the various assumptions
used, for example technology, plant size, plant design, load factor, base year, exchange rate, discount rates,
etc., the LCOE computed in this study was compared to that computed by EPRI (2010) and Eskom (2011).
EPRI (2010) computed a LCOE for a pulverised coal plant of 4 856MW (gross) with FGD, using a load factor
of 85%. The cost estimates computed by EPRI (2010) were based on most recent EPRI studies (US), which
were adjusted for the South African case. Constant dollar estimates were computed by the EPRI and the
base year was also 2010 as in this study. In spite of the plant performance and financial assumption
differences between this study and that conducted by EPRI, the overall lifetime LCOE from both studies is
not diversely different. EPRI computed an overall lifetime LCOE of R590.8/MWh while in this study
COALPSCA estimated a LCOE of R554.2/MWh.
On the other hand, even though Eskom computed the LCOE for Kusile, the entity does not disclose the
financial parameters used, neither the breakdown of the LCOE nor the exact method of assessment. In its
Annual Report for 2011, Eskom benchmarked Kusile’s LCOE to the same base year and exchange rate as
that of EPRI (2010) and reported a value of USD79/MWh, which translates to R584.6/MWh if the exchange
rate used by EPRI is adopted (7.4ZAR (South African Rand)/US dollar). In the same Annual Report, Eskom
reports the LCOE for Kusile to be USD73/MWh (not benchmarked) (Eskom, 2011), which translates to
R540.2/MWh at an exchange rate of 7.4ZAR/US dollar as used by EPRI. The LCOE of R554.2/MWh
- 179 -
computed by COALPSCA in the current study is therefore more comparable to the estimate by Eskom of
R540.2/MWh, though not diversely different from the estimate by EPRI (2010) of R590.8/MWh. In 2012
Eskom reported that it had altered the levelised cost model and that the computations made earlier would
change once the Board had approved them (Eskom, 2012c).
On another note, net present value analysis - an economic measure for examining cash outflows (costs) and
cash inflows (revenues) collectively - was also conducted. The NPV simulation output is presented in Figure
7.2 and is accompanied by the NPV output in Table 7.5. Before tax and after tax NPVs were computed. A
tax rate of 28% (NERSA, 2013) was used for the after tax computations. The COALPSCA Model (Figure 7.2)
indicates a fast declining negative NPV (after tax) from year 2010 up until year 2015, which is mainly as a
result of the incurred capital cost coupled with low revenues owing to low plant capacity. From year 2015
up until year 2024, the plant is at full capacity but it is unable to generate enough revenue to cover the
private costs, hence the negative NPV. From year 2025 onwards the NPV becomes incrementally positive.
The cumulative PV revenue (Table 7.5) is estimated at about 274 billion Rands, while the cumulative PV
cost comprising capital, FGD and O&M costs is estimated at about 181.8 billion Rands. The NPV (before tax)
of generating coal in Kusile is therefore estimated at about 92.2 billion Rands (NPV after tax is about 66.4
billion Rands). The positive NPV shows that the investment is economical. Had the NPV been negative it
would have indicated that the returns are worth less than the cash outflows and therefore not a good
investment. A zero NPV would have made the investor indifferent as to whether to make the investment. It
must, however, be noted that the LCOE and the NPV as reported above, exclude the environmental and
societal costs linked with generating electricity in Kusile. The life-cycle externality costs of producing
electricity in Kusile are presented in the following section.
- 180 -
Present value outputs
300 B
R
207.5 B
115 B
22.5 B
-70 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Cumulative PV costs : Baseline
Cumulative PV revenue : Baseline
"NPV (after tax)" : Baseline
Figure 7.2: Present value costs and revenues
7.2.3
Externalities inventory
The externalities quantified and monetized in the COALPSCA Model consist of water use, water pollution,
fatalities and morbidity, ecosystem services loss, air pollution (i.e. classic air pollutants) and GHG emissions.
The externalities were assessed in the entire coal-fuel cycle, excluding the transmission and use phases. The
coal mining phase, plant construction phase with its associated upstream phases, transportation phase,
plant operation phase, and the waste disposal phase were therefore investigated. Water use in the coalfuel cycle over the lifetime of Kusile is presented in Table 7.6.
The water use indicator represents water consumption in the various coal-fuel cycle phases. The COALPSCA
Model estimates water consumption of about 1.1 billion m3 over the life-cycle and lifetime of Kusile. About
37% of the water is consumed during the coal mining phase (an estimate that incorporates coal washing),
while plant operation consumes about 31% of the lifetime water use by Kusile (Table 4.3). The FGD system
also consumes significant quantities of water which constitutes about 22% of Kusile’s lifetime water use.
The FGD system is estimated to consume an annual amount of 5.1 million m 3 of water once Kusile reaches
full capacity. The estimate is comparable with NINHAM SHAND’s (2007) estimates of 5.5 million m3 for a
5400MW plant or 4.9 million m3 if one corrects the estimate to Kusile’s actual size, yielding about 3.9%
variation. On the other hand, plant construction and waste disposal use less water of about 9% and 2%,
- 181 -
respectively. The model further estimates that once Kusile reaches full capacity, annual water consumption
over its fuel cycle will amount to about 21 million m3.
Table 7.6: Coal-fuel cycle water use (Million m3) over Kusile’s lifetime
Coal-fuel cycle phase
Model output
Mining
Plant operation
FGD system
Plant construction & materials inputs
Waste disposal
Total (life-cycle & lifetime)
408.2 (36.6%)
342.5 (30.7%)
248.3 (22.3%)
95.5 (8.6%)
19.4 (1.7%)
1 114
The fatalities and morbidity indicators represent the injuries and deaths that arise in the coal-fuel cycle. For
the baseline scenario, central estimates of fatality and injury rates from various sources linked with the
various coal-fuel cycle phases were applied (see Table 6.4 in the methods section). Table 7.7 presents the
fatalities and morbidity output. The model estimates that approximately 503 deaths are likely to be
suffered by the general public and by occupational personnel over the whole life-cycle and lifetime of
Kusile. The plant operation phase contributes about 90% to the lifetime fatalities, while coal mining
contributes about 10%. No estimates of fatalities on the construction phase could be obtained from Eskom
or from the literature. Construction phase materials inputs fatalities were insignificant at less than 0.1%.
Concerning injuries, the model estimates that approximately 928 persons are likely to be injured over the
lifetime of Kusile. The coal mining phase contributes about 77% to the lifetime injuries, while the plant
operation phase contributes about 23% (this figure includes injuries associated with limestone
procurement). The construction phase material procurement contributes the least at less than 0.1% (this
figure excludes estimates of injuries linked with the construction of the plant due to lack of data). Based on
the model output, the coal mining phase is more prone to injuries than deaths, whereas the plant
operation phase is more prone to deaths than injuries.
Table 7.7: Coal-fuel cycle fatalities and morbidity over Kusile’s lifetime
Coal-fuel cycle phase
Coal mining
Materials inputs
Plant operation & limestone production
Total (life-cycle & lifetime)
Fatalities & morbidity (Units - Persons)
Deaths
Injuries
49 (9.7%)
716 (77.2%)
1 (<0.1%)
<1 (<0.1%)
454 (90.3%)
503
- 182 -
211 (22.7%)
928
Concerning land use, the model focuses on the land area associated with the coal mine and the power
plant. Both areas were used for grazing and crop production, so the mining of coal for Kusile and the
generation of power in Kusile changes the current land uses and the associated benefits derived from their
use. Since the focus of the study is on a specific plant of a particular size, the year on year land area
associated with the power plant and the coal mine is fairly constant. Table 7.8 presents the land uses. The
coal mine occupies an area of 6 817 hectares which was mainly allocated to maize production. 70% of the
land (4 771.9 hectares) was allocated to maize production, with the remaining 30% (2 045.1 hectares) being
under grazing/grasslands (Wolmarans & Madallie, 2011). The power plant occupies an area of 1 456
hectares of which about 96% and 4% were allocated to maize production and grazing/grasslands,
respectively (NINHAM SHAND (2007); Eskom (2010b).
Table 7.8: Coal-fuel cycle land use (Hectares)
Coal-fuel cycle phase
Coal mining
Plant operation
Crop production
Irrigated land
Grasslands
4771.9
1404
Total
Grazing
2045.1
6817
1456
52
Another class of environmental indicators concerns air pollution loads, namely classic air pollutants and
GHGs. Three main classic air pollutants were considered, namely emissions of SO2, NOx and PM originating
from coal transportation, plant construction, plant operation, FGD system and waste disposal. In the
baseline scenario, mainly central estimates of emission factors from a number of sources were applied (see
Table 7.8 in the methods section). Air pollution loads in the coal-fuel cycle over the lifetime of Kusile are
presented in Table 7.9. Concerning coal transportation, the emissions estimated by the COALPSCA Model
reflect coal transportation to Kusile in the early years by road, and then once the conveyor has been
established, transportation mainly by the conveyor, with the remainder transported by road (Synergistics
Environmental Services and Zitholele Consulting, 2011). The model estimates coal transportation emissions
of SO2, NOx and PM of about 42 000 tons, 35 000 tons and 2 700 tons, respectively over the lifetime of
Kusile.
Table 7.9: Coal-fuel cycle classic air pollutant loads over Kusile’s lifetime
Coal-fuel cycle phase
Coal transportation
Construction material transportation
Plant operation
FGD system (limestone transport)
Waste disposal
Total (life-cycle & lifetime)
Units
Tons
Tons
Million tons
Tons
Tons
Million tons
Classic air pollutants
SO2
42 061
5
1.6
5 313
11 521
1.7
- 183 -
%
2.5
<0.01
96.5
0.3
0.7
NOx
35 490
42
3.9
2 744
5 952
3.9
%
0.9
<0.01
98.9
0.1
0.2
PM
2 679
2
0.4
253
548
0.4
%
0.7
<0.01
99.1
0.1
0.1
Coal combustion phase emissions of SO2, NOx and PM are estimated at about 1.6, 3.9 and 0.4 million tons,
respectively over the lifetime of Kusile. The SO2 emissions were adjusted to reflect installation of the FGD
system. The coal combustion phase as shown in Table 7.9 is the highest emitter of the studied classic air
pollutants in the coal-fuel chain. This phase contributes over 95% of Kusile’s lifetime SO2, NOx and PM
emissions. The model further estimates the annual emissions of SO2, NOx and PM from coal combustion in
the order of 33.8, 81.1 and 7.9 thousand tons, respectively once Kusile reaches full capacity (Table 7.10).
NINHAM SHAND (2007) in an EIA for a 5400MW plant, i.e. a plant that is slightly higher than the actual size
of Kusile which is 4800MW, estimates annual tonnages of SO2, NOx and PM of about 36.4 (if one corrects
for 90% SO2 removal efficiency), 87.4 and 7.9 thousand tons, respectively (Table 6.10). Riekert and Koch
(2012) corrected the estimates reported in NINHAM SHAND (2007) for Kusile’s actual size (i.e. 4800MW)
and reports SO2, NOx and PM of about 32.3 (if one corrects for 90% SO2 removal efficiency), 77.7 and 7.8
thousand tons, respectively (Table 7.10). The COALPSCA Model estimates are thus more comparable to the
estimates by Riekert and Koch (2012) with less than 4% variation.
In addition, four trace metal emissions from the coal combustion phase, namely arsenic, chromium, lead
and nickel were estimated. Though the year-on-year trace metals emissions are about a ton, with the
exception of arsenic which is quite low, they are more lethal than most of the other classic air pollutants.
Chromium, nickel, lead and arsenic were estimated at about 46 tons, 35 tons, 24 tons and 4 tons,
respectively over the lifetime of Kusile.
Table 7.10: Annual emissions of classic air pollutants - coal combustion (Thousand t)
Classic air pollutant
SO2
NOx
PM
Model outcomes
NINHAM SHAND
(2007)
%
variation
Riekert & Koch
(2012)
%
variation
33.8
81.1
7.9
36.4
87.4
7.9
7.1
7.2
0
32.3
77.7
7.8
4.4
4.1
1.3
Concerning the FGD system, the emissions estimated by the model reflect emissions associated with the
transportation of limestone to Kusile through electric railway. The model estimates limestone
transportation emissions of SO2, NOx and PM of about 5 300t, 2 700t and 253t, respectively over the
lifetime of Kusile. On another note, the disposal of waste, mainly flyash, which will be transported to the
disposal site through a conveyor, emits SO2, NOx and PM2.5 of about 11 500t, 6 000t and 500t, respectively
(Table 6.9). Waste disposal emits less classic air pollutants mainly because of the efficiency of the conveyor,
which is estimated to consume about 31 786 MWh/Year once Kusile is fully operational. The transportation
of construction materials releases insignificant quantities of the studied classic air pollutants. In addition,
- 184 -
no data were available on the direct emission from the construction of the plant, so the construction phase
air pollution estimates are underestimated.
The main GHGs investigated in the coal-fuel chain are CO2, CH4, and N2O. The various GHGs’ global warming
potentials were used to convert the GHGs into their CO2 equivalence. Five main coal-fuel chain phases were
investigated, namely coal mining and transportation, plant construction, plant operation, FGD system and
waste disposal. The coal-fuel cycle GHG pollutant loads over the lifetime of Kusile are presented in Table
7.11. The model estimates emissions of about 1 583 million tons of CO2e over the coal-fuel cycle and
lifetime of Kusile. About 85% of the GHGs emanate from the combustion phase while coal mining and
transportation contribute about 13%. Plant construction, FGD operation and waste disposal each generate
GHGs of about or less than 1%. The coal combustion phase is thus the main source of GHGs in the coal-fuel
chain. Annual emissions of CO2 and CO2e from coal combustion, once Kusile reaches full capacity, are
estimated by COALPSCA at 27.9 and 28 million tons, respectively (Table 7.12). These estimates are
comparable with NINHAM SHAND’s (2007) estimates of 29.9 and 36.8 million tons for a 5400MW plant or
26.6 and 32.7 million tons if one corrects the estimates to Kusile’s actual size (Table 7.12). In the following
sub-section the quantified externalities are monetized.
Table 7.11: Coal-fuel cycle greenhouse gas pollutant loads over Kusile’s lifetime
GHGs - mainly CO2, CH4 & N2O
Coal-fuel cycle phase
Coal mining & transportation
Construction material & transportation
Plant operation
FGD system
Waste disposal
Total (life-cycle & lifetime)
%
CO2e (Million t)
210.5
8.9
1 348.9
13.5
1.2
1 582.9
13.3
0.6
85.2
0.9
0.1
Table 7.12: Annual emissions of greenhouse gases - coal combustion (Million t)
NINHAM SHAND
NINHAM SHAND
%
Classic air pollutant
Model outcomes
(corrected)
variation
(2007)
CO2
CO2e
7.2.4
27.9
28
29.9
36.8
6.7
23.9
26.6
32.7
%
variation
4.7
14.4
Externality costs
The externalities quantified in the previous section, are monetized in this sub-section. The damage cost
estimates or economic values of the studied externalities are discussed while presenting the results. For the
baseline scenario, all damage costs were escalated at the rate of population growth which is 1.1% (Statistics
South Africa, 2011). The coal-fuel cycle externality cost over the lifetime of Kusile is presented in Table 7.13.
- 185 -
Concerning water, the water price is a fundamental indicator of the availability and cost of supplying water
(Van der Zaag & Savenjie, 2006), nonetheless, in South Africa though water is a resource in critical supply,
the administered water price does not signal the state of water scarcity or reflect the opportunity cost of
the resource. The opportunity cost of water use in Kusile was estimated by Inglesi-Lotz & Blignaut (2012).
The estimates from this study formed the basis of opportunity cost analysis in the coal-fuel chain. The
baseline scenario society-wide opportunity cost of water use in the coal-fuel chain is estimated at about 1
474 billion Rands over the lifetime of Kusile. About 31% of the cost stems from the operation of the plant
while its ancillary water using activities, namely FGD system and waste disposal account for about 23% and
2%, respectively. Coal mining and washing account for approximately 37% of the coal-fuel cycle lifetime
cost while plant construction accounts for about 7%.
Table 7.13: Coal-fuel cycle externality cost (Billion Rands) over Kusile’s lifetime
Coal-fuel cycle phase
Water use
Water
pollution
Fatalities
&
morbidity
Coal mining/transport
551.4 (37.4%)
0.3 (99.9%)
0.0 (20.5%)
Ecosyste
m loss
Classic air
pollutant
GHGs
Grand %
per phase
c/kWh
5.3
5.7 (1.3%)
31.2 (13.3%)
27.3%
37
0.0 (<0.1%)
1.0 (0.4%)
4.6%
6
450.6 (98.3%)
200 (85.3%)
51.2%
70
0.6 (0.1%)
2 (0.8%)
15.6%
21
1.3 (0.3%)
0.2 (0.1%)
1.3%
2
458.2
234.4
21.1%
10.8%
(86.3%)
Plant construction
98.0 (6.6%)
0.0 (0.1%)
0.0 (<0.1%)
Plant operation
462.6 (31.4%)
X
0.2 (79.4%)
FGD system
335.4 (22.8%)
X
X
26.2 (1.8%)
X
X
0.3
0.2
Waste disposal
Total
cost
externality
Grand cost
per
Grand
%
externality
per
1473.5
0.8
(13.7%)
6.1
136
2 172.7
67.8%
<0.1%
<0.1%
0.3%
Turning to water pollution, the water pollution externality was monetised only for the coal mine and for
Kusile’s raw material requirements for building the coal plant. The water pollution associated with the
direct construction and operation of Kusile was excluded, owing to Eskom’s stated zero effluent discharge
policy (NINHAM SHAND, 2007)). Regarding the water pollution damage cost, adapted to this study were
direct damage costs of sulphate pollution from various industries on other water users in the eMalahleni
area, estimated by Van Zyl et al. (2002). The shortcomings of the Van Zyl et al. (2002) study include its focus
on sulphate and not all pollutants, its focus on impacts in the catchment and not downstream and lack of
address of natural/environmental uses. Owing to the Van Zyl study’s shortcomings, the estimates
computed in the current study are considered conservative. The COALPSCA Model estimates the water
pollution externality at about 0.3 billion Rands and is more or less wholly associated with the mining of coal
(99.9%).
- 186 -
Concerning the economic value for morbidity, cost estimates estimated using the cost-of-illness approach
by Van Horen (1997), through discussions with public health practitioners in South Africa were transferred
to this study by adjusting the values for inflation and some form of internalization. The economic values for
mortality were based on valuation of changed life expectancy, obtained from the NEEDS (2007) and NewExt
(2004) studies. The values were adjusted to reflect the disparity of income levels between the EU and South
Africa and to cater for inflation and some form of internalization. For the baseline scenario, central cost
estimates for morbidity and mortality were used. The COALPSCA Model estimates monetary estimates for
morbidity and mortality of about 0.2 billion Rands over the life-cycle and lifetime of Kusile. About 79% of
this cost is attributable to the plant operation phase while approximately a one-fifth is attributable to the
coal mine.
Pertaining to land use, the extraction of the coal resource and the establishment and operation of the coal
plant will lead to loss of farmlands and grasslands. The opportunity cost of these activities is therefore the
forgone benefits derived from agricultural production and ecosystem services generated by grasslands. The
market price of maize and the maize yield per hectare for dry and irrigated land were used to compute the
foregone benefit from maize production, while the value of ecosystem goods and services generated by
grasslands was adapted from a study undertaken by Blignaut et al. (2010) for the Maloti–Drakensberg
mountain range of South Africa. The model estimates an ecosystem services loss of about 6 billion Rands
over the life-cycle and lifetime of Kusile. About 86% of the lifetime loss is linked with the surface coal mine
while the remainder is attributable to the power plant and its ancillary activities. Since the power plant and
associated structures occupy a few hectares, most of the power station related ecosystem services loss is
associated with waste disposal.
On another note, the health cost of air pollution in the coal fuel chain was estimated based on damage
costs per ton of a specific local air pollutant adapted from NEEDS (2007; 2008; 2009), Sevenster et al.
(2008) and from AEA Technology Environment (2005). The original estimates are based on value of a life
year (VOLY) lost due to air pollution, estimated using change of life expectancy. Basing the valuations of air
pollution mortality on the change of life expectancy, as opposed to a valuations based on accidental death
or a small change in the probability of dying is more appealing because the approach automatically factor in
the constraint that humans die only once irrespective of pollution, it offers a unified framework for time
series, cohort and intervention studies plus directly yields the life expectancy change as a time integral of
the observed mortality rate (Rabl, 2006). In addition, change in life expectancy is further favourable
because respondents during surveys show too much difficulty understanding small probability variations
while a change in life expectancy is well understood (NewExt, 2003). The estimates were adjusted to reflect
- 187 -
the disparity of income levels between the EU and South Africa and to cater for inflation and some form of
internalization. For the baseline scenario central cost estimates of the three main classic air pollutants
studied were used. The model estimates the health cost of air pollution at about 458 billion Rands over the
life-cycle and lifetime of Kusile. The coal combustion phase makes up most of the air pollution health cost
at 98%, followed by the coal mining phase at about 1%. Waste disposal, FGD system and plant construction
make up the least of the lifetime air pollution cost of about 0.3%, 0.1% and less than 0.1%, respectively.
Turning to global warming, the global warming damage cost associated with GHG releases in the coal fuel
chain was quantified through the application of a range of CO2 damage cost estimates, ranging between
R5.86/tCO2 – R820.56/tCO2, adapted from a study by Blignaut (2012). Blignaut (2012) highlights that the
arguable most probable range of the global warming damage cost is given by the market, median and high
damage rates (i.e. damage cost ranging between R104.98/tCO2 – R177.94/tCO2). These are the damage
costs used in this study. For base case analysis, the market estimate was used (i.e. R109.89/tCO2). The
GHGs were adjusted to reflect their global warming potential alongside that of carbon dioxide. The National
Treasury in South Africa also plans to impose a tax on emitters of greenhouse gases of R120/ton of carbon
dioxide equivalence (CO2e) above the tax-free threshold (BDFM Publishers, 2013d). The National Treasury’s
proposed tax therefore falls within the range used in this study and its impacts are specially discussed later
in chapter 7 together with other carbon tax scenarios. The model estimates a CO2e global warming damage
cost of about 234 billion Rands over the coal-fuel cycle and lifetime of Kusile. The coal combustion phase
generates the majority of this cost (85%), followed by the coal mining phase at about 13%. GHG damages
from the transportation and use of limestone in the FGD system are about 0.8% while even lower estimates
are associated with plant construction and waste disposal.
The total coal-fuel cycle externality cost over the lifetime of Kusile is estimated at about 2 173 billion Rands
(Table 7.13). Most of the externality cost stems from three groups of externalities, namely the water use
externality, air pollution health cost and the global warming damage cost which accounts for about 68%,
21% and 11%, respectively. Table 6.13 further discloses the grand distribution of the total externality cost
per coal-fuel cycle phase over the lifetime of Kusile. The plant combustion phase, FGD system and waste
disposal house about 51%, 16% and 1.3% of the total coal-fuel cycle externality cost, respectively. The
operation phase with its ancillary activities (i.e. FGD system and waste disposal), accordingly accounts for
about two thirds of the externality cost. A significant amount of the total coal-fuel cycle externality cost
also stems from the mining and transportation of coal, which accounts for almost a third (27%) of the cost.
The plant construction phase houses about 5% of the lifetime cost. Three main coal-fuel cycle phases thus
contribute the most to the total externality cost over the lifetime of Kusile, namely plant operation, FGD
- 188 -
system operation and the mining and transportation of coal. Collectively the three phases make up about
94% of the lifetime externality cost.
Based on Kusile’s lifetime electricity production of about 1.6 billion MWh, the base case construction phase
externality cost is about 6c/kWh (see Table 7:13). There are, however, no studies locally to compare this
estimate with and the two international studies (European Commission, 1995; 1999b) that study the
construction phase do not report explicitly the externality cost linked with this phase. The externality cost
of mining and transporting coal to Kusile is about 37c/kWh while that of the power plant (including waste
disposal but excluding FGD system) is 72c/kWh. The FGD system externality cost is about 21c/kWh while
the FGD system and the power plant combined produce an externality cost of 93c/kWh, which is about
100% of the electricity price that will prevail at the end of the simulation (93c/kWh). The base case coal-fuel
cycle externality cost thus amounts to 136c/kWh (see Table 7:13) (when converted to US cents/kWh it is
about 19c/kWh) and falls in the middle range of the international externality cost studies reported in Table
4.3, while clearly above most of the local studies (in Table 4.5) because of the inclusion of more
externalities and coal-fuel cycle phases. The combined externality cost estimate of Nkambule & Blignaut
(2012), Riekert and Koch (2012), Inglesi-Lotz and Blignaut (2012), and Blignaut (2012) reported in Table 4.5
(i.e. 4.23 – 25.66 in US cents/kWh or in South African Rands 31c/kWh – 188c/kWh), shows a rather higher
externality cost compared to the base case estimate computed in the current study, irrespective of that
more externalities and fuel-cycle stages are being included in this study because the four collective studies’
estimates are low to high estimates whereas the COALPSCA Model value is a baseline value.
A further look at Table 7.13 discloses the plant operation phase externality cost to be connected with water
use, air pollution and GHG emissions, while the coal mining and transportation externality cost is mainly
associated with water usage, air pollution, GHG emissions and loss of ecosystem services due to the
disruptive nature of a surface mine on land. On the other hand, the FGD system externality cost mainly
stems from water use.
Interestingly, the installation of the FGD system increases water use while curbing SO2 emissions. So in
order to explore this interesting trade-off between water use externality and human health cost savings,
simulated was the lifetime air pollution health cost and opportunity cost of water use with and without the
installation of the FGD system. Since the FGD system is linked with the coal combustion phase in that the
air pollution health cost savings are revealed in the coal combustion phase, the air pollution health cost and
the water use externality cost were quantified for the coal combustion phase jointly with the FGD system
and waste disposal phases. Table 7.14 reports the outcomes of this application.
- 189 -
Without the installation of the FGD system, the air pollution health cost and water use externality cost are
estimated at about 1 472 and 489 billion Rands, respectively. Fitting the power plant with an FGD system
reduces the air pollution health cost to about 453 billion Rands while increasing the water use externality
cost to about 824 billion Rands. These outcomes disclose that the installation of the FGD system introduces
an extra water use externality cost of about 335 billion Rands while creating air pollution health cost
savings of about 1 019 billion Rands. For this reason, the installation of the FGD system is a sensible effort
(on the grounds of externality cost versus externality cost savings) since its air pollution health cost savings
outweigh the water use externality cost it introduces (positive net change of about 684 billion Rands).
Water is, however, a scarce resource in South Africa and human health is without doubt valuable, so the
country and its people need to decide what it is willing to forego in order to gain the other. Give-up water
in exchange for clean air and hence gain better human health or vice versa. On the other hand, in order for
one to reach a final conclusion about the economic viability of the FGD system, the private and externality
costs associated with the FGD system need to be fully paid off by the savings.
Table 7.14: FGD system or not, costs and savings (Billion Rands) over Kusile’s lifetime
Externality cost
Water use externality cost
Air pollution health cost
Net change
7.2.5
With FGD system
No FGD system
A
B
824.2
C
452.5
488.8
D
1471.7
FGD system installation extra
externality cost or savings
B - A = -335.4 (cost)
D - C = 1019.2 (savings)
Savings + cost = 683.8
Social cost
In the previous sections the private costs and the externality costs of energy were reported. In this section
both costs are examined jointly since the true cost of energy is composed of not only the price of electricity
that is reflected on electric bills (i.e. private costs) but also the less obvious negative impacts of electricity
generation on third parties, for example, on the environment and on society. A more holistic accounting of
the full cost of energy thus embraces both the private costs and the externality costs and is known as the
social cost of energy (Greenstone & Looney, 2012).
A number of economic indicators were computed and reported in the previous sections, including the
LCOE, the NPV and externality costs. In this section, three additional economic indicators are computed,
namely the levelised externality cost of energy (LECOE), the levelised social cost of energy (LSCOE) and the
social net present value (SNPV). The LECOE and the LSCOE were estimated in a similar manner as done for
the LCOE – thus they are measured in R/MWh. The LECOE and LSCOE outcomes are reported in Table 7.15.
The levelised externality cost of energy is estimated by the model at about R1 370.8/MWh. The LECOE,
- 190 -
when added to the LCOE (computed earlier amounting to R554.2/MWh), yields a levelised social cost of
energy (LSCOE) of about R1 925/MWh. The LCOE thus reflects about 29% of the true cost of coal while the
externality cost makes up approximately 71% of the social cost of energy. A little over two thirds of the true
cost of electricity therefore does not reflect on the balance sheet of the utility and is borne by society.
Comparing the LECOE estimated in this study of R1 370.8/MWh, to the four collective studies’ (i.e.
Nkambule & Blignaut (2012), Riekert and Koch (2012), Inglesi-Lotz and Blignaut (2012), and Blignaut (2012))
externality cost for Kusile conducted for the year 2010, of between R310/MWh – R1 880/MWh (2010
values), the collective studies’ estimates are comparable but slightly higher compared to the base case
value computed by COALPSCA as they are low to higher range estimates. Another reasons that could
slightly elevate the four collective studies’ externality costs above those computed in this study could be
the downward adjustment of the opportunity cost of water in the current study (attributed to the fact that
renewable technologies are not yet on large enough scales enabling them to uptake/utilize the water or to
generate electricity analogous to Kusile), and the air pollution health costs and fatalities and morbidity
costs internalization that was accounted for in this study.
Table 7.15: Levelised externality and social cost of energy (R/MWh) over Kusile’s lifetime
Present value output
Model output
Levelised externality cost of energy (LECOE)
Levelised social cost of energy (LSCOE)
1 370.8
1 925.0
The social net present value (SNPV) on the other hand, is synonymous with social benefit-cost analysis in
that it aims to compare the benefits and costs of a project/action by taking into consideration both the
private and externality costs and benefits. The SNPV approach discounts these costs over the lifetime of the
investment to arrive at a present value measure. By definition, the SNPV ought to also incorporate positive
externalities. In this study the SNPV reflects the present value of investing in Kusile (i.e. private
benefits/returns less private costs) less the present value of the externality costs.
The selected present value output is presented in Table 7.16, accompanied by the present value simulation
output in Figure 7.3. The NPV before tax of generating coal in Kusile was earlier reported as 92.2 billion
Rands (NPV after tax is about 66.4 billion Rands), this, coupled with the cumulative PV externality cost of
449.5 billion Rands yields a SNPV (before tax) of -357.3 billion Rands (SNPV after tax is about -383.2 billion
Rands). Figure 7.3 shows a negative SNPV after tax (alike for before tax) throughout the lifetime of Kusile,
highlighting that year on year Kusile will be unable to generate enough revenue to cover the negative
- 191 -
externalities it imposes on third parties. The earlier computed positive NPV shows that the investment is
economical and passes a private cost benefit analysis, but when the externality effects of the investment on
third parties are incorporated, the project is no longer acceptable as it does not generate positive net social
benefits, but significant externalities that impose a large externality cost.
Table 7.16: Selected present value output (Billion Rands)
Present value output
Model outcomes
NPV (before tax)
NPV (after tax)
Cumulative PV externality cost
SNPV (before tax)
SNPV (after tax)
92.2
66.7
449.5
-357.3
-383.2
Social NPV
500 B
R
250 B
0
-250 B
-500 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
"NPV (after tax)" : Baseline
Cumulative PV externality cost : Baseline
"Social NPV (after tax)" : Baseline
Figure 7.3: Social NPV
In the previous sections the private costs and the externality costs of energy were reported. In this section
both costs are examined jointly since. A more holistic accounting of the full cost of energy thus embraces
both the private costs and the externality costs and is known as the social cost of energy (Greenstone &
Looney, 2012).
- 192 -
7.3
Model validation
Model validation involves repeated actions of testing and establishing confidence in the model (Forester &
Senge, 1980; Sterman et al., 1988). This process runs through the entire process of model building,
beginning with model conceptualization up until implementation of policy recommendations (Forester &
Senge, 1980; Sterman et al., 1988). Based on models being simple representations of actual-world
situations, they can, however, never be fully validated (Sterman, 2000) and in addition, no particular test
can completely verify a model but the confidence in a model is improved as the model passes a range of
tests (Forrester & Senge, 1980). Forrester (1961) furthermore emphasizes that model validation ought to
be judged with reference to a particular purpose, that is, detached from purpose, model validity is
worthless. This is considered important for system dynamics models because they are built to fulfill a
purpose (Holling, 1978; Barlas & Carpenter, 1990).
In system dynamics, the internal structure of the model needs to be validated first, followed by validation
of model behaviour. The accuracy of model behaviour is only meaningful once adequate confidence on
model structure has been established beforehand (Barlas, 1989; Barlas, 1994). This sensible order of model
validation is not difficult to comprehend, since the usefulness of a system dynamics model lies in its
capability to relate patterns of behaviour of a system to the structures that underlie the system (QudratUllah, 2012). That is, system dynamics models are causal models, which seek to understand how the
internal structure of a system helps to create visible patterns of behaviour of a system. Hence structural
validity comes first, followed by behaviour validity which seeks to establish how well the model generated
behaviour mirrors the behaviour of a real system. Model validation thus seeks to establish – whether the
model is acceptable, given its purpose (Goodall, 1972; Forrester and Senge, 1980; Zebda, 2002); and to
establish the degree of confidence to place in the model based on inferences of an actual system (Curry,
Deuermeyer & Feldman, 1989; Barlas, 1994; Sterman, 2000). Though lack of formal validation tools is
regularly the critique of system dynamics methodology (Barlas, 1994), the literature discloses a number of
validation tests which are described below. An explanation is also given of how they were used in this
study.
7.3.1
Structural validity
Structural validity concerns establishing validity with regards to the internal structure of the model. These
tests involve comparing model structure versus knowledge of the real system or versus general knowledge
of the system as evidenced by literature (Barlas, 1994). Five direct structure validation tests were
introduced by Forrester and Senge (1980) for system dynamics, namely boundary accuracy, structure
- 193 -
verification, dimensional consistency, parameter verification, and extreme condition tests and were
conducted in this study. The model boundary was discussed in section 6.5, so it is not reported here.
7.3.1.1
Structure verification test
This test concerns comparing model structure versus the actual system structure/knowledge in the
literature. It assesses the consistency of model structure with the descriptive knowledge of the real system
actuality modelled (Forrester & Senge, 1980). For structural verification three approaches were used.
Firstly, when developing the causal relationships in the model, Eskom- and Kusile-specific data were used,
that is, available knowledge of the (currently under construction) system. The conceptual model of the
modelled system was presented by the causal loop diagram in the methods section. It was shown that
investment in electricity generation increases plant capacity, which boots power generation, which then
generates revenues and profits for utility owners. At the same time an increase in power generation
triggers an increase in resource input use, for example coal which increases the fuel cost, which together
with other material/resource requirements, increases the private costs of generating power. On another
note, the increase in power generation directly and indirectly (e.g. through upstream services) produces
negative externalities, for example GHG emissions, classic air pollutants, injuries, fatalities, water pollution,
loss of ecosystem services and water consumption externality. These burdens, coupled with the likely
damages they impose on humans and on the environment, signify the externality costs which, together
with the private costs of generating power (capital, fuel and O&M costs), intensifies the social cost, which
then negatively affects the revenues and profits earned by utility owners. The causal relationships of the
COALPSCA Model were founded on available knowledge of the real system, and for that reason they served
as a form of empirical structure validation (Zebda, 2002).
Secondly, all the stock variables by definition should either be positive or zero, but not negative. So as the
stocks approaches zero so should the outflows from the stocks. This test was conducted for the stocks and
flows in the COALPSCA Model. Thirdly, the validity of each of the model equations against available
knowledge was conducted by directly comparing each of the model equations with the (currently under
construction) real system (empirical) and with generalized knowledge of the system existing in the
literature (theoretical). As an example of how model equations were evaluated see, Table 7.17. In the light
of these tests, the COALPSCA Model was found to be a reasonable, simplified match of the real-world
system.
- 194 -
Table 7.17: Examples of structure test
Model equation
Available knowledge on real system
CConsump  ((GEP * MWhtoKWh) * HR) / CEC/ kgtoton
GEP  ( FCC * POH )  ( DFCA * POH ) * LF 
7.3.1.2
Coal consumption (CConsump) is the amount of coal consumed by the
plant and is a function of the coal energy content (CEC), heat rate of the
plant (HR) and gross electricity production (GEP).
For a system that is evaluated at the farm gate, gross electricity
production is the quantity of power produced by the plant and is not net
of the amount of power internally consumed by the plant. Gross
electricity production is a function of the plant operating hours (a
variable in the developed model that was adjusted for the time the
plant will be shut down for maintenance, i.e. energy availability factor),
load factor and the plant functional capacity [in the developed model,
functional capacity was separated into functional capacity during
construction (FCC) and desired functional capacity after construction
(DFCA)].
Dimensional consistency test
The dimensional consistency test intends to establish the unit’s uniformity of all model equations. That is,
for each model equation, the measurement units of all the variables in it must be dimensionally consistent
without including scaling parameters that in the real world have no meaning (Forrester & Senge, 1980;
Sterman, 2000). So the measurement units enable the checking of dimensional consistency of model
equations. Accordingly, the dimension of input variables of the COALPSCA Model equations was examined.
In addition, the menu item Model>Units Check was used to check the COALPSCA Model equations in
totality.
7.3.1.3
Parameter verification test
The parameter verification test concerns the conceptual and numerical evaluation of constant parameters
of the model against knowledge of the actual system. It assesses the consistency of the model parameters
against the system’s descriptive and numerical knowledge (Forrester & Senge, 1980). The values allocated
to COALPSCA Model parameters were obtained from existing knowledge of the system, coupled with
available numerical data on Kusile and its associated processes. As an illustration, Table 7.18 presents the
main input parameters, baseline values and data sources used in the power generation sub-model (a
comprehensive list of parameters used in the COALPSCA Model are presented in table form in the methods
section, after the discussion of each sub-model).
- 195 -
Table 7.18: Selected parameters, values and data sources - power generation
Variable
Coal energy content
Days per year
Fraction of electricity consumed
internally
Heat rate
Hours per day
Load factor
Planned investment in plant capacity
table
7.3.1.4
Units
Baseline value
Data source
MJ/kg
Day/Year
19.22
365
Eskom, 2010a.
Dmnl/Year
0.075
Eskom communication, 2012.
MJ/kWh
h/Day
Dmnl/Year
9.769
24
0.9
EPRI, 2010.
MW/Year
Time series
Eskom communication, 2012.
Eskom communication, 2012.
NINHAM SHAND. 2007
Calculated based on Eskom (2012b).
Extreme condition test
This test concerns assigning extreme values to certain parameters and evaluating the plausibleness of the
model reproduced behaviour versus knowledge/anticipation of what may take place in comparable
conditions in real life. For the model to pass this test, it must demonstrate logical behaviour under extreme
conditions (Forrester & Senge, 1980). Two extreme condition tests are presented in this section.
In extreme condition test 1, the planned investment in plant capacity table was set to zero, which in reality
means no capital investment into electricity production, therefore no capacity construction and
consequently no electricity production and no GHGs being emitted by the power plant (e.g. CO2). The
COALPSCA Model outcomes for this condition are presented in Figure 7.4 and are in agreement with this
extreme condition.
- 196 -
Capacity construction start
2000
30 B
1500
MW/Year
R/Year
Capacity investment
40 B
20 B
10 B
0
2010 2015
1000
500
2020
2025
2030 2035 2040
Time (Year)
2045
2050
2055
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
2060
Capacity investment : Baseline
Capacity investment : Extreme condition test 1
Capacity construction start : Baseline
Capacity construction start : Extreme condition test 1
Coal combustion CO2 emissions
30 M
30 M
22.5 M
ton/Year
MWh/Year
Gross electricity production
40 M
20 M
10 M
0
2010 2015
15 M
7.5 M
2020 2025
2030 2035 2040 2045
Time (Year)
2050 2055
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
2060
Gross electricity production : Baseline
Gross electricity production : Extreme condition test 1
Coal combustion CO2 emissions : Baseline
Coal combustion CO2 emissions : Extreme condition test 1
Figure 7.4: COALPSCA Model behaviour under extreme condition test 1
In extreme condition test 2, the unit cost of coal was grown to R1 100/ton from the baseline value of
R210/ton. In reality, with such escalation in the price of coal, the fuel cost (or coal cost) will soar high,
increasing the private costs of power generation by a great margin, because the fuel cost is a major cost
component of the generation cost, making up over 46% of the utility’s generation cost in the base case (see
Table 7.5). The higher private cost would negatively affect the profitability of coal-based power and the
utility’s ability to repay its debt. The simulation results in Figure 7.5 accurately depict this extreme
condition.
- 197 -
Plant private costs
60 B
15 B
45 B
R/Year
R/Year
Coal cost
20 B
10 B
5B
30 B
15 B
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Coal cost : Baseline
Coal cost : Extreme condition test 2
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Plant private costs : Baseline
Plant private costs : Extreme condition test 2
NPV (before tax)
NPV (after tax)
100 B
50 B
0
0
R
100 B
R
200 B
-100 B
-50 B
-200 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
"NPV (before tax)" : Baseline
"NPV (before tax)" : Extreme condition test 2
-100 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
"NPV (after tax)" : Baseline
"NPV (after tax)" : Extreme condition test 2
Figure 7.5: COALPSCA Model behaviour under extreme condition test 2
The COALPSCA Model behaviour under extreme conditions mimics the anticipated behaviour of the actual
system under comparable extreme conditions. The model therefore passes the extreme condition test and
model validity is improved. Behaviour validity is discussed in the following section.
7.3.2
Behaviour validity
Behaviour validity seeks to establish how well the model produced behaviour matches the behaviour of the
real system (Barlas, 1996.) The focus is on patterns. Among the behaviour validation tools are the
behaviour sensitivity test, reference test, modified-behaviour prediction test and a face validity test.
7.3.2.1
Face validity test, reference test and modified-behaviour prediction test
A face validity test can be used when simulation models are applied to operational problems. In this test,
experts evaluate the closeness of the model and its outcomes to the real system (Zebda, 2002). A face
validity test, could not, however, be undertaken because Kusile power station is still under construction,
and therefore not yet in operation. On another note though, one can argue that it had been evaluated
since the thesis and the articles produced from it were evaluated by experts when sent to external
- 198 -
examiners or journals. In a reference test, the model is simulated a few years back and the model outcomes
are compared to historical data. However, given that Kusile is still under construction and that Eskom was
not willing to share data relevant to this research on existing plants, the reference test could not be
performed. A modified-behaviour prediction test is only possible if data on the modified patterns of the
real system can be sourced - in which case the model then passes the test on condition it can mimic the
modified behaviour (Forrester & Senge, 1980). Data absence/inaccessibility, as explained earlier, prevented
execution of these tests.
7.3.2.2
Behaviour sensitivity test/sensitivity analysis
The behaviour sensitivity test seeks to uncover the parameters the model is responsive to and questions
whether the real system would also display higher responsiveness to the said parameters (Barlas, 1994).
This test is synonymous with sensitivity analysis. The aim of sensitivity analysis therefore is to study the
effects of variations in model assumptions on model results (Saltelli et al., 2000). The assumptions may be
about parameter values or feedback loops and they portray uncertain information that cannot be gathered
from real life observations. System dynamics model parameters are subject to uncertainty, so sensitivity
analysis is a noteworthy task for the reliability of simulation results (Hekimoğlu & Barlus, 2010).
The frequently used system dynamics approaches to deal with uncertainty include univariate sensitivity
analysis and multivariate sensitivity analysis (Pruyt, 2007). Two types of sensitivity analysis were conducted
in this study, namely univariate and multivariate sensitivity analysis. Univariate (one-way/one-at-a-time)
sensitivity analysis was conducted through varying the value of one parameter at a time while holding all
other parameters constant at their base case value. This form of sensitivity analysis can highlight the most
influential parameters in the model outputs but it is insufficient for a complete investigation of the model
in nonlinear and complex models due to that nonlinear relationships among model components may
produce unanticipated output change when simultaneous changes in more than one parameter values
occur (Sterman, 2000). For this reason univariate sensitivity analysis is often followed by multivariate
sensitivity analysis, which assesses the effects of simultaneous change of several variables on model
outputs (Monte Carlo simulation).
Monte Carlo simulation also known as multivariate sensitivity simulation examines the future likelihood of
output variables of importance through running a large amount of simulations by repeatedly drawing
samples from probability distributions of uncertain variables. Given the uncertain parameters, confidence
bounds are utilised for demonstrating model outputs. Like with any long-term analysis there is uncertainty
about the costs and the technical factors in a coal-fired power plant. The sensitivity analysis in this study
- 199 -
focused on the load factor, discount rate, cost growth rates of all private costs in the model (e.g. coal,
limestone, water, O&M and capital costs), cost growth rates of all damage cost estimates and the
sensitivity of the model outcomes to lower and higher range estimates. Uncertainties concerning such
variables are a reality for energy markets (International Energy Agency, 2010) so all these parameters are
important input variables that can affect among other factors the production of coal-based power, total
generating cost of power, LCOE, LECOE and the financial viability of coal based power. To assess the impact
of these parameters on selected model outcomes, minimum and maximum values were assigned to each of
them along with a random distribution over which to vary them. Table 7.19 shows the range of values
assigned to each of the uncertain parameters with the exception of lower and higher range damage costs
estimates which are shown in Figure 7.20. The number of simulations was set at 400 and the random
uniform distribution was used.
Table 7.19: Minimum and maximum parameter values versus baseline values
Parameter
Discount rate
Private cost growth rates (i.e.,
coal, O&M & capital cost, etc.)
Load factor
Damage cost growth rates
Units
Baseline value
Minimum – Maximum
0.08 (8%)
0.04 – 0.12 (4% to 12%)
Dimensionless/Year (%)
0.001 (0.1%)
-0.05 – 0.05 (-5% to 5%)
Dimensionless (%)
Dimensionless/Year (%)
0.9 (90%)
0.011 (1.1%)
0.85 – 0.95 (85% to 95%)
-0.0055 – 0.0165
Dimensionless (%)
Concerning the lower and higher range damage costs estimates which are shown in Table 7.20, apart from
univariate and multivariate sensitivity analysis I also conducted manual sensitivity testing (i.e. through
changing the value of a constant one at a time and simulating) because it delivers more insightful findings
that are key to this study. The outcomes of this exercise are reported last after univariate and multivariate
sensitivity analysis.
Table 7.20: Lower and higher range damage cost estimates versus baseline values
Variable
Unit morbidity value
