Wells To Wire: Sarah Marie Jordaan

Sarah Marie Jordaan
Wells to Wire
Life Cycle Assessment of
Natural Gas-Fired Electricity
Wells to Wire
Sarah Marie Jordaan
Wells to Wire
Life Cycle Assessment of Natural Gas-Fired
Electricity
Sarah Marie Jordaan
Advanced and International Studies
The Johns Hopkins University
Washington, DC, USA
ISBN 978-3-030-71970-8 ISBN 978-3-030-71971-5 (eBook)

https://doi.org/10.1007/978-3-030-71971-5
© Springer Nature Switzerland AG 2021

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Wells to Wire presents an unbiased,
comprehensive examination of the state of
knowledge for life cycle assessments (LCAs)
of natural gas-fired electricity, covering a
suite of environmental impact categories. An
exploration of the life cycle environmental
impacts of gas-fired electricity is used to
introduce the field of LCA, advancements in
methods and data, and the limitations
thereof. Natural gas, particularly as a fuel for
electricity generation, serves as a dichotomy
within energy and environmental systems
analysis. While the cleanest burning fossil
fuel, it is not without impacts, making it an
excellent case study for introducing life cycle
assessment.
With an introduction to the field of LCA and a
focus on natural gas-fired electricity, the
reader is presented with the state of the art in
life cycle data and scientific debate related to
this product system. The author elucidates
data and methodological challenges inherent
to the field of LCA, exemplified using
published research and with new analyses.
The text explores how to conduct LCA,
describing the analysis from the simplified
perspective of a numerator and denominator.
With each chapter, the complexity of
undertaking a LCA of gas-fired power is
unravelled beyond a simple fraction to the
expansive network of infrastructure examined
in this type of research.
Students, instructors, LCA practitioners, and
energy professionals will benefit from not
only the introduction to data and methods,
but also this useful summary of the state of
the art in the field. Policymakers and the
interested public can learn more about the
implications of LCA results for
decision-support and the commentary about
the economics of natural gas and its role as a
bridge fuel. This book provides a useful
reference and a springboard for researchers
and experts interested in specializing in LCA,
natural gas, or both.
To my loved ones and to those who helped me
along the way.
Foreword
The USA has seen unprecedented and long-needed change in the power sector
over the last decade, with natural gas surpassing coal as the dominant generation
source and now rapidly succumbing to renewable energy sources as least-cost, high-
reliability sources for the energy transition. With such immense changes to the sector,
the need for solid, quantitative, analyses of impacts and trade-offs are more impor-
tant than ever, especially as the sun continues to rise on renewable power, energy
storage, and social justice. Dr. Jordaan’s Wells to Wire expertly demonstrates the
extent to which the now well-established field of life-cycle assessment continues to
hold immense potential to resolve these issues across multiple dimensions—span-
ning the environment to economic to social aspects of our energy infrastructure. This
book emerges presciently at a time when understanding the full dimension of impacts
associated with natural gas is something that is vital to explore allay concerns for the
environment and society expand in tandem.
Life cycle assessment has been around for around half a century and has method-
ologically anchored our understanding of the environmental impacts of the electric
sector. It is not unusual for analyses to greatly simplify the infrastructure effects of
energy development. Wells to Wire expertly asks the right questions and cleverly
wields large datasets to engender more accurate, much-needed analyses embodying
the innate complexity of natural gas systems. For example, Jordaan presents methods
and analysis of land-use for different energy sources with a focus on reclamation and
fuel supply infrastructure, elements that have not previously been considered. In this
book, life-cycle assessment has been skillfully used for some much-needed analyses
for consistent comparisons across energy types. This is not surprising given Jordaan’s
deep scholarship in land-based comparisons of energy systems, in part, which have
quantified the immense land requirements of fossil fuels compared to renewables.
Beyond land, the importance of understanding climate impacts of energy choices
has never been more important. Natural gas has been called a bridge fuel to a more
sustainable energy future. While true in the short-term, the quantitative assessments
embodied in Wells to Wire bring to question the environmental and economic sustain-
ability of natural gas serving as a bridge fuel on the longer term. A bridge should
span a physical obstacle without blocking passage underneath. While natural gas
ix
x Foreword
may serve as a baseload of reliability in the short term, it may ultimately fail in facil-
itating a rapid passage to a sustainable future over the long term. Decision-makers
and developers can confidently turn to these pages for an empirical justification of
their choices and a reconciliation of outcomes.
The transition from fossil fuel to renewable energy is both critically needed, and
needs tools to be developed and put into effect so that we can examine unbiased
assessments that can guide policy decisions as this transition occurs. While natural
gas continues to play a role in the power sector, it must be scaled-back and ultimately
replaced with clean energy altogether. This book gives insight into how to balance
challenging decisions in the face of natural gas as an increasingly dominant fuel for
electricity generation. It is only occasionally that a book like this one comes on the
scene to provide an end-to-end combination of scientific and economic analyses to
inform policy decisions about the future of an energy technology.
The challenges the world now faces in climate change are real. It is a new world
where natural gas must be fully accounted for. Dr. Jordaan’s Wells to Wire ends with
a salient recognition of the actors, entities, and industries who increasingly bear the
onus to prove the holistic sustainability of natural gas and, if unable, to decide what
role natural gas will serve.
Dan Kammen
James and Catherine Lau Distinguished Professor of Sustainability
University of California
Berkely, USA
Former Science Envoy, US State Department
Coordinating Lead Author, Intergovernmental Panel on Climate Change
Acknowledgements
The completion of this book would not have been possible without the support from
two Johns Hopkins University awards. A Discovery Award funded several years
of research on emissions from the life cycle of natural gas-fired electricity in the
USA. My collaborator and co-Principal Investigator Scot Miller and his doctoral
student Leyang Feng provided valuable input into the life cycle assessment completed
by Sakineh Tavakkoli under my supervision. While the results of that study were
not published in this book, valuable insights from our collaboration informed this
research. Several students in my research group—Energy Technology and Policy
Assessment (ETAPA)—provided data support in early stages of the work that was
funded by a Johns Hopkins Catalyst Award. In addition to Sakineh’s efforts on that
publication, Shreya Rangarajan and Mike Yu collected data on natural gas prices.
Andrew Ruttinger has been a big help on global research related to topics discussed
in this book, and while that analysis will be published elsewhere, the collaboration
also informed the content and vice versa. The work was improved by collabora-
tions with the National Renewable Energy Laboratory on land use (namely Garvin
Heath, Jordan Macknick and others) as well as with Hendrix College (Matt Moran and
Maureen McClung). The research on the greenhouse gas outcomes of liquified natural
gas was completed with colleagues from my prior collaborations at the University
of Calgary (Adebola Kasumu, Vivian Li, James Coleman, and Jeanne Liendo). My
collaborations on the water use of electricity generation with Lauren Patterson and
Laura Diaz Anadon provided a solid basis for the related discussion in this book. I was
fortunate to receive well-informed, constructive criticism that improved my research
during numerous invited seminars (namely Harvard, Princeton, George Washington
University, and the University of Maryland). Daniel Kammen provided the fore-
word, for which I am very grateful and honored. My friend and collaborator Rebecca
Hernandez provided a very needed helping hand during the last stretch, as did my
family (especially my mother). Finally, the American Center for Life Cycle Assess-
ment provides an annual forum through which I gain insights from other experts (for
example, from my colleagues from the National Energy Technology Laboratory).
While I note these organizations and collaborators, the research results and analysis
presented here are my own and do not represent their views (except where they are
directly cited).
xi
Contents
1 An Introduction to Life Cycle Assessment of Natural Gas-Fired

Electricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Global Evolution of Natural Gas Production and Use in Electricity
Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Life Cycle Assessment as a Framework to Understand
Environmental Impacts of Gas-Fired Power . . . . . . . . . . . . . . . . . . . . . . . . . 3
LCA Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
What Will Be Covered in This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2 LCA Framework, Methods, and Application . . . . . . . . . . . . . . . . . . . . . . 13
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
General Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Goal and Scope Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Inventory Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Impact Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Interpretation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3 The Denominator: Natural Gas Production, Throughput,
and Electricity Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Infrastructure Lifetime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Natural Gas Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Facility Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Power Plant Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
xiii
xiv Contents
4 Life Cycle Impact Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Impact Categories and Characterization Models . . . . . . . . . . . . . . . . . . . . . 46
A Focus on Improving Individual Impact Categories . . . . . . . . . . . . . . . . . 51
Greenhouse Gas Emissions and Other Air Pollution . . . . . . . . . . . . . . . 51
Water Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Quantification of Life Cycle Land Use Impacts . . . . . . . . . . . . . . . . . . . 57
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5 Global Markets and Competitiveness of Gas-Fired Power . . . . . . . . . . 67
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Levelized Cost of Electricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
The Economics of Natural Gas Production . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Production Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Natural Gas Markets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Additional Costs Associated with LNG . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Implications of LNG for Life Cycle Assessment of Gas-Fired Power . . . . 75
Liquefaction to Power Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Electricity Generation to End Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
6 Tackling Uncertainty Across the Life Cycle of Gas-Fired Power . . . . 85
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Uncertainty in LCA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Data Sources, Precision, and Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Probabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Approaching Uncertainty in LCA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Monte Carlo Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Break-Even (Parametric) Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Scenario Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
7 Natural Gas as a Bridge Fuel? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Compatibility with Paris Agreement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Political, Economic, and Technological Uncertainties . . . . . . . . . . . . . . . . . 105
Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Chapter 1
An Introduction to Life Cycle Assessment
of Natural Gas-Fired Electricity
Abstract The electric sector is undergoing a transition to a more environmentally

sustainable energy system, and electricity generated from natural gas has played
a role to date as a lower carbon alternative to coal. While it combusts the cleanest
burning fossil fuel, natural gas-fired power is not without environmental impacts. Life
cycle assessment offers an internationally standardized methodology to quantify the
environmental burdens from materials extraction to waste disposal. Environmental
impacts of gas-fired power can thus be robustly examined using this approach. The
primary goal of the book is to introduce the reader to the field of life cycle assessment
with an unbiased, comprehensive examination of the state of knowledge of natural
gas-fired electricity. Due to the irreducible variability of infrastructure types across
an expansive network associated with this product system, the focus of the analyses
laid out in this book is on the fuel cycle and electricity generation. Wider systems
boundaries that include upstream material extraction and waste disposal are consid-
ered as applicable. This chapter introduces the premise of the book and lays out the
subsequent chapters for the readers. Insights are developed on not only how to under-
take a life cycle assessment of gas-fired power, but also how to assess uncertainty,
emerging markets and their implications and the role of natural gas as a bridge fuel.
Introduction
Natural gas is the cleanest burning fossil fuel, but the full scope of its environmental
implications for the electric sector remains unclear. Displacing coal with natural gas
can result in reduced greenhouse gas emissions from the power sector, yet the fugitive
methane emissions from the natural gas supply chain may reverse such benefits if
left unmanaged. Similarly, switching from coal to gas in the electric sector results
in a net reduction of water consumption. This reduction does not capture the full
picture: environmental benefits are not experienced by all regions impacted by such
a transition—the spatial patterns of water used for hydraulic fracturing over time
have not yet been thoroughly investigated across shale gas producing regions. These
two examples shed light on how the overall environmental impacts of the sector
remain unresolved, amongst scientists, policymakers, and the broader public.
© Springer Nature Switzerland AG 2021 1

S. M. Jordaan, Wells to Wire,
https://doi.org/10.1007/978-3-030-71971-5_1
2 1 An Introduction to Life Cycle Assessment of Natural Gas …
Life cycle assessment (LCA) is an analytical tool that can be used to examine the
environmental impacts of a product or process from cradle to grave. New research
has been initiated in recent years to understand the life cycle environmental implica-
tions of changes in the natural gas and power sectors; yet results remain disputed in
many cases due to challenges with uncertain data, rapidly evolving operations and
technologies, and assumptions. The most prominent example in the past decade is
the influence of methane leaks from natural gas production systems on the life cycle
of natural gas-fired electricity. One notable study in 2011 suggested these upstream
emissions may reverse the environmental benefits of electricity generated from coal
compared to natural gas [1]. These results were quickly disputed [2] with subsequent
studies noting the uncertainty in upstream emissions and how this reversal would
only hold true if emissions were substantially underestimated [3]. More recently,
Alvarez et al. [4] found that 2015 emissions may indeed have been underestimated,
suggesting environmental benefits may be overestimated (or even not realized) if
the global warming potential is considered over 20-year time horizons. Tanaka et al.
[5] suggest that—even if that were the case under some conditions—the metrics
presently used for comparing different greenhouse gases are limited and using more
robust metrics affirms the environmental benefits of natural gas relative to coal.
The goal of this book is to introduce the field of LCA using an unbiased,
comprehensive examination of the state of knowledge of natural gas-fired electricity,
covering a suite of environmental impact categories. To meet this goal, the book
has three objectives: (1) to introduce readers to the field of LCA using natural gas-
fired electricity as a case study; (2) to delineate the state-of-the-art in life cycle
data, research, and scientific debate related to natural gas-fired electricity, and (3) to
elucidate data and methodological challenges inherent to the field of LCA, exem-
plified through the case of natural gas-fired power. We will explore how to conduct
LCA, describing the analysis from the perspective of a numerator and denominator.
Natural gas, particularly as a fuel for electricity generation, creates a dichotomy
within energy and environmental systems analysis. While the cleanest burning fossil
fuel, it is not without impacts across supply chains, making it an excellent case study
for introducing life cycle assessment. While many books address either natural gas,
natural gas-fired electricity, LCA of natural gas vehicles, or LCA as a field in and of
itself [6–10], this book combines an introduction to the field of LCA using natural
gas-fired electricity as an example.
In this first chapter, the evolution and growth of the natural gas sector will be
explored to illuminate the significance of this book. Basic concepts and terms asso-
ciated with LCA will be introduced, including the four methodological stages (goal
and scope definition, inventory analysis, impact assessment, and interpretation). The
natural gas supply chain through use in power generation will be explained, following
a natural gas molecule from extraction from the ground through combustion in a
power plant and the delivery of electricity to end use. Finally, each chapter of the
book will be summarized to provide the reader with the big picture of the book.
Global Evolution of Natural Gas Production and Use … 3
Global Evolution of Natural Gas Production and Use

in Electricity Generation
The natural gas sector is undergoing global changes, with the recent expansion of
liquefied natural gas trade (LNG) and the unlocking of substantial shale gas reserves.
The combination of hydraulic fracturing and vertical drilling represents an innovation
that has unlocked enormous shale reserves, leading to the 34 trilling cubic feet of
natural gas produced in 2019—the highest annual production recorded in the United
States and 9% greater than domestic consumption [11]. Global demand for natural
gas in the electric sector has been growing in recent years, exemplified by the coal-to-
gas transition in the sector in the United States. The share of U.S. electricity generated
from natural gas increased from 19% in 2005 to 24% and 33% in 2010 and 2015
respectively [12] initiated by the shale boom and the subsequent shift from coal to
gas in the electric sector.
Up until the 2019–2020 pandemic, natural gas was expanding globally via lique-
fied natural gas (LNG) trade and the oversupply had created global glut. The market
had already experienced a slowdown in 2019, which was confounded by an addi-
tional demand shock with COVID-19 [13]. The International Energy Agency (IEA)
expects a progressive recovery to pre-crisis levels in 2021 and, despite 75 billion
m3 of lost growth from 2019 to 2025, foresees an average growth rate of 1.5%.
While industrial growth is the primary contributor, even the lowest emitting scenario
anticipates medium-term growth in natural gas demand from the power sector up
until 2030. This IEA scenario—the Sustainable Development Scenario (SDS)—is
fully aligned with the Paris agreements goals [14]. Natural gas consumption in this
scenario increases to an annual average rate of 0.9% until the end of the 2020s. After
2030, however, declines are anticipated in the SDS, in contrast to the other scenarios
which all anticipate steady growth. Although natural gas-fired electricity generation
declines, capacity grows compared with today as gas expands its role as a provider
of power system flexibility.
Life Cycle Assessment as a Framework to Understand

Environmental Impacts of Gas-Fired Power
Both renewables and natural gas are set to play a substantial role in the future electric
grid. Coal without carbon capture and storage (CCS) comes under fire due to impacts
to the environment. Although natural gas is expected to grow in even the lowest carbon
scenarios because it is the cleanest burning fossil fuel, it remains a fossil fuel (if
extracted from oil and gas reservoirs). As a result, overall impacts of gas-fired power
are especially scrutinized and require balanced and unbiased technical analysis. LCA
provides a standardized, broadly accepted method to complete balanced, yet rigorous
examinations of the impacts. Several chapters of this book—notably Chap. 4—will
include such comparisons to illustrate useful application of the method.
LCA methodology has two key strengths that serve to better understand and
mitigate the environmental impacts of natural gas in the electric sector, providing
a method to produce unbiased results [9]. First, the method is internationally stan-
dardized through ISO, providing a robust and comprehensive way to understand the
cradle-to-grave impacts of products and processes. LCA strives to deliver superior
environmental outcomes: once the impacts have been quantified, mitigation options
can be identified and implemented. Second, radically different product systems that
deliver similar services can be compared. Importantly for this book, electricity gener-
ated from natural gas can be compared to that generated from coal, nuclear, or
renewables.
Like a LCA of any other product or process, analyses of gas-fired power must
grapple with the field’s limitations [9]. Concerns about “hired guns” modifying results
to endorse their products has been an ongoing issue and certainly true for the fossil
fuel sector. Similarly, LCA results can be used inappropriately in product marketing;
for example, results promoting a product may be showcased while hiding uncertain-
ties or focusing on favorable results. The limitations of LCA are rooted in inconsis-
tencies in system boundaries, inputs, functional units. Recent harmonization studies
found large differences in system boundaries and inputs in LCAs of gas-fired power
[15, 16]. Additionally, even changes in the functional unit from electricity generated
to electricity delivered (the latter including the transmission and distribution (T&D))
can result in substantial changes in results if inefficiency T&D is high enough [17].
These criticisms are truly exemplified by the case of natural gas since assumptions
are often disputed between industry and other stakeholders. Methane leaks from
natural gas production systems provide a good example—their magnitude is often
downplayed by industry, questioned by scholars, and disputed by environmentalists
[3, 18–20]. The analysis of natural gas production and electricity generation using
LCA is particularly challenged in this regard, as it involves a complex network of
expansive infrastructures. Leaks may occur across hundreds of thousands of different
components, ranging from production sites, through pipelines, midstream sites and
infrastructures, with further inefficiencies throughout the electric grid. Even the most
detailed studies to date report different populations of infrastructure types for the fuel
cycle of natural gas power (Table 1.1). The fuel cycle encompasses the part of the
life cycle of gas-fired electricity that is directly related to the extraction and transport
of natural gas up to the power plant.
LCA Methodology
Due to the complexity of the infrastructure involved with the life cycle of gas-fired
electricity in addition to vested interests, the standardized LCA methodology holds
importance in developing robust results. The method involves four stages, as outlined
below:
LCA Methodology 5
Table 1.1 Example infrastructure counts published for the natural gas supply chain, referred to
here as the fuel cycle
Fuel cycle stage Heath et al. [21] EPA GHGI Tavakkoli et al. [23] Unit
[22]
Production wells 500,000 (natural 251,076 84,155 (hydraulic Well
gas) (hydraulic fracturing)
fracturing)
Gathering 200,000 403,863 160,126 Mile
pipelines
Gathering stations 10,000 5276 2550 Station
Storage facilities 400 349 461 Underground
(compressor facility unless
stations) noted
Processing plants 580 plants 667 plants 801 plants Plant
Transmission 10,000 1844 stations 1720 stations Station unless
stations compressors noted
Transmission 320,000 301,164 229,296 Mile
pipelines
1. Goal and Scope definition. The product and purpose of the LCA are deter-
mined in this LCA stage. Practitioners select the system boundaries and the
functional unit, the latter serving as the basis of comparison for the analysis (the
denominator).
2. Inventory analysis. The second stage of a LCA involves the compilation of
relevant data inputs and outputs. For example, resource inputs and environmental
releases will be quantified for each part of a product’s life cycle.
3. Impact assessment. The environmental impacts associated with the inputs,
outputs, and emissions are calculated and, where possible, aggregated to indices.
4. Interpretation. The interpretation stage is geared towards informing decisions.
The results of the LCA are evaluated relative to the goal and scope definition.
Uncertainty assessment is a critical component of the interpretation, to inform
decision-makers of how results may change under different conditions.
A life cycle model is developed while implementing the methodology. It is impor-
tant to distinguish the LCA model (Fig. 1.1a) from the LCA method (Fig. 1.1b). The
LCA model provides guidance on how practitioners complete the inventory anal-
ysis, primarily in the form of a process flow diagram. A process flow diagram is
a schematic that describes the key processes involved in the life cycle. Figure 1.1a
presents an example of a generic process flow diagram, while a detailed schematic
developed specifically for gas-fired power will be presented in Chap. 2.
The LCA model is developed iteratively by implementing the four stages of the
LCA method (Fig. 1.1b), with a detailed process flow diagram being developed in
the Goal and Scope Definition stage. A critical component of this first stage of LCA
is the selection of a functional unit. The functional unit is defined by the Interna-
tional Organization for Standardization (ISO) as: the quantified performance of a
product system, process, or organization for use as a reference unit. As the product
of focus for this book is LCA of gas-fired electricity, the functional unit is typically
selected as a unit of electricity generated, such as kilowatt-hour (kWh) or Megawatt-
hour (MWh). Care must be taken as the analysis proceeds in appropriate selection
and characterization of the functional unit. While it may appear straightforward,
substantial differences have been noted depending on whether the functional unit is
electricity generated or delivered [17].
Materials aquisition Products

Materials
Formulation, processing,
Co-products
and manufacturing
Transportation and
Water pollution
distribution
Energy
Use of product Air pollution
Recycling, if applicable Solid waste

Natural resources
(water, land, etc)
Waste management Other environmental
impacts (e.g. biodiversity)
(a)
1. Goal and scope

Performance improvements
• Product development and
2. Inventory 4. Interpretation improvement
Data • Strategic planning
analysis of results
• Policy implementation
• Marketing
3. Impact
assessment
(b)
Fig. 1.1 a, b LCA models and methods. a Life cycle model (adapted from Fava et al. [24]).
b LCA procedure, including the four phases of the LCA methodology (adapted from International
Organization for Standardization (ISO) 2006 [25])
What Will Be Covered in This Book 7
What Will Be Covered in This Book
This book centers on the part of the life cycle that includes the natural gas fuel
cycle up through electricity generation. The analyses may consider additional lique-
faction stages, depending on the scope of the analyses being discussed. Electricity
generated from LNG imports will be examined in Chap. 5 (Fig. 1.2).
To envision the life cycle of natural gas-fired electricity, the reader is invited to
imagine the path of a natural gas extracted from the ground. When extracted, natural
gas is typically a mixture, the majority of which is methane (Table 1.2). It may
be extracted through a well that produces primarily gas or primarily oil; the latter
is referred to as associated gas and is viewed as a co-product. Depending on the
composition of the hydrocarbons produced from the well site, some purification and
separation may be required via processing. Even the heating value of the natural gas—
representing the amount of energy that is released upon combustion—is variable,
ranging from 42–55 Megajoules per kilogram (MJ/kg) [26] (Table 1.2).
Gas production
and supply chain Processing Midstream infrastructure End use
(e.g. storage, compressor
Extraction stations, liquefaction)
Ocean transport
(if applicable)
Fig. 1.2 Simplified representation of the life cycle of natural gas products
Table 1.2 Typical compositions of natural gas [27]. Note that some estimates may be based on
volume; for example, Heath et al. used an aggregate U.S. national estimate of 78.8% methane by
volume
Components Chemical composition Range (mole %)
Methane CH4 70–90
Ethane C2 H6 0–20
Propane C3 H8 0–20
Butane C4 H10 0–20
Carbon dioxide CO2 0–8
Oxygen O2 0–0.2
Nitrogen N2 0–5
Hydrogen sulfide H2S 0–5
Rare gases A, He, Ne, Xe Trace
Once the natural gas has been extracted from the ground, it is transported through
pipelines to a processing facility (if required) where impurities are removed and/or
products are separated. The rest of the infrastructure through which the gas travels
may include underground storage, one or more compressor stations (which increase
the pressure of the gas and enable its flow through pipelines), and other midstream
infrastructure such as liquefaction for peak shaving (storage to regulate demand). If
the natural gas is to be exported to an international market, then it may be liquefied for
ocean transport, then regasified and transported once it reaches the target destination.
The target destination depends on the end use, and here we examine electricity
generation. As a result, the natural gas would end up in a power plant in the region
of interest to the practitioner (defined in the goal and scope definition). In this book,
several examples of how the choice of region can influence LCA results of gas-fired
electricity will be examined.
The analyses presented here focus specifically on the quantification of the fuel
cycle and electricity generation of gas-fired power, unless noted otherwise (such as in
Chap. 4, where impacts are harmonized and compared to other electricity generation
sources). For greenhouse gas emissions, for example, ignoring the material extraction
stage for renewables like solar and wind would falsely result in zero emissions [15].
For natural gas, these emissions contribute less than 1% of emissions; however,
consistency should be applied in comparative analyses. While the focus of this book
is on the operations of the fuel cycle and electricity generation, to help readers better
characterize a highly complex system more accurately, practitioners are encouraged
to develop the scope of their analyses comprehensively. In particular, consistency
and transparency is crucial component of credible, comparative LCAs.
Summary
With each chapter, the complexity of undertaking a LCA of gas-fired power will be
unraveled beyond a simple fraction to the expansive network of infrastructure that
must be considered. Readers of this book may be graduate students or specialized
researchers interested in either LCA or the natural gas sector. LCA practitioners and
professionals in the energy sector will benefit from not only the data and methods
presented, but also the summary of the state-of-the-art research in the field. As readers
may find particular chapters more useful than others depending on their interests,
content of each chapter is described in the following summary.
After the present introductory chapter, Chap. 2 comprehensively covers the four
methodological stages of LCA in depth (goal and scope definition, inventory analysis,
impact assessment, and interpretation), relating the method to the example of natural
gas. Readers interested in learning the basics of LCA will strongly benefit from this
chapter, or those who are interested in understanding how LCA methods are applied to
natural gas specifically. While broad LCA concepts are explored, additional details
regarding applying these concepts to natural gas are discussed. The chapter has
specialized sections related to the choice of functional unit, allocation methods, and
Summary 9
available impact assessment methods (as well as a discussion about how to choose
from the available methods) [28–31].
Chapter 3 examines how to quantify the energy flows through the supply chain
(also referred to as throughput), from how much natural gas can be extracted from a
well, the natural gas throughput that travels through each infrastructure type, leading
up to the amount of natural gas combusted at the power plant and the amount of
electricity generated. Critical to the implementation of LCA is the choice of func-
tional unit, which serves as the denominator in the final calculations. For the case of
gas-fired power, the functional unit is the unit of electricity generated (e.g. Megawatt-
hour or MWh). Methods to quantify estimated ultimate recovery from natural gas
wells will be reviewed, including a compilation of published data. Different infras-
tructure assets are examined along with the throughput of natural gas through each
facility type. Last, data specific natural gas-fired generation is used as an example
to demonstrate the calculations used to convert each life cycle phase to the basis of
a functional unit (unit of electricity generated) for the final LCA results. The vari-
ability in efficiency across electricity generation technologies and heat rates across
generating units are reviewed, using real-life data examples from power plants in the
U.S.A and aggregate country-level statistics internationally.
Life cycle inventories and impact assessment methods used in LCAs of gas-
fired power will be covered in Chap. 4. Across impact categories, natural gas faces
many of the same challenges as other product systems in LCA: data. Due to the
high levels of variability and uncertainty inherent to the natural gas sector, impact
results remain disputed amongst scientists and policymakers alike. The typical life
cycle approach of characterizing results into a suite of aggregate impact categories
is therefore less valuable for decision-makers in this particular sector, due to high
levels of controversy surrounding particular assumptions employed in LCA models.
As a result, deep dives into specific life cycle impacts and their inventories will be
taken for greenhouse gas emissions, air pollution, water consumption, and land use.
Recent analyses for each impact will be reviewed, along with a critical analysis of the
state of data and impact methods. Specific attention will be paid to highlight where
additional methodological and data improvements are required.
Chapter 5 will cover the evolution of natural gas economics and global markets,
and how these factors influence LCA results. Shale gas reserves have been unlocked
across the planet with the combination of horizontal drilling and hydraulic frac-
turing. At the same time, the global trade of natural gas has been increasing with
a burgeoning liquefied natural gas (LNG) industry. These twin developments have
resulted in new political, economic, and environmental dynamics across the supply
chain from the exporting to importing nations [32]. The addition of new processes
in the natural gas supply chain (e.g. liquefaction, ocean transport, and regasification)
can have substantial impacts on life cycle results. The economics of natural gas will
be reviewed followed by a discussion of present and emerging markets for natural
gas (with emphasis on LNG) and their implications for LCA.
Uncertainty is a critical component of LCA methodology for practitioners because
it enables decision-makers to understand how impacts may change under different
conditions. Chapter 6 will include an overview of the importance of uncertainty
assessment in LCA, followed by a review of methods frequently employed by LCA

practitioners, and ending with examples specific to natural gas-fired electricity. The
chapter will provide an introduction to uncertainty assessment, starting with sensi-
tivity analysis followed by a detailed discussion of the most commonly used uncer-
tainty assessment tools in LCA [32]. The methods will be introduced along with a
review of how they are employed in LCAs of gas-fired power [8].
The final chapter will address the overall sustainability of natural gas and its
potential role as a bridge fuel, using LCA as a framework for the discussion. The
conclusion will not only summarize the book’s overall findings in the context of
natural gas as a bridge fuel, but it will also revisit remaining research areas that
readers may want to pursue. Only one end use will be covered in this manuscript,
for example, but results for the fuel cycle may be useful for analyses that focus on
other end uses.
Demand for natural gas by the electric sector is presently anticipated to grow over
the coming decades, pointing to a need to understand the environmental implications.
This book presents a comprehensive overview of LCA as a systematic tool to build
this understanding, illuminating present findings and areas of inquiry.
References
1. Howarth, R.W., Santoro, R., Ingraffea, A.: Methane and the greenhouse-gas footprint of natural
gas from shale formations. Clim. Change 106, 679–690 (2011)
2. O’Sullivan, F., Paltsev, S.: Shale gas production: potential versus actual GHG emissions. MIT
Joint Program on the Science and Policy of Global Change, p. 234 (2012)
3. Brandt, A.R., Heath, G.A., Kort, E.A., O’Sullivan, F., Ptron, G., Jordaan, S.M., Tans, P., Wilcox,
J., Gopstein, A.M., Arent, D.: Methane leaks from North American natural gas systems. Science
343, 733–735 (2014)
4. Alvarez, R.A., Zavala-Araiza, D., Lyon, D.R., Allen, D.T., Barkley, Z.R., Brandt, A.R., Davis,
K.J., Herndon, S.C., Jacob, D.J., Karion, A.: Assessment of methane emissions from the US
oil and gas supply chain. Science eaar7204 (2018)
5. Tanaka, K., Cavalett, O., Collins, W.J., Cherubini, F.: Asserting the climate benefits of the
coal-to-gas shift across temporal and spatial scales. Nat. Clim. Change 9, 389–396 (2019)
6. Lyons, W.C., Plisga, G.J.: Standard Handbook of Petroleum and Natural Gas Engineering.
Elsevier (2011)
7. Nigge, K.: Life cycle assessment of natural gas vehicles: development and application of
site-dependent impact indicators, vol. 6. Springer Science & Business Media (2012)
8. Matthews, H.S., Hendrickson, C.T., Matthews, D.H.: Life cycle assessment: quantitative
approaches for decisions that matter. Retrieved June 2018, 1 (2018)
9. Tillman, A.M., Baumann, H.: The Hitchhikers guide to LCA. Studentlitteratur AB, Lund,
Sweden (2004)
10. Schenck, R., White, P.: Environmental Life Cycle Assessment: Measuring the Environmental
Performance of Products. American Center for Life Cycle Assessment Vashon, Washingto
(2014)
11. EIA Natural gas explained: Where our natural gas comes from. Energy Informainistra-
tion (EIA). https://www.eia.gov/energyexplained/natural-gas/where-our-natural-gas-comes-
from.php. Accessed on 14 Jan 2021
12. Energy Information Administration: Form EIA-923 detailed data with previous form data.
Accessed on Jan 08 2019
References 11
13. IEA Gas 2020. International Energy Agency (IEA) (2020)

