You are on page 1of 88

A Systems

Approach to
Modeling the
Water-Energy-
Land-Food Nexus
A Systems
Approach to
Modeling the
Water-Energy-
Land-Food Nexus
System Dynamics Modeling and
Dynamic Scenario Planning

Volume II

Bernard Amadei

MOMENTUM PRESS, LLC, NEW YORK


A Systems Approach to Modeling the Water-Energy-Land-Food Nexus:
System Dynamics Modeling and Dynamic Scenario Planning, Volume II

Copyright © Momentum Press®, LLC, 2019.

All rights reserved. No part of this publication may be reproduced, stored


in a retrieval system, or transmitted in any form or by any means—­
electronic, mechanical, photocopy, recording, or any other—except for
brief quotations, not to exceed 400 words, without the prior permission
of the publisher.

First published by Momentum Press®, LLC


222 East 46th Street, New York, NY 10017
www.momentumpress.net

ISBN-13: 978-1-94708-354-7 (print)


ISBN-13: 978-1-94708-355-4 (e-book)

Momentum Press Sustainable Structural Systems Collection

Collection ISSN: 2376-5119 (print)


Collection ISSN: 2376-5127 (electronic)

Cover and interior design by Exeter Premedia Services Private Ltd.,


Chennai, India

10 9 8 7 6 5 4 3 2 1

Printed in the United States of America


Abstract

Water, energy, land, and food (WELF) resources are critical components
in the overall discourse on sustainable human development. Over the next
50 years, rapid population, urbanization, and economic growth world-
wide will create unprecedented demands for such resources, as well as for
health, transportation, waste disposal, communication, and other services.
The discussion on how to meet human needs for water, energy, land,
and food and how to guarantee their respective securities has changed
over time. The traditional way has been to look at all four sectors in iso-
lation, regardless of whether one is interested in supply and demand,
infrastructure planning and design, resource management and allocation,
and/or governance. Since 2011, however, there has been a new empha-
sis on understanding the interdependency of the four sectors through the
so-called WELF nexus.
The approach presented in this book responds to the overall agree-
ment in the WELF nexus literature that the management and allocation of
water, energy, land, and food resources at the community level need to be
examined in a more systemic, multidisciplinary, participatory, and prac-
tical manner while seeking to increase synergies and reduce trade-offs.
Such an integrated approach is not yet mainstream among those involved
in the science and policy decision aspects of the nexus. This book was
written to explore the value proposition of that approach.
This two-volume book describes a flexible and adaptive system-based
methodology and associated guidelines for the management and alloca-
tion of community-based WELF resources. Volume 1 reviews the existing
literature about the nexus and focuses on defining the landscape in which
it operates. The proposed methodology is also outlined. Volume 2 explores
the quantitative and qualitative modeling of the nexus and landscape using
system modeling tools including system dynamics. It presents a road map
for the formulation, simulation, selection, and ranking of possible inter-
ventions, as well as the development of possible intervention plans.
vi  •   Abstract

The proposed methodology is designed to serve as a guide for differ-


ent groups of scientists, engineers, policy makers, and decision makers
when selecting intervention strategies for the management and allocation
of WELF resources within the broader context of community develop-
ment. The methodology focuses mostly on WELF-related issues in small-
scale and low-income communities where resources are limited and
securing resources is critical to their short- and long-term livelihood and
development.

KeyWords

adaptive; behavior; community; context; complexity; development


­projects; patterns; participation; reflective practice; systems approach;
systems thinking; structure; satisficing; water-energy-land-food nexus;
system dynamics
Contents

List of Figures ix
List of Tables xiii
Preface xv
5   A System Dynamics Approach to the Nexus 1
5.1 Introduction 1
5.2 Hard Systems Approaches to the Nexus 3
5.3  Systems Thinking 14
5.4  System Dynamics 18
5.5  System Dynamics Modeling 33
5.6  Combining Methods of Analysis 44
5.7  Participatory Model Building 49
5.8  Concluding Remarks 52
5.9 References 53
6   System Dynamics Models of the Nexus 67
6.1 Introduction 67
6.2  Model WELF-02 73
6.3  Model WELF-03 76
6.4 From Resources to Services: Model WELF-04 79
6.5  Model WELF-05 83
6.6  The WELF-G Model 85
6.7  Population and Capacity Module 87
6.8  Food Resources Module 95
6.9  Land Resources Module 95
6.10  Water Resources Module 108
6.11  Energy Resources Module 116
6.12  Sensitivity Analysis 118
6.13 References 120
viii  •   Contents

7   Exploring System-Level Interventions 123


7.1  Intervention Dynamics 124
7.2  Formulating Intervention Scenarios 128
7.3 Evaluating and Selecting Intervention Scenarios 147
7.4  Multiobjective Decision Analysis 150
7.5  Decision Making 158
7.6 From Selecting to Implementing Intervention Scenarios 168
7.7 References 169
8   Conclusions 177
8.1  An Integrated Approach 177
8.2 Challenges 181
8.3 References 183
Appendix A 185
Example A1 185
Example A2 187
Example A3 189
Example A4 192
Example A5 194
Appendix B 197
About the Author 209
Index 211
List of Figures

Figure 5.1. Causal loop diagram showing several cause-and-­effect


loops at play in a disadvantaged neighborhood in
Costa Rica. 23
Figure 5.2. Basic building blocks of stock-and-flow diagrams:
Stocks (reservoir or conveyor), flow (inflow and
outflow) with clouds, converters, and connectors. 24
Figure 5.3. Stock-and-flow diagram illustrating the d­ ynamics
at play between a population and the WELF
­(aggregated) resources upon which it depends. 26
Figure 5.4. Causal loop diagram illustrating the limits to success
dynamic across the water, energy, and agriculture
sectors of the nexus. 32
Figure 5.5. Tragedy of the commons causal loop diagram showing
competition for water resources between the electricity
and food sectors. 33
Figure 5.6. Different components of system dynamics modeling. 34
Figure 6.1. Diagram showing model complexity vs. usefulness.
Note the ­threshold beyond which adding c­ omplexity
­significantly reduces (1) or has l­imited impact on
(2) ­usefulness. 72
Figure 6.2. Stock-and-flow diagram illustrating the dynamics at
play between a population and the WELF resources on
which it depends. All four resources are disaggregated
and considered in an array form. 75
Figure 6.3. Stock-and-flow diagram illustrating the dynamics at
play between a population and the WELF resources
upon which it depends. (a) Links between the water,
energy, land, and food resource modules and the ­
x  •   List of Figures

 population module; (b) Nuts and bolts of the water


resources module (all other modules have the same
­structure); (c) Population module and calculation of
the overall weighted WELF nexus security. 77
Figure 6.4.  System dynamics model of the methodology proposed
by ­researchers at the University of Virginia at
­Charlottesville (UVC). (a) Modular r­ epresentation
of the ­different types of capacity; (b) Dynamics of
the ­technical capacity ­module; (c) ­Determining
the TML and TRL for selected water s­ ervices
source, ­procurement, s­ torage, distribution, and
­treatment  options. 82
Figure 6.5.  System dynamics of the Pardee RAND food–energy–
water security index. (a) Each sub-index is
represented by a m
­ odule; (b) Breakdown of the
water module sub-index model. 84
Figure 6.6.  Five modules considered in the WELF-G system
dynamics model. 86
Figure 6.7a. Population chain distribution by age group and gender. 88
Figure 6.7b. System dynamics representation of the health part
of the population submodule. 90
Figure 6.7c. Agricultural workforce and employment part of the
population submodule. 92
Figure 6.8.  System dynamics model to determine community
capacity to p­ rovide services of Water Supply and
Wastewater Treatment (WWT). 93
Figure 6.9.  System dynamics representation of the food module.
(a) Plant-based food; (b) Meat-based food. 97
Figure 6.10a. System dynamics representation of the land zoning
­submodule. 99
Figure 6.10b. System dynamics representation of the
soil/vegetation submodule. 100
Figure 6.10c. System dynamics representation of the
residential–rural land submodule. (c1) Homes;
(c2) Infrastructure. 103
Figure 6.10d. System dynamics representation of the agricultural
land ­submodule. (d1) Agricultural land dynamic;
(d2) Arable land productivity; (d3) L
­ ivestock unit
­productivity. 107
List of Figures   •   xi

Figure 6.10e. System dynamics representation of the industrial/­


commercial land submodule. 109
Figure 6.10f.  System dynamics representation of the wild land
­submodule. 110
Figure 6.11. System dynamics representation of the water
balance submodule. 113
Figure 6.12. System dynamics representation of the wastewater
­submodule. 115
Figure 6.13. System dynamics representation of the energy
balance submodule. 117
Figure 7.1.  Dynamic between current state, desirable state, and
­sustainability state of a community using a goal-seeking
archetype. Bi-flows are used to model possible increases
or decreases of the current and desirable states. 127
Figure 7.2.  Scenario trumpet representation of formative
scenario analysis. 131
Figure 7.3.  Dynamic of interaction between insiders and outsiders
using an oscillation archetype. Bi-flows are used to
model any positive or negative adjustments in the
interaction between insiders and outsiders. 137
Figure A1.1.  Variation of population and WELF resource security
over 50 years. 185
Figure A1.2.  Variation of (a) population and (b) WELF resources
security over 50 years. Sensitivity analysis was
carried out by varying the WELF supply rate from 0.1
to 0.3/year and the WELF resource consumption rate
­ranging between 1 and 2 units/pp/year. Run 1: 0.1, 1;
Run 2: 0.1, 2; Run 3: 0.2, 1; Run 4: 0.2, 2;
Run 5: 0.3, 1; Run 6: 0.3, 2. 186
Figure A1.3.  Time-dependent variation of (a) population and
(b) WELF resources security mean values and
50 percent confidence level. 187
Figure A2.1.  Variation of (a) Population and (b) WELF Nexus
Security NS, and Water, Energy, Land, and Food
Security over 50 Years. 188
Figure A3.1. Variation of (a) population and (b) WELF Nexus
Security NS, and water, energy, land, and food
securities over 50 years. 191
Figure A4.1. Variation of TML and TRL over 50 years. 193
xii  •   List of Figures

Figure A5.1. Variation of water, energy, food, and FEW


securities over 50 years. 195
Figure A5.2. Time-dependent variation of FEW nexus security
index mean value and 100 percent confidence level. 195
List of Tables

Table 5.1. Characteristics of the energy, water, and land use


models in CLEWS 7
Table 5.2. Fourteen habits of a systems thinker according to the
Waters Foundation 16
Table 5.3. System archetypes and related community
development and FEW/WELF nexus examples 29
Table 5.4. Tests used to gain confidence in system dynamics
models. This table was constructed based on the work
of Sterman (2000, pp. 859–61) and Forrester and
Senge (1980) 37
Table 5.5. Questions that may arise when creating a dynamic
hypothesis and formulating a system dynamics model 39
Table 7.1. Common strategic framework in the management of
development projects 132
Table 7.2. Relationship between scenario development processes
and scenario outcomes across geographic scales 135
Table 7.3. Example of cross-impact balance (CIB) analysis
involving three groups of stakeholders 139
Table 7.4. Possible storyline components in WELF scenario
planning144
Table 7.5. Example of multicriteria utility assessment matrix
for a project in Nepal 152
Table B1.1.  Variables, parameters, and multipliers used in the
population submodule (Figures 6.7a–c) 197
Table B1.2.  Variables, parameters, and multipliers used in the
community capacity assessment submodule
(Figure 6.8) 199
xiv  •   List of Tables

Table B2. Variables, parameters, and multipliers used in the


food resources module (Figures 6.9a,b) 201
Table B3. Variables, parameters, and multipliers used in the
land resources module (Figures 6.10a–f) 201
Table B4. Variables, parameters, and multipliers used in the
water resources module (Figures 6.11 and 6.12) 205
Table B5. Variables, parameters, and multipliers used in the
energy resources module (Figure 6.13) 208
Preface