Unit mortality value
Unit opportunity of water use
Steel embodied water
Water requirements of a surface mine (in litres/ton)
Unit damage cost of sulphate pollution from coal mining
Unit damage cost of sulphate pollution from steel production
Unit damage cost of sulphate pollution from Al & concrete
production
Maize yield per hectare (dry land)
Unit damage cost SO2
Unit damage cost NOx
Unit damage cost PM
Unit damage cost CO2
- 200 -
Units
R/person
R/person
R/m3
m3/ton
l/ton
R/ton
R/ton
R/ton
Lower
9 130
69 285
669
200
431
0.19
0.58
0.14
Base case
25 434
245 438
1 001
225
469
0.27
0.79
0.31
Higher
59 998
771 700
1 331
250
581
0.34
0.99
0.48
ton/ha
R/ton
R/ton
R/ton
R/ton
3.5
29 025
26 735
116 739
104.98
4.25
51 619
41 952
227 175
109.89
5
86 778
64 689
402 332
177.94
The univariate sensitivity analysis outcomes for variations in discount rate, load factor, private cost growth
rates, damage cost growth rates, and lower and higher damage costs estimates are presented in Figure 7.6,
7.7 7.8, 7.9 and 7.10, respectively in form of confidence bounds for selected output variables while the
multivariate sensitivity analysis outcomes are reported in Figure 7.11. In all the figures the base case run is
shown by the solid (blue) line (i.e., run name sensitivity).
Focusing on Figure 7.6 which shows the sensitivity of the LCOE and NPV after tax to variations in discount
rate, the base case run indicates a fast declining negative NPV after tax from year 2010 up until year 2015,
which is mainly as a result of the incurred capital cost coupled with low revenues owing to low plant
capacity. From year 2015 up until year 2024, the plant is at full capacity but it is unable to generate enough
revenue to cover the private cost, hence the negative NPV. From year 2025 onwards the NPV becomes
incrementally positive and by the end of the simulation (year 2060) it is estimated at 66.4 billion Rands.
Given the uncertainties in the discount rate, the 100% confidence bounds suggest that the NPV after tax
could range from R5 billion to R217 billion by the end of the simulation while the LCOE could range from
R378/MWh to R775/MWh by the end of the simulation (the base case run for the LCOE by the end of the
simulation is estimated at about R554/MWh). Figure 7.6 therefore reveals a wide band of uncertainty on
the simulated LCOE and NPV after tax but coal-based power could still be a viable enterprise.
Sensitivity
50%
75%
95%
Sensitivity
50%
75%
100%
"NPV (after tax)"
300 B
150 B
3750
0
2500
-150 B
1250
-300 B
2010
95%
100%
Levelised cost of energy
5000
2023
2035
Time (Year)
2048
2060
0
2010
2023
2035
Time (Year)
2048
2060
Figure 7.6: Confidence bounds for discount rate (range: 0.04 to 0.12) on selected model outcomes
Turning our attention to Figures 7.7 which shows the sensitivity of model outcomes to variations in private
cost growth rates, the confidence bounds of the NPV after tax and the LCOE show the same general
patterns as in Figure 7.6. The 100% confidence bounds unveil that the total generation cost (i.e., cumulative
private costs), LCOE and the NPV after tax could range between R217 billion – R980 billion (base run
cumulative private costs are estimated at R410 billion), R464/MWh - R808/MWh and between R4 billion –
- 201 -
R90 billion, respectively by the end of the simulation. The selected model outcomes are therefore sensitive
to variations in cost growth rates but their bands of uncertainty are narrow than those effected by
variations in discount rates. Also the project is still economically viable.
Sensitivity
50%
75%
95%
Sensitivity
50%
75%
100%
Cumulative private costs
1T
750 B
3750
500 B
2500
250 B
1250
0
2010
Sensitivity
50%
75%
95%
100%
Levelised cost of energy
5000
2023
95%
2035
Time (Year)
2048
2060
0
2010
2023
2035
Time (Year)
2048
2060
100%
"NPV (after tax)"
90 B
45 B
0
-45 B
-90 B
2010
2023
2035
Time (Year)
2048
2060
Figure 7.7: Confidence bounds for private cost growth rates (range: -0.05 to 0.05) on selected model
outcomes
On the other hand, the plant load factor which is the ratio of power produced by a power plant over the
theoretical maximum it could produce at full capacity over a time period is a key variable to the economics
of power generation as it is useful for predicting the amount of electric power production per unit of
generating capacity that would earn revenues to cover the generation cost of a power plant. There is
generally an inverse relationship between the load factor and the plant-costs/ LCOE, because the higher
the load factor, the lower the generation cost per MWh due to that the higher the load factor the more
electricity is produced and the more the private costs of the plants are distributed across the electricity,
basically making power production cheaper. The response of the plant private costs and LCOE to variations
in load factor is shown in Figure 7.8. Figure 7.8 displays even more narrow bands of uncertainty on the
simulated private costs and LCOE than those in Figure 7.6 and 7.7. This could be partly attributable to that
Kusile is planned to be a base load power station (Eskom, 201c; Eskom communication, 2013), so it is likely
- 202 -
going to be operated at higher load factors. The univariate sensitivity analysis outcomes highlight the
important drivers of the generation cost of coal-based power to be the discount rate, cost growth rates and
the load factor in descending order.
Sensitivity
50%
75%
95%
Sensitivity
50%
75%
100%
Cumulative private costs
500 B
375 B
3750
250 B
2500
125 B
1250
0
2010
95%
100%
Levelised cost of energy
5000
2023
2035
Time (Year)
2048
0
2010
2060
2023
2035
Time (Year)
2048
2060
Figure 7.8: Confidence bounds for load factor (range: 0.85 to 0.95) on selected model outcomes
Now turning to the externality related uncertainties, given the uncertainties in the damage cost growth
rates of +/- 50% of the base case growth rate (i.e. base case 0.011%), the 100% confidence bounds suggest
that the coal-fuel cycle externality costs, levelised externality cost of energy and social NPV after-tax could
range between R24 billion –R72billion (base run at R55 billion), R1 131/MWh to R1 470/MWh (base run at
R1 370/MWh) and –R303 billion to –R416 billion to (base run at -R383.2 billion), respectively by the end of
the simulation (Table 7.9) while sensitivity of coal-fuel cyle externality costs, levelised externality cost of
energy and social NPV after-tax could range between R42 billion – R68 billion, R1 046/MWh to R1
693/MWh and –R276 billion to –R490 billion in the case of low and high damage costs estimates by the end
of the simulation (Table 7.10). The selected model outcomes are therefore sensitive to variations in
damage cost growth rates but their bands of uncertainty are narrow than those to variations in low and
high damage costs estimates.
- 203 -
Sensitivity
50%
75%
95%
Sensitivity
50%
75%
100%
60 B
3750
40 B
2500
20 B
1250
0
2010
Sensitivity
50%
75%
2023
95%
2035
Time (Year)
2048
2060
0
2010
Sensitivity
50%
75%
100%
Levelised social cost of energy
9000
-125 B
4500
-250 B
2250
-375 B
2023
100%
2023
95%
2035
Time (Year)
2048
2035
Time (Year)
2048
2060
100%
"Social NPV (after tax)"
0
6750
0
2010
95%
Levelised externality cost of energy
5000
"Coal-fuel cycle externality costs"
80 B
2035
Time (Year)
2048
2060
-500 B
2010
2023
2060
Figure 7.9: Confidence bounds for damage cost growth rates (range: -0.0055 to 0.0165) on selected model
outcomes
- 204 -
Sensitivity
50%
75%
95%
Sensitivity
50%
75%
100%
"Coal-fuel cycle externality costs"
80 B
60 B
4500
40 B
3000
20 B
1500
0
2010
Sensitivity
50%
75%
2023
95%
2035
Time (Year)
2048
0
2010
2060
Sensitivity
50%
75%
100%
Levelised social cost of energy
10,000
-125 B
5000
-250 B
2500
-375 B
2023
100%
2023
95%
2035
Time (Year)
2048
2060
2035
Time (Year)
2048
2060
100%
"Social NPV (after tax)"
0
7500
0
2010
95%
Levelised externality cost of energy
6000
2035
Time (Year)
2048
2060
-500 B
2010
2023
Figure 7.10: Confidence bounds for low and high damage costs estimates (range: Table 7.20) on selected
model outcomes
The effects of the simultaneous change in discount rate, load factor, private cost growth rates, damage cost
growth rates and lower and higher damage costs estimates, on selected model variables are presented in
Figure 7.10 in form of confidence bounds. For instance the 100% confidence bounds suggest that the total
private costs ranges between R200 billion - R1 trillion while the LCOE and NPV after tax range from R261
billion – R1 088 billion and between -R207 billion to -R907billion, respectively by the end of the simulation.
So though the confidence bounds show the same general patterns as in the univariate analyses, Figure 7.11
shows slightly wider bands of uncertainty on all simulated outputs than any of the univariate sensitivity
analysis. The combined uncertainty in all the uncertain parameters translates into a more uncertainty in
selected model results by the end of the simulation.
- 205 -
Sensitivity
50%
75%
95%
Sensitivity
50%
75%
100%
1.5 T
3750
1T
2500
500 B
1250
0
2010
Sensitivity
50%
75%
2023
2035
Time (Year)
95%
2048
0
2010
2060
Sensitivity
50%
75%
100%
Levelised externality cost of energy
6000
15,000
3000
10,000
1500
5000
Sensitivity
50%
75%
100%
2023
95%
2023
95%
2035
Time (Year)
2048
2060
100%
Levelised social cost of energy
20,000
4500
0
2010
95%
Levelised cost of energy
5000
Cumulative private costs
2T
2035
Time (Year)
2048
2060
2035
Time (Year)
2048
2060
0
2010
2023
2035
Time (Year)
2048
2060
100%
"Social NPV (after tax)"
0
-250 B
-500 B
-750 B
-1 T
2010
2023
Figure 7.11: Confidence bounds for of all uncertain parameters on selected model outcomes
(multivariate)
Now turning to the manual sensitivity testing of lower and higher range damage costs estimates which was
discussed earlier, the sensitivity analysis outcomes of this exercise are reported in Table 7.21 and later in
Table 7.22. The tables report the findings over the lifetime of Kusile. Table 7.21 shows the total coal-fuel
cycle externality cost over the lifetime of Kusile to range from a low value of about R1 450 billion to a high
value of R3 279 billion (base case at R2 173 billion). The base case estimate is therefore approximately 33%
higher and 34% lower than the lower and higher estimates, respectively.
- 206 -
The lower range damage cost scenario is expected to lower the social cost of power generation and
consequently improve the attractiveness of coal-based power by lowering the externality cost and hence
improving the social NPV of the project. The levelised externality cost of energy (LECOE) is estimated by the
model to range from a low value of R908/MWh to a high value of R2 052/MWh (baseline R1 371/MWh).
Comparing the LECOE to the four collective studies’ (i.e. Nkambule & Blignaut (2012), Riekert and Koch
(2012), Inglesi-Lotz and Blignaut (2012), and Blignaut (2012)) externality cost for Kusile conducted for the
year 2010, of between R310/MWh – R1 880/MWh, this study’s estimates are comparable but higher due to
the inclusion of more externalities and fuel-cycle phases.
The lower range scenario lowers the base case levelised social cost of energy from R1 925/MWh to R1
462/MWh while the higher estimate increases it to R2 606/MWh. The LCOE was earlier estimated as
R554.2/MWh, so with the use of the lower and higher range scenarios the LCOE was found to reflect about
38% or 21%, respectively of the true cost of coal, while the externality cost makes up the remainder. About
two-thirds to three-quarters of the true cost of electricity therefore do not reflect on the balance sheet of
the utility and are borne by society. With the use of the lower range scenario, the social NPV (after tax)
improves from its base case value of –383 billion to –R231 billion but is still negative and therefore coalbased power is still unattractive in social terms. The higher range scenario further worsens the social NPV
(after tax) of coal-based power to –R606 billion. So with the use of lower or higher damage cost estimates,
Kusile will still be unable to generate enough revenue to cover the negative externalities it imposes on third
parties.
Table 7.21: Lower and higher range lifetime externality costs versus baseline over the lifetime of Kusile
Externality
Total coal-fuel cycle externality cost
Externality cost as % of base case
Levelised externality cost of energy
Levelised social cost of energy
Social NPV (after tax)
Units
Lower
Base case
Higher
R billion
%
R/MWh
R/MWh
R billion
1 449.8
33%
908.0
1 462.2
-231.4
2 172.7
0%
3 279.0
34%
2 051.6
2 605.8
-606.4
1 370.8
1 925.0
-383.2
Regarding the externality costs per phase obtained over the life-time of Kusile shown in Table 7.22, the
overall life-time externality cost of mining and transporting coal to Kusile ranges between R377 and R974
billion (base case R594 billion). Water consumption makes up over 90% of the coal mining and
transportation externality cost, followed by global warming damage cost and then ecosystem service loss.
Based on Kusile’s lifetime coal consumption of about 0.870 billion tons as estimated by COALPSCA, the
externality cost linked with the mining and transportation of coal translates into between R434 – R1120/t
- 207 -
(base case R683/t). A value that is noticeably higher than that of the earlier South African studies discussed
in section 4.3.5, because of the inclusion of more externalities and a higher price of carbon. Based on,
Kusile’s lifetime electricity production of about 1.6 billion MWh, the externality cost of mining and
transporting coal to Kusile ranges between 24c/kWh - 61c/kWh (base case 37c/kWh). Based on the
electricity tariff that will prevail at the end of the simulation (i.e. 93c/kWh), the externality cost will be
between 26% - 66% of the electricity price (base case 40%).
Concerning the power plant, the externality cost considers the externalities from the direct combustion of
coal plus those from waste disposal. The FGD process will be reported on separately hereafter. The lifetime
externality cost from the operation of the power plant ranges between R784 and R1 715 billion. Water
consumption makes up about 40% of the externality cost, followed by air pollution human health cost and
global warming damages. Based on lifetime power generation, the power plant externality cost translates
to between 49c/kWh - 107c/kWh (base case 72c/kWh) which represents between 53% - 115% (base case
77%) of the electricity tariff that will prevail at the end of the simulation. The externality cost in c/kWh
(ZAR), when converted to US cents/kWh, ranges between 7c/kWh - 15c/kWh (base case 10/KWh) and falls
within a fair range with both the international and local studies reported on Tables 4.3 and 4.5,
respectively.
Table 7.22: Lower and higher range life-time externality costs per phase versus baseline
Coal mining
&
transportation
Externality
Construction
Plant operation
&
waste disposal
FGD system
operation
Low
Base
High
Low
Base
High
338.6
551.4
908.2
326.7
488.8
0.2
0.3
0.4
0.0
0.0
0.1
0.05
0.2
0.5
5.3
5.3
5.3
0.72
0.8
0.9
3.3
5.7
9.5
265.2
451.9
739.1
Low
Base
Not annual but over entire
construction phase
High
Low
Base
High
61.3
98.0
138.6
0.0
0.0
0.0
0.0
0.0
0.0
Billion Rands
Water use
Water
pollution
Fatalities &
morbidity
Ecosystem
loss
Classic air
pollutants
GHGs
Total
c/kWh
649.9
224.1
335.4
445.9
0.0
0.3
0.6
1.0
0.0
0.0
29.8
31.2
50.5
191.3
200.2
324.2
1.9
2
3.2
1.0
1.0
1.6
377.3
593.9
974.0
783.9
1141.9
1,714.7
226.4
338
450.1
62.3
99
140.2
24
37
61
49
72
107
14
21
28
4
6
9
The operation of the FGD system on the other hand, produces a lifetime externality cost that ranges
between R226 – R450 billion (base case R338 billion). Owing to data unavailability, only water- and airpollution related externalities were considered. Water consumption dominates the externality cost more so
- 208 -
than in the power plant (about 99%), followed by global warming damages and air pollution human health
costs. The FGD system externality cost ranges between 14c/kWh - 28c/kWh (base case 21c/kWh) which
represents between 15% - 30% (base case 23%) of the electricity tariff that will prevail at full capacity. The
FGD system and the power plant combined produce an externality cost that ranges between 63c/kWh –
135c/kWh (base case 93c/kWh). Fitting Kusile with an FGD system will therefore generate approximately a
third of the externality cost of the power station. The entire power station (i.e. coal combustion, FGD
operation and waste disposal) represents between 68% - 145% (base case 100%) of the electricity tariff at
full capacity.
Lastly, with regards to the plant construction phase, COALPSCA estimates the externality cost of
constructing Kusile to range between R62 - R140 billion (base case 99 billion). Water consumption
dominates the externality cost (over 98%), followed by global warming damage cost while the other
externality costs are very low. The construction phase externality cost translates to between 4c/kWh 9c/kWh (base case 6c/kWh). There are, however, no studies locally to compare these estimates with and
the two international studies (European Commission, 1995; 1999b) that did study the construction phase,
do not report explicitly the externality cost linked with this phase.
Summarizing the fuel-cycle externality costs reported above, the coal-fuel cycle externality cost ranges
between 91c/kWh – 205c/kWh with the base case at 136c/kWh (when converted to US cents/kWh it
ranges between 12c/kWh - 28c/kWh with the base case at 19c/kWh). The plant combustion phase with
waste disposal comprised most of the externality cost (49c/kWh - 107c/kWh), followed by coal mining and
transportation (24c/kWh - 61c/kWh), then the FGD system (14c/kWh - 28c/kWh), and lastly the
construction phase (4c/kWh - 9c/kWh). The externality cost generated by the model falls within the range
of the international studies reported on Table 4.3 and the local studies in in Table 4.5, but is slightly higher
than those that study the entire coal-fuel chain owing to the inclusion of more externalities and coal-fuel
cycle phases.
7.4
Policy analysis
In this section an evaluation is conducted of the COALPSCA Model outcomes under two potential policies
that could be faced by coal-based power utilities. The first policy is linked to Eskom’s concern that domestic
coal prices will soar to export levels, and is named the export parity coal pricing (EPP) policy. The second
policy concerns a form of internalizing the externality cost of coal-based power generation through carbon
taxation, and is named the carbon tax policy. Both policies are discussed further below. Sixteen scenarios
were formulated to evaluate the implementation effects of the two policies. Table 7.23 presents a
- 209 -
summary of these scenarios while the following sections provide a discussion of the policies, scenarios and
model outcomes.
Table 7.23: Policy scenarios
Scenarios
Name
Baseline
Export parity coal pricing
Carbon tax at 10%
Carbon tax at 5%
Carbon tax at 0.1%
7.4.1
Abbreviation
Export parity coal pricing
(R/ton)
Carbon tax
(R/ton of CO2e)
Base
EPPP600
EPPP700
EPPP800
CT100
CT120
CT150
CT200
CT100
CT120
CT150
CT200
CT100
CT120
CT150
CT200
Baseline – 210
600
700
800
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
Baseline
0
0
0
0
100 - growth rate 10%
120 - growth rate 10%
150 - growth rate 10%
200 - growth rate 10%
100 - growth rate 5%
120 - growth rate 5%
150 - growth rate 5%
200 - growth rate 5%
100 - growth rate 0.1%
120 - growth rate 0.1%
150 - growth rate 0.1%
200 - growth rate 0.1%
Export parity coal pricing scenarios
The sale of coal domestically at export parity prices is a looming danger facing the South African industry
and especially the energy sector as the main user of coal. The looming price threat can be explained by
various factors, including the emergence of viable export markets competing for Eskom’s low-grade coal
and under-investment in the coal mining industry which might result in coal shortage and therefore a rise in
coal price, i.e. coal shortage is forecasted to commence after 2018 (Creamer Media, 2013). It is therefore
imperative to explore the probable implications of various coal price regimes on the cost of generating
coal-based power.
In general, Eskom secures most of its coal through long-term (mostly cost-plus) contracts, as discussed in
section 2.3.4. Furthermore, personal communication with Joubert reaffirmed that Eskom sourced the bulk
of its coal through long-term coal supply agreements and that the amount of this coal is roughly less than
or equal to 80% of each power station’s annual coal usage and coal usage over its expected lifetime. This
percentage might, however, vary dependent on the power station’s expected position in the fleet ‘merit
order’ which is mainly centred on the variable cost of a power plant. In this regard, in the process of
adjusting the number of plants to service instantaneous electricity demand, a lower-fuel-cost plant will be
dispatched earlier than a high-fuel-cost plant. Since long-term coal supply agreements imply a very high risk
to the supplier, most of the agreements follow the cost-plus return on investment approach with some cost
- 210 -
efficiency incentives. The coal prices paid for through long-term supply agreements may or not be higher
than market prices for analogous coal grades (Eskom communication, 2013).
While coal volume flexibility is useful to all power stations, it needs to be less for low-fuel-cost plants.
Based on the discussion above, the short-term, more flexible contracts might provide 20% of the annual
coal requirement, to augment the long-term, annual-volume contracts (i.e. 80%) and might be subject to
export competition. For the purchaser, a balance needs to be struck between flexibility and cost, because
flexibility implies more risk for the supplier and often a higher price for the purchaser. Kusile, as a base-load
power station, will be contracted on long-term, cost-plus return on investment agreements and, depending
on national electricity demand growth being higher/lower than currently envisaged, it could imply
variances, necessitating some flexibility with the shorter-term, flexible contracts being exposed to export
opportunities (Eskom communication, 2013).
In the light of the above discussion, 25% of the annual coal usage in Kusile was exposed to export
competition as a worst-case scenario. The export parity coal pricing policy reflects three coal pricing
regimes, namely R600/ton (Creamer media, 2013), R700/ton and R800/ton, which were introduced from
year 2022 in the model. The export parity price outcomes are represented in Figure 7.12.
The export price regimes are expected to increase the fuel cost, total generation cost and therefore also
the LCOE. This is so because these variables are positively related to the LCOE. The Figure shows that
increasing the price of coal has an immediate and fairly noticeable effect on fuel costs, which will amplify
the cost of generating power and possibly power tariffs. The levelised fuel cost increases from the baseline
value of R117/MWh to R142/MWh, R149/MWh and R155/MWh in the case of the EPPP600, EPPP700 and
EPP800 scenarios, respectively. The figure continues to show the positive effect of the coal price regimes on
the LCOE which increases by +/-7%. Given that the EPPP600, EPPP700 and EPPP800 scenarios, show
domestic coal price increases of about 65%, 70% and 73%, respectively, whilst they effect +/-7% sensitivity
to the LCOE, the LCOE is thus not that response to the fuel prices, which could be mainly explained by that
only 25% of the coal requirements of Kusile were exposed to export competition. The LCOE arranges
between R580/MWh - 593/MWh for the EPPP scenarios
The NPV (after tax) simulation output in Figure 7.12 displays that the export price regimes diminish the
attractiveness of coal-based power as shown by the declining NPV as the unit cost of coal rises. For
example, the rise in fuel prices decreases the baseline NPV (after tax) from R66.4 billion to between R57.3 –
R60.4 billion. In summary, should coal prices rise to export parity price levels the total generation cost of
- 211 -
coal-based power would increase but not by a great margin, given the 25% exposure to export competition.
Under this exposure Kusile will still be a viable enterprise if one chooses to disregard the externality costs
associated with such an investment.
LCOE
6,000
150
4,500
R/MWh
R/MWh
Levelised fuel cost
200
100
50
3,000
1,500
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised fuel cost : EPPP800
Levelised fuel cost : EPPP700
Levelised fuel cost : EPPP600
Levelised fuel cost : Baseline
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised cost of energy : EPPP800
Levelised cost of energy : EPPP700
Levelised cost of energy : EPPP600
Levelised cost of energy : Baseline
NPV after tax
70 B
35 B
R
0
-35 B
-70 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
"NPV (after tax)" : EPPP800
"NPV (after tax)" : EPPP700
"NPV (after tax)" : EPPP600
"NPV (after tax)" : Baseline
Figure 7.12: Export parity price outcomes
7.4.2
Carbon tax scenarios
With discussions being underway on the imposition of carbon taxes to producers of greenhouse gases, it is
a looming threat facing the South African industry, and especially the energy sector as the main producer of
GHGs. For its energy needs, South Africa is largely dependent on coal - a source of energy known for its high
emissions of GHGs, especially CO2. As a form of mitigating the risk of climate change, National Treasury has
opted to impose a carbon tax on emitters of GHGs. Companies would not pay for the entire emissions they
cause, but a tax-free exemption threshold and offset to a maximum of 90% is proposed, to minimize
negative impact on local firms’ competitiveness and also to lighten the burden of higher energy prices on
households (National Treasury, 2013).
- 212 -
The carbon tax proposal was revised recently, with the new proposal raising the threshold beyond which
the tax is payable, and suggesting subsidies to invest in low-carbon technologies, among other issues. The
proposed tax by National Treasury is R120/ton of CO2 equivalence beyond the tax-free threshold (National
Treasury, 2013) which is normally 60% for all companies (BDFM Publishers, 2013d). Eskom will pay for only
40% of its CO2e emissions, that is, a tax-free exemption threshold of 60% (National Treasury, 2013),
resulting in an effective tax rate of R48/ton of CO2e. Eskom could invest in offset investments of its own,
and such investments could be subtracted from its tax liability to a maximum of 10%. This results in a
minimum tax liability of R36/ton of CO2e. Also, in acknowledgment of the complexity of quantifying,
reporting on, and policing the emissions of a gas that is odourless and colourless, the tax is going to be
imposed as a fossil-fuel input tax, based on carbon contents of fossil-fuels like coal, natural gas and crude
oil (Urban Earth, 2012; Urban Africa, 2013). For the purposes of this study, the tax would therefore have
been a fuel input tax based on coal consumption, but since GHG emissions were estimated in the current
study, the tax was imposed on CO2e emissions.
In contrast to the proposed tax by National Treasury, Robbie Louw, the director of Promethium, and
Harmke Immink concur that, based on South Africa’s GHG emissions, R100/ton is more sensible
(Esterhuizen, 2013). The tax proposed by Treasury is also not well received by business enterprises, who
foresee the tax rate raising the cost of doing business in the country. Treasury plans effecting the proposed
carbon tax beginning 2015 (BDFM Publishers, 2013d) and increasing it annually by 10% within the first
phase, which is from 2015 to 2019 (Esterhuizen, 2013; National Treasury, 2013). Following this phase, the
tax will be revised to a new tax rate and to lower tax-free thresholds which will be effective from 2020
(National Treasury, 2013). It is therefore important to explore the likely implications of various carbon tax
regimes on the cost of generating power from coal. Four carbon tax regimes, R100/ton of CO2e, R120/ton
of CO2e, R150/ton of CO2e, and R200/ton of CO2e, were introduced in the COALPSCA Model. The carbon tax
scenarios at 10%, 5% and 0.1% shown in Table 7.23, escalate these four tax regimes at an annual rate of
10%, 5% and 0.1% over the entire lifetime of Kusile, respectively. In addition, in all the scenarios, Kusile was
assumed to be charged for only 30% of its combustion phase emissions.
The carbon tax regimes are expected to increase the cost of generating power in a coal-based plant, by
imposing a new operational cost, namely carbon tax cost. The carbon tax also generates revenue for the
government that can be put to a number of uses, such as support of cleaner sources of electricity. The
model unveiled that the proposed tax rate of R120/ton of CO2e, generates revenue over the lifetime of
Kusile that ranges between R47 billion at 0.1% growth rate, R1 283 billion at 10% growth rate (R216 billion
at 5% growth rate). Given that the government plans to lower tax-free thresholds beginning 2020 (National
- 213 -
Treasury, 2013), and perhaps revise the tax rate upwards, these revenue streams as estimated by the
model, are lower-end estimates for the growth rates in question. Overall for all carbon tax regimes (low to
high) at growth rates of 10%, 5% and 0.1%, the model estimates revenues ranging between R1 069 – R2
138 billion, R180 – R361 billion, R39 – R79 billion, respectively.
The power plant externality costs of the entire GHG emissions (100%) at the low, median and high damage
cost of carbon listed in Table 7.20 (i.e. based on Blignaut, 2012) were computed (growth rate 0.1%) and
based on the resultant GHG emissions externality costs, carbon tax scenarios (listed in Table 7.23) were
explored that recouped the various costs. The 100% GHG emissions externality costs at the unit damage
cost of CO2 of R104.98/ton, R109.89/ton and R177.94/ton amounts to R142 billion, R148 billion, R240
billion, respectively. Previously, it was stated that all four carbon tax regimes at growth rates of 10%, 5%
and 0.1%, yielded government revenues ranging between R1069 – R2138 billion, R180 – R361 billion, R39 –
R79 billion, respectively, so evidently none of the four carbon tax regimes at the growth rate of 0.1%
recovers any of the 100% GHG emissions externality costs while the carbon tax regimes at the growth rate
of 10% will overly recoup the 100% GHG emissions externality costs. The revenues generated by CT100,
CT120, CT150 and CT200 at 5% growth rate are R180 billion, R216 billion, R270 billion and R361 billion,
respectively. So the CT100 scenario at 5% growth rate more than recoups the low and median 100% GHG
emissions externality costs computed using the unit damage cost of CO2 of R104.98/ton and R109.89/ton,
respectively while the CT150 scenario at 5% growth rate more than recoups the high 100% GHG emissions
externality cost computed using the unit damage cost of CO2 of R177.94/ton. These findings therefore
suggest that regardless of the carbon tax regimes the growth rate of the carbon tax regimes need to be
carefully selected since it greatly alters the resultant payable GHG externality cost or government revenues,
which negatively affect the financial viability of coal-based power plants. Based on these findings growing
any of the carbon tax regimes at 10% nullify the fact that coal-power plants pay for only 30% of their GHG
emissions, this is so because such a higher growth rate tend to more than recoup the 100% GHG emissions
externality costs (depending on the unit damage cost of CO2 effected). Regarding the unit carbon tax
regimes and carbon tax growth rates combined, the findings suggest that the carbon tax should preferably
be lower than R150/ton of CO2 and be grown at a growth rate lower than 5%, or else coal-based power
plants pay way above 100% of their GHG emissions.
The carbon tax regimes in Table 7.23 impose a new operational cost (i.e. carbon tax cost) which is expected
to increase the cost of generating power in a coal-based plant, which should amplify the LCOE. The
simulation output of the new operational cost (i.e. carbon tax cost) at various tax regimes and growth rates
is shown in Figure 7.13. The figure depicts the increasing effect of the levelised carbon tax per tax regime,
- 214 -
from the low growth rate of 0.1% to the higher growth rate of 10%. Increasing the tax regimes at 10% is
shown to have the greatest impact on the generation cost of coal-based power, with levelised carbon tax
estimates ranging between R179/MWh – R358/MWh.
Levelised carbon tax cost
Levelised carbon tax cost
Levelised carbon tax cost
50
200
400
37.5
150
300
25
100
200
12.5
50
100
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised carbon tax cost : CT100 at 1
Levelised carbon tax cost : CT120 at 1
Levelised carbon tax cost : CT150 at 1
Levelised carbon tax cost : CT200 at 1
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised carbon tax cost : CT100 at 5
Levelised carbon tax cost : CT120 at 5
Levelised carbon tax cost : CT150 at 5
Levelised carbon tax cost : CT200 at 5
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised carbon tax cost : CT100 at 10
Levelised carbon tax cost : CT120 at 10
Levelised carbon tax cost : CT150 at 10
Levelised carbon tax cost : CT200 at 10
Figure 7.13: Carbon tax cost at various tax regimes and growth rates
Figure 7.14 shows the LCOE for all the carbon tax scenarios, the carbon tax inclusive LCOE estimates range
between R575/MWh – R913/MWh, effecting sensitivity on the LCOE of about +/-39%, +/-16% and +/-7%
when grown at 10%, 5% and 0.1% growth rates, respectively. With LCOE for renewable technologies such
as wind plants and solar based-power, estimated at about R658/MWh – R1052/MWh depending on wind
classes and R1517/MWh – R2026/MWh depending on technology type and storage hours, respectively,
wind energy will quickly become cost-competitive with coal based power, if any of the four tax regimes are
escalated at either 5% or 10%.
- 215 -
LCOE at 10%
LCOE at 5%
5000
5000
3750
3750
3750
2500
1250
2500
2500
1250
1250
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised cost of energy : CT100 at 1
Levelised cost of energy : CT120 at 1
Levelised cost of energy : CT150 at 1
Levelised cost of energy : CT200 at 1
Levelised cost of energy : Baseline
R/MWh
5000
R/MWh
R/MWh
LCOE at 0.1%
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
0
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
Levelised cost of energy : CT100 at 10
Levelised cost of energy : CT120 at 10
Levelised cost of energy : CT150 at 10
Levelised cost of energy : CT200 at 10
Levelised cost of energy : Baseline
Levelised cost of energy : CT100 at 5
Levelised cost of energy : CT120 at 5
Levelised cost of energy : CT150 at 5
Levelised cost of energy : CT200 at 5
Levelised cost of energy : Baseline
Figure 7.14: Carbon tax effects on LCOE
Figure 7.15 depicts the NPV (after tax) simulation output at various tax regimes and growth rates. The
figure shows the weakening financial viability of coal-based power as shown by the declining NPV as the
carbon tax rates and growth rates are increased. The tax regimes at the 10% growth rate show the largest
change of the baseline value (R66.4 billion) which decreases to between R24.1 billion (at CT100) and -R18.3
billion (at CT200). The CT200 tax regime at 10% growth rate is the only scenario that yields a negative NPV,
so for the most part, if coal-based power utilities should be charged the studied regimes at 5% or 0.1% and
only for about one third of the GHGs they emitted, coal-based power production would still be a viable
enterprise, with NPV ranging between R42 – R54 billion or R56 – R61 billion, respectively. The financial
viability of coal-based power would, however, be profoundly worsened by any of the studied tax regimes
when perpetually grown at 10% over the lifetime of the plant.
- 216 -
NPV at 5% growth rate
NPV at 0.1% growth rate
NPV at 10% growth rate
35 B
35 B
35 B
0
0
0
R
70 B
R
70 B
R
70 B
-35 B
-35 B
-35 B
-70 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
-70 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
-70 B
2010 2015 2020 2025 2030 2035 2040 2045 2050 2055 2060
Time (Year)
"NPV (after tax)" : CT100 at 1
"NPV (after tax)" : CT120 at 1
"NPV (after tax)" : CT150 at 1
"NPV (after tax)" : CT200 at 1
"NPV (after tax)" : Baseline
"NPV (after tax)" : CT100 at 5
"NPV (after tax)" : CT120 at 5
"NPV (after tax)" : CT150 at 5
"NPV (after tax)" : CT200 at 5
"NPV (after tax)" : Baseline
"NPV (after tax)" : CT100 at 10
"NPV (after tax)" : CT120 at 10
"NPV (after tax)" : CT150 at 10
"NPV (after tax)" : CT200 at 10
"NPV (after tax)" : Baseline
Figure 7.15: NPV at various tax regimes and growth rates
7.5
Summary
In this chapter the COALPSCA Model baseline outcomes were analyzed based on sixteen economic and
environmental/societal indicators. Ten of these indicators were economic indicators representing the
performance of the plant and the cost incurred by plant developers and the community at large. The
remaining six indicators were environmental indicators reflecting the six main categories of externalities
quantified and monetized in this study. This analysis was followed by model validation tests and lastly,
policy analysis. The key findings of the study are summarized below.
Baseline results: Based on model settings, the main factors influencing the behaviour of electricity
generation are – (i) investment in plant capacity; (ii) load factor; (iii) plant operating hours, and (iii) profits.
The model showed that the behaviour of the resource inputs into power generation (e.g. coal
consumption) and plant construction (e.g. steel) follows the same dynamics as that of power generation
and construction schedule, respectively. The environmental indicators analysis also established that most
of the indicators studied in the coal-fuel chain, (e.g. air emissions (classic air pollutants, trace metals and
GHGs), water use, fatalities, morbidity and waste production), with the exception of the construction
phase, mainly follow similar dynamics as that of power generation, while those from the construction
phase mainly follow the construction schedule behaviour.
Concerning the costs of generating power (private costs), the model showed that the main determinants of
generation cost and the LCOE are fuel and capital costs. These costs are therefore the main factors
determining the viability of coal-based power. The model estimates a base case lifetime generation cost of
Kusile of about 411 billion Rands, with the LCOE at about R554/MWh. NPV analysis was also performed and
- 217 -
it was positive, indicating that investing in Kusile is economical in private terms, but after attaching
economic values to the studied environmental indicators (externalities), the model estimated a negative
social NPV throughout the lifetime of Kusile.
The model estimated the base case coal-fuel cycle externality cost over the lifetime of Kusile to be about 2
173 billion Rands. The base case externality cost per kWh sent out is therefore about 136c/kWh specifically 72c/kWh is attributable to the power plant and waste disposal phases, 37c/kWh stems from the
coal mining and transportation phase, while the FGD system and plant construction contribute 21c/kWh
and 6c/kWh, respectively. Most of the externality cost stems from three types of externalities, namely
water use (68%), air pollution health cost (21%) and global warming damage cost (11%), and from three
coal-fuel cycle phases, namely plant operation (51%), coal mining and transportation (27%) and FGD system
operation (16%). Collectively the three phases make up about 94% of the lifetime externality cost.
Finally, since the installation of the FGD system increases water use while curbing SO2 emissions, this
interesting trade-off between water use and SO2 emissions was explored by studying the air pollution
health cost and opportunity cost of water use with and without the installation of the FGD system. The
installation of the FGD system was found to be a sensible effort (on the grounds of externality cost versus
externality cost savings) since its air pollution health cost savings outweigh the water use externality cost it
introduces. Water is, however, a scarce resource in South Africa and human health is without doubt
valuable, so the country and its people need to decide what it is willing to forego in order to gain the other.
Give-up water in exchange for clean air and hence gain better human health or vice versa. On the other
hand, in order for one to reach a final conclusion about the economic viability of the FGD system, the
private and externality costs associated with the FGD system need to be fully paid off by the savings.
Model validation: Five structural validity tests were performed in this study, namely structure verification,
dimensional consistency, boundary adequacy, extreme condition and parameter verification tests. In the
light of these tests the COALPSCA Model was found to be a reasonable simplified match of the real-world
system. Behaviour validity tests were also conducted. Like with any long-term analysis there is uncertainty
about the costs and the technical factors in a coal-fired power plant. To learn how uncertainty in parameter
estimates translates into uncertainty in simulated model outputs, sensitivity analysis was conducted first by
using univariate sensitivity analysis followed by multivariate sensitivity analysis. Sensitivity analysis in this
study focused on the load factor, discount rate, cost growth rates of all private costs in the model (e.g. coal,
limestone, water, O&M and capital costs), cost growth rates of all damage cost estimates and the
sensitivity of the model outcomes to lower and higher range estimates. Uncertainties concerning such
- 218 -
variables are a reality for energy markets so all these parameters are important input variables that can
affect among other factors the production of coal-based power, total generating cost of power, LCOE,
LECOE and the financial viability of coal based power.
Given the individual uncertainties in discount rate, load factor and private cost growth rates, univariate
sensitivity analysis highlight the important drivers of the generation cost of coal-based power to be the
discount rate, cost growth rates and the load factor in descending order. While given the individual
uncertainties in damage cost growth rates, and lower and higher damage costs estimates, univariate
sensitivity analysis highlight narrow bands of uncertainty to variations in damage cost growth rates than
those to variations in low and high damage costs estimates. The effects of the simultaneous change in
discount rate, load factor, private cost growth rates, damage cost growth rates and lower and higher
damage costs estimates disclosed that though the confidence bounds of the multivariate analysis show the
same general patterns as in the univariate analyses, the multivariate analysis outcomes show slightly wider
bands of uncertainty on all simulated outputs than any of the univariate sensitivity analysis. The combined
uncertainty in all the uncertain parameters translates into a more uncertainty in selected model results by
the end of the simulation.
Policy analysis: The COALPSCA Model outcomes were evaluated under two potential policy scenarios that
could be faced by coal-based power utilities, namely carbon taxation and the pricing of domestic coal at
export parity price levels. Fifteen scenarios characterized by the two policies at various price regimes and
growth rates, were defined and evaluated with reference to the baseline scenario. Due to the amount of
coal exposed to export competition (25%), the total generation cost and consequently the LCOE were found
to be fairly responsive to the export coal price regimes (+/-7%). Under this exposure, coal-based power
production would still be a viable enterprise if one chose to disregard the externality costs associated with
such an investment.
Conversely, the total generation cost of coal-based power was found to be moderately to severely
impacted by the carbon tax regimes, depending on the rate at which they were grown. The four tax regimes
were found to effect +/-39%, +/-16% and +/-7% sensitivity on the LCOE when grown at 10%, 5% and 0.1%
growth rates, respectively, with carbon tax inclusive LCOE estimates ranging between R575/MWh R913/MWh. Coal-based power production would still be a viable enterprise if power utilities were charged
the studied tax regimes at 5% or 0.1%, and only for one third of the GHGs they emitted. The financial
viability of coal-based power would, however, be profoundly worsened by any of the studied tax regimes
when perpetually grown at 10% over the lifetime of the plant.
- 219 -
On another note, the power plant externality costs of the entire GHG emissions (100%) were also
computed using the unit damage costs of CO2 listed in Table 7.20 (i.e. based on Blignaut, 2012) grown at
0.1% and based on the resultant GHG emissions externality costs, carbon tax scenarios (listed in Table 7.23)
were explored that recouped the various costs. The findings: (i) suggested that regardless of the carbon tax
regimes the growth rate of the carbon tax regimes need to be carefully selected since it greatly alters the
resultant payable GHG externality cost or government revenues, which negatively affect the financial
viability of coal-based power plants; (ii) disclosed that growing any of the carbon tax regimes at 10% nullify
the fact that coal-power plants pay for only 30% of their GHG emissions; and (iii) further suggested that the
carbon tax should at most be preferably lower than R150/ton of CO2 and be grown at a growth rate lower
than 5%, or else coal-based power plants pay way above 100% of their GHG emissions.
Lastly, not all basic aspects of reality are considered by the model, partly owing to the lack of data and
anticipated model complication of including certain parameters (e.g. ecosystem services lost upstream of
the power plant, excluding those linked with the coal mine). Regardless of the limitations, the model does
indeed provide a reasonable, simplified demonstration of the use of system dynamics for the assessment of
electricity generation, fuel-cycle externalities and social cost associated with a coal-based power plant.
- 220 -
CHAPTER 8:
8.1
CONCLUSION, LIMITATIONS AND RECOMMENDATIONS
The research conducted in this study and main findings
The primary concern of this study was to understand coal-based power generation and its interactions with
resource inputs, private costs, externalities, externality costs and hence its consequent economic, social
and environmental impacts over its lifetime and fuel cycle, through the application of a system dynamics
approach along a life-cycle viewpoint. A model that assesses power generation and the social cost of
generating power in the coal-fuel chain named COALPSCA was therefore developed. The purpose of
developing the model was twofold - firstly, to aid energy decision makers with a tool for making informed
energy supply decisions that consider the financial viability of power generation technologies, but also the
socio-environmental consequence of the technologies. Secondly, to aid coal-based power developers with
a useful tool for detecting the main drivers of the burdens and costs in the system which should yield vital
socio-economic-environmental tradeoff information that can be beneficial to them.
Early in this thesis a historical review was conducted of the schools of economic thought and system
dynamics, with the ultimate aims of determining the schools of economic thought that underpins this study
and its links with system dynamics. The review of economic thought disclosed that the main concepts in
this study, namely production, externalities and social cost are rooted in neoclassical and environmental
economics, particularly, in welfare economic theory, theory of production and Pareto efficiency.
Neoclassical and environmental economics were therefore the main economic disciplines that provided the
theoretical base for this study. The ontology, epistemology and methodology of both neoclassical and
environmental economics were discussed to be realist, objective and quantitative, respectively, and hence
to fall within the positivist research paradigm of Guba and Lincoln’s classification. The proposed modelling
approach (system dynamics) was found to share many elements that are consistent with the two economic
disciplines that underpin this study, for instance, ontology and epistemology elements and the use of
quantitative techniques, but in addition, to offer more features such as a complex unitary approach with
the ability to deal with large number of elements and many interactions between elements, experimental
approach and empirical solutions, case study approach instead of using abstractions to develop models,
problem-orientated approach, transdisciplinarity methods, confidence based on model structure over
coefficient accuracy, emphasis on understanding system’s structure and our assumptions about it as
opposed to focusing on predictions, non-linear structures, dynamic structures, disequilibrium approach and
focus on closed loop information feedback structures.
- 221 -
A review of power generation assessment tools and their application was conducted. The review disclosed
that an assortment of methods and tools have been adopted by researchers to evaluate power generation
technologies contingent on the specific aims and scopes of the applications. The tools were grouped into
three broad categories of methods, namely financial, impact, and systems analysis methods. The review of
financial measures disclosed that different financial measures are suited for different computations, but
generally cost-effective energy projects are those with lowest LCOE, LCC, simple payback period and
discounted payback period plus those with high IRR, MIRR and NPV. The review of impact analysis tools
discloses that various tools are suited for identifying, quantifying and monetizing externalities. Depending
on the aims of the investigations and a number of issues surrounding the research (such as time and
financial constraints and availability/unavailability of previous primary valuation studies) various
researchers employed various impact analysis tools. On the other hand, the review of the systems analysis
tools unveiled that various systems models are designed for different purposes (e.g. modelling energy
system, and/or economic system, and/or ecological system), different technologies (e.g. renewable energy,
non-renewable energy or both), different scales of analyses (e.g. national, regional or global) and different
sizes of energy systems.
Concerning the application of the tools in the power sector, the review disclosed that in the past three
decades many studies have been undertaken on electric sector costs in both developed and developing
nations. Earlier externality studies used the abatement cost and bottom-up approaches to derive
externality costs estimates while recent studies used the bottom-up approach and benefit transfer
technique to estimate externality cost of power generation. The studies differ in terms of the types of
externalities they consider, the fuel-cycle stage(s) they investigate and they do not factor in the longstanding repercussions of the technologies on the environment and social systems. The most investigated
externalities internationally and locally are climate change and human health impacts associated with
airborne pollution from coal combustion. Both locally and internationally more attention is still paid to the
power generation phase. These differences in scope affected the outcomes of the studies, made comparing
them difficult, and highlighted the need for comprehensive externality investigations that widen the range
of externalities studied, that consider the various fuel-cycle stages and that embrace the long-term
repercussions of the technologies on the environmental and social systems.
The literature on systems analysis models disclosed that various models have been developed for
addressing various energy-related issues (for instance, locally the models are mainly used for modelling
energy supply and demand, projecting GHG emissions and studying climate change mitigation options),
however, as evidenced by the review not all facets/features of power generation had been studied by the
- 222 -
researchers, for example, the models were not tailored to specific coal-based power generation
technologies, did not address social cost, nor permit deeper understanding of coal-based power generation
and its interactions with resource inputs, private costs, externalities, externality costs and hence its
consequent economic, social and environmental impacts over its lifetime and life-cycle. Also environmental
focus was evidently on quantifying direct GHG emissions from the coal combustion phase, thus numerous
combustion phase and upstream burdens could still be incorporated and monetized to advance coal energy
analysis. Finally, no study was found that used a system dynamics approach to assess power generation and
the social cost of coal-based power generation over its lifetime and fuel cycle.
The modelling steps suggested by Roberts et al. (1983), Ford (1999), and Sterman (2000), namely problem
formulation, dynamic hypothesis, model formulation (structure and equations), model validation and policy
design and evaluation, were followed in developing and validating the COALPSCA Model. The model was
used for:

Understanding coal-based power generation and its interactions with resource inputs, private
costs, externalities, externality costs and hence its consequent economic, social and environmental
impacts over its lifetime and life-cycle;

Aiding energy decision makers with a visual tool for making informed energy supply decisions that
consider the financial viability and the socio-environmental consequences of power generation
technologies;

Aiding coal-based power developers with a useful tool with a clear interface and graphical outputs
for detecting the main drivers of costs and sources of socio-environmental burdens in the system
which should yield vital socio-economic-environmental tradeoff information;

Understanding the impacts of various policy scenarios on the viability of coal-based power
generation; and,

Validating the model since no historical data existed on Kusile power station.
The main findings of this research are as follows:

Pertaining to the private costs – the baseline scenario disclosed that investing in Kusile was
economical and that fuel and capital costs were the main cost components determining the
financial viability of coal-based power. The baseline model estimated the lifetime generation cost
of power in Kusile at about 411 billion Rands, while the LCOE was estimated at about R554/kWh.

The externalities inventory analysis unveiled the plant operation phase as the highest water using
phase in the coal-fuel chain, with the combustion phase and FGD system using about 31% and 22%,
- 223 -
respectively of the coal-fuel cycle water requirements. Water use in the coal mining phase was also
found to be high (37%), making the coal-fuel cycle a large yet hidden water user. Another
important outcome from the inventory output is that the coal mining phase was found to be more
prone to injuries than deaths whereas the plant operation phase was found to be more prone to
deaths than injuries. Human safety is therefore a serious problem in these two phases. Concerning
air pollution loads, CO2e emissions were estimated at about 1 583 million tons over the coal-fuel
cycle and lifetime of Kusile, with low SO2 emissions due to the installation of the FGD system. Over
85% of the air pollutants emanated from the combustion phase.

Concerning the externalities - the model estimated the total coal-fuel cycle externality cost over the
lifetime of Kusile to range between R1 450 billion – R3 379 billion (baseline R2 173 billion) or
between 91c/kWh – 205c/kWh sent out (baseline 136c/kWh). Specifically, 49c/kWh - 107c/kWh
(baseline 72c/kWh) is attributable to the power plant and waste disposal phases, 24c/kWh 61c/kWh (baseline 37c/kWh) is linked with coal mining and transportation while the FGD system
and plant construction contribute 14c/kWh - 28c/kWh (baseline 21c/kWh) and 4c/kWh - 9c/kWh
(baseline 6c/kWh), respectively. Most of the externality cost stems from three types of
externalities, namely water use, air pollution health cost and global warming damages, in
descending order, and from three coal-fuel cycle phases, namely plant operation, coal mining and
transportation and FGD system, in descending order. The externality cost generated by the model
when converted to US cents/kWh it ranges between 12c/kWh - 28c/kWh with the base case at
19c/kWh, so it falls within the range of the international studies reported on Table 4.3 and the local
studies in in Table 4.5, but is slightly higher than those that study the entire coal-fuel chain owing
to the inclusion of more externalities and coal-fuel cycle phases.

The social cost analysis unveiled that about two-thirds to three-quarters of the true cost of coalbased electricity is not reflected in the balance sheet of the utility but is borne by society.
Accounting for the life-cycle burdens of coal-derived electricity thus conservatively doubles to
triples the price of electricity, making renewable energy sources like wind and solar attractive.

With regard to policy evaluation (i.e. carbon tax policy and the pricing of domestic coal at export
parity price levels) – owing to the amount of coal exposed to export competition (25%), the total
generation cost and consequently the LCOE was found to be fairly responsive to the export coal
price regimes (+/-7%). Coal-based power production was still found to be a viable enterprise under
the export parity price regimes. Conversely, the total generation cost of coal-based power was
found to be moderately to severely impacted by the carbon tax regimes (+/-39%), depending on
the rate at which they were grown (10%, 5% or 0.1%). Enforcing any of the studied carbon tax
regimes at 5% or 0.1% to only about one third of the GHGs emissions, would still make coal-based
- 224 -
power generation a viable enterprise, while tax escalation at 10% would profoundly worsen the
financial viability of coal-based power, quickly making renewable energy (especially wind energy)
cost-competitive with coal-based power. In the event that both policies are faced simultaneously
by power utilities, coal-based power will become even more costly further encouraging market
penetration of cleaner sources of energy. Carbon taxation as a policy instrument to mitigate
climate change will therefore bring great market penetration of clean technologies in the near
future if carefully planned and implemented.
On another note, the power plant externality costs of the entire GHG emissions (100%) were also
computed using the unit damage costs of CO2 listed in Table 7.20 (i.e. based on Blignaut, 2012)
grown at 0.1% and based on the resultant GHG emissions externality costs, carbon tax scenarios
(listed in Table 7.23) were explored that recouped the various costs. The findings: (i) suggested that
regardless of the carbon tax regimes the growth rate of the carbon tax regimes need to be carefully
selected since it greatly alters the resultant payable GHG externality cost or government revenues,
which negatively affect the financial viability of coal-based power plants; (ii) disclosed that growing
any of the carbon tax regimes at 10% nullify the fact that coal-power plants pay for only 30% of
their GHG emissions; and (iii) further suggested that the carbon tax should at most be preferably
lower than R150/ton of CO2 and be grown at a growth rate lower than 5%, or else coal-based
power plants pay way above 100% of their GHG emissions.
8.2
COALPSCA Model limitations
The COALPSCA Model, while it attempted to incorporate most of the important aspects of power
generation in a coal-fired power plant and its links with economic, social and environmental issues, it does
not capture all the intrinsic aspects. The limitations of the model include:

The exclusions of important burdens due to lack of data, such as fatalities and injuries linked with
plant construction, water pollution linked with the power plant and FGD system, noise pollution
and damages to roads;

The exclusion of some burdens due to the anticipated and unnecessary complications they could
pose, such as ecosystem services lost upstream of the power plant though not including those
linked with the coal mine (e.g. ecosystem services lost due to resource requirements for building
and operating the plant);
- 225 -

The exclusion of the influence of electricity demand on plant investment (i.e. investment in plant
capacity was exogenously modelled), due to the limited scope of the model, i.e. model focused on
a single plant;

Inflating damage cost values to the base year of this study (2010) using the consumer price index,
which likely underestimated the costs, for instance, the inflation of medical costs likely exceeds
that of the normal basket; and lastly,