14. International Energy Agency, (Sustainable Development Scenario, World Energy
Outlook. https://www.iea.org/reports/world-energy-model/sustainable-development-sce
nario. Accessed on 3 Mar 2020
15. O’Donoughue, P.R., Heath, G.A., Dolan, S.L., Vorum, M.: Life cycle greenhouse gas emis-
sions of electricity generated from conventionally produced natural gas: systematic review and
harmonization. J. Ind. Ecol. 18, 125–144 (2014)
16. Heath, G.A., O’Donoughue, P., Arent, D.J., Bazilian, M.: Harmonization of initial estimates of
shale gas life cycle greenhouse gas emissions for electric power generation. Proc. Natl. Acad.
Sci. USA 111, 3167 (2014)
17. Surana, K., Jordaan, S.M.: The climate mitigation opportunity behind global power transmis-
sion and distribution. Nat. Clim. Change 9, 660–665 (2019)
18. Hauck, M., Steinmann, Z., Laurenzi, I.J., Karuppiah, R., Huijbregts, M.: How to quantify
uncertainty and variability in life cycle assessment: the case of greenhouse gas emissions of
gas power generation in the US. Environ. Res. Lett. 9, 074005 (2014)
19. Tollefson, J.: Methane leaks erode green credentials of natural gas. Nature 493, 12 (2013)
20. Atherton, E., Risk, D., Fougere, C., Lavoie, M., Marshall, A., Werring, J., Williams, J.P.,
Minions, C.: Mobile measurement of methane emissions from natural gas developments in
northeastern British Columbia, Canada. Atmos. Chem. Phys. 17 (2017)
21. Heath, G., Warner, E., Steinberg, D., Brandt, A.: Estimating US Methane Emissions from the
Natural Gas Supply Chain. Approaches, Uncertainties, Current Estimates, and Future Studies
(2015)
30. EPA.: U. S. Inventory of US greenhouse gas emissions and sinks: 1990–2017 (2019)
23. Tavakkoli, S., Feng, L., Miller, S., Jordaan, S.M.: The implications of generation efficiencies
and supply chain leaks for the life cycle greenhouse gas emissions of natural gas-fired electricity
in the United States (under review)
24. Fava, J.A., Denison, R., Jones, B., Curran, M.A., Vignon, B., Selke, S.: A technical framework
for life cycle assessment: workshop report; August 18–23, 1990; Society of Environmental
Toxicology and Chemistry (SETAC): (1991)
25. International Organization for Standardization, (ISO). 14040/44 International standard: Envi-
ronmental Management–Life Cycle Assessment. International Organisation for Standardiza-
tion, Geneva, Switzerland (2006)
26. Hauschild, M.Z., Rosenbaum, R. K., Olsen, S.I.: Life Cycle Assessment. Springer, Berlin
(2018)
27. Jolliet, O., Saadé-Sbeih, M., Shaked, S., Jolliet, A., Crettaz, P.: Environmental Life Cycle
Assessment. CRC Press (2015)
28. Klöpffer, W., Grahl, B.: Life Cycle Assessment (LCA): A Guide to Best Practice. Wiley (2014)
29. Curran, M.A.: Life Cycle Assessment Handbook: a Guide for Environmentally Sustainable
Products. Wiley (2012)
30. World Nuclear Association Heat Values of Various Fuels. https://www.world-nuclear.org/inf
ormation-library/facts-and-figures/heat-values-of-various-fuels.aspx. Accessed on 14 Jan 2021
31. Siažik, J., Malcho, M.: Accumulation of primary energy into natural gas hydrates. Proc. Eng.
192, 782–787 (2017)
32. Kasumu, A., Li, V., Coleman J.W., Liendo J., Jordaan, S.M.: Country-level life cycle assessment
of liquefied natural gas trade for electricity generation. Environ. Sci Tech. 52(4), 1735–1746
(2018)
Chapter 2
LCA Framework, Methods,
and Application
Abstract Life cycle assessment (LCA) offers a robust, internationally standardized

method for determining environmental costs and benefits of product systems. For
credible and consistent results, the broadly accepted method should be applied with
transparent reporting of data, assumptions, and results. In this chapter, the four stages
of the LCA method—goal and scope definition, inventory analysis, impact assess-
ment, and interpretation—are explained in detail, using gas-fired power to exemplify
how to develop a process-based LCA.
Introduction
Life cycle findings can result in reversals of environmental conclusions about prod-
ucts or the discovery of potential environmental improvements [1]. For example,
potential negative consequences of early lead batteries for use in electric vehicles
(EV) were discovered using a life cycle approach before widespread commercializa-
tion in the 1990s [2]. With powerful economic forces at play in the sectors involved
across the life cycle of products, the concern over the misuse of LCA for corporate
gain has been an ongoing challenge for the field [3]. Ever since the development
and broader use of LCA—with the first being completed by Coca-Cola company in
the form of a Resource and Environmental Profile Analysis (REPA)—the methods
have been a contentious point [3]. The contention was not because of Coca-Cola
specifically, but more so because the choice of methods (and assumptions) can yield
biased results that can be used incorrectly for marketing purposes by corporate and
other interests. Growing concerns about spin doctors and hired guns (trained experts
capable of manipulating the data to promote specific products and outcomes) led to
the need for standardizing methods, particularly as LCAs were applied more broadly
to powerful sectors such as energy and fossil fuels. As a result, the methodological
approach has undergone a rigorous standardization process through the International
Organization for Standardization (ISO) [4].
LCA follows a four-step methodology that includes (1) goal and scope definition,
(2) inventory analysis, (3) impact assessment and (4) interpretation. The International
Organization for Standardization (ISO) describes each of these four iterative steps,

14 2 LCA Framework, Methods, and Application
which will be discussed in-depth in this chapter [3, 4]. The standardization process led
to a robust methodology, a specific cookbook which holds practitioners accountable
by requiring comprehensive analysis and transparency to certify their LCAs. While
there are several types of life cycle approaches, such as economic input output LCAs,
we focus on process based LCAs which examine specific processes within a defined
system boundary. Such a scope permits a comprehensive examination of a complex
fuel cycle and life cycle throughout this book. This chapter will provide an overview
of the ingredients for each life cycle phase, combined into a recipe for each, to
simplify the process for those who are new to the field. Each stage will be explained
relative to gas-fired electricity.
General Framework
The central theme of this chapter will be to cover each stage of the LCA approach, as
internationally accepted via the International Organization for Standardization (ISO),
an organization that seeks broad consensus to develop standards (Fig. 2.1). The first
ISO standard for LCA was published in 1996 with a later revision in 2006. While
clearly there was a dire need for consensus, the standardization process received early
criticisms, importantly that it was premature and risked stagnating the development
of LCA since it was early in its adoption [5, 6].
Such criticism holds true in some regards and are yet to be fully resolved. For
example, data availability and science has progressed in leaps and bounds since
1996 (and even 2006) so many methods currently being developed were not even
conceived when the standards were initially formulated. That said, the ISO frame-
work was developed for flexibility (with ongoing updates and amendments) and
met important goals of ensuring the accountability of companies in publishing their
LCAs. Regardless, many organizations do not complete certification and practitioners
should always keep a critical eye on the methods and assumptions being employed.
1. Goal and scope

Performance improvements
• Product development and
2. Inventory 4. Interpretation improvement
Data • Strategic planning
analysis of results
• Policy implementation
• Marketing
3. Impact
assessment
(b)
Fig. 2.1 LCA methodology (developed from ISO 2006)

General Framework 15
The four stages of methodology will be covered in depth, including [3, 4, 7–10]:
1. Goal and Scope Definition. The product and purpose of the LCA are deter-
mined in this stage. Practitioners select the product system, delineate the system
boundaries and determine the functional unit, the latter serving as the basis of
comparison for the analysis. The functional unit in LCA is the factor used to
normalize the environmental impacts. For example, one kilowatt-hour (kWh) or
one Megawatt-hour (MWh) may be used as the functional unit for a life cycle
study of greenhouse gas emissions from natural gas-fired electricity genera-
tion, where results will be presented in grams of CO2 equivalent released for
each unit of electricity generated (e.g., 1 kWh). Alternatively, if the study were
focused on natural gas vehicles, a more appropriate functional unit would be a
passenger-mile travelled.
2. Inventory Analysis. The second stage of LCA involves the compilation of rele-
vant data inputs and outputs and quantifying them on the basis of a functional
unit. For example, resource inputs and environmental releases will be quantified
for each part of a product’s life cycle. An inventory analysis would require the
inputs of each process within the system boundaries and relating them to the
functional unit by considering the material and energy flows across the system.
For example, the emissions from the natural gas supply chain are related to the
functional unit by the amount of natural gas required by the power plant for
each unit of electricity generated (also known as a heat rate of a power plant,
which may be expressed as mmBTU of gas consumed for MWh generated, for
instance).
3. Impact Assessment. The goal of the third stage, impact assessment, is to under-
stand and evaluate the significance and magnitude of environmental impacts
caused by each of the life cycle stages within the system boundaries for the
product in question. The environmental impacts associated with the inputs,
outputs, and emissions are calculated and, where possible, aggregated to indices.
For example, greenhouse gas emissions released throughout the life cycle of
gas-fired electricity involve not only carbon dioxide (CO2 ) but non-combustion
sources like methane (CH4 ) leaks. Impact assessment will involve character-
izing these emissions as carbon dioxide equivalents (CO2 e) so that they can be
considered in aggregate.
4. Interpretation. The interpretation stage is geared towards informing decisions.
Interpretation involves an evaluation of results relative to the goal and scope
of the LCA in order to generate conclusions and/or recommendations for deci-
sionmakers (e.g., report commissioners). Uncertainty assessment is a critical
component of the interpretation, to inform decision-makers of how results may
change under different conditions. Studies examining the life cycle impacts
of gas-fired electricity are subject to high levels of variability and uncertainty
due to an expansive network of infrastructure that occurs across heterogeneous
landscapes, watersheds, and geologies. Emphasis on quantifying and commu-
nicating the implications of this variability and uncertainty is highly important
for impactful results.
Before a more comprehensive look at LCA methodology, it is important to distin-

guish LCA from other types of analyses; after all, there is a degree of overlap. Table
2.1 presents a comparison of LCA to other tools, outlining the differentiating factors.
Note that there are distinct differences between methods but also clear links.
Understanding the similarities and differences between LCA and other environ-
mental assessment tools underscores the importance of clearly delineating and docu-
menting decisions about the goal and scope of the LCA. For example, LCAs may
Table 2.1 Comparison of LCA to other environmental assessment tools. This table is adapted
from Jolliet et al. [9], with slight adjustments to acknowledge that the scope of LCA and carbon
footprinting are defined according to the respective methodologies; for example, only a portion of
the life cycle may be examined and will depend on the system boundaries selected
Tool Object of Scale and Substances and Basis for Basic elements
study scope impacts comparison
Life cycle Product or Global or Many Function of Mass balance;
assessment service regional; substances (as the product multimedia
life cycle defined) or service model* ; effects
(as assessment
defined)
Carbon Product Global, life Greenhouse Product Mass balance;
footprint activity or cycle (as gases; climate function, global warming
company defined) change activity, or potential
company
Water footprint Product Local or Water Product Water balance;
activity or regional; consumption; function, consumption;
company most water related activity, or competition;
important exposure company adaptation
life cycle
stages
Material flow Raw material Regional Single or Given time Mass balance;
analysis or compound or national; multiple and region material flow
material materials tracking
life cycle
Environmental New Local Highly Local Highly variable
impact localized scale; local variable carrying
assessment activity activity capacity
Risk assessment Installation Local or Relevant Maximum Multimedia
or chemical regional; substances; level of risk model* ; effects
substance selected toxicity assessment
stage
Substance flow Polluting Regional Single Given time Mass balance;
analysis substance or global; substance and region multimedia
substance model*
cycle
* Multimedia models refer to mathematical models that describe the rate and transport of pollutants
under specific environmental conditions (e.g., assuming environmental media such as air and water
as uniform and steady state)
General Framework 17
focus on one impact category (e.g. greenhouse gas emissions), resulting in overlap
with another assessment tool (e.g., carbon footprinting). The ISO standard promotes
the examination of multiple impacts to discourage practitioners from cloaking of
hidden negative impacts. As a result, limitations should be acknowledged appropri-
ately so that the audience can appropriately interpret results vis-à-vis the goal and
scope.
Goal and Scope Definition
Any LCA should commence with a clearly defined goal and scope. What is the intent
of the analysis? Who is the audience? Indeed, the goal and scope definition stage of
LCA provides critical direction for the subsequent stages of the LCA: the product
system, the system boundaries, and the functional unit are all determined during this
stage. The recipe for this stage is threefold, involving defining the product system,
the functional unit, as well as the inventory items and impacts to be analyzed. The
steps included in the recipe are summarized in this section.
(1) Define the product system that is being studied
Before commencing to undertake a LCA of gas-fired electricity, or any other for

that matter, the product system must be defined. A product system comprises of
all processes within the boundaries of the analysis. It can be considered the collec-
tion of unit processes (the smallest processes that can be disaggregated) connected
by material and energy flows resulting in the product or service under study [11].
LCA examines a product system as the combination of all sub-processes required
to produce one functional unit. The challenge is that the number of these processes
could become quite large and their delineation could be interpreted as arbitrary. For
example, a natural gas production site requires a drilling rig to drill the well and that
rig requires metals and other materials in its construction. Those manufacturing facil-
ities also require electricity, which may come from a variety of sources depending
on the location. Then one must look to the construction of those power plants, and
so on. A robust LCA thus necessitates defining the system boundaries clearly and
diligently, delineating what processes and options are modeled. In our case, we are
examining the collective products and processes that result in electricity generated
from natural gas.
A process diagram, or initial flow chart, serves as a necessary and useful guide
for the inventory analysis and subsequent steps of the LCA methodology. Figure 2.2
details the systems boundaries that are generally considered as a guide for the analyses
examined in this book, noting that they may be modified to discuss specific appli-
cations. As is true for any LCA, the system boundaries may be revisited throughout
the LCA process, refining them as the analysis is completed.
18
Other materials extraction, transportation of materials, construction, etc.

System boundaries
Combustion
• Combustion emissions
Production
Production Gathering Storage Processing Transmission
• Well drilling • Pipelines • Wells • Processing plants • Pipelines
• Completion • Compressor stations • Compressor stations • Acid gas removal • Compressor stations
• Workover • Gathering and boosting • Routine Maintenance • Compressor exhaust • Transmission and
• Liquids unloadings espisodic events • Flaring at stations LNG transport/receipt distribution (T&D) losses
• Pneumatic devices • Venting at stations
• Production • Pneumatic devices • Liquefaction
• Production compressor • Tanker Berthing &
Deberthing
• LNG Ocean Transport
• Pipelines and
distribution
Waste management, decommissioning, end-of-life, etc.
Fig. 2.2 Illustrative system boundaries of the LCAs examined in this book
2 LCA Framework, Methods, and Application
Goal and Scope Definition 19
(2) Determine the functions and the functional unit
The functional unit is an important decision, as it serves as the basis for comparisons
to be made to other systems that can result in a substitutable product [12]. For
travel using different modes (e.g., car, plane, and bike), a useful functional unit
could be passenger-mile traveled. Impacts, such as global warming, can then be
quantified on the basis of a common functional unit across different products with
the same function. For example, greenhouse gas emissions in terms of carbon dioxide
equivalents (CO2 e) can thus be quantified for each travel option per passenger-mile
traveled. In our case of gas-fired power, it makes sense to use a unit of electricity
as the basis for comparison—the impacts can then be compared to other types of
electricity generation.
Figure 2.3 illustrates the position of the functional unit in life cycle results; it is
essentially the denominator in the analysis. Specifically, the impact can be viewed as
the numerator in a life cycle result (e.g., kilogram of emissions) with the denominator
as the functional unit (e.g., Megawatt-hour of electricity generated) (Table 2.2).
Functions can be slightly different but lead to substantially different results.
(3) Select the inventory items and/or impacts studied
The inventory items (i.e., the processes to be analyzed) are determined by leveraging
the process flow diagram. A full determination of the data and results to be quantified
in a LCA involves initial decisions surrounding the impact categories that will be
analyzed (e.g., resource use, ecological consequences, and human health) and the
impact assessment methods to be implemented (see the section on impact assessment
for more information).
Impact
Life cycle result
Funconal unit
Fig. 2.3 Illustrative diagram explaining the role of the functional unit in the life cycle result. The
calculation of a life cycle result hinges on the choice of functional unit
Table 2.2 Examples of functional units and their use in life cycle results
Product system Impact of interest Example functional unit Life cycle result
CO2 e
Natural gas-fired Global warming, Unit of electricity MWh
electricity estimated using carbon generated; for e.g.,
dioxide equivalent Megawatt-hour (MWh)
(CO2 e)
m3 o f water
Natural gas fuel Water consumption, Thousand cubic feet of Mc f o f naturalgas
supply measured in cubic metres natural gas (Mcf)
(m3 )
Inventory Analysis
Life cycle inventory (LCI) is the life cycle phase where the consumption of resources
and the quantities of waste flows and environmental pollutants (e.g. emissions) are
estimated and attributed to each process within the defined system boundary of a
product’s life cycle [12]. This stage can be tricky due regional variability (e.g., in
operations, impacts, the relative amount of co-products to which one can allocate
the impacts), but also temporal variability (e.g., times during which processes and
impacts occur as well as their duration). In the inventory analysis, specific terms are
required to understand the procedure. Importantly, a unit process is a key concept for
LCA; it is the smallest element for which the inputs and outputs are quantified [13].
The inventory analysis is considered one of the last objective steps of LCA [14],
since the process hinges on data compilation and calculations. Specifically, it involves
compiling inputs and outputs across the life cycle of a product; for example, the tons
of steel required for construction and facilities equipment, the cubic feet of natural
gas produced over the life of a facility, or the lifetime Megawatt-hours of electricity
generated by a power plant. Like the other phases of LCA, there is a recipe for the
inventory analysis; however, it is more detailed than the goal and scope definition,
involving eight rather than three steps.
(1) Preparing for data collection
Preparation for data collection should be based on goal and scope definition, with
the process flow diagram as a helpful guide to show which data are to be the focus
of efforts. The system boundary delineates the processes to be examined, and the
quantitative data to be collected for each unit process (e.g., power plant). The “gold
standard” would be primary data, or data measured specific to the location and
duration of the study [14]. Secondary data is often used, with analysts leveraging
credible data, such as peer-reviewed literature, governmental and intergovernmental
reports. There are often data gaps and uncertainties, so experts and industry may be
consulted where data are limited. In doing so, analysts should be well prepared and
researched regarding available information on the processes to be discussed [3]. In
consulting experts, analysts are cautioned to ensure that ethical standards are met
in terms of involvement of human subjects but also issues of confidentiality and
anonymity are considered before stakeholder engagement.
(2) Data collection
While it might seem daunting for the first-time LCA practitioner, data collection can
be unpacked into easily managed chunks: the unit processes. Using the process flow
diagram as a guide, the product system can be broken into unit processes (see Fig. 2.4
for an example of gas-fired power). The LCA process is iterative, so practitioners
should iterate across stages and ensure they are aligned. Here, the goal and scope
definition must align with the inventory analysis. The goal and scope may contract
due to low data availability or even expand if relevant, interesting data are readily
available. In one study of LNG use in electricity generation, the system boundaries
Inventory Analysis 21
Input flows, e.g.: Unit process: natural Output flows, e.g.:

• Metals for equipment • Emissions to air
• Energy use gas producon • Emissions to water
• Waste
Intermediate flow: natural gas produced
Input flows, e.g.: Unit process: Output flows, e.g.:
• Energy use gathering sites • Emissions to water
• Waste
Intermediate flow: natural gas compressed
• Energy use gathering pipelines • Emissions to water
• Waste
Intermediate flow: natural gas transported
Input flows, e.g.: Output flows, e.g.:

Unit process: • Emissions to air
• Metals for equipment
• Energy use processing • Emissions to water
• Waste
Intermediate flow: processed natural gas

• Energy use transmission sites • Emissions to water
• Waste
Intermediate flow: natural gas compressed
Input flows, e.g.: Output flows, e.g.:

Unit process: • Emissions to air
transmission pipelines • Emissions to water
• Energy use
• Waste
Intermediate flow: natural gas delivered
Input flows, e.g.: Unit process: power Output flows, e.g.:

• Emissions to air
• Energy use
plants • Emissions to water
• Waste
Fig. 2.4 Conceptual illustration of the inventory analysis for the life cycle of natural gas-fired
electricity (without LNG). Note that the intermediate flows in this example are natural gas; however,
natural gas may be leaked so the value will decrease through each stage
were expanded as the influence of transmission and distribution losses along power
lines were found to be an important factor [15]. The process flow diagram must be
revised to ensure that any changes are reflected.
(3) Data Validation
Questions about data uncertainty in LCA are real, not to mention the need for
careful calculation, leading to the importance of data validation. Matthews et al.
[14] correctly recommend starting with common sense order of magnitude checks.
Practitioners can and should compare to published LCA results and other secondary
data to ensure that the results make sense. Finding a major difference may or may not
be due to a mistake: it is the responsibility of the analyst to determine the root cause
of the discrepancy. Could it, for example, be a technological improvement relative
to older operations? It is important to document and note factors that may affect data
quality and relevance, such as the vintage of the data, and certainly to triple-check
calculations as the analysis proceeds. Practitioners should always check data inputs
and results with existing knowledge: are the orders of magnitude consistent? If they
are not, then there may be either an error or even a potentially important reason.
(4) Co-product allocation
Quite often, multiple products are produced from one unit process, leaving questions
regarding which product is responsible for the environmental impacts.
Allocation in LCA is a technique that can be used to manage such questions by
proportional assignment of quantities of inputs and outputs to various co-products.
Allocation can be based on a variety of quantities associated with the process; for
example, it could be volumetric or based on mass.
Allocation procedures are particularly relevant for natural gas since wells often
produce co-products along with natural gas, such as natural gas liquids and oil.
Additionally, natural gas is often used for combined heat and power (CHP). For the
case of CHP, a portion of the inputs and outputs would be associated to the electricity
while the rest would be associated with the useful heat.
If there were two products from a well (e.g., oil and natural gas) then the impacts
could be allocated based on mass, energy, or volume of liquid produced. Using
any of these methods, the impact allocated to a product should be the fraction of
the total amount produced. Physical quantities (like mass or energy) are generally
recommended; however, others can be used in the absence of accepted physical
quantities (e.g., price of the product). Equations (2.1) and (2.2) provide examples of
how impacts may be allocated based on energy.
0.17kgoil × 45 kg
MJ
λoil = oil
= 0.17 (2.1)
0.17kgoil × 45 kg
MJ
oil
+ 0.76kg naturalgas × 50 kgnaturalgas
MJ
0.76kgnaturalgas × 50 kgnaturalgas
MJ
λgas = = 0.83 (2.2)
0.17kgoil × 45 kg
MJ
oil
+ 0.76kgnaturalgas × 50 MJ
kgnaturalgas
The allocation factor calculated in Eq. (2.1) means that 0.17 times the inputs
and outputs should be allocated to oil using this method. Similarly, the allocation
factor calculated in Eq. (2.2) means that 0.83 times the inputs and outputs should be
allocated to oil using this method.
Analysts should be clear when differentiating co-products and waste. It would
be incorrect to allocate impacts to a waste stream; however, it is correct to allocate
impacts to useful co-products. For example, heat from electricity generation can be
either waste heat or used as a product. For the former, the cooling methods (e.g., air
cooling, ponds, or towers) used in thermoelectric power generation generally result
in what is considered waste heat since it remains unused for any purpose. In this
case, allocation procedures are not applicable to a waste stream, and all impacts are
assigned to the electricity generation. If recovered rather than wasted, heat may be
used for space heating (i.e., combined heat and power), or industrial processes. If the
Inventory Analysis 23
Input flows Unit process Output flows Emissions and waste

Intermediate flows
Input flows Unit process Output flows Input flows from Intermediate flow
Unit process
technosphere (products and funcons)
Intermediate flows
Input flows Unit process Output flows Input flows from nature
(a) (b)
Fig. 2.5 Unit processes and quantifying inputs and outputs. a conceptualizes unit processes within
the process flow diagram, and b illustrates what types of data are required for each unit process.
Figures adapted from Matthews et al. [14] and ISO 14040 [4]
heat is used as a product, allocation procedures should be applied, and impacts will
be proportionally assigned to each product.
(5) Relating data to the unit process
Data must then be scaled to each of the unit processes described in the process flow
diagram (Fig. 2.5). First, input and output data for all the activities (unit processes)
within the system boundary must be normalized to each intermediate flow. Second,
the flows that link the activities of different unit processes are calculated based on
the requirements to produce one functional unit.
(6) Relating data to the functional unit
Each unit process then needs to be related to the functional unit in order to obtain the
final results. This means that the intermediate flows (e.g., the amount of natural gas
combusted in a power plant to generate a unit of electricity) can be used to scale the
inputs and outputs for each unit process that are required to produce one functional
unit. For each unit process, the impact (numerator) should now be on the basis of
a unit of electricity, such as kWh or MWh (denominator). The resource use and /or
emissions to the environment can then be summed up for the whole system. For
gas-fired power, heat rates and efficiencies are useful for relating the data from the
fuel cycle to the functional unit.
(7) Data aggregation
Once the inputs and outputs for each unit process have been related to the functional
unit, data can then be aggregated in preparation for the impact assessment stage. An
example is presented in Table 2.3 where impacts for three (albeit broad) life cycle
stages have been related to the functional unit of electricity generated (MWh) and
can therefore be summed to determine the life cycle result.
Table 2.3 An example inventory of natural gas-fired power, including raw materials acquisition,
materials transport, and power plant operations (without carbon capture and storage). Assumptions
for both the fuel cycle and electricity generation have drawn from Skone et al. [16] assuming a plant
efficiency of 51% HHV. The units for the fuel cycle are in kilograms (kg) of the noted substance
emitted per MWh
Flow Natural gas fuel cycle Electricity Unit
Extraction Processing Transmission Total generation
CO2 7.18 16.41 3.93 27.57 350 kg/MWh

CH4 0.95 0.11 0.70 1.78 0.00856 kg/MWh
Impact Assessment
Life cycle impact assessment (LCIA) is the process of quantifying life cycle results
on the basis of the potential contributions of resource extraction and environmental
releases in terms of a subset of their potential impacts [17]. The results of the LCIA
are presented in specific impact categories (e.g., climate change, eco-toxicity, ozone
depletion, resource scarcity) that are often aggregated indices. A wide variety of envi-
ronmental impact categories are examined in LCA (Table 2.4), each being related
to varying spatial scales from local to global. With spatially aggregated and undif-
ferentiated inventories, determining impacts at scales lower than global will remain
technically unfeasible. For some cases where impact categories are determined at
global scales, additional spatiotemporal analyses may not be required even when
more granular data are available. However, if spatiotemporal differentiation is deter-
mined to be important during the Goal and Scope Definition and inventories permit,
then methods should be developed accordingly.
The recipe for LCIA comes with mandatory and optional elements. The manda-
tory part of LCIA includes selection, classification, and characterization. During the
first step, selection, relevant impact categories are chosen, along with the relevant
impact indicators and the methods of aggregation (or characterization). The second
step, classification, involves assigning inventory results to specific impact categories.
For example, carbon dioxide and other greenhouse gases would be assigned to
climate change. The last mandatory step, characterization, involves using the inven-
tory results to calculate the category indicator result. For climate change, for example,
the use of global warming potentials (GWPs) enables the representation of green-
house gas emissions on an equivalent basis to carbon dioxide (generally referred
to as carbon dioxide equivalents (CO2 e)). To convert other greenhouse gases on an
equivalent basis as carbon dioxide, the IPCC publishes GWP that can be applied
with simple multiplication. To convert methane (CH4 ) to CO2 e, one must simply
multiply the mass emitted by the most recently published GWP. The IPCC presently
recommends 36 and 87 for 20-year time horizons for 100-year and 20-year time hori-
zons respectively. Tables 2.5 and 2.6 present simple characterization results using
these GWPs. Similar methods should be applied for characterization in other impact
categories.
Impact Assessment 25
Table 2.4 Summary of life cycle impact categories, their scales and examples of LCA data
Impact category Scale Examples of LCI data
Climate change Global Carbon dioxide (CO2 ), methane (CH4 )
Stratospheric ozone depletion Global Chlorofluorocarbons (CFCs),
hydrochlorofluorocarbons (HFCs),
Methyl Bromide (CH3 Br)
Acidification Regional, local Sulfur Oxides (SOx ), Nitrogen oxides
(NOx ), hydrochloric acid (HCl)
Eutrophication Local Phosphate (PO4 ), Ammonia (NH4 )
Photochemical smog Local Non-methane hydrocarbon (NMHC)
Terrestrial toxicity Local Toxic chemicals reported with lethal or
harmful concentrations to terrestrial
animals and ecosystems
Aquatic toxicity Local Toxic chemicals reported with lethal or
harmful concentrations to fish and
aquatic ecosystems
Human health Global, regional, local Total releases to water, air, soil;
exposure and impacts to humans
Resource depletion Global, regional, local Quantity of mineral or fossil fuel used,
can be relative to availability
Land use Global, regional, local Quantity disposed of in a landfill or land
modifications such as transformation or
occupation impacts
Water use Regional, local Water consumed or withdrawn
Adapted from prior work (e.g., U.S. EPA 2006; Matthews et al. [14])
Table 2.5 An example of characterization for a combined cycle natural gas-fired power plant using
a 100-year GWP
Flow Natural gas fuel cycle Electricity Total Unit
Extraction Processing Transmission generation
CO2 7.18 16.41 3.93 350 377.57 kgCO2 e/MWh

CH4 34.38 4.14 25.28 0.31 64.22 kgCO2 e/MWh
Total 41.56 20.55 29.20 350.31 441.80 kgCO2 e/MWh
Assumptions are based on but adapted from Skone et al. [16]
Table 2.6 An example of characterization for a combined cycle natural gas-fired power plant using
a 20-year GWP. Assumptions are based on Skone et al. [16]
Flow Natural gas fuel cycle Electricity Total Unit
Extraction Processing Transmission generation
CO2 7.18 16.41 3.93 350 377.57 kgCO2 e/MWh