This two-volume book is a follow-up to my book titled A Systems


Approach to Modeling Community Development Projects published by
Momentum Press in 2015. More specifically, it addresses the application
of systems thinking and systems tools to model the complex dynamics
at play when considering the security of water, energy, land, and food
(WELF) resources at the community level. Managing these resources and
their allocation has become critical in the overall discourse on sustain-
able development as humanity is facing global challenges related to rapid
urbanization, population growth, economic growth, and climate change.
The discussion on how to meet human needs for water, energy, land,
and food, and how to guarantee their respective securities has changed
over time. The traditional way has been to look at all four sectors in iso-
lation, regardless of whether one is interested in supply and demand,
infrastructure planning and design, resource management and allocation,
and/or governance. Since 2011, however, there has been a new empha-
sis on understanding the interdependency of the four sectors through the
so-called WELF nexus. The literature on the WELF nexus is abundant
especially about the “why” and “what” of an integrated approach to the
management and allocation of WELF resources. Less clear is the “how”
of that management, the “who” of who should be involved in it, and the
importance of scale and context (“where” and “when”) when designing
appropriate management interventions that more often require decision
makers to select trade-offs rather than synergistic interventions.
The WELF nexus is addressed in this book using an integrated or
systemic approach. It has been my experience that the concept of sys-
temic approach resonates well with individuals who pride themselves in
showing that they are forward thinkers. Very often, however, they stop
short of applying systems tools to address complex challenges and fall
back into their more comfortable deterministic mindsets. Shallow sys-
tem thinking combined with an absence of follow-up decision making
xvi  •   Preface

remains an ­intellectual exercise with no meaningful results. This book


was designed to show aspiring systems thinking decision makers the value
­proposition of qualitative and quantitative systems tools in addressing
complex problems.
At the outset, this book acknowledges that WELF nexus-related
problems belong to a class of complex problems that cannot be readily
formulated and modeled. These problems are sometimes referred to in
the literature as messy or ill-defined problems. The book also emphasizes
that the complexity and uncertainty of nexus-related problems should
be embraced and acknowledged, with all the advantages and limitations
this entails. Failure to recognize these attributes is equivalent to using
the wrong tools to solve a problem: it will lead to making the same mis-
takes that have been made in the past when addressing the sectors of the
nexus in isolation. These mistakes are emerging consequences of using a
traditional reductionist, deterministic, and rigid mindset that is based on
simplifying complex problems, breaking the problems into pieces, finding
experts to solve each piece, and putting the solutions side by side without
due consideration of initial links between the problems and links between
the proposed solutions. This traditional approach to tackling development
projects also makes the false claim that clear and optimal solutions to
complex and uncertain issues are possible.
This book is not meant to be an introductory book on systems or com-
munity development. It should be seen instead as a manual that provides
a methodology to address specific WELF resource-related challenges in
a specific context and at a given scale. The methodology emphasizes that
the management and allocation of WELF resources unfolds in a landscape
consisting of multiple interacting systems and subsystems. Furthermore,
it is a dynamic process that must be addressed in a participatory and mul-
tidisciplinary manner. Finally, the methodology requires decision makers
to become systems thinkers and use multiple system-related tools such
as system dynamics combined with other analysis tools such as causal
analysis, cross-impact analysis, network analysis, and multi-criteria
­
­decision analysis.
This book emphasizes that the management and allocation of WELF
resources is always part of a community development story with a ­narrative
describing how community members envision progressing from a current
to a desired state of development. What is the story? How does it manifest
itself at the community level? Can a systemic structure be created to model
current community issues and behavior patterns and propose scenarios of
intervention to address these issues? Simply put, these questions cannot
be answered by decision makers who favor quick technical solutions to
Preface   •   xvii

every problem. The management and allocation of WELF resources in the


context of community development is a journey that takes time. This book
is more about the methodology or road map necessary for that journey to
unfold in order to create long-lasting resource security and benefits.
This book is the third of a series of three books that I have written
about small-scale community development. As I have progressed in try-
ing to understand the multiple dimensions of that complex field of study
and have purposely worked on small-scale community projects, my
approach to community development has become more holistic and cer-
tainly less engineering like as time goes on. That journey has also helped
me to reinforce that human development is foremost about people, their
­empowerment, and their transformation. More importantly, it is about the
human story.
After reading this book or parts thereof, I hope that you will be more
attentive to development stories. They are worth listening to. You may
also find yourself in most of them. I hope that this book will be useful to
you and your projects. The approach used in this book is not written in
stone. Feel free to develop new approaches and share them with the rest
of us.
I want to thank Momentum Press for giving me the opportunity to
publish my work. I also want to thank Ms. Allison Goldstein for ­editing
this book and three of my colleagues for reviewing the manuscript and pro-
viding feedback. Finally, I want to thank my wife Robin and our ­children
Elizabeth Ann and Alex for their support, patience, and love.

Bernard Amadei
Boulder, December 01, 2018
CHAPTER 5

A System Dynamics
Approach to the Nexus

This chapter reviews first different soft and hard system-based formula-
tions that have been proposed in the literature to model the interaction
between several components of the FEW/WELF nexus. It then presents a
value proposition for using qualitative and quantitative system ­dynamics
simulation models to capture the dynamics at play across the nexus, includ-
ing interactions across the nexus components and their interactions with
social, natural, infrastructure, and economic systems. The chapter presents
the characteristics of systems thinking and reviews basic system dynamics
tools, systems archetypes, and the different steps involved in the system
dynamics modeling methodology. As a hard systems approach, system
dynamics has unique characteristics that warrant its use when ­simulating
how the different sectors of the nexus unfold and interact when subjected
to different inputs and constraints. In turn, the simulations help decision
makers to make management decisions and select interventions across the
nexus and at the community level. The chapter concludes with a discussion
of the value added by using social network analysis and the importance of
engaging in participatory modeling to develop system dynamics models.

God grant me the serenity to deal with ill-defined and messy systems
over which I have little or no control, courage to improve systems
over which I have some control, and wisdom to know the difference.
—adapted from the Serenity Prayer, R. Niebuhr

5.1 Introduction

Volume 1 emphasized the value proposition that the landscape in which


the FEW/WELF nexus unfolds and community development takes
2  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

place is a complex and adaptive landscape and argued that it should be


­handled as such when managing and allocating water, energy, land, and
food resources. The landscape consists of multiple complex, ­complicated,
or simple systems and subsystems, with each providing functions and
­purposes with different degrees of intra- and inter-connectivity. This
global picture suggests that adopting a systems mindset, rather than a
deterministic sectorial mindset, is better for appraising the landscape and
formulating dynamic hypotheses around critical issues that are at play in
the landscape. It also implies that the management and allocation of FEW/
WELF resources within the overall context of community development
must be conducted with an integrated (e.g., systemic) and participatory
perspective by considering the following dynamics: (i) the water, energy,
and food sectors and their supply chains (from acquisition, to production
and processing, to consumption, and ultimately to waste disposal), includ-
ing their dependencies; (ii) the links between these sectors and land; and
(iii) how the nexus as a whole interacts with human/social, natural, infra-
structure, and economic systems.
It should be noted that the FEW/WELF literature has mostly empha-
sized why an integrated or systems approach should be used to understand
the nexus dynamics (Alcamo 2015; McCormick and Kapustka 2016;
Wolfe et al. 2016; among others). This, in turn, has helped decision and
policymakers to become aware that: (i) synergies and trade-offs between
sectors of the nexus need to be considered in the management and alloca-
tion of water, energy, land, and food resources, and (ii) decisions at one
spatial or temporal scale may impact decisions at other scales. Finally,
it has become clear that the FEW/WELF nexus cannot be separated and
analyzed independently from the landscape in which it unfolds.
Compared to the “why,” the “what” and especially the “how” of the
integrated perspective of the FEW/WELF nexus is still a work in progress
in the literature where different versions of both soft and hard systems
approaches have been advocated. As noted by Checkland and Poulter
(2006), both approaches have different characteristics. A soft systems
approach to a problem focuses more on the “systemic” process of inquiry
and “problem situations,” but not on the structure of the problems per se.
On the other hand, a hard systems approach includes the development of
systems models (qualitative and quantitative) to address both the systemic
and structural aspects of the problem being analyzed.
The soft systems approach to the nexus has been formulated and
presented in many different forms. Graphical representations such as
those shown in Figures 1.1 and 1.2 (and multiple variations of it) abound
and have been proposed, for instance, to capture the two- or three-way
A System Dynamics Approach to the Nexus   •  3

i­nteractions between water, energy, and food and how these sectors
interact with other factors such as land, climate, the environment, or
­

waste (WBCSD 2009; Bazilian et al. 2011; Hoff 2011; FAO 2014; Wake-
ford et al. 2015; Sohofi et al. 2016; among others). Double entry matrixes
have also been proposed to show links between the different sectors of the
nexus and their indicators (Flammini et al. 2014; Ferroukhi et al. 2015;
Wakeford et al. 2015; among others). In Chapter 3, a soft approach was
used, for instance, to emphasize the types of impacts that exist across the
nexus components (Table 3.5). In Chapter 4, the emphasis was on map-
ping the impacts that exist among the four system groups at play in the
landscape (Table 4.2).
A soft systems approach is clearly a step toward considering all sec-
tors of the nexus in isolation. However, it stops short of being able to
quantitatively model the dynamics and structure of the nexus in depth and
how its sectors interact with the landscape and its systems. This could be
handled better using a hard systems approach with system dynamics mod-
eling tools, as discussed in Volume 2 of this book.

5.2 Hard Systems Approaches to


the Nexus

Various hard system-based formulations have been proposed in the liter-


ature to model the combined systemic and structural aspects of the inter-
actions among components of the FEW/WELF nexus. They fall over a
spectrum that ranges from global to sector-specific formulations. Among
all existing hard systems formulations, those that use the system dynamics
modeling methodology (discussed further in this chapter) seem to offer the
most value in modeling the management and allocation of FEW/WELF
resources over time.

5.2.1  Global Formulations

Since 2011, the development of hard system models of the complex


and adaptive landscape in which the FEW/WELF nexus unfolds and
­community development takes place has happened in parallel to for-
mulations used in the closely related field of international development
and aid. With some modifications, both approaches could be combined
to ­integrate systems thinking into nexus studies and in the context of
community development.
4  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

As described, for instance, in the book A Systems Approach to Mod-


elling Community Development Projects (Amadei 2015), there has been a
strong interest in the global development literature in adopting a systems
perspective to account for the interactions of socioeconomic-political and
technical issues. The book, Aid on the Edge of Chaos, by Ramalingam
(2014) provides an excellent introduction to the scientific study of com-
plex systems and reviews a wide range of applications of systems thinking
in global development and aid-related topics.
Ramalingam’s book builds on several other authors’ observations that
in development and aid, complexity and systemic behavior are the norm
and not the exception and therefore should be embraced instead of ignored
(Rihani and Geyer 2001; Rihani 2002; Mavotras 2010; Porter  2011;
­Breslin 2004; Neely 2015; Wu 2016). Despite international development
and aid’s slow progress toward incorporating systems thinking and the
science of complexity, some studies over the past 10 years have clearly
demonstrated the benefits of using systems thinking and systems mod-
eling, primarily when making decisions and developing systemic policy
solutions at the regional or country scale (Reynolds et al. 2003; Hjorth and
Bagheri 2006; Scoones et al. 2007; Newell et al. 2011; USAID 2011a,b;
2014; Williams and Britt 2014; Fowler and Dunn 2014; Neely 2015;
Downes and Bishop 2016; among others).
The methodologies associated with the studies mentioned above,
however, have a limited range of applications if they were to be used
for making decisions and proposing solutions in the management and
allocation of FEW/WELF resources. First, they are not designed for
­making decisions and proposing solutions at the community scale, which
is the scale of interest in this book. Second, they are qualitative since
they mostly emphasize feedback mechanisms (causal loops, as discussed
below) to represent how different sectors interact. This in turn implies
that any decisions made and solutions proposed using these methodolo-
gies should be considered with caution, since no quantitative analyses are
conducted, and the decisions and solutions are therefore subjective. This
limitation was emphasized, for instance, by Sohofi et al. (2016) regard-
ing the use of causal loops to study the various types of trade-offs at
play when trying to ensure simultaneous success in water security, energy
security, and food security.
On the quantitative side, several system-based frameworks have been
developed to model global development and conduct global forecasting
using system dynamics in the so-called field of futures research. They can
be classified as world modeling frameworks, according to the ­terminology
proposed by Hughes (2016). One of these frameworks is the World3-03
A System Dynamics Approach to the Nexus   •  5