While a concerted effort was made to solicit and use South African based data in computing most
of the externality costs, for placing value on air pollution related human health effects, used were
damage cost estimates from international studies which were adjusted for income differences,
inflation and currency exchange to the South African context. Moreover, there are limitations of
adjusting and transferring externality costs from secondary data, such as carrying forward errors
from previous studies (i.e. judgment and potential bias). Conducting primary research on the
externalities studied in this study was mostly, however, impossible due to that the power plant and
the coal mine are under construction. This was mitigated by using mainly published literature and
focusing on a range of externality cost estimates, instead of point estimates.
8.3
What could be done to improve the COALPSCA Model and energy research
In spite of these limitations, the model does indeed provide a reasonable, simplified demonstration of the
employment of a system dynamics approach to the assessment of electricity generation, resource/material
inputs, externalities and the social cost associated with a coal-based power plant over its lifetime and fuel
cycle, in a transparent manner. In addition, it provides - (i) coal-based power developers with a useful tool
for detecting the main drivers of burdens and costs in the system, which should yield vital socio-economicecological tradeoff information; and (ii) energy decision makers with a tool for making informed energy
supply decisions that consider not only the financial feasibility of power-generation technologies, but also
the socio-environmental consequences of the technologies.
What could have been more interesting to inform the energy supply debate, would have been to conduct
similar social cost assessments for renewable energy technologies, for instance, wind and solar, as initially
planned in the conception days of this research and comparing them to the outcomes of this study.
However, time and financial restraints did not allow for such to be conducted in this research. The
development of similar social cost assessment models for alternative energy sources, both renewables and
non-renewables, is therefore recommended, because all power generation alternatives are associated with
varying socio-economic-environmental effects and private costs. In this regard, social cost assessment may
help with capacity expansion decisions because of its ability to evaluate the trade-offs of electricity
- 226 -
generation alternatives. It will moreover limit the politicization of capacity expansion plans, by encouraging
energy planners to be transparent about their assumptions, which will likely stimulate public debate among
various stakeholder groups and possibly shape future capacity expansion decisions based on a broader
consensus.
Future research on externalities of energy technologies can also be improved by practically verifying the
growth rate of the damage cost of various externality burdens through conducting surveys which solicit the
various damage costs over some period of time. Another issue that can be explored in future works, is
conducting primary research in the South African context of air pollution impacts on human health. Such an
in-depth research could be a lengthy process, owing to the necessity of understanding the dispersion of
pollutants, their ultimate deposition and the responsiveness of humans to various doses of pollution. In
addition, placing a value on human-health-impairment/loss-of-a-human-life might prove difficult as
respondents could lack the knowledge of fully understanding what is being valued.
The COALPSCA Model was built for a single Eskom power plant, namely Kusile power station, and it
therefore excluded all of Eskom’s coal-based power plants or coal-fired power plants in South Africa. There
is thus the possibility of customizing the COALPSCA Model for the entire country through considering the
specificity of other coal plants, thereby enabling the assessment of the social cost of coal-based power
production in South Africa. Customizing the model to incorporate all of the country’s power plants will also
– (i) enable the exploration of the private and externality costs of retrofitting Eskom’s/the-country’s coalfired power plants with FGD systems; and (ii) how doing so will affect coal-based power’s attractiveness
and, most importantly, (iii) implications of the retrofits on water consumption. In addition, customizing the
model will also facilitate an evaluation of the factual country-level implications of coal-based power
utilities’ exposure to carbon taxation and export parity coal price levels.
8.4
Way forward for the South African government
The harshest way forward for the South African government in its address of the serious impacts of coalbased electricity generation, would be to reform the pricing system for coal-based electricity in the country
such that all the externality costs are properly reflected in the price. Doing so, will however, result in
serious socio-economic consequences, so the government needs to be strategic in its externality
internalization approach in order to minimize the socio-economic impacts, say by focusing its attention on
the main burdens in the coal-fuel chain such as water use, air pollution human health effects and climate
change effects due to GHG emissions, as revealed by the COAPSCA model.
- 227 -
Since the analysis conducted in this work unveils that most of the externality cost in the coal-fuel chain is
linked with water use (over 65%), lowering the water consumption of existing and new power plants and
coal mines, necessitates policy changes at both the local and national levels, which will force these
dominant water users to reduce their water consumption to minimal levels, using existing and affordable
technology. Existing power plants in the country, for instance, mostly uses wet-recirculating cooling system,
so existing plants can be required to upgrade to dry cooling systems over some sensible time period. By
doing so, the price of power will rise to mirror the costs of retrofits, which will make less water demanding
power generation sources more competitive. Alternatively, the South African government may directly
internalize the externality cost of water use for all water users in the country through pricing water well.
This will necessitate estimating the opportunity cost of water use for a number of industries/water-users,
which in the most part has been done for a number of dominant water users, so what the government
need to do is to take these studies serious, and do what is necessary to channel water use to efficient uses.
Concerning air pollution human health effects, which is the second largest externality in the coal-fuel chain
(over 21%) and which is mainly associated with the power plant – the government can take action by
requiring retrofits of all existing plants with FGD device over some reasonable period of time, and as well as
requiring new plants to be fitted with this device. The ideal device should preferably be dry FGD in order to
minimize water consumption. Such retrofits will not only safeguard human health but will assist in
reinstating the balance between clean and dirty power generation sources, and will encourage eventual
transformation of the existing fleet of dirty power stations to extensively more sustainable power
technologies.
With regards to the third largest externality cost in the coal fuel chain, namely climate change effects due
to GHG emissions (over 11%) – the South African government has taken action and intends to internalize
the externality cost of carbon emissions on producers of GHGs beginning 2015, through a carbon tax of
R120/ton of CO2e emissions. As explained in the policy evaluation section, companies will not pay for the
entire GHGs they emit, so the effectiveness of this tax in internalizing the externality cost of carbon
emissions will depend on how low the tax-free thresholds are and the rate at which the tax will be grown.
The analysis conducted in this study disclosed that enforcing any of the studied carbon tax regimes at 5% or
0.1% to only about one third of the GHGs emissions, would still make coal-based power generation a viable
enterprise, while tax escalation at 10% would profoundly worsen the financial viability of coal-based power,
and encourage market penetration of cleaner sources of energy. At the 10% tax growth rate, renewable
energy (especially wind energy) quickly becomes cost-competitive with coal-based power. Carbon taxation
- 228 -
as a policy instrument to mitigate climate change will therefore bring great market penetration of clean
technologies in the near future if carefully planned and implemented.
Lastly, though accounting for the life-cycle burdens of coal-based electricity generation was found to
double to triple the price of electricity, making non-fossil fuel sources such as wind energy attractive, all
electricity generation technologies are accompanied by undesirable side-effects at some point in their fuel
cycles, so comparative analyses of life-cycle social costs of all power generation sources in South Africa are
necessary to offer guidance to future energy policy development.
- 229 -
REFERENCES
AEA Technology Environment. 2005. Damages per tonne emission of PM2.5, NH3, SO2, NOx and VOCs from
each EU-25 member state (excluding Cyprus) and surrounding seas. Available at: http://www.cafecba.org/assets/marginal_damage_03-05.pdf (accessed on 14 August 2011).
AES Corporation. 1993. An overview of the IDEAS MODEL: a dynamic long-term policy simulation model of
US. Arlington, VA: Energy Supply and Demand.
African Development Bank (AfDB). 2009. Executive summary of South Africa: Environmental impact
assessment
for
the
Medupi
power
plant
project
of
Eskom.
Available
at:
http://www.afdb.org/fileadmin/uploads/afdb/Documents/Environmental-and-Social-Assessments/ESIA%2
0Ex%20Summary%20of%20Medupi%20Coal%20Power%20Plant%20July%201%20revised%20Final-ram-1.pdf (accessed on 10 September 2011).
Airshed Planning Professionals. 2006. Air quality impact assessment for the proposed new coal-fired power
station (Kendal North) in the Witbank area.
Available at: http://www.golder.com (accessed on 11
December 2012).
Alter, N. & Syed, S.H. 2011. An empirical analysis of electricity demand in Pakistan. International Journal of
Energy Economics and Policy, 1(4):116-139.
Ampofo-Anti, N.L. 2008. Life cycle assessment: applications and implications for the greening of the South
African construction sector. CSIR, Pretoria, South Africa.
Anand, S., Vrat, P. & Dahiya, R.P. 2005. Application of a system dynamics approach for assessment and
mitigation of CO2 emissions from the cement industry. Journal of environmental management, 79(4):38398.
Andries, J. M., Janssen, M. A. & Ostrom, E. 2004. A framework to analyze the robustness of socio-ecological
systems
from
an
institutional
perspective.
Ecology
and
Society,
http://www.ecologyandsociety.org/vol9/iss1/art18 (accessed on 14 July 2011).
- 230 -
9,
18.
Available
at:
Androulidakis, I. & Karakassis, I. 2005. Evaluation of the EIA system performance in Greece, using quality
indicators. Environ Impact Assess Rev, 26:242–56.
Armstrong, J. E. & Harman, W.W. 1980. Strategies for conducting technology assessment, Boulder,
Colorado, Westview Press.
ARPEL. 2010. Guidelines for an environmental impact assessment (EIA) process. Available at:
http://www.arpel.org/media/apps/library/63/files/GUIDELINE%2010.pdf (accessed on 1 December 2012).
Arrow, K.J. & Scitovsky, T. 1969. Readings in welfare economics. London: George Allen and Unwin.
Arvanitidis, P.A. 2006. A framework of socioeconomic organisation: redefining original institutional
economics along critical realist philosophical lines. 46th Congress of the European Regional Science
Association (ERSA). Volos, 30 August – 3 September 2006. Available at: http://www-sre.wu-wien.ac.at
/ersa/ersaconfs/ersa06/papers/575.pdf (accessed on 18 June 2010).
Ashourian, M.H., Cherati, S.M., Mohd Zin, A.A., Niknam, N., Mokhtar, A.S. & Anwari M. 2013. Optimal green
energy management for island resorts in Malaysia. Renewable Energy, 51:36-45.
Asiedu, Y. & Gu, P. 2010. Product life cycle cost analysis: State of the art review. International Journal of
Production Research, 36(4):883-908.
Aslani, A., Helo, P. & Naaranoja, M. 2014. Role of renewable energy policies in energy dependency in
Finland: system dynamics approach. Applied Energy, 113:758–765.
Assefa, G. & Frostell, B. 2006. Technology assessment in the journey to sustainable development. IN
MUDACUMURA, G., MEBRATU, D. & HAQUE, M. S. (Eds.) Sustainable development policy and
administration USA, Taylor and Francis Group.
Aucamp, I., Woodborne, S., Perold, J., Bron, A. & Aucamp, S.M. 2011. Looking beyond impact assessment to
social sustainability. In: Vanclay F, Esteves AM, editors. New directions in social impact assessment:
conceptual and methodological advances. Cheltenham: Edward Elgar.
Auerhahn, K. 2008. Dynamic systems simulation analysis: a planning tool for the new century. Journal of
Criminal Justice, 36:293-300.
- 231 -
Australian Academy of Technological Sciences and Engineering (ATSE). 2009. The hidden costs of electricity:
Externalities of power generation in Australia. Available at: http://www.atse.org.au (accessed on 5 January
2011).
Ayres, C. E. 1962. The Theory of Economic Progress: a Study of the Fundamental Economic Development and
Cultural Change, 2nd ed. New York: Schocken.
Babbie, E. & Mouton, J. 2001. The practice of social research. South African edition. Cape Town: Oxford
University Press.
Backus, G., Green, J. & Masevice, A. 1979. FOSSIL 79: documentation, Dartmouth College, Hanover NH,
Resource Policy Center.
Ball, M., Wietschel, M. & Rentz, O. 2007. Integration of a hydrogen economy into the German energy
system: an optimising modelling approach. International Journal of Hydrogen Energy, 32(10–11):1355–
1368.
Bandyopadhyay, P., Groo, A., Hartley, M., LeBrun, J. & Moazed, A. 2008. Renewable energy for BHP Billiton.
Unpublished Master’s Dissertation. Ann Arbor: University of Michigan.
Bardhan, S. 2012. Assessment of water resource consumption in building construction in India. In:
Villacampa, Y. & Brebbia, C.A. (eds.) Ecosystems and Sustainable Development VIII. UK: WIT Press.
Barisa, A., Romagnoli, F., Blumberga, A. & Blumberga, D. 2014. Future biodiesel policy designs and
consumption patterns in Latvia: a system dynamics model. Journal of Cleaner Production, xxx:1-12.
Barker, A. & Jones C.E. 2013. A critique of the performance of EIA within the offshore oil and gas sector.
Environ Impact Assess Rev, 43:31–9.
Barker, A. & Wood, C. 1999. An evaluation of EIA system performance in eight EU countries. Environ Impact
Assess Rev, 18(4):387–404.
- 232 -
Barlas, Y. & Carpenter, S. 1990. Philosophical roots of model validation: two paradigms. System Dynamics
Review, 6(2):148–166.
Barlas, Y. 1989. Multiple tests for validations of system dynamics type of simulation models. European
Journal of Operational Research, 42(1):59–87.
Barlas, Y. 1994. Model validation in system dynamics. In: Wolstenholme, E., Monaghan, C. (eds.),
Proceedings of the 1994 International System Dynamics Conference. System Dynamics Society, Scotland:
Sterling.
Barlas, Y. 1996. Formal aspects of model validity and validation in system dynamics. System Dynamics
Review, 12:183-210.
Barney, G. O., Eberlein, R., Qu, W. & Sharma, P. D. 1995. The Threshold 21 Sustainable Development Model.
System Dynamics 1:21-29.
Bassi, A.M. 2006. Modelling US. energy with Threshold 21 (T21). Millennium Institute, Arlington, VA.
Available at: http://www.systemdynamics.org/conferences/2006/proceed/papers/BASSI205.pdf (accessed
21 May 2011).
Bassi, A.M., Shilling, J.D. & Herren, H.R. 2007. Informing the US. energy policy debate: How can we deal
with
rising
demand
and
constrained
supply?
Available
at:
http://www.millennium-
institute.org/resources/elibrary/papers/T21USAreport.pdf (accessed 13 April 2011).
Batabyal, A.A., Kahn J.R. & O’Neill, R.V. 2003. On the Scarcity Value of Ecosystem Services. Journal of
Environmental Economics and Management, 46:334-352.
Baumol, W.J. & Oates, W.E. 1988. The theory of environmental policy. 2nd ed. Cambridge, U.K.: Cambridge
University Press.
Baumol, W.J. & Oates, W.E. 1993. The theory of environmental policy. Cambridge, U.K.: Cambridge
University Press.
- 233 -
Beed, C. & Beed, C. 2006. Alternatives to economics: Christian socio-economic perspectives. Lanham, MD:
University Press of America.
Bekele, G. & Palm, B. 2009. Wind energy potential assessment at four typical locations in Ethiopia. Appl
Energy, 86(3):388–96.
Berg, I. V. 1994. Technology assessment in Europe. A documentation of TA research establishments.
Karlsruhe, Germany, Kernforschungszentrum.
Berglund, C. & Soderholm, P. 2006. Modelling technical change in energy system analysis: analysing the
introduction of learning-by-doing in bottom-up energy models. Energy Policy, 34:1344–1356.
Berkhout, F. & Goudson, A. P. 2003. Inducing, shaping, modulating: perspectives on technology and
environmental policy. IN BERKES, F., LEACH, M. & SCOONES, I. (Eds.) Negotiating environmental change:
new perspectives from social science. Cheltenham: Edward Elgar.
Berloznik, R. & Van Langenhove, L. 1998. Integration of Technology Assessment in R&D Management
Practices. Technological Forecasting and Social Change, 58:23-33.
Bernow, S., Biewald, B. & Marron, D. 1991. Full-cost dispatch: incorporating environmental externalities in
electric system operation. The Electricity Journal, 4(2):20–33.
Beyond
Petroleum
(BP).
2012.
BP
Statistical
Review
of
World
Energy.
Available
at:
http://www.bp.com/liveassets/bp_internet/globalbp/globalbp_uk_english/reports_and_publications/statis
tical_energy_review_2011/STAGING/local_assets/pdf/statistical_review_of_world_energy_full_report_201
2.pdf (accessed 11 April 2013).
Bilginsoy, C. 1994. Quesnay's tableau economique: analytics and policy implications. Oxford Economic
Papers, 46:519-533.
Bjureby, E., Britten, M., Cheng, I., Kaźmierska, M., Mezak, E., Munnik, V., Nandi, J., Pennington, S., Rochon,
E., Schulz, N., Shahab, N., Vincent, J. & Wei, M. 2008. The true cost of coal: how people and the planet are
paying the price for the world’s dirtiest fuel. Available at: http://www.greenpeace.org/australia/en whatwe-do/climate/resources/reports/the-true-cost-of-coal/ (accessed on 3 July 2011).
- 234 -
Black and Veatch. 2011. 1: Levelized cost of energy calculation. Methodolgy and sensitivity. Available at:
http://www.efchina.org/Attachments/Report/reports-efchina-20110128-en/Levelized%20Cost%20of%20
Energy%20Calculation_BV_EN.pdf (accessed on 14 August 2011).
Blaxter, L., Hughes, C. & Tight, M. 2006. How to research. 3rd ed. Maidenhead, Berkshire, England: Open
University Press, McGraw-Hill.
Blignaut, J., Koch, S., Riekert, J., Inglesi-Lotz, R. & Nkambule, N. 2011. The external cost of coal-fired power
generation:
the
case
of
Kusile.
Available
at:
http://www.greenpeace.org/africa/Global/africa/
publications/coal/FULL%20SCIENTIFIC%20PAPER%20139%20pages.pdf (accessed on 18 December 2011).
Blignaut, J., Mander, M., Schulze, R., Horan, M., Dickens, C., Pringle, K., Mavundla, K., Mahlangu, I., Wilson,
A., McKenzie, M. & McKean, S. 2010. Restoring and managing natural capital towards fostering economic
development: Evidence from the Drakensberg, South Africa. Ecological Economics, 69:1313–1323.
Blignaut, J.N. & Hassan, R.M. 2001. A Natural Resource Accounting Analysis of the contribution of mineral
resources to sustainable development in South Africa. South African Journal of Economic and Management
Sciences, SS(3):S-1- S-39.
Blignaut, J.N. & King, N. 2002. The externality cost of coal combustion in South Africa. Forum for Economics
and Environment Conference 2002, Cape Town. Available at: http://www.anthonyturton.com/admin/m
y_documents/my_files/Economic_costs_of_coal_combustion_in_RSA.pdf (accessed 8 February 2011).
Blignaut, J.N. 2012. Climate change: the opportunity cost of Medupi and Kusile power stations. Journal of
Energy in Southern Africa, 23(4):67-75.
Blignaut, J.N., Chitiga-Mabugu, M.R. & Mabugu, R.M. 2005. Constructing a greenhouse gas emissions
inventory using energy balances: the case of South Africa for 1998. Journal of Energy in Southern Africa, 16
(3):21-32.
Bloomberg New Energy Finance (BNEF). 2011. Chile levelised cost of energy. Available at:
http://docs.nrdc.org/energy/files/ene_11052401a.pdf (accessed on 20 January 2012).
- 235 -
Boardman, A., Greenberg, D., Vining, A. & Weimer, D. 2006. Cost-benefit analysis: concepts and practice.
New Jersey: Prentice Hall.
Bohlweki Environmental, 2006. Environmental impact report for the proposed establishment of a new coalfired power station in the Lephalale Area, Limpopo Province. Final Report - REF NO: 12/12/20/695.
Midrand, South Africa.
Boldeman, L. 2007. The cult of the market: Economic fundamentalism and its discontents. Australia: ANU E
Press.
Booras, G. & Holt, N. 2004. EPRI. Pulverized coal and IGCC plant cost and performance estimates.
Gasification Technologies, Washington.
Bosch, P. & Kuenen, J. 2009. Greenhouse gas efficiency of industrial activities in EU and non-EU. TNO build
environment and geosciences report. TNO-034-UT-2009-01420_RPT-ML. Utrecht, The Netherlands.
Botes, F. 2006. Impact of transport pricing in South Africa on freight transport costs. Available at:
http://www.hsrc.ac.za/research/output/outputDocuments/5616_Botes_Theimpactoftransport.pdf (accessed on 17 January 2013).
Boyce, J.W. 2011. Differences in environmental economics and ecological economics: A contrast of
resource management and sustainability of forests. Available at: http://www.africaimpactgroup.com/wpcontent/uploads/Research_Environmental_Economics_VS_Ecological_Economics.pdf (accessed 13 May
2013).
Branker, K., Pathak, M.J.M. & Pearce, J.M. 2011. A review of solar photovoltaic levelized cost of electricity.
Renewable and Sustainable Energy Reviews, 15:4470-4482.
Brannon, I. 2005. What is a life worth? Available at: http://object.cato.org/sites/cato.org/file
s/serials/files/regulation/2004/12/v27n4-8.pdf (accessed on 14 September 2014).
Brennan, G. & Moehler, M. 2010. Neoclassical economics. Available at: http://www.moehler.org/files
/Neoclassical%20Economics.pdf (accessed 18 February 2013).
- 236 -
Brooks, H. 1994. Technology assessment. In Salomon, J.J., Sagasti, F.R. & Sachs-Jeantet (Eds.) The uncertain
quest: Science, Technology and Development. Tokyo, Japan, United Nations University Press.
Burdge, R.J. 2003. Benefiting from the practice of social impact assessment. Impact Assess Proj Apprais,
21(3):225–9.
Burrell, G. & Morgan, G. 1979. Sociological paradigms and organizational analysis: elements of the
sociology of corporate life. London: Heinemann Educational.
Business Day and Financial Mail (BDFM) Publishers. 2013a. China’s top utilities protest plan to ban lowgrade coal imports. Available at: http://www.bdlive.co.za/world/asia/2013/05/30/chinas-top-utilitiesprotest-plan-to-ban-low-grade-coal-imports (accessed on 18 June 2013).
Business Day and Financial Mail (BDFM) Publishers. 2013b. Eskom seeks ‘coal pact’ to avert extra tariff
hikes. Available at: http://www.bdlive.co.za/business/energy/2012/11/28/eskom-seeks-coal-pact-to-avertextra-tariff-hikes (accessed 21 January 2013).
Business Day and Financial Mail (BDFM) Publishers. 2013c. Numsa to protest power price increase.
Available at: http://www.bdlive.co.za/business/energy/2013/01/14/numsa-to-protest-powerprice-increase
(accessed on 26 January 2013).
Business Day and Financial Mail (BDFM) Publishers. 2013d. Lower carbon tax still risk to growth. Available
at: http://www.bdlive.co.za/national/science/2013/05/03/lower-carbon-tax-still-risk-to-growth (accessed
on 15 April 2013).
Butler, E. 2011. The condensed wealth of nations and the incredibly condensed theory of moral sentiments.
Available at: http://www.adamsmith.org/sites/default/files/resources/condensed-WoN.pdf (accessed 11
June 2013).
Canelas, L., Almansa, P., Merchan, M. & Cifuentes, P. 2005. Quality of environmental impact statements in
Portugal and Spain. Environ Impact Assess Rev, 25:217–25.
Canter, L.W. 1996. Environmental impact assessment. McGraw-Hill International Editions.
- 237 -
Cashmore, M., Gwilliam, R., Morgan, R., Cobb, D. & Bond, A. 2004. The interminable issue of effectiveness:
substantive purposes, outcomes and research challenges in the advancement of environmental impact
assessment theory. Impact Assess Project Appraisal, 22(4):295–310.
CEFIC. 1997. Technology assessment: a tool towards sustainable chemical industry. European Chemical
Industry Council, Brussels.
Cepeda, M. & Finon, D. 2013. How to correct for long-term externalities of large-scale wind power
development by a capacity mechanism? Energy Policy, 61:671–685.
Chamber of Mines. 2011. Mining: Chamber of Mines CEO comments on coal supply to Eskom. Available at:
http://www.bullion.org.za/documents/Chamber-Media-Statement-on-Coal-Supplies.pdf (accessed on 19
February 2011).
Chamber
of
Mines.
2012.
South
African
mining
industry
snapshot
2011.
Available
at:
http://www.chamberofmines.org.za (accessed on 7 March 2013).
Checkland, P. 1981. Systems thinking, systems practice. Chichester: John Wiley and Sons.
Chernick, P. & Caverhill, E. 1989. The valuation of externalities from energy production, delivery, and use
fall, 1989 update. A Report to the Boston Gas Company, PLC Inc., Boston, Massachusetts.
Chi, K.C., Reiner, D.M. & Nuttall, W.J. 2009. Dynamics of the UK natural gas industry: system dynamics
modelling and long-term energy policy analysis. EPRG Working Paper 0913. Available at:
http://www.econ.cam.ac.uk/dae/repec/cam/pdf/cwpe0922 (accessed 11 June 2013).
Christensen, P., Kornov, L. & Nielsen, E. 2005. EIA as Regulation: Does it Work? J Environ Plann Manage,
48(3):393–412.
Clarke, L.B. 1996. Externalities and coal-fired power generation. IEAPER/29. IEA Coal Research, London.
Available at: http://www.caer.uky.edu/iea/ieaper29.shtml (accessed 17 February 2010).
Coaltech. 2009. Coaltech transport investigation. Available at: http://www.coaltech.co.za (accessed on 15
August 2011).
- 238 -
Coates, J. F. 2001. A 21st century agenda for technology assessment. Technological Forecasting and Social
Change, 67:303-308.
Colander, D., Holt, R. P. & Rosser, J. B. 2004. The changing face of mainstream economics. Review of
Political Economy, 16(4):485–500.
Common, M. 1988. Environmental and Resource Economics: An Introduction. London: Longman.
ContiTech, A.G. 2013. 340 million tons of good reasons for climate protection. Available at:
http://www.contitech.de/pages/news/aktuelles/090928_foerdergurte_en.html (accessed on 2 February
2013).
Cooper, D.R. & Schindler, P.S. 2006. Business research methods. 9th ed. New York, NY: McGraw-Hill.
Cormos, C.C. 2012. Integrated assessment of IGCC power generation technology with carbon capture and
storage (CCS). Energy, 42:434-445.
Cornes, R. & Sandler, T. 1986. The theory of externalities, public goods and club goods. Cambridge:
Cambridge University Press.
Cosic, B., Krajacic, G. & Duic, N. 2012. A 100% renewable energy system in the year 2050: The case of
Macedonia. Energy, x(x):1-8.
Cosmi, C., Cuomo, V., Macchiato, M., Mangiamele, L., Masi, S. & Salvia, M. 2000. Waste management
modelling by MARKAL model: a case study for Basilicata Region. Environmental Modelling and Assessment,
5(1):19–27.
Council for Geoscience. 2010. Mine water management in the Witwatersrand gold fields with special
emphasis on acid mine drainage. Available at: http://www.dwaf.gov.za/Documents/ACIDReport.pdf
(accessed on 12 September 2011).
Coyle, G. 2000. Qualitative and quantitative modelling in system dynamics: some research questions. Syst
Dyn Rev, 16:225–244.
- 239 -
Cracker, T.D. & Rogers III, A.J. 1971. Environmental Economics. Hinsdale: The Dryden Press.
Crawford, R.H. 2011. Life cycle water analysis of an Australian residential building and its occupants.
Available at: http://www.conference.alcas.asn.au/2011/Crawford%20LCA%202011%20revised%20final.pdf
(accessed on 10 February 2012).
Creamer Media. 2013. Eskom concerned that domestic coal prices may migrate to export levels. Available
at: http://m.engineeringnews.co.za (accessed on 25 April 2013).
Crepea, A. T. 1995. A systems engineering approach to technology assessment. Journal of Clinical
Engineering, 20, 297-303.
Creswell, J.W. 2008. Research design: qualitative, quantitative and mixed methods approaches. 3rd ed.
Thousand Oaks: Sage.
Crotty, M. 1998. The foundations of social research: Meaning and perspective in the research process.
Thousand Oaks, CA: Sage.
Curry, G. L., Deuermeyer, B. L., & Feldman, R. M. 1989. Discrete simulation. Oakland: Holden-Day.
Damle, P. 2003. A system dynamics model of the integration of new technologies for ship systems.
Unpublished Master’s Dissertation. Virginia, State University.
Darling, S.B., You, F., Veselka, T. & Velosa, A. 2011. Assumptions and the levelized cost of energy for
photovoltaics. Energy Environmental Science, x:x-x.
Dastkhan, H. & Owlia, M.S. 2014. What are the right policies for electricity supply in Middle East? A regional
dynamic integrated electricity model for the province of Yazd in Iran. Renewable and Sustainable Energy
Reviews, 33:254–267.
Davidsen, P.I., Sterman, J.D. & Richardson, G.P. 1990. A petroleum life cycle model for the United States
with endogenous technology, exploration, recovery, and demand. System Dynamics Review, 6:66–93.
- 240 -
Davis, G.A. & Owens, B. 2003. Optimizing the level of renewable electric R&D expenditures using real
options analysis. Energy Policy, 31:1589–1608.
Davis, J. B. 2005. Heterodox economics, the fragmentation of the mainstream and embedded individual
analysis, in Garnett, R. & Harvey, J. (eds.), The future of heterodox economics, essays in honour of Paul Dale
Bush. Ann Arbor: University of Michigan Press.
Davis, J.B. 2006. The nature of heterodox economics. Post-Autistic Economics Review, 40:23–30.
Davis, S., Cohen, B., Hughes, A., Durbach, I. & Nyatsanza, K. 2010. Measuring the rebound effect of energy
efficiency initiatives for the future: A South African case study. Available at: http://www.sanedi.org.za/wpcontent/uploads/2012/02/ee/Rebound%20Effect%20-%20Main%20report%20-%20Final.pdf (accessed on
17 May 2013).
Davison, J. 2007. Performance and costs of power plants with capture and storage of CO2. Energy,
32:1163–1176.
De Piante, H. A. 1997. A technology assessment primer for management of technology. International
Journal of Technology Management, 13:615-638.
De Souza, P.F., Du Plessis, G.J. & Mackintosh, G.S. 2002. An evaluation of the suitability of the limestone
based side stream stabilisation process for stabilisation of waters of the Lesotho Highlands Scheme.
Available at: http://www.wrc.org.za/Knowledge%20Hub%20Documents/Water%20SA%20Journals/Manus
cripts/2002/05/WaterSA_2002_05_4.pdf (accessed on 12 January 2013).
De Vroey, M. 1975. The transition from classical to neoclassical economics: a scientific revolution. Journal of
Economic Issues, 9(3):415-439.
Dean, E. B. 1995. Parametric cost deployment. Proceedings of the Seventh Symposium on Quality Function
Deployment, June, Nov., MI, US, 27± 34. Debate with T21.
Dekker, J., Nthontho, M., Chowdhury, S. & Chowdhury S.P. 2012. Economic analysis of PV/diesel hybrid
power systems in different climatic zones of South Africa. Electrical Power and Energy Systems, 40:104–
112.
- 241 -
Denholma, P. & Margolis, R.M. 2007. Evaluating the limits of solar photovoltaics (PV) in traditional electric
power systems. Energy Policy, 35:2852–2861.
Denzin, N. K. & Lincoln, Y. S. 1994. Introduction: entering the field of qualitative research. In Denzin, N.K. &
Lincoln, Y.S. (eds.) Handbook of qualitative research. Thousand Oaks, CA: Sage.
Denzin, N. K. & Lincoln, Y. S. 2000. Introduction: The discipline and practice of qualitative research. In
Denzin, N.K. & Lincoln, Y.S. (eds.), Handbook of qualitative research (2nd ed.). Thousand Oaks, CA: Sage.
Department of Energy. 2009. Digest of South African energy statistics 2009. Available at:
http://www.energy.gov.za/files/media/explained/2009%20Digest%20PDF%20version.pdf (accessed on 4
July 2011).
Department of Energy. 2010. South African energy synopsis 2010. Available at: http://www.energy.gov.z
a/files/media/explained/2010/South_African_Energy_Synopsis_2010.pdf (accessed on 13 July 2011).
Department of Environmental Affairs (DEA). 2010. Government Gazette 18 June 2010. Amendment of
regulations in terms of Chapter 5 of the National Environmental Management Act, 1998. Available at:
https://www.environment.gov.za/sites/default/files/gazetted_notices/nema_amendment_g31144rg8904g
on659.pdf (accessed on 16 December 2013).
Department of Environmental Affairs and Development Planning. 2006. NEMA environmental impact
assessment regulations, guideline and information document series: Guideline on the interpretation of the
listed
activities.
Chief
Directorate:
Environmental
and
Land
Management.
Available
at:
http://www.westerncape.gov.za/text/2006/12/dea&dp_nema_eia_guideline_interpretations_listed_activit
ies_nov2006.pdf (accessed on 29 August 2013).
Department of Environmental Affairs and Tourism (DEAT). 1998. EIA Regulations: Implementation of
Sections 21 22 and 26 of the Environment Conversation Act (No.78 of 1989), Guideline Document, April
1998. Available at: http://www.deat.gov.za/PolLeg/GenPolicy/eia.htm (accessed on 11 August 2013).
Department of Environmental Affairs and Tourism (DEAT). 2006a. Guideline 4: public participation, in
support of the EIA regulations 2005. Government Gazette, Vol 491, No. 28854, 19 May 2006.
- 242 -
Department of Environmental Affairs and Tourism (DEAT). 2006b. Presentation on Revised National
Environmental Management Act: EIA regulation. Pretoria: Department of Environmental Affairs and
Tourism (DEAT).
Department of Environmental Affairs and Tourism (DEAT). 2006c. General Notice 1469 of 2006: Guidelines
3. General guidelines to the environmental impact assessment regulations. Government Gazette, 2006.
Department of Environmental Affairs and Tourism (DEAT). 2006d. Regulations in terms of Chapter 5 of the
National
Environmental
Management
Act,
1998.
Available
at:
https://www.environment.go
v.za/sites/default/files/gazetted_notices/nema_enviroauthorisation_regulations_g28753rg8453gon385.pdf
(accessed on 22 August 2013).
Department of Environmental Affairs and Tourism (DEAT). 2008a. Government Gazette 13 June 2008.
Amendment of regulations in terms of Chapter 5 of the National Environmental Management Act, 1998. 13
June
2008.
Available
at:
https://www.environment.gov.za/sites/default/files/gazetted_notices/ne
ma_amendment_g31144rg8904gon659.pdf (accessed on 9 August 2013).