CH4 83.08 10.00 61.08 0.74 155.20 kgCO2 e/MWh
Total 90.26 26.41 65.01 350.74 532.78 kgCO2 e/MWh
The optional steps (normalization, grouping, and weighting) are also dependent on
the study and methodology applied. Normalization involves calculating the magni-
tude of the category indicator results relative to the total in a locality, region or
globally, for example. Grouping involves assigning impact categories to groups of
similar impacts. Finally, weighting impacts means that the LCA practitioner applies
a relative value to the importance of the impacts, allowing integration across all
impact categories. There are several published impact assessment methods that can
be applicable during the LCIA phase [7–10]. Such pre-developed impact assessment
methods often apply some form of weighting impacts.
LCIA is a cornerstone stage of LCA, but like the other stages, it is not without
challenges [18]. First, it is considered subjective relative to the inventory analysis
stage, based on science but combined with values. Second, a comprehensive focus and
boundary is often lacking, making the LCIA generic rather than providing meaningful
results specific to the environment. Third, related to the subjectivity of this LCA
stage, indicators are not equally important or precise. This limitation typically leads
to subjective methods, such as weighting. Fourth, data and methods are typically
emerging and not completely mature, making it challenging to characterize impacts
accurately (this will be more comprehensively addressed in Chap. 4). Finally, impact
results are subject to high levels of uncertainty, particularly where methods/data are
lacking. Such uncertainty is exemplified in the assumed linearity of many of the
impact characterization and other methods.
Interpretation
The interpretation stage of a LCA requires the practitioner to make clear what
the results mean for the decision-maker. Interpretation involves a subset or all the
following:
• Studying results in relation to Goal and Scope Definition to determine whether
conclusions can be made that are consistent with study scope.
• Undertaking the appropriate sensitivity or uncertainty analyses.
• Checking back to ensure your system boundaries are appropriate.
While a practitioner may be discouraged if there appears to be no major results
(e.g., two projects have virtually the same impact), I encourage two interpretations.
First, no major finding is a finding: if one can conclusively state that the life cycle
results of two projects are comparatively similar, this finding may be meaningful and
should be interpreted with this in mind. Second, interpretation also involves uncer-
tainty assessment. Practitioners should ensure to view with a critical eye both data
quality and potential uncertainties in terms of how they may influence results. Uncer-
tainty analysis such as breakeven analysis can provide insight into more valuable
interpretation (such concepts are explored in depth in Chap. 6).
Interpretation 27
LCA often stops at the inventory stage, skips over impacts and goes right to the
interpretation stage due to the aforementioned limitations in LCIA [14]. While this
may appear to be a limitation, such inventory results can be quite conclusive and
informative; for example, if one alternative consumes a greater number of resources.
Regardless, the interpretation in LCA should occur at every stage [12].
The recipe for interpretation involves the following steps.
• Identifying significant issues, for example, determining important findings and
identifying critical methodological choices.
• Evaluation of these issues considering the results.
• Establishing confidence in results and testing their robustness.
• Conclusions, recommendations, and reporting.
There are numerous quantitative methods for establishing confidence in results
and testing their robustness, ensuring both completeness and consistency [3, 14].
First, a dominance (or gravity) analysis identifies the heaviest hitting life cycle activ-
ities; for example, the generation stage for gas-fired electricity generation results
in the dominant contribution to greenhouse gases. Similarly, a contribution analysis
determines the largest contributing process or intervention to each impact category
or environmental load to the total impact.
There are many well-established methods for testing the robustness of the results of
LCA models. Completeness checks involve identifying data gaps and determining the
completeness of impact assessment. Consistency checks involve evaluating whether
the methodological choices were appropriate relative to goal and scope definition.
Uncertainty assessment is a useful tool to check for the influence of uncertain data
on the results and can capture the effects of multiple uncertain inputs (to be covered
in Chap. 6). Sensitivity analyses—often considered an option for uncertainty assess-
ment—can be used to identify individual inputs that are critical to the robustness of the
results. Variation analysis determines the influence of alternative scenarios. Related,
breakeven analysis can parametrically determine the input assumptions required for
results of two alternatives to break even or to be equivalent. And finally, data quality
assessment examines data gaps as well as approximates data, and appropriates data
as required.
Quantitative uncertainty assessment is a critical component of the interpretation
stage; however, the purpose is to interpret results for decision-makers. A decision-
maker analysis involves determining who has power in relation to each part of the
results since LCA results often impact a variety of different sectors; specifically, who
has the power to reduce impacts. Sometimes, it is clear for whom the LCA is being
interpreted; however, the report commissioners may be interested in how to leverage
results.
Conclusion
Environmental analyses of gas-fired power, due to the controversy over being the
cleanest burning fossil fuel and substantial economic interests, reap enormous bene-
fits from the robust, standardized methods that LCA brings to the table. At the same
time, simplified or streamlined LCA approaches simply are less applicable for the
product system of gas-fired power due to the irreducible variability of the expansive
network of infrastructure and known uncertainties in some impact categories (to be
elucidated in subsequent chapters). While known LCA models and databases may
include simplified representations of gas-fired power, they will not represent the full
scope of variability and uncertainty in both inputs and life cycle results. For better
representation both variability and uncertainty, process based LCA provides a robust
approach that can incorporate comprehensive datasets towards a more realistic exam-
ination of the natural gas fuel cycle and the power plants where the combustion of
the gas occurs.
References
1. Hellweg, S., Mila i Canals, L.: Emerging approaches, challenges and opportunities in life cycle
assessment. Science 344, 1109–1113
2. Lave, L.B., Hendrickson, C.T., McMichael, F.C.: Environmental implications of electric cars.
Science 268, 993–995 (1995)
3. Tillman, A.M., Baumann, H.: The Hitchhikers Guide to LCA. Studentlitteratur AB, Lund,
Sweden (2004)
4. International Organization for Standardization (ISO).: 14040 International standard: Envi-
ronmental Management–Life Cycle Assessment–Principles and Framework. International
Organisation for Standardization, Geneva, Switzerland (2006)
5. Pryshlakivsky, J., Searcy, C.: Fifteen years of ISO 14040: a review. J. Clean. Prod. 57, 115–123
(2013)
6. Hertwich, E.G., Pease, W.S.: ISO 14042 restricts use and development of impact assessment.
Int. J. Llfe Cycle Assess. 3, 180 (1998)
7. Curran, M.: A. Life Cycle Assessment Handbook: a Guide for Environmentally Sustainable
Products. Wiley (2012)
8. Klöpffer, W., Grahl, B.: Life Cycle Assessment (LCA): a Guide to Best Practice. Wiley (2014)
9. Jolliet, O., Saadé-Sbeih, M., Shaked, S., Jolliet, A., Crettaz, P. Environmental Life Cycle
Assessment. CRC Press (2015)
10. Hauschild, M.Z., Rosenbaum, R.K., Olsen, S.I.: Life Cycle Assessment. Springer, Berlin (2018)
11. Life Cycle Initiative Glossary of Life Cycle Terms.: https://www.lifecycleinitiative.org/resour
ces/life-cycle-terminology-2/#p. Accessed on 7 Jan 2021
12. Rebitzer, G., Ekvall, T., Frischknecht, R., Hunkeler, D., Norris, G., Rydberg, T., Schmidt, W.,
Suh, S., Weidema, B.P., Pennington, D.W.: Life cycle assessment: part 1: framework, goal and
scope definition, inventory analysis, and applications. Environ. Int. 30, 701–720 (2004)
13. Life Cycle Initiative; United Nations Environmental Programme Glossary of Life Cycle Terms.
Accessed on 28 Mar 2020
14. Matthews, H.S., Hendrickson, C.T., Matthews, D.H.: Life Cycle Assessment: Quantitative
Approaches for Decisions that Matter. Retrieved June 2018, 1
15. Kasumu, A.S., Li, V., Coleman, J.W., Liendo, J., Jordaan, S.M.: Country-level life cycle assess-
ment of greenhouse gas emissions from liquefied natural gas trade for electricity generation.
Environ. Sci. Technol. 52(4), 1735–1746 (2018)
References 29
16. Skone, T.J., Schivley, G., Jamieson, M., Marriott, J., Cooney, G., Littlefield, J., Mutchek, M.,
Krynock, M., Shih, C.Y.: Life Cycle Analysis: Natural Gas Combined Cycle (NGCC) Power
Plants. Keystone Logic and the National Energy Technology Laboratory (2018)
17. Pennington, D.W., Potting, J., Finnveden, G., Lindeijer, E., Jolliet, O., Rydberg, T., Rebitzer,
G.: Life cycle assessment part 2: current impact assessment practice. Environ. Int. 30, 721–739
(2004)
18. Sala, S., Pant, R., Hauschild, M., Pennington, D.: Research needs and challenges from science
to decision support. Lesson learnt from the development of the international reference life
cycle data system (ILCD) recommendations for life cycle impact assessment. Sustainability 4,
1412–1425 (2012)
Chapter 3
The Denominator: Natural Gas
Production, Throughput, and Electricity
Generation
Abstract To arrive at a life cycle result, not only impacts must be calculated, but
the denominator (or functional unit) must be estimated. For each unit process within
the life cycle boundaries, the amount of material that passes through must be esti-
mated; for natural gas production systems, we quantify the natural gas extracted by
production sites and the gas that travels through each facility in the fuel cycle (i.e., the
throughput). The results for each unit process in the fuel cycle is then related to the
functional unit—a unit of electricity generated—using heat rates or efficiencies of
natural gas-fired power plants. In this chapter, methods used to estimate the denomi-
nator of life cycle results for gas-fired power are explained, from estimated ultimate
recover of natural gas from a well to facility throughput, to electricity generation.
The extensive network of natural gas infrastructure that is present in the fuel cycle
of gas-fired power results in an additional layer of variability in LCA, which should
be considered by practitioners interested in this product system.
Introduction
The selection of the denominator—the functional unit—in life cycle assessment

(LCA) is a crucial decision that can be highly influential on the outcome of the results.
LCAs are completed for a function that is fulfilled by a product system, where the
functional unit is the unit of output to which all environmental impacts across the
life cycle are related [1–5]. It is the very factor used to normalize the environmental
impacts. For our case of gas-fired power, the functional unit is generally one unit of
electricity generated; for example, one kilowatt-hour (kWh) or one Megawatt-hour
or MWh. The function that our system provides is thus the generation of electricity.
We note that the inclusion of power transmission and distribution (T&D) implies a
very subtle difference in the functional unit: the unit of electricity delivered (rather
than generated).
This chapter involves a comprehensive examination of the quantification of the
functional unit and how it is related to intermediate flows, from the supply chain of
natural gas through electricity generation at a power plant. Recall that in the inventory
analysis stage of LCA (described in Chap. 2), data are first analyzed at the scale of
each of the unit processes described in the process flow diagram. Input/output data
32 3 The Denominator: Natural Gas Production, Throughput …
for all unit processes are then normalized to each intermediate flow. The flows that
link the activities of different unit processes are calculated based on the requirements
to produce one functional unit.
For simplicity, we will commence by separating the unit processes associated with
the natural gas supply chain from those associated with electricity. Impacts of the
natural gas supply chain are first defined based on a unit of natural gas, which can
then be related to the functional unit using the amount of natural gas consumed at the
power plant for each unit electricity generated; for example, using the heat rate of
the power plant or the efficiency. For production facilities, the amount of natural gas
produced (cubic feet, for example) can be used to relate the impacts to the functional
unit (Megawatt-hour, for example) using either metric. The lifetime throughput—or
the total flow of natural gas that has passed through a facility over its life—is the
measure that can be related to the functional unit for each unit process in the natural
gas supply chain.
This chapter will examine the relevant energy throughputs for typical unit
processes of interest to our system boundaries. Data specific natural gas-fired gener-
ation will be used as an example to demonstrate the calculations used to relate each
life cycle phase to the basis of a functional unit (unit of electricity generated) for the
final LCA results. The variability in efficiencies across technologies and heat rates
across generating units will be discussed, using real-life data examples from power
plants in the U.S.A and aggregate country-level statistics internationally.
Infrastructure Lifetime
In the absence of a complete primary dataset that details the lifetime throughput of
each facility (a real challenge for many processes that are still ongoing or that are
under development), one of the most important parameters is the producing lifetime
of the infrastructure. Annual or monthly throughput can then serve as the basis for
the lifetime generation. Such estimates for lifetime throughput for each facility are
highly sensitive to the assumption about the operational life of the facility and the
month or year chosen to be representative. Of course, it is a major challenge to predict
the exact duration of time that a facility will be in operation or exactly when it will be
decommissioned. While methods to evaluate uncertainty will be examined in detail
in Chap. 6, it is critical to acknowledge and investigate the influence of uncertainty for
facility throughput and ultimately the functional unit. In consultation with experts,
we compiled assumptions for an analysis of gas-fired power in the Barnett Shale
region of Texas that can serve as a useful baseline [6] (Table 3.1).
Data collection enabled a range of estimates to be determined (Table 3.1), which
can serve as inputs for an uncertainty analysis (see Chap. 6 for more details). Note that
a regionally aggregated value represented transmission pipeline throughput. In the
absence of site-level data, such aggregate assumptions can inform LCAs; however,
practitioners should seek to quantify variability over the population in the region and
timeframe under study whenever possible to improve the accuracy of the results.
Natural Gas Extraction 33
Table 3.1 Sample sizes (n), population sizes (N), and infrastructure lifetime assumptions used to
estimate the functional unit of lifetime throughput for each life cycle phase in the seven-county
study area of the Barnett Shale in Texas, United States
Life cycle stage Lifetime Throughput unita Lifetime (years)b Range (years)c
throughput
estimator
Production sites EUR mcf 25 10–30
Gathering sitesd EUR mcf 25 10–30
Gathering EUR mcf 25 10–30
pipelinese
Processing sites Facility mcf 30 10–50
throughput
Transmission Total mcf 30 10–50
sites production
(study area)
Transmission Total mcf 30 10–50
pipelinesf production
(study area)
Power plants Lifetime MWh 30 10–50
generation
a Mcf refers to thousand cubic feet and MWh refers to Megawatt hour
b The lifetime of facilities was estimated assuming a 30 year economic lifetime [5, 7, 8] except
for wells (and gathering infrastructure connected to wells) which were assumed to have 25 years
of production [9, 10]. Note that to estimate EUR, well-level production curves were fit to real
data rather than multiplying an average production rate by the estimated lifetime; therefore, the
production decline over the lifetime of the well is represented in our estimates.
c Facility lifetimes were assumed to range from the minimum of 10 years to a maximum of 50 years,
except for wells (and gathering infrastructure connected to wells), which have lifetimes assumed to
range from 10 as a conservative estimate to 30 years [8–12]
d Gathering sites typically service more than one well. To account for this, the average number of
wells per gathering site was calculated for the study area (=88). Each well was assumed to produce
for 25 years. The lifetime throughput was then calculated by multiplying the average wells per site
by the average EUR per well.
e Each production site has an associated gathering pipeline, connecting the production site to
transmission pipelines. The EUR for gathering pipelines is taken to be the cumulative EUR of
the associated wells
f Transmission pipelines were considered collectively for the analysis with production from the
seven-county study area used as the estimate of throughput
Natural Gas Extraction
For wells and infrastructure that is directly related to wells (production sites, gath-
ering pipelines, and gathering sites), Estimated Ultimate Recovery (EUR) can be
used to relate the facility to the functional unit. EUR provides estimates of well-level
natural gas production over the well’s producing life. If a well is no longer producing,
the lifetime is established. More frequently than not, wells are still producing at
the time of an analysis. Further complicating the matter, wells may appear to have
completed production but may be temporarily shut-in or recompleted, so they may

continue to produce natural gas. Like other infrastructure elements, this means that
the full lifetime is unknown as the end of production will be in the future and subject
to technological and economic conditions. That said, informed assumptions can
support the analysis along with uncertainty assessments. In general, there are two
options used to define the EUR: the use of prior estimates or the analysis of specific
wells relevant to the study for the time at hand. The former has generally been the
convention [13]; however, more recent studies have commenced estimating EUR for
wells within the geographic scope of the analysis [6].
In the past, EUR has been estimated using initial production rates and decline
curve characteristics [14]. Such methods have yielded useful results (Table 3.2) that
can serve as data inputs for LCAs. Decline curve analysis is the most common
approach to evaluating the performance of reservoir and for calculating EUR. A
variety of methods have been developed, such as the fitting of exponential and hyper-
bolic curves to existing production datasets. Such methods may result in over- or
under-estimation of well-level lifetime recovery. For example, exponential decline
curves will underestimate production for wells producing in geologic formations
where natural gas does not flow as easily as those where conventional production is
economically viable (e.g. shale gas) [60]. Further confounding the challenge, studies
often rely on only one value for EUR—a practice that overlooks how well-level vari-
ability may influence results and other assumptions that are points in a probability
distribution of which we cannot be certain.
More recently, new methods provide a path to more accurate results for shale gas
wells [19], which we will use as an example for EUR calculations. This curve-fitting
method that has been found to more accurately predict gas production from shale
gas wells has been more recently used in LCA studies. While this method has been
selected as most applicable to horizontal wells, we note that other methods may be
more suitable for vertical and directional wells.
The amount of gas recovered from a shale gas well (EUR) is calculated by:

EUR = M × RF = Mk t/τ (3.1)
Table 3.2 A sample of published Estimated Ultimate Recoveries (EURs) for natural gas wells in
the United States. EUR is a prediction of the natural gas produced over the life of a well
EUR (Bcf) INTEK [15] Baihly [16] Warlick [17] EIA-AEO [18]
Lifetime 30 30 – 30
Barnett 1.42 2.989 2–2.5 0.255–2.174
Marcellus 1.18 – 4.2–4.4 2.045–13.574
Haynesville 3.57 5.915 >6.5 5.261–10.162
Fayetteville 2.07 1.390 2.1–3.5 –
Woodford 2.98 1.696 2–5 0.0055–2.524
Eagle Ford – 3.793 – 0.0010–1.530
where M is the original mass of gas contained in the reservoir (converted to a volu-
metric basis), RF is the dimensionless recovery factor (a number between 0 and unity
that describes the proportion of gas that can be recovered), t is the number of months
the well has been producing (or the well lifetime in months), τ is the interference
time (months), and k is a constant depending on the gas composition, temperature,
and initial reservoir and well flowing pressures pi and pf . The interference time τ
is the point in time where the gas flow between fractures drops below the original
reservoir pressure.
For a given well in any given producing basin, the history of monthly volumetric
production in thousand cubic feet (Mcf) can be estimated for an assumed well life-
time. As an example, Patzek [19] found that for the Barnett shale the values of pi
and pf are 3500 psi and 500 psi, respectively, resulting in k = 0.645. Using these
values, monthly production data can be used to fit a production curve, estimating τ
and M simultaneously for every well in the sample. As previously noted, the lifetime
remains an uncertain value. Various lifetimes (e.g.,10, 15, 20, 25, and 30 years) may
be assumed to represent potential economic lifetimes for the wells in the dataset but
also to account for uncertainty. Testing the results against these assumptions about
production lifetimes is an example of a sensitivity analysis (for more information on
this approach to understanding uncertainty, refer to Chap. 6).
To estimate the parameters used for fitting the curve (M and τ ), the Levenberg–
Marquardt algorithm [61] was used to minimizes the error between real data and the
fitted curve, as defined in the following objective function:

m(t) − Mk t/τ (3.2)
where m(t) is an array of the monthly cumulative natural gas production for each well
and t is the corresponding month. M and τ are then optimized so that the absolute
value of objective function is minimized.
In order to complete such an analysis, actual production data for the length of
each well’s producing life is required. Such data may be acquired from public data,
such as from the Energy Information Administration’s Annual Energy Outlook [18]
or commercial databases that may sell license to the data. Enverus Drilling Info [20]
is an example of the latter; it is a database that retains historical well production data
and characteristics provided by state agencies, with quality screens applied at the
state-level to remove anomalies and provide consistency across jurisdictions.
EUR may serve to define the intermediate flow for numerous unit processes. In
the aforementioned study of the Barnett shale of Texas, site-level EUR (including
co-location of wells) was directly employed as the denominator to calculate the life
cycle land use intensity for production sites and gathering pipelines (Fig. 3.1). There
may be some infrastructure that has less than ideal throughput data. For example,
throughput data for compressor stations have been scarce – horsepower is more
readily available than estimates for the actual natural gas that passes through the
36
Fig. 3.1 Distribution of estimated ultimate recovery (EUR, in billion cubic feet (bcf)) for the 400 sampled sites from the seven-county study area in the Barnett
using methods from Patzek [19]
3 The Denominator: Natural Gas Production, Throughput …
infrastructure. For infrastructure that services numerous production sites, like gath-
ering infrastructure, the average number of wells per unit of infrastructure can be
used to relate EUR to the throughput of the infrastructure.
It must be noted that drilling multiple producing wells at one site may not only
result in a higher amount of product, but it may also diminish impacts through reduced
infrastructure, emissions, and land use (for example). EUR can be used to estimate
throughput for the case of single or multiple wells on the same site, using monthly
production data for an individual well or as the sum of the individual EURs for all
co-located wells. The temporal lifetime of a producing well is uncertain and can
influence the results. To test sensitivity of our results to differing assumptions of
average well lifetime, we calculated well-level EUR across a range of assumed well
lifetimes.
The sensitivity of the EUR estimate to assumption about well lifetime is demon-
strated. The “single wells” portion of the plot assumes sites that have only one well
whereas the “co-located wells” portion of the plot accounts for observed occurrences
where multiple wells were drilled within a given site’s boundaries. Whether the single
well or co-located well approach is most applicable depends on the impact of interest.
For land area requirements, considering the co-location of wells is important since
the efficiency of land use is higher. Emissions analysis is another story, however,
since the drilling and completion of wells may add to the results. Additionally, even
for land, the area required for sites with multiple wells may be higher than that with
single wells; analyses of the number of wells per site is encouraged. Data showed
that 291 sites from our sample had one well, while the remainder ranged from 2 to 11
wells per site in the study area for the year 2009. The boxes in the figure represent the
interquartile range, the line in the box represents the median, the whiskers represent
the maximum and minimum, and the blue diamond represents the mean.
Because there is indeed uncertainty in well lifetime for regions like the Barnett
where wells may have been producing for more than a decade. The influence of
uncertain lifetimes on results is further investigated in Chap. 6.
It is important to note that the example here is specific to horizontally drilled
shale gas in the Barnett shale of Texas; however, the approach is broadly applicable
across regions. For example, Tavakkoli et al. (in review) expanded the analysis to
all horizontal, hydraulically fractured natural gas wells that were producing in the
United States in 2005, 2010, and 2015. In the future, such analyses may be expanded
to conventional, as well as other vertical and directional wells.
Facility Throughput
For unit processes other than wells with throughput that cannot be defined using
EUR, there are three primary ways that facility throughput can be estimated:
• Facility: an individual facility is used to estimate throughput.
• Aggregated to a region: numerous facilities within a defined region.
• Product system: all facilities that fall within the scope of a full product system.
Other infrastructure associated with domestic consumption of natural gas include

processing facilities, transmission lines and the associated compressor stations, lique-
faction, ocean transport and regasification (as required), and power plants. For each
of these infrastructure types, methods to estimate throughput are described below.
While lifetime throughput can be calculated as total throughput for the operational
lifetime, working capacity can also be leveraged in the determination of throughput.
For facilities where periodic data are available, estimates for the lifetime
throughput can be estimated. Taking annual data as an example, the lifetime
throughput of natural gas that travels through the infrastructure can be approximated
as the product of annual throughput for the year of question and the infrastructure
element’s assumed lifetime. Where throughput data are not available, analyses may
rely on regional production. For example, in Jordaan et al. [6] transmission sites were
treated differently than processing facilities because site-level throughput and disag-
gregated data were not publicly available. The total natural gas production in 2009
from the 7-county study area in the Barnett shale of Texas was divided by the number
of transmission sites in the study area to estimate an average annual throughput for
each site. The lifetime throughput was then multiplied by the assumed life of the
facility. Similarly, the annual throughput for all transmission pipelines in the study
area was assumed to be the study area’s annual production of natural gas.
Transmission compressor stations are located every 50–60 miles to boost and
maintain the pressure of natural gas [21], as a result, a unit of natural gas might
be compressed multiple times between the source of generation and power plants.
Determining the throughput of natural gas compression facilities is less straightfor-
ward than for other infrastructure since the relationship is not one-on-one between
the installed capacity or horsepower (HP) and the amount of natural gas that flows
through that station [22]. Specifically, natural gas may flow through many compressor
stations before it reaches the end use. Data on actual throughput may be lacking, as
is the case for the United States. Installed horsepower and capacity may serve as
substitutes. In 2007, for example, a total of 16.9 million installed horsepower as
well as a total throughput capability of 881 Bcf per day was reported by the Energy
Information Administration [23], translating to 19,027 mcf HP−1 y−1 .
In addition to transmission compressor stations, there are three other compression
infrastructure elements in natural supply chain: production, gathering, and storage
compressor stations. Since natural gas may be compressed several times in different
stations, this approach is recommended for using relevant throughput reported at the
same facility or location. Such data can provide useful placemarks, but the lack of
data availability points to a need to improve datasets that characterize less studied
infrastructure types which are becoming increasingly important in different aspects
of LCA, such as compressor stations. For example, Tavakkoli et al. [22] made the
reasonable assumption that the throughput of the storage sites is representative of
the throughput of compressor stations located at the storage sites.
Power Plant Operation 39
Power Plant Operation
Arguably the most important part of a LCA of natural gas-fired electricity is the
efficiency of combustion at the power plant, which can be used to convert each life
cycle phase to the basis of a functional unit (unit of electricity generated) for the final
LCA results. The functional unit for power plants can be calculated as electricity
generation in a baseline year—selected in the goal and scope definition stage of the
LCA—multiplied by the assumed lifetime. All upstream impacts associated with the
natural gas fuel supply depend on the amount of natural gas required to generate
this electricity. Efficiencies are generally expressed as a percentage, represented by
the useful energy that comes out of a system (in this case, a power plant) divided
by the energy consumed by the system. Thermodynamically, they cannot be greater
than 100% and the highest efficiency natural gas boilers hover around 85% and the
highest net efficiency natural gas power plants reaching just over 60% [24].
Heat rates are another expression of efficiency, specifically for electrical genera-
tors or power plants [25], that measure the efficiency at which fuel can be converted
into electricity. Specifically, for gas-fired power, the value would be a measure of
the amount of natural gas (e.g., in cubic feet or British Thermal Units (BTUs)) used
to generate a unit of electricity (e.g., kWh). Efficiencies expressed as a heat rate
can easily be converted into percentages by dividing the equivalent BTU content
for a kWh (3412 BTU). A heat rate of 10,500 Btu/kWh would be equivalent to an
efficiency of 33%. Heat rates provide a simple conversion by which unit processes
associated with the fuel supply can be related to the functional unit. LCA practi-
tioners should be consistent regarding whether the data they use in their analyses
is higher or lower heating value (HHV or LHV) and report which type of is used.
HHV is the amount of heat released from combustion of a fuel (initially at 25 °C)
relative to when products have returned to the initial temperature, accounting for the
latent heat of vaporization. LHV, on the other hand, does not consider the heat of
vaporization of water and considers a final temperature of 150C. Using the latter
artificially inflates the efficiency of the process and may be misleading if not clearly
reported or used consistently.
The efficiencies of the full power plant fleet, including those specific to natural
gas, are not static over time. For natural gas, it will depend on the type of generation
technology. The two most common categories can be categorized as simple cycle and
combined cycle (Fig. 3.2) [26]. Simple cycle involves a gas combustion turbine that
exhausts the gas into the atmosphere. The exhaust gas for can be used in an exchanger
to pre-heat the air that passes from the compressor to the combustor, thus creating
a more efficient system (often referred to as regenerative cycle). Combined cycle
involves using a heat recovery steam generator to re-use waste energy in a steam
turbine. The latter has been recognized as among the highest efficiency options for
thermal power generation, with operation efficiencies reaching into the 60% range in
recent years. In addition to these options, steam generation and internal combustion
engines are also used for electricity generation.
40
Fuel Heat recovery steam

generator
Inlet air
Gas turbine
Gas Compression Turbine

Exhaust gas
Steam turbine
Condenser
(a) (b)
Fig. 3.2 Illustrative examples of gas-fired electricity generation technologies ( adapted from Fadok [27])
3 The Denominator: Natural Gas Production, Throughput …
Power Plant Operation 41
The technology employed in gas-fired generation becomes a critical factor in