“world model” of Meadows and coworkers, which looked at different


scenarios or pathways of world development by considering the links
between economic growth and the environment (Meadows et al. 2004). It
succeeded three earlier “limits to growth” models (World2, World3, and
World3-91) developed at MIT starting in the 1970s. A more recent for-
mulation of World3-03 can be found in Bossel (2007c, Chapter 6) who
used modern system dynamics tools to model the various systems con-
sidered in World3-03. The modules considered by Bossel (2007c) include
population, industry, employment, services, agriculture, soil fertility, food
production, environmental pollution, and nonrenewable resources. The
World3-03 model was also recalibrated using more recent data from 1995
to 2002 around industrial and service output, population, birth rate, life
expectancy, and arable land (Pasqualino et al. 2015).
Another system-based world modeling framework has been devel-
oped by the Pardee Center for International Futures (IFs) at the U
­ niversity
of Denver in Colorado. This system dynamics model accounts for links
between various systems and subsystems involved in development such
as economics, agriculture, population, energy, education, health, envi-
ronment, infrastructure, and socio-political systems (Hughes 2016). The
framework, built on solid up-to-date database, is designed to explore
various future trends and “what-if” (or “what-happens-if”) projections or
scenario planning at the country level (for 186 countries) and over a time-
frame as far into the future as 2100 (Hughes 1999; Hughes and Hillebrand
2006; Hughes 2016). The framework can also be used for regional and
global predictions and forecasting (Hughes et al. 2015). Version 7.31 of
the IFs software is available on the Pardee Center website (www.pardee.
du.edu, September 2018).
In parallel to the world modeling frameworks mentioned above are
system-based frameworks that have been developed to model and forecast
changes associated with human–environment interactions and global envi-
ronmental changes. They can be classified as integrated assessment mod-
eling (IAM) frameworks, according to the terminology used by Hughes
(2016). One of them, called TARGETS, which stands for Tool to Assess
Regional and Global Environmental and Health Targets for Sustainability,
was developed by Rotmans and deVries (1997). It consists of five inter-
acting dynamic submodels covering the areas of population and health,
energy, land and food, water, and biogeochemical cycles. Another IAM
model called ANEMI, developed by Davies and Simonovic (2010, 2011),
considers “eight [interacting] sectors of the society-biosphere-climate
system—climate, carbon cycle, land use, population, surface water flow,
water use, water quality, and the economy” (Davies and Simonovic 2010).
6  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Other integrated assessment modeling frameworks have been pro-


posed in the literature. Jaeger and Henrichs (2008) conducted a review of
about 80 such frameworks to assess environmental changes at the scale
of Europe. The frameworks are divided into three groups: (i) frameworks
with a thematic focus (e.g., water, energy, food and agriculture, land use,
biodiversity, transport, wastewater, forests); (ii) frameworks based on
geographic scale, with a focus on either the full area covered (extent) or at
a more detailed level (resolution); and (iii) frameworks based on either a
single analytical method (e.g., equilibrium, empirical/statistical, dynamic
modeling, interactive) or several complimentary methods (hybrid). A sep-
arate group of frameworks consists of those that are integrated such as the
IFs and TARGETS frameworks mentioned above. Additional information
about integrated assessment modeling frameworks can be found on the
website of the International Institute of Applied Systems Analysis (IIASA,
www.iiasa.ac.at).
Another global dynamic modeling framework that explores various
strategy and policy scenarios in an integrated manner—more specifically
across the FEW/WELF nexus, including the effect of climate change—is
called CLEWS, which stands for Climate-Land-Energy-Water Strategy
(Bazilian et al. 2011). It integrates three existing frameworks that can
be used to create intervention scenarios and carry out policy ­analysis
around water, energy, and land resources and the impact of climate on
these resources. The frameworks include: (i) the long-range energy alter-
natives planning system (LEAP; www.energycommunity.org) model for
energy policy analysis; (ii) the water evaluation and planning system
(WEAP; www.weap21.org) for water management and policy analysis;
and (iii) the global agro-ecological zones (GAEZ, http://fao.org/nr/gaez/
en/) model for land use. The three models are part of three modules in the
CLEWS ­framework that allow the water, energy, and land resources to
interact where the output of one module becomes the input of another one.
Examples of how this framework can be applied can be found in Hermann
(2012), Hermann et al. (2012), Howells et al. (2013), Welsch et al. (2014),
and Dale et al. (2015).
The CLEWS framework allows for scenario-based analysis around
issues such as energy shortage, water shortage, and/or food shortage while
considering climate effects. It also considers the links between supply
and demand in each sector and across sectors. Table 5.1 summarizes the
characteristics of the three types of analysis considered in CLEWS. The
CLEWS framework has mostly been used at scales much larger than the
local scale of interest in this book. Nevertheless, the framework is worth
mentioning because it represents an attempt at linking several sectors of
A System Dynamics Approach to the Nexus   •  7

Table 5.1.  Characteristics of the energy, water, and land use models
in CLEWS
Energy Water Land Use
MESSAGE (IAEA,

IIASA)
WEAP(SEI) plus
Model GAEZ(IIASA)
LEAP (SEI)
­several others
OSeMOSYS(KTH)

Small scale to
From small island Local water
Scale of country analysis
systems to large systems based on
operation (flexible grid
country analysis geographical data
cells sizes)

• Demand • Climatic data • Climatic


­(current/future, • Land cover data data (plus
load curves) • Soil data and ­projections)
• Existing + water available • Land cover
planned power • Water data
plants consumption • Soil data
Input
• Imports and • Desalination and
exports, and hydropower
resource
availability
• GHG emission
factors
• Future optimal • Water availability • Crop map
energy mix under different (most suitable
under different scenarios for all crops per area)
Output
conditions points in a • Crop calendar
and
• Future GHG modeled system • Future water
results
emissions demand
• Costs • Fertilizer
demand
Source: http://indico.ictp.it/event/a11197/session/40/contribution/30/material/0/0.
pdf (accessed November 15, 2017).

the nexus and the climate. It should be noted that the CLEWS approach
does not use system dynamics per se but rather a combination of mass
balance and linear programming. Another limitation of CLEWS is that,
although it allows the output of one of the three modules to become
the input of one or two of the other modules, no feedback mechanisms
8  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

between the modules are taken into consideration; only feedbacks within
the water, energy, and land modules are considered.
Another framework similar to CLEWS (and with the same limita-
tions) is the WEF Nexus Tool 2.0 proposed by Daher and Mohtar (2015).
Used at the country level, it starts with an assessment of country con-
ditions related to available food products, water sources (precipitation,
surface, groundwater, graywater, desalination), and use (domestic, agri-
cultural, and industrial); energy sources (nonrenewable and renewable)
and use (water, agriculture, transportation); and types of import. The user
selects a scenario with clear objectives where these different variables
interact. For instance, the scenarios could be used to explore different
levels of a country’s food/water/energy self-sufficiency and dependence
on import.
Finally, a more recent quantitative systems approach framework is
proposed by the Millennium Institute around the 17 Sustainable Develop-
ment Goals (SDSN 2015). The Integrated Sustainable Development Goals
(iSDG) planning model uses a system dynamics modeling methodology to
explore development scenarios linking strategies and policy as a country
progresses toward reaching the 17 SDGs with projections as far into the
future as 2030 and beyond. The iSDG model builds on the functionalities
of the Threshold 21 (T21) model (Barney 2002) and involves 30 different
sectors (10 social, 10 economic, and 10 environmental), each consisting
of interconnected stocks (variables) and flows with possible feedback pro-
cesses. Version 1.2 of the software is available on the Millennium Institute
website (www.isdgs.org; September 2018).
It should be noted that the global frameworks mentioned above can
all handle high levels of complexity and allow for high levels of decision
making around nexus issues. However, all of them have been mostly used
to model the complex dynamics of change and development at the global,
regional, or country scale. Furthermore, they all reveal how difficult it is to
obtain data to quantify the links considered in the frameworks. With some
modifications, however, they could be applied to model the dynamics at
play in the complex but smaller-scale landscape shown in Figure 4.1a and
account for the two-, three-, or four-way interactions across the nexus and
the complex interactions between the nexus and social, natural, infrastruc-
ture, and economic systems.

5.2.2 Sector-Specific Formulations

On the other end of the spectrum of frameworks from the global ones
described above are those that focus on the specific management of water,
A System Dynamics Approach to the Nexus   •  9

energy, or food systems, but in a sectoral manner (with one or two inter-
reacting sectors). Like the global frameworks mentioned above, these
frameworks address issues at the country, regional, or global scales.
Sohofi et al. (2016) give an excellent review of many of these different
topical frameworks and models.
It should be noted that, despite being focused on the management of
specific resources, many of the sector-specific formulations also incor-
porate, in a broad manner, other factors interacting with the resources
being considered. For instance, most water resource management frame-
works include socioeconomic and environmental issues and factors such
as energy and agricultural demand, pollution, land use, climate change,
and so on. Likewise, food resource management frameworks account for
land and soil resources and water resources. Without going into the details
of these frameworks and models, the more notable ones that use system
dynamics are listed below.

5.2.2.1  Water Resources Management

System-based frameworks have been suggested by many authors for the


management of water resources (Hamilton 1969; Biswas 1976). The WEAP
framework mentioned above is an example of a more recent ­framework
(Yates et al. 2015). Extensive reviews of system dynamics models pro-
posed in the water resources management literature have been carried out
by Simonovic and Rajasekaram (2004); Ahmad and Simonovic (2004);
Simonovic (2008); Ahmad and Prashar (2010); Mirchi et al. (2012); and
Zhuang (2014).
In general, there is no one unified integrated model of water resources
management, as all models have been developed to meet specific goals.
The models can, however, be regrouped into several categories or
themes. For instance, Mirchi et al. (2012) classify existing models into
“(i) ­predictive simulation models; (ii) descriptive integrated models; and
(iii) ­participatory and shared vision models.” Zhuang (2014) distinguishes
four groups of system dynamics models that are used for the ­“examination
of management strategies for regional water resources, evaluation of
­management strategies for utility companies, simulation of ­hydrological
processes, and explanation of stakeholder participation.” Examples of
­system dynamics modeling being applied to water resources management
at different scales (country, urban, or basin/watershed) can be found in
Guo et al. (2001); Xu et al. (2002); Khan et al. (2007); Chung et al. (2008);
Winz et al. (2009); Cheng (2010); Gober et al. (2011); Takahashi (2012);
10  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Dawadi and Ahmad (2012); Winz and Brierley (2013); and Turner et al.
(2016), among others.
The WorldWater model is another example of a system-based
framework that explores strategies of water management at large scales
(Simonovic 2003); it is based on the World3 model mentioned above.
Examples of how this model is applied at the country scale (e.g.,
­CanadaWater model) or within regions of a country are described by
Simonovic and Rajasekaram (2004) and Ahmad and Simonovic (2004).

5.2.2.2 Energy Resources Management

Like for water, system-based frameworks have been proposed in the


literature for the management of energy resources. Aside from the
LEAP framework (SEI 2013) mentioned above, other frameworks
using methods such as linear and dynamic programming, equilibrium
models, optimization, scenario analysis, and agent-based modeling
have been proposed to model the dynamics of energy policies, a good
review of which can be found in Qudrat-Ullah (2016). As mentioned
by this author, system dynamics quantitative and qualitative modeling
has been shown to have a strong value proposition to capture the com-
plex interactions at play in the management of energy resources and
account for uncertainties, nonlinear forms of behavior, feedback mech-
anisms, delays, and so on. In fact, as noted by Ford (1996), system
dynamics modeling for energy is not new and has been used as a tool in
defining energy policy since the 1970s (see Naill 1992) around several
themes such as the design and assessment of regional energy and elec-
tricity polices, the effects of privatization, defining energy efficiency
and management, assessing electricity markets, considering the expan-
sion of generation capacity, and assessing energy-related emissions and
environmental impact. A review of these system dynamics models can
be found in Naill (1992); Ford (1996); Moumouni et al. (2014); and
Qudrat-Ullah (2016).
As with water resources management, there is no such thing as one
unified integrated model of energy resources management. ­Different sys-
tem dynamics models have been proposed in the literature for ­different
purposes such as analyzing fossil fuel and renewable energy resources;
modeling electricity supply and demand/consumption in the residential,
agricultural, transportation, and industrial/commercial ­sectors; exploring
and planning energy efficiency, reliability, security, and independence in
one or several energy sectors; comparing renewable and nonrenewable
A System Dynamics Approach to the Nexus   •  11

sources of energy supply and demand; and predicting trends of pollution


and greenhouse gas (GHG) emissions for different scenarios of energy
­supply and demand at different scales (Ford 1996; Kiani et al. 2010; Jeon
and Shin 2014). Zhuang (2014) provides a review of about 60 system
dynamics models used in energy resource management that are ­subdivided
based on whether policy analysis or investment analysis is considered.
A limited selection of well-documented case studies worth exploring
(among many others) that illustrate the range of application of existing
system models can be found in: (i) Chi et al. (2009); Newell et al. (2011);
Moumouni et al. (2014); Robalino-Lopez et al. (2014); and Aslani et al.
(2014) at the scale of countries such as the UK, ­Australia, Niger, Ecuador,
and Finland, respectively; (ii) Ansari and Seifi (2013) pertaining to energy
consumption and CO2 emissions related to the cement industry in Iran;
and (iii) Feng et al. (2013) at the city scale.