Department of Environmental Affairs and Tourism (DEAT). 2008b. Review of the Effectiveness and Efficiency
of the Environmental Impact Assessment (EIA) system in South Africa. Pretoria: Department of
Environmental Affairs and Tourism.
Department of Environmental Affairs and Tourism. 2004. Cost benefit analysis: Integrated environmental
management series 8. Available at: http://www.environment.gov.za (accessed on 11 July 2013).
Department of Environmental Affairs and United Nations Environment Programme. 2013. Green economy
modelling report for South Africa: focus on the sectors of Natural Resource Management, Agriculture,
Transport and Energy. Available online: https://www.environment.gov.za/sites/default/files/docs/greeneco
nomy_modelingreport.pdf (accessed on 15 August 2014).
Department of Mineral Resources (DMR). 2003. A review of dolomite and limestone industry in South
Africa. Available at: http://www.dme.gov.za (accessed on 13 July 2011).
- 243 -
Department of Mineral Resources (DMR). 2012. South Africa’s mineral industry 2011/2012. Directorate:
Mineral Economics. Available at: http://www.dmr.gov.za/publications/south-africas-mineral-industrysami.html (accessed 21 April 2013).
Department of Mineral Resources. 2009. Unpublished Coal Statistics. Pretoria.
Department of Minerals and Energy (DME). 2005. Dolomite and limestone in South Africa: supply and
demand 2005. Available at: http:// http://www.dme.gov.za (accessed on 1 June 2011).
Department of Minerals and Energy (DME). 2007. Annual report 2006/2007. Available at: http://www.info
.gov.za/view/DownloadFileAction?id=72590 (accessed on 29 January 2013).
Department of Minerals and Energy (DME). 2008. Annual report 2007/08. Available at: http://www.info.
gov.za/view/DownloadFileAction?id=93533 (accessed on 28 June 2011).
Department of Minerals and Energy (DME). 2009. Annual report 2008/09. Available at: http://www.pmg.
org.za/report/20091027-department-mineral-resources-annual-report-200809 (accessed on 29 January
2013).
Department
of
Minerals
and
Energy
(DME). 2010.
Annual
report
2009/10.
Available
at:
http://www.dmr.gov.za/publications/annual-report/summary/26-annual-reports/331-dme-annual-report-2
00910-.html (accessed on 29 January 2013).
Department of Natural Resources & Public Service Commission of Wisconsin (DNR & PSCOW). 2007.
Integrated gasification combined-cycle technology: Benefits, costs, and prospects for future use in
Wisconsin. Available at: http://psc.wi.gov/initiatives/cleanCoal/documents/IGCCFinalReport.pdf (accessed
on 6 October 2012).
Department of Water Affairs (DWA). 2011. Development of a reconciliation strategy for the Olifants river
water supply system: Final reconciliation strategy report. Available at: http://www.dwaf.gov.za/Projects/Ol
ifantsRecon/Documents/Main%20Reports/ORRS%20Final%20Reconciliation%20Strategy%20Report.pdf
(accessed on 11 December 2012).
- 244 -
Dequech, D. 2007. Neoclassical, mainstream, orthodox, and heterodox economics. Journal of Post
Keynesian Economics, 30(2):279-302.
Downs, H. W. & Hansen, R.W. 1998. Estimating farm fuel requirements. Fact sheet No. 5.006. Available at:
http:// http://ext.colostate.edu.(accessed on 6 February 2013).
Du Pisani, J.A. & Sandham, L.A. 2006. Assessing the performance of SIA in the EIA context: a case study of
South Africa. Environ Impact Assess Rev, 26:707–24.
Du Plessis, P. 2011. The potential of solar process heat for South African industry. Unpublished Master’s
Dissertation. UCT, CA.
Durbin, P. T. & Rapp, F. 1983. Philosophy and technology. Boston, D. Reidel.
Dutkiewicz, R.K. & De Villiers, M.G. 1993. Social cost of electricity production. Engineering research. Report
for the National Energy Council, Pretoria.
Dyner, I., Ochoa, C. & Franco, C.J. 2011. SD for assessing electricity market integration. Available at:
http://www.systemdynamics.org/conferences/2011/proceed/papers/P1181.pdf
(accessed
on
23
September 2014).
Ea Energy Analyses, Hagman Energy, COWI. 2008. Congestion management in the Nordic market –
evaluation of different market models. Nordic Council of Ministers; 2008. Available at: http://eaenergianalyse.dk/reports/730_congestion_management_in_the_nordic_market-evaluation_of_different
_market_models.pdf (accessed on 8 September 2013).
Ea Energy Analyses. 2007. 50% Wind Power in Denmark in 2025. Available at: http://eaenergianalyse.dk/reports/642_50_per_cent_wind_power_in_Danmark_in_2025_July_2007.pdf
(accessed
on 23 December 2011).
Ea Energy Analyses. 2008. Large scale wind power in New Brunswick – a regional scenario towards 2025.
Available at: http://www.uwig.org/LargeScaleWindPowerinNBAugust%202008.pdf (accessed on 11 Januray
2011).
- 245 -
Eagly, R.V. 1974. The structure of classical economic theory. Oxford: University Press.
Eberhard. A. 2011. The future of South African coal: market, investment, and policy challenges. Working
Paper No.: 100. Freeman Spogli Institute for International Studies, Stanford. Available at: http://iisdb.stanford.edu/pubs/23082/ WP_100_Eberhard_Future_of_South _African_Coal.pdf (accessed on 16
December 2011).
Eichner, A.S. & Kregel, J. A. 1975. An essay on post-keynesian theory: a new paradigm in economics. Journal
of Economic Literature, 13(4):1293-1314.
Ekelund, R.B. & Herbert, R.F. 1975. A history of economic theory and method. New York: McGraw-Hill.
Electric Power Research Institute (EPRI). 2010. Power generation technology data for Integrated Resource
Plan of South Africa. Palo Alto, CA.
Energy Information Administration. 1995. Electricity generation and environmental externalities: case
studies. Available at: ftp://ftp.eia.doe.gov/pub/electricity/ external.pdf (accessed on 16 March 2011).
Energy Technology Systems Analysis Program. 2007. ETSAP Workshop and Training Course. Brasilia, Grand
Bittar Hotel, November 19-23, 2007. Available at: http://www.iea-etsap.org/web/Workshop/Brazil
_11_2007/Dual-Remme.PDF (accessed on 5 June 2011).
EO Miners. 2011. Test Site 2. Witbank coalfields. Available at: http://www.eo-miners.eu/test_sites
/ts_testsite2_witbank.htm (accessed on 18 August 2011).
Epstein, P.R., Buonocore, J.J., Eckerle, K., Hendryx, M., Stout III, B.M., Heinberg, R., Clapp, R.W., May, B.,
Reinhart, N.L., Ahern, M.M., Doshi, S.K. & Glustrom, L. 2011. Full cost accounting for the life cycle of coal.
Ann. N.Y. Acad. Sci, 1219:73–98.
Eriksson, O. & Frostell, B. 2001. An approach to sustainability assessment of energy systems. IN SILVERIA, S.
(Ed.) Building sustainable energy systems: Swedish experiences. Stockholm, Sweden, Swedish National
Energy Administration.
Eskom Communication. 2012. Personal communication with Gina Downes. Eskom Megawatt Park.
- 246 -
Eskom Communication. 2013. Personal communication with Joubert Deon. Electronically.
Eskom. 2007. Annual Report 2007. Available at: http://financialresults.co.za/2007/eskomar2007_fin/dow
nloads/eskom_ar2007.pdf (accessed on 28 January 2013).
Eskom. 2009a. Eskom’s response to water for growth & development framework. Available at:
http://www.dwaf.gov.za/WFGD/documents/2009Summit/Eskom.pdf (accessed on 11 December 2012).
Eskom. 2009b. Annual report 2007. Available at: http://www.eskom.co.za (accessed on 28 January 2013).
Eskom. 2010a. Integrated report 2010. Available at: http://financialresults.co.za/2010/eskomar2010/d
ownloads/eskom_ar2010.pdf (accessed on 15 June 2011).
Eskom. 2010b. Kusile power station. Available at: http://www.eskom.co.za/c/article/58/kusile-powerstation/ (accessed on 7 July 2011).
Eskom. 2010c. Integration of wind and solar resources and the role of Kusile. Available at: http://www.
eskom.co.za/ (accessed on 6 August 2013).
Eskom. 2011. Integrated report 2011. Available at: http://financialresults.co.za/2011/eskom_ar2011/down
loads/eskom-ar2011.pdf (accessed on 10 February 2012).
Eskom. 2012a. Integrated results presentation for the year ended 31st March 2012. Available
at:http://www.eskom.co.za/content/Eskomresults31Mar2012media_presFINAL~3.pdf (accessed on 17 February 2013).
Eskom. 2012b. Integrated report 2012. Available at: http://financialresults.co.za/2012/eskom _ar2012/inte
grated-report/downloads/full_downloads/01_eskom integrated_report2012.pdf (accessed on 13 February
2013).
Eskom. 2012c. Service functions: Group capital. Available at: http://financialresults.co.za/2012/ eskom_a
r2012/divisional-report/group-capital.php (accessed on 18 February 2013).
- 247 -
Eskom. 2013a. Eskom summary document: Applications for postponement of the Minimum Emissions
Standards (MES) for Eskom’s coal and liquid fuel-fired power stations. Available at: http://www.iliso.co
m/2013-12-20/ESKOM%20Applications%20-%20Summary_Final_V1_2013%2012%2019_KL.pdf
(accessed
on 1 February 2014).
Eskom. 2013b. Tariffs & Charges. Available at: http://www.eskom.co.za/c/53/tariffs-and-charges (accessed
on 16 December 2013).
Esterhuizen, I. 2013. Carbon pricing idea gaining steam as global leaders recognise climate risks. Available
at: http://www.engineeringnews.co.za/article/carbon-pricing-idea-gaining-steam-as-global-leaders-recogn
ise-climate-risks-2013-07-23 (accessed on 29 August 2013).
European Commission. 1995. ExternE: Externalities of energy. Vol. 3: coal and lignite. Luxembourg: Office
for Official Publications of the European Communities.
European Commission. 1999a. Guidelines for the assessment of indirect and cumulative impacts as well as
impact interactions. Luxemburg: Office for Official Publications of the European Communities.
European Commission. 1999b. ExternE: externalities of energy. Volume 7-10. Luxembourg: Office for Official
Publications of the European Communities.
European Commission. 2004. New elements for the assessment of external costs from energy technologies.
Available at:
http://www.ier.uni-stuttgart.de/forschung/projektwebsites/newext/
(accessed on 21
December 2012).
European Commission. 2005. ExternE: externalities of energy, methodology 2005 update. Available at:
http://ec.europa.eu/research/energy/pdf/kina_en.pdf (accessed 17 March 2011).
European Commission. 2008. Guide to cost benefit analysis of investment project. DG Regional Policy,
European Commission, Brussels.
Evans, D. J. & Sezer, H. 2005. Social discount rates for member countries of the European Union. Journal of
Economic Studies, 32(1):47-59.
- 248 -
EVRAZ. 2009. Annual Report 2009. Available at: http://www.evrazhighveld.co.za/annual_reports/Highv
eld_Steel_and_Vanadium_Annual_Report_2009.pdf (accessed on 11 February 2013).
ExternE-Pol. 2005. ExternE-Pol externalities of energy: extension of accounting framework and policy
applications. Available at: http://www.externe.info (accessed on 21 December 2012).
Faaij, A., Meuleman, B., Turkenburg, W., Van Wijk, A., Bauen, A., Rosillo-Calle, F. & Hall, D. 1998.
Externalities of biomass based electricity production compared with power generation from coal in the
Netherlands. Biomass and Energy, 14(2):125-147.
Fan, Y., Yang, R.G. & Wei, Y.M. 2007. A system dynamics based model for coal investment. Energy,
32(6):898–905.
Fankhauser, S., & D. McCoy, 1995: Modelling the economic consequences of environmental policies. In.
Principles of environmental and resource economics: a guide to decision makers and Students. H. Folmer, L.
Gabel, J. Opschoor, (eds.). Aldershot, UK: Edward Elgar.
Feng, Y.Y., Chen, S.Q. & Zhang L.X. 2013. System dynamics modeling for urban energy consumption and CO2
emissions: a case study of Beijing, China. Ecological Modelling, 252:44– 52.
Feng, Y.Y., Chen, S.Q. & Zhang, L.X. 2013. System dynamics modeling for urban energy consumption and
CO2 emissions: a case study of Beijing, China. Ecol. Modell., 252:44–52.
Ferguson, C.E. 1972. Microeconomic theory. Homewood, Illinois: Richard D. Irwin.
Fiddaman, T.S. 1997. Feedback complexity in integrated climate–economy models. Cambridge, MA, MIT:
Sloan School of Management.
Fishbone, L.G. & Abilock, H. 1981. MARKAL, a Linear-Programming Model for Energy Systems Analysis:
Technical Description of the BNL Version. International Journal of Energy Research, 5:353-375.
Fisher, L. 2011. The treatment of socio economic impact in environmental impact statements. Unpublished
Master’s Dissertation. School of Environmental Sciences, University of East Anglia. Available at:
- 249 -
http://www.uea.ac.uk/env/all/teaching/eiaams/pdf_dissertations/2011/Fisher_Laura_20 11.pdf (accessed
on 21 June 2013).
Fleischer, T., Decker, M. & Fiedeler, U. 2005. Assessing emerging technologies - methodological challenges
and the case of nanotechnologies. Technological Forecasting and Social Change, 72:1112-1121.
Fleury, A., Fichtner, W. & Rentz, O. 2003. Energy system analysis models for sustainable energy production.
Environmental research, engineering and management, 4(26):15-20.
Flood, R.L. & Jackson, M.C. 1991. Creative problem solving: Total systems intervention. Chichester: John
Wiley & Sons.
Ford, A. 1999. Modelling the environment. An introduction to system dynamics modelling of environmental
systems. Washington D.C.: Island Press.
Ford, A., Vogstad, K. & Hilary, F. 2007. Simulating price patterns for tradable green certificates to promote
electricity generation from wind. Energy Policy, 35:91-111.
Forrester, J. W. 1987. Lessons from system dynamics modelling. System Dynamics Review, 3:136-149.
Forrester, J.W 1958. Industrial dynamics - a major breakthrough for decision makers. Harvard Business
Review, 36(4):37-66.
Forrester, J.W. & Senge, P.M. 1980. Tests for building confidence in system dynamics models. TIME Studies
in the Management Science, 14: 209–228.
Forrester, J.W. 1961. Industrial Dynamics. Cambridge, Massachusetts: MIT Press.
Forrester, J.W. 1971. Principles of systems. Wright-Allen Press, Cambridge.
Forrester, J.W. 1975a. Industrial dynamics – after the first decade. In: Collected papers of Jay W. Forrester.
Wright-Allen Press, Cambridge.
- 250 -
Forrester, J.W. 1975b. Industrial Dynamics: A major breakthrough for decision makers. In: Collected papers
of Jay W. Forrester. Wright-Allen Press, Cambridge.
Forrester, J.W. 2000. From the ranch to system dynamics: an autobiography. In Bedeian, A.G. Management
laureates: a collection of autobiographical essays, vol. 1. Greenwich, CT: JAI Press
Forrester, J.W. 2003. Dynamic models of economic systems and industrial organizations. Syst Dyn Rev,
19:331–345.
Fox, W. & Bayat, M.S. 2007. A guide to managing research. Cape Town: Juta.
Friedrich, R. & Voss, A. 1993. External costs of electricity generation. Energy Policy, 21:114-122.
Fritz, M. 2004. Ludwig von Mises: a scholar who would not compromise. Available at:
http://www.mises.org/dai ly/1700 (accessed on 12 May 2013).
Fuller, S.K. & Petersen, S.R. 1996. Life cycle costing manual for the federal energy management program.
NIST handbook 135. Prepared for US Department of Energy.
Gabriel, S. 2003. Heterodox economic theory: introduction to political economy. Available at:
https://www.mtholyoke.edu/courses/sgabriel/heterodox_defined.htm (accessed on 17 February 2013).
Gaffen, M., Naude, C., Lombaard, P., Maasdorp, G., Taylor, A. & Pretorius, J. 2000. A quantitative analysis of
the full cost associated with motor vehicle use in South Africa. Report for the Department of Transport,
funded by USAID.
Galinis, A. & Van Leeuwen, M.J. 2000. A CGE model for Lithuania: the future of nuclear energy. Journal of
Policy Modelling, 22:691-718.
Georgakellos, D.A. 2010. Impact of a possible environmental externalities internalisation on energy prices:
the case of the greenhouse gases from the Greek electricity sector. Energy Economics, 32:202–209.
Ghader, S. F., Azadeh, M.A. & Zahed, S.M. 2006. Modelling and forecasting the electricity demand for major
economic sectors in Iran. Information Technology Journal, 5(2):260-266.
- 251 -
Ghosh, S. 2002. Electricity consumption and economic growth in India. Energy Policy, 30:125-129.
Giancarlo, V. 2011. II Tableau économique di François Quesnay. Contributions to Political Economy,
30(1):99-101.
Gibson, R.B., Winfield, M., Markvart, T., Gaudreau, K. & Taylor, J. 2008. An analysis of the Ontario power
authority’s consideration of environmental sustainability in electricity system planning. Studies in Ontario
electricity policy series Paper No. 2. University of Waterloo and York University.
Glasson J, Heaney D. 1993. Socio-economic impacts: the poor relations in British Environmental Impact
Statements. J Environ Plan Manag; 36(3):335–43
Glasson, J., Therivel, R. & Chadwick, A. 2005. Introduction to Environmental Impact Assessment. 3rd ed.
London: Routledge.
Glasson, J. 2001. Socio-economic impacts 1: overview and economic impacts. In: Morris P, Therivel R,
editors. Methods of environmental impact assessment. 2nd ed. London: UCL Press; 2001.
Glasson, J. & Heaney. D. 1993. Socio-economic impacts: the poor relations in British Environmental Impact
Statements. J Environ Plan Manag, 36(3):335–43.
Glasson, J. 2001. Socio-economic impacts 1: overview and economic impacts. In: Morris P, Therivel R,
editors. Methods of environmental impact assessment. 2nd ed. London: UCL Press; 2001.
Goldblatt, M., Gelb, S. & Davies, G. 2002. Macroeconomics and sustainable development in Southern
Africa: synthesis report of South African study. Development Bank of South Africa. Midrand, South Africa.
Goodall, D.W. 1972. Building and testing ecosystem models. In J.N.J. Jeffers (ed.), Mathematical models in
ecology. Oxford: Blackwell.
Grant, W.E., Pederson, E.K. & Marin, S.L. 1997. Ecology and natural resource management. Systems analysis
and simulation. New York: John Wiley and Sons Inc.
- 252 -
Gray, I.M. & Edwards-Jones, G. 1999. A review of the quality of environmental impact assessments in the
Scottish forest sector. Forestry, 72(1):1-10.
Greenstone, M. & Looney, A. 2012. Paying too much for energy? The true costs of our energy choices.
Available at: http://www.amacad.org/publications/daedalus/12_spring_greenstone_looney.pdf (accessed
on 20 April 2013).
Greves, D. & Schreiber, B. 1993. Engineering costing techniques in ESA. Available at: http://esapub.esriu.
esa.it.pointtobullet/greves1.html (accessed on 4 March 2013).
Gronow, T. 2001. Material flow models in environmental policy planning. Case: pulp and paper industry.
Unpublished Doctoral Thesis. Jyvaskyla Studies in Business and Economics 13. University of Jyvaskyla.
Grubler, A. 1998. Technology and global change. Cambridge, Cambridge University Press.
Grubler, A., Nakicenovic, N. & Nordhaus, W.D. 2002. Technological change and the environment,
Washington, DC:Resources for the Future.
Guba, E.G. & Lincoln, Y.S. 1981. Effective evaluation: improving the usefulness of evaluation results through
responsive and naturalistic approaches. San Francisco: Jossey-Bass.
Guba, E.G. & Lincoln, Y.S. 1994. Competing paradigms in qualitative research. In: Denzin, N.K. & Lincoln, Y.S.
(eds.) Handbook of qualitative research. Thousand Oaks, CA: Sage.
Guba, E.G. & Lincoln, Y.S. 2005. Paradigmatic controversies, contradictions, and emerging confluences. In:
Denzin, N.K. & Lincoln, Y.S. (eds.) The Sage handbook of qualitative research. 3rd ed. Thousand Oaks, CA:
Sage.
Guh, S., Xingbao, C., Poulos, C., Qi, Z., Jianwen, C., von Seidlein, L., Jichao, C., Wang, X., Zhanchun, X.,
Nyamete, A., Clemens, J. & Whittington, D. 2008. Comparison of cost-of-illness with willingness-to-pay
estimates to avoid shigellosis: evidence from China. Health Policy and Planning, 23:125-136. Available from:
http://heapol.oxfordjournals.org/content/23/2/125.full.pdf+html (Accessed: 2011-05-12).
Hahn, F. H. 1982. The neo-Ricardians. Cambridge Journal of Economics, 6:353–74.
- 253 -
Hahn, F. H. 1984. Equilibrium and Macroeconomics. Oxford: Basil Blackwell.
Hammond, D. 2005. Philosophical and ethical foundations of systems thinking. TripleC, 3(2):20-27.
Harrison, M. 2010. Valuing the future: the social discount rate in cost-benefit analysis. Visiting researcher
paper, Productivity Commission, Canberra. Harvard Business Review, 37–66.
Haw, M. & Hughes, A. 2007. Clean energy and development for South Africa: Background data. Energy
Research Centre University of Cape Town and British High Commission South Africa. South Africa. Available
at: http://www.erc.uct.ac.za/Research/publications/07HawHughes%20Clean%20energy%20&%20develop
ment%20-%201.pdf (accessed on 15 May 2012).
Hearps, P., McConnell, D., Sandiford, M. & Dargaville, R. 2011. Renewable energy technology cost review.
Available at: http://www.earthsci.unimelb.edu.au/_rogerd/Renew_Energy_Tech_Cost_Review.pdf (accessed on 15 May 2012).
Heggedal, A.M. 2006. Investment in new transmission capacity between Estonia and Finland – effects on
the electricity market and welfare. Unpublished Master’s Dissertation. Department of Economics and
Natural
Resource
Management,
Norwegian
University
of
Life
Sciences.
Available
at:
http://www.eabalmorel.dk/files/download/Trainingandeducation/Investment%20in%20new%20transmissi
on%20capacity%20between%20Estonia%20and%20Finland.pdf (accessed on 21 June 2013).
Heinma K, Poder T. 2010. Effectiveness of Environmental Impact Assessment system in Estonia. Environ
Impact Assess Rev, 30:272–7.
Heinma, K. & Poder, T. 2010. Effectiveness of Environmental Impact Assessment system in Estonia. Environ
Impact Assess Rev, 30:272–7.
Hekimoğlu, M. & Barlas, Y. 2010. Sensitivity analysis of system dynamics models by behavior pattern
measures. Available at: http://www.systemdynamics.org/conferences/2010/proceed/papers/P1260.pdf
(accessed on 16 June 2014).
- 254 -
Hendrickson, C.T., Horvath, A., Joshi, S. & Lave, L.B. 1998. Economic input-output models for environmental
life cycle assessment. Environ. Sci. Technol., 32(47):184-191.
Herna´ndez-Moro, J. & Martı´nez-Duart, J.M. 2013. Analytical model for solar PV and CSP electricity costs:
present LCOE values and their future evolution. Renewable and Sustainable Energy Reviews, 20:119–132.
Heylighen, F. & Joslyn, C. 1992. What is systems theory? Available at: http://pespmc1.vub.ac.be/SYST
HEOR.html (accessed on 12 January 2013).
Hildebrandt, L. & Sandhamb, L.A. 2014. Social impact assessment: the lesser sibling in the South African EIA
process? Environmental Impact Assessment Review, 48:20–26.
Hiles, D.R. 1999. Paradigms lost - paradigms regained. Paper presented to 18th International Human
Science Conference, July 26 - 29, Sheffield, UK. Available from: http://www.psy.dmu.ac.uk/drh
iles/Paradigms/Lost.htm (accessed on 15 August 2011).
Hjorland, B. & Nicolaisen, J. 2005. The epistemological lifeboat: epistemology and philosophy of science for
information scientists. Available at: http://www.iva.dk/jni/lifeboat/ (accessed on 12 January 2013).
HM Treasury. 2003. The green book: appraisal and evaluation in central government. London: HM Treasury.
Hodge, B., Aydogan-Cremaschi, S., Blau, G., Pekny, J. & Reklaitis, G. 2008. A prototype agent-based
modelling approach for energy system analysis. 18th European Symposium on Computer Aided Process
Engineering – ESCAPE 18. Computer Aided Chemical Engineering, 25:1071-1076.
Hodgson, G. M. 1988. Economics and Institutions. Oxford: Polity.
Hoffmann, B., Hafele, S. & Karl, U. 2013. Analysis of performance losses of thermal power plants in
Germany - a system dynamics model approach using data from regional climate modeling. Energy, 49:193203.
Hoffmann, B.S. & Szklo, A. 2011. Integrated gasification combined cycle and carbon capture: A risky option
to mitigate CO2 emissions of coal-fired power plants. Applied Energy, 88:3917–3929.
- 255 -
Hohmeyer, O. 1988. Social costs of energy consumption. Berlin: Springer-Verlag.
Hohmeyer, O. 1992. The social costs of electricity-renewables versus fossil and nuclear energy.
International Journal of Solar Energy, 11:231-250.
Holling, C.S. 1978. In: Adaptive environmental assessment and management. John Wiley & Sons, New York,
NY 377. Homage to Mises. Hillsdale College.
Hoosen, F. 2010. An investigation into the role of public participation in achieving social justice: a case
study of EIAs undertaken (under old and new regulations) in south Durban. Unpublished Master’s
Dissertation. Braamfontein: University of Witwatersrand.
Hou, Y., Xia, X., Zhang, Z., Lou, Z., Zhang, F. & Xin, C. 2009. Research on system dynamics simulation of coal
resources in Hebei province. Coal Engineering, 1:68–71.
Hughes, A., Haw, M., Winkler, H., Marquard, A. & Merven, B. 2007. Energy modelling: a modelling input
into the Long Term Mitigation Scenarios process, LTMS Input Report 1. Energy Research Centre, Cape
Town. Available at: http://www.erc.uct.ac.za/Research/publications/07Hughesetal-LTMSEnergy_Modellin
g.pdf (accessed on 11 July 2013).
Hugo, M.L. 2004. Environmental management: an ecological guide to sustainable living in Southern Africa.
Brooklyn Square: Ecoplan.
Hugo, M.L. 2004. Environmental management: an ecological guide to sustainable living in Southern Africa.
Brooklyn Square: Ecoplan.
Hugo, M.L. 2004. Environmental management: an ecological guide to sustainable living in Southern Africa.
Brooklyn Square: Ecoplan.
Huppes, G. & Suh, S. 2002. Towards global IOA. Available at: http://www.lcaconf.alcas.asn.au/Papers/12
(accessed on 8 February 2011).
Icyk, B. 2006. At what cost? A comparative evaluation of the social costs of selected electricity generation
alternatives in Ontario. Unpublished Master’s Dissertation. Canada: University of Waterloo. Available at:
- 256 -
http://www.environment.uwaterloo.ca/research/greenpower/projects/documents/Icyk
summary.pdf
(accessed on 16 April 2011).
IndexMundi. 2012a. South Africa hydraulic cement production by year. Available at: http://www.ind
exmundi.com/minerals/?country=za&product=cement&graph=production (accessed on 29 January 2013).
IndexMundi. 2012b. South Africa aluminium, primary production by year. Available at: http://www.indexm
undi.com/minerals/?country=za&product=aluminum&graph=production (accessed on 29 January 2013).
InEnergy. 2010. Cement and Concrete Institute: concrete industry greenhouse gas emissions. Northcliff,
South Africa. Available at: http://www.cnci.org.za (accessed on 29 July 20012).
Inglesi-Lotz, R. & Blignaut, J.N. 2012. Estimating the opportunity cost of water for the Kusile and Medupi
coal-fired electricity power plants in South Africa. Journal of Energy in Southern Africa, 23(4):74-84.
Integrated Resource Plan (IRP). 2011. Integrated resource plan for electricity 2010-2030. Revision 2, final
report. Available at: http://www.energy.gov.za/IRP/irp%20files/IRP2010_2030_Final_Report_20110325.p
df (accessed on 13 January 2012).
Intergovernmental Panel on Climate Change (IPCC). 1995. Summary for policy-makers: second assessment
report: working group III draft report. Available from: http://www.ipcc.ch/pdf/climate-changes-1995/ipcc2nd-assessment/2nd-assessmenten.pdf (accessed: 2011-03-10).
Intergovernmental Panel on Climate Change (IPCC). 2000. Sectoral economic costs and benefits of GHG
mitigation. Geneva: IPCC.
International Energy Agency Greenhouse Gas R&D Programme (IEA GHG). 2004. Improvements in power
generation with post-combustion capture of CO2. Report PH4/33, IEA Greenhouse Gas R&D Programme,
Cheltenham, UK. Available at: http://ccs101.ca/assets/Documents/iea-improvement-in-powe (accessed on
11 June 2011).
International Energy Agency. 2007. Tracking industrial energy efficiency and CO2 emissions: In support of
the G8 plan of action, Volume 1. Paris, France: OECD Publishing.
- 257 -
International Energy Agency. 2010. Projected costs of generating electricity. 2010 Edition. Available at:
http://www.iea.org.za (accessed on 29 July 2011).
International Panel on Climate Change (IPCC). 1996. IPCC guidelines for national greenhouse gas
inventories: reference manual, Volume 3. United Nations Environment Program, the Organisation for the
Co-operation and Development, the International Energy Agency and the Intergovernmental Panel on
Climate Change. Cambridge: Cambridge University Press.
International Panel on Climate Change (IPCC). 2001. Climate change 2001. In: R.T. Watson & Core Writing
Team (eds.). Third assessment report of the International Panel on Climate Change. Geneva: Cambridge
University Press.
International Panel on Climate Change (IPCC). 2007a. Climate change 2007: synthesis report. Contribution
of work groups I, II, and III to the fourth assessment report of the Intergovernmental Panel on Climate
Change. IPCC. Geneva, Switzerland.
International Panel on Climate Change (IPCC). 2007b. 2006 IPCC guidelines for national greenhouse gas
inventories. Volume 3-Industrial processes and product use. Available at: http://www.ipcc.ch/ (accessed
on 13 January 2011).
International Panel on Climate Change (IPCC). 2007c. 2006 IPCC guidelines for national greenhouse gas
inventories. Volume 2-Energy. Available at: http://www.ipcc.ch/ (accessed on 13 January 2011).
Jackson, M. 1990. Beyond a system of systems methodologies. Journal of the Operational Research Society,
41(8):657–668.
Jackson, M. 1993. Social theory and operational research practice. Journal of the Operational Research
Society, 44:563–577.
Jackson, M. 2003. Systems thinking: creative holism for managers. Wiley, Chichester.
Jalava, K., Pasanen, S., Saalasti, M. & Kuitunen. M. 2010. Quality of environmental impact assessment:
Finnish EISs and the opinions of EIA professionals. Impact Assess Project Appraisal, 28(1):15–27.
- 258 -
Jay, S., Jones, C., Slinn, P. & Wood, C. 2007. Environmental Impact Assessment: Retrospect and Prospect.
Environ Impact Assess Rev, 27(4):287–300.
Jegarl, S., Baek, J., Jang, D. & Ryu, C. 2009. Impact assessment of CO2 mitigation options in Korea using
energy system analysis model. Energy, 1:3747–3754.
Jensen, S.G. & Meibom, P. 2008. Investments in liberalised power markets: gas turbine investment
opportunities in the Nordic power system. International Journal of Electrical Power & Energy Systems,
30(2):113-124.
Jeong, S., Kim, K., Park, J., Lim, D. & Lee, S. 2008. Economic comparison between coal-fired and liquefied
natural gas combined cycle power plants considering carbon tax: Korean case. Energy, 33:1320-1330.
Johnson, S. T., Taylor, T. & Ford, D. 2006. Using system dynamics to extend real options use: insights from
oil and gas industry. Conference of the System Dynamics Society. Available at: http://www.system
dynamics.org/conferences/2006/proceed/papers/FORD340.pdf (accessed 13 July 2010).
Jorgensen, A.A. 2010. Transport costs and the relevance of externalities. Available at: http://www.rra.co
.za/?p=16756 (accessed on 25 June 2011).
Joshi, S. 2000. Product environmental life-cycle assessment using input-output techniques. Journal of
Industrial Ecology, 3 (2 & 3):95-120.
Joskow, P.L. 1992. Weighing environmental externalities: Let’s do it right. Electricity Journal, 5:53-67.
Kannan, R., Leong, K.C., Osman, R. & Ho, H.K. 2007. Life cycle energy, emissions and cost inventory of
power generation technologies in Singapore. Renewable and Sustainable Energy Reviews, 11:702–715.
Karlsson, K. & Meibom, P. 2008. Optimal investment paths for future renewable based energy systems:
using the optimisation model Balmorel. International Journal of Hydrogen Energy, 33(7):1777-1787.
Karmis, M. 2005. A study of increased use of renewable energy resources in Virginia. Virginia Center for
Coal and Energy Research, Blacksburg, Virginia. Available at: http://www.energy.vt.edu/Publications/Incr_U
se_Renew_Energy_VA_rev1.pdf (accessed on 11 July 2013).
- 259 -
Kelman, S. 2008. Discussion paper on procurement issues in South Africa that affect growth and
development. Available at: http://www.cid.harvard.edu/southafrica (accessed on 11 July 2013).
Kidd, M. & Retief, F. 2009. Environmental Assessment. In: Strydom H, King N, editors. Fuggle and Rabie's
Environmental Management in South Africa. Cape Town: Juta Publishing.
Kim, S. 2003. Research paradigms in organizational learning and performance: compiting modes of inquiry.
Information Technology, Learning and Performance Journal, 21(1):9-18.
Kim, S. 2007. Evaluation of negative environmental impacts of electricity generation: neoclassical and
institutional approaches. Energy Policy, 35:413-423.
Kincheloe, J. L. & McLaren, P. L. 1994. Rethinking critical theory and qualitative research. In Denzin, N.K. &
Lincoln, Y.S. (eds.), Handbook of qualitative research. Thousand Oaks, CA: Sage.
Kincheloe, J. L. & McLaren, P. L. 2000. Rethinking critical theory and qualitative research. In Denzin, N.K. &
Lincoln, Y.S. (eds.), Handbook of qualitative research (2nd ed.). Thousand Oaks, CA: Sage.
Kiviluoma, J. & Meibom, P. 2009. Influence of wind power, plug-in electric vehicles and heat storages on
power system investments. Energy - the International Journal, xx:xx-xx.
Kochi, I., Donovan, G. H., Champ, P. A. & Loomis, J.B. 2010. The economic cost of adverse health effects
from wildfire-smoke exposure: a review. International Journal of Wildland Fire, 19(7):803-817.
Kopp, P. & Prud’homme, R. 2007. The internalisation of external costs i n the transportation system.
Discussion paper - 12th ACEA SAG meeting –December 2007. Available at: http://www.acea.be/uploads
/publications/SAG_12_ExternalCostsInternalisation.pdf (accessed on 13 August 2014).
Kornbluth, K., Greenwood, J., Jordan, E., McCaffrey, Z. & Erickson, P.A. 2012. Economic feasibility of
hydrogen enrichment for reducing NOx emissions from landfill gas power generation alternatives: a
comparison of the levelized cost of electricity with present strategies. Energy Policy, 41:333–339.
- 260 -
Kouvaritakis, N., Soria, A. & Isoard, S. 2000. Endogenous learning in world post-kyoto scenarios: application
of the POLES model under adaptive expectations. International Journal of Global Energy Issues, 14(1–
4):222–248.
Kruger, E. & Chapman, O.A. 2005. Quality aspects of Environmental Impact Assessment Reports in the Free
State Province, South Africa. South African Geographical Journal, 87: 52-57.
Kruopienė J, Zidoniene S, Dvorioniene J.2009. Environmental Impact Assessment in Lithuania. Environ
Impact Assess Rev, 29(5):305–9.
Kruopienė, J., Zidoniene, S. & Dvorioniene, J. 2009. Environmental Impact Assessment in Lithuania. Environ
Impact Assess Rev, 29(5):305–9.
Krupnick, A., Cropper, M., Alberini, A, Simon, N., O’Brien, B. & Goeree, R. 2002. Age, health and the
willingness to pay for mortality risk reductions: a contingent valuation survey of Ontario residents. Journal
of Risk and Uncertainty, 24(22): xx-xx.
Kucukaksoy, I. 2011. Adam Smith’s conceptual contributions to international economics: based on the
Wealth of Nations. Business and Economic Horizons, 4(1):108-119.
Kuhn, Thomas S. 1970. The structure of scientific revolutions. 2nd ed. Chicago: University of Chicago Press.
Lafarge, J.D. 2011. Implications on recycling activities in the cement and concrete sector. Available at:
https://docs.google.com/viewer?a=v&q=cache:FqLHVfr9zUAJ:www.lebensministerium.at/dms/lmat/umwe
lt/abfall-ressourcen/workshop251011/daul/Implications%2520on%2520Recycling%2520Activities%2520in
%2520the%2520cement%2520and%2520concrete%2520sector.pdf+&hl=en&gl=za&pid=bl&srcid=ADGEESg