understanding all life cycle impacts, as all upstream impacts associated with the
fuel supply are dependent on the amount of gas combusted in the generation of
electricity. If the LCA is conducted for a region, for example, there may be a mix of
these technologies employed. Further, the generation fleet in the region may evolve
over time.
The United States exemplifies the case, as well as many subregions in the country
since the 2009 shale boom has initiated a transition from coal-fired power to gas
generation. Increasing efficiencies combined low natural gas prices have made gas-
fired power competitive as baseload power. Combined cycle gas turbines (CCGT)
were generally not used as baseload prior to the shale boom due to the volatility
in natural gas prices. CCGT became among the capacity additions of choice, along
with increasingly competitive renewable options (i.e., wind and solar). For example,
the gas-fired generation fleet has been transitioning to a more efficient baseline as
combined cycle power plants increased from 75 to 85% between 2005 and 2015
[28], increasing the average operating efficiencies from 33 to 45% [29]. Over time,
efficiencies can be expected to generally improve as technologies mature, making
the overall trends in life cycle environmental impacts a moving target with new tech-
nological innovation. The Energy Information Administration publishes information
about heat rates in of the power generation types in the United states as well as data
reported by individual power plants that can characterize efficiencies [28].
Decisions about power plant heat rates or efficiencies are extremely influential in
the estimation of life cycle results, since they are used to relate the upstream impacts
associated with the fuel supply to the functional unit [30, 31]. Importantly, the goal
and scope definition should bound the geographic scope under study. Individual
power plants may be characterized [22, 23], or regional averages may be employed
[30, 31], depending on the question at hand. Overall, the heat rates employed should
reflect the variability of power plants within the scope of the study. Studies should
seek to reflect the differences in efficiencies across regions and particularly to charac-
terize the variability across individual power plants (where data permits). Investiga-
tions of the variability of efficiency across regions yield large differences (Table 3.3).
Conclusions
After providing context around the complexity of the fuel cycle of gas-fired power,
the need for detailed examinations of the expansive and evolving network of
infrastructure should be clearer.
Natural gas throughput is challenged by large populations of infrastructure that
require substantial data collection and analysis of large datasets. Each life cycle phase
requires tailored approaches and assumptions to model lifetime throughput as either
the production, consumption, or flow through the infrastructure and facilities. On top
of the fuel cycle, understanding the life cycle impacts of natural gas-fired electricity
may involve the examination of an individual power plant or a region with more than
Table 3.3 Select power generation efficiencies with potential upgrades for global fleet and the top
10 countries in terms of CO2 reduction potential for improving efficiency (estimates published by
General Electric in 2016) [32]
Country 2015 gas average efficiency Potential CO2 reductions
generation (%) efficiency with (million metric
Upgrades (%) tonnes)
World 5,713,194 39 43 203
Russian 564,068 26 30 45
Federation
United States 1,316,652 45 48 34
Japan 427,836 45 48 12
Saudi Arabia 160,229 32 36 11
Iran 184,498 43 46 6
United Arab 107,746 34 37 4
Emirates
China 134,041 39 42 4
Korea 163,691 44 46 4
Egypt 152,635 45 48 4
India 153,579 45 48 4
one (or many) power plants. As a result, understanding the heat rates or efficiencies
of the power plants under study is critical. The required upstream infrastructure,
material requirements across the life cycle, through each life cycle stage up to and
including waste management provide a more complete picture and require that the
practitioner relate impacts to both fuel cycle and power generation. Practitioners
should include specific details to be accurate: the technologies considered, the time
horizon, and the geographic area under study for each supply chain segment. Like
any other product system, data challenges should be expected. Data are not always
available or may not be reliable, however, particularly at the global scale [31]. For
example, while heat rates are published annually for the United States and can even
be calculated for each power plant, high resolution global data was not available in
the form of time series data at the time that this book was written. A robust LCA will
include a detailed report of data used in the analysis but also data gaps with relevant
uncertainty analyses to ensure the results can be correctly interpreted.
References
1. Matthews, H.S., Hendrickson, C.T., Matthews, D.H.: Life Cycle Assessment: Quantitative
Approaches For Decisions That Matter. Retrieved June 2018, 1
2. International Organization for Standardization, (ISO).: 14040 International Standard: Envi-
ronmental Management–Life Cycle Assessment–Principles and Framework. International
Organisation for Standardization, Geneva, Switzerland (2006)
References 43
3. Rebitzer, G., Ekvall, T., Frischknecht, R., Hunkeler, D., Norris, G., Rydberg, T., Schmidt, W.,
Suh, S., Weidema, B.P., Pennington, D.W.: Life cycle assessment: part 1: framework, goal and
scope definition, inventory analysis, and applications. Environ. Int. 30, 701–720 (2004)
5. Schenck, R., White, P.: Environmental Life Cycle Assessment: Measuring the Environmental
Performance of Products. American Center for Life Cycle Assessment Vashon, Washington
(2014)
6. Jordaan, S.M., Heath, G.A., Macknick, J., Bush, B.W., Mohammadi, E., Ben-Horin, D., Urrea,
V., Marceau, D.: Understanding the life cycle surface land requirements of natural gas-fired
electricity. Nat. Energy 1 (2017)
7. Texas Gas Transmission, L.: Frequently Asked Questions. www.txgt.com/Safety.aspx?id=
1447. Accessed on 12 Dec 2016
8. Clark, C.E.; Han, J.; Burnham, A.; Dunn, J.B.; Wang, M.: Life-Cycle Analysis of Shale Gas
and Natural Gas. Argonne Nat Lab (2012)
9. Texas Railroad Commission. Oil and Gas Well Records—Online. https://www.rrc.state.tx.
us/oil-gas/research-and-statistics/obtaining-commission-records/oil-and-gas-well-records-onl
ine/. Accessed on 2 June 2017
10. Logan, J., Heath, G.A., Macknick, J., Paranhos, E., Boyd, W., Carlson, K.: Natural Gas and the
Transformation of the U.S. Energy Sector: Electricity, NREL/TP-6A50–55538. Joint Institute
for Strategic Energy Analysis (JISEA) (2012)
11. Williams Transco Central Penn Line South Pipeline Lifetime. Accessed on 29 2015
12. Texas Railroad Commission Memorandum: change in determination administrative policy for
gas well classification (2016)
Sci. 111, E3167–E3176 (2014)
14. Lee, W.J., Sidle, R.: Gas-reserves estimation in resource plays. SPE Econ. Manage. 2, 86–91
(2010)
15. INTEK Review of Emerging Resources: U.S. Shale Gas and Shale Oil Plays. Prepared for the
Energy Information Administration: Washington, D.C. (2011)
16. Baihly, J., Altman, R., Malpani, R., Luo, F.: Study assesses shale decline rates. The American
Oil & Gas Reporter, 5 (2011)
17. Warlick, D.: A current view of the top 5 US gas shales. Oil & Gas Financial Journal, 1 (2010)
18. EIA Annual Energy Outlook 2020. Department of Energy (2020)
19. Patzek, T.W., Male, F., Marder, M.: Gas production in the Barnett Shale obeys a simple scaling
theory. Proc. Natl. Acad. Sci. 110, 19731–19736 (2013)
20. Enverus Drilling Info Oil and gas production data (2020)
21. AGA How Does the Natural Gas Delivery System Work? https://www.aga.org/natural-gas/del
ivery/how-does-the-natural-gas-delivery-system-work-/. Accessed on 3 July 2020
22. Tavakkoli, S., Feng, L., Miller, S., Jordaan, S.M.: The implications of generation efficiencies
and supply chain leaks for the life cycle greenhouse gas emissions of natural gas-fired electricity
in the United States (under review)
23. EIA Natural gas compressor stations on the interstate pipeline network: developments since
1996. Office of Oil and Gas (2007)
24. Larson, A.: World’s most–efficient combined cycle plant: EDF Bouchain. Power (Magazine)
161, 22–23 (2017)
25. EIA What is the efficiency of different types of power plants? https://www.eia.gov/tools/faqs/
faq.php?id=107&t=3
26. Forsthoffer, W.E.: Forsthoffer’s best practice handbook for rotating machinery. Elsevier (2011)
27. Fadok, J.: Advanced gas turbine materials, design and technology. In: Advanced Power Plant
Materials, Design and Technology. Elsevier, pp. 3–31 (2010)
28. EIA Form EIA-923 detailed data. https://www.eia.gov/electricity/data/eia923/. Accessed on 8
Jan 2019
29. EIA Table 8.2. Average Tested Heat Rates by Prime Mover and Energy Source, 2009–2019.
https://www.eia.gov/electricity/annual/html/epa_08_02.html. Accessed on 14 Apr 2020
ment of greenhouse gas emissions from Liquefied natural gas trade for electricity generation.
sion and distribution. Nat. Clim. Chang. 9, 660–665 (2019)
32. GE GE Global Power Plant Efficiency Analysis. General Electric (GE) (2016)
Chapter 4
Life Cycle Impact Assessment
Abstract Across impact categories, electricity generated from natural gas faces
many of the same challenges as other product systems in life cycle assessment (LCA):
data. Due to the high levels of variability and uncertainty inherent to the natural gas
sector, life cycle impact results remain under investigation amongst scientists and
policymakers. The typical life cycle approach of characterizing results into aggre-
gate impact categories therefore can be less valuable for decision-makers in this
sector. The limitations to the assumptions and the depth of the analysis as well as
the impacts and processes of importance may not be obvious using such aggre-
gate indices. Consequently, specific life cycle impacts and their inventories will
be discussed after the presentation of aggregate results, including deep dives into
air emissions (including greenhouse gases), water consumption and withdrawal, and
land use. State-of-the-art methods for these example mid-point impact categories will
be reviewed, along with a critical analysis of the state of data and important chal-
lenges for future research. Where relevant, results for gas-fired power are presented as
comparisons to other electricity generation sources for context. A dilemma remains
for gas-fired power. Studies that include comprehensive suites of impact categories
provide an important picture of trade-offs thereby reducing the potential to overlook
negative consequences. Alternatively, focusing on one impact category using an in-
depth investigation that considers site-level information enables analysts to account
better for the inherent complexity of the product system.
Introduction
Environmental impacts can be defined as either positive or negative impacts to the

environment resulting from human intervention. LCA is a powerful tool to under-
standing the environmental impacts of products and processes. As discussed in
Chap. 2, a life cycle result is presented in terms of a numerator and a denomi-
nator. The numerator and denominator are quantitative representations of an envi-
ronmental impact and the functional unit, respectively. The life cycle inventory (LCI)
involves establishing all elementary flows that are relevant to a product system, in this
case, natural gas [1]. The life cycle impact assessment (LCIA) aims to assess the

46 4 Life Cycle Impact Assessment
contribution of each elementary flow (e.g., emission or resource use) to their respec-
tive environment impacts. To do so, impacts are placed in impact categories, which
often serve as indices of total impact. In fact, this is one of the mandatory steps of
LCIA—the selection of impact categories, category indicators and characterization
methods. LCIA and LCI for comprehensive LCAs are thus closely tied.
An important distinction must be made for the LCIA between midpoint and
endpoint indicators. Endpoint indicators relate to what is called “areas of protection”
which refer to what society may seek to protect, namely: the natural environment,
human health, and natural resources [2, 3]. An endpoint assessment will calculate the
magnitude of harm caused to these areas of protection by the life cycle of a product.
Alternatively, midpoint indicators characterize impacts between the emission and
the endpoint; for example, using global warming potential to quantify the emissions
of greenhouse gases in terms of carbon dioxide equivalents.
Natural gas-fired electricity involves a highly diffuse upstream natural gas supply
chain spread out over vast geographies, with a similarly expansive set of power plants
that use different technologies and are both influenced by and influencing their oper-
ating environments. As a result, the spatiotemporal disconnect between inventories
and impacts have created largest challenges for LCA in providing meaningful results
across impact categories. Many environmental impacts examined within LCAs are
fundamentally local, yet LCA often ignores spatial differentiation in inputs and
results [4]. Even for global challenges like climate change, the local context matters
for natural gas due to the site- and facility-level variability of elementary flows such
as atmospheric emissions.
In this chapter, a detailed overview of LCIA as it relates to gas-fired electricity
as well as outstanding questions that practitioners face in its implementation are
provided. Specific examples will be discussed to illuminate methodological gaps
and advancements in LCIA of relevance to natural gas-fired power. After intro-
ducing popular impact assessment models, three key life cycle impact categories
are reviewed in depth regarding their relationship with spatiotemporal assessment:
greenhouse gas implications of gas-fired power, water consumption implications of
the coal-to-gas transition of the electric sector, and the spatial land requirements and
impacts of gas-fired power and its alternatives. While many valuable impact assess-
ment models exist, such as the tool for the reduction and assessment of chemical
and other environmental impacts (TRACI) [5], the limits of knowledge are sought in
this chapter. The goal is to cover how present methods may be employed but also to
identify areas of future research for practitioners to improve upon present practice.
Impact Categories and Characterization Models
The steps in LCIA—the selection of impact categories, their classification, and char-
acterization as well as non-mandatory components—require a firm understanding
of how well developed the impact categories under question are. The International
Impact Categories and Characterization Models 47
Organization for Standardization (ISO) has LCA standards that recommend the selec-
tion should be comprehensive and span the environmental issues of importance to the
study in question. While such a case is optimal, challenges remain in many aspects of
methodological development—particularly for the case of gas-fired power—where
facility-level variability is high, and the development has immense coverage over
different regions and ecosystems. These challenges are manifested in how accurately
LCA practitioners can express the elementary flows (e.g., release of pollutants) in
terms of the implications to the endpoint impact categories (e.g., harm to human
health, the natural environment, or to natural resources).
While many impact categories are mature, many continue to develop, or remain
in need of development Sala et al. [6], leaving a lot of room for practitioners to
improve both data and methods (Fig. 4.1). Many of these challenges take root in not
only facility-level variability, but also in the fact that the different unit processes in
LCA can occur across many different regions and environments. The local environ-
ment may, for example, already be experiencing either high or low concentrations
of specific pollutants which may mean that different ecological or human health
thresholds may be closer to their limits. Specifically, the incremental impact from
product development may have a high or low impact on the baseline for where the
processes are occurring. The baseline is also not static. The amount of nitrogen in a
local area, for example, can vary with rainfall. Pre-existing local pollution loading
influences how harmful any new releases can be to the environment. These chal-
lenges are true for all products examined in LCA and are exemplified by natural gas
due to its diffuse geographical dispersion and evolution in operational practices and
technologies over time. In fact, these very questions make the impact assessment
stage of LCA particularly tricky—LCA often stops at the inventory stage and goes
right to the interpretation.
Regardless of these challenges, there have been powerful arguments for using LCA
to examine a comprehensive set of environmental impact categories. Most impor-
tantly, companies may choose to focus on only one impact that makes their product
look environmentally superior while (perhaps purposely) overlooking another impact
category with very negative environmental consequences [1, 7]. In fact, studies are
often commissioned to examine only one impact. As a result, more comprehensive
coverage is correctly encouraged (even required) to be certified as compliant to the
ISO standard.
Perhaps most referenced is the Intergovernmental Panel on Climate Change
(IPCC) method to characterize greenhouse gas emissions on a common basis (carbon
dioxide equivalents) using Global Warming Potential (GWP). Climate impacts of
fossil fuels have been increasingly disputed with growth of their use and broader
acceptance of climate science. As a result, such standardization is an enormous asset
to the field of LCA. The IPCC model, however, only examines greenhouse gases;
many other impact assessment models exist. Popular impact assessment models and
their impact categories are presented in Table 4.1 [7].
As recommended by ISO, LCAs of natural gas often examine environmental
impacts using the wide-reaching characterization models outlined in Table 4.1 but
are similarly challenged by questions of accuracy and broad applicability. And even
Midpoint impacts Endpoint impacts
48
Climate change Climate change
Ozone depletion Ozone depletion

Human health
Human toxicity, cancer Human toxicity, cancer
Human toxicity, non-cancer Human toxicity, non-cancer
PM/respiratory inorganics PM/respiratory inorganics
Ionising radiation (human health) Ionising radiation (human health)
Ionising radiation (ecosystems) Ionising radiation (ecosystems)

Mature methods
Photochemical ozone formation Photochemical ozone formation
Natural environment Methods in development
Environmental
cause and Methods in need of development
Eutrophication, aquatic Eutrophication, aquatic
Eutrophication, terrestrial Eutrophication, terrestrial
Ecotoxicity, freshwater Ecotoxicity, freshwater

Ecotoxicity, marine Ecotoxicity, marine
Ecotoxicity, terrestrial Ecotoxicity, terrestrial
Land use Land use

Natural resources
Resource depletion: water Resource depletion: water
Fossil and mineral depletion Fossil and mineral depletion
Renewable resources Renewable resources
Fig. 4.1 Impact categories are constantly improving, and many have yet to reach maturity. An illustrative depiction of this progress is reflected, with impact
assessment methods qualitatively determined to be mature, in development, or in need of development (adapted from Sala et al. [6])
4 Life Cycle Impact Assessment
Table 4.1 Popular impact assessment models Matthews et al. [7]
Climate Ozone Respiratory Human Ionizing Ecotoxicity Ozone Acidification Terrest. Aquatic Land Resource
change depletion inorganics toxicity radiation formation eutrophication eutrophication use consumption
CED X
CML2002 X X X X X X X X X X X
Eco-indicator X X X X X X X X X X
99
EDIP X X X X X X X X X X X
2003/EDIP976
EPS 2000 X X X X X X X X X X X X
Impact Categories and Characterization Models
Impact 2002+ X X X X X X X X X X X
IPCC X
LIME X X X X X X X X X X X
LUCAS X X X X X X X X X X
MEEuP X X X X X X X X X X
ReCiPe X X X X X X X X X X X X
Swiss X X X X X X X X X X X X
Ecoscarcity 07
TRACI X X X X X X X X X X
USEtox X X
49
these impact assessment models generally focus on midpoint categories rather than
endpoint. Generalized assessments can be useful in broad technological compar-
isons, however. Table 4.2 presents LCIA results from a 2015 study, where electricity
generated from natural gas can be broadly compared against other technologies [8].
With some exceptions, natural gas-fired power ranks firmly second to coal but higher
than other alternatives for nearly all impact categories. Exceptions include a higher
ranking of particulate matter for natural gas relative to coal and lower land occupation
relative to specific renewable sources. Studies that include broader, more compre-
hensive suites of impact categories are unlikely to fully capture the full variability
across regions and sites.
Table 4.2 Results for midpoint categories for different electricity generation technologies (adapted
from Hertwich et al. [8]). Results include waste disposal, well pre-production, and materials used
in the construction of infrastructure and are reported on the basis of a kilowatt-hour of electricity
generated. The geographic scope of the study is global
Greenhouse Particulate Land
Impact category Ecotoxicity Eutrophication
gases matter occupation
kg 1,4-DCB 2
Unit kg CO2 e. kg PM10 eq. kg P eq. ma
eq.
Subcritical wo CCS 0.933 0.0003 0.0096 0.0005 0.0204
IGCC wo CCS 0.791 0.0002 0.008 0.0004 0.0177
SCPC wo CCS 0.871 0.0003 0.0089 0.0005 0.0191
Subcritical w CCS 0.263 0.0004 0.0149 0.0007 0.0291
IGCC w CCS 0.201 0.0002 0.0109 0.0006 0.0238
SCPC w CCS 0.236 0.0004 0.0136 0.0006 0.0268
NGCC wo CCS 0.527 0.0008 0.0063 0 0.0005
NGCC w CCS 0.247 0.0009 0.0081 0 0.0007
Poly-Si ground 0.057 0.0001 0.0032 0 0.0099
Poly-Si roof 0.0575 0.0001 0.0037 0 0.0018
CIGS ground 0.0195 0 0.0005 0 0.0097
CIGS roof 0.0243 0 0.0007 0 0.0006
CdTe ground 0.0161 0 0.0005 0 0.01
CdTe roof 0.0206 0 0.0007 0 0.0006
CSP–trough 0.0227 0 0.0001 0 0.009
CSP–tower 0.033 0.0001 0.0004 0 0.014
Reservoir 1 0.0788 0.0002 0.0002 0 0.0044
Reservoir 2 0.0056 0 0 0 0.0262
Wind onshore 0.0084 0 0.0003 0 0.0003
Wind offshore steel 0.0114 0 0.0004 0 0.0003
Wind offshore gravity-
0.0111 0 0.0004 0 0.0003
based
Impact Categories and Characterization Models 51
While some ranges are presented in the study, the wide-ranging variability has yet
to be captured in a full assessment with hundreds of thousands of wells producing
in the United States alone. The spatiotemporal heterogeneity combined with the
uncertainty in impacts give rise to questions about the methodology and data inputs
specific to individual impact categories.
A Focus on Improving Individual Impact Categories
When both data inputs and life cycle results are highly variable and influenced by
region, spatial differentiation becomes even more important for the impact assess-
ment stage of LCA. Impacts related to air pollution, water consumption and ecosys-
tems require the integration of spatiotemporal considerations to provide context for
decision-makers. States may import natural gas for electricity generation, resulting
in impacts due to hydraulic fracturing in other regions. Even within a water rich state,
highly localized water demand may be incurred for drilling during a time when water
is scarce.
While LCA was not conventionally a spatially explicit method [4], new research
has been proving not only the need but also the possibility of improving both spatial
and temporal dimensions of LCA Jordaan et al. [9]. These impact categories can
then have contextual relevance for different regions, which is critical for the LCA
practitioner to communicate during the interpretation stage. Potting and Hauschild
[10] presented three categories of special LCA: site-generic, site-dependent, and site-
specific. Site-generic LCAs are the most common —they lack any spatial information
and assume homogeneity. Site-dependent LCAs can involve different scales of spatial
resolution, resulting in questions regarding what resolution should be applied (e.g.,
informed landscape type or resolution of the data [11]). Finally, site-specific LCAs
model individual sources and local responses. There is a trade-off between the amount
of site-specific information that is used in a LCA and the ability of decision-makers
to apply LCA results more broadly to a product system—the results may only apply
to that site or region.
Temporal accounting presents similar challenges for decisions related to the
impact assessment stage of LCA of natural gas-fired electricity. LCA quantifies
environmental impacts at a baseline year and according to time horizons selected
during the initial goal and scope definition stage [12], and results generally provide
a snapshot of impacts and resource consumption. Often data inputs were published
or acquired at different years and may be outdated.
Greenhouse Gas Emissions and Other Air Pollution
Recent questions about the life cycle greenhouse gas implications of natural gas-fired
electricity have centered upon whether upstream fugitive emissions may reverse
conclusions about the climate benefits compared to coal-fired power. The origins
of the debate around fugitive emissions take root in U.S. Environmental Protec-
tion Agency’s early estimates. They were used in several published LCAs. Studies
increasingly found that the emissions inventories were outdated, underestimating
actual emissions [13]. The search to uncover the magnitude of these emissions
ensued, with major updates by way of independent, published work and subsequent
revisions to many of the outdated EPA estimates. Such research has clearly had
substantial implications for decision-making. Another critical assumption in these
analyses is the time horizon used to calculate the carbon dioxide equivalence of
methane through Global Warming Potentials, which are based on 20-year and 100-
year time horizons. Since methane has higher warming potential relative to carbon
dioxide and an atmospheric lifetime of approximately 12 years relative to carbon
dioxide’s lifetime (the latter is mostly absorbed on the scale of centuries, but a
portion remains for 1000s of years) [13, 14]. The trade-offs between 20- and 100-
year time horizons remain under investigation and are discussed further in Chap. 6
on uncertainty analysis.
LCAs focused on energy technologies have contributed to uncovering relatively
hidden and uncertain contributors to emissions along energy supply chain and high-
lighted their importance for further investigation [15–18]. Emissions across the life
cycle are not always characterized with the same level of precision and accuracy [8].
More recently, the potential for uncertain fugitive methane emissions to evaporate
the benefits of natural gas-fired electricity when compared to coal has resulted in
international debate [16, 18–20].
Among related issues, the question of what time horizon should be explored
when estimating the relative global warming potential between methane emissions
and carbon dioxide remains unresolved. The greenhouse gas emissions associated
with gas-fired electricity have been variable, with results in this section representing
domestic production (the import of liquefied natural gas by tanker is examined in
Chap. 5). Life cycle phases characterized in process based LCAs include production,
processing, transportation, and storage, through end use (Fig. 4.2).
Greenhouse gas emissions from each of the life cycle stages of natural gas-
fired electricity are subject to substantial site- and project-level variability by region
and depending on many factors, such operational practices [21]. Local operational
practices inform whether mitigation measures are implemented. Petroleum geology
also plays into the equation. Site-level variability for facilities becomes even more
pronounced for oil and gas production due to the heterogeneity of geologic reser-
voirs [22–24]. Locational variability extends much farther than just regional differ-
ences in oil and gas reservoirs: greenhouse gas emissions from life cycle stages are
also influenced by national infrastructure differences, but these differences are not
well understood and may be measured differently in different jurisdictions [25]. For
natural gas production, even primary data from measurements are not always directly
comparable due to different reporting thresholds for facility-level emissions (e.g., 25
kilotonnes (kt) CO2 e a year for the U.S.; 10 kt for British Columbia (BC); 50 kt
for Alberta) [17]. Certainly, the methane leaks quantified in different LCAs have
been shown to be variable [26]. To address inconsistencies, the National Renewable
Energy Laboratory developed an analytical procedure that they applied specifically
to electricity generation technologies called harmonization [26–29]. The procedure
A Focus on Improving Individual Impact Categories 53
Fig. 4.2 Life cycle stages generally considered in process based LCAs of gas-fired power, both
fuel cycle and electricity generation [18]
involved modifying models, assumptions, and boundaries for consistency. Even once
key parameters and systems boundaries are harmonized, variability is reflected in life
cycle results (Fig. 4.3).
The results in Fig. 4.3 show little difference between life cycle emissions of elec-
tricity generated from the combustion of shale gas compared to that from conventional
natural gas. At the same time, the uncertainty of methane emissions remains. Most
notably, recent LCAs highlight the need to improve data and estimates related to
fugitive methane emissions [17, 20, 22, 30]. One of the most important contribu-
tions of LCA to energy research is the identification of data gaps and uncertainties
and their relative contributions to life cycle emissions estimates. Another reason to
undertake a LCA is to promote stakeholder dialogue, which has been particularly
effective for the energy sector regarding improving data reporting by both industry
and government alike. Such initiatives have been ongoing to improve not only data
but also mitigation.
The variability noted in greenhouse gas emissions transcends to other air pollu-
tants. One of the largest benefits of natural gas-fired electricity if used instead of
coal is the reduction of air pollutants at the stack. While the reductions are clear,
natural gas-fired electricity is not without impacts [31] and the air quality implica-
tions of high-density hydraulic fracturing have yet to be resolved (particularly with
diesel trucks used during drilling and volatile organic compounds released during
operations). Air pollutants such as ozone precursors, for example, have already been
shown to have air quality impacts and their impacts to human health are heavily
dependent on ambient air quality [32, 33].
Convenonal
Unconvenonal
Shale US domesc
Fig. 4.3 Harmonized life cycle estimates for different types of natural gas, compiled by Heath et al.
[27]. The study focused on comparing existing LCAs that consider shale gas and conventional, with
point estimates for unconventional and domestic for comparison. Results include waste disposal,
well pre-production, and materials used in the construction of infrastructure. The boxes in the
figure represent the interquartile range (IQR), which is the difference between the third quartile
(Q3) and the first quartile (Q1). The line in the box represents the median, the whiskers represent
the maximum and minimum. The x represents the mean
Even deploying higher efficiency combined cycle natural gas options may not
result in environmental benefits across all impact categories relative to coal [8, 33].
Combined cycle natural gas plants (see Chap. 3) have been reported to result in
higher NOx emissions than their coal counterparts, resulting in higher terrestrial
acidification potential (Table 4.2). Further, applying carbon capture and storage to
natural gas may increase all other impact categories by 20–80%, in part due to the
parasitic load and need for increased natural gas. It is important to note, however,
that many studies have reported close alignment between the NOx emissions from
natural gas and coal-fired power (0.2–3.9 kg NOx/MWh for coal options versus 0.2–
3.8 kg NOx/MWh for natural gas alternatives) [8], suggesting close attention should
be paid to regional differences in technologies and fuel quality in such comparisons.
Water Consumption
Typical LCAs focus on how the net life cycle water consumption of different types of
electricity generation may compare. Examining product systems can be informative;
however, water implications of the spatiotemporal patterns of energy decisions cannot
be captured by examining product systems alone. Importantly for gas-fired power,

hydraulic fracturing requires more water in comparison to conventional gas produc-
tion. Most studies point to the fact that water requirements for hydraulic fracturing
are relatively insignificant when compared to that for power generation based on each
Megawatt-hour generated [34, 35]. But LCAs of water consumption for energy are
also spatiotemporally dependent but often presented in spatially aggregate bar-charts,
without mapping spatially differentiated maps.
Water resource requirements can face highly localized impacts, which may result
in scarcity constraints. These impacts are typically characterized without connection
to spatially explicit inventories and ignore the temporal variability in streamflow.
Even those studies that attempt to relate water consumption with availability
typically rely upon administrative boundaries that represent political jurisdiction
rather than ecological systems [35–37]. The spatiotemporal dynamics of the water
consumed by energy decisions are exemplified by a case study of the coal-to-gas
transition of the electric sector of Pennsylvania [9, 38].
Results demonstrated that watershed-scale results are highly variable when using
spatially differentiated inventories of water consumption for two life cycle stages of
coal and gas-fired electricity generation. The life cycle stages included fuel extraction
and electricity generation in Pennsylvania from 2009 to 2012. We found that water-
shed-scale water consumption may be reversed from a net increase to decrease (and
vice versa) when the total water consumption is compared to the water consumed
specifically for life cycle stages of electricity generation. The finding reinforces the
importance of further developing spatially resolved inventories for LCAs of energy
technologies (Fig. 4.4).
Life cycle studies have generally focused on a specific region using national
political boundaries or subnational political boundaries [38–40]; however, water
impacts from energy technologies can be highly variable across and within each
of these regions (as shown by our example). An important emerging question for
LCA practitioners is: what spatial resolution should be employed to quantify water
use? Ecological scales, such as watersheds, are most relevant for understanding local
impacts to availability and ecosystems [41]. Yet, most water consumption data are
aggregated at political scales such as counties, states, provinces, or a country. More
recently, research has moved towards developing site-level inventories that can be
aggregated at the watershed scale [9, 38].
Another topic critical to developing a better understanding for water impacts in
LCA relates to the temporal variation of water use (both consumption and withdrawal)
compared to the availability of water. Water stress metrics have only recently been
connected to consumption in LCA with the “available water remaining” or AWARE
method, which quantifies the amount of fresh water available in each region as the
freshwater supply minus consumption (both human and environmental demand) [41–
43]. More remaining water means the burdens from new consumption are smaller, as
the potential for water deprivation caused by new consumption can be contextualized
with a consistent method [41–43]. With the emergence of life cycle inventories
of water consumption at the spatial resolution of watersheds and at the temporal
resolution of months (Fig. 4.5), future research may better connect inventories and
Fig. 4.4 Watershed-level results for the coal-to-gas transition in Pennsylvania [9]. a presents the
total water consumption changes associated with the coal to gas transition (Mgal) from 2009 to 2012
including fossil extraction and power generation (updated from Patterson et al. [38]); b power end
product water consumption changes (Mgal) for the coal-to-gas transition in the power sector where
the fuel assessment is limited to the fuel consumed by the sector from 2009 to 2012. The bars for
each watershed represent the magnitude of the increase or decrease in water consumption for each
industrial activity included within the scope of the study: gas extraction (Hydraulic Fracturing—HF
Extract), coal extraction (Coal Mine), natural gas power (NG Power), and coal power (Coal Power).
The color of each watershed represents the total increase or decrease in water consumption between
2009 and 2012
Fig. 4.5 Monthly water demand for two life cycle stages of coal and gas-fired electricity generation
in Pennsylvania for the coal to gas transition from 2009 to 2012 (fuel extraction and electricity
generation) for four watersheds. Water consumption across watersheds in Pennsylvania can found
using our online tool: https://nicholasinstitute.duke.edu/hydraulic-fracturing/
impacts. Such methods would provide a more robust decision-making context for
water managers in the face of stream-flow variability and its confluence with spikes
in water demand from energy development [38].
Quantification of Life Cycle Land Use Impacts
While many studies have examined the land use of energy technologies [22, 44–
50], recent research has noted the need for more robust comparative analyses of the
land use of electricity generation due to the inconsistent application of methods and
assumptions. It is therefore unsurprising that the spatial requirements of natural gas-
fired electricity have been less examined from a life cycle perspective in comparison
to other impact categories. In fact, land use provides a useful example of an impact
category that is less vetted in comparison to others.
The methodology employed to quantify land use in LCA can be complex: there are
many different impacts that can be considered and quantifying their value to society
can be challenging [45, 46, 51, 52]. In prior literature that builds upon the framework
for LCIA of land use developed by the United Nations Environment Programme
(UNEP)/Society of Environmental Toxicology and Chemistry (SETAC) Life Cycle