5.2.2.3 Agro-Food and Land Resources Management

A fair number of system-based frameworks have been proposed to


make decisions and develop policies in the fields of agriculture, land
use planning, forestry, and natural resources management. Compared to
frameworks that use linear and equilibrium models (sometimes called
“neoclassical” models), system-based frameworks can capture the
­multiple dimensions, complexity, and uncertainty associated with these
different fields. The global agro-ecological zones (GAEZ) model is a
good example of a decision-making tool for agricultural development
and resource security.
As with water and energy resources, system dynamics models have
been used to model different issues related to land and soil (e.g., land
productivity, soil erosion, soil nutrient and water-holding management,
reclamation strategies), food security (e.g., supply and demand, produc-
tion to consumption, effect of scale, resilience to adverse events), and
smallholder crop-livestock farming practices and associated economic
development. Discussions on system dynamics modeling and usefulness
for addressing agro-food and land resources management in specific case
studies can be found, for instance, in Monasterolo et al. (n.d.); Giraldo
et al. (2008); and Turner et al. (2016). It should be noted that the food
­production component of the World3 model developed at MIT in the
1970s (mentioned above) is often cited as a starting point in the ­agro-food
and land resource management system dynamics models p­ roposed in
the literature.
12  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

5.2.2.4 Other Related Frameworks

Qualitative and quantitative system-based frameworks have also been


developed to model the dynamics in other areas related to development
and aid, such as health, state instability, and corruption. These topics are
mentioned here because they have potential to interact with one or several
sectors of the nexus and play a role in the management of FEW/WELF
resources. For instance, in the field of health, integrated models have
been used to capture interactions among the key building blocks of health
systems, which, per the World Health Organization (WHO), include: ser-
vice delivery; health workforce; information; medical products, vaccines,
and technologies; financing; and leadership governance (de Savigny and
Taghreed 2009). More specifically, system-based models have been sug-
gested to model the connections between health and other human and social
activities (e.g., economic and infrastructure development) that contribute
indirectly to the health and well-being of communities (Newman et al.
2003; Lebel 2003; McDonnell et al. 2004; Rwashana and Williams 2008).
A systems approach has also been used to model chronic disease preven-
tion and the interaction of multiple diseases (Janssen and Martens 1997;
Homer and Hirsch 2006; Jones et al. 2006) and ­epidemics (Zhou 2014).
In the context of policy and health, Newman (2010) describes several
examples of in-country applications of systems thinking to development
to understand the consequences of policy decisions at the country level on
community health. More information about the use of the system dynam-
ics modeling methodology in health can be found in the proceedings of
a workshop on strengthening health systems in low- and middle-income
countries (Future Health Systems 2014).
Another interesting system-based approach was used by Choucri
et al. (2005) to examine possible drivers of state instability, resilience, and
capacity (institutional, technical, social, financial, and so on). The ratio-
nale is that the stability of the state is driven by a multitude of interacting
social, economic, and geopolitical pressures, which when combined may
reach a tipping point beyond which state resilience is exceeded and con-
flict arises. It is not difficult to envision how water, energy, land, and/or
food insecurity could contribute to that dynamic.
Another domain that is being addressed in a systemic way is corrup-
tion. This is an important topic in the management of resources across the
nexus, as it has potential to undermine the effectiveness of technical inter-
ventions and policy decisions. One example of using system dynamics
to model corruption is provided by Ullah (2012), who offers an ­excellent
A System Dynamics Approach to the Nexus   •  13

literature review of models of corruption and a framework to model


­corruption specifically in Pakistan.

5.2.2.5 Remarks

The different hard system formulations discussed above represent a


valuable and somewhat overwhelming portfolio of existing models that
address one or several components of the nexus in great depth and capture
possible interconnections. They can indeed be valuable to modelers and
decision makers in developing custom-made models to address specific
issues. The author benefitted a lot from exploring in detail some system
dynamics models and associated case studies available in the literature.
In many instances, however, the system-based models reported in the
literature lack documentation and transparency, especially when written
by research-oriented academics. They often lack a clear discussion on the
what, why, where, when, who, and how of model development and the
decision process involving different groups of stakeholders. The mod-
els can also be hard to reproduce. As a result, nonspecialists who are not
familiar with systems thinking and modeling—especially those who will
be affected by decisions made based on these models—may find these
models hard to comprehend. For that reason, a conscious decision was
made to propose, in the forthcoming chapters (Chapters 6 and 7) of this
book, generic system dynamics models that balance simplicity and com-
prehensiveness. The models were purposely designed to offer more breath
than depth, in order to be more easily integrable in the overall methodol-
ogy described in the methodology flowchart of Figure 2.2. Having said
that, nothing prevents modelers and decision makers from adding more
details to these generic models when addressing specific issues of interest.
Before describing the proposed system-based formulation and dis-
cussing how it can be used as a tool to make decisions and propose inter-
ventions in the management and allocation of resources across the nexus
and at the community level (Chapter 7), it is important to understand
different aspects of the value proposition of using systems thinking in
addressing the nexus and the dynamics at play between the nexus and the
systems shown in Figures 4.1a,b. The rest of this chapter presents (i) a
discussion of what systems thinking (or nexus thinking) is, (ii) the unique
habits and skills of systems thinkers, (iii) a review of system dynam-
ics tools, and (iv) a description of the different steps involved in system
dynamics modeling.
14  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

5.3  Systems Thinking

5.3.1 The Systems Thinker

The challenge decision makers face when trying to manage and allocate
water, land, and food resources at the community level is to be able to
come up with “good enough” decisions that account for several factors
such as community dynamics, the multiple links across the nexus, the
nexus’s interactions with the various systems of the landscape, and the
intra- and inter-connectivity among these systems. Trying to account for
all these factors simultaneously belongs to a class of problems that can-
not be easily formulated and modeled, and which are sometimes referred
to as “wicked” problems (Rittel and Weber 1973); the terms “messy”
or “ill-defined” are sometimes used instead. Wicked problems can be
­interpreted as “malignant,” “vicious,” “tricky,” or “aggressive.” Para-
phrasing Rittel and Weber, unlike “tamed” problems, wicked ones have
unique characteristics:

• They are difficult to formulate and can be defined differently


depending on context and scale (physical and temporal).
• There is no finish line, i.e., “wicked problems have no stopping
rules.”
• There are “no true or false answers.”
• They are hard to evaluate, since they create unintended conse-
quences that are unique.
• They are irreversible and attempting to correct some unintended
consequences may create more disruptions.
• There are multiple solutions to the problems that require making
compromises between multiple objectives.
• Caution, flexibility, and adaptability are needed when selecting the
most appropriate solutions if such solutions can even be found.
• They are unique and cannot be regrouped into classes with similar
characteristics and solutions.
• They are meant to be discovered in their unique ways.
• They can be the cause or effect of other wicked problems.

Wicked problems, referred to as messy or ill-defined problems in


the rest of this book, create challenges for those faced with making man-
agement decisions and selecting solutions to address these problems.
In ­general, there is no singular way to handle such problems, although
A System Dynamics Approach to the Nexus   •  15

s­ everal opinions have been proposed in the literature on how to approach


them (Vennix 1999; Metlay and Sarewitz 2012; Stahl and Cimorelli 2012).
Chapter 7 will discuss more specifically the consequences of dealing with
complex and ill-defined problems when deciding on interventions in the
management of resources across the FEW/WELF nexus and at the com-
munity level.
Simply put, messy or ill-defined problems cannot be approached with
traditional linear and reductionist (Cartesian) thinking tools and a com-
mand and control mindset (Holling and Meffe 1996). Systems (or inte-
grated) thinking represents a tangible alternative to reductionist thinking
to handle the complexity and uncertainty inherent in these kinds of prob-
lems. As defined by Richmond (1994), “systems thinking is the art and
science of making reliable inferences about behavior by developing an
increasingly deep understanding of underlying structure.” Another defini-
tion proposed by Sterman (2006) sees systems thinking as a new mindset
with both depth and breadth and “an iterative learning process in which
we replace a reductionist, narrow, short-term, static view of the world with
a holistic, broad, long-term, dynamic view, reinventing our policies and
institutions accordingly.” In the context of the FEW/WELF nexus, nexus
thinking is synonymous to systems thinking.
Even though systems thinking sounds forward-thinking and is often
used in public discourse and debates in science, engineering, politics, and
economics, its value proposition to deliver better and more integrated
solutions and policy decisions to overcome complex problems is still not
mainstream. As remarked by Dent (2001), terms such “systems approach,
systems thinking, and systems view” are not commonly understood.
Another limiting factor is that systems thinking requires the “thinker”
to embrace a new mindset and adopt different habits from those used
in reductionist thinking. These habits are best described by the Waters
Foundation (2016) based in Pittsburgh, Pennsylvania, and are listed in
Table 5.2. Even though these habits were developed in the context of inte-
grating systems thinking in K–12 education, they apply to a wide range of
situations where the thinker is faced with complex issues associated with
messy or ill-structured problems (Benson and Marlin 2017). These habits
can also be understood as thinking strategies (visual, listening and speak-
ing, and kinesthetic) that a decision maker, such as someone who makes
decisions about the management of FEW/WELF resources, might want
to follow to address complex problems. The importance of integrating
the habits in deciding on interventions across the nexus will be discussed
­further in Chapter 7.
16  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Table 5.2.  Fourteen habits of a systems thinker according to the Waters


Foundation

• Seek to understand the big picture


• Observe how elements within systems change over time, generating
patterns and trends
• Recognize that a system’s structure generates its behavior
• Identify the circular nature of complex cause and effect relationships
• Make meaningful connections within and between systems
• Change perspectives to increase understanding
• Surface and test assumptions
• Consider an issue fully and resist the urge to come to a quick
conclusion
• Consider how mental models affect current reality and the future
• Use understanding of system structure to identify possible leverage
actions
• Consider short-term, long-term and unintended consequences of
actions
• Pay attention to accumulations and their rates of change
• Check results and change actions if needed: “successive approximation”
• Recognize the impact of time delays when exploring cause and effect
relationships
Source: http://watersfoundation.org/ (accessed March 13, 2015).

5.3.2  Value Proposition

Once new habits have been developed, systems thinking offers clear
advantages to the thinker compared to deterministic and linear thinking.
According to various authors (e.g., Richmond 1994, 2004; Kim 2000;
Sweeney 2001; Sterman 2006; Stroh 2015), systems thinking helps the
thinker to:

• see the world in wholes instead of snapshots;


• look at the world using different perspectives by stepping out of
self-inflicted boxes and considering mental models other than
deeply engrained ones;
• sense how well parts of systems work together and form structures
and patterns;
• acknowledge relationships between system components from mul-
tiple perspectives, including circular causation, rather than from a
cause–effect linear chain of reaction viewpoint;
A System Dynamics Approach to the Nexus   •  17

• look at events not as separate from each other but instead as parts of
patterns of behavior, which themselves are created by some internal
structure resulting from patterns and modes of thought;
• understand the dynamic, adaptable, unpredictable, and changing
nature of life including the effect of time and delays (information
and materials);
• understand how one small event can influence another ­(positively
or negatively) and the associated consequences of such i­ nteractions;
• identify leverage points in a system, i.e., critical components and/or
links where certain actions yield the most return;
• understand that what is happening depends on where one is in the
system and one’s attitude toward and perception of that system;
• challenge assumptions through mental models;
• become aware of and accept one’s bounded rationality—that is,
a need to make decisions without knowing all the facts due to
­complexity; and
• realize that complexity is not an obstacle but an opportunity to step
out of the boxes that one has created when describing the world.