jCGAqAKEIwA03LMbgJ83V7aoe_B3NxLw-Krkx7tnzRef89hgdNuaPT24uBlTjusl2yoqbbhx0xGu8t0QK7DhSjDLn
rKp4Q928yD9AnuYbg3RA9ySt97t8eSvj5T0RmxnYxXPt&sig=AHIEtbQkk97UjUxlVRsfa0-35U_3HQFdxQ
(acessed on 29 January 2013).
LaHaye, L. 2008. The concise encyclopedia of economics: Mercantilism. Available at: http://www.econli
b.org/library/Enc/Mercantilism.html (accessed on 18 May 2013).
- 261 -
Lambert, T., Gilman, P. & Lilienthal P. 2006. Micropower system modelling with HOMER. In: Farret FA,
Simões MG, (eds.), Integration of alternative sources of energy. Wiley-IEEE Press.
Lammers, P.J. 2009. Energy and nutrient cycling in pig production systems. Unpublished Doctoral Thesis.
Iowa: Iowa State University. Available at: http://lib.dr.iastate.edu (accessed on 6 February 2013).
Lana. 2010. Electricity price hike in South Africa. Available at: http://www.southafricaweb.co.za/article/elec
tricity-price-hike-south-africa (accessed on 16 January 2013).
Landefeld, J.S. & Seskin, E.P. 1982. The economic value of life: linking theory to practice. AJPH, 72(6):555566.
Lane, D. & Oliva, R. 1998. The greater whole: towards a synthesis of system dynamics and soft systems
methodology. European Journal of Operational Research, 107(1): 214–235.
Lane, D.C. 1994. With a little help from our friends - how system dynamics and soft OR can learn from each
other. System Dynamics Review, 10(2-3):101–134.
Lane, D.C. 2000. Should system dynamics be described as a 'hard' or 'deterministic' systems approach?
Systems Research and Behavioural Science, 17:3–22.
Lane, D.C. 2001. Rerum cognoscere causas: Part I – How do the ideas of system dynamics relate to
traditional social theories and the voluntarism/determinism debate? System Dynamics Review, 17(2):97–
118.
Laszlo, A. & Krippner, S. 1998. Systems Theories: Their Origins, Foundations, and Development. In J.S.
Jordan (ed.), Systems theories and a priori aspects of perception. Amsterdam: Elsevier.
Lawson, T. 2013. What is this ‘school’ called neoclassical economics? Cambridge Journal of Economics, 37:
947–983.
Lee, N. 2000. Reviewing the quality of environmental assessments. In: Lee N, George C, editors.
Environmental assessment in developing and transitional countries. Chichester: Wiley & Sons.
- 262 -
Leeth, J.D. & Ruser, J. 2003. Compensating wage differentials for fatal and non-fatal injury risk by race and
gender. Journal of Risk and Uncertainty, 27(3): xx-xx.
Lenzen, M. & Murray, S. 2001. A modified ecological footprint method and its application to Australia.
Ecological Economics, 37(2):229-255.
Lenzen, M. 2001. Errors in conventional and input-output-based life-cycle inventories. Journal of Industrial
Ecology, 4(4):127-148.
Leontief, W. 1986. Input-output economics. 2nd ed. New York: Oxford University Press.
Li, F., Dong, S., Li, Z., Li, Y., Li, S. & Wan, Y. 2012. The improvement of CO2 emission reduction policies based
on system dynamics method in traditional industrial region with large CO2 emission. Energy Policy, 51:683–
695.
Lindboe, H.H., Werling, J., Kofoed-Wiuff, A. & Bregnbaek, L. 2007. Impact of CO2 quota allocation to new
entrants in the electricity market. Ea Energy Analyses, Energy Modelling; 2007. Available at: http://www2.
mst.dk/common/Udgivramme/Frame.asp?http://www2.mst.dk/Udgiv/publications/2007/978-87-7052-494
-0/html/default_eng.htm (accessed on 23 March 2011).
Liu, W., Lund, H. & Mathiesen, B.V. 2011. Large-scale integration of wind power into the existing Chinese
energy system. Energy, 36:4753-4760.
Lloyd, P. & Cook, A. 2005. Methane release from South African coalmines. The Journal of The South African
Institute of Mining and Metallurgy, 105:483–490.
Lluch, E. & Argemi, L. 1994. Physiocracy in Spain. History of Political Economy, 26(4):613-627.
Loschel, A. 2002. Technological change in economic models of environmental policy: a survey. Ecological
Economics, 43:105–126.
Loucks, D.L., Van Beek, E., Stedinger, J.R., Dijkman, J.P.M. & Villars, M.T. 2005. Water resources systems
planning and management: An introduction to methods, models and applications. Available at: http://ecom
mons.cornell.edu/bitstream/1813/2804/21/00_intro.pdf. (accessed on 7 June 2013).
- 263 -
Loulou, R., Goldstein, G. & Noble, K. 2004. Documentation for the MARKAL family of models. Available at:
http://www.iea-etsap.org/web/MrklDoc-II_MARKALMACRO.pdf (accessed on 4 January 2013).
Lund, H. & Mathiesen, B.V. 2009. Energy system analysis of 100% renewable energy systems—the case of
Denmark in years 2030 and 2050. Energy, 34:524–531.
Lund, H. 2006. Large-scale integration of optimal combinations of PV, wind and wave power into the
electricity supply. Renewable Energy, 31:503–515.
Lund, H. 2007. Renewable energy strategies for sustainable development. Energy, 32(6):912–919.
Lund, H., Duic, N., Krajacic, G. & Carvalho, M.G. 2007. Two energy system analysis models: a comparison of
methodologies and results. Energy, 32:948–954.
Ma, S. & Jin, X. 2010. Simulation on the Emission of SO2 and NOx from Coal-Fired Power Plants using MM5SMOKE-CMAQ. North China Electric Power University, Baoding, China.
Magnusson, 2003. Mercantilism. In Samuels, W.J., Biddle, J.E. & Davis, J.B. (eds.) A companion to the history
of economic thought. Malden, UK: Blackwell.
Mallah, S. & Bansal, N.K. 2010. Nuclear and clean coal technology options for sustainable development in
India. Energy, 35:3031–3039.
Manis, J. 2005. An inquiry into the nature and causes of the wealth of nations by Adam Smith: an electronic
classics series publication. Available at: http://www2.hn.psu.edu/faculty/jmanis/adam-smith/wealthnations.pdf (accessed on 13 August 2013).
Mankiw, N.G. 2012. Principles of economics. Sixth edition. Fort Worth, TX: The Dryden Press.
Mao, G., Dai, X.,Wang, Y., Guo, J., Cheng, X., Fang, D. et al. 2013. Reducing carbon emissions in China:
industrial structural upgrade based on system dynamics. Energy Strategy Reviews 2:199-204.
Marshall, A. 1890. The principles of economics. London: Macmillan.
- 264 -
Marx, K. 1952. A history of economic theories, ed. Karl Kautsky and trans. Terence McCarthy. New York: The
Langland Press.
Massachusetts Institute of Technology (MIT). 2007. The future of coal: options for a carbon constrained
world. Available at: http://web.mit.edu/coal/The_Future_of_Coal.pdf (accessed on 27 June 2013).
Mbhele, P. 2009. The quality of EIA reports for housing developments in the Nkangala district of the
Mpumalanga province, South Africa. Unpublished Master’s Dissertation. Potchefstroom: North-West
University.
McGee, R.W. 1989. The economic thought of David Hume. Hume Studies, 15(1):184-204.
Meek, Ronald L. 1973. Studies in the labour theory of value. 2nd ed. London: Laurence and Wishart.
Merriam, S.B. 2002. Qualitative research in practice: Examples for discussion and analysis. San Francisco:
Jossey-Bass.
Mertens, D. 2002. Handbook of Mixed Methods in Social and Behavioural Research, Chapter Mixed
Methods and the Politics of Human Research: The Transformational and Emancipatory Perspective, pp.
135–164. Thousand Oaks: SAGE Publications.
Messner, S. 1997. Endogenized Technological Learning in an Energy Systems Model. Journal of Evolutionary
Economics, 7:291–313.
Millennium Institute. 2010. T21 integrated development model. Available at: http://www.millenniuminstitute.org (accessed on 3 November 2013).
Mineweb. 2013. China's looming low-grade coal ban boon for Australia, SA coal. Available at:
http://www.mineweb.com/mineweb/content/en/mineweb-energy?oid=194470&sn=Detail (accessed on 27
June 2013).
Ministry of Environment and Tourism. 1997. Environmental Impact Assessment Guidelines. Harare:
Government Printers.
- 265 -
Ministry of Environment and Tourism. 1997. Environmental Impact Assessment Guidelines. Harare:
Government Printers.
Mirowski, P. 1987. The philosophical bases of institutional economics. Journal of Economic Issues,
XXI(3):1001-1038.
Mishan, E.J. 1960. A survey of welfare economics, 1939-1959. Economic Journal, 70:197-256.
Mishan, E.J. 1969. The relationship between joint products, collective goods and external effects. Journal of
Political Economy, 77(3):329-348.
Mishra, S.K. 2009. Estimation of externality costs of electricity generation from coal: an OH-MARKAL
extension dissertation. Unpublished Doctoral Thesis. Canada: The Ohio State University. Available at:
http://etd.ohiolink.edu/view.cgi/Khadka%20Mishra%20Shruti.pdf?%20osu12597033 37 (accessed on 8 July
2011).
Mokheseng, B. 2010. An investigation into life-cycle costing as a comparative analysis approach of energy
systems. Available at: http://www.conference.csir.co.za (accessed on 26 April 2013).
Moolman, C., Blignaut, J. & Van Eyden, R. 2006. Modelling the marginal revenue of water in selected
agricultural commodities: a panel data approach. Agrekon, 45(1): 78–88.
Moore, M. 1999. Estimating irrigators' ability to pay for reclamation water. Land economics, 75(4): 652–
578.
Morthorst, P.E., Jensen, S.G. & Meibom, P. 2005. Investment and pricing in a liberalised electricity market.
Riso National Laboratory. Available at: http://www.risoe.dk/rispubl/SYS/syspdf/ris-r-1519.pdf (accessed on
12 March 2013).
Mrozek, J.R. & Taylor, L.O. 2002. What determines the value of life? A meta-analysis. Journal of Policy
Analysis and Management, 21(2): 253–270.
- 266 -
Munnick, V., Hochmann, G. & Hlabane, M. 2009. The social and environmental consequences of coal
mining. South African case study: Final draft. Available at: http://www.bothends.org/uploaded_files/2ca
se_study_South_Africa.pdf (accessed on 13 June 2011).
Munster, M. 2009. Energy system analysis of waste-to-energy technologies. Unpublished Doctoral Thesis.
Denmark: Aalborg University.
Murphy, J., Allen, P.G., Stevens, T.H. & Weatherhead, D. 2004. A meta-analysis of hypothetical bias in
stated preference valuation. Environmental and Resource Economics, 30(3): 313–325.
Musango, J.K., Brent, A.C. & Tshangela, M. (in press). Green economy transitioning of the South African
power sector: A system dynamics analysis approach. Development Southern Africa.
Musango, J.K., Brent, A.C., Amigun, B., Pretorius, L., & Müller, H. 2012. A system dynamics approach to
technology sustainability assessment: The case of biodiesel developments in South Africa. Technovation,
32, 6398651.
Naicker, K., Cukrowska, E. & McCarthy, T.S. 2003. Acid mine drainage from gold mining activities in
Johannesburg, South Africa and environs. Environmental Pollution, 122: 29–40.
National Energy Regulator of South Africa (NERSA). 2013. In the matter regarding revenue application Multi year price determination 2013/14 to 2017/18 (MYPD3) by Eskom Holdings Soc Limited (‘Eskom’).
Available at: http://www.eskom.co.za/content/NersaReasonsforDecision.pdf (accessed on 3 July 2013).
National Pollutant Inventory (NPI). 2012. National pollutant inventory emission estimation technique
manual for fossil fuel electric power generation version 3.0. Available at: http://www.npi.gov.au (accessed
on 23 January 2013).
National Renewable Energy Labouratory. 2008. HOMER: The optimization model for distributed power.
Available at: http:// http://www.nrel.gov (accessed on 9 September 2013).
National Treasury. 2004. National Treasury PPP manual, module 4: PPP feasibility study. Available at:
http://www.dwaf.gov.za (accessed on 11 July 2013).
- 267 -
National Treasury. 2013. Carbon tax policy paper: Reducing greenhouse gas emissions and facilitating the
transition to a green economy. Available at: http://www.treasury.gov.za (accessed on 10 December 2013).
Naughten, B., 2003. Economic assessment of combined cycle gas turbines in Australia: some effects of
micro-economic reform and technological change. Energy Policy, 31(3):225-245.
Navrud, S & Ready, R. 2007: Environmental value transfer: issues and methods. Dordrect, The Netherlands:
Springer.
Navrud, S. 2004. Value transfer and environmental policy. Chapter 5 (pp. 189-217) in Tietenberg, T. and H.
Folmer (eds.) 2004: The International Yearbook of Environmental and Resource Economics 2004/2005. A
survey of Current Issues. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing.
NEEDS. 2008. External costs per unit of emission. Deliverable No. 1.1 - RS 3a, NEEDS integrated project,
Priority 6.1, sub-priority 6.1.3.2.5: Excel sheets, an appendix to New Energy Externalities Developments for
Sustainability (NEEDS) 2008 Stuttgart: IER, University of Stuttgart, 2008.
New Energy Externalities Developments for Sustainability (NEEDS). 2007. Final report on the monetary
valuation of mortality and morbidity risks from air pollution. Deliverable for WP6 of RS1b of the New
Energy Externalities Developments for Sustainability (NEEDS) project. IER, University of Stuttgart.
New Energy Externalities Developments for Sustainability (NEEDS). 2009. Sixth framework programme.
Integrated project - Priority 6.1: Sustainable Energy Systems and, more specifically, Sub-priority 6.1.3.2.5:
Socio-economic tools and concepts for energy strategy. Deliverable n° 2.1 - RS 3a - Value transfer techniques
and expected uncertainties. Project no: 502687. Available at: http://www.needs-project.org/do
cs/Needs_Rs3a_D2.1.doc (accessed on 6 July 2012).
NewExt. 2004. New Elements for the Assessment of External Costs from Energy Technologies”. Funded
under the EC 5th Framework Programme (1998 – 2002), Thematic programme: energy, environment and
sustainable development, Part B: energy; generic activities: 8.1.3. Externalities ENG1-CT2000-00129.
Newman, P. C. 1962. The development of economic thought. New York: Prentice-Hall.
- 268 -
Nhamo, G. & Inyang, E. 2011. Framework and tools for environmental management in Africa. Available at:
http://www.codesria.org (accessed on 4 May 2013).
NINHAM SHAND. 2007. Environmental impact assessment process: proposed coal-fired power station and
associated infrastructure in the Witbank area. Final environmental impact report. Sandton: Eskom.
Njobeni, S. 2010. Eskom cleans up Medupi emissions. Available at: http://allafrica.com/stories/20
1001060259.html (accessed on 21 June 2013).
Nkambule, N. & Blignaut, J.N. 2012. The external costs of coal mining: the case of collieries supplying Kusile
power station. Journal of Energy in Southern Africa, 23(4):85-93.
Nordhaus, W.D. 1993. Reflections on the economics of climate change policy. Journal of economic
perspectives, 7(4)11-25. [Online] Available from: http://www.jstor.org/pss/ 2138498 (accessed on 15
August 2011).
NZEC. 2009. Carbon dioxide capture from coal-fired power plants in china. Summary report for NZEC work
package 3 September, 2009. Available at: http://www.nzec.info/en/assets/Reports/Techno-economic-Co
mparison-WP3-Final-English.pdf (accessed on 12 January 2013).
Oak Ridge National Labouratory and Resources for the Future (ORNL & RfF). 1994. External costs and
benefits of fuel cycles. A study for the US Department of Energy and the Commission of the European
Communities. Report No. 1-8. Washington: McGraw-Hill/Data Institute.
Ochoa, P. & Van Ackere, A. 2009. Policy changes and the dynamics of capacity expansion in the Swiss
electricity markets. Energy Policy, 37:1983–1998.
Odeh, N.A. & Cockeril, T.T. 2008. Life cycle analysis of UK coal fired power plants. Energy Conversion and
Management, 49:212–220.
Office of Technology Assessment. 1994. Studies of the environmental costs of electricity. OTA-ETI-134.
Washington: US Government Printing Office.
- 269 -
Østergaard, P.A. 2009. Reviewing optimisation criteria for energy systems analyses of renewable energy
integration. Energy, 34:1236–1245.
Ottinger, R.L., Wooley, D.R., Robinsson, N.A., Hodas, D.R. & Babb, S.E. 1991. Environmental costs of
electricity. New York: Oceana Publications.
Owen, A.D. 2004. Environmental externalities, market distortions and the economics of renewable energy
technologies. The Energy Journal, 25(3):127-156.
Palladian Publications. 2013. South Africa’s cement industry shaping up. Available at: http://www.world
cement.com/news/cement/articles/South_Africa%E2%80%99s_cement_industry_is_shaping_up.aspx (accessed on 29 January 2013).
Palm, E. & Hansson, S.O. 2006. The case for ethical Technology Assessment (eTA). Technological
Forecasting and Social Change, 73:543-558.
Paltsev, S., Jacoby, H.D., Reilly, J.M., Ejaz, Q.J., Morris, J., O’Sullivan, F., Rausch, S., Winchester, N. & Kragha,
O. 2011.The future of US. natural gas production, use, and trade. Energy Policy, 39:5309–5321.
Panagopoulos, T. 2009. Linking forestry, sustainability, and aesthetics. Ecological Economics, 68:2485-2489.
Papatheodorou, Y.E. 1990. Energy in the Greek economy: A total energy approach at the macro level.
Energy Economics, 12:377-395.
Park, K., Shin, D. & Yoon, E.S. 2011. The cost of energy analysis and energy planning for emerging fossil fuel
power plants based on the climate change scenarios. Energy, 36:3606-3612.
Pasaoglu, G. 2006. A simulation model for the long-term analysis of decentralized electricity market.
Bogazici University.
Patton, M. 1990. Qualitative evaluation and research methods. Newbury Park: Sage.
- 270 -
Pauw, K. 2007. Economy-wide modelling: an input into the Long Term Mitigation Scenarios process, LTMS
input report 4. Energy Research Centre, Cape Town. Available at: http://www.erc.uct.ac.za/Researc
h/publications/07-Pauw-LTMSEconomy-wide_modelling.pdf (accessed on 6 May 2011).
Pearce, D. 1995. Costing the environmental damage from energy production. In: Rooke, D. (ed.), Energy for
the future. London: E & FN Spon.
Pearce, D. 2003. Conceptual framework for analyzing the distributive impacts of environmental policies.
Available at: http://www.ucl.ac.uk/~uctpa36/oecd%20distribution.pdf (accessed on 17 August 2014).
Pearce, D., Bann, C. & Georgiou, S. 1992. The social cost of fuel cycles. Report to the UK Department of
Trade and Industry, HMSO, London.
Pearce, D.W. & Turner, R.K. 1990. Economics of natural resources and the environment. London: Harvester
Wheatshef.
Perdicoúlis, A. & Glasson, J. 2006. Causal networks in EIA. Environmental Impact Assessment Review,
26:553–569.
Pereira, A.J.C. & Saraiva, J.T. 2010. Building generation expansion plans - A decision aid approach to use in
competitive electricity markets. Power generation, transmission, distribution and energy conversion. 7th
Mediterranean Conference and Exhibition.
Pereira, A.J.C. & Saraiva, J.T. 2011. Generation expansion planning (GEP) – A long-term approach using
system dynamics and genetic algorithms (GAs). Energy, 36(8):5180–5199.
Perman, R., Ma, Y., McGilvray J. & Common, M. 1999. Natural resource and environmental economics. 2nd
ed. London: Longman.
Peterson K. 2010. Quality of environmental impact statements and variability of scrutiny by reviewers.
Environ Impact Assess Rev, 30:169–76.
Phillips, A. 1995. The tableau Économique as a simple Leontief model. The Quarterly Journal of Economics,
69(1):137-144.
- 271 -
Pigou, A.C. 1920. The economics of welfare. Available at: http://www.econlib.org/library/NPDBooks/Pigo
u/pgEW3.html (accessed on 16 March 2011).
Pinheiro, H.J., Boshoff, H.P. & Pretorius, C.C. 1997. Analysis of coal product samples of producing South
African collieries. Pretoria: CSIR.
Pinho, P., Maia, R. & Monterrosa, A. 2007. The quality of Portuguese Environmental Impact Studies: the
case of small hydropower projects. Environ Impact Assess Rev, 26(3):189–205.
Plack, M.M. 2005. Human nature and research paradigms: theory meets physical therapy practice. The
Qualitative Report, 10(2):223-245. Available at: http: //www.nova.edu/ssss/QR/QR10-2/plack.pdf (accessed
on 30 May 2011).
Pojer, S.M. n.d. From mercantilism to Adam Smith: the evolution of the modern capitalist system. Available
at: https://www.google.co.za/#q=From+mercantilism+to+Adam+Smith%3A+the+ecolution+of +the+moder
n +capitalist+system (accessed on 8 November 2011).
Polonen, I., Hokkanen, P. & Jalava, K. 2010. The effectiveness van the Finnish EIA system – what works,
what doesn't and what could be improved? Environ Impact Assess Rev, 31(2):120–8.
Ponterotto, J.G. & Grieger, I. 2007. Effectively communicating qualitative research. Journal of Counseling
Psychology, 35(3):404-430.
Ponterotto, J.G. 2005. Qualitative research in counseling psychology: a primer on research paradigms and
philosophy of science. Journal of Counseling Psychology, 52(2):126-136.
Pouris, A. 1987. The cost of photovoltaic cells in South Africa. Energy, 12:117-121.
Pretorius, K. 2009. Coal mining and combustion: internalising the cost for a fair climate change debate.
Federation for a Sustainable Environment.
Pruyt, E. 2006. What is system dynamics? A paradigmatic inquiry. Available at: http://www.system
dynamics.org/conferences/2006/proceed/papers/PRUYT177.pdf (accessed on 2 September 2011).
- 272 -
Pruyt, E. 2007. The EU-25 power sector: A system dynamics model of competing electricity generation
technologies. Available at: http://www.systemdynamics.org/conferences/2007/proceed/papers/PRUYT24
6.pdf (accessed on 11 May 2011).
Pruyt, E. 2013. Small system system dynamics models for big issues: triple jump towards real-world
dynamic complexity. Available at: http://www.slideshare.net/DelftOpenEr/system-dynamics-37532848
(accessed on 17 August 2014).
Pulles, W., Boer, R. & Nel, S. 2001. A generic water balance for the South African coal mining industry.
Water Research Commission Report No. 801/1/01.
Qudrat-Ullah, H. & Davidsen, P.I. 2001. Understanding the dynamics of electricity supply, resources and
pollution: Pakistan’s case. Energy, 26:595–606.
Qudrat-Ullah, H. 2012. On the validation of system dynamics type simulation models. Telecommun Syst,
51:159–166.
Qudrat-Ullah, H. 2013. Understanding the dynamics of electricity generation capacity in Canada: a system
dynamics approach. Energy, 59:285-294.
Rabl, A. 2006. Analysis of air pollution mortality in terms of life expectancy changes: relation between time
series, intervention and cohort studies. Environmental health: a global access science source 2006, 5:1.
Available at: http://web.me.com/arirabl/Site/Publications_files/ Rabl06%20LE%20framework.pdf (accessed
on 14 August 2011).
Rafaj, P. & Kypreos, S. 2007. Internalisation of external cost in the power generation sector: analysis with
global multi-regional MARKAL model. Energy Policy, 35:828-843.
Raico, R. 1995. The Australian school and classical liberalism. Available at: http://mises.org/ete
xts/austrianliberalism.pdf (accessed on 14 August 2011).
Rehman, S., El-Amin, I.M., Ahmad, F., Shaahid, S.M., Al-Shehri, A.M. & Bakhashwain, J.M. 2007. Feasibility
study of hybrid retrofits to an isolated off-grid diesel power plant. Renew Sustain Energy Rev, 11(4):635–53.
- 273 -
Rendon-Sagardi, M.A., Sanchez-Ramirez, C., Cortes-Robles, G., Alor-Hernandez, G. & Cedillo-Campos, M.G.
2014. Dynamic analysis of feasibility in ethanol supply chain for biofuel production in Mexico. Applied
Energy, 123:358–367.
Republic of South Africa. 1998. National Environmental Management Act 107 of 1998. Government
Gazette, No. 19519.
Republic of South Africa. 2006. Regulations in terms of the National Environmental Management Act, 1998,
No.R. 385. Department of Environmental Affairs and Tourism (DEAT).
ESTScreen International. 2012. RETScreen clean energy project analysis software. Available at:
http://en.openei.org/wiki/RETScreen_International_Clean_Energy_Decision_Support_Centre (accessed on
16 May 2013).
Retief, F. & Chabalala. B. 2009. The cost of environmental Impact Assessment (EIA) in South Africa. J
Environ Assess Policy Manag, 11(1):51–68.
Richardson, G. P. & Anderson, D. F. 1980. Toward a pedagogy of system dynamics. System Dynamics, TIMS
Studies in the Management Sciences, 14:91-106.
Richardson, G. P. & Pugh, A. L. 1981. Introduction to system dynamics modelling with DYNAMO. Cambridge,
Productivity Press.
Riekert, J. & Koch, S.F. 2012. Projecting the external health costs of a coal-fired power plant: the case of
Kusile. Journal of Energy in Southern Africa, 23(4):52-66.
Robalino-Lopez, A., Mena-Nieto, A. & Garcia-Ramos, J.E. 2014. System dynamics modeling for renewable
energy and CO2 emissions: a case study of Ecuador. Energy for Sustainable Development, 20:11–20.
Roberts, E.B. 1978. System dynamics – an introduction. In: Roberts EB (ed.) Managerial applications of
system dynamics. Waltham: Pegasus Communications Inc.
- 274 -
Roberts, N. H., Anderson, D. F., Deal, R. M., Grant, M. S. & Shaffer, W. A. 1983. Introduction to computer
simulation: the system dynamics modelling approach. Reading, MA: Addison-Wesley.
Robertshaw, J.E., Mecca, S.J. & Rerick, M.N. 1978. Problem solving: a systems approach. New York:
Petrocelli Books Inc.
Rommelse, G. 2010. The role of mercantilism in Anglo-Dutch political relations, 1650–74. Economic History
Review, 63(3):591–611.
Rosenberg, N. 1960. Some institutional aspects of the wealth of nations. Journal of Political Economy,
68(6):557-570.
Roth, I.F. & Ambs, L.L. 2004. Incorporating externalities into a full cost approach to electric power
generation life-cycle costing. Energy, 29:2125-2144.
Rothbard, M.N. 2010. Mercantilism: a lesson for our times? Available at http://mises.org/daily/4304
(accessed 2 July 2013).
Rowe, R.D., Lang, C.M., Chestnut, L.G., Latimer, D.A., Rae, D.A., Bernow, S.M. & White, D.E. 1995. New York
state environmental externalities cost study. New York: Oceana Publications.
Rozylow, M. 2013. A cost-benefit analysis of the first nuclear power plant in Poland. Unpublished Master’s
Dissertation. Poland: Aarhus University.
Rubin, E.S., Rao, A.B. & Chen, C. 2005. Comparative assessments of fossil fuel power plants with CO2
capture and storage. In: Rubin, E.S., Keith, D.W. & Gilboy, C.F. (eds.), Proceedings of the 7th International
Conference on Greenhouse Gas Control Technologies (GHGT-7), Vancouver, Canada, September 5–9, 2004.
Volume 1: Peer-reviewed papers and overviews, Elsevier, 2005 (in press).
Sagar, A. D. & Holdren, J. P. 2002. Assessing the global energy innovation system: some key issues. Energy
Policy, 30:465-469.
Sahu, N.C. & Nayak, B. 1994. Niche Diversification in Environmental/Ecological Economics. Ecological
Economics, 11:9-19.
- 275 -
Saltelli, A., Chan, K. & Scott, E.M. 2000. Sensitivity analysis. Chichester: John Wiley & Sons.
Salvia, M., Cosmi, C., Macchiato, M. & Mangiamele, L. 2002. Waste management system optimisation for
Southern Italy with MARKAL model. Resources, Conservation and Recycling, 34(2):91–106.
Sanchez, J., Barquin, J., Centeno, E. & Lopez-Pea, A. 2008. A multidisciplinary approach to model long-term
investments in electricity generation: combining system dynamics, credit risk theory and game theory.
Available at: http://www.researchgate.net/publication/224325392_A_multidisciplinary_approach_to_mod
el_long-term_investments_in_electricity_generation_Combining_System_Dynamics_credit_risk_theory_an
d_game_theory/links/02e7e526b4759e9674000000 (accessed on 11 September 2014).
Sandham, L.A., Carroll, T.H. & Retief, F.P. 2010. The contribution of Environmental Impact Assessment (EIA)
to decision making for biological pest control in South Africa – the case of Lantana camara. Biol Control,
55(2):141–9.
Sandham, L.A., Siphugu, M.V. & Tshivhandekano, T.R. 2005. Aspects of Environmental Impact Assessment
(EIA) practice in the Limpopo Province – South Africa. AJEAM-RAGEE, 10:50–65.
Sandham, L.A. & Pretorius, H.M. 2008. A review of EIA report quality in the North West Province of South
Africa. Environ Impact Assess Rev, 28(4–5):229–40.
Sandham, L.A., Carroll, T.H. & Retief, F.P. 2010. The contribution of Environmental Impact Assessment (EIA)
to decision making for biological pest control in South Africa – the case of Lantana camara. Biol Control,
55(2):141–9.
Sandham, L.A., Hoffmann, A.R. & Retief, F.P. 2008. Reflections on the quality of mining EIA reports in South
Africa. J South Afr Inst Min Metall, 108:701–6.
Sandham, L.A., van Heerden, A.J., Jones, C.E., Retief, F.P. & Morrison-Saunders, A.N. 2013. Does enhanced
regulation improve EIA report quality? Lessons from South Africa. Environmental Impact Assessment
Review, 38:155–162.
- 276 -
SANS 14040. 2006. South African National Standard. Environmental management – life cycle assessment principles and framework. Pretoria: Standards South Africa, SABS. Available at https://www.sabs.co.za
(accessed on 5 January 2013).
Saunders, M., Lewis, P. & Thornhill, A. 2009. Research methods for business students. 5th ed. Harlow, Essex:
Pearson Education.
Saysel, A.K. & Hekimoglu, M. 2010. Dynamics of CO2 mitigation in electric power industry through
replacements and early retirements. Institute of Environmental Sciences, Bogazici University, Bebek,
İstanbul.
Available
at
http://www.systemdynamics.org/conferences/2010/proceed/papers/P1233.pdf
(accessed on 12 July 2013).
Schnelker, D. L. 2006. The student as bricoleur: making sense of research paradigms. Teaching and Teacher
Education, 22:42-57.
Schuman, M. & Cavanagh, R. 1982. A model conservation and electric power plan for the Pacific Northwest,
Appendix 2: Environmental costs. NCAC, Seattle, WA.
Scientific Applications International Corporation. 2006. Life cycle assessment: Principles and practice.
Available at http://www.epa.gov/nrmrl/std/lca/pdfs/chapter1_frontmatter_lca101.pdf (accessed on 3
March 2010).
Selgin, G.A. 1988. Praxeology and understanding: an analysis of the controversy in Austrian Economics. The
Review of Austrian Economics, 2(1):19-58.
Sevenster, M., Croezen, H., Van Valkengoed, M., Markowska, A. & Donszelmann, E. 2008. External costs of
coal: global estimate. Available at: http://www.cedelft.eu/publicatie/external_costs_of_ coal/878?PHPSESS
ID=f138219238c72e8038a0a5694354af1d (accessed on 1 July 2011).
Sherwani, A.F., Usmani, J.A & Varun. 2010. Life cycle assessment of solar PV based electricity generation
systems: A review. Renewable and Sustainable Energy Reviews, 14:540–544.
Shields, M. D. & Young, S. M. 1991. Managing product life cycle costs: an organizational model. Journal of
Cost Management, xx:39-52.
- 277 -
Shih, Y. & Tseng, C. 2014. Cost-benefit analysis of sustainable energy development using life-cycle cobenefits assessment and the system dynamics approach. Applied Energy, 119:57-66.
Shogren, J. & Stamland, T. 2001. Skills and the value of life. Journal of Political Economy, 110(x): xx-xx.
Short, W., Packey, D. & Holt, T. 1995. A manual for the economic evaluation of energy efficiency and
renewable energy technologies. NREL/TP-462-5173.
Singh, G. 2008. Mitigating environmental and social impacts of coal mining in India. Available at:
http://www.ismenvis.nic.in/My_Webs/Digital_Library/GSingh/Mitigating%20Environmental%20and%20Soc
ial%20Impacts%20of%20Coal%20Mining%20in%20India.pdf (accessed on 11 July 2011).
Singleton, T.C. 2010. The decision to install flue gas desulphurisation on Medupi power station:
Identification of environmental criteria contributing to the decision making process. Unpublished Doctoral
Thesis. South Africa: University of the Witwatersrand. Available at: http://wiredspace.wits.ac.za/
bitstream/handle/10539/8387/Tyrone%20C%20Singleton_9601499FINAL.pdf?sequence=1 (accessed on 1
January 2013).
Smith, A. & Stirling, A. 2008. Socio-ecological resilience and socio-technical transitions: critical issues for
sustainability governance, Brighton, STEPS Centre, University of Sussex.
South Africa. 1989. Environment Conservation Act, Act no. 73 of 1989. Pretoria: Government Printer.
South Africa. 1997a. Regulations regarding activities identified under Section 21(1) of the Environment
Conservation Act, 1989 (Act No. 73 of 1989). Government Gazette, 18261, September 5 (R 1183). 1997a.
South Africa. 1997b. The identification under Section 21 of activities which may have a substantial
detrimental effect on the environment. Government Gazette, 18261, September 5 (R 1182). 1997b.
South Africa. 2006. Regulations in term of Chapter 5 of the National Environmental Management Act, 1998.
Government Gazette, 28753, April 21. (Notice R 385).
- 278 -
South African Iron & Steel Institute. 2013. Iron and crude steel production in South Africa. Available at:
http://www.saisi.co.za/ironsteelproduction.php (accessed on 29 January 2011).
Southern African Institute for Environmental Assessment. 2004. Environmental impact assessment in
Southern Africa: Summary report. Windhoek: Southern African Institute for Environmental Assessment.
Sovacool, B.K. 2008. Renewable energy: economically sound, politically difficult. The Electricity Journal,
21(5):1040-6190.
Spalding-Fecher, R. & Matibe, D.K. 2003. Electricity and externalities in South Africa. Energy Policy, 31:721734.
Spath, P.L., Mann, M.K. & Kerr, D.R. 1999. Life cycle assessment of coal-fired power production. Available
at: http://efile.mpsc.state.mi.us/efile/docs/16077/0065.pdf (accessed on 27 June 2011).
Spegel, H.W. 1991. The growth of economic thought. 3rd ed. US: Duke University Press.
Standing Committee on Environment and Sustainable Development. 2011. Statutory Review of the
Canadian Environmental Assessment Act, Standing Committee on Environment and Sustainable
Development. Ottawa: Parliament of Canada.
Stanfield, J.R. 1999. The scope, method, and significance of original institutional economics. Journal of
Economic Issues, XXXIII(2):231-255.
Statistics South Africa (SSA). 2012. National accounts: Mineral accounts for South Africa: 1980-2010.
Discussion document: D0405.2. Available at: www.statssa.gov.za (accessed on 5 March 2013).
Statistics South Africa (SSA). 2010. National accounts: Mineral accounts for South Africa: 1980–2007.
Available at: http://www.statssa.gov.za/Publications/D04052/D040522007.pdf (accessed on 13 July 2011).
Statistics South Africa. 2011. Mid-year population estimates 2011. Available at: http://www.statssa.gov
.za/publications/P0302/P03022011.pdf (accessed on 23 February 2013).
Sterman, J. D., Richardson, G. P. & Davidsen, P. I. 1988. Modelling the estimation of petroleum resources in
the United States. Technological Forecasting and Social Change, 33:219-239.
- 279 -
Sterman, J.D. 1981. The energy transition and the economy: a system dynamics approach. Massachusetts
Institute of Technology.
Sterman, J.D. 2000. Business dynamics: systems thinking and modelling for a complex world. New York:
McGraw-Hill/Irwin.
Stern, N. 2007. The economics of climate change: The Stern review. Cambridge: Cambridge University Press.
Stern, N. 2008. The economics of climate change. Am. Econ. Rev. 98: 1–37.
Steyn, M. & Minnitt, R.C.A. 2010. Thermal coal products in South Africa. The Journal of The Southern African
Institute of Mining and Metallurgy, 10:593–599.
Stone, A. & Bennett, K. n.d. A bulk model of emissions from South African diesel commercial
vehicles.Energy
Research
Institute
(ERI),
University
of
Cape
Town.
Available
at:
http
://www.erc.uct.ac.za/Research/publications-pre2004/01Stone-Bennett_Diesel_emissions.PDF (accessed on
13 June 2011).
Sundqvist, T. 2000. Electricity externality studies — do the numbers make sense? Unpublished Doctoral