Initiative working group [11, 53–56], three types of land use impacts have been
identified: transformation, occupation, and permanent impacts. Land transformation
(LT ) arises from the change of land to a new type, typically for an intended human
activity such as energy developent, quantified in terms of area (e.g. meters squared
or m2 ). Land occupation (LO ) results from the human use of land over a duration of
time with original land type being a potential outcome post-reclamation, quantified in
terms of area-time (e.g., m2 -year). Permanent land use refers to irreversible impacts to
ecosystems when an ecosystem will not fully recover after use. Land transformation
and occupation are defined in Eqs. (4.1) and (4.2).
LT = A × Q (4.1)
Lo = A × t × Q (4.2)
where A is the area that is transformed or occupied, Q is land quality, and t is the
amount of time a piece of land is occupied. The difficulties in quantifying these
indicators are two-fold: the highly subjective concept of quality and the reliance
of results on the location of the land disturbance. A developer may look at a land-
scape with many projects and admire the multi-use while a conservationist may
look at the same landscape with disdain. The quality aspect will differ for different
biogeographical regions, impacts, and perspectives, thus requiring a more robust and
specific definition when used in practice (e.g., biodiversity or ecosystem services).
Much LCA research has focused on the concept of quality; however, researchers
may never agree on what it should encompass until values are made explicit within
the assessments. Specific challenges to the methodology relate directly to the fact
that ecosystem impacts are generally more acute locally and that there are many
different impacts that may result from the conversion of land. LCA thus has faced
a challenging dilemma as a field: practitioners often argue that the quantification of
land area alone is insufficient in understanding impacts (or the “quality” of land pre-
and post-transformation), so the investigation of land has often stagnated as even
understanding land transformation requires substantial analysis.
Recent methodological developments, including machine learning of aerial and
satellite images, have enabled improved quantification of the life cycle land require-
ments of gas-fired electricity but results generally remain focused on better quanti-
fying the amount of land required (i.e., transformation) [22]. The system boundaries
using such novel methods only capture the fuel cycle and electricity generation to
date. Unlike greenhouse gas emissions, upstream materials extraction has been found
to only have negligible impacts for renewable electricity; for example, upstream land
area associated with solar photovoltaic (PV) is estimated to be less than 1% of the
area related to electricity generation [57]. While specific to the Barnet Shale of
Texas, the applicability of new methods to other gas producing regions may in the
future illuminate variability associated with the fuel supply. When including the fuel
supply of natural gas—which is not applicable for renewable electricity—results
Table 4.3 Land use intensity of renewable electricity generation facilities reported by the National
Renewable Energy Laboratory [22, 47–49]. The annual results (m2 /(MWh/year)) and results for an
assumed a 30-year lifetime (m2 /MWh) for the average land use intensity of U.S. renewable electricity
generation facilities are reported. The number of projects sampled in the study are represented by
sample size n. Note that rooftop solar would not require additional land for the use phase of the life
cycle
Type of electricity generation Annual direct areaa (m2 /(MWh/year)) Direct area (m2 /MWh)
Natural gas-fired power – 0.62
Small Photovoltaicb 12 0.40
Large Photovoltaicc 13 0.43
Concentrating solar power 11 0.36
Geothermal 0.57 0.019
Wind 0.77 0.026
a Direct area refers to land permanently or temporarily disturbed due to physical infrastructure
development
b Small is defined as greater than 1 MW, but smaller than 20 MW
c Large is defined as greater than 20 MW, all ground mounted
suggest that the gas-fired power offers no real advantage over renewable sources in
this region. Our recent research determined that the total life cycle land requirements
of natural gas-fired electricity for the Barnett shale of Texas was 0.62 m2 /MWh
(95% confidence intervals ±0.01 m2 /MWh). We found that total land use was domi-
nated by midstream infrastructure, particularly pipelines (74%). More surprisingly,
we demonstrated that renewable energy may require less land when time horizons
representative of the economic life of the technologies are applied (Table 4.3).
Though renewable energy is often considered as more destructive to landscapes
when compared to gas-fired power, the application of new methods may reveal more
evidence that confirms that this may not always hold true when upstream impacts—
notably the fuel cycle—are considered. While some of the present set of results
sometimes appear contradictory, different methods have been applied to compare
renewable energy with non-renewable that treat time differently. The impacts of
the different temporal assumptions on results are exemplified in Table 4.3. There
is a clear need to robustly examine the amount of land required to maintain energy
infrastructure across the life cycle of energy technologies using clear and transparent
spatiotemporal data and consistent assumptions. More specifically, the timescales
at which different types of renewable energy becomes equivalent to fossil energy
warrant investigation.
Going beyond simply the land area required for electricity generation, the “qual-
ity” aspect can represent many valued aspects of land, from carbon mitigation to
biodiversity to cultural values. This impact category remains nascent relative to
many others, which can be exemplified by ecosystem services. Even beyond LCA,
it is well-recognized that there is no consensus for valuing ecosystem services [58].
Rugani et al. [59] noted the diversity of methods that can be used to accomplish this
goal. Among the many approaches that can be used to monetize ecosystem services
in LCA, the Economics of Ecosystems and Biodiversity (TEEB) framework has

been broadly applied. This well-established framework [60, 61] assigns values to
four major categories: provisioning (e.g. food production), regulating (e.g. carbon
storage), habitat (e.g. gene pool protection), and cultural services (e.g. recreation
and aesthetic). Ecosystem values for the Chihuahuan Desert region were recently
estimated by Taylor et al. [62, 63] and applied within a comparative LCA of elec-
tricity generated from natural gas, solar, and wind [64]. Particularly when consid-
ering time within the calculations, the results illuminate interesting trends in how
each technology compares over time (Fig. 4.6).
The average land use intensity of gas-fired generation for the region was esti-
mated to be 0.31 m2 /MWh with a standard deviation of 0.30 m2 /MWh (Fig. 4.6).
Results for the region were just under half of those for the recent study in the Barnett
shale of Texas, suggesting that regional variability can be influential in life cycle
quantifications of land use for this pathway. The average land use intensity of solar
was determined to be 0.68 m2 /MWh (standard deviation of 0.31 m2 /MWh), but the
breakeven with natural gas is under 10 years. The average ecosystem services cost
per MWh for gas-fired generation is estimated to be 0.63 $/MWh with a standard
Fig. 4.6 a–d. Land use and ecosystem services costs of energy projects in the Chihuahuan Desert
Region. a, b present results for the land use intensity and c, d focus on ecosystem services. Figures
republished with permission from Jordaan et al. (in press) [64]
Table 4.4 Estimated time of

Ecosystems Time (years)
restoration for select
ecosystem types [65] Agricultural land, pioneera vegetation <5
Species-poor grasslands, mature pioneer 5–25
vegetation
Oligotrophicb vegetation, species-rich 25–50
marshland, grasslands
Species-rich forests, shrubs, hedgerows 50–200
Immature peatbogs, old dry grasslands 200–1000
Mature peat bogs, old growth forests 1000–10,000
a Pioneer vegetation refers to plants early in the succession after a
piece of land has been cleared
b Nutrient-poor
deviation of 0.60 $/MWh, with the cumulative average only reaching a breakeven
with solar power just under 50 years (assuming no reclamation). Wind remains
the smallest ranking for both metrics, with averages of be 0.064 m2 /MWh (standard
deviation of 0.041 m2 /MWh) and 0.12 $/MWh (standard deviation of 0.084 $/MWh)
respectively for land use intensity and ecosystem services costs.
Further complicating the characterization of land use impacts, there has been
no broadly accepted methodology regarding the treatment of time horizons in LCAs
specific to those LCAs focused on land use of energy technologies [64, 65]. To clarify,
should a horizon refer to the amount of time that a feature remains on the landscape or
the amount of time it takes for the land to be restored? The latter more conventional,
LCA approach suggest a ‘relaxation time’ representing the time required for the
ecosystem to re-grow. The results are again spatially-dependent and will rely on
the ecoregion where the energy development occurs (Table 4.4) Restoration times
for ecosystems may hold substantial differences in duration relative to the facility
lifetimes noted in Chap. 3.
The issue of choosing the appropriate time horizon for land use elicits more
subjective questions. At what point does one consider the land restored? Again, this
is a value judgment where the answer depends on the stakeholder. Neither restoration
time nor facility lifetime is necessarily correct, for example, but the assumption
should be made explicit. Even once an ecosystem has re-grown, stakeholders may
or may not feel satisfied with the restoration.
Conclusions
In conclusion, there is a trade-off made by the LCA practitioner in the goal and scope
definition in LCAs of gas-fired power that should be made clear in the communication
of results. Broad and comprehensive analyses that include a wide set of impact cate-
gories aggregated to end-point indices can be valuable because they are less likely to
overlook specific impacts. At the same time, they may lack depth in the examination
of the fuel supply, or they may be limited to a small, manageable population of infras-
tructure that may not represent the full variability of the product system. Conversely,
in-depth analyses of specific facilities and infrastructure across the supply chain can
advance the understanding of regional impacts or inter-regional variability. Deci-
sions about spatial scale, boundaries, and detail should be made during the goal and
scope definition stage and aligned with the purpose of the LCA. Of course, the inven-
tory stage may reveal that initial decisions regarding spatial details must be revised
according to data quality and availability.
These initial decisions can serve to contribute improved LCIA methods and are
notable for the case of this product system that covers large tracts of lands and extends
over diverse regions. Regional and site-level variability of impacts across projects
is particularly pronounced for the natural gas sector, pointing to the importance of
robust uncertainty assessment and careful data quality critiques (to be addressed in
Chap. 6). If a LCA practitioner chooses metrics that are site generic to characterize
the entire product system, the results for specific impact categories that depend on
local environmental conditions are unlikely to hold significance or meaning for the
regions where impacts occur (e.g., water consumed by natural gas extraction in a
specific watershed). Decisions should be clearly aligned with the LCA purpose and
refined according to data availability as the LCA proceeds. There is opportunity for
improving methods with either approach.
References
2. Capaz, R.S., Seabra, J.: Life cycle assessment of biojet fuels. In: Biofuels for Aviation. Elsevier,
pp. 279–294 (2016)
3. Finnveden, G., Hauschild, M.Z., Ekvall, T., Guinée, J., Heijungs, R., Hellweg, S., Koehler, A.,
Pennington, D., Suh, S.: Recent developments in life cycle assessment. J. Environ. Manage.
91, 1–21 (2009)
4. Reap, J., Roman, F., Duncan, S., Bras, B.: A survey of unresolved problems in life cycle
assessment. Int. J. Life Cycle Assess. 13, 374–388 (2008)
5. Bare, J.C.: TRACI: The tool for the reduction and assessment of chemical and other
environmental impacts. J. Indus. Ecol. 6(3-4), 49–78.
6. Sala, S., Pant, R., Hauschild, M., Pennington, D.: Research needs and challenges from science
to decision support. Lesson learnt from the development of the international reference life
cycle data system (ILCD) recommendations for life cycle impact assessment. Sustainability 4,
1412–1425 (2012)
approaches for decisions that matter. Retrieved June 2018, 1
8. Hertwich, E.G., Gibon, T., Bouman, E.A., Arvesen, A., Suh, S., Heath, G.A., Bergesen, J.D.,
Ramirez, A., Vega, M.I., Shi, L.: Integrated life-cycle assessment of electricity-supply scenarios
confirms global environmental benefit of low-carbon technologies. Proc. Natl. Acad. Sci. USA.
112, 6277–6282 (2015)
9. Jordaan, S.M., Patterson, L.A., Anadon, L.D.: A spatially-resolved inventory analysis of the
water consumed by the coal-to-gas transition of Pennsylvania. J. Clean. Prod. (2018)
References 63
10. Potting, J., Hauschild, M.: Spatial differentiation in life cycle impact assessment: a decade of
method development to increase the environmental realism of LCIA. Int. J. Life Cycle Assess.
11, 11–13 (2006)
11. Lindeijer, E.: Review of land use impact methodologies. J. Clean. Prod. 8, 273–281 (2000)
12. Tillman, A.M., Baumann, H.: The Hitchhikers Guide to LCA. Studentlitteratur AB, Lund,
Sweden (2004)
13. Ehhalt, D., Prather, M., Dentener, F., Derwent, R., Dlugokencky, E.J., Holland, E., Isaksen,
I., Katima, J., Kirchhoff, V., Matson, P.: Atmospheric chemistry and greenhouse gases. In
Climate Change 2001: The Scientific Basis, Third Assessment Report. Working Group I,
Intergovernmental Panel on Climate Change (2001)
14. Archer, D., Brovkin, V.: The millennial atmospheric lifetime of anthropogenic CO2 . Clim.
Change 90, 283–297 (2008)
15. Brandt, A.R., Farrell, A.E.: Scraping the bottom of the barrel: greenhouse gas emission conse-
quences of a transition to low-quality and synthetic petroleum resources. Clim. Change 84, 241
(2007)
17. Searchinger, T., Heimlich, R., Houghton, R.A., Dong, F., Elobeid, A., Fabiosa, J., Tokgoz, S.,
Hayes, D., Yu, T.H.: Use of U.S. croplands for biofuels increases greenhouse gases through
emissions from land-use change. Science 319, 1238–1240 (2008)
18. Konschnik, K., Jordaan, S.M.: Reducing fugitive methane emissions from the North American
oil and gas sector: a proposed science-policy framework. Clim. Policy 18(9), 1133–1151 (2018)
19. Alvarez, R.A., Pacala, S.W., Winebrake, J.J., Chameides, W.L., Hamburg, S.P.: Greater focus
needed on methane leakage from natural gas infrastructure. Proc. Natl. Acad. Sci. U. S. A. 109,
6435–6440 (2012)
343, 733–735 (2014)
23. O’Sullivan, F., Paltsev, S.: Shale Gas Production: Potential Versus Actual GHG Emissions,
Report 234. MIT Joint Program on the Science and Policy of Global Change (2012)
24. Montgomery, J.B., O’Sullivan, F.M.: Spatial variability of tight oil well productivity and the
impact of technology. Appl. Energy 195, 344–355 (2017)
25. Coleman, J.W., Kasumu, A., Liendo, J., Li, V., Jordaan, S.M.: Calibrating liquefied natural
gas export life cycle assessment: Accounting for legal boundaries and post-export markets.
Canadian Institute of Resources Law, Occasional Paper, (49) (2015)
26. National Renewable Energy Laboratory Life Cycle Assessment Harmonization. https://www.
nrel.gov/analysis/life-cycle-assessment.html. Accessed on 30 May 2018
Sci. U. S. A. 111, 3167 (2014)
28. Heath, G.A., Mann, M.K.: Background and reflections on the life cycle assessment harmoniza-
tion project. J. Ind. Ecol. 16 (2012)
30. Reed, A.: The Natural Gas Sector: Life-Cycle Greenhouse Gas and Air Quality Issues. Nova
Science Publishers, Inc., New York (2016)
31. Kemball-Cook, S., Bar-Ilan, A., Grant, J., Parker, L., Jung, J., Santamaria, W., Mathews, J.,
Yarwood, G.: Ozone impacts of natural gas development in the Haynesville Shale. Environ.
Sci. Technol. 44, 9357–9363 (2010)
32. Litovitz, A., Curtright, A., Abramzon, S., Burger, N., Samaras, C.: Estimation of regional air-
quality damages from Marcellus Shale natural gas extraction in Pennsylvania. Environ. Res.
Lett. 8, 014017 (2013)
33. Gibon, T., Arvesen, A., Hertwich, E.G.: Life cycle assessment demonstrates environmental
co-benefits and trade-offs of low-carbon electricity supply options. Renew. Sustain. Energy
Rev. 76, 1283–1290 (2017)
34. Mielke, E., Anadon, L.D., Narayanamurti, V.: Water consumption of energy resource extraction,
processing, and conversion. Belfer Center for Science and International Affairs (2010)
35. Scown, C.D., Horvath, A., McKone, T.E.: Water footprint of US transportation fuels. Environ.
Sci. Technol. 45, 2541–2553 (2011)
36. Jordaan, S.M., Diaz Anadon, L., Mielke, E., Schrag, D.P.: Regional water implications of
reducing oil imports with liquid transportation fuel alternatives in the United States. Environ.
Sci. Technol. 47, 11976–11984 (2013)
37. Grubert, E.A., Beach, F.C., Webber, M.E.: Can switching fuels save water? A life cycle quan-
tification of freshwater consumption for Texas coal-and natural gas-fired electricity. Environ.
Res. Lett. 7, 045801 (2012)
38. Patterson, L.A., Jordaan, S.M., Diaz Anadon, L.: A Spatiotemporal Exploration of Water
Consumption Changes Resulting from the Coal-to-Gas Transition in Pennsylvania. Belfer
Center for Science and International Affairs, Harvard Kennedy School and Nicholas Institute
for Environmental Policy Solutions, Duke University (2016)
39. Jiang, M., Hendrickson, C.T., VanBriesen, J.M.: Life cycle water consumption and wastewater
generation impacts of a Marcellus shale gas well. Environ. Sci. Technol. 48, 1911–1920 (2014)
40. Mutel, C.L., Pfister, S., Hellweg, S.: GIS-based regionalized life cycle assessment: how big is
small enough? Methodology and case study of electricity generation. Environ. Sci. Technol.
46, 1096–1103 (2011)
41. Boulay, A., Bare, J., Benini, L., Berger, M., Lathuillière, M.J., Manzardo, A., Margni, M.,
Motoshita, M., Núñez, M., Pastor, A.V.: The WULCA consensus characterization model for
water scarcity footprints: assessing impacts of water consumption based on available water
remaining (AWARE). Int. J. Life Cycle Assess. 23, 368–378 (2018)
42. Ansorge, L., Beránková, T.: LCA water footprint AWARE characterization factor based on
local specific conditions. Euro. J. Sustain. Dev. 6, 13 (2017)
43. Lee, U., Xu, H., Daystar, J., Elgowainy, A., Wang, M.: AWARE-US: quantifying water stress
impacts of energy systems in the United States. Sci. Total Environ. 648, 1313–1322 (2019)
44. Hernandez, R.R., Hoffacker, M.K., Field, C.B.: Efficient use of land to meet sustainable energy
needs. Nat. Clim. Change 5, 353 (2015)
45. Jordaan, S.M.: The land use footprint of energy extraction in Alberta. University of Calgary.,
Calgary, Alberta (2010)
46. Ciliberti, C., Jordaan, S.M., Smith, S.V., Spatari, S.: A life cycle perspective on land use and
project economics of electricity from wind and anaerobic digestion. Energy Policy 89, 52–63
(2016)
47. Ong, S., Campbell, C., Denholm, P., Margolis, R., Heath, G.: Land-Use Requirements for Solar
Power Plants in the United States. National Renewable Energy Laboratory, Golden, CO (2013)
48. Denholm, P., Hand, M., Jackson, M., Ong, S.: Land use requirements of modern wind power
plants in the United States (2009)
49. Ong, S., Campbell, C., Heath, G.: Land Use for Wind, Solar, and Geothermal Electricity
Generation Facilities in the United States. A report from the National Renewable Energy
Laboratory to the Electric Power Research Institute, National Renewable Energy Laboratory
(2012)
50. Jordaan, S.M., Keith, D.W., Stelfox, B.: Quantifying land use of oil sands production: a life
cycle perspective. Environ. Res. Lett. 4, 024004 (2009)
References 65
51. Canals, L.M., Clift, R., Basson, L., Hansen, Y., Brandão, M.: Expert workshop on land use
impacts in life cycle assessment. 12–13 June 2006 Guildford, Surrey (UK). Int. J. Life Cycle
Assess. 11, 363–368 (2006)
52. i Canals, L.M., Bauer, C., Depestele, J., Dubreuil, A., Knuchel, R.F., Gaillard, G., Michelsen,
O., Mller-Wenk, R., Rydgren, B.: Key elements in a framework for land use impact assessment
within LCA (11 pp). Int. J. Life Cycle Assess. 12, 5–15 (2007)
53. Spitzley, D.V., Tolle, D.A.: Evaluating land-use impacts: selection of surface area metrics for
life-cycle assessment of mining. J. Ind. Ecol. 8, 11–21 (2004)
54. Koellner, T., Scholz, R.W.: Assessment of land use impacts on the natural environment. Int. J.
Life Cycle Assess. 13, 32–48 (2008)
55. Koellner, T., De Baan, L., Beck, T., Brandão, M., Civit, B., Margni, M., i Canals, L.M., Saad,
R., De Souza, D.M., Müller-Wenk, R.: UNEP-SETAC guideline on global land use impact
assessment on biodiversity and ecosystem services in LCA. Int. J. Life Cycle Assess. 18,
1188–1202 (2013)
56. De Baan, L., Alkemade, R., Koellner, T.: Land use impacts on biodiversity in LCA: a global
approach. Int. J. Life Cycle Assess. 18, 1216–1230 (2013)
57. Murphy, D.J., Horner, R.M., Clark, C.E.: The impact of off-site land use energy intensity
on the overall life cycle land use energy intensity for utility-scale solar electricity generation
technologies. J. Renew. Sustain. Energy 7, 033116 (2015)
58. de Souza, D.M., Lopes, G.R., Hansson, J., Hansen, K.: Ecosystem services in life cycle assess-
ment: a synthesis of knowledge and recommendations for biofuels. Ecosyst. Serv. 30, 200–210
(2018)
59. Rugani, B., de Souza, D.M., Weidema, B.P., Bare, J., Bakshi, B., Grann, B., Johnston, J.M.,
Pavan, A.L.R., Liu, X., Laurent, A.: Towards integrating the ecosystem services cascade frame-
work within the life cycle assessment (LCA) cause-effect methodology. Sci. Total Environ. 690,
1284–1298 (2019)
60. De Groot, R.S., Wilson, M.A., Boumans, R.M.: A typology for the classification, description
and valuation of ecosystem functions, goods and services. Ecol. Econ. 41, 393–408 (2002)
61. Sukhdev, P., Wittmer, H., Schroter-Schlaack, C., Nesshover, C., Bishop, J., Ten Brink,
P., Gundimeda, H., Kumar, P., Simmons, B.: Mainstreaming the economics of nature. A
synthesis of the approach, conclusions and recommendations of TEEB. Go to www. teebweb.
org/Swanson T & Johnston S (1999) Global Environmental Problems and International
Environmental Agreements: The Economics of International Institution Building (2010)
62. Taylor, N.T., Davis, K.M., Abad, H., McClung, M.R., Moran, M.D.: Ecosystem services of the
big bend region of the Chihuahuan desert. Ecosyst. Serv. 27, 48–57 (2017)
63. Taylor, N.T., Davis, K.M., Abad, H., McClung, M.R., Moran, M.D.: Corrigendum to
“ecosystem services of the big bend region of the chihuahuan desert”. Ecosyst. Serv. 27, 48–57
(2017). Ecosyst. Serv. 30, 192 (2018)
64. Jordaan, S.M., Lee, J., McClung, M., Moran, M.: Quantifying the ecosystem services of elec-
tricity generation in the U.S. Chihuahuan Desert using a life cycle perspective. J. Indus. Ecol.
(in press)
65. Scholz, R.: Assessment of land use impacts on the natural environment. Part 1: an analytical
framework for pure land occupation and land use change (8 pp). Int. J. Life Cycle Assess. 12,
16–23 (2007)
66. Charpentier, A.D., Bergerson, J.A., MacLean, H.L.: Understanding the Canadian oil sands
industry’s greenhouse gas emissions. Environ. Res. Lett. 4, 014005 (2009)
67. Zavala-Araiza, D., Lyon, D.R., Alvarez, R.A., Davis, K.J., Harriss, R., Herndon, S.C., Karion,
A., Kort, E.A., Lamb, B.K., Lan, X., Marchese, A.J., Pacala, S.W., Robinson, A.L., Shepson,
P.B., Sweeney, C., Talbot, R., Townsend-Small, A., Yacovitch, T.I., Zimmerle, D.J., Hamburg,
S.P.: Reconciling divergent estimates of oil and gas methane emissions. Proc. Natl. Acad. Sci.
U. S. A. 112, 15597–15602 (2015)
68. Turconi, R., Boldrin, A., Astrup, T.: Life cycle assessment (LCA) of electricity generation
technologies: overview, comparability and limitations. Renew. Sustain. Energy Rev. 28, 555–
565 (2013)
69. Jordaan, S.M., Diaz-Anadon, L., Schrag, D.: Temporal Aspects of Water Use in Shale Gas
Development: Implications for LCA. American Center for Life Cycle Assessment: Chicago
Illinois (2011)
70. Smil, V.: Energy at the Crossroads: Global Perspectives and Uncertainties. MIT Press,
Cambridge, MA (2003)
71. Fthenakis, V., Kim, H.C.: Land use and electricity generation: a life-cycle analysis. Renew.
Sustain. Energy Rev. 13, 1465–1474 (2009)
72. Sproul, E., Barlow, J., Quinn, J.C.: Time value of greenhouse gas emissions in life cycle
assessment and techno-economic analysis. Environ. Sci. Technol. 53, 6073–6080 (2019)
73. O’Hare, M., Plevin, R.J., Martin, J.I., Jones, A.D., Kendall, A., Hopson, E.: Proper accounting
for time increases crop-based biofuels’ greenhouse gas deficit versus petroleum. Environ. Res.
Lett. 4, 024001 (2009)
74. Schenck, R., White, P.: Environmental life cycle assessment: measuring the environmental
performance of products. American Center for Life Cycle Assessment Vashon, Washington
(2014)
Chapter 5
Global Markets and Competitiveness
of Gas-Fired Power
Abstract The combination of hydraulic fracturing and horizontal drilling has

decreased and stabilized prices of natural gas since 2009. Gas-fired power has subse-
quently become a more competitive option and has since become the dominant
generation source in the United States, holding promise to similarly displace coal in
other regions that depend on higher carbon fuels. On the long term, this newfound
competitiveness presently faces challenges with plummeting costs of renewables and
marginal increases expected in gas prices. Further confounding these challenges,
the emerging liquefied natural gas (LNG) market has fierce competition amongst
suppliers. This chapter examines the competitiveness of natural gas vis-à-vis other
generation alternatives, the economic landscape facing the LNG market as well as
the life cycle environmental implications of this bourgeoning industry.
Introduction
The economics of gas-fired power have improved over time with higher efficiency
combustion and more recently with more stable, low natural gas prices. These more
stable prices stem from the shale gas reserves that have been unlocked in regions
across the world since 2009 by the combination of horizontal drilling and hydraulic
fracturing. With these improvements, natural gas has been more competitive in elec-
tricity markets. Prices dropped with the expansion of reserves and more efficient
extraction, making gas an attractive alternative to coal for baseload power (particu-
larly when coupled with high efficiency power generation). Before the shale boom,
the largest challenge to natural gas-fired electricity was rooted in variable natural gas
prices which made investors wary when comparing to stable, cheap coal. Post-boom,
the increasing number of highly efficient natural gas power plants have profited from
stable, low prices in the United States.
The resulting low costs of producing natural gas have made global trade more
attractive where landed natural gas prices are high enough. A burgeoning liquefied
natural gas (LNG) industry has emerged because of technological advances and
landed natural gas prices in regions where demand growth has outpaced supply.
These twin developments—the shale boom and emerging markets—have resulted in

68 5 Global Markets and Competitiveness of Gas-Fired Power
economic opportunities related to the potential for international trade of LNG [1].
These emerging developments have implications for the life cycle environmental
impacts of gas-fired power. The import and export of natural gas in the form of LNG
requires new processes in the natural gas supply chain to ensure that the natural
gas can be transported safely and economically. Natural gas must be cooled to a
liquid state (i.e., liquefied) at approximately −260 F so that it can be reasonably
shipped and stored using smaller volumes compared to the gaseous state [2]. Once
the LNG arrives at its destination, it is then regasified and transported to the end uses.
These new processes required for transportation of natural gas in liquid form (e.g.
liquefaction, ocean transport, and regasification) and its use in the importing nation
can have substantial impacts on life cycle results.
The traditional perspective on natural gas markets has been evolving, however,
with increasing global LNG trade that has created links between previously discon-
nected markets. The favorability of the economics of LNG varies by natural gas
producing region but also by importing nation and the associated landed natural gas
prices[3]. Prior to recent demand growth during 2010–2020, the disparate global
natural gas markets were typically viewed as three regional markets with similar
trends in prices, namely North America, Europe and Asia–Pacific [4]. More recently,
the prospect of the convergence of these markets with the potential of natural gas
becoming a globally traded commodity has sparked debate [5].
This chapter covers the evolution of the economics of gas-fired power, covering
the levelized costs of electricity, natural gas production costs, markets and prices.
The discussion is then expanded to examine the evolution of global markets and the
resulting implications for the life cycle of gas-fired power. While the discussion of
life cycle assessment (LCA) will focus on greenhouse gas emissions, insights are
provided into the implications for other impact categories.
Levelized Cost of Electricity
The evolution of the economics of natural gas-fired electricity over time can be
evaluated most simply using the levelized cost of electricity (LCOE). LCOE is the
cost of generating a unit of electricity, informing estimates of the revenues required
to build and operate a power plant over a specified economic lifetime (i.e., a cost
recovery period) [6]. It depends on both capital and Operations and Maintenance
(O&M) costs, the financial parameters of the project (e.g., interest rate, plant life,
etc.) and the plant’s capacity factor. LCOE is generally used as a convenient metric
to estimate the overall competitiveness of a technology or a project. It is defined as
follows.
LCOE = FC + VC (5.1)
Levelized Cost of Electricity 69
where LCOE is the cost of electricity, FC are the fixed costs, and VC are the variable
costs. The fixed costs (FC) do not depend on fuel, rather the total overnight costs and
O&M costs (Eq. 5.2).
CAPEX × CCF + FOM

FC = (5.2)
CF × MWnet
CAPEX represents the plant’s capital cost and mainly depends on the sizes of
different equipment which are generally a function of the flow rates. The total capital
requirement of the plant includes the direct costs of equipment as well as the indirect
costs such as contingencies. CCF is the capital charge factor (or the fixed charge
factor), which is used to annualize capital costs over the economic lifetime of the
power plant. CCF depends on the plant life and financial variables such as discount
rate, as defined shortly in Eq. (5.5). FOM represents the fixed O&M costs. The fixed
costs (FC) are adjusted according to the capacity factor, since a plant does not operate
at its full output for the entire year. CF is the capacity factor or the effective fraction of
time the power plant operates at full capacity in a year (average plant output divided
by its peak output), and MWnet is the net capacity or the net power that the plant
can generate.
VC = FUEL × HR + VOM (5.3)
The variable costs are defined as the sum of the fuel cost (FUEL) accounting for
the heat rate (or the amount of fuel used in operations per unit electricity generated)
and the Variable O&M (VOM) (Eq. 5.3). For the case of natural gas-fired electricity,
the variable prices in natural gas before the boom were highly influential in LCOE
calculations, comprising a substantial portion of the costs [7]. Fuel costs remain an
important factor in evaluating the competitiveness of gas-fired power.
Combining Eqs. (5.2 and 5.3) provides a more detailed calculation of the LCOE,
including the inputs to FC and VC (Eq. 5.4). Representative units are included for
additional clarity; for example, the unit for LCOE is $/MWh.
⎡
⎤

CAPEX $ × CCF 1
+ FOM $
$ year year
LCOE =⎣ ⎦
MWh 8760 year × CF × MWnet
hour
⎡ ⎤
$
$ HR BTU VOM MWyear
+ ⎣FUEL × kWh
+ ⎦ (5.4)
BTU 1000 8760 houryear
The capital charge rate (CCF) used to annualize CAPEX payments is defined in
Eq. (5.5).
r
CCF = +r (5.5)
(r + 1)T − 1
Table 5.1 Example results

Interest rate, r (%) Number of years, T CCF (%)
for different assumptions
applied in calculating the 5 25 7
capital charge factor (CCF) 10 30 11
15 20 16
where r is the interest rate and T is the number of time periods used to estimate the
capital recovery (e.g. 20 one-year increments over 20 years). For reference, several
calculted CCFs are presented in Table 5.1.
Even when ranging finanicial assumptions, estimates suggest that natural gas
remains a solid competitor against all other technologies given low natural gas prices
and highly efficienct technologies (Fig. 5.1). The upper range estimates for LCOE
in Fig. 5.1 represent the case with carbon capture and storage (CCS). The techno-
logical innovation that enabled the shale boom—hydraulic fracturing and horizontal
drilling—has derisked the fuel from past volatility that may have deterred invest-
ments. Gas-fired power is now competitive enough to supply a baseload as well as
a flexible response to intermittent renewable power. The method to estimate LCOE
results presented in Fig. 5.1 follows the general idea of equations presented in this
section with improvements [8]. For example, the National Renewable Energy Labo-
ratory (NREL) publishes a model with their Annual Technology Baseline (ATB) that
delineates high, low, and average capacity factors. Estimates for LCOE in the ATB
model vary as a function of the capacity factor. Their mode also includes the tax rate,
depreciation rate, inflation rate, and debt to equity ratios.
Understanding how LCOE may change over time provides valuable insights into
its future viability, particarly in a carbon-constrained scenario (Fig. 5.2). Notably,
estimates for solar’s LCOE have plummetted from $360/MWh to $40/MWh over the
past decade. Figure 5.2 examines the case of CCS implementation, thus reflecting
higher costs for fossil options. Natural gas generation has experienced increasing
efficiencies (thus lower heat rates) in combination with lower fuel prices and a larger
2020 ATB LCOE Range by Technology for 2018 Based on R&D Financial Assumpons
Biopower $86 $115
Hydropower $49 $96
Geothermal $58 $424
Alternative Energy
Solar - CSP $89 $111