The value proposition of systems thinking can also be expressed in


terms of the various skills that the systems thinker can use to address
­complex problems. According to Richmond (1997), the skills include:

• dynamic thinking (instead of static equilibrium thinking) that accounts


for how issues and problems change and develop patterns over time;
• system-as-cause thinking (instead of system-as-effect thinking) that
consists of finding causes to a problem or issue that resides within
the system, instead of driven by external forces;
• forest thinking (instead of tree-by-tree thinking) that looks at trends
within a system instead of focusing on specific system parts (which
could result in analysis paralysis);
• operational thinking (instead of factor thinking) that explores how
behavior is generated through the structure of the system and its
components;
• closed-loop thinking (instead of linear thinking) that considers
causal loops within a system and sees a circular instead of linear
one-way causality between cause and effect; and
• quantitative and scientific thinking, where models of problems
­consist of quantifiable (but not always measurable) components
and can be tested to see whether they match what is being observed
in the real world and, if needed, require correction.
18  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

To that list, Richmond (2004) later added the skills of nonlinear


thinking (i.e., no proportional relationship between cause and effect) and
empathic thinking (e.g., sharing, understanding). Additional skills sug-
gested by researchers other than Richmond include being able to:

• balance critical and creative thinking (Vaughan 2013), which can


be understood as balancing left- and right-brain thinking;
• make the distinction between thinking leading to first-order change
(i.e., specific change that does not change the entire system)
and second-order change (i.e., global change with large impact;
­Watzlawick et al. 1974);
• recognize whether systems thinking, or deterministic thinking, is
more appropriate to address a specific issue (the two forms of think-
ing are not mutually exclusive and can, in fact, complement each
other); and
• distinguish between what systems thinking is and is not. Specifically,
systems thinking is not about: (i) analysis that involves breaking
down a problem into bite-size, manageable pieces with the overall
intent of comprehending each part individually; (ii) focusing on
detailed complexity by simulating thousands of variables and com-
plex arrays of details (analysis paralysis); and (iii) making things less
complicated, perfect, and simple in order to avoid the complexity.

Acquiring the habits and developing the skills described above are
essential before any attempt can be made to formulate in a systemic way
the various dynamic hypotheses and associated mental models outlined
at the end of the participatory appraisal discussed in Chapter 4. They are
also needed when developing qualitative and quantitative system dynam-
ics models to explain these hypotheses.

5.4  System Dynamics

5.4.1 Characteristics

System dynamics is a relatively new branch of systems science


­(Wolstenholme 1990; Umpleby and Dent 1999; Schwaninger 2009) that
originated with the work of Dr. Jay Forrester at MIT in the 1950s and
1960s. Within systems science, it is seen as a milestone in the overall
evolution in the application of systems thinking and the development of
tools to address complex issues in a wide range of disciplines such as
engineering, ­business and economics, health, planning, management, and
A System Dynamics Approach to the Nexus   •  19

so on. The books Urban Dynamics (Forrester 1969) and World Dynamics
(Forrester 1971) helped develop what came to be known as the system
dynamics approach to complex problems. The approach was subsequently
used by a variety of researchers including Dr. Donella H. Meadows and
coworkers in several studies showing the impact of population growth,
industrial growth, pollution, and degradation of the environment on world
systems; a review of these studies can be found in Meadows (2008).
A simple way to portray system dynamics is to state that it studies
“how systems change over time” (Ford 2010). The System Dynamics
Society (2016) defines system dynamics as

a computer-aided approach to policy analysis and design. It applies


to dynamic problems arising in complex social, managerial,
­economic, or ecological systems—literally any dynamic systems
characterized by interdependence, mutual interaction, information
feedback, and circular causality.

Landmark books that have promoted the applications of system dynam-


ics in solving complex problems include The Fifth Discipline by Senge
(1994) and Business Dynamics by Sterman (2000). Other interesting texts
on various applications of systems thinking and system dynamics include
those by Ford (2010) on modeling environmental processes; H ­ argrove
(1998) on health sciences; Vennix (1996) and Richardson and Andersen
(2010) on group model building and decision making; Robinson (2001) on
climate sciences; Hannon and Ruth (2001a) on modeling biological sys-
tems; Pidd (2004, Part III) on management science; and Stroh (2015) on
social change. Other interesting books with multidisciplinary applications
of system dynamics include those of Wolstenholme (1990); Hannon and
Ruth (2001b); Bossel (2007a,b,c); Richmond et al. (2010); Pruyt (2013);
and Ghosh (2017). Finally, the reader might be interested in reading the self-
study guide System Dynamics Road Maps, available through the Creative
Learning Exchange website (2015), to learn more about system dynamics
and how it compares with other systems science traditions.
System dynamics has unique characteristics that warrant its use in
modeling the FEW/WELF nexus issues addressed in this book:

• It is a method that can be used to study how systems continuously


change over time due to possible changes in their components,
relationships between components, and the overall direction of the
systems. The method allows for both qualitative and quantitative
modeling.
20  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

• It requires that the problem(s) at stake be clearly formulated,


mapped, and approached iteratively.
• Models of system dynamics are defined by closed boundaries
(causally closed models) where endogenous components and
­
­factors (those originating from within) are assumed to form the
system structure and predominantly dictate the behavior of the sys-
tems. Models are designed to be self-contained in terms of cause
and effect inside their boundaries so that no external (exogeneous)
influence needs to be considered explicitly. In other words, system
dynamics models are designed to contain the components that are
important to explain the dynamic behavior of systems including
their internal rules (Vennix 1996; Sterman 2000; Richardson and
Andersen 2010).
• Nonlinearities in the system are included in the form of first order dif-
ferential equations that are solved numerically (e.g., finite ­difference
method).
• Information feedback mechanisms in the system can be included in
the form of interconnected closed feedback loops and circular cau-
sality, allowing trends in a system to be reinforced and balanced.
This can help in explaining the counter-intuitive behavior of some
systems, indirect effects, and emerging system properties.
• The method emphasizes that the structure of systems (i.e.,
their components, mutual interactions, and interaction with the
­environment) affects their continuous behavior. Combining rein-
forcing and balancing feedback loops, various dynamic patterns
can be simulated and used to model different behavioral patterns
of system changes such as growth, decay, overshoot, oscillations,
­equilibrium, ­randomness, and chaos. As the structure of a system
changes, so is its behavior.
• More emphasis is placed on the structure of a system (its aggregated
nature) than on trying to figure out the details of all its components.

Like all modeling methods, system dynamics is not a panacea for


handling the complexity of every ill-structured and messy problem and
has its own share of limitations. For example, Umpleby and Dent (1999)
note that:

System dynamicists focus on modeling some observed system.


They deal with the issue of knowledge acquisition, but only in
terms of how one understands what is happening in the referent
system. For them, the process of understanding is encompassed
A System Dynamics Approach to the Nexus   •  21

by the methodology of modeling. They do not assume that the


philosophy of knowledge needs to be considered. They are con-
cerned with verifying their models with historical data and helping
decision makers improve their understanding of a referent system.
They are not concerned with cognition as a problem in itself.

As noted by Richardson (1999), system dynamics emphasizes


c­ ontent, objectivity, and patterns of behavior in well-defined boundaries.
Thus, it has strong limitations when applied more specifically to the mod-
eling of social systems and when trying to: (i) account for social systems’
variability, context, values, meaning, and overall subjectivity; (ii) model
singular decisions and events and their impact; and (iii) model the inter-
action of social systems across boundaries. Overcoming such limitations
requires combining system dynamics with other system science methods
(Schwaninger 1997).
Another limitation of system dynamics emphasized by Ahmad and
Simonovic (2004) is that it can only handle systems and processes that
vary in time. To work with processes that vary in time and space requires
combining system dynamics with other methods that emphasize spatial
variations such as geographic information systems (GIS; Ahmad and
Simonovic 2004).

5.4.2 Basic Components

Without going into the details of system dynamics (which can be found
in the various references mentioned at the beginning of this section), it is
important to summarize the basic components that make system dynamics
useful for modeling complex systems.
One of the characteristics of system dynamics modeling mentioned
above is that it captures the feedback mechanisms inherent to complex
systems using two types of cause-and-effect circular causations: reinforc-
ing and balancing feedback loops (Richardson 1999). Reinforcing (R)
feedback loops are used to model self-reinforcing feedback processes.
Reinforcing loops are prevented from growing or declining forever with
balancing (B) loops, which create self-correcting processes that lead to
stability and equilibrium and reaching a goal or objective. In addition to
these two basic components, a delay may be added to model the effect of
time in linking causes and effects or any adjustment processes. In the man-
agement of FEW/WELF resources, delays can be associated with the time
it takes different groups of stakeholders to make decisions ­(information
22  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

delays), the time it takes to implement processes (material delays), or time


needed for various processes (e.g., supply chains, services) to unfold.
As described in the system dynamics literature, various combinations
of reinforcing and balancing loops and delays can be created to model
the behavior patterns of complex systems and unique repetitive patterns
called archetypes. The models (sometimes referred to as c­onceptual
­models) can be represented in an object-oriented form of causal loop
­diagrams or stock-and-flow diagrams. Such diagrams represent useful
tools in (i) depicting how parts of a system interact and create patterns of
behavior, (ii) communicating the dynamic of systems, and (iii) designing
and planning interventions to address issues faced by the system. Note that
causal loop and stock-and-flow diagrams can sometimes be combined in
the form of so-called hybrid diagrams.

5.4.3 Causal Loop Diagrams

Causal loop diagrams (CLDs) show how elements of a feedback mecha-


nism interact in a causal manner. They are influence diagrams consisting
of two basic links:
A −−−+> B
and
A −−−−> B
Both indicate that variable A influences variable B. The first link with
a (+) polarity sign indicates that A and B move in the same directions
(∂B/∂A > 0); i.e., both A and B increase or decrease simultaneously. The
second link with a (–) polarity sign indicates that A and B move in oppo-
site directions (∂B/∂A < 0); i.e., B decreases as A increases or B increases
as A decreases. In some causal loop diagrams in the literature, the (+) and
(–) signs attached to the arrows can be replace by the letters “s” (for same)
and “o” (for opposite), respectively.
A possible delay (due to information or material) between the action
of A and its effect on B is represented by the following link:
A −−║−−> B
In general, causal loops are created by combining the three types of
links mentioned above. All loops with links that have only positive signs
or a combination of positive signs and an even number of negative signs
are reinforcing (R), whereas loops with links that have an odd number of
negative signs are balancing (B). Guidelines for drawing CLDs can be
found in Ghosh (2017).
Figure 5.1 shows an example of a causal loop diagram that was used
to model the dynamics at play in a disadvantaged neighborhood in Costa
A System Dynamics Approach to the Nexus   •  23

Capacity to deal
with critical issues
+
Capacity to invest
R4 in education
+ + +
Effectiveness of Degree of civic
community participation R5
organization + Average real
+ + income
Average level of +
B1 education + +
+
R2 Sense of +
Political community
leverage − Spare time R6 Community
+
productivity
+ Community
+ R3 employment R7 +
Successful
community projects +
Community Health
+
+ infrastructure

Figure 5.1.  Causal loop diagram showing several cause-and-effect loops at play
in a disadvantaged neighborhood in Costa Rica.
Source:  Developed by students in the 2014 Global Sustainability Fellows
program conducted by the Sustainability Laboratory at Earth University in
Costa Rica.

Rica. The model looked at critical issues of importance to the ­community


members such as education, participation, organization, and infrastruc-
ture. Multiple interconnected variables are at play in this diagram and
form six reinforcing loops and one balancing loop. Delays (║) appear in
reinforcing loops R4 and R5. A narrative is attached to each loop. For
instance, reinforcing loop R3 can be expressed as follows:

As a result of poor infrastructure, people currently allocate a lot


of time to everyday activities such as fetching water or bringing
their children to faraway schools. This means that they have less
time available for economic or civic participation. Low civic par-
ticipation leads to less effective community organization, which
in turn makes it more difficult to implement meaningful projects.
This makes it even less likely that they enhance the state of their
infrastructure. (Ben-Eli 2016)

Causal loop diagrams such as Figure 5.1 are useful for mapping,
inferring, and visualizing what contributes to growth, decline, delay, or
stability, and are mostly used at the strategy level. They show, in a con-
densed form, relationships, trends, and connections, as well as causal
feedback mechanisms in a system. Causal loop diagrams are not used to
conduct numerical simulations of systems. They help in laying out the
24  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

d­ ifferent structural components of a system in a conceptual manner and


show how they interact dynamically in a qualitative manner. In short, they
are useful “for communication, not for simulation” (Ford 2010). Despite
that limitation, causal loop diagrams provide a strong value proposition
in the participatory-based management of FEW/WELF resources, as they
help the various stakeholders of Figure 2.1 to develop a common under-
standing of issues they might be facing.