Thesis. Sweden: Lulea University of Technology.
Sundqvist, T. 2002. Power generation choice in the presence of environmental externalities. Unpublished
Doctoral Thesis. Sweden: Lulea University of Technology. Available at: http://epubl.ltUSe/14021544/2002/26/LTU-DT-0226-SE.pdf (accessed on 15 August 2011).
Swanepoel, E. 2008. Once ‘frightening’ EIA backlog will be cleared this year, State claims.
Synergistics Environmental Services & Zitholele Consulting. 2011. Environmental impact assessment, water
use license and a waste management license for the proposed New Largo Colliery. MDEDET Ref: 17/2/3N41.
Tacconi, L. 1998. Scientific methodology for ecological economics. Ecological Economics, 27:91-105.
- 280 -
Tan, B., Anderson, E., Dyer, J. & Parker, G. 2010. Evaluating system dynamics models of risky projects using
decision trees: alternative energy projects as an illustrative example. System Dynamics Review, 26:1-17.
Tan, R. & Culaba, A. 2002. Environmental life-cycle assessment: a tool for public and corporate policy
development. Washington: American center for life cycle assessment. Available at: http://www.imamu.e
dUS/topics/IT/IT%206/A%20Tool%20for%20Public%20and%20Corporate%20Policy%20Development.pdf
(accessed on 15 August 2011).
Tashakkori, A. & C. Teddlie. 1998. Mixed Methodology. Combining qualitative and quantitative approaches,
Volume 46 of Applied Social Research Methods Series. Thousand Oaks: SAGE Publications.
Taviv, R., Trikam, A., Lane, T., O’Kennedy, K., Mapako, M. & Brent, A.C. 2008. Population of the LEAP system
to model energy futures in South Africa. Available at: http://researchspace.csir.co.za/dspace/bitstream
/10204/2378/1/Taviv_2008.pdf (accessed on 2 July 2011).
Tech Environmental. 2009. Noise study for the Mississippi power company IGCC power plant and coal
mining project. Kemper County, Mississippi.
Thopil, G.A. & Pouris, A. 2010. An overview of the electricity externality analysis in South Africa within the
international context. South African Journal of Science, 106(11/12):1-6.
Tietenberg, T. 1992. Environmental and natural resource economics. New York: Harper Collins.
Tol, R. 2005. The marginal damage costs of carbon dioxide emissions: an assessment of the uncertainties.
Energy Policy, 33(16):2064–2074.
Tol, R. 2009. The economic effects of climate change. Journal of Economic Perspectives, 23(2):29–51.
Tran, T. A. & Daim, T. 2008. A taxonomic review of methods and tools applied in technology assessment.
Technological Forecasting and Social Change, 75, 1396-1405.
Tran, T. A. 2007 Review of methods and tools applied in technology assessment literature. PICMET
Proceedings, 5-9 August, Portland, Oregon, USA.
- 281 -
Treloar, G.J. & Love, P.E.D. 2000. A hybrid life cycle assessment method for construction. Construction
Management and Economics, 18(1):5-10.
Turpie, J., Winkler, H. & Midgley, G. 2004. Economic impacts of climate change in South Africa: a
preliminary assessment of mitigated damage costs. In: Blignaut, J.N. & de Wit, M.P. (ed.). Sustainable
options: economic development lessons from applied environmental resource economics in South Africa.
Cape Town: University of Cape Town Press.
Turpie, J., Winkler, H., Spalding-Fletcher, R. & Midgley, G. 2002. Economic impacts of climate change in
South Africa: a preliminary analysis of the unmitigated damage costs. Southern Waters and Energy and
Development
Research
Centre.
Available
from:
http://drupal6dev15.econsultant.co.za/sites/dru
pal6dev15.econsultant.co.za/files/Economic/Impact/of/Climate/change/Feb02.pdf (accessed on 15 August
2011).
Turton, A. 2008. Three strategic water quality challenges that decision-makers need to know about and
how the CSIR should respond. CSIR conference: Science Real and Relevant. Pretoria, South Africa.
Tzoumis, K. 2007. Comparing the quality of environmental impact statements by agencies in the United
States since 1998 to 2004. Environ Impact Assess Rev, 27(1):26–40.
Unger, T. & Ahlgren, E.O. 2005. Impacts of a common green certificate market on electricity and CO 2emission markets in the Nordic countries. Energy Policy, 33:2152–2163.
Uninterruptible Power Supplies Direct. 2012. Long-term coal supply a worry: Eskom. Available at:
http://www.upsdirect.co.za/index.php?option=com_content&view=article&id=269:long-term-coal-supplya-worry-eskom-&catid=78:eskom (accessed on 21 June 2013).
United Nations Environment Programme. 2002. Environmental impact assessment training resource
manual. Geneva: UNEP.
United States Agency International Development (USAID). 2008. Workers’ compensation in the republic of
South
Africa.
Available
at:
http://ssreform.treasury.gov.za/Publications/Workers'%20Compensatio
n%20in%20the%20Republic%20of%20South%20Africa%20(USAID,%202008).pdf (accessed on 13 August
2014).
- 282 -
United States Energy Information Administration (USEIA). 2012. Levelized cost of new generation resources
in the Annual Energy Outlook. 2012. Available at: http://www.eia.gov/forecasts/aeo/pdf/electricity
_generation.pdf (accessed on 22 July 2013).
Urban Africa. 2013. South Africa’s carbon tax – a means to an end? Available at: http://www.urban
africa.net/blogs/south-africas-carbon-tax-means-end/ (accessed on 10 December 2013).
Urban Earth. 2012. South African carbon snapshot. Available at: http://urbanearth.co.za/system/files/priva
te/Urban%20Earth_SA%20Carbon%20Snapshot_0.pdf (accessed on 11 November 2013).
US. Department of Energy. 2009. Kemper County Integrated Gasification Combined-Cycle (IGCC) project,
Draft Environmental Impact Statement (EIS) (DOE/EIS-0409D). Kemper County, Mississippi.
Van Der Vyver, F. 2008. The quality of Environmental Impact Reports for Explosive Industry Projects in
South Africa. Unpublished Master’s Dissertation. Potchefstroom: North-West University.
Van der Zaag, P. & Savenije, H.H.G. 2006. Water as an economic good: the value of pricing and the failure of
markets. Available at: http://urbanearth.co.za/system/files/priva te/Urban%20Earth_SA%20Carbon%20
Snapshot_0.pdf (accessed on 17 March 2013).
Van Horen, C. 1997. Cheap energy – at what cost? Externalities in South Africa’s electricity sector. In
Counting the social costs: electricity and externalities in South Africa. Van Horen, C. (ed.). Cape Town: Elan
Press and UCT Press.
Van Zyl, H., Raimondo, J. & Leiman, T. 2002. Energy supply sector – coal mining. WWF macroeconomic
reforms and sustainable development in South Africa. Development Bank of South Africa. Midrand, South
Africa.
Vardar, N. & Yumurtaci, Z. 2010. Emissions estimation for lignite-fired power plants in Turkey. Energy
Policy, 38:243–252.
Varun, I.K.B. & Ravi, P. 2009. LCA of renewable energy for electricity generation systems—a review.
Renewable and Sustainable Energy Reviews, 13:1067–1073.
- 283 -
Vattenfall Research and Development Magazine. 2011. Theme: energy system analysis. Available at: http://
www.vattenfall.com/en/file/R_D_Magazine_Jan_2011_16995667.pdf (accessed on 11 January 2012).
Veblen, T. 1919. The Place of Science in Modern Civilisation and Other Essays. New York: Huebsch.
Venema, H.D. & Barg, S. 2003. The full costs of thermal power production in eastern Canada. Available at:
http://www.iisd.org/pdf/2003/energy_fca_canada.pdf (accessed on 11 June 2013).
Venkatachalam, L. 2007. Environmental economics and ecological economics: where they can converge?
Ecological Economics, 61:550-558.
Ventana Systems. 2002. Vensim® User’s Guide Version 5. United States of America: Ventana Systems, Inc.
Ventana Systems. 2003. Vensim® Standard Professional DSS Tutorial Harvard, USA, 60 Jacob Gates Road.
Viscusi, K. & Aldy, J.E. 2003. The value of a statistical life: a critical review of market estimates throughout
the world. Journal of Risk and Uncertainty, 27(1): 5-76.
Viscusi, W.K. 2003. The Value of life: estimates with risks by occupation and industry. Available at:
http://www.law.harvard.edu/programs/olin_center/papers/pdf/422.pdf (accessed on 27 September 2014).
Vogstad, K., 2005. Combining system dynamics and experimental economics to analyse the design of
tradable green certificates. Proceedings of the 38th Hawaii International Conference on System Sciences –
2005.
Available
at:
http://www.computer.org/csdl/proceedings/hicss/2005/2268/02/22680058a.pdf
(accessed on 23 September 2014).
Vogstad, K., Botterud, A., Maribu, K.M. & Jensen, S.G. 2002. The transition from fossil fuelled to a
renewable
power
supply
in
a
deregulated
electricity
market.
http://www.systemdynamics.org/conf erences/2002/proceed/papers/Vogstad1.pdf
Available
at:
(accessed on 23
September 2014).
Von Bertalanffy, L.V. 1952. Problems of life: an evaluation of modern biological thought. New York: Wiley.
Von Bertalanffy, L.V. 1955. General system theory. Main Currents in Modern Thought, 11:75-83.
- 284 -
Von Bertalanffy, L.V. 1968. General system theory: essays on its foundation and development. New York:
George Braziller.
Von Bertalanffy, L.V. 1973. General systems theory: foundations, development, applications. New York:
George Braziller.
Wallace, S.W. & Ziemba, W.T. 2005. Applications of stochastic programming. MPS-SIAM book series on
optimization. Cambridge University Press, Cambridge.
Wang, X., Kurdgelashvili, L., Byrne, J. & Barnett, A. 2011. The value of module efficiency in lowering the
levelized cost of energy of photovoltaic systems. Energy, 36:3606-3612.
Wassung, N. 2010. Water scarcity and electricity generation in South Africa Part 1: water use in the coal-toelectricity process. Unpublished Master’s Dissertation. Stellenbosch: University of Stellenbosch.
Weaver, A. 2003. EIA and sustainable development: key concepts and tools, in Tarr, P. ed., Environmental
Impact Assessment in Southern Africa. Windhoek: Southern African Institute for Environmental Assessment.
Weaver, K. & Olson, J.K. 2006. Understanding paradigms used for nursing research. Journal of Advanced
Nursing, 53(4):459-469.
Weintraub, E.R. 2002. Neoclassical economics, in Concise encyclopaedia of economics, Library of Economics
and Liberty, Liberty Fund.
Weisser, D. 2007. A guide to life-cycle greenhouse gas (GHG) emissions from electric supply technologies.
Energy, 32:1543–1559.
Wester, P., Merrey, D.J. & De Lange, D. 2003. Boundaries of consent: Stakeholder representation in river
basin management in Mexico and South Africa. World Development, 31(5):797–812.
Wilber, C.K. & Harrison, R.S. 1978. The methodological basis of institutional economics: Pattern model,
storytelling, and holism. Journal of Economic Issues, XII(1):61-89.
- 285 -
Winkler, H., Hughes, A., Marquard, A., Haw, M. & Merven, B. 2011. South Africa’s greenhouse gas emissions
under business-as-usual: The technical basis of ‘growth without constraints’ in the Long-Term Mitigation
Scenarios. Energy Policy, 39:5818–5828.
Wolmarans, M. & Medallie, M. 2011. New largo colliery environmental scoping report. Report No.
S0403/NL/SR02. Synergistics Environmental Services, Johannesburg, South Africa.
Wolstenholme, E.F. 2003. The use of system dynamics as a tool for intermediate level technology
evaluation: three case studies. Journal of Engineering and Technology Management, 20:193-204.
Wood, C. 2003. Environmental Impact Assessment: A Comparative Review. 2nd edition. Harlow, England:
Pearson.
Wood, C. 1999. Pastiche or Postiche? Environmental impact assessment in South Africa. South African
Geographical Journal, 81(1):52-59.
Wood, G., Glasson, J. Becker, J. 2006. EIA Scoping in England and Wales: Practitioner Approaches,
Perspectives and Constraints. Environmental Impact Assessment Review, 26:221-241.
World Coal Association (WCA). 2006. Coal facts. Available at: http://www.worldcoal.org/resources/coalstatistics/ (accessed on 18 August 2010).
World Coal Association (WCA). 2007. Coal facts. Available at: http://www.worldcoal.org/resources/coalstatistics/ (accessed on 18 August 2010).
World Coal Association (WCA). 2008. Coal facts. Available at: http://www.worldcoal.org/resources/coalstatistics/ (accessed on 18 August 2010).
World Coal Association (WCA). 2009. Coal facts. Available at: http://www.worldcoal.org/resources/coalstatistics/ (accessed on 18 August 2010).
World Coal Association. 2010. Coal facts. Available at: http://www.worldcoal.org/resources/coalstatistics(accessed 1 September 2011).
- 286 -
World Energy Council. 2004. Comparison of energy systems using life cycle assessment. A special report of
the World Energy Council, ISBN 0 946121 16 8. London: United Kingdom.
Yu, S. & Wei, Y. 2012. Prediction of China’s coal production-environmental pollution based on a hybrid
genetic algorithm-system dynamics model. Energy Policy, 42: 521–529.
Zebda, A. 2002. Using cost-benefit analysis for evaluating decision models in operational research. Journal
of American Academy of Business, xx:106–114.
Zhang, Z.X. & Baranzini, A. 2004. What do we know about carbon taxes? An inquiry into their impacts on
competitiveness and distribution of income. Energy Policy, 32:507-518.
Zitholele Consulting. 2011. Environmental impact assessment and waste management license application
for the proposed extension for the ash disposal facility at Kusile power station. Background information
document. Available at: http://www.zitholele.co.za (accessed on 20 February 2013).
Zweibel, K., Mason, J. & Fthenakis, V. 2008. A solar grand plan. Scientific American, 298:64–73.
- 287 -
APPENDICES
Appendix A: COALPSCA Model equations
A1: Power generation sub-model equations
1-Internal consumption rate=Conversion factor-Fraction of electricity consumed internally
Units: Dmnl
Capacity construction start=Capacity investment/Unit capital cost
Units: MW/Year
Capacity investment=(Planned investment in plant capacity table(Time))*Unit capital cost
Units: R/Year
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal energy content=19.22
Units: MJ/kg
Conversion factor=1
Units: Dmnl
Cumulative gross electricity production=INTEG (Gross electricity production, 0)
Units: MWh
Cumulative net electricity production=INTEG (Net electricity production, 0)
Units: MWh
Days per year=365
Units: Day/Year
Desired functional capacity after construction=IF THEN ELSE(Time<=2015,0,Plant capacity during and after
construction as planned*Effect of profitability on desired functional capacity)
Units: MW
Effect of profitability on desired functional capacity=IF THEN ELSE(Time<=2015,0,Function for effect of
profitability on desired functional capacity(Expected profitability))
Units: Dmnl
Energy availability factor=0.94
Units: Dmnl
Expected profitability= INTEG (Change in expected profitability, 0)
Units: Dmnl
FINAL TIME=2060
Units: Year
- 288 -
Fraction of electricity consumed internally=0.075
Units: Dmnl
Function for effect of profitability on desired functional capacity ([(-30,0)-(1,1.6)],(-30,0),(-0.75,0),(-0.5,0),(0.25,0),(0,0.75),(0.25,0.878),(0.5,0.987),(0.75,1.061),(1,1.091))
Units: Dmnl
Functional capacity during construction=IF THEN ELSE(Time<=2015,Plant capacity during and after
construction as planned, 0)
Units: MW
Gross electricity production=(((Functional capacity during construction*Plant operating hours)+(Desired
functional capacity after construction*Plant operating hours))*Load factor)*MWh/MW*h
Units: MWh/Year
Heat rate=9.769
Units: MJ/kWh
Hours per day=24
Units: h/Day
INITIAL TIME=2010
Units: Year
kg to ton=1000
Units: kg/ton
Load factor=0.9
Units: Dmnl
MWh to kWh=1000
Units: kWh/MWh
MWh/MW*h=1
Units: MWh/(MW*h)
Net electricity production=Gross electricity production*1-Internal consumption rate
Units: MWh/Year
New capacity=IF THEN ELSE(Time<=2014,Plant capacity construction/Plant construction time, 0)
Units: MW/Year
Planned
investment
in
plant
capacity
table
([(2010,0)(2060,2000)],(2010,800),(2011,800),(2012,800),(2013,800),(2014,1600),(2015,0),(2016,0),(2017,0),(2018,0),
(2019,0),(2020,0),(2021,0),(2022,0),(2023,0),(2024,0),(2025,0),(2026,0),(2027,0),(2028,0),(2029,0),(2030,0),
(2031,0),(2032,0),(2033,0),(2034,0),(2035,0),(2036,0),(2037,0),(2038,0),(2039,0),(2040,0),(2041,0),(2042,0),
(2043,0),(2044,0),(2045,0),(2046,0),(2047,0),(2048,0),(2049,0),(2050,0),(2051,0),(2052,0),(2053,0),(2054,0),
(2055,0),(2056,0),(2057,0),(2058,0),(2059,0),(2060,0))
Units: MW/Year
- 289 -
Plant capacity construction= INTEG (Capacity construction start-New capacity, 800)
Units: MW
Plant capacity during and after construction as planned= INTEG (New capacity, 800)
Units: MW
Plant construction time=1
Units: Year
Plant operating hours=Days per year*Hours per day*Energy availability factor
Units: h/Year
Unit capital cost= INTEG (Change in capital cost, Capital cost/Plant size)
Units: R/MW
A2: Generation cost sub-model equations
Capacity investment=(Planned investment in plant capacity table(Time))*Unit capital cost
Units: R/Year
Capital cost=1.185e+011
Units: R
Capital
cost
escalation
table
([(2010,0)(2060,0.002)],(2010,0.001),(2011,0.001),(2012,0.001),(2013,0.001),(2014,0.001),(2015,0),(2016,0),(2017,0),
(2018,0),(2019,0),(2020,0),(2021,0),(2022,0),(2023,0),(2024,0),(2025,0),(2026,0),(2027,0),(2028,0),(2029,0),
(2030,0),(2031,0),(2032,0),(2033,0),(2034,0),(2035,0),(2036,0),(2037,0),(2038,0),(2039,0),(2040,0),(2041,0),
(2042,0),(2043,0),(2044,0),(2045,0),(2046,0),(2047,0),(2048,0),(2049,0),(2050,0),(2051,0),(2052,0),(2053,0),
(2054,0),(2055,0),(2056,0),(2057,0),(2058,0),(2059,0),(2060,0))
Units: Dmnl/Year
Capital investment rate=Capacity investment
Units: R/Year
Change in capital cost=(Capital cost escalation table(Time))*Unit capital cost
Units: R/MW/Year
Change in coal cost=Unit coal cost*Coal cost escalation
Units: (R/ton/Year)
Change in limestone cost=Unit limestone cost*Limestone cost escalation
Units: (R/ton/Year)
Change in other FGD O&M cost=Other FGD O&M costs*Other O&M costs escalation
Units: R/Year
Change in other variable O&M costs=Other variable O&M costs*Other O&M costs escalation
Units: (R/Year)
Change in transport cost=Unit transport cost*Transport cost escalation
- 290 -
Units: (R/ton/km/Year)
Change in water cost=Unit water cost*Water cost escalation
Units: (R/m3/Year)
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal cost=Coal consumption*Unit coal cost
Units: R/Year
Coal cost escalation=0.001
Units: Dmnl/Year
Conversion factor=1
Units: Dmnl
Cumulative capital cost escalated=INTEG (capital investment rate, 0)
Units: R
Cumulative PV FGD operation cost=INTEG (PV FGD operation cost, 0)
Units: R
Cumulative PV fixed O&M costs=INTEG (PV fixed O&M costs, 0)
Units: R
Cumulative PV fuel cost=INTEG (PV fuel cost, 0)
Units: R
Cumulative PV net electrity production=INTEG (PV net electricity production, 1)
Units: MWh
Cumulative PV variable O&M costs=INTEG (PV variable O&M costs, 0)
Units: R
Discount rate=0.08
Units: Dmnl
FGD operation cost=FGD water cost+Limestone cost+Other FGD O&M costs year
Units: R/Year
FGD water consumption=FGD water consumption per MWh*Gross electricity production
Units: m3/Year
FGD water consumption per MWh=0.145
Units: m3/MWh
FGD water cost=FGD water consumption*Unit water cost
Units: R/Year
FINAL TIME=2060
- 291 -
Units: Year
Fixed O&M cost escalation=0.001
Units: Dmnl/Year
Fixed O&M costs year=Fixed O&M costs/Year
Units: R/Year
Fixed O&M costs=INTEG (PV fixed O&M cost, 8.93e+008)
Units: R
Gross electricity production=(((Functional capacity during construction*Plant operating hours)+(Desired
functional capacity after construction*Plant operating hours))*Load factor)*MWh/MW*h
Units: MWh/Year
INITIAL TIME=2010
Units: Year
Levelised capital cost=Cumulative capital cost escalated/Cumulative PV net electrity production
Units: R/MWh
Levelised cost of energy=Levelised fuel cost+Levelised O&M costs+Levelised FGD operaton cost+Levelised
capital cost
Units: R/MWh
Levelised FGD operaton cost=Cumulative PV FGD operation cost/Cumulative PV net electrity production
Units: R/MWh
Levelised fixed O&M costs=Cumulative PV fixed O&M costs/Cumulative PV net electrity production
Units: R/MWh
Levelised fuel cost=Cumulative PV fuel cost/Cumulative PV net electrity production
Units: R/MWh
Levelised O&M costs=Levelised variable O&M costs+Levelised fixed O&M costs
Units: R/MWh
Levelised variable O&M costs=Cumulative PV variable O&M costs/Cumulative PV net electrity production
Units: R/MWh
Limestone consumption=Limestone consumption per hour*Plant operating hours
Units: ton/Year
Limestone consumption cost=Limestone consumption*Unit limestone cost
Units: R/Year
Limestone consumption per hour=70
Units: ton/h
Limestone cost=Limestone consumption cost
Units: R/Year
- 292 -
Limestone cost escalation=0.001
Units: Dmnl/Year
Limestone transport cost=Unit transport cost*Limestone transportation distance*Limestone consumption
Units: R/Year
Limestone transportation distance=120
Units: km
Net electricity production=Gross electricity production*1-Internal consumption rate
Units: MWh/Year
Other FGD O&M costs year=Other FGD O&M costs/Year
Units: R/Year
Other FGD O&M cost= INTEG (Change in other FGD O&M cost, 1.705e+008)
Units: R
Other O&M costs escalation=0.001
Units: Dmnl/Year
Other variable O&M costs year=Other variable O&M costs/Year
Units: R/Year
Other variable O&M costs= INTEG (Change in other variable O&M costs, 7.26e+008)
Units: R
Overnight cost=Capital cost/Plant size
Units: R/MW
Planned
investment
in
plant
capacity
table
([(2010,0)(2060,2000)],(2010,800),(2011,800),(2012,800),(2013,800),(2014,1600),(2015,0),(2016,0),(2017,0),(2018,0),
(2019,0),(2020,0),(2021,0),(2022,0),(2023,0),(2024,0),(2025,0),(2026,0),(2027,0),(2028,0),(2029,0),(2030,0),
(2031,0),(2032,0),(2033,0),(2034,0),(2035,0),(2036,0),(2037,0),(2038,0),(2039,0),(2040,0),(2041,0),(2042,0),
(2043,0),(2044,0),(2045,0),(2046,0),(2047,0),(2048,0),(2049,0),(2050,0),(2051,0),(2052,0),(2053,0),(2054,0),
(2055,0),(2056,0),(2057,0),(2058,0),(2059,0),(2060,0))
Units: MW/Year
Plant operating hours=Days per year*Hours per day*Energy availability factor
Units: h/Year
Plant size=4800
Units: MW
Plant water consumption=Plant water consumption per MWh*Gross electricity production
Units: m3/Year
Plant water consumption per MWh=0.2
Units: m3/MWh
- 293 -
Plant water cost=Plant water consumption*Unit water cost
Units: R/Year
Present value factor=((Conversion factor+Discount rate)^Year of cost(Time))
Units: Dmnl
PV FGD operation cost=FGD operation cost/Present value factor
Units: R/Year
PV fixed O&M cost=Fixed O&M costs*Fixed O&M cost escalation
Units: R/Year
PV fixed O&M costs=Fixed O&M costs year/Present value factor
Units: R/Year
PV fuel cost=Coal cost/Present value factor
Units: R/Year
PV net electricity production=Net electricity production/Present value factor
Units: MWh/Year
PV variable O&M costs=Variable O&M costs/Present value factor
Units: R/Year
Transport cost escalation=0.001
Units: Dmnl/Year
Unit capital cost=INTEG (Change in capital cost, Capital cost/Plant size)
Units: R/MW
Unit coal cost=INTEG (Change in coal cost, 210)
Units: R/ton
Unit limestone cost=INTEG (Change in limestone cost, 335)
Units: R/ton
Unit transport cost=INTEG (Change in transport cost, 1.22)
Units: R/ton/km
Unit water cost=INTEG (Change in water cost, 0.7)
Units: R/m3
Variable O&M costs=Plant water cost+Other variable O&M costs year
Units: R/Year
Water cost escalation=0.001
Units: Dmnl/Year
Year=1
Units: Year
- 294 -
Year
of
cost
([(2010,0)(2060,60)],(2010,1),(2011,2),(2012,3),(2013,4),(2014,5),(2015,6),(2016,7),(2017,8),(2018,9),(2019,10),(2020
,11),(2021,12),(2022,13),(2023,14),(2024,15),(2025,16),(2026,17),(2027,18),(2028,19),(2029,20),(2030,21),(
2031,22),(2032,23),(2033,24),(2034,25),(2035,26),(2036,27),(2037,28),(2038,29),(2039,30),(2040,31),(2041,
32),(2042,33),(2043,34),(2044,35),(2045,36),(2046,37),(2047,38),(2048,39),(2049,40),(2050,41),(2051,42),(
2052,43),(2053,44),(2054,45),(2055,46),(2056,47),(2057,48),(2058,49),(2059,50),(2060,51))
Units: Dmnl
A3: Morbidity and fatalities sub-model equations
Al=Al per MW*Capacity construction start
Units: ton/Year
Al in million tons=Al/Tons to million tons
Units: million tons/Year
Change in morbidity value=Unit morbidity value*Escalation of damage cost
Units: R/person/Year
Change in mortality value=Unit mortality value*Escalation of damage cost
Units: R/person/Year
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal consumption in million tons=Coal consumption/Tons to million tons
Units: million tons/Year
Coal-fuel cycle fatalities & morbidity costs=Fatalities & morbidity costs (coal mining)+Fatalities & morbidity
costs (construction)+Fatalities & morbidity costs (power generation)
Units: R/Year
Concrete=Concrete per MW*Capacity construction start
Units: ton/Year
Concrete in million tons=Concrete/Tons to million tons
Units: million tons/Year
Deaths from coal mining=Fatalities per million tons of coal mined*Coal consumption in million tons
Units: person/Year
Deaths limestone production=Fatalities per million tons of limestone*Limestone in million tons
Units: person/Year
Deaths material inputs production=(Fatalities per million tons of Al*Al in million tons)+(Fatalities per million
tons of concrete*Concrete in million tons)+(Fatalities per million tons of steel*Steel in million tons)
Units: person/Year
Deaths power generation=Fatalities per MWh*Gross electricity production
Units: person/Year
- 295 -
Escalation of damage cost=0.011
Units: Dmnl/Year
Fatalities & morbidity costs (coal mining)=Fatalities cost (coal mining)+Morbidity cost (coal mining)
Units: R/Year
Fatalities & morbidity costs (construction)=Fatality cost due to material inputs production+Morbidity cost
due to material inputs production
Units: R/Year
Fatalities & morbidity costs (power generation)=Fatalities & mortality costs limestone production
(FGD)+Fatality cost power generation+Morbidity cost power generation
Units: R/Year
Fatalities & mortality costs limestone production (FGD)= Fatality
production+Morbidity cost due to limestone production
Units: R/Year
cost
due
to
limestone
Fatalities cost (coal mining)=Deaths from coal mining*Unit mortality value
Units: R/Year
Fatalities per million tons of Al= 3.17428
Units: persons/million tons
Fatalities per million tons of coal mined=0.056
Units: person/million tons
Fatalities per million tons of concrete=0.159
Units: person/million tons
Fatalities per million tons of limestone= 0.290698
Units: person/million tons
Fatalities per million tons of steel=0.321374
Units: person/million tons
Fatalities per MWh=2.6e-007
Units: person/MWh
Fatality cost due to limestone production=Deaths limestone production*Unit mortality value
Units: R/Year
Fatality cost due to material inputs production= Deaths material inputs production*Unit mortality value
Units: R/Year
Fatality cost power generation=Deaths power generation*Unit mortality value
Units: R/Year
FINAL TIME=2060
Units: Year
- 296 -
Gross electricity production=(((Functional capacity during construction*Plant operating hours)+(Desired
functional capacity after construction*Plant operating hours))*Load factor)*"MWh/MW*h"
Units: MWh/Year
INITIAL TIME=2010
Units: Year
Injuries from coal mining=Injuries per million tons of coal mined*Coal consumption in million tons
Units: person/Year
Injuries limestone production=Injuries per million tons of limestone*Limestone in million tons
Units: person/Year
Injuries material inputs production=(Injuries per million tons of Al*Al in million tons)+(Injuries per million
tons of concrete*Concrete in million tons)+(Injuries per million tons of steel*Steel in million tons)
Units: person/Year
Injuries per million tons of Al=19.9114
Units: person/million tons
Injuries per million tons of coal mined= 0.823
Units: person/million tons
Injuries per million tons of concrete=0.995
Units: person/million tons
Injuries per million tons of limestone=1.33721
Units: person/million tons
Injuries per million tons of steel=2.01589
Units: person/million tons
Injuries power generation=Injury rate per MWh*Gross electricity production
Units: person/Year
Injury rate per MWh=1e-007
Units: person/MWh
Limestone consumption=Limestone consumption per hour*Plant operating hours
Units: ton/Year
Limestone in million tons=Limestone consumption/Tons to million tons
Units: million tons/Year
Morbidity cost (coal mining)=Injuries from coal mining*Unit morbidity value
Units: R/Year
Morbidity cost due to limestone production=Injuries limestone production*Unit morbidity value
Units: R/Year
- 297 -
Morbidity cost due to material inputs production=Injuries material inputs production*Unit morbidity value
Units: R/Year
Morbidity cost power generation=Injuries power generation*Unit morbidity value
Units: R/Year
Steel=Steel per MW*Capacity construction start
Units: ton/Year
Steel in million tons=Steel/Tons to million tons
Units: million tons/Year
Tons to million tons=1e+006
Units: ton/million tons
Unit morbidity value=INTEG (Change in morbidity value, 25434)
Units: R/person
Unit mortality value= INTEG (Change in mortality value, 245438)
Units: R/person
A4: Water consumption sub-model equations
Al=Al per MW*Capacity construction start
Units: ton/Year
Al embodied water=8.8e-005
Units: m3/ton
Ash produced per ton of coal burnt=0.293
Units: ton/ton
Change in the opportunity cost of water use=Unit opportunity cost of water use*Escalation of damage cost
Units: R/m3/Year
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal-fuel cycle externality cost of water use=Opportunity cost of water use (construction)+Opportunity cost
of water use (power generation)+Opportunity cost of water use in disposing Kusile's waste+Opportunity
cost of water use in FGD+Opportunity cost of water use in the New Largo colliery (coal mining)
Units: R/Year
Concrete=Concrete per MW*Capacity construction start
Units: ton/Year
Concrete embodied water=26.352
Units: m3/ton
Dry waste Kusile=Ash produced per ton of coal burnt*Coal consumption
Units: ton/Year
- 298 -
Escalation of damage cost=0.011
Units: Dmnl/Year
Factor
curbing
construction
([(2010,0)(2060,1)],(2010,1),(2011,1),(2012,1),(2013,1),(2014,1),(2015,0),(2016,0),(2017,0),(2018,0),(2019,0),(2020,0)
,(2021,0),(2022,0),(2023,0),(2024,0),(2025,0),(2026,0),(2027,0),(2028,0),(2029,0),(2030,0),(2031,0),(2032,0)
,(2033,0),(2034,0),(2035,0),(2036,0),(2037,0),(2038,0),(2039,0),(2040,0),(2041,0),(2042,0),(2043,0),(2044,0)
,(2045,0),(2046,0),(2047,0),(2048,0),(2049,0),(2050,0),(2051,0),(2052,0),(2053,0),(2054,0),(2055,0),(2056,0)
,(2057,0),(2058,0),(2059,0),(2060,0))
Units: Dmnl
FGD water consumption=FGD water consumption per MWh*Gross electricity production
Units: m3/Year
FINAL TIME=2060
Units: Year
INITIAL TIME=2010
Units: Year
Litres to m3=1000
Units: l/m3
Opportunity cost of water use (construction)=Opportunity cost of water use in producing material inputs of
constructing Kusile+Opportunity cost of water use in constructing Kusile
Units: R/Year
Opportunity cost of water use (power generation)=Unit opportunity cost of water use*Plant water
consumption
Units: R/Year
Opportunity cost of water use in constructing Kusile=Unit opportunity cost of water use*Water
requirements of constructing Kusile (curbed)
Units: R/Year
Opportunity cost of water use in disposing Kusile's waste=Unit opportunity cost of water use*Water usage
in disposing waste
Units: R/Year
Opportunity cost of water use in FGD=Unit opportunity cost of water use*FGD water consumption
Units: R/Year
Opportunity cost of water use in producing material inputs of constructing Kusile=Unit opportunity cost of
water use*Water requirements of construction materials
Units: R/Year
Opportunity cost of water use in the New Largo colliery (coal mining)=Unit opportunity cost of water
use*Water requirements of a surface mine
Units: R/Year
- 299 -
Plant water consumption=Plant water consumption per MWh*Gross electricity production
Units: m3/Year
Steel=Steel per MW*Capacity construction start
Units: ton/Year
Steel embodied water=225
Units: m3/ton
Unit opportunity cost of water use=INTEG (Change in the opportunity cost of water use, 1001)
Units: R/m3
Water requirements of a surface mine=Water requirements of a surface mine (in m3/ton)*Coal
consumption
Units: m3/Year
Water requirements of a surface mine (in litres/ton)=469
Units: l/ton
Water requirements of a surface mine (in m3/ton)=Water requirements of a surface mine (in
litres/ton)/Litres to m3
Units: m3/ton
Water requirements of constructing Kusile=4.12392e+006
Units: m3/Year
Water requirements of constructing Kusile (curbed)=(Water requirements of constructing Kusile*Factor
curbing construction(Time))
Units: m3/Year
Water requirements of construction materials=Water usage Al+Water usage concrete+Water usage steel
Units: m3/Year
Water usage Al=Al embodied water*Al
Units: m3/Year
Water usage concrete=Concrete embodied water*Concrete
Units: m3/Year
Water usage in disposing waste=Water usage per ton of solid waste disposed*Dry waste Kusile
Units: m3/Year
Water usage per ton of solid waste disposed=0.076
Units: m3/ton
Water usage steel=Steel embodied water*Steel
Units: m3/Year
A5: Water pollution sub-model equations
Al=Al per MW*Capacity construction start
- 300 -
Units: ton/Year
Change in damage cost of sulphate pollution (Al & concrete production) =Unit damage cost of sulphate
pollution from Al & concrete production*Escalation of damage cost
Units: R/ton/Year
Change in damage cost of sulphate pollution (coal mining)=Unit damage cost of sulphate pollution from
coal mining*Escalation of damage cost
Units: R/ton/Year
Change in damage cost of sulphate pollution (steel production)= Unit damage cost of sulphate pollution
from steel production*Escalation of damage cost
Units: R/ton/Year
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal-fuel cycle water pollution externality cost=Damage cost of sulphate pollution from coal
mining+Damage cost of sulphate pollution from Kusiles' raw material requirements
Units: R/Year
Concrete=Concrete per MW*Capacity construction start
Units: ton/Year
Damage cost of sulphate pollution from Al & cement production=(Unit damage cost of sulphate pollution
from Al & concrete production*Al)+(Unit damage cost of sulphate pollution from Al & concrete
production*Concrete)
Units: R/Year
Damage cost of sulphate pollution from coal mining=Unit damage cost of sulphate pollution from coal
mining*Coal consumption
Units: R/Year
Damage cost of sulphate pollution from Kusiles' raw material requirements=Damage cost of sulphate
pollution from Al & cement production+Damage cost of sulphate pollution from steel production
Units: R/Year
Damage cost of sulphate pollution from steel production=Unit damage cost of sulphate pollution from steel
production*Steel
Units: R/Year
Escalation of damage cost=0.011
Units: Dmnl/Year
FINAL TIME=2060
Units: Year
INITIAL TIME=2010
Units: Year
Steel=Steel per MW*Capacity construction start
- 301 -
Units: ton/Year
Unit damage cost of sulphate pollution from Al & concrete production=INTEG (Change in damage cost of
sulphate pollution (Al & concrete production), 0.31)
Units: R/ton
Unit damage cost of sulphate pollution from coal mining=INTEG (Change in damage cost of sulphate
pollution (coal mining), 0.27)
Units: R/ton
Unit damage cost of sulphate pollution from steel production=INTEG (Change in damage cost of sulphate