Solar - PV Dist. Residenal $88 $137
Solar - PV Dist. Commercial $65 $102
Solar - Ulity PV $29 $47
Offshore Wind $84 $172
Land-Based Wind $28 $127
Nuclear $76
Coal $72 $166
Natural Gas $30 $96
$0 $50 $100 $150 $200 $250 $300 $350 $400 $450

Levelized Cost of Energy ($/MWh)
Fig. 5.1 Range of LCOE estimates by technology (NREL ATB [8]). The range of fossil fuel
estimates – natural gas and coal – include options with and without carbon capture and storage
Levelized Cost of Electricity 71
R&D
90
Coal (with CCS)
80
70 Nuclear
60
Natural gas (CC with CCS)
LCOE 50 Hydropower
($/MWh)
40 Offshore wind
30
Onshore wind
20
Utility-scale solar photovoltaic

10
0
2020 2025 2030 2035 2040 2045 2050
Filtered by: Scenario - Moderate; Financial Case - R&D; Cost Recovery Period - 30 years
Fig. 5.2 Forecasted changes in LCOE for key technologies using NREL’s 2020 annual technology
baseline (ATB) [8]. The natural gas LCOE represents combined cycle (CC) technology with carbon
capture and storage (CCS). Similarly, the coal LCOE also includes CCS
share of baseload power supply (thus increasing capacity factors). Accounting for
these changes in a simplified analysis of LCOE, the overall effect is decreasing costs
of gas-fired power in recent years (see Chap. 3 for more information). Regardless,
the substantial advances in renewable technologies have led to a more competitive
landscape.
While the short-term decreases in costs since the shale boom are notable, longer-
term price forecasts examine a number of potential outcomes. The Energy Informa-
tion Administration’s reference case, for example shows a modest increase in prices,
but remaining just under $4/mmBTU [9]. In the absence of technological improve-
ments, gas-fired power costs adjust accordingly. Along the same lines, IEA forecasts
similar increases. Such modest increases are reflected in LCOE forecasts, as shown
by Fig. 5.2, where fuel costs increase from $2.81 to $3.99 between 2020 and 2050.
Clearly, one can not precisely predict exactly how the price of natural gas will evolve
in the United States or elsewhere, suggesting a critical role for uncertainty analysis
(methods to be discussed further in Chap. 6). Similar trends in LCOE are noted for
combined cycle gas-fired generation without CCS, but with costs reduced by 33%.
The Economics of Natural Gas Production
One of the critical inputs for LCOE is the price of natural gas, which informs the
variable costs. But similar to electricity generation, natural gas will not be produced
if the costs are not adequately surpassed by the profits from selling the natural gas.
Production Costs
As defined by the Energy Information Administration, the costs of natural gas produc-
tion include drilling, completion, construction, and operation [10]. For onshore
natural gas production, drilling comprises 30–40% of total well costs, including
tangible costs that are capitalized and depreciated over time (e.g., well casing) and
intangible costs that can be expensed (e.g., fuel and drilling fluids). Completion costs,
which also include tangible and intangible costs, comprise 55–70% of the total well
costs. Specialized workers who perform fracturing and other activities related to
natural gas production systems connect the supply to the transportation system. The
construction of facilities contributes 7–8% of the total costs and includes activi-
ties such as road construction, site preparation, surface equipment for operations
(e.g., storage tanks and separation units). Operating costs experience high variability
across wells and depend on several factors such as location and well productivity.
They include well maintenance costs, but also those costs incurred in delivering the
gas to a purchase point or pricing hub.
While cost-benefit analyses capture the aforementioned costs, external and recla-
mation costs are generally not realized up front in financial decisions [11, 12]. Costs
of reclamation may be as low as tens of thousands of dollars, extending to a million
with only a fraction of these costs covered by bonds [12]. Indeed, the average cost
of plugging a well alone in Pennsylvania depends on factors such as depth, site
conditions, and geographic location, and finally decommissioning and reclamation
averaging $37,000 per well but ranging from approximately $1000 to $1,000,000 per
well (presented in 2015 USD) [13].
Natural Gas Markets
Global natural gas markets have experienced substantial changes in recent years,
even before the COVID-19 pandemic. Natural gas has not historically been a global
commodity, with three regional markets of natural gas, namely North America,
Europe and Asia–Pacific. Recently, increasing demand for natural gas from China
and India has resulted in more attractive options for sellers, with prices even spiking
over $16/mmBTU with the additional demand shock of nuclear shut-downs post-
Fukushima (Fig. 5.3). The price trends reveal the importance of the two price mech-
anisms used in different regions involved with the international trade of natural gas
The Economics of Natural Gas Production 73
Fig. 5.3 Landed natural gas prices in international markets
[4]. North America uses the gas-on-gas mechanism that responds to supply and
demand and competitive gas market spot prices. With the recent shale boom, gas
prices have become more stable due to over-supply. In the Asia–Pacific region and to
a lesser extent Europe natural gas prices are indexed to oil market spot prices nudging
prices higher, particularly when combined with a sharp demand spike on top of
steady growth. More recently, the COVID-19 pandemic has created a demand shock,
decreasing demand and prices in key regions. The International Energy Agency
presently anticipates that after this shock will be a time of recovery proceeded by
slow but steady growth.
Total natural gas trade volumes increased by 2.9% in 2019, with 53% by pipeline
and the remaining 47% by LNG [14]. The shift is notably towards LNG growth;
in 2013, approximately 70% of the natural gas was traded by pipelines, while the
remaining 30% was traded by LNG carriers [15]. In 2019, three countries, the
United States, Russia and Australia, accounted for more than 75% of the incre-
mental supply with growth placing the United States as the world’s third-largest
LNG supplier after Qatar and Australia. That said, gas prices fell in 2019 with supply
surpassing demand in key consuming regions. Importantly for the burgeoning LNG
sector, LNG spot prices fell but only by 3% in major importers—China, Japan, and
Korea—since their gas prices continue to have a dependence on oil indexation in the
LNG import contracts [14].
The COVID-19 pandemic has resulted in what the International Energy Agency
refers to as “Unchartered Macroeconomic Territory” with an unprecedented set of
global restrictions, including lockdowns, that have substantially impacted supply and
demand of natural gas. The result for natural gas is an expected 4% decrease in global
natural gas demand for 2020, followed by a slow recovery. By 2025, even LNG trade
is still expected to increase by 21% relative to a 2019 baseline. The one thing that is
certain regarding a potential recovery from COVID-19 at the time of this writing is
the uncertainty, not only for the natural gas sector but for the world.
Additional Costs Associated with LNG
The additional costs due to liquefaction, shipping, and regasification can influence
the viability of LNG trade and use in electricity generation. Landed natural gas price
can be volatile, highlighting the importance of understanding pricing systems and
how they may change over time. Table 5.2 presents both domestic costs as well the
difference between costs of shipping through regasification and landed natural gas
prices in the importing country. Natural gas prices and costs associated with LNG
trade are representative of the case where the gas is shipped from the United States to
a variety of regions. Data represent average prices from 04/2017–11/2019, reflecting
Bloomberg data available to the author at the time of data collection. The purchase
price reflects the U.S. Henry Hub natural gas price (EIA) for the same time period
and the liquefaction costs were published in a recent study [16]. Comparing the total
for the exporting columns to the difference in the importing section of Table 5.2
Table 5.2 Generalized LNG economics, including domestic and international supply and demand,
average 04/2017–11/2019
Importing Exporting country (example of the Transport through sale
country United States)
Purchase Liquefaction Total Gas Price Regasification Shipping Difference
price
China 2.67–4.09 2.00–4.00 4.67–8.09 9.05 −0.5 −1.87 6.67
India 9.04 −0.5 −1.83 6.71
Japan 9.24 −0.5 −2.13 7.1
South 9.23 −0.5 −1.74 6.93
Korea
Chile 9.24 −0.5 −1.97 7.36
Singapore 9.23 −0.5 −1.85 6.77
Argentina 8.16 −0.5 −1.05 6.61
Spain 6.45 −0.5 −0.83 5.12
Great 6.67 −0.5 −0.82 5.34
Britain
The Economics of Natural Gas Production 75
suggests only a very small margin for LNG to remain competitive without additional
incentives.
The price of natural gas clearly has substantial implications for the profitability of
its use in electricity generation, bringing into question to how it fares against available
alternatives across regions—a question that warrants more comprehensive analyses
than this chapter can provide. Of course, the difference between the costs and price
has evolved over time, with variability primarily driven by natural gas prices but also
shipping to a much lesser degree (Fig. 5.4). Compared to the domestic costs incurred
by operators in the exporting country (Table 5.2), these findings point to the need for
long-term off-take agreements and contracts in final investment decisions related to
this industry.
Implications of LNG for Life Cycle Assessment of Gas-Fired

Power
The evolution of global natural gas markets has been influential in LCA of gas-fired
power, primarily due to the emergence of LNG opportunities [4, 17, 18]. Results are
influenced by both the location of the potential markets for LNG (mostly regions with
growing demand outpacing existing supply) as well as the potential exporters. There
is not only site-level variability in environmental and economic factors that depend
on where the natural gas is extracted and consumed but also varying technologies
and operations for the different sectors across life cycle stages.
The implications of expanding LNG trade for electricity generation will be
discussed in this section, with emphasis on greenhouse gas emissions. In addition
to reviewing recent works, emissions from liquefaction, transport, and regasification
will be examined, enabling a more comprehensive characterization of uncertainty.
The discussion is separated into two parts linked to downstream life cycle stages: the
first part focuses on the liquefaction through to the delivery to the power plant and
the second part analyses the electricity generation up to the end use (Table 5.3).
Liquefaction to Power Plant
The emissions incurred by the additional supply chain segments for electricity gener-
ated from LNG trade include liquefaction, tanker berthing and reberthing, ocean
transport, regasification, and losses through the natural gas pipeline network to the
power plant. Natural gas transmission and distribution experience losses through
the pipeline network, which must be differentiated from power sector transmission
and distribution (T&D) losses. The latter is discussed in the next section. These
additional segments result in higher life cycle emissions when compared to the
case where the gas is combusted domestically (assuming equivalent power plant
76
12 China
India
Japan
South Korea
10 Chile
Singapore
Argenna
Spain
8 Great Britain
$/MMBTU
4
Fig. 5.4 The difference between landed natural gas prices and costs incurred by shipping and regasification
5 Global Markets and Competitiveness of Gas-Fired Power
Implications of LNG for Life Cycle Assessment of Gas-Fired Power 77
Table 5.3 The LNG life cycle segments analyzed in this study presented in this table are organized
by upstream and downstream of liquefaction, and whether the values used are shared across all
countries or unique to each country (Adapted from Kasumu et al. [19])
Upstream Downstream (part 1) Downstream (part 2)
Liquefaction to power plant Electricity generation up to end
use
Construction Liquefaction Power Plant Operations
Drilling Tanker berthing and deberthing Electricity T&D Losses
(Direct)
Fracturing LNG ocean transport
Completion LNG Regasification
Flaring Natural gas T&D losses
Lease energy (production)
Plant emissions (processing)
Vented CO2
Fugitive well
Fugitive plant
Workovers
Liquids unloadings
Compression
Fugitive NG transmission
efficiency). Exporting natural gas using liquefaction, transport by tanker and regasi-
fication has already been found to contribute approximately 15–20% of the total life
cycle emissions of natural gas-fired electricity (not including the T&D losses in the
power sector), depending on the input assumptions (Table 5.4) [19].
Liquefaction alone contributes on the order of 10% to the life cycle emissions.
The average of published emissions for liquefaction amount to 5.83 g of carbon
dioxide per Megajoule of natural gas (g-CO2-e/MJ) with a standard deviation of
1.93 [18]. The variability in emissions associated with the liquefaction component
of the life cycle has been reported as relatively small, particularly in comparison to
the other factors that drive the variability across the life cycle, such as the efficiency
of electricity generation and transmission and distribution power losses [19]. In
Table 5.4, a range of results from Kasumu et al. [4] are provided which is specific
to export of LNG from Canada that is used for combustion in gas-fired electricity
in specific markets. From a review of potential markets, twelve potential importers
were identified: Belgium, Germany, Spain, the United Kingdom, Turkey, Argentina,
Brazil, China, India, Japan, South Korea, and Taiwan. A static assumption about
the emissions associated with power generation is show in Table 5.4, which will be
further explored in the next section.
In a world of perfect data, each life cycle segment associated with liquefaction
through to end use would vary by country; however, disaggregated data were not
Table 5.4 Comparison of life cycle stages of gas-fired power, including the additional LNG segments (adapted from results presented in the supporting
78
information of Kasumu et al. [4]). Results presented in grams of CO2 equivalent per kilowatt-hour
Baseline Construction Drilling Fracturing Completion Flaring Lease energy Plant Vented Fugitive Fugitive
(production) emissions CO2 well plant
(processing)
Howarth 0.0 0.0 0.0 75.7 0.0 0.0 37.9 43.8 0.0 0.0
Burnham 25.6 48.0 0.0 0.0 29.1 0.0 0.0 0.0 0.0 0.0
Jiang 1.1 1.8 3.1 10.1 85.0 0.0 37.7 0.0 0.0 0.0
Venketash 0.0 0.0 0.0 0.0 85.0 0.0 37.7 0.0 0.0 0.0
Stephenson 0.0 3.5 1.3 9.9 33.1 0.0 0.0 0.0 0.0 0.0
Weber 1.3 1.6 2.6 9.5 4.7 25.3 0.0 9.5 21.4 14.2
Fulton 93.9 0.0 0.0 0.0 25.8 0.0 0.0 0.0 0.0 0.0
NETL 0.8 0.0 0.0 10.2 0.0 0.0 25.1 0.0 14.2 2.6
JISEA 17.4 0.0 0.0 24.9 7.5 0.0 23.6 0.0 0.0 0.0
Laurenzi and Jersey 0.0 0.0 0.0 77.1 0.0 0.0 10.8 0.0 0.0 0.0
BC—10kt threshold 0.0 0.0 0.0 23.2 0.0 0.0 23.6 0.0 0.0 0.0
HHS
GHGenius 4.0.3 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0
HHS—BC—2012
Skone 2012 0.1 0.0 0.0 8.7 0.0 0.0 4.6 0.0 10.4 2.4
Marcellus
Skone 2012 Barnett 0.7 0.0 0.0 7.4 0.0 0.0 4.6 0.0 10.4 2.3
Hultman et al 0.0 0.0 0.0 39.9 1.5 0.0 0.0 0.0 18.5 5.4
(continued)
Table 5.4 (continued)
Workovers Liquids Compression Fugitive Liquefaction LNG Tanker LNG Power Plant Total
unloadings transmission transport Berthing and Regasificatio n Operations
Deberthing
0.0 5.2 0.0 20.9 64.7 32.8 1.5 17.7 388 688
0.0 0.0 0.0 20.9 64.7 32.8 1.5 17.7 388 628
0.0 0.0 12.3 0.0 64.7 32.8 1.5 17.7 388 656
0.0 0.0 12.3 0.0 64.7 32.8 1.5 17.7 388 640
0.0 0.0 15.4 0.0 64.7 32.8 1.5 17.7 388 568
9.5 0.0 3.2 15.0 64.7 32.8 1.5 17.7 388 623
0.0 0.0 33.2 0.0 64.7 32.8 1.5 17.7 388 658
36.2 0.0 3.4 18.2 64.7 32.8 1.5 17.7 388 615
0.0 0.0 23.6 0.0 64.7 32.8 1.5 17.7 388 602
0.0 0.0 38.0 0.0 64.7 32.8 1.5 17.7 388 631
0.0 0.0 6.6 0.0 64.7 32.8 1.5 17.7 388 558
0.0 11.3 6.2 20.6 64.7 32.8 1.5 17.7 388 543
24.2 0.0 20.6 16.7 64.7 32.8 1.5 17.7 388 592
26.2 0.0 21.6 16.7 64.7 32.8 1.5 17.7 388 595
Implications of LNG for Life Cycle Assessment of Gas-Fired Power
39.9 0.0 0.0 17.2 64.7 32.8 1.5 17.7 388 627
79
available for this level of granularity for several downstream processes at the time of
the analysis. For the segments that are not available in disaggregated, granular form,
we relied on the U.S. Department of Energy/National Energy Technology Laboratory
(DOE/NETL) study for common values that could be reasonably applied across all
countries.
In the absence of operational values related to LNG export, the DOE/NETL
study informed our estimates for emissions arising from natural gas liquefaction,
tanker berthing and deberthing, LNG regasification [15]. Similarly, we used emis-
sions factors for ocean transport from the DOE/NETL report [20], and applied them
to transportation distances from Kitimat, BC to prospective export destinations for
the ocean transport values. For these point estimates, we applied the medium values
in the DOE/NETL report for the baseline which provides low, medium, and high
estimates of the emissions factor. The values were weighted according to the market
potential of each country, though results disaggregated by country are provided in
the next section.
The value presented in the base dataset represents an average that is weighted
by the market potential of each country. This was estimated by first determining the
ocean distances in nautical miles from the export port (Kitimat, BC) to a selected
LNG regasification terminal in the destination country. Using the ocean distance from
the export port, the emission factor from the DOE/NETL report was then adjusted
to determine the LNG transport emission factor of each destination. Finally, we
calculated a weighted average LNG transport emission factor (representative of the
expected trade from Kitimat) based on the size of the electricity generation of each
country.
Electricity Generation to End Use
Before the recent growth in LNG trade, analysts would primarly concern themselves
with domestic production of natural gas and, without the LNG segments, consider
domestic consumption in the power sector. Such an analysis is relatively simple in
comparison to the international context, particularly when considering multiple coun-
tries. Of course, efficiencies and T&D losses will vary within a country; however, in
cases like the United States, data required to estimate unit-level efficiencies and heat
rates are relatively well-published by organizations such as the Energy Information
Administration [21, 22]. Country-level efficiencies and T&D losses are more chal-
lenging to access at the international scale, particularly when seeking more granular
data within each country [23]. Regardless, data for many countries is published by
organizations such as the World Energy Council [24].
To better understand the segments associated with electricity generation through
end use, my colleagues and I [4] examined two important factors observed to vary
across countries: power plant efficiency and electricity T&D losses. Table 5.5 shows
how these results were found to differ by country in 2016. While it is valuable
Implications of LNG for Life Cycle Assessment of Gas-Fired Power 81
Table 5.5 Base values of country specific values describing natural gas plant efficiency, T&D
losses, and emissions expected from ocean transport of LNG from Kitimat, BC, to each respective
location
Country NG Plant Eff (HHV) Direct T&D losses (%) Power Plant operations
adjusted for NG Plant Eff
Belgium 0.51 0.05 351.6
Germany 0.44 0.04 405.5
Spain 0.56 0.10 323.7
UK 0.53 0.08 342.4
Turkey 0.51 0.15 351.8
Argentina 0.46 0.15 387.5
Brazil 0.45 0.15 399.1
China 0.46 0.06 395.0
India 0.41 0.20 444.2
Japan 0.47 0.05 381.1
South Korea 0.44 0.03 406.3
Taiwan 0.52 0.04 343.0
to understand the overall trends in variability, we note that robust sets of country-
specific data would better capture regional variation, highlighting the importance of
continuing research and data generation related to this topic.
The full results, including all segments, demonstrate that the burgeoning LNG
sector adds substantial implications for LCA (Fig. 5.5). Notably, power generation
efficiencies are subject to substantial variation across countries, and even across
individual plants. Further, T&D losses mean that more than one kilowatt hour
of electricity needs to be generated for each kilowatt hour consumed, since elec-
tricity is lost in transit. This compensatory electricity generation, in turn, results
in higher greenhouse gas emissions. These factors were found to be substantial
contributors to country-level variability in LNG export emissions when examining
the implications of combusting the fuel to generate electricity (Fig. 5.5) [4].
Conclusions
The natural gas sector has been undergoing a substantial transformation since the
shale boom began in 2009; the technological innovation resulting from the combi-
nation of hydraulic fracturing and horizontal drilling has influenced how natural
gas is used across the world. Not only have the economics changed, but the overall
viability as a fuel for power generation has become more attractive due to low price
and increasingly efficient generation technologies. This new economic environment
has encouraged the emergence of a LNG market, which creates new sources of natural
gas for countries that do not have large resource endowments as well as new demand
82
1000
750
Life cycle emissions (gCO2e/kWh)
500
Argentina Belgium Brazil China Germany India Japan South Korea Spain Taiwan Turkey UK U.S. - 55% U.S. - 46.4% U.S. w/o LNG U.S. w/o LNG
(State of Art) (46.4%) (55%)
Fig. 5.5 Life cycle results for liquefied natural gas (LNG) export from Canada vary significantly based on country-level differences in infrastructure [4].
Emissions are calculated based on a unit of electricity delivered, including power transmission and distribution. The boxes in the figure represent the interquartile
range (IQR), which is the difference between the third quartile (Q3) and the first quartile (Q1). The line in the box represents the median, the whiskers represent
the maximum and minimum
Conclusions 83
for natural gas producers. Regardless, the future of natural gas is not without both
economic and environmental challenges. Renewable power is expected to continue
to become increasingly competitive, including when coupled with storage options
that can reduce or eliminate intermittency issues. The additional costs associated
with LNG lower the bar for more competitive options in the electric sector, casting
doubt on its role in the global power sector. While these challenges remain, evalua-
tions of the overall competitiveness of gas-fired power should account for important
advantages such as flexibility and dispatchability. Even in domestic analyses, the
LCOE of gas-fired power is expected to incrementally increase while renewable
options continue to plummet. While LCOE is an informative indicator, it can be
highly misleading as it fails to account for the costs of intermittent or inflexible oper-
ation [25]. System level analyses are needed to better characterize how such costs
may evolve in the future, considering changes in the grid as well as the advantages
natural gas may offer.
With the evolving economics of not only natural gas but also power generation, the
LNG market that has emerged has implications for our understanding of LCA of gas-
fired electricity generation. The additional environmental burdens are incurred from
the new supply chain segments that are required. Specifically, liquefaction through
regasification adds to life cycle emissions and other impact categories remain in ques-
tion. Of critical importance to the environmental outcomes of this sector is a better
understanding of the methane leaks across the supply chain of natural gas. Green-
house gases alone will not capture the full implications of the sector to environmental
sustainability; analyses focused on individual and combined impact categories can
provide important insights. The development and publication of comprehensive, site-
level, international datasets that characterize all supply chain segments have yet to
be confirmed and would support the development of important contributions.
References
1. Boersma, T., Jordaan, S.M.: Whatever happened to the golden age of natural gas? Energ. Trans.
1, 5 (2017)
2. EIA Natural gas explained: Liquefied natural gas: (2020)
3. Maxwell, D., Zhu, Z.: Natural gas prices, LNG transport costs, and the dynamics of LNG
imports. Energ. Econ. 33, 217–226 (2011)
5. Barnes, R., Bosworth, R.: LNG is linking regional natural gas markets: evidence from the
gravity model. Energ. Econ. 47, 11–17 (2015)
6. EIA Levelized Cost and Levelized Avoided Cost of New Generation Resources in the Annual
Energy Outlook 2020. (2020)
7. Breeze, P.: Chapter 10—the cost of electricity generation from natural gas-fired power plants.
In: Breeze, P. (ed.) Gas-Turbine Power Generation, pp. 93–98. Academic Press (2016)
8. Akar, S., Beiter, P., Cole, W., Feldman, D., Kurup, P., Lantz, E., Margolis, R., Oladosu, D.,
Stehly, T., Rhodes, G.: 2020 Annual Technology Baseline (ATB) Cost and Performance Data
for Electricity Generation Technologies: (2020)
9. EIA Annual Energy Outlook 2020 with projections to 2050: (2020)

10. EIA, U.S.: Trends in US oil and natural gas upstream costs. In: US Energy Information
Administration (2016)
11. Davis, L.W.: Policy Monitor—bonding requirements for US natural gas producers. Rev.
Environ. Econ. Policy 9, 128–144 (2015)
12. Mitchell, A.L., Casman, E.A.: Economic incentives and regulatory framework for shale gas
well site reclamation in Pennsylvania. Environ. Sci. Technol. 45, 9506–9514 (2011)
13. Kang, M., Mauzerall, D.L., Ma, D.Z., Celia, M.A.: Reducing methane emissions from
abandoned oil and gas wells: strategies and costs. Energ. Policy 132, 594–601 (2019)
14. BloombergNEF/International Gas Union/Snam 2020 Global Gas Report 2020: (2020)
15. GIIGNL The LNG Industry in 2013; Report for International Group of Liquefied Natural Gas
Importers. In: International Group of Liquefied Natural Gas Importers (GIIGNL) (2014)
16. Songhurst, B.: LNG Plant Cost Reduction 2014–18. (2018)
17. Coleman, J.W., Kasumu, A.S., Liendo, J., Li, V., Jordaan, S.M.: Calibrating liquefied natural
gas export life cycle assessment: accounting for legal boundaries and post-export markets. In:
CIRL Occasional Paper #49. Canadian Institute of Resources Law (2015)
18. Abrahams, L.S., Samaras, C., Griffin, W.M., Matthews, H.S.: Life cycle greenhouse gas emis-
sions from US liquefied natural gas exports: implications for end uses. Environ. Sci. Technol.
49, 3237–3245 (2015)
19. Kasumu, A., Li, V., Coleman, J.W., Liendo, J., Jordaan, S.M.: Country-level life cycle
assessment of liquefied natural gas trade for electricity generation. Submitted
20. Skone, T., Engineer, P.S.E.: Role of alternative energy sources: Natural gas technology
assessment. In: NETL Office of Strategic Energy Analysis and Planning (2012)
21. Energy Information Administration, Form EIA-923 detailed data with previous form data.
Accessed 08 Jan 2019
22. Energy Information Administration, Form EIA-860 detailed data. https://www.eia.gov/electr
icity/data/eia860/. Accessed 25 Oct 2017
24. World Energy Council Energy Efficiency Indicators. https://wec-indicators.enerdata.net/.
Accessed 30 May 2018
25. Joskow, P.L.: Comparing the costs of intermittent and dispatchable electricity generating
technologies. Am. Econ. Rev. 101, 238–241 (2011)
Chapter 6
Tackling Uncertainty Across the Life
Cycle of Gas-Fired Power
Abstract Uncertainty assessment is a critical component of the interpretation stage

of life cycle assessment (LCA) for practitioners because it enables decision-makers
to understand how impacts may change under different conditions. LCAs of gas-
fired power embody the importance of uncertainty assessment. On the one hand,
the large networks of infrastructure result in irreducible variability of both process
efficiency and localized environmental impacts. On the other hand, large questions
remain over the magnitude of leaks in the supply chain of the fuel cycle, further
confounded by quickly evolving operational practices and regulatory environments.
This chapter will include an overview of the importance of uncertainty assessment in
LCA exemplified by gas-fired power, followed by a summary of frequently employed
methods illustrated with examples of their use related to this product system.
Introduction
In navigating every-day life decisions, one is frequently faced with uncertain data
in deliberations, from booking tickets to an outdoor concert that may or may not be
rained out to making investments in the stock market. By uncertainty, I refer to the lack
of complete information or data required to complete a full decision analysis. From
a more quantitative perspective, one can envision uncertainty as a deviation from
the (perhaps unrealistic idea) of completely deterministic knowledge about a system
under study [1]. From either qualitative or quantitative perspective, the importance
of acknowledging uncertainty is embedded in making more robust decisions. When
investing in outdoor concert tickets, one knows that the weather on a particular day in
the future is uncertain. Similarly, a stock market can behave in unexpected ways that
are driven by unforeseen events. More informed decisions can be made by observing
trends and evaluating probability such that likely outcomes can be discerned and
incorporated into planning. If an outdoor event occurs on a day with 60% change of
precipitation, one might bring an umbrella. Similarly, if the demand for a particular
technology is anticipated to grow exponentially due to an expanding new market,
it might be a good time to invest. In either case, one can make an informed deci-
sion, but the outcome is uncertain. These concepts underlying many of our everyday

86 6 Tackling Uncertainty Across the Life Cycle of Gas-Fired Power
choices transcend to the interpretation stage of LCA when the practitioner trans-
lates the results in meaningful ways for the decision-maker. An informed decision
should acknowledge uncertainty, necessitating a robust understanding of uncertain
inputs on the outcomes of decisions. Gas-fired power exemplifies the need for uncer-
tainty assessment in LCA, due to the high variability of inputs and uncertain impacts
inherent to sectors across the life cycle.
Uncertainty is not to be confused with variability, which refers to the inherent
heterogeneity in data associated with a system [2, 3]. The former may be reduced
with improved data (e.g., unknown emissions from the natural gas supply chain)
whereas the latter cannot be reduced but rather better characterized with superior
information (e.g., the diversity of operational efficiencies present in natural gas power
plants operating today). Both uncertainty and variability are critical concepts for
better understanding decisions in LCA and, more broadly, comprehending the full
distribution of outcomes and their probabilities enables decision-makers to be well-
prepared for potential outcomes. Defining these distributions for gas-fired power
requires an investigation of the populations of different facilities across the underlying
network of infrastructure.
The two most important concepts for investigating uncertainty in decision-making
about any product system are utility and probability. Utility has historically been
viewed in terms of monetary value but is more commonly defined as the overall
value one places on the outcomes; here, we focus on the environmental outcomes of
electricity generated from natural gas. Indeed, as uncertainty assessment proceeds,
one should continually revisit fundamental questions. What is the value to society
or stakeholders of understanding the uncertainties under consideration? Can one
demonstrate which uncertainties may impact results in ways that are important
to decisions? Probability is a quantitative statement about chance associated with
different outcomes that are important to a decision. If data inputs are unknown in a
LCA model, probability becomes a powerful way to understand potential influences
on the results.
In this chapter, uncertainty assessment in LCA is introduced, along with founda-
tional concepts of probability and methods that are utilized to characterize uncertainty
in LCA. Embedded in each section are examples specific to the life cycle of gas-
fired power. Basic concepts of uncertainty and LCA are first introduced, followed by
the methods and their application. It is important to note that uncertainty assessment
generally requires the LCA practitioner to identify, characterize, analyze (uncertainty
and sensitivity analyses), and communicate uncertainty to support recommendations
for their audience, which may be comprised of the public, specific stakeholders,
report commissioners, or relevant decision-makers [4].
Uncertainty in LCA
Uncertainty is present throughout the LCA process. Igos et al. (2019) developed a
useful guide for practitioners in the identification of uncertainties, recommending
Uncertainty in LCA 87
Table 6.1 Examples of uncertainty in LCA, classified into the location and nature of uncertainty
[4]
LCA stage Examples Context Model Quantity Epistemic Ontic
structure
Goal and Definition of the
scope functional unit:
– Consistency with X X
goal and scope
– Representativeness X X
of performance
characteristics
– Variability of X X
performance
characteristics
Definition of system
boundaries:
– Technological X X
representativeness
– Geographical X X
representativeness
– Temporal X X
representativeness
Definition of LCI
modeling framework:
goal and scope
– Consistency of X X
multi-functionality
solving approach
Definition of LCIA
methodology and
indicators:
goal and scope
of environmental
impacts
Life cycle For each foreground
inventory unit process:
of the flows
considered
of the relationships
(continued)
Table 6.1 (continued)