5.4.4 Stock-and-Flow Diagrams

Another way of describing the dynamics of systems is to use stock-and-


flow diagrams that consist of combinations of several building blocks
shown in Figure 5.2. Using the terminology used in the STELLA Architect
software by isee Systems, these building blocks are defined as follows:

• Stocks (or levels) correspond to accumulations of something


­(concrete or abstract) that can be measured at one point in time.
Concrete accumulations may include water behind a dam, a human

Reservoir Conveyor

Cloud Faucet Stock Stock 1

Inflow Outflow

Flow 1 Stock 2 Flow 2 Stock 4

Flow 3 Stock 5
Stock 3
Connector

Converter Converter
(Function)
Figure 5.2.  Basic building blocks of stock-and-flow diagrams: Stocks (reservoir
or conveyor), flow (inflow and outflow) with clouds, converters, and connectors.
A System Dynamics Approach to the Nexus   •  25

population, trees in the forest, and so on. Abstract accumulations


are subjective and may include a level of happiness, well-being,
fear, and so on. A stock can be expressed as a reservoir or level
(Forrester 1969) in and out of which information flows. It can
also be represented as a conveyor to account for the time it takes
for information to pass through the stock. If a stock represents,
for instance, the trees in a forest, the conveyor will allow for the
time it takes tree seedlings to mature before they can be harvested.
Stocks are state variables: they define the current state of a system.
• Flow (inflow, outflow, or bi-flow) is represented in the form
of pipelines (with a faucet controlling the flow rate). It refers to
activities that cause information or materials to change over time
(e.g., number of births per year, inflation rate, flow of a river, cash
flow, carbon emission and sequestration, rate of cutting or planting
trees). Flow—also considered as flux or rate (Forrester 1969)—in
turn, results in changes (dynamic behaviors) in the stock accumu-
lations and in the entire system. Flows are control variables; they
create change in the state of a system.
• Clouds indicate infinite sources or sinks somewhere outside of the
system boundaries.
• Converters are used to convert or transform information from
one stock-and-flow path into information driven by another
­stock-and-flow path, or to feed information into an existing flow.
A converter can also represent a stock if there is no flow in and
out of the stock. They are converting variables. Converters can
be described in a functional form that depends on time or a con-
nected variable. In this case a (~) appears inside the converter and
denotes a graphical function.
• Connectors indicate the transmission or linking of actions and
information (i.e., causal connections) between variables such as
stock-to-flow, flow-to-flow, or between converters. One or ­several
variables can provide input to and have influence on another
­variable through connectors.

Figure 5.3 shows an example of a stock-and-flow diagram illustrat-


ing the dynamics at play between a population and the WELF resources
upon which it depends. These two state variables are assumed to change
(i.e., flow) over time. Births add to the population (a reinforcing loop)
and create a compounding (exponential growth) behavior, ceteris paribus.
Likewise, deaths decrease the population (a balancing loop) and create
a draining (exponential decline) behavior, ceteris paribus. The overall
26  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

being born dying

other Population other


factors
factors
birth
WELF death
resources
birth security
rate br effect effect death
resources resources rate dr
security on
normal desired security on normal
birth rate
birth WELF death rate death
rate resources rate

supplying consuming
WELF WELF initial WELF
resources resources initial resources
population per person
WELF
Resources

WELF
resources initial
WELF Population WELF WELF
gap
resources resources resources
supply consumption rate
rate per person
desired WELF desired
resources WELF
per person resources

Figure 5.3.  Stock-and-flow diagram illustrating the dynamics at play between a


population and the WELF (aggregated) resources upon which it depends.

increase or decrease of the population depends on the birth and death


rates, their relative values, and the current size of the population. In the
population part of the stock-and-flow diagram, the dominant loop dictates
the behavior of the population—that is, growth or decay.
In the WELF resources part of the stock-and-flow diagram, the
resources are assumed to be supplied at a certain rate by some unspec-
ified process. The consumption rate of the WELF resources depends on
the population and the consumption rate per person. The population is
seeking a desired level of available WELF resources per person, which is
assumed to be higher than the initial level. The ratio between the current
amount of WELF resources and the desired amount of WELF resources
(which vary with time due to changes in the population) is considered here
as a potential measure of resources security and ranges between 0 and
1. It is assumed to affect the birth and death rates in some manner along
with other factors. Figure 5.3 makes use the of the “ghost” icon function
for the population variable. It is a useful feature in STELLA Architect for
creating shortcuts to stocks and flows at several locations in the model and
­reducing any model cluttering.
A simple stock-and-flow diagram such as Figure 5.3 helps to visual-
ize how the population interacts with WELF resources. In this aggregated
A System Dynamics Approach to the Nexus   •  27

example, no distinctions were made about the different types of resources.


More complicated versions of the model, defined hereafter as WELF-01,
will be presented in Chapter 6.
In general, stock-and-flow diagrams such as that shown in Figure 5.3
map systems and help visualize flow, accumulation, delay, and dissipa-
tion processes (qualitatively). They also allow for numerical ­(quantitative)
simulations and parametric or sensitivity studies and can therefore be
used at the operation level. Stock-and-flow diagrams are used for “both
communication and simulation” (Ford 2010). As we will see in the next
­chapter, stock-and-flow diagrams are powerful tools that can be used to
model complex interactions in the FEW/WELF nexus and between the
nexus and the systems in the landscape of Figure 4.1a.

5.4.4 Archetypes

As mentioned in Chapter 4, out of all the systems at play in the land-


scape, human/social systems (e.g., populations, households, c­ ommunities,
institutions) are the most complex to analyze. Inferring the structure
­responsible for the observed behavior patterns of human systems as they
interact among themselves and with other systems (natural, infrastructure,
economic) is not easy. However, the complexity of this form of reverse
analysis can be somewhat reduced by considering systems archetypes,
sometime called generic structures (Senge 1994).
In general, archetypes clearly demonstrate that the structure of sys-
tems controls their behavior (Meadows 2008). They can be interpreted as
traps or grooves forcing a system to produce the same answer (intended
or unintended) under the same conditions. They can also be interpreted as
creating attractors of various strengths, i.e., levels where systems tend to
return after undergoing changes. In short, archetypes create “habits” (good
or bad), which in turn define the character of the system and, ultimately,
its destiny. Recognizing archetypes presents an opportunity to identify
places and leverage points in systems where structures and habits can be
changed, which over time may cause the systems to adopt new behavior
patterns (Meadows 2008).
The rationale for using archetypes in analyzing human/social systems
is that despite being seemingly complex, their patterns of behavior are
not always completely random. For instance, recurring (generic) patterns
have been found to manifest themselves in different cultures and contexts.
Archetypes allow for recognizing such patterns. It should be noted that
there are many nongeneric and context-specific patterns of human behav-
ior that are also at play alongside the generic ones mentioned above.
28  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Archetypes have been discussed extensively in the systems ­literature


(Goodman and Kleiner 1994; Senge et al. 1994; Kim 2000; Braun 2002;
Stroh 2015; among others) and have been shown to help tremendously
in the reverse analysis of human systems’ behavior patterns. Key arche-
types often mentioned in the systems literature include balancing process
with delay, limits to growth, shifting the burden, eroding goals, ­escalation,
success to the successful, tragedy of the commons, fixes that fail or back-
fire, growth and underinvestment (fixed or drifting standards), accidental
adversaries, and attractiveness principles (Senge 1994). Each archetype
is further described in Table 5.3 along with examples related to the inter-
action of human systems with the FEW/WELF nexus. Various authors
have proposed specific causal loop diagrams to represent the dynamic
behind each archetype of Table 5.3 and for different domains of applica-
tion (Senge et al. 1994; Kim 2000; Kim and Anderson 2007; Stroh 2015).
For example, Mirchi et al. (2012) propose a series of causal loop diagrams
illustrating archetypes that could be at play in the management of water
resources. On the other hand, Zhang et al. (2014) consider the interactions
among the sustainable development goals (SDGs) and identify several
archetypes at play in that interaction.
Although archetypes have traditionally been used to qualitatively
model various forms of dynamics at play among human systems or
between human systems and other systems, they can also be applied
to model various forms of dynamics at play between the sectors of the
FEW/WELF nexus. This use of archetypes was explored by Sohofi et al.
(2016), who showed how the limits to success [growth] and accidental
adversaries archetypes of Table 5.3 are at play in the circular causalities
and interactions across sectors of the FEW nexus. An illustrative ­example
proposed by these authors is shown in Figure 5.4. In this figure, the ­limits
to s­uccess archetype appears in the loop combinations (R1, B1) for
energy ­production, (R2, B2) for water production, and (R3, B3, B4) for
agriculture production. The accidental adversaries archetype in Figure
5.4 deals with competition and associated trade-offs related to simultane-
ously ­trying to increase water, energy, and agriculture production when
these three sectors depend on each other and the success of one of them
negatively affects the success of the other two. Figure 5.4 also includes
the effect of climate change on energy, water, and agriculture production
potential and land availability.
Another example of system archetypes being applied at the nexus
level was proposed by Newell et al. (2011). The example uses the tragedy
of the commons archetype (Hardin 1968; 1994) to capture the dynamic of
competition for water resources between the electricity and food sectors.
Table 5.3.  System archetypes and related community development and FEW/WELF nexus examples
Archetypes Description and dynamic Examples
(from Braun 2002)
Limits to “A reinforcing process of accelerating growth (or - Agricultural economy is limited by unavailability of
growth expansion) will encounter a balancing process as water, energy, infrastructure
the limit of that system is approached … - Increase in poor population with availability of water,
­Continuing efforts will produce diminishing energy, and food requires more services
returns as one approaches the limits” - Community services limited by lack of trained personnel
and lack of technology spare parts
- Water demand creates more energy demand (R) which is
limited by energy and water reserves (B).
- See Figure 5.4
Shifting the “A problem symptom can be resolved either by - Relying on a charity model of development instead of
burden using a symptomatic solution or applying a fun- capacity development
damental solution ... Once a symptomatic solution - Dependency on NGOs, governments and outsiders to
is used, it alleviates the problem symptom and address local FEW/WELF issues
reduces ­pressure to implement a fundamental - Quick fixes made to water and energy infrastructure
solution, a side effect that undermines fundamen- without considering long-term performance
tal solutions” - Solving one immediate problem without considering
unintended consequences
(Continued)
A System Dynamics Approach to the Nexus   •  29
Table 5.3. (Continued)
Archetypes Description and dynamic Examples
(from Braun 2002)
Eroding (or “A gap between a goal and an actual condition - Community identifies high goals for itself. Over time,
drifting) goals can be resolved in two ways: by taking correc- the goals cannot be met because: (i) goals were too
tive action to achieve the goal, or by lowering ­complex with and did not match the capacity of the
the goal … When there is a gap between a goal ­community; (ii) the interest of the community erodes
and a condition, the goal is lowered to close the away with time
gap. Over time, lowering the goal will deteriorate - External assistance and services to the community
­performance” decrease over time due to a decreasing commitment
from government and community leaders
Escalation “One party’s actions are perceived by another party - Two ethnic groups live side by side and compete for
to be a threat, and the second party responds in a resources (water, energy, land, and food)
similar manner, further increasing the threat ... - Two local institutions compete for a same project and/or
The two balancing loops create a reinforcing limiting funding
­figure-8 effect, resulting in threatening actions by
both parties that grow exponentially over time”
Success to the “If one person or group (A) is given more resources - Due to corruption one group has access to more water,
successful than another equally capable group (B), A has energy, land and food resources
a higher likelihood of succeeding … A’s initial - Men have access to more resources than women
­success justifies devoting more resources to A,
30  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

further widening the performance gap between the


two groups over time”
Tragedy of the “If the total usage of a common resource becomes - Overuse of shared natural resources, overgrazing, ­fishing,
commons too great for the system to support, the commons fertilizers, and deforestation
will become overloaded or depleted and everyone - There is no agreement as to how to share the resources
will experience diminished benefits” - See Figure 5.5
Fixes that fail “A quick-fix solution can have unintended - A water pumping and distribution system is installed
or backfire ­consequences that exacerbate the problem … without consideration for long term performance.
The problem symptom will diminish for a short ­Systems break down very quickly
while and then return to its previous level, or - Political promises fail to materialize
become even worse over time” - Lack of accountability in development planning and execution
- Water shortage requires using more energy which creates
unintended pollution that increases the water shortage
Growth and “If a system is stretched beyond its limit, it will - Community development is going well but c­ ommunity
underinvest- compensate by lowering performance standards, members and/or outsiders underestimate what’s
ment which reduces the perceived need for investment. ­necessary to carry out the development forward
It also leads to lower performance, which further - Community does not invest in its own resources and has
justifies underinvestment over time” not enough capacity
Accidental “When teams or parties in a working relationship - Conflict emerges during development projects between
Adversaries misinterpret the actions of each other because of community members and NGO and/or government
misunderstandings, unrealistic expectations or because of misunderstanding, lack of initial shared
performance problems, suspicion and mistrust vision, and lack of capacity in adapting to change
erode the relationship. If mental models fueling the - Conflict resolution was not included in project planning
deteriorating relationship are not challenged, all
parties may lose the benefits of their synergy”
A System Dynamics Approach to the Nexus   •  31
Effect of climate change on Effect climate change Effect of climate change on Effect of climate change
energy production potential on land availability agriculuture production on water production
potential potential
<Water price>
<Water
<Required Land
<Water production cost>
land> availability Water
demand> <Total energy
investment
<Total water produced>
Water price produced>
Energy Water
Energy availability production R2
Energy Agriculture
investment potential Water revenue
production cost Water production
availability potential
Energy B2
Energy revenue R1 Energy Agriculture Total agriculture demand Total water
production investment R3 production produced
Energy required
Cost of potential Water demand for water
capital in Total energy B3 Agriculture production
energy sector produced B1 revenue B4
Agriculture Energy demand
Total capital in Water required for water demand Energy
in other sectors
energy sector energy production price
<Energy price>
Agriculture Water
Water demand <Energy
Technology mix energy demand production cost
Technology
in other sectors in agriculture production cost>
<Energy sector
mix in energy
investment>
sector Technology mix in <Energy
<Agriculture Required land water sector availability>
investment>
<Water Total capital in Cost of capital in
Required land for investment> water sector
other purposes water sector