pollution (steel production), 0.79)
Units: R/ton
A6: Ecosystem services loss sub-model equations
Change in maize price=Unit maize price*Escalation of damage cost
Units: R/ton/Year
Change in the value of ecosystem goods & services=Unit value of ecosystem goods & services generated by
grasslands*Escalation of damage cost
Units: R/ha/Year
Coal-fuel cycle cost of lost ecosystem services=Ecosystem services lost due to coal mining+"Ecosystem
services lost due to plant construction & operation"
Units: R/Year
Ecosystem services lost due to coal mining=Forgone benefit from grasslands due to coal mining+Forgone
benefit from maize cultivation due to coal mining
Units: R/Year
Ecosystem services lost due to plant construction & operation=Forgone benefit from grasslands due to
building and operating plant+Forgone benefit from maize cultivation due to building and operating plant
Units: R/Year
Escalation of damage cost=0.011
Units: Dmnl/Year
FINAL TIME=2060
Units: Year
Forgone benefit from grasslands due to building and operating plant=Power plant area under grazing*Unit
value of ecosystem goods & services generated by grasslands
Units: R/Year
Forgone benefit from grasslands due to coal mining=Mining area under grazing/grasslands*Unit value of
ecosystem goods & services generated by grasslands
Units: R/Year
- 302 -
Forgone benefit from maize cultivation due to building and operating plant=(Maize production (dry
land)*Unit maize price)+(Maize production (irrigated land)*Unit maize price)
Units: R/Year
Forgone benefit from maize cultivation due to coal mining=Maize production*Unit maize price
Units: R/Year
INITIAL TIME=2010
Units: Year
Maize production=Mining area under maize production*Maize yield per hectare
Units: ton/Year
Maize production (dry land)=Power plant area under dry land maize production*Maize yield per hectare
(dry land)
Units: ton/Year
Maize production (irrigated land)=Power plant area under irrigated maize production*Maize yield per
hectare (irrigated land)
Units: ton/Year
Maize yield per hectare=10
Units: ton/ha
Maize yield per hectare (dry land)=4.25
Units: ton/ha
Maize yield per hectare (irrigated land)=10
Units: ton/ha
Mining area under grazing/grasslands=2045.1
Units: ha/Year
Mining area under maize production=4771.9
Units: ha/Year
Power plant area under dry land maize production=1404
Units: ha/Year
Power plant area under grazing=3744
Units: ha/Year
Power plant area under irrigated maize production=52
Units: ha/Year
Unit maize price=INTEG (Change in maize price, 1600)
Units: R/ton
Unit value of ecosystem goods & services generated by grasslands=INTEG (Change in the value of
ecosystem goods & services, 510)
Units: R/ha
- 303 -
A7: Air pollution sub-model equations
Arsenic content in coal= 2.95
Units: mg/kg
Arsenic damages=Coal combustion arsenic & compounds emissions*Unit damage cost arsenic
Units: R/Year
Arsenic emission factor in kg/PJ=Constant arsenic*(((Arsenic content in coal/Weight fraction of ash in
coal)*PM emitted per GJ heat input)*GJ to PJ/mg to kg)^(Exponent arsenic)
Units: kg/PJ
Change in damage cost of nickel=Unit damage cost nickel*Escalation of damage cost
Units: R/ton/Year
Change in damage cost per ton of arsenic=Unit damage cost arsenic*Escalation of damage cost
Units: R/ton/Year
Change in damage cost per ton of chromium=Unit damage cost chromium*Escalation of damage cost
Units: R/ton/Year
Change in damage cost per ton of lead=Unit damage cost lead*Escalation of damage cost
Units: R/ton/Year
Change in NOx damage cost=Unit damage cost NOx*Escalation of damage cost
Units: R/ton/Year
Change in PM damage cost=Unit damage cost PM*Escalation of damage cost
Units: R/ton/Year
Change in SO2 damage cost=Unit damage cost SO2*Escalation of damage cost
Units: R/ton/Year
Chromium content in coal=57.02
Units: mg/kg
Chromium damages=Coal combustion chromium & compounds emissions*Unit damage cost chromium
Units: R/Year
Chromium emission factor in kg/PJ=Constant chromium*(((Chromium content in coal/Weight fraction of
ash in coal)*PM emitted per GJ heat input)*GJ to PJ/mg to kg)^(Exponent chromium)
Units: kg/PJ
Coal combustion air pollution health damages=Coal combustion NOx damages+Coal combustion PM
damages+Coal combustion SO2 damages
Units: R/Year
Coal combustion arsenic & compounds emissions=(Arsenic emission factor in kg/PJ*Coal consumption in
PJ)/kg to ton
Units: ton/Year
- 304 -
Coal combustion chromium & compounds emissions=(Chromium emission factor in kg/PJ*Coal
consumption in PJ)/kg to ton
Units: ton/Year
Coal combustion heavy metals damages=Arsenic damages+Chromium damages+Lead damages+Nickel
damages
Units: R/Year
Coal combustion lead & compounds emissions= (Lead emission factor in kg/PJ*Coal consumption in PJ)/kg
to ton
Units: ton/Year
Coal combustion nickel & compounds emissions=(Nickel emission factor in kg/PJ*Coal consumption in
PJ)/kg to ton
Units: ton/Year
Coal combustion NOx damages=Coal combustion NOx emissions*Unit damage cost NOx
Units: R/Year
Coal combustion NOx emissions=Gross electricity production*Emission factor NOx (coal)
Units: ton/Year
Coal combustion PM damages=Coal combustion PM emissions*Unit damage cost PM
Units: R/Year
Coal combustion PM emissions=Gross electricity production*"Emission factor PM (coal)"
Units: ton/Year
Coal combustion SO2 damages= Coal combustion SO2 emissions*Unit damage cost SO2
Units: R/Year
Coal combustion SO2 emissions=Gross electricity production*Emission factor SO2 (coal)
Units: ton/Year
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal consumption in PJ=(Coal energy content in GJ/ton*Coal consumption)/GJ to PJ
Units: PJ/Year
Coal energy content=19.22
Units: MJ/kg
Coal energy content in GJ/ton=Coal energy content*(kg to ton/MJ to GJ)
Units: GJ/ton
Coal road transport NOx damages=Coal road transport NOx emissions*Unit damage cost NOx
Units: R/Year
Coal road transport NOx emissions=Emissions NOx in grams/g to ton
- 305 -
Units: ton/Year
Coal road transport PM damages=Coal road transport PM emissions*Unit damage cost PM
Units: R/Year
Coal road transport PM emissions=Emissions PM in grams/g to ton
Units: ton/Year
Coal road transport SO2 damages=Coal road transport SO2 emissions*Unit damage cost SO2
Units: R/Year
Coal road transport SO2 emissions=Emissions SO2 in grams/g to ton
Units: ton/Year
Coal road transportation distance=2.21484e+007
Units: km/Year
Coal transportation air pollution health cost=Coal road transport NOx damages+Coal road transport PM
damages+Coal road transport SO2 damages+Conveyor coal transport air pollution damages
Units: R/Year
Coal-fuel cycle air pollution human health cost= Coal transportation air pollution health cost+FGD system
air pollution health cost +Plant construction air pollution health cost+Plant operation air pollution health
cost+Waste disposal air pollution health cost
Units: R/Year
Constant arsenic=2.73
Units: Dmnl
Constant chromium=2.47
Units: Dmnl
Constant lead=2.87
Units: Dmnl
Constant nickel=2.84
Units: Dmnl
Conversion factor=1
Units: Dmnl
Conveyor coal transport air pollution damages=Conveyor coal transport NOx damages+Conveyor coal
transport PM damages+Conveyor coal transport SO2 damages
Units: R/Year
Conveyor coal transport NOx damages= Conveyor coal transport NOx emissions*Unit damage cost NOx
Units: R/Year
Conveyor coal transport NOx emissions=NOx emissions per MWh*Electricity use by conveyor
Units: ton/Year
- 306 -
Conveyor coal transport PM damages=Conveyor coal transport PM emissions*Unit damage cost PM
Units: R/Year
Conveyor coal transport PM emissions=PM emissions per MWh*Electricity use by conveyor
Units: ton/Year
Conveyor coal transport SO2 damages=Conveyor coal transport SO2 emissions*Unit damage cost SO2
Units: R/Year
Conveyor coal transport SO2 emissions=SO2 emissions per MWh*Electricity use by conveyor
Units: ton/Year
Distanced travelled (construction materials)=Number of road trips*Transportation distance (round trip)
Units: km/Year
Electricity use by conveyor=Conveyor electricity use per ton-km*Conveyor transported coal*Conveyor
length
Units: MWh/Year
Emission factor NOx (coal)=0.00228
Units: ton/MWh
Emission factor NOx (transportation)=13.04
Units: g/km
Emission factor PM (coal)=0.000221
Units: ton/MWh
Emission factor PM (transportation)=0.68
Units: g/km
Emission factor SO2 (coal)=0.00095
Units: ton/MWh
Emission factor SO2 (transportation)=1.66
Units: g/km
Emissions NOx in grams=Coal road transportation distance*Emission factor NOx (transportation)
Units: g/Year
Emissions PM in grams= Coal road transportation distance*Emission factor PM (transportation)
Units: g/Year
Emissions SO2 in grams=Coal road transportation distance*"Emission factor SO2 (transportation)"
Units: g/Year
Escalation of damage cost=0.011
Units: Dmnl/Year
Exponent arsenic=0.85
Units: Dmnl
- 307 -
Exponent chromium=0.58
Units: Dmnl
Exponent lead= 0.8
Units: Dmnl
Exponent nickel=0.48
Units: Dmnl
FGD system air pollution health cost=Limestone transportation NOx damages+Limestone transportation
PM damages+Limestone transportation SO2 damages
Units: R/Year
FINAL TIME=2060
Units: Year
Fly ash fraction of total ash=0.2
Units: Dmnl
Fraction of fly ash emitted=(Conversion factor-(Particulate collection efficiency/100))
Units: Dmnl
g to ton=1e+006
Units: g/ton
GJ to PJ=1e+006
Units: GJ/PJ
Gross electricity production=(((Functional capacity during construction*Plant operating hours)+(Desired
functional capacity after construction*Plant operating hours))*Load factor)*"MWh/MW*h"
Units: MWh/Year
INITIAL TIME=2010
Units: Year
kg to ton=1000
Units: kg/ton
Lead content in coal=20.38
Units: mg/kg
Lead damages=Coal combustion lead & compounds emissions*Unit damage cost lead
Units: R/Year
Lead emission factor in kg/PJ=Constant lead*(((Lead content in coal/Weight fraction of ash in coal)*PM
emitted per GJ heat input)*GJ to PJ/mg to kg)^(Exponent lead)
Units: kg/PJ
Limestone transportation electricity use=Conveyor
consumption*Limestone transportation distance
- 308 -
electricity
use
per
ton-km*Limestone
Units: MWh/Year
Limestone transportation NOx damages=Limestone transportation NOx emissions*Unit damage cost NOx
Units: R/Year
Limestone transportation NOx emissions=NOx emissions per MWh*Limestone transportation electricity use
Units: ton/Year
Limestone transportation PM damages= Limestone transportation PM emissions*Unit damage cost PM
Units: R/Year
Limestone transportation PM emissions=PM emissions per MWh*Limestone transportation electricity use
Units: ton/Year
Limestone transportation SO2 damages=Limestone transportation SO2 emissions*Unit damage cost SO2
Units: R/Year
Limestone transportation SO2 emissions=SO2 emissions per MWh*Limestone transportation electricity use
Units: ton/Year
Material transportation NOx damages=Material transportation NOx emissions*Unit damage cost NOx
Units: R/Year
Material transportation NOx emissions= (Distanced travelled (construction materials)*Emission factor NOx
(transportation))/g to ton
Units: ton/Year
Material transportation PM damages=Material transportation PM emissions*Unit damage cost PM
Units: R/Year
Material transportation PM emissions=(Distanced travelled (construction materials)*Emission factor PM
(transportation))/g to ton
Units: ton/Year
Material transportation SO2 damages=Material transportation SO2 emissions*Unit damage cost SO2
Units: R/Year
Material transportation SO2 emissions=(Distanced travelled (construction materials)*Emission factor SO2
(transportation))/g to ton
Units: ton/Year
mg to kg=1e+006
Units: mg/kg
MJ to GJ=1000
Units: MJ/GJ
Nickel content in coal=25.69
Units: mg/kg
Nickel damages=Coal combustion nickel & compounds emissions*unit damage cost nickel
- 309 -
Units: R/Year
Nickel emission factor in kg/PJ=Constant nickel*(((Nickel content in coal/Weight fraction of ash in coal*PM
emitted per GJ heat input)*GJ to PJ/mg to kg)^(Exponent nickel)
Units: kg/PJ
NOx emissions per MWh=0.00389
Units: ton/MWh
Particulate collection efficiency=99.8
Units: Dmnl
Plant construction air pollution health cost=Plant construction raw material transportation damages
Units: R/Year
Plant construction raw material transportation damages=Material transportation NOx damages+Material
transportation PM damages+Material transportation SO2 damages
Units: R/Year
Plant operation air pollution health cost=Coal combustion air pollution health damages+Coal combustion
heavy metals damages
Units: R/Year
PM emissions per MWh=0.000358
Units: ton/MWh
PM emitted per GJ heat input=Weight fraction of ash in coal*Fly ash fraction of total ash*Fraction of fly ash
emitted *ton to kg/"Coal energy content in GJ/ton"
Units: kg/GJ
SO2 emissions per MWh=0.00753
Units: ton/MWh
ton to kg=1000
Units: kg/ton
Unit damage cost arsenic=INTEG (Change in damage cost per ton of arsenic, 339976)
Units: R/ton
Unit damage cost chromium=INTEG (Change in damage cost per ton of chromium, 133866)
Units: R/ton
Unit damage cost lead=INTEG (Change in damage cost per ton of lead, 6.79953e+006)
Units: R/ton
Unit damage cost nickel=INTEG (Change in damage cost of nickel, 16149)
Units: R/ton
Unit damage cost NOx=INTEG (Change in NOx damage cost, 41952)
Units: R/ton
- 310 -
Unit damage cost PM=INTEG (Change in PM damage cost, 227175)
Units: R/ton
Unit damage cost SO2=INTEG (Change in SO2 damage cost, 51619)
Units: R/ton
Waste disposal air pollution health cost=Waste disposal NOx damages+Waste disposal PM damages+Waste
disposal SO2 damages
Units: R/Year
Waste disposal electricity use=Conveyor electricity use per ton-km*Dry waste Kusile*Distance travelled
Kusile waste
Units: MWh/Year
Waste disposal NOx damages=Waste disposal NOx emissions*Unit damage cost NOx
Units: R/Year
Waste disposal NOx emissions=NOx emissions per MWh*Waste disposal electricity use
Units: ton/Year
Waste disposal PM damages=Waste disposal PM emissions*Unit damage cost PM
Units: R/Year
Waste disposal PM emissions=PM emissions per MWh*Waste disposal electricity use
Units: ton/Year
Waste disposal SO2 damages=Waste disposal SO2 emissions*Unit damage cost SO2
Units: R/Year
Waste disposal SO2 emissions=SO2 emissions per MWh*Waste disposal electricity use
Units: ton/Year
Weight fraction of ash in coal=29.6
Units: Dmnl
A8: Global pollutants sub-model equations
Al=Al per MW*Capacity construction start
Units: ton/Year
Al C2F6 embodiment=0.04
Units: kg/ton
Al CF4 embodiment=0.8
Units: kg/ton
Al CO2 embodiment=5301
Units: kg/ton
Al per MW=0.419
Units: ton/MW
- 311 -
Al production CO2e emissions=((CO2 Al production)+(C2F4 Al production*Global warming potential
C2F6)+(CF4 Al production*Global warming potential CF4))/kg to ton
Units: ton/Year
C emission factor for diesel=20.2
Units: ton/TJ
C2F4 Al production=Al*Al C2F6 embodiment
Units: kg/Year
Capacity construction start=Capacity investment/Unit capital cost
Units: MW/Year
Carbon %=0.425
Units: Dmnl
Carbon oxidation factor=0.99
Units: Dmnl
CF4 Al production=Al*Al CF4 embodiment
Units: kg/Year
CH4 emission factor for heavy duty diesel vehicles=1.97917e-008
Units: ton/l
CH4 emission m3/ton=0.014
Units: m3/ton
CH4 emission per ton of coal=CH4 emission m3/ton*Density of bituminous coal
Units: kg/ton
CH4 steel production=Steel*Steel CH4 embodiment
Units: kg/Year
Change in CO2 damage cost=Unit damage cost CO2*Escalation of damage cost
Units: R/ton/Year
CO2 Al production=Al*Al CO2 embodiment
Units: kg/Year
CO2 emission factor=(1/Coal energy content in kJ/ton)*Carbon %*Gravimetric factor converting C to
CO2*Carbon oxidation factor
Units: ton/kJ
CO2 emission per MWh=IF THEN ELSE(Gross electricity production<1e-006,0,Coal combustion CO2
emissions/Gross electricity production)
Units: ton/MWh
CO2 steel production=Steel*Steel CO2 embodiment
Units: kg/Year
- 312 -
Coal combustion CO2 damages=Coal combustion CO2 emissions*Unit damage cost CO2
Units: R/Year
Coal combustion CO2 emissions=CO2 emission factor*Coal consumption in kJ
Units: ton/Year
Coal combustion CO2e damages (N2O)=Coal combustion CO2e emissions (N2O)*Unit damage cost CO2
Units: R/Year
Coal combustion CO2e emissions (N2O)=N2O emissions per MWh*Gross electricity production*Global
warming potential N2O
Units: ton/Year
Coal consumption=(((Gross electricity production*MWh to kWh)*Heat rate)/Coal energy content)/kg to ton
Units: ton/Year
Coal consumption in kJ=Coal energy content in kJ/ton*Coal consumption
Units: kJ/Year
Coal energy content=19.22
Units: MJ/kg
Coal energy content in kJ/ton=Coal energy content*(kg to ton/MJ to kJ)
Units: kJ/ton
Coal mining & transportation global warming damages= Coal mining CO2e damages+Coal road transport
global warming damages+Conveyor coal transport damages
Units: R/Year
Coal mining CO2e damages=Unit damage cost CO2*Coal mining CO2e emissions (CH4)
Units: R/Year
Coal mining CO2e emissions (CH4)=((CH4 emission per ton of coal*Coal consumption)/kg to ton)*Global
warming potential CH4
Units: ton/Year
Coal road transpoirt CO2e damages=Unit damage cost CO2*Coal road transport CO2e emissions (CH4)
Units: R/Year
Coal road transport CO2 damages=Coal road transport CO2 emissions*Unit damage cost CO2
Units: R/Year
Coal road transport CO2 emissions=Diesel consumption in TJ*C emission factor for diesel*Diesel oxidation
factor*Gravimetric factor converting C to CO2
Units: ton/Year
Coal road transport CO2e damages=Unit damage cost CO2*Coal road transport CO2e emissions (N2O)
Units: R/Year
- 313 -
Coal road transport CO2e emissions (CH4)=(CH4 emission factor for heavy duty diesel vehicles*Diesel
consumption in litres (coal transport))*Global warming potential CH4
Units: ton/Year
Coal road transport CO2e emissions (N2O)=(N2O emission factor for heavy duty diesel vehicle*Diesel
consumption in litres (coal transport))*Global warming potential N2O
Units: ton/Year
Coal road transport global warming damages=Coal road transport CO2 damages+Coal road transport CO2e
damages+Coal road transpoirt CO2e damages
Units: R/Year
Coal road transportation distance=2.21484e+007
Units: km/Year
Coal-fuel cycle global warming damage cost=Coal mining & transportation global warming damages"+FGD
system global warming damages+Plant construction global warming damages+Plant operation global
warming damages+Waste disposal global warming damages
Units: R/Year
Concrete=Concrete per MW*Capacity construction start
Units: ton/Year
Concrete CO2e embodiment=119.72
Units: kg/ton
Concrete per MW=158.758
Units: ton/MW
Concrete production CO2e emissions=(Concrete*Concrete CO2e embodiment)/kg to ton
Units: ton/Year
Construction materials CO2e damages=(Unit damage cost CO2*Steel production CO2e emissions)+(Unit
damage cost CO2*Al production CO2e emissions)+(Unit damage cost CO2*Concrete production CO2e
emissions)
Units: R/Year
Conversion factor=1
Units: Dmnl
Conveyor coal transport CO2 damages=Unit damage cost CO2*Conveyor coal transport CO2 emissions
Units: R/Year
Conveyor coal transport CO2 emissions=CO2 emission per MWh*Electricity use by conveyor
Units: ton/Year
Conveyor coal transport CO2e damages=Unit damage cost CO2*Conveyor coal transport CO2e emissions
(N2O)
Units: R/Year
- 314 -
Conveyor coal transport CO2e emissions (N2O)= N2O
conveyor*Global warming potential N2O
Units: ton/Year
emissions
per
MWh*Electricity
use
by
Conveyor coal transport damages=Conveyor coal transport CO2 damages+Conveyor coal transport CO2e
damages
Units: R/Year
Conveyor electricity use per ton-km=0.0002
Units: MWh/ton/km
Conveyor length=42
Units: km
Conveyor transported coal=(Conversion factor-Fraction of coal transported by road)*Coal consumption
Units: ton/Year
Density of bituminous coal=732
Units: kg/m3
Diesel consumption (TJ)=(Diesel consumption in litres (construction)*Energy density of diesel)/MJ to TJ
Units: TJ/Year
Diesel consumption in litres (coal transport)=Coal road transportation distance*Truck fuel consumption in
l/km
Units: l/Year
Diesel consumption in litres (construction)=Distanced travelled (construction materials)*Truck fuel
consumption in l/km
Units: l/Year
Diesel consumption in TJ=(Diesel consumption in litres (coal transport)*Energy density of diesel)/MJ to TJ
Units: TJ/Year
Diesel oxidation factor= 0.99
Units: Dmnl
Distance travelled Kusile waste=30
Units: km
Distanced travelled (construction materials)=Number of road trips*Transportation distance (round trip)
Units: km/Year
Dry waste Kusile=Ash produced per ton of coal burnt*Coal consumption
Units: ton/Year
Electricity use by conveyor=Conveyor electricity use per ton-km*Conveyor transported coal*Conveyor
length
Units: MWh/Year
Energy density of diesel=38.46
- 315 -
Units: MJ/l
Escalation of damage cost=0.011
Units: Dmnl/Year
FGD system global warming damages=Limestone transportation CO2 damages+Limestone transportation
CO2e damages (N2O)+Limestone use damages (FGD)
Units: R/Year
FINAL TIME=2060
Units: Year
Fraction of coal transported by road=0.27
Units: Dmnl
Global warming potential C2F6= 12200
Units: Dmnl
Global warming potential CF4=7390
Units: Dmnl
Global warming potential CH4=23
Units: Dmnl
Global warming potential N2O=310
Units: Dmnl
Gravimetric factor converting C to CO2=3.667
Units: Dmnl
Gross electricity production=(((Functional capacity during construction*Plant operating hours)+(Desired
functional capacity after construction*Plant operating hours))*Load factor)*MWh/MW*h
Units: MWh/Year
INITIAL TIME=2010
Units: Year
kg to ton=1000
Units: kg/ton
Limestone consumption=Limestone consumption per hour*Plant operating hours
Units: ton/Year
Limestone transportation CO2 damages=Unit damage cost CO2*Limestone transportation CO2 emissions
Units: R/Year
Limestone transportation CO2 emissions=CO2 emission per MWh*Limestone transportation electricity use
Units: ton/Year
Limestone transportation CO2e damages (N2O)=Unit damage cost CO2*"Limestone transportation CO2e
emissions (N2O)
- 316 -
Units: R/Year
Limestone transportation CO2e emissions (N2O)=N2O emissions per MWh*Limestone transportation
electricity use*Global warming potential N2O
Units: ton/Year
Limestone transportation distance=120
Units: km
Limestone transportation electricity use=Conveyor
consumption*Limestone transportation distance
Units: MWh/Year
electricity
use
per
ton-km*Limestone
Limestone use CO2 damages=Limestone use CO2 emissions*Unit damage cost CO2
Units: R/Year
Limestone use CO2 emission factor=439.71
Units: kg/ton
Limestone use CO2 emissions=(Limestone use CO2 emission factor*Limestone consumption)/kg to ton
Units: ton/Year
Limestone use damages (FGD)=Limestone use CO2 damages
Units: R/Year
Main material inputs=Al+Concrete+Steel
Units: ton/Year
Material transportation CO2e damages=(Unit damage cost CO2*Materials transportation CO2
emissions)+(Unit damage cost CO2*Materials transportation CO2e emissions (CH4))+(Unit damage cost
CO2*Materials transportation CO2e emissions (N2O))
Units: R/Year
Materials transportation CO2 emissions=C emission factor for diesel*"Diesel consumption (TJ)"*Diesel
oxidation factor *Gravimetric factor converting C to CO2
Units: ton/Year
Materials transportation CO2e emissions (CH4)=(CH4 emission factor for heavy duty diesel vehicles*"Diesel
consumption in litres (construction))*Global warming potential CH4
Units: ton/Year
Materials transportation CO2e emissions (N2O)=(N2O emission factor for heavy duty diesel vehicle*Diesel
consumption in litres (construction))*Global warming potential N2O
Units: ton/Year
MJ to kJ=1000
Units: MJ/kJ
MJ to TJ=1e+006
Units: MJ/TJ
- 317 -
N2O emission factor for heavy duty diesel vehicle=9.16667e-009
Units: ton/l
N2O emissions per MWh=1.15e-005
Units: ton/MWh
N2O steel production=Steel*Steel N2O embodiment
Units: kg/Year
Number of road trips=Main material inputs/Truck capacity
Units: Dmnl/Year
Plant construction global warming
transportation CO2e damages
Units: R/Year
damages=Construction
materials
CO2e
damages+Material
Plant operation global warming damages=Coal combustion CO2 damages+Coal combustion CO2e damages
(N2O)
Units: R/Year
Steel=Steel per MW*Capacity construction start
Units: ton/Year
Steel CH4 embodiment=100
Units: kg/ton
Steel CO2 embodiment=2710
Units: kg/ton
Steel N2O embodiment=100
Units: kg/ton
Steel per MW=50.721
Units: ton/MW
Steel production CO2e emissions=((CO2 steel production)+(CH4 steel production*Global warming potential
CH4)+(N2O steel production*Global warming potential N2O))/kg to ton
Units: ton/Year
Transportation distance (round trip)=100
Units: km
Truck capacity=31
Units: ton
Truck fuel consumption in l/km=0.35
Units: l/km
Unit damage cost CO2=INTEG (Change in CO2 damage cost, 109.89)
Units: R/ton
- 318 -
Waste disposal CO2 damages=Unit damage cost CO2*Waste disposal CO2 emissions
Units: R/Year
Waste disposal CO2 emissions=CO2 emission per MWh*Waste disposal electricity use
Units: ton/Year
Waste disposal CO2e damages (N2O)=Unit damage cost CO2*Waste disposal CO2e emissions (N2O)
Units: R/Year
Waste disposal CO2e emissions (N2O)=N2O emissions per MWh*Waste disposal electricity use*Global
warming potential N2O
Units: ton/Year
Waste disposal electricity use=Conveyor electricity use per ton-km*Dry waste Kusile*Distance travelled
Kusile waste
Units: MWh/Year
Waste disposal global warming damages=Waste disposal CO2 damages+Waste disposal CO2e damages
(N2O)
Units: R/Year
A9: Social cost sub-model equations
Capacity investment=(Planned investment in plant capacity table(Time))*Unit capital cost
Units: R/Year
Change in expected profitability=(Unit profitability-Expected profitability)/Time to adjust profit
Units: Dmnl/Year
Coal cost=Coal consumption*Unit coal cost
Units: R/Year
Coal-fuel cycle air pollution human health cost= Coal transportation air pollution health cost+FGD system
air pollution health cost +Plant construction air pollution health cost+Plant operation air pollution health
cost+Waste disposal air pollution health cost
Units: R/Year
Coal-fuel cycle cost of lost ecosystem services=Ecosystem services lost due to coal mining+Ecosystem
services lost due to plant construction & operation
Units: R/Year
Coal-fuel cycle externality cost of water use=Opportunity cost of water use (construction)+Opportunity cost
of water use (power generation)+Opportunity cost of water use in disposing Kusile's waste+Opportunity
cost of water use in FGD+Opportunity cost of water use in the New Largo colliery (coal mining)
Units: R/Year
Coal-fuel cycle externality costs=Coal-fuel cycle air pollution human health cost+Coal-fuel cycle cost of lost
ecosystem services+Coal-fuel cycle externality cost of water use"+"Coal-fuel cycle fatalities & morbidity
costs+Coal-fuel cycle global warming damage cost+Coal-fuel cycle water pollution externality cost
Units: R/Year
- 319 -
Coal-fuel cycle fatalities & morbidity costs=Fatalities & morbidity costs (coal mining)+Fatalities & morbidity
costs (construction)+Fatalities & morbidity costs (power generation)
Units: R/Year
Coal-fuel cycle global warming damage cost=Coal mining & transportation global warming damages+FGD
system global warming damages+Plant construction global warming damages+Plant operation global
warming damages+Waste disposal global warming damages
Units: R/Year
Coal-fuel cycle water pollution externality cost=Damage cost of sulphate pollution from coal
mining+Damage cost of sulphate pollution from Kusiles' raw material requirements
Units: R/Year
Cumulative capital cost escalated=INTEG (capital investment rate, 0)
Units: R
Cumulative private costs=INTEG (Private cost rate, 0)
Units: R
Cumulative PV air pollution cost=INTEG (PV air pollution cost, 0)
Units: R
Cumulative PV costs=Cumulative capital cost escalated+Cumulative PV fuel cost+Cumulative PV fixed O&M
costs+Cumulative PV variable O&M costs+Cumulative PV FGD operation cost
Units: R
Cumulative PV ecosystem services loss=INTEG (PV ecosystem services loss, 0)
Units: R
Cumulative PV externality cost=Cumulative PV air pollution cost+Cumulative PV ecosystem services
loss+Cumulative PV externality cost of water use+Cumulative PV fatalities & morbidity cost+Cumulative PV
global warming damages+Cumulative PV water pollution externality
Units: R
Cumulative PV externality cost of water use=INTEG (PV external cost of water use, 0)
Units: R
Cumulative PV fatalities & morbidity cost=INTEG (PV fatalities & morbidity cost, 0)
Units: R
Cumulative PV FGD operation cost=INTEG (PV FGD operation cost, 0)
Units: R
Cumulative PV fixed O&M costs=INTEG (PV fixed O&M costs, 0)
Units: R
Cumulative PV fuel cost=INTEG (PV fuel cost, 0)
Units: R
Cumulative PV global warming damages=INTEG (PV global warming damages, 0)
Units: R
- 320 -
Cumulative PV net electrity production=INTEG (PV net electricity production, 1)
Units: MWh
Cumulative PV revenue=INTEG ( PV revenue, 0)
Units: R
Cumulative PV variable O&M costs=INTEG (PV variable O&M costs, 0)
Units: R
Cumulative PV water pollution externality=INTEG (PV water pollution externality, 0)
Units: R
Cumulative revenue=INTEG (Revenue rate, 0)
Units: R
Electricity
price
table
([(2010,0)(2061,1500)],(2010,413.1),(2011,516.8),(2012,599.49),(2013,655.1),(2014,707.51),(2015,764.11),(2016,825
.24),(2017,891.26),(2018,892.15),(2019,893.04),(2020,893.93),(2021,894.83),(2022,895.72),(2023,896.62),(
2024,897.51),(2025,898.41),(2026,899.31),(2027,900.21),(2028,901.11),(2029,902.01),(2030,902.91),(2031,
903.82),(2032,904.72),(2033,905.62),(2034,906.53),(2035,907.44),(2036,908.34),(2037,909.25),(2038,910.1
6),(2039,911.07),(2040,911.98),(2041,912.89),(2042,913.81),(2043,914.72),(2044,915.64),(2045,916.55),(20
46,917.47),(2047,918.39),(2048,919.3),(2049,920.22),(2050,921.14),(2051,922.06),(2052,922.99),(2053,923
.91),(2054,924.83),(2055,925.76),(2056,926.68),(2057,927.61),(2058,928.54),(2059,929.47),(2060,930.4))
Units: R/MWh
Expected profitability=INTEG (Change in expected profitability, 0)
Units: Dmnl
Externality cost switch= 0
Units: Dmnl
FGD operation cost=FGD water cost+Limestone cost+Other FGD O&M costs year
Units: R/Year
FINAL TIME=2060
Units: Year
Fixed O&M costs year=Fixed O&M costs/Year
Units: R/Year
Gross electricity production=(((Functional capacity during construction*Plant operating hours)+(Desired
functional capacity after construction*Plant operating hours))*Load factor)*MWh/MW*h
Units: MWh/Year
INITIAL TIME=2010
Units: Year
Levelised air pollution cost=Cumulative PV air pollution cost/Cumulative PV net electrity production
Units: R/MWh
- 321 -
Levelised cost of energy=Levelised fuel cost+Levelised O&M costs+Levelised FGD operaton cost+Levelised
capital cost
Units: R/MWh
Levelised ecosystem services loss=Cumulative PV ecosystem services loss/Cumulative PV net electrity
production
Units: R/MWh
Levelised externality cost of energy=Levelised air pollution cost+Levelised ecosystem services
loss+Levelised fatalities & morbidity cost+Levelised global warming damages+Levelised water pollution
externality+Levelised water use externality
Units: R/MWh
Levelised fatalities & morbidity cost=Cumulative PV fatalities & morbidity cost/Cumulative PV net electrity
production
Units: R/MWh
Levelised global warming damages=Cumulative PV global warming damages/Cumulative PV net electrity
production
Units: R/MWh
Levelised social cost of energy=Levelised cost of energy+Levelised externality cost of energy
Units: R/MWh
Levelised water pollution externality=Cumulative PV water pollution externality/Cumulative PV net electrity
production
Units: R/MWh
Levelised water use externality=Cumulative PV externality cost of water use/Cumulative PV net electrity
production
Units: R/MWh
Net electricity production=Gross electricity production*1-Internal consumption rate
Units: MWh/Year
NPV (after tax)=NPV (before tax)*Tax rate factor
Units: R
NPV (before tax)=Cumulative PV revenue-Cumulative PV costs
Units: R
Planned
investment
in
plant
capacity
table
([(2010,0)(2060,2000)],(2010,800),(2011,800),(2012,800),(2013,800),(2014,1600),(2015,0),(2016,0),(2017,0),(2018,0),
(2019,0),(2020,0),(2021,0),(2022,0),(2023,0),(2024,0),(2025,0),(2026,0),(2027,0),(2028,0),(2029,0),(2030,0),
(2031,0),(2032,0),(2033,0),(2034,0),(2035,0),(2036,0),(2037,0),(2038,0),(2039,0),(2040,0),(2041,0),(2042,0),
(2043,0),(2044,0),(2045,0),(2046,0),(2047,0),(2048,0),(2049,0),(2050,0),(2051,0),(2052,0),(2053,0),(2054,0),
(2055,0),(2056,0),(2057,0),(2058,0),(2059,0),(2060,0))
Units: MW/Year
Plant private costs=Capacity investment+Coal cost+FGD operation cost+Fixed O&M costs year+Variable
O&M costs
- 322 -
Units: R/Year
Present value factor=((Conversion factor+Discount rate)^Year of cost(Time))
Units: Dmnl
Private cost rate=Plant private costs
Units: R/Year
Profits (after tax)=Profits (before tax)*Tax rate factor
Units: R
Profits (before tax)=Cumulative revenue-Cumulative private costs
Units: R
PV air pollution cost=Coal-fuel cycle air pollution human health cost/Present value factor
Units: R/Year
PV ecosystem services loss=Coal-fuel cycle cost of lost ecosystem services/Present value factor
Units: R/Year
PV external cost of water use=Coal-fuel cycle externality cost of water use/Present value factor
Units: R/Year
PV fatalities & morbidity cost=Coal-fuel cycle fatalities & morbidity costs/Present value factor
Units: R/Year
PV global warming damages=Coal-fuel cycle global warming damage cost/Present value factor
Units: R/Year
PV revenue=Revenue/Present value factor
Units: R/Year
PV water pollution externality=Coal-fuel cycle water pollution externality cost/Present value factor
Units: R/Year
Revenue=Electricity price table(Time)*Net electricity production
Units: R/Year
Revenue rate=Revenue
Units: R/Year
Social NPV (after tax)=NPV (after tax)-Cumulative PV externality cost
Units: R
Social NPV (before tax)=NPV (before tax)-Cumulative PV externality cost
Units: R
Tax rate factor= 0.72
Units: Dmnl
Time to adjust profit=1
- 323 -
Units: Year
Unit coal-fuel cycle externality cost=IF THEN ELSE(Gross electricity production<1e-006,0,Coal-fuel cycle
externality costs/Gross electricity production)
Units: R/MWh
Unit cost of prodution=IF THEN ELSE(Gross electricity production<=1e-006,0,Plant private costs/Gross
electricity production)
Units: R/MWh
Unit profitability=IF THEN ELSE(Time>2015,(IF THEN ELSE(Externality cost switch=1,(((Electricity price table
(Time)*Tax rate factor)-(Unit cost of prodution+Unit coal-fuel cycle externality cost))/Electricity price
table(Time)), (((Electricity price table(Time)*Tax rate factor)-Unit cost of prodution)/Electricity price
table(Time)))),0)
Units: Dmnl
Variable O&M costs=Plant water cost+Other variable O&M costs year
Units: R/Year
Year
of
cost
([(2010,0)(2060,60)],(2010,1),(2011,2),(2012,3),(2013,4),(2014,5),(2015,6),(2016,7),(2017,8),(2018,9),(2019,10),(2020
,11),(2021,12),(2022,13),(2023,14),(2024,15),(2025,16),(2026,17),(2027,18),(2028,19),(2029,20),(2030,21),(
2031,22),(2032,23),(2033,24),(2034,25),(2035,26),(2036,27),(2037,28),(2038,29),(2039,30),(2040,31),(2041,
32),(2042,33),(2043,34),(2044,35),(2045,36),(2046,37),(2047,38),(2048,39),(2049,40),(2050,41),(2051,42),(
2052,43),(2053,44),(2054,45),(2055,46),(2056,47),(2057,48),(2058,49),(2059,50),(2060,51))
Units: Dmnl
- 324 -
Was this manual useful for you? yes no
Thank you for your participation!

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

Related manuals

Download PDF

advertisement