LCA stage Examples Context Model Quantity Epistemic Ontic
structure
– Variability of the X X
relationships with
reference flow
of the background
processes used
of data used for
each flow
– Variability of each X X
flow
Life cycle For each LCIA
impact indicator:
assessment – Representativeness X X
of the characterized
substances
of the modeling
structure
– Variability of the X X
relationships
of each CF
– Variability of each X X
CF
that as comprehensive a list as possible should be developed (Table 6.1) [4]. In doing
so, uncertainties were placed in different categories, identifying where in the LCA
process uncertainties may be present.
There are five key uncertainty categories in this approach: context, model struc-
ture, quantity, epistemic, and ontic. Context uncertainty takes root in the normative
choices made by the LCA practitioner that reflect their beliefs or objectives; for
example, the choice of functional unit and impact weighting can influence outcomes.
Model structure uncertainty relates to the mismatch between the LCA model and
reality. A very common example is how generic characterization factors do not gener-
ally represent local conditions. Process-based models tend to be more representative
but are limited by the system boundaries and available data. Uncertainties in process-
based results may be induced, for example, by assuming linearity between dose and
response in the LCIA stage or in the choice of input assumptions made in building
an inventory. For the latter, such uncertainty prevails across model types; however,
they may be amplified in process models where only a portion of the supply chain is
analyzed (as defined by the system boundaries). While greater model complexity may
reduce uncertainty in comparison to a simple process-based model, the uncertainty
Uncertainty in LCA 89
with the more extensive input data requirements is likely to reciprocally increase [5].
Additionally, choices about allocation methods can add to uncertainty.
Uncertainty in any data used for either inventory analysis or LCIA can be catego-
rized as quantity uncertainty. Quantity uncertainty often guides the decisions about
uncertainty analysis in LCA [6]. Such uncertainty is often handled using the methods
outlined later in this chapter and its management requires research into what is known
about the data inputs. Uncertainties may arise from errors in measurement, where
measurement devices may result in differing results. One can imagine that such
uncertainty may become important for methane emissions across the natural gas
supply chain where numerous measurement techniques are employed, resulting in a
wide range of results. Parameter uncertainty refers to the uncertainty in all input data.
Such uncertainty may, for example, refer to missing or incomplete data. Temporal
uncertainty may arise if data are not current. Geospatial data refers to how different
operations, designs, and geographical environments may influence results.
Clearly, it may be possible that the knowledge is simply lacking in the domain
or additional analysis is required to reduce uncertainty; such uncertainty is catego-
rized here as epistemic. The final category—referred to here as ontic—is related to
irreducible uncertainty induced by inherent variability.
While it is important to recognize each type of uncertainty that may appear
throughout the LCA process (reporting those that may be influential to the outcomes),
those uncertainties that can be explored and better understood using quantitative
analyses will be the focus of this chapter. Such uncertainties fall within the quan-
tity uncertainty category—those associated with data inputs—and have widely used
tools in LCA available for their management.
Data Sources, Precision, and Accuracy
LCA, just like other many other models, uses both quantitative and qualitative data.
Data is fundamentally bound to uncertainty analysis in LCA, captured within the
quantity uncertainty category in Table 6.1. Understanding the data is thus a critical
part of LCA, making the interpretation to decision-makers more robust.
Data can come from a variety of sources, primary or secondary. Primary data is
collected from the actual sources, measured for the study at hand, and represent the
gold standard in LCA data [6]. Due to the extensive nature of LCA methodologies,
it can be challenging to obtain all required data from primary sources. As a result,
data is often compiled (in part or full) from sources that have already been published
or those that are already available. While collecting and compiling data, however,
it is critical to re-assess how certain the inputs are. In the past decade, one major
scientific and policy debate has been about the magnitude of methane escaping across
natural gas production systems [7]. In 2009, a major publication noted that if these
escaping emissions were greater than 3.2% of the produced natural gas, then the
benefits of electricity from gas compared to coal may be negated [8]. The issue was
further inflamed by a 2011 study which used published data to show that this was
already the case [9]; yet, the study used unrepresentative data and did not account
for the efficiency improvements of gas-fired power compared to coal. Since then, a
number of studies suggest that this reversal has not occurred [10–13] and ongoing
research points to unresolved uncertainties. [7, 14] While new measurement studies
are improving the state-of-knowledge, [15] it is important to note that the epistemo-
logical uncertainty in emissions has led to substantial quantity uncertainty related to
LCAs in this field. Many publications relied upon outdated estimates with question-
able representativeness [16]. A similar challenge exists for comparisons of the water
consumed across the life cycle of gas-fired and coal-fired power: several often-used
data sources are known to be outdated, with examples going back as far as 1994
[17, 18].
Such data challenges speak to two critical questions related to LCA data and
results. LCA practitioners should first ask themselves how accurate, or close to the
real values, these numbers are. Second, analysts should consider how precise the
results are, or whether repeated measurements would present the same or different
results. The choice of uncertainty analysis should be informed by considering these
two questions.
Probabilities
Probability—a measure of the degree to which an outcome is certain or uncertain—is

a fundamental concept to consider as uncertainty is investigated in LCA or otherwise.
It can be considered a quantitative statement about the uncertainty associated with
the event under consideration [19]. As such, it is critical to understand what is
known about specific quantities or data that are being used in models to better inform
how to either reduce or evaluate the implications of the uncertainty. One can imagine
a simple example of rolling a six-faced die. Of course, the probability of rolling a
two is one of six. A die landing with the two facing upwards can be considered an
outcome. The act of rolling the die is considered an event. The more we study an
event, the better our probabilistic reasoning about the outcomes; for example, one
might continue rolling the die and recording the outcome to better understand it’s
probability (or use some simple math, considering there are six faces).
Probabilities have traditionally been viewed as objective statements about
outcomes determined mathematically. Importantly, working with probabilities neces-
sitates specific rules. The most important for this chapter are outlined below, but this
short list is not exhaustive [20].
• Probabilities must be equal to or greater than 0 and equal to or less than 1:
probabilities cannot be negative or greater than 1.
• Probabilities must add up: if outcomes are mutually exclusive (if outcome A
happens then B cannot happen), then the probability of one or the other is the
sum.
Probabilities 91
Fig. 6.1 Months in production for active and inactive Barnett shale gas wells as of 11/1/2016 [21]
• Total probability must equal 1: for a set of outcomes (mutually exclusive) that
cover all possibilities, then their sum must be equal to 1.
Histograms provide a very useful first step towards visualizing probabilities, once
a sample or population of data has been obtained. They enable analysts to visualize
the distribution of a numeric variable (e.g. height of students in their freshman year
of college).
Numeric results are grouped into bins and display the frequency of outcome for
each event. The x-axis reflects the range of known values (outcomes) for the data
quantity (event) in question. The y-axis reflects either the frequencies (counts) or
relative frequencies determined for each outcome (percentage of the total).
Let’s begin with an empirical probability distribution, which is built from the
empirical results represented in histograms. An empirical probability distribution
describes observations of outcomes from empirical data (Fig. 6.1).
Recall from Chap. 3 the discussion surrounding uncertainty in facility lifetimes
in gas-fired power. A practical example can be illustrated by a deep dive into under-
standing the uncertainty in a natural gas production well’s lifetime. Given the lack of
multi-decadal operational experience with hydraulically fractured shale gas wells,
the uncertainty inherent to estimates of the lifetimes of shale gas wells is widely
recognized. Generally, it is assumed that a well will produce for at least 10 years, but
estimates range up to 30 years [22]. Regardless, estimates have been published that
are as low as three years [23]. To investigate this matter for horizontally drilled wells
in the Barnett as of 2009, the difference between the date of first production and
last production for all active horizontal wells that were producing as of 11/1/2016
are presented in Table 6.1 compared to those wells that are inactive. An empirical
probability distribution can provide the same information (Table 6.2 and Fig. 6.1).
The active wells included in the figure were still in production at the time of the
last report, so the ultimate lifetime is yet unknown. The wells that were inactive may
Table 6.2 Years in production of Barnett shale gas wells producing as of 11/1/2016 and those
reported as inactive as of the same date [21]
Barnet shale gas wells
Horizontal active wells Horizontal inactive wells
Sample size (n) 11,943 3536
Mean lifetime (years) 7.2 4.9
Minimum lifetime (years) 0.2 0.1
Lifetime (years), 25th percentile 5.3 2.8
Maximum lifetime (years) 24.7 14.4
Standard Deviation (years) 2.5 2.7
be temporarily shut-in and reactivated. The number of months was calculated as the
difference between the first and last months of production, as reported by the Texas
RRC [21].
The empirical distribution is informative in terms of better understanding the
lifetime of wells. What we can determine is that the majority of wells were still
producing at the time of the analysis, so low estimates (e.g. 3 years) reported by
specific studies are not representative for the region. And the results could have had
profound implications. Using a 25-year lifetime with a representative portion of wells
co-located on the same site, we estimated the mean life cycle land-use intensity of
Barnett natural gas used for electricity generation in 2009 was 0.62 m2 /MWh. The
extreme case of a three-year well lifetime with no co-location of wells resulted in an
estimate of 2.2 0.62 m2 /MWh, equivalent to 3.5 times the base case of 25 years.
By exploring the uncertainty in data, the low case was shown to be highly unlikely
and unrepresentative of wells producing in the region (Table 6.1 and Fig. 6.1). Specif-
ically, the 11,943 horizontal shale gas wells in the Barnett that were producing at the
end of 2016 had an average well lifetime of 7.2 years but had not yet reached the end
of their lives. This means that cumulative production will increase continuously until
the end of the producing lifetime of the well (or until temporarily shut-in). The 3536
inactive wells had an average producing life of 4.9 years; however, wells may be
temporarily shut-in or permanently shut-in so it is unclear whether 4.9 is representa-
tive of their producing lifetimes. Without additional evidence, this analysis suggest
that the range of 10–30 years is reasonable, acknowledging that the full well lifetime
for these wells is yet unknown.
The empirical distribution above provides a discrete interpretation from which one
can derive information; however, probabilities can also be represented as continuous
functions. Such representations inform the development of probability distribution,
which can be represented as probability distribution function F(x) were x is the
uncertain parameter under question. Numerous types of probability distributions are
outlined in Fig. 6.2a. To apply a distribution to data implies that the analyst has
Probabilities 93
Probability density
Distribuon Shape funcon (pdf)
Uniform
Median and mean,

since a normal
distribuon
Triangular
Normal
Cumulave distribuon
funcon (cdf)
x
(a) (b)
Fig. 6.2 (a, b). a presents a three examples of simple distributions that represent probability density
functions. b compares a probability density function (top) to a cumulative distribution function
(bottom). The latter is the integral of the former
sufficient information about the distribution; however, it should be understood that

using any distribution is assuming knowledge about the data and system at hand.
For clarity, the probability density function (pdf) and cumulative distribution func-
tion (cdf) should be differentiated as they are two ways that the probability distri-
bution of an uncertain variable [24] A pdf, denoted as f(x) in Fig. 6.2b, describes
the probability that the variable will be equal to x. The total area under the curve is
equal to 1, meaning that all potential outcomes are covered. A cdf, denoted as F(x)
in Fig. 6.2b, describes the probability that the variable under question will be less
than or equal to x. The value of F(x) ranges between 0 and 1 for all x and represents
the integral of the pdf.
Approaching Uncertainty in LCA
There are a number of approaches to understand uncertainty in LCA, ranging from

qualitative to quantitative (Table 6.3). Qualitative can span from a simple discussion
of sources to more complex semi-quantitative methods. Significance heuristics, for
example, can rely on pre-set thresholds. For example, an analyst may choose to
consider comparative results that have less than a 20% difference as inconclusive.
The pedigree matrix approach uses data quality indicators to develop (subjective)
quantitative representations of uncertainty in inventories (Table 6.3).
While the qualitative forms of uncertainty assessment can be useful in some cases,
this section focuses on quantitative uncertainty assessment, starting with a detailed
discussion of the most common uncertainty assessment tools in LCA (Monte Carlo
Simulation, sensitivity analysis, parametric analysis, and scenario analysis). The
methods will be introduced along with a review of how they are employed in LCA.
Table 6.3 Approaches to uncertainty analysis (adapted from Matthews et al. [25])
Type Approach Description
Qualitative Discussion of sources Written description of uncertainties
without quantification
Semi- quantitative Significance heuristics Use of pre-set thresholds that serve as
“rules of thumb” to compare outcomes
Pedigree matrix Subjective assessment of data quality
that informs the development of
parametric uncertainty factors
Quantitative Uncertainty propagation using Use of probability distributions to
probability distributions develop probabilistic results expressed
as pdfs or cdfs
Sensitivity analysis Use of a parametric analysis to
examine how changing the value of a
variable influences result
Uncertainty propagation using Compiling of various datasets and
ranges inputs and evaluating how results range
using known values
The application of these uncertainty assessment techniques to natural gas-fired power

will be examined. Worked examples for each type of method will be provided from
previous research on natural gas, along with a discussion of available software that
can be used to complete such analyses.
Monte Carlo Simulation
Monte Carlo Simulation was first formalized during the Manhattan Project—the U.S.
effort to develop nuclear weapons coinciding with the first electronic computers [26,
27]. Manhattan Project scientists John von Neumann and Stanislaw Ulam applied the
method to many nuclear weapons problems to manage uncertainty and subsequently
informed its development. While first applied to nuclear weapons, the method is now
used broadly across disciplines with many applications, LCA being one example.
Imagine an equation with a number of input variables that are uncertain. Monte Carlo
Simulation will develop a distribution of results by producing the results over many
iterations where the input distributions are randomly sampled.
For the case of LCA, imagine all inputs can be represented with data distribu-
tions. The model calculations would be repeated random data point selections from
input distributions and completing a calculation many times (e.g. 10,000) to obtain
a distribution of what the life cycle results might look like. There are generally two
sampling methods employed [28]. In addition to simple random sampling, the Latin
hypercube approach is stratified and requires fewer simulations for similar levels of
Approaching Uncertainty in LCA 95
precision, ensuring that the upper and lower bounds of the distribution are sampled.
We focus on the former here to demonstrate the approach.
Pdfs are often applied to the data to run the simulation and are preferred, but
empirical distributions may also be used. If the full population of data is known, for
example, empirical distributions will be more accurate as all outcomes have been
represented. In the absence of data, analysts can apply distributions, but they should
be cautioned that such practice should be based in knowledge and expertise. Expert
elicitation can be more formal (using a structured, replicable approach to acquiring
information from a group of experts) or informal (self-assessment, brainstorming
or expert engagement without structure) [28]. If information is known about the
variable—for example, several data points—then it should be used to inform the
modeler’s choice.
Due to the diverse components of the natural gas production system and large
population of gas-fired power plants, Monte Carlo Simulation is a very useful way
to examine both irreducible variability across facility types as well as uncertainty
inherent to several related questions.
In a recent co-authored publication, I employed this method to determine the
life cycle land use intensity of gas-fired electricity generation using infrastructure
within Barnett shale region of Texas. We used empirical pdfs in our analysis as
distributions were informed by the complete or near-complete population of data
for all infrastructure included in the analysis except for production sites and gath-
ering pipelines. [21] For 10,000 simulations, random samples were selected from
the empirical distributions for each input variable to calculate life cycle land-use
intensity.
Using the Monte Carlo simulation (representative of 2009), we were able to deter-
mine that the mean life cycle land intensity can be estimated at 0.62 m2 /MWh
when considering the co-location of wells with 95% confidence intervals of
±0.01 m2 /MWh (Fig. 6.3). Upon comparison with other datasets, our results suggest
that the land-use intensity associated with several renewable energy sources may be
on the same (or even lower) order of magnitude when compared with natural gas-fired
electricity. As shown in Fig. 6.3, midstream infrastructure like pipelines is influential
in the results.
Sensitivity Analysis
A sensitivity analysis is a process of varying inputs to test how analysis is affected

by changes in assumptions. It is a very useful tool when the analyst does not have
a high degree of confidence in the result, when there is high variability in the input
data, and when the data for a particular element are missing or deficient [6].
There are many ways to represent the results of a sensitivity analysis, from
explaining them in text or a table to a more sophisticated tornado diagram (Fig. 6.4).
As noted, one metric that is particularly influential on life cycle results is the life-
time of a facility (or the lifetime of production for natural gas wells). Informed by
96
100
10-1
10-2
10-3
Life cycle land use intensity (m2/MWh)

10-4
Production sites Gathering pipelines Gathering sites Processing sites Transmission pipelines Transmission sites Power plants Total
n = 400 n = 361 n = 30 n = 32 n=1 n = 25 n=6 10,000

repetitions
Fig. 6.3 Land-use intensity for infrastructure elements from the seven-county study area of Barnett Shale, Texas for life cycle (normalized by generated
electricity). The boxes in the figure represent the interquartile range, the line in the box represents the median, the whiskers represent the maximum and
minimum, and the blue diamond represents the mean. On the x axis, n denotes the number of infrastructure elements sampled for each life cycle stage
6 Tackling Uncertainty Across the Life Cycle of Gas-Fired Power
Well lifeme (30-10 yrs)
Processing (50-10 yrs)
Transmission sites (50-10 yrs)
Transmission pipelines (50-10 yrs)
Power Plant (50-10 yrs)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

Life cycle land use intensity (m2/MWh)
Fig. 6.4 Results of a sensitivity analysis presented as a tornado diagram. The sensitivity of the
results for the life cycle land-use intensity to facility lifetime was explored
literature and discussions with experts, we determined the results of the analysis
were particularly sensitive to the production lifetime of wells, resulting in a range of
89–169% of the baseline.
Break-Even (Parametric) Analysis
A break-even analysis is a parametric assessment that is commonly used in finance,

where an analyst will determine the number of units or dollars of revenue are required
for cost recovery. In the case of an environmental breakeven analysis, one evaluates
the value an uncertain parameter must reach in order for one product to be envi-
ronmentally comparable with another. The method is particularly useful for cases in
comparative LCA where results have been inconclusive and data inputs to which the
model is sensitive are uncertain.
First, recent comparative LCAs of coal and natural gas have raised questions
regarding the possibility that fugitive emissions may be high enough to eliminate
climate benefits of electricity generated from natural gas relative to coal [9, 14].
In addition to the high level of uncertainty related to methane leaks in natural gas
production systems, a critical component of this debate is the choice of time horizon
used to characterize the global warming potential of methane relative to carbon
dioxide. A 100-year time horizon is typically used, but 20-year horizons—which
are more important for near-term warming—have often been excluded from LCAs.
Shorter time horizons amplify the short-term impacts of methane relative to carbon
dioxide, arguably better capturing potential tipping points related to climate impacts
on shorter time scales. Focusing on short-term effects, however, will overlook the
longer term impacts of carbon dioxide. Carbon dioxide has atmospheric lifetimes
that range from hundreds to thousands of years [29] while methane resides in the
atmosphere for just over ten [30]. A more comprehensive approach may explore the
question by parametrically varying the time horizon or using breakeven analyses that
examine the amount of methane that must leak from the natural gas supply chain to
negate the climate benefits in the power sector for specific time horizons (Fig. 6.5).
Fig. 6.5 Environmental break-even point where the methane emissions become high enough to
reverse the climate benefits of coal-fired relative to natural gas-fired electricity Skone et al. [31]
Regardless of present estimates of methane leaks across natural gas production

systems, we know that if they become large enough, the environmental impacts of
electricity generated from natural gas breakeven with coal. Using a parametric anal-
ysis, the National Energy Technology Laboratory estimated that breakeven methane
emissions would have to reach 5.8 and 13.3% of natural gas produced for 20-year
and 100-year GWP respectively to breakeven with estimated emissions from the life
cycle emissions from the U.S. coal-fired power fleet. Similarly, emissions would
have to reach 4.4 and 10.0% of natural gas produced for 20-year and 100-year GWP
respectively to breakeven with estimated emissions from the life cycle emissions
from advanced coal-fired power. While estimates of current practice on average were
1.4%, their findings illuminate the importance of regional differences and variability
for the life cycle of gas-fired power as some high emitting regions already exceed
the 20-year GWP breakeven.
Scenario Analysis
LCA practitioners are often faced with an uncertain future. How can LCA account for
different future outcomes that may or may not come to pass? LCA is meant to inform
decision-making, making scenarios a useful way to evaluate how impacts may differ
relative to potential unknown outcomes. Scenarios are not predictions of the future
as it would be impossible to do so. Accepting what is possible, scenarios present
reasonable depictions of possible future outcomes that serve to understand uncertain
outcomes [32–34]. They are used broadly in planning [33] and can be normative
in that they inform what the future should be, from the eyes of the analyst [35]. In
the context of uncertainty assessment, scenarios are an approach that can be used in
policy analysis to manage uncertainty related to factors external to the system that
may have unanticipated effects [1]. Walker et al. [1] suggest that scenarios are best
for exploring assumptions that are not verifiable and that are beyond simple statistical
uncertainty as the mechanisms leading to the outcomes are not well understood. As
a result, while probabilities can be assigned to scenarios, it may be challenging to
present verifiable probabilities.
With a burgeoning LNG industry, potential exports and the development of
markets present a useful case for scenarios. It is unclear, for example, where the
natural gas will arrive, whether it will be used for power generation, and whether
or not one can consider it beneficial in comparison to the alternative. To understand
better the implications of these uncertainties, scenarios provide insight (though not
predictions). First, a review of potential markets in 2018 presented twelve potential
markets and focused on five that were expecting the highest demand growth, the
highest landed gas prices, and the market sizes: China, Taiwan, Japan, South Korea,
and India. These five countries represented the majority of global LNG imports,
amounting to 75% at the time of the study [36] (Fig. 6.6).
As the influence of LNG use in electricity generation depends highly on the
grid mix and counterfactual (i.e., what would have been generated in the absence
Fig. 6.6 Effect of emissions displacement scenarios on life cycle results for countries importing
Canadian natural gas, assuming the end use is electricity (Figure developed by Kasumu et al. [36]).
The y-axis of the figure describes the change in life cycle greenhouse gas emissions (kgCO2 e/MWh)
of LNG import), potential futures were explored in a recent study of LNG export
from Canada. Three emissions displacement scenarios included (1) total electricity
mix cross-section, (2) dispatchable electricity, and (3) marginal electricity. First, the
cross-section represents the electricity generated from all sources in the year of study
(2010). Second, dispatchable electricity refers to nuclear, hydro, coal, oil and natural
gas. Finally, marginal generation was considered the long-run marginal and assumed
to be the most abundant at the time of the study (coal for China, India South Korea,
and Taiwan, and oil Japan).
Results showed that LNG generally resulted in environmental benefits for all
countries with the exception of Japan prior to Fukushima, when nuclear comprised a
larger share of electricity generation. Overall, if the displacement of greenhouse gas
emissions is to be quantified, then it hinges on assumptions about the electricity that is
displaced. It is uncertain whether the electricity displaced in the future will be higher
or lower carbon, making it unclear whether LNG will in fact result in emissions
benefits. Potential outcomes illustrate the usefulness of exploring alternative futures
with scenarios.
Conclusions
The rapid pace of technological and operational change involved in sectors across
the life cycle of gas-fired power combined with their known variabilities and uncer-
tainties make it extremely challenging to produce accurate, representative LCAs of
the full sector and all impacts. A comprehensive LCA and uncertainty analysis of
even the most recent year would relatively quickly become outdated, meaning that
understanding the impacts of this product system presently requires ongoing inves-
tigation. At present, it would be a fool’s errand to try to resolve all questions related
to uncertainty and variability at once with a large dynamic network of infrastructure
that embodies the concept of a moving target. Due to the variable environmental
conditions (e.g., water availability) and facility operations, site-level information
for this sector is highly valuable for understanding environmental outcomes. For
example, only a few impacts and regions studied were presented in this chapter,
but each component of the supply chain is subject to variability across facilities and
environmental conditions. Additionally, the burgeoning LNG market opens relatively
unexplored areas of inquiry, which include uncertainty in the source, destination, and
ultimate use of natural gas. As a result, uncertainty analysis provides a critical tool
that can capture the influence of upstream variability in infrastructure on life cycle
results but also illuminate the implications of uncertain global markets.
References
1. Walker, W.E., Harremoës, P., Rotmans, J., Van Der Sluijs, Jeroen, P., Van Asselt, M.B., Janssen,
P., Krayer von Krauss, Martin, P.: Defining uncertainty: a conceptual basis for uncertainty
management in model-based decision support. Integr. Assess. 4, 5–17 (2003)
References 101
2. EPA Exposure factors handbook 2011 edition. (2011)

3. EPA Uncertainty and Variability (2020)
4. Igos, E., Benetto, E., Meyer, R., Baustert, P., Othoniel, B.: How to treat uncertainties in life
cycle assessment studies? Int. J. Life Cycle Assess. 24, 794–807 (2019)
5. van Zelm, R., Huijbregts, M.A.: Quantifying the trade-off between parameter and model
structure uncertainty in life cycle impact assessment. Environ. Sci. Technol. 47, 9274–9280
(2013)
approaches for decisions that matter. Retrieved June 1 (2018)
343, 733–735 (2014)
8. Alvarez, R.A., Pacala, S.W., Winebrake, J.J., Chameides, W.L., Hamburg, S.P.: Greater focus
needed on methane leakage from natural gas infrastructure. Proc. Natl. Acad. Sci. U. S. A. 109,
6435–6440 (2012)
9. Howarth, R.W., Santoro, R., Ingraffea, A.: Methane and the greenhouse-gas footprint of natural
gas from shale formations. Clim. Change 106, 679–690 (2011)
11. Logan, J., Heath, G.A., Macknick, J., Paranhos, E., Boyd, W., Carlson, K.: Natural gas and
the transformation of the U.S. energy sector: electricity, NREL/TP-6A50–55538. Joint Inst.
Strateg. Energ. Anal. (JISEA) (2012)
Sci. 111, E3167–E3176 (2014)
13. National Renewable Energy Laboratory Life Cycle Assessment Harmonization.: https://www.
nrel.gov/analysis/life-cycle-assessment.html. Accessed 30 May 2018.
15. Allen, D.T., Torres, V.M., Thomas, J., Sullivan, D.W., Harrison, M., Hendler, A., Herndon,
S.C., Kolb, C.E., Fraser, M.P., Hill, A.D.: Measurements of methane emissions at natural gas
production sites in the United States. Proc. Natl. Acad. Sci. 110, 17768–17773 (2013)
16. Lyon, D.R.: Methane emissions from the natural gas supply chain. In: Environmental and
Health Issues in Unconventional Oil and Gas Development, pp. 33–48. Elsevier (2016)
18. Gleick, P.H.: Water and Energy. Annu. Rev. Energy Env. 19, 267–299 (1994)
19. Jordaan, I.: Decisions under uncertainty: probabilistic analysis for engineering decisions.
Cambridge University Press, New York, NY (2005)
20. Feldman, R.M., Valdez-Flores, C.: Applied probability and stochastic processes. Springer
Science & Business Media (2009)
electricity. Nat. Energ. 2(10), 804–812 (2017)
22. Clark, C.E., Han, J., Burnham, A., Dunn, J.B., Wang, M.: Life-cycle analysis of shale gas and
natural gas. Argonne National Lab. (2012)
23. Stephenson, T., Valle, J.E., Riera-Palou, X.: Modeling the relative GHG emissions of
conventional and shale gas production. Environ. Sci. Technol. 45, 10757–10764 (2011)
24. Stowell, S.: Probability distributions. In: Using R for statistics, pp. 87–98. Springer, Berlin
(2014)
approaches for decisions that matter. Retrieved June 1, 2016 (2015)
26. Harrison, R.L.: In Introduction to monte carlo simulation; AIP conference proceedings. Am.
Inst. Phys. 1204, 17–21 (2010)
27. Benov, D.M.: The Manhattan Project, the first electronic computer and the Monte Carlo method.
Monte Carlo Methods Appl. 22, 73–79 (2016)
28. Firestone, M., Fenner-Crisp, P., Barry, T., Bennett, D., Chang, S., Callahan, M., Burke,
A., Michaud, J., Olsen, M., Cirone, P.: Guiding principles for Monte Carlo analysis. US
Environmental Protection Agency, Washington, DC (1997)
29. Archer, D., Brovkin, V.: The millennial atmospheric lifetime of anthropogenic CO2 . Clim.
Change 90(3) 283–297 (2008)
30. Schiermeier, Q.: Global methane levels soar to record high. Nature (2020)
31. Skone, T.J., Marriott, J., Littlefield, J.: Life cycle greenhouse gas emissions: natural gas and
power production (2016)
32. Chermack, T.J., Lynham, S.A.: Definitions and outcome variables of scenario planning. Hum.
Resour. Dev. Rev. 1, 366–383 (2002)
33. Schwartz, P.: The art of the long view: planning for the future in an uncertain world. Doubleday,
New York (1996)
34. Schwartz P., Ogilvy J.: Plotting your Scenarios: An introduction to the art and process of
scenario planning. Global Business Network, Emeryville (2004)
35. Andreescu, L., Gheorghiu, R., Zulean, M., Curaj, A.: Understanding normative foresight
outcomes: Scenario development and the ‘veil of ignorance’effect. Technol. Forecast. Soc.
Chang. 80, 711–722 (2013)
Chapter 7
Natural Gas as a Bridge Fuel?
Abstract The question of whether natural gas can act as a bridge fuel has yet to be
resolved. Proponents focus specifically on the role that gas-fired power can play in
displacing coal while opponents point to methane leaks across the supply chain and
longer time horizons when renewable alternatives become more competitive. After
the past decade in the United States, there have been conditions where natural gas
acted as a bridge fuel by displacing coal when renewable power was less competitive.
In this chapter, the bridge fuel question is revisited by determining the pre-conditions
required for natural gas to serve as a bridge to a more sustainable, low carbon future.
Moving forward, the role will become more challenging, requiring industries across
the life cycle to participate in efforts and demonstrate deep emissions cuts.
Introduction
Natural gas has often been touted as a bridge fuel, offering an opportunity for a path
from a more coal-dominated future to a low carbon future [1]. At the same time,
critics see a “crack in the natural gas bridge” with abundant natural gas replacing all
other options, cheap energy resulting in growth in the sector, and increasing methane
emissions across the supply chain [2, 3]. The central question addressed here is: can
natural gas serve as a bridge fuel to a sustainable and low carbon future?
There is validity to both sides of the argument. On the one hand, the reversal
of climate benefits of switching from coal to gas-fired electricity generation is
commonly understood to be unlikely due to the higher efficiencies of gas-fired power
[4] and the range of accepted estimates for upstream methane leaks [1, 5, 6]. Evidence
exists that there may be cases when reversal is possible on shorter time scales, [7] but
the climate benefits over timescales and regions of using natural gas instead of coal
have been demonstrated [8]. On the other hand, an abundant supply of natural gas in
the absence of any mitigation policies may not be a climate benefit on the longer term
as lower carbon alternatives become less costly. Renewable alternatives are becoming
increasingly competitive (although not yet ubiquitously), with the combination
of solar and storage holding particular promise for addressing intermittency [9].
A recent study even showed that natural gas offers no significant reduction between