Figure 5.4.  Causal loop diagram illustrating the limits to success dynamic across the water, energy,
and agriculture sectors of the nexus.
Source:  Sohofi et al. (2016), reproduced with permission from Springer. Redrawn using the Vensim
32  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

software.
A System Dynamics Approach to the Nexus   •  33

+
Amount of Viability of
electricity R electrictiy
generated generation industry
+ + +
Amount of
water used + B
by electricity
generation industry Total Water Water
water available reserves
used – per user + limit
Amount of
water used +
by food, feed, B
and fibre industries
+ + +
Amount of Viability of
food, feed, R food, feed,
and fibre and fibre industries
produced
+

Figure 5.5.  Tragedy of the commons causal loop diagram showing competition
for water resources between the electricity and food sectors.
Source:  Adapted from Newell et al. (2011), with permission from the authors.
The delay and “water reserves limit” were added to the original figure.

This dynamic, shown in Figure 5.5, describes how water consumption by


the energy and food sectors has the potential to create diminishing returns
for both sectors if not jointly managed. Figure 5.5 assumes that the total
amount of water available is limited.
There are several other examples of causal loop diagrams
­representing archetypes at play in various sectors of the FEW/WELF
nexus. For instance, Turner et al. (2016) show how the shifting the
burden and fixes that backfire archetypes can be applied to model
the importance of using long-term rather than short-term solutions in
food and natural resources management. Finally, Zhuang (2014) offers
examples of archetypes at play in the interaction between water and
energy systems.

5.5  System Dynamics Modeling

5.5.1 Modeling Steps

In general, system dynamics models are built in steps of increasing com-


plexity. Sterman (2000) recommends that models are built following a
road map consisting of five interactive activities: problem articulation
34  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

and boundary selection, dynamic hypothesis formulation, simulation and


model formulation, testing of the model, and policy design and evaluation.
Ford (2010) suggests a similar but more detailed road map consisting of
eight steps combined into six key activities: (i) problem f­amiliarization
(step 1); (ii) problem definition (step 2); (iii) model formulation by con-
structing stock-and-flow diagrams (step 3) and causal loop diagrams
(step 4); (iv) parameter estimation (step 5); (v) simulation to explain the
problem being addressed (step 6); and (vi) simulation analysis consisting
of sensitivity analysis (step 7) and policy analysis (step 8). These steps are
illustrated in Figure 5.6.
Both Ford and Sterman emphasize the cyclical and iterative process
between the different components of system dynamics modeling as shown
in Figure 5.6. The first two steps suggested by Sterman and Ford are about
defining the problems or issues being modeled, developing an initial men-
tal model, and proposing dynamic hypotheses. In the management and
allocation of water, energy, land, and food resources at the community

Step 1
Get acquainted
Steps 7 & 8 (problem familiarization)
Sensitivity analysis &
policy analysis

Step 2
Be specific about the
Step 6 dynamic problem
Simulation to (problem definition)
explain the problem

Steps 3 & 4
Step 5 Stock-and-flow diagram
Estimate and causal loop diagram
the
parameters (model formulation)

Quantitative system dynamics Qualitative system dynamics

Figure 5.6.  Different components of system dynamics modeling.


Source: Ford (2010), with permission from Island Press.
Note: The components are divided into those involved in qualitative modeling
(right-hand side) and those involved in quantitative modeling (left-hand side).
A System Dynamics Approach to the Nexus   •  35

level, problem familiarization, definition, articulation and the formulation


of mental models, and dynamic hypotheses for each problem and issue
were already discussed at the end of Chapter 4. They are the deliverables
and the outcome of the participatory appraisal, which must be carried out
using a systemic perspective.
Figure 5.6 also shows that the system dynamics modeling process is
divided into qualitative and quantitative modeling. The first two activities
suggested by Sterman and the first three of Ford’s plan can be interpreted
as the qualitative and conceptual components of system dynamics mod-
eling. The other activities emphasize the quantitative dimension of that
modeling. Whether qualitative or quantitative modeling is used depends
largely on what the system analysis is used for, the availability of data and
information about the system components, and the participating audience
(Wolstenholme 1999).

5.5.2 Setting Expectations

Before addressing in more detail the qualitative and quantitative aspects


of system dynamics modeling, a few remarks need to be made about what
to expect from system dynamics models. First, one should not expect the
models to be validated, since they are virtual representations of reality.
They are not reality itself, and there is no such thing as a “good model.”
However, there is such a thing as a “useful and sound model” in the sense
that the model is useful, sound, and consistent in simulating reality and,
to a certain extent, observed behavior patterns (Forrester 1961; Sterman
2000). As noted by Barlas (1996), the validity of a model can be described
in its “usefulness with respect to some purpose,” keeping in mind that “the
usefulness of the purpose itself” needs to be formulated, as well. Since
system behavior is dictated by structure, validating a model is also about its
usefulness with respect to some structure and its environment (i.e., the con-
text and scale discussed in Chapter 3). In the management of water, energy,
land, and food resources, the purpose would be to address any real situa-
tion that the community is facing or expects to face when managing these
resources, such as managing the availability and accessibility of drinking
and irrigation water and the consistent functioning of water infrastructure
over a period of one or several years. The usefulness of that purpose is
obvious regarding the community’s survival, resilience, well-being, and
economic development at a certain scale and over some time.
Second, as one goes through the different steps in the system dynam-
ics framework, confidence in the model’s usefulness is likely to grow
­(Barlas 1996). Confidence starts with having a clear understanding of what
36  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

is expected in activities of model building, mentioned above. The books


of Sterman (2000); Richmond (2004); and Fisher (2011) provide detailed
recommendations for model building and “good modeling ­ practices”
using multiple practical examples.
Finally, building confidence in model development is as learning
­process (Lane 1999) that requires continuous feedback between the model
itself and the real world that the model is meant to represent (Sterman
2000). Simply put, models evolve and need to be updated. Starting with an
initial model, system dynamics modeling “creates new opportunities for
learning that improve the accuracy, coherence, and complexity of mental
models” (Doyle et al. 2002). This feedback mechanism can be understood
as an ongoing “reality check” to prevent the model from going astray.
Confidence in the model can also be gained by conducting several tests
that have been proposed by system dynamicists. These tests are listed in
Table 5.4. A more detailed review of these tests can be found in Forrester
and Senge (1980); Barlas (1996); Sterman (2000); and Mirchi et al. (2012)
offer examples of tests that can be used to verify system dynamics models
for water resources management.

5.5.3  Qualitative System Dynamics Modeling

The qualitative system dynamics modeling of key issues identified in the


participatory appraisal of the landscape in which the nexus unfolds and
community development takes place calls for several key questions to be
answered by modelers as models are being formulated. Table 5.5 gives a
possible list of questions that explore the why, what, who, where, when,
and how of the modeling process.
An important topic to consider when creating dynamics models is
what constitutes endogenous components and factors (i.e., originating
from within) that need to be included in the dynamic model and exogenous
components (i.e., originating from without) that can be set aside in order
to model and understand the key issues of interest. Some components may
need to be excluded, at least during the first step in the ­iterative modeling
process of Figure 5.6. They may be included later or on an as-needed basis
(Ford 2010). Endogenous components might be related to the various sec-
tors of the nexus and how they interact with the social, natural, infrastruc-
ture, and economic systems in the landscape of Figure 4.1a. Exogenous
components might be related to the influence of other communities and pol-
icy decisions dictating how the interactions take place. In system dynam-
ics, exogenous components are not accounted for explicitly. If critical, they
Table 5.4.  Tests used to gain confidence in system dynamics models. This table was constructed based on the work of Sterman
(2000, pp. 859–61) and Forrester and Senge (1980)
Tests Purpose Rationale
Boundary adequacy Appropriateness of selected boundary to address Boundaries control how systems are framed,
the problem of interest. Does the outcome of the what is in (endogenous) the model and what is
model change with the selected boundaries? out (exogenous)

Structure assessment Assessment of how well the structure of the system Inconsistencies create confusion. Violation of
has been formulated in terms of possible incon- conservation of mass and energy will lead to
sistencies, conservation laws, unit consistencies, erroneous answers. More details in the model
appropriate level of aggregation to capture the may or may not be necessary to understand
dynamic of the problem of interest the problem of interest
Dimensional consistency Consistency of units between flow and stock Garbage in, garbage out
Parameter assessment Making sure that model parameters are relevant Reality check, identify what is qualitative and
and represent something real in the problem being quantitative
addressed. Objective or subjective assessment of
parameters and their variation is necessary
Extreme conditions Test the robustness of the model and see if it Identify limitations of the model. How far can
­provides logical responses when parameters are we push the model?
given extreme values
(Continued )
A System Dynamics Approach to the Nexus   •  37
Table 5.4. (Continued )
Tests Purpose Rationale
Behavior reproduction Compare model prediction (qualitative and quan- If the prediction is not adequate, the model
titative) to reference behavior of the problem of needs to be changed and parameters reevalu-
interest. Forrester and Senge (1980) introduce ated. It makes no sense to use the model for
four additional sub-tests. Helps build confidence predicting future behavior
in using model for predicting future behavior
Behavior analysis and Identify unexpected forms of behavior and The model may show expected or unexpected
surprise behavior ­anomalies by changing model structure and forms of behavior, some more realistic
assumptions than others
Sensitivity analysis Test the model to see how the system responds to The world is not deterministic and all parame-
the variation of one or several critical parameters. ters entering into a system are not fixed
The variation can be described statistically
Policy analysis Help identify new policies that will help the Old policies are responsible for the current
system perform better in the future. Forrester and behavior. New policies must be introduced for
Senge (1980) introduce the policy and extreme change to take place
policy tests
38  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS
A System Dynamics Approach to the Nexus   •  39

Table 5.5.  Questions that may arise when creating a dynamic hypothesis
and formulating a system dynamics model
Why Why is a dynamic model constructed?
What What problem and behavior (reference mode) is being modeled
and over what time frame and spatial scale?
What methods other than (or complimentary to) dynamic
modeling can be used to model the problem?
What would happen if the problem were or were not
addressed?
What solutions have been attempted in the past to address the
problem and what were their outcomes?
What are the components of the problem being addressed and
their connections?
What range of responses can be expected from the model?
How How will the model complement the traditional steps of project
management?
How have the components of the model interacted in the past?
How will community members be involved in building,
reviewing, and updating the model?
How will the model recommendations be presented to the
community members?
Who Who is participating in developing the model (insiders and
outsiders)?
When When should the model be integrated into project management?
When should the model be started, evaluated, modified, and
updated?
Where Where should model development and community interaction
take place (office, community)?

can be made endogenous to the model by extending its boundary. Another


option is to account for them implicitly by considering user-defined graph-
ical functions to model their influence on endogenous parameters.
Another related concern in building a system dynamics model of
the landscape and its components is to decide on its spatial and temporal
boundaries. Boundaries are critical to system dynamics modeling, as they
determine how systems are framed (Richmond 2004). They differentiate
between “what is in and what is out, what is deemed relevant and irrelevant,
what is important and unimportant, what is worthwhile and what is not,
who benefits and who is disadvantaged” (Williams 2008). As discussed in
Chapter 3, landscape boundaries are likely to coincide with geographic/
natural physical boundaries. For instance, the boundary of a watershed
40  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