104 7 Natural Gas as a Bridge Fuel?
2020–2050 in a carbon constrained future [10]. Further, if ageing nuclear power

plants retire and are replaced by natural gas, there is risk of increasing greenhouse
gases in the electric sector [11]. Finally, if upstream methane emissions from natural
gas productions go unchecked, there are cases where the environmental benefits can
indeed be reversed. That said, the evidence shows this to be unlikely.
The conclusion that natural gas results in environmental benefits hinges on three
pre-conditions. First, to be true, natural gas would always have to replace either coal-
or oil-fired generation (without carbon capture and storage) with analyses placed in
a global and short-term context. The second pre-condition is that the only environ-
mental impact that matters is greenhouse gas emissions, so no other impact cate-
gory would play a role in decision-making. Last, the natural gas in question would
have to always be used in the electric sector. As a result, these three pre-conditions
are required for the environmental benefits to be realized: natural gas-fired power
typically results in net global environmental benefits specific to greenhouse gas
reductions in the electric sector when provably replacing higher-carbon electricity.
As a result, we are left with four outstanding questions that must be explored to
determine whether or not natural gas can serve as a bridge fuel. First, what are the
climate implications with regard to long-term choices considering all alternatives
(from lower carbon to higher carbon)? The answer is determined by exploring the
compatibility of the evolving sector with the Paris Agreement. Second, could there
be unpredicted events or uncertainties which may disrupt the growth trajectory of
gas-fired power? Certainly, there could be. The field thus presents an ongoing line
of inquiry for practitioners to examine results vis-à-vis political, economic, and
technological uncertainties. Third, given the high variability and uncertainty inherent
to this product system, are there any potential trade-offs in other impact categories?
This question requires ongoing exploration that encompasses both breadth and depth
of inquiry, to ensure that environmental impacts are not being overlooked. Finally,
natural gas has certainly played a role as a bridge fuel in the past decade in the United
States. At what point in time does natural gas cease to act as a bridge fuel?
Compatibility with Paris Agreement
For gas-fired power to be compatible with the Paris Agreement on a life cycle basis,
both natural gas transported solely by pipeline and also as LNG by tanker warrant
attention. It has been shown and is generally accepted that in the United States,
natural gas can contribute to Paris goals [12]. Due to the upstream emissions and
the fact that there is a limit to the total amount of coal that can be replaced, long-
term, deep decarbonization will likely require ongoing policy efforts in the electric
sector and beyond. Policy efforts that support innovation of low carbon electricity
generation—from research and development through demonstration, deployment,
and diffusion—will be most compatible with climate goals. The International Energy
Agency (IEA), for example, prepares a sustainable development scenario (SDS) each
year that is compatible with Paris goals but does not eliminate natural gas [13]. This
Compatibility with Paris Agreement 105
compatibility, however, requires what the IEA refers to as “immediate and major
reductions in methane emissions” across natural gas production systems [14]. To
exemplify the urgency, emissions reductions in natural gas flaring required to realize
the SDS are already falling behind schedule (meaning they are not on track to meet
targets), [15] underscoring the need to address fuel supply emissions across segments.
Despite recognized potential environmental benefits of its use in the electric sector,
[16] LNG may prove to be incompatible with Paris goals. The same questions about
supply chain emissions apply to LNG as to natural gas: for compatibility, these
emissions must be reduced. Recall that one pre-condition is LNG-fired electricity
replacement of higher carbon electricity without carbon capture, utilization, and
storage (CCUS). Considering limits to supply and demand specifically for LNG that
can displace coal, it has already been shown that on the long-term electricity fueled
by LNG is not compatible with Paris goals [17]. Specifically, prior to 2030, the
potential is LNG-limited in that the supply of LNG will not be sufficient to replace
coal. Over longer time horizons, the rapid decline in coal limits the opportunities
for its replacement with LNG. As a result, the environmental benefits of further
expansion come into question post 2030; though they are (of course) dependent
on the evolution of the sector and the electricity generation mix. Whether or not
environmental benefits are realized depends on where the natural gas ends up and
how it is consumed. In turn, how and where the natural gas is used depends on the
evolution of markets. What countries will have the highest demand? At this point,
the increasing resistance to U.S. LNG in Europe implies the greatest demand growth
is likely to continue from the Asia–Pacific.
While gas and LNG-fired electricity has been shown to be incompatible with
climate goals if unmitigated, solutions often exist where impacts can be reduced. For
example, affordable carbon capture and storage would be a game changer in terms
of the implications of emissions growth related to gas-fired power. The SDS already
assumes that substantial CCUS reduces emissions from gas-fired power, amounting
to approximately 27 billion metric tons of carbon dioxide (GtCO2 ) cumulatively by
2070 and over 1 GtCO2 annually in 2070. Another disruptive technology that might
revolutionize the gas sector (the most recent being the technologies that resulted
in the shale boom) could economically convert natural gas into hydrogen with non-
combustion storage or utilization of the leftover carbon. Hydrogen can generate elec-
tricity without carbon dioxide emissions upon combustion. To ensure that there are
no unanticipated trade-offs across the supply chain (e.g., electricity use for the sepa-
ration of hydrogen from water using hydrolysis), life cycle assessment can inform
decisions with objective analyses.
Political, Economic, and Technological Uncertainties
Further confounding those statistical and scientific uncertainties that can be better
understood with inquiry, there are some uncertainties that remain beyond the grasp of
full characterization. With a growing global market, new political tensions and policy
directives become more challenging to predict. Across the life cycle, companies are
involved with fast-paced technological innovation that could substantially disrupt the
present status quo, from satellite measurement of methane emissions to CCUS. The
COVID-19 pandemic paints another picture of unanticipated circumstances. Such
uncertainties point to the importance of going beyond more conventional quantitative
analyses to the development of scenarios to characterize potential futures, as explored
in Chap. 6.
The natural gas sector has substantially expanded with new markets in electricity
with coal-to-gas shifts and climate concerns. These markets were driven by both
markets and technological innovation that enabled low costs. But in the past decade,
the costs of renewable options have become increasingly more competitive, signi-
fying that the longer-term viability of natural gas will depend on more challenging
economic and environmental benchmarks. Already, wind and solar can contend with
natural gas-fired generation in specific locations and it is anticipated that plummeting
costs will continue [18, 19]. As the competitive landscape evolves, how different
jurisdictions manage their share of supply chain emissions becomes more important.
Already, jurisdictions are requiring lower supply chain emissions. In 2020, the Euro-
pean Commission adopted a methane strategy as part of the European Green New
Deal, which places explicit focus on the need to reduce emissions from the oil and
gas sector [20]. Such initiatives point to a future where suppliers will be required
to reduce emissions in order to meet demand in specific markets. Pressure is also
being applied to reduce emissions produced by generators, in addition to ongoing
initiatives to reduce grid inefficiencies that result in power losses between the gener-
ator and consumer. In 2020, the World Bank reported that there were 61 carbon
pricing schemes in operation or scheduled for implementation, covering 22% of
emissions (up from 20% in 2019, despite the effects of the pandemic). Countries are
commencing to include the reduction of grid inefficiencies that occur during power
transmission and distribution in the Nationally Determined Contributions that they
submit as parties to the Paris Agreement, at times building upon quite substantial
grid modernization programs [21].
Disruptive technological change can substantially divert reality from our
embedded understanding of both economic and environmental impacts. The shale
boom and resulting coal-to-gas shift in the U.S. electric sector is an obvious example
of how a disruptive technology—the combination of hydraulic fracturing with hori-
zontal drilling—can radically decrease and stabilize costs of a generation source.
Higher efficiency power generation that uses combined cycle natural gas turbines
has improved both economic and environmental performance, with a surprising
change towards increased natural gas as baseload power (Chaps. 3 and 5). Beyond
presently deployed technologies, operators across the supply chain and jurisdictions
can play a role in reducing life cycle environmental impacts of gas-fired power. For
example, oil and gas companies can seek to implement cost-effective technologies
that either reduce or eliminate leaks. What makes the question particularly compli-
cated and interesting is that these operators face a highly variable and evolving
Political, Economic, and Technological Uncertainties 107
regulatory environment related to the deployment of new technologies and reduc-

tion of environmental impacts more broadly. Governments can require and imple-
ment more robust mitigation and measurement programs, to which companies may
choose to respond with leadership, simple participation, or non-compliance. Future
research should explore how technologies may disrupt the present systems them-
selves, looking beyond technologies that mitigate the environmental impacts of the
current product system. Plummeting costs of renewable power presents one example.
Comparisons of dominant natural gas production systems should be considered in
light of new and emerging technologies that may serve as a replacement. The Inter-
national Energy Agency, for example, points to the importance of not only reducing
emissions from core natural gas operations but also increasing private investments
in hydrogen, biomethane, and advanced biofuels [13].
All bets have been off since the beginning of the COVID-19 pandemic, which
illustrated precisely how unpredictable exogenous events can upend any reasonable
forecast of the energy sector. As economies recover from the pandemic, the percep-
tion of future outcomes regarding growth in markets has substantially evolved. The
vulnerability of an emerging LNG industry has been revealed through these chal-
lenges, which have been unprecedented in the past century. Demand in both China
and India is still anticipated to grow in the future after the present downturn, but
fierce competition will exist between supplying nations as growth resumes. Large
suppliers and incumbents such as Qatar, Australia, and the United States have a
strong foothold, presenting a challenging environment for smaller incumbents such
as Canada. With vaccines increasingly available, even a possible COVID-resistant
future should not dull the memory of how an unanticipated event can propel the most
reasonable predictions into oblivion (reinforcing how scenarios can provide guidance
in the face of an uncertain future). Beyond more conventionally understood uncer-
tainties and variabilities, exogenous events can profoundly impact the evolution of
the energy sector, particularly with a burgeoning industry like LNG.
Future Research
The expansive network of infrastructure associated with natural gas production

systems results in irreducible uncertainty (especially in the absence of sector-wide
monitoring with adequately reported impacts) and often uncharacterized variability
(due to lack of adequate tracking and reporting for all infrastructure and facilities).
This leaves the practitioner with a unique challenge: should the analysis prioritize
breadth or depth when there are so many highly uncertain moving parts? I would
argue that there is a need for both. One the one hand, International Organization
for Standardization (ISO) recommends examining a comprehensive suite of impact
categories in life cycle assessment (LCA), reducing the potential for environmental
trade-offs in different categories being overlooked (purposefully or not). Yet, the anal-
ysis of gas-fired power illuminates the challenges with taking only a broad approach
(Chap. 4).
If a broad-sweeping approach to understanding all impact categories is taken, very

specific and important issues may be overlooked. One prime example is the methane
emissions from natural gas supply chains. Careful attention must be paid to improving
the data that is utilized in LCA to represent these emissions, which have been increas-
ingly studied. Focused studies on greenhouse gas emissions enable a greater under-
standing for how to investigate these uncertain systems within LCA. With ongoing
improvements and change in operational practice and technologies, it is useful to
examine specific changes in detail.
To fully understand the implications of natural gas as a bridge fuel to a sustainable,
low carbon future, one must fully evaluate all relevant impact categories. Moving
beyond greenhouse gas emissions alone, there may be unanticipated trade-offs in
other environmental impact categories relative to coal and other generation types.
In Chap. 4, a large number of broad impact categories was explored, followed by
detailed examinations of specific impact categories to underscore the importance of
both breadth and depth of analyses for this topic. For example, an investigation of the
water implications of the coal-to-gas transition in the electric sector demonstrated that
the upstream natural gas production can cause regionalized water demand and acute
impacts in specific locations [22, 23]. At the same time, evaluating a comprehensive
suite of impact categories inarguably can support the identification of trade-offs. Gas-
fired power may produce fewer life cycle greenhouse gas emissions than coal, but the
outcomes discussed in Chap. 4 showed that natural gas-fired power has been ranked
higher than coal for particulate matter equivalents when compared on the basis of a
Megawatt-hour of electricity generated.
It is not only the known impact categories for natural gas that are uncertain, but
the less vetted impact categories can be highly uncertain and influential to LCA
results. It took a full analysis spanning several years to determine the life cycle
surface land requirements of gas-fired power in only one small region of Texas [24].
The results suggest that when the fuel cycle is considered, the full life cycle land
requirements of natural gas is unlikely to offer the perceived benefits over renewable
alternatives over the lifetime of power plants.
A practitioner may opt for a very specific, detailed LCA of gas-fired power that
focuses on one impact category specific to one region; this strategy can clearly
be as informative as a broad analysis of the product system that considers a more
comprehensive suite of impact categories. Due to the constant evolution of the
involved sectors and emerging markets, both approaches hold value to the advance-
ment of societal knowledge about the environmental implications of electricity
generated from natural gas.
Conclusions
Recent natural gas investments and forecasts of growth are indicative of a sector
that is likely to increase regardless of compatibility with the Paris agreement and the
COVID-19 pandemic. This prospect raises the question that we seek to answer: can
Conclusions 109
natural gas serve as a bridge fuel during a transition to a low carbon energy system?
I would argue that it depends. Perhaps the more salient question, since it is possible
under certain technological and economic conditions for natural gas to be a bridge
fuel, is: will it? Provided companies, governments, and international agencies step
up to the plate to both understand emissions and reduce them, natural gas—as the
cleanest burning fossil fuel—can become cleaner. For any benefits to be realized with
present technologies and fleets, gas-fired must replace coal concurrent to making cuts
to emissions across the life cycle. Such emission cuts should encompass mitigation
options from technologies that cut and reduce fugitive emissions to increasing power
plant efficiency, deploying carbon capture and storage technologies, and eliminating
unnecessary grid inefficiencies. Technological deployment at such a scale requires
concerted efforts. To reach goals met in the IEA-SDS, for example, 35 GW of gas-
fired power with CCUS would have to be deployed with no large-scale project in
operation to date. It is not impossible to get on track to meet this goal; already, of
the sixteen CCUS projects in development, five relate to electricity generated from
natural gas [25].
Policymakers, companies, and other stakeholders across the world are taking
initiative in support of reducing life cycle impacts of gas-fired power. In addition
to the EU Green New Deal, other major developed producers, are implementing
regulations to better understand and reduce emissions [12, 20, 26, 27]. Companies
are cooperating on efforts to reduce emissions with initiatives such as the Oil and
Gas Climate Initiative [14]. The number and coverage of carbon pricing systems are
growing globally, in addition to increasing support for carbon capture and storage
and grid modernization. Going beyond the power sector’s emissions, recent frame-
works have been proposed to leverage the Paris Agreement to ensure that methane
emissions are better addressed as ambitions are increased [28]. While the complexity
of the evolution of the sector is apparent in a patchwork of emissions performance,
supply chain stages, and regulatory environments, what is clear is that a more rigid
environmental expectation is emerging.
It is important to note that this book examines only electricity as an end use.
What about other markets that have received less attention or emerging markets that
may play a role? Understanding the potential for natural gas to serve as bridge fuel
requires an exploration of the climate implications, the potential for unanticipated
environmental impacts, potential disruptions, and, finally, acknowledging the impacts
beyond the electric sector. It is widely recognized that the electric sector cannot be
the only sector to transition to low carbon fuels [13]. While the end use examined
here was electricity, results for the fuel supply are applicable for other end uses (e.g.,
industrial use or home heating); however, they must be related to the appropriate
functional unit. Life cycle assessment serves as a valuable tool to provide insights
into both unintended environmental consequences and impactful solutions.
Questions related to the life cycle of gas-fired electricity provide a complex and
evolving area of inquiry for LCA practitioners due to the fast-paced, multi-faceted
industry. The emergence of new markets can have profound impacts on environ-
mental outcomes; so can mitigation options that are specific to individual supply
chain segments like power transmission and distribution. While the complexity of
the expansive network with many different infrastructure components adds inter-
esting questions of variability, the industries will continue to evolve. Natural gas
production, power generation, and energy transportation networks will continue to
improve operational practices and technologies, exemplified by the evolving shale
gas sector in the United States and beyond. In conclusion, natural gas ceases to act
as a bridge fuel for the electric sector when the pre-conditions are no longer met.
Deeper emissions cuts that support decarbonization goals necessitate looking beyond
the electric sector alone, placing onus on all industries involved with the life cycle of
gas-fired power to prove the viability of natural gas as a bridge fuel moving forward.
References
1. Levi, M.: Climate consequences of natural gas as a bridge fuel. Clim. Change 118, 609–623
(2013)
2. McJeon, H., Edmonds, J., Bauer, N., Clarke, L., Fisher, B., Flannery, B.P., Hilaire, J., Krey, V.,
Marangoni, G., Mi, R.: Limited impact on decadal-scale climate change from increased use of
natural gas. Nature 514, 482–485 (2014)
3. Davis, S.J., Shearer, C.: Climate change: a crack in the natural-gas bridge. Nature 514, 436–437
(2014)
4. Zhang, X., Myhrvold, N.P., Hausfather, Z., Caldeira, K.: Climate benefits of natural gas as a
bridge fuel and potential delay of near-zero energy systems. Appl. Energ. 167, 317–322 (2016)
Sci. U. S. A. 111, 3167 (2014)
343, 733–735 (2014)
7. Zavala-Araiza, D., Lyon, D.R., Alvarez, R.A., Davis, K.J., Harriss, R., Herndon, S.C., Karion,
A., Kort, E.A., Lamb, B.K., Lan, X., Marchese, A.J., Pacala, S.W., Robinson, A.L., Shepson,
P.B., Sweeney, C., Talbot, R., Townsend-Small, A., Yacovitch, T.I., Zimmerle, D.J., Hamburg,
S.P.: Reconciling divergent estimates of oil and gas methane emissions. Proc. Natl. Acad. Sci.
U. S. A. 112, 15597–15602 (2015)
8. Tanaka, K., Cavalett, O., Collins, W.J., Cherubini, F.: Asserting the climate benefits of the
coal-to-gas shift across temporal and spatial scales. Nat. Clim. Chang. 9, 389–396 (2019)
9. Frew, B., Cole, W., Denholm, P., Frazier, A.W., Vincent, N., Margolis, R.: Sunny with a chance
of curtailment: Operating the US grid with very high levels of solar photovoltaics. iScience 21,
436–447 (2019)
10. Woollacott, J.: A bridge too far? The role of natural gas electricity generation in US climate
policy. Energy Policy 147, 111867 (2020)
11. Jordaan, S.M.: Resilience for power systems amid a changing climate. Bull. At. Sci. 1–7 (2018)
12. Peters, J.C.: Natural gas and spillover from the US clean power plan into the paris agreement.
Energy Policy 106, 41–47 (2017)
13. IEA The Oil and Gas Industry in Energy Transitions.: International Energy Agency (IEA)
(2020)
14. IEA Methane Tracker 2020.: International Energy Agency (IEA) (2020)
15. IEA Flaring Emissions: (2020)
16. Kasumu, A.S., Li, V., Coleman, J.W., Liendo, J., Jordaan, S.M.: Country-level Life cycle assess-
References 111
17. Yang, S., Ravikumar, A.P., Hastings-Simon, S.: Global liquefied natural gas industry expansion
may imperil paris agreement temperature targets (2020)
18. International Renewable Energy Agency, (IRENA) Renewable Capacity Statistics 2019:
International Renewable Energy Agency (2018)
19. Taylor, M., Ralon, P., Ilas, A.: The power to change: Solar and wind cost reduction potential
to 2025. In: The International Renewable Energy Agency (2016)
20. European Commission Communication from the Commission to the European Parliament, the
Council, the Economic and Social Committee and the Committee of the Regions on an EU
strategy to reduce methane emissions 1–21 (2020)
23. Patterson, L.A., Jordaan, S.M., Diaz Anadon, L.A.: Spatiotemporal exploration of water
consumption changes resulting from the coal-to-gas transition in Pennsylvania. In: Belfer
Center for Science and International Affairs, Harvard Kennedy School and Nicholas Institute
for Environmental Policy Solutions, Duke University (2016)
25. IEA Natural Gas-Fired Power.: International Energy Agency (IEA) (2020)
26. O’Connell, E., Risk, D., Atherton, E., Bourlon, E., Fougère, C., Baillie, J., Lowry, D., Johnson,
J., Helmig, D., Lamb, B.: Methane emissions from contrasting production regions within
Alberta, Canada: Implications under incoming federal methane regulations. Elementa: Sci.
Anthropocene 7 (2019)
27. Konschnik, K., Jordaan, S.M.: Reducing fugitive methane emissions from the North American
oil and gas sector: a proposed science-policy framework. Clim. Policy 1–19 (2018)
28. Pekkarinen, V.: Going beyond CO2 : Strengthening action on global methane emissions under
the UN climate regime. Rev. Eur.Comp. Int. Environ. Law (2020)
Index
A – approaches, 20, 22, 28

Acidification – definition, 20
– characterization/quantification, 25 Critical review, 9, 45
impact indicator and characterization
factors, 25, 49, 54
D
Data
B secondary and primary, 20, 21, 52, 89
Break even analysis, 27 Data availability, 14, 20, 38
Data collection, 20, 32, 41, 74
Displacing coal, 103
C Dominance analysis, 27
Capital charge rate, 69
Capital costs, 69
Capital expenditures (CAPEX), 69 E
Carbon Capture and Storage (CCS), 3, 24, Ecosystem services, 58–61
54, 104, 105, 109 Ecosystem services values, 58–60
Carbon Capture, Utilization, and Storage Efficiency, 4, 9, 23, 24, 31, 32, 37, 39, 41,
(CCUS), 105, 106, 109 42, 54, 67, 70, 77, 80, 81, 85, 86, 90,
Carbon dioxide, 15, 24, 25, 52, 77, 97, 105 103, 106, 109
Carbon dioxide equivalent, 15, 19, 24, 46, Electricity delivered, 4, 31
47, 52 Electricity generated, 1–4, 6, 9, 15, 17, 19,
Carbon footprint, 16 20, 23, 31, 32, 39, 45, 50, 60, 69, 75,
Climate change 86, 97, 98, 100, 109
– carbon footprint, 16 Electricity mix, 100
– characterization, 24 Endpoint impact, 47
– greenhouse effect, 24, 47 Estimated Ultimate Recovery (EUR), 9, 33–
– greenhouse effect, 24 37
– impact indicator and characterization Eutrophication, 49
factors, 24
Coal, 1–4, 46, 50, 52–54, 56, 67, 89, 90, 97,
98, 100, 103–105, 108, 109 F
Comparative analysis, 8, 57 Facility lifetime, 33, 61, 91, 97
Compressor stations, 8, 35, 38 Facility throughput, 31–33, 37
Contribution analysis, 27 Freshwater consumption, 55
Co-products, 7 Fuel cycle, 1, 4, 5, 7, 8, 10, 14, 23–25, 28,
Co-products and allocation, 20, 22 31, 41, 42, 53, 58, 59, 85, 108
https://doi.org/10.1007/978-3-030-71971-5
114 Index
Functional unit L
– reference flow, 88 Land use
– area, 37, 58
– characterization using ecosystem
G services, 60, 61
Gathering pipelines, 5, 33, 35, 95 Levelized cost of electricity, 68–72, 83
Gathering sites, 33 Life Cycle Assessment (LCA)
Geographical system boundary, 87 – definition and limitations, 2, 4, 17
Global Warming Potential (GWP), 2, 24, 25, – functional unit, 4, 5, 9, 15, 19, 31, 32,
46, 47, 52, 97, 98 88
Goal and scope definition, 2, 5, 8, 13, 15, 17, – history
20, 24, 26, 27, 39, 41, 51, 61, 62 – early applications, 86
Greenhouse gases, 2, 16, 24, 27, 45–47, 51, – product system, 4, 6, 9, 13, 15, 17, 19,
83, 104 20, 28, 31, 37, 45, 51, 54, 55, 62, 85,
Grid, 3, 4, 83, 99, 106, 109 86, 100, 108
Grid inefficiencies, 106, 109 – unit process, 17, 20, 22, 23, 32, 39, 47,
87
Life Cycle Impact Assessment (LCIA)
H – characterization factors, 88
Heating value, 7 – impact categories, 24, 26, 45, 46, 50,
Higher Heating Value (HHV), 24, 39, 81 57
– impact indicators, 26
Life cycle interpretation, 5, 13, 15, 26, 86
I Life Cycle Inventory (LCI)
Impact assessment, 2, 5, 8, 9, 13, 15, 16, 19, – allocation, 20
23, 24, 26, 27, 46–51 – data collection, 20
Impact categories – stages, 20, 46
– endpoint, 47, 50
Liquefied Natural Gas (LNG), 3, 7, 9, 20, 21,
– greenhouse gas emissions, 9, 68, 104,
52, 67, 68, 73–75, 77, 78, 80–83, 99,
108
100, 104, 105, 107
– impact indicators, 46
Lower Heating Value (LHV), 39
– indicator model and characterization
factors, 24
– land use, 9, 25
– midpoint, 50 M
– NOx, 25, 54 Midpoint impacts, 46, 50
– SOx, 25 Monte Carlo Simulation, 93–95
– water consumption, 9, 45, 51, 54, 108
– water use, 25
International Energy Agency (IEA), 3, 71, N
73, 74, 104, 105, 107 Natural gas
– sustainable development scenario – composition, 7
(SDS), 3, 104, 109 – energy content, 2
Interpretation, 2, 5, 8, 13, 15, 26, 27, 47, 51, – prices, 41, 67, 68, 70, 73–76
85, 86, 89, 92 Natural gas infrastructure
ISO standard – gathering pipelines, 5, 33, 35, 95
– LCIA mandatory components – gathering sites, 33
– – characterisation, 24 – gathering stations, 5
– – classification, 24 – processing, 5, 8, 33, 38, 52
– – impact categories selection, 24 – production, 4, 5, 17, 31, 33, 35, 37, 41,
– LCIA optional elements 50, 52, 54, 95, 107
– – grouping, 26 – storage, 5, 8, 38, 52
– – normalization, 26 – transmission pipelines, 5, 32, 33, 38
– – weighting, 26 – transmission sites, 33, 38
Index 115
– transmission stations, 5 T
Nuclear power, 104 T&D losses, 77, 80, 81
Technical system boundary
co-products, 20
O Temporal system boundary
Operational efficiency, 86 and time horizon, 51
Transmission and distribution (T&D)
(power sector), 4, 21, 31, 77, 106,
P 109
Photovoltaic energy, 59
Pipelines, 4, 8, 59, 73, 75, 95, 104
Potential environmental impact, 24 U
Power plants, 2, 4, 8, 9, 15, 17, 20, 23–25, Uncertainty
28, 31–33, 38, 39, 41, 42, 46, 67–69, – breakeven analysis, 26, 27
75, 77, 79–81, 86, 95, 109 – in LCA, 9, 10, 21, 28, 85–87, 89, 90,
Prices, 22, 67–75, 81 93, 100
Processing, 7, 24, 25, 38 – Monte Carlo Simulation, 93–95
Production sites (natural gas), 4, 17 – scenario analysis, 99
Product system, 4, 9, 13, 45, 54, 55 – sensitivity analysis, 10, 27, 86
– types, 88, 89, 95
Uncertainty analysis, 26, 32, 34, 42, 52, 71,
R
86, 89, 90, 94, 99, 100
Resource and Environmental Profile Anal-
ysis (REPA), 13
Resource demand
– energy resources, 81 V
– land use, 57 Valuation (weighting), 26, 88
S W
Scenario, 3, 93, 98–100, 104, 106, 107 Waste disposal, 1, 50, 54
Sensitivity analysis, 10, 35, 93–95, 97 Water consumption, 1, 9, 16, 19, 45, 46, 51,
Site-level, 32, 35, 38, 45, 52, 55, 62, 75, 83, 54–57
100 Watershed, 15, 55–57, 62
System boundaries, 4, 5, 14–16, 20, 23, 26, Water use, 25, 55
87 Wind power, 8, 41, 59, 60

Wells To Wire: Sarah Marie Jordaan

Uploaded by

Copyright:

Available Formats

Wells To Wire: Sarah Marie Jordaan

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wells To Wire: Sarah Marie Jordaan

Uploaded by

Copyright:

Available Formats

Sarah Marie Jordaan

ISBN 978-3-030-71970-8 ISBN 978-3-030-71971-5 (eBook)

© Springer Nature Switzerland AG 2021

1 An Introduction to Life Cycle Assessment of Natural Gas-Fired

4 Life Cycle Impact Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

Abstract The electric sector is undergoing a transition to a more environmentally

© Springer Nature Switzerland AG 2021 1

Global Evolution of Natural Gas Production and Use

Life Cycle Assessment as a Framework to Understand

Materials aquisition Products

Recycling, if applicable Solid waste

1. Goal and scope

What Will Be Covered in This Book

assessment in LCA, followed by a review of methods frequently employed by LCA

13. IEA Gas 2020. International Energy Agency (IEA) (2020)

Abstract Life cycle assessment (LCA) offers a robust, internationally standardized

© Springer Nature Switzerland AG 2021 13

1. Goal and scope

Fig. 2.1 LCA methodology (developed from ISO 2006)

Before a more comprehensive look at LCA methodology, it is important to distin-

Goal and Scope Definition

Before commencing to undertake a LCA of gas-fired electricity, or any other for

Other materials extraction, transportation of materials, construction, etc.

Waste management, decommissioning, end-of-life, etc.

(2) Determine the functions and the functional unit

Input ﬂows, e.g.: Unit process: natural Output ﬂows, e.g.:

Input ﬂows, e.g.: Output ﬂows, e.g.:

Input ﬂows, e.g.: Unit process: Output ﬂows, e.g.:

Input ﬂows, e.g.: Output ﬂows, e.g.:

Input ﬂows, e.g.: Unit process: power Output ﬂows, e.g.:

Input ﬂows Unit process Output ﬂows Emissions and waste

CO2 7.18 16.41 3.93 27.57 350 kg/MWh

CO2 7.18 16.41 3.93 350 377.57 kgCO2 e/MWh

CO2 7.18 16.41 3.93 350 377.57 kgCO2 e/MWh

The selection of the denominator—the functional unit—in life cycle assessment

seven-county study area used as the estimate of throughput

Natural Gas Extraction

completed production but may be temporarily shut-in or recompleted, so they may

Other infrastructure associated with domestic consumption of natural gas include

Power Plant Operation

Fuel Heat recovery steam

Gas Compression Turbine

The technology employed in gas-fired generation becomes a critical factor in

Environmental impacts can be defined as either positive or negative impacts to the

© Springer Nature Switzerland AG 2021 45

Impact Categories and Characterization Models

Climate change Climate change

Ozone depletion Ozone depletion

Human toxicity, non-cancer Human toxicity, non-cancer

PM/respiratory inorganics PM/respiratory inorganics

Ionising radiation (human health) Ionising radiation (human health)

Ionising radiation (ecosystems) Ionising radiation (ecosystems)

Eutrophication, aquatic Eutrophication, aquatic

Eutrophication, terrestrial Eutrophication, terrestrial

Ecotoxicity, freshwater Ecotoxicity, freshwater

Land use Land use

Fossil and mineral depletion Fossil and mineral depletion

Renewable resources Renewable resources

A Focus on Improving Individual Impact Categories