(catchment, drainage basin) may be appropriate in identifying the dynamic


of a community and how its members use water, energy, food, and land
if the community operates in an isolated manner and within the confines
of the watershed. In other cases, administrative boundaries (villages,
­counties, wards, etc.) might be more relevant. Finally, temporal boundaries
may include seasonal, monthly, or yearly nexus-related activities. In cases
where boundaries cannot easily be defined, artificial boundaries may need
to be selected. This might be the case when the boundaries of the water,
energy, food, and land components do not coincide; administrative bound-
aries are more realistic; or transboundary issues are at play in the landscape.
It should be remembered that the conclusions reached in modeling
any system depend greatly on the context and the selected boundaries. As
the complexity of a model grows, boundaries may need to be expanded
to see if the conclusions about the system’s behavior remain the same or
change. This is referred to as the boundary adequacy test by Forrester and
Senge (1980) and Sterman (2000; Table 5.4).
The last step in the qualitative part of system dynamics modeling is to
formulate the model in terms of stock-and-flow or causal-loop diagrams
(steps 3 and 4 in Figure 5.6). The basic building blocks (i.e., stock, flow,
conveyor) of the model must be selected first. Selecting these components
is based on: (i) what is accumulating (reservoirs); (ii) what processes are
flowing in and out of the reservoirs (flow); and (iii) what processes control
the flow (reinforcing or balancing). Both stocks and flows can be physical
or nonphysical in nature. Following the selection of the model’s building
blocks, decisions need to be made on how these blocks are connected,
paired, and dependent on each other through direct (linear) causality,
reciprocal causality, closed-loop causality, feedback mechanisms (rein-
forcing and balancing), and possible delay.
The challenge in building stock-and-flow diagrams to model the
dynamics at play in the nexus and between the nexus and the different
systems of the landscape is to select building blocks and connections that:
(i) best simulate the mental models and hypotheses associated with each
issue outlined at the end of the appraisal phase described in Chapter 4,
and (ii) are consistent with the various questions raised in Table 5.5. In
constructing stock-and-flow models, rules must be followed (Richmond
2004). First, care must be taken to respect unit consistency—that is, each
flow in or out of a reservoir must use the same units of measure as the
reservoir itself, except that flow is measured “per unit of time.” Likewise,
all stocks attached to a given flow path must use the same unit of measure.
Second, conservation laws of mass and energy must be respected.
Causal loop diagrams (CLDs) can provide an alternative way to
conceptualize and comprehend the interaction of components at play in
A System Dynamics Approach to the Nexus   •  41

complex dynamic systems (Wolstenholme 1990). They may offer addi-


tional insights about the nature and intensity of feedback mechanisms and
how reinforcing and balancing processes unfold in such systems. They
can be built before, during, or after constructing stock-and-flow diagrams,
depending on the problem being analyzed and the modeler’s preferences.
At the end of the qualitative part of system dynamics modeling, causal
loop and stock-and-flow diagrams have been constructed, key model
­variables and their relationships have been identified (at least prelimi-
narily), and boundaries and initial conditions have been selected. Several
tests must then be conducted to build confidence in the model. According
to Forrester and Senge (1980) and Sterman (2000), these tests include
(i) evaluating the adequacy of the selected boundaries, (ii) assessing the
model structure for inconsistencies with the real problem structure, and
(iii) ensuring a proper level of aggregation. Both (ii) and (iii) are part of
the structure assessment test listed in Table 5.4.

5.5.4  Quantitative System Dynamics Modeling

In system dynamics, the jump from qualitative to quantitative modeling can


be a difficult one, as it implies that the variables included in the model (Step 5
in Figure 5.6; Step 3 in Sterman’s model) can be ­measured ­objectively,
i.e., that they are backed up with “hard” data, per the t­erminology used by
Luna-Reyes and Andersen (2003). As discussed in Chapters 3 and 4, such
data may only be available for certain measures of water, energy, land,
and food security; some components of the social, natural, infrastructure,
and economic systems; and certain links between technical measures and
nontechnical (social, economic) ones. There is a consensus in the literature
pertaining to the system dynamics modeling of the nexus (see Section 5.2)
that not enough specific data are available to model the complexity of the
issues being addressed. That limitation often forces modelers to use aggre-
gated data that may not have any meaning in the real world.
In addition, in the landscape and its components, most of the ­variables
are qualitative (i.e., backed up with soft data), which makes quantitative
system dynamics modeling more difficult to interpret. As we will see in
the system dynamics models of Chapter 6, it is possible to overcome this
­limitation by using semi-quantitative measures and using ­multiplying
­factors and graphical functions (built-in and built-up) to account for
­subjective judgments and the effects of intangible issues. Another
approach is to use proxies or constructs that relate a qualitative variable to
other ­variables for which quantitative data are available (Luna-Reyes and
Andersen 2003). In any case, all data used in the quantitative ­modeling
of the landscape or some of its components must be checked using the
42  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

dimensional ­consistency and parameter assessment tests suggested by


Sterman (2000) and mentioned in Table 5.4.
The value proposition of using quantitative models instead of just
qualitative models has been a topic of discussion in the system dynamics
literature (Vennix 1996; Mirchi et al. 2012). On one hand, proponents of
qualitative models argue that although quantitative models can provide
a complete understanding of problems being addressed, appropriate and
accurate hard data are often difficult to include in quantitative models and
as a result, quantitative models may not be reliable due to data ­uncertainty.
On the other hand, proponents of using qualitative and quantitative
­models together accept that despite such uncertainty, quantitative models
still add value in decision making (Homer and Oliva 2001). Indeed, for
some, quantitative modeling is essential for a complete understanding of
­complex problems (Dhawan et al. 2011).
Yet what represents “complete understanding” is subject to interpre-
tation, especially when applied to messy and ill-defined problems such as
those involved in the nexus (Vennix 1996). Since, as discussed in this book,
we are more interested in a process of satisficing than optimizing when
dealing with complex systems such as communities, we can argue that
qualitative system dynamics modeling provides a real value ­proposition
for (i) creating a learning environment, (ii) understanding communities
better, and (iii) making more intelligent project management decisions.
Of course, the fact that there are positive aspects to qualitative modeling
should not preclude the use of quantitative modeling when hard quanti-
tative data are available for some parts of the system. Using quantitative
modeling should lead to a fuller understanding of the problem (Vennix
1996). In community development projects, the quantitative data collected
during community appraisal can be included in the quantitative modeling
of tangible issues faced by the community such as water, energy, land, and
food management and job creation. However, it is likely that intangible
social and cultural issues can only be qualitatively modeled.
If a quantitative model of the landscape (or parts of it) and its compo-
nents can be developed, the next step is to try to simulate and reproduce
as accurately as possible the current (or even past) landscape, the mental
models, and the reference modes of behavior, i.e., the behavior patterns and
trends observed across the FEW/WELF nexus and at the community level
(Step 6 in Figure 5.6; Step 4 in Sterman’s model). This calibration process
is referred to as the behavior reproduction test mentioned in Table  5.4;
it is an evaluation (sometimes called validation) of the modeling p­ rocess
and serves to build confidence. Based on the outcome of that evaluation,
the model can be modified, restructured, and ultimately improved. More
A System Dynamics Approach to the Nexus   •  43

c­ omplex data and information can then be included once further confi-
dence has been built around the model. Once confidence is built, ­multiple
scenario simulations are considered. The system’s overall behavior is
studied under various assumptions (i.e., “what if ” or “what happens if ”
simulations), possible leverage points are identified, and possible forms of
unintended consequences are explored. In Table 5.4, this is referred to as
the behavior analysis and surprise behavior test (Sterman 2000).
Ultimately, a new and better way of understanding the landscape sys-
tem (or parts of it) and its components (i.e., building a new mental model)
emerges, and a decision can be made about whether the model is “robust”
enough to generate meaningful patterns (Ford 2010). The robustness of a
model can be assessed using the extreme conditions test, as suggested by
Sterman (2000), where the model’s behavior is tested under conditions far
from equilibrium. This helps define the application range of the model, its
limitations, and any surprise behavior it may engender. In the landscape,
the quantitative modeling of water, energy, food, and land issues and other
critical issues at the community level can be tested under the extreme con-
ditions that the community may have faced in the past. The response of the
community to such challenges can help predict how the community may
respond to more extreme conditions in the future, such as those associated
with natural hazards. This, in turn, can help in identifying ahead of time the
vulnerability of the communities to such events and planning accordingly.
Once in place, a robust model of the landscape and its components
becomes a very useful tool to explore scenarios that are created by
changing the values of one or several parameters. As noted by Houghton
et al. (2014), the model can be tested for variables that follow certain
­probabilistic distributions (e.g., Monte Carlo analysis). Sensitivity analy-
sis tests (Sterman 2000) can be carried out to analyze how a system such
as the landscape responds to variations of one or several critical param-
eters. Such analyses may, in turn, contribute to developing new WELF-­
related policies (Step 8 in Figure 5.6), i.e., “chosen course[s] of action
significantly affecting a large number of people” (Simonovic and Fahmy
1999). As suggested by Forrester and Senge (1980), extreme conditions
and sensitivity analysis tests can help to identify new future policies that
emphasize synergies and reduce trade-offs among sectors of the nexus.
Examples may include deciding on using water for a specific purpose
such as industry instead of farming; addressing the competition for water
resources between the energy (electricity production) and food sectors;
prioritizing big water projects over small ones; and so on.
Finally, a robust systemic model of the landscape can help to select
modes of intervention among alternatives, analyze how the landscape and
44  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

its components react to the chosen modes of intervention, and identify


possible unintended consequences that may emerge if these interventions
are adopted. As discussed in Chapter 7, system dynamics modeling of the
landscape is a valuable approach when deciding whether one mode of inter-
vention is better suited than another to address issues at the level of the
landscape, the community, and/or the nexus. An excellent example illus-
trating the steps mentioned above can be found in Adeniran and Bamiro
(2015), who used system dynamics to develop scenarios to aid short- and
long-term strategic planning of a municipal water system in Nigeria. In this
example, eight different scenarios were considered. Similarly, Simonovic
and Rajasekaram (2004) used the CanadaWater system ­dynamics model
to explore 12 scenarios to “investigate policy options in the areas of fresh-
water availability, wastewater treatment, economic growth, population
growth, energy production and food production” in Canada and at the coun-
try level. Other detailed case studies worth exploring include those reported
by Atherton (2013) and those mentioned in Section 5.2.

5.6  Combining Methods of Analysis

As will be seen in the next chapter, system dynamics is a powerful method


that can capture multiple dynamics at play across the nexus, includ-
ing the interaction among the nexus components and their interactions
with the social, natural, infrastructure, and economic systems shown in
­Figure 4.1a,b. But it is not only capable of capturing these ­dynamics, as it
can indeed be complemented with other methods that are also s­ ystemic in
nature. Graphical representations such as mind maps and problem trees (see
Chapter 4), cross-impact analysis, and network ­analysis can complement
system dynamics. Other methods (not used in this book) could include
probabilistic cross-impact and Monte Carlo analysis. ­Several authors
have also explored the combination of system dynamics with artificial
neural networks (Abdelbari and Shafi 2017; Ghosh 2017), agent-based
modeling (Borshchev and Filippov 2004; Nava Guerrero et al. 2016); and
geographic information systems (Ahmad and Simonovic 2004). As noted
by Ritchie-Dunham (1997), combining methods of analysis has a strong
value proposition for better capturing the structure of complex problems
being analyzed.

5.6.1 Cross-Impact Analysis

As discussed in Section 3.6, cross-impact analysis can be used to describe


the influence/impact and dependence/sensitivity of several interacting
A System Dynamics Approach to the Nexus   •  45

variables. Among the various cross-impact analysis formulations pro-


posed in the literature, the one retained here is that of Godet (2001) and
Arcade et al. (2014). The formulation was used in Section 3.6 to analyze
semi-quantitatively the interactions of the water, energy, land, and food
sectors of the nexus. The same methodology was used in Chapter 4 to map
the double causality that exists between the social, natural, infrastructure,
and economic systems at play at the community level and possible mutual
interactions among subsystems in each system.
Being able to identify which variables influence others and which are
dependent on others is critical in developing system dynamics ­models.
As suggested by Arcade et al. (2014), this can be accomplished using
graphical representations such as those of Figures 3.3 and 3.4, which plot
the influence vs. dependence of each variable being considered in the
cross-impact analysis.

5.6.2 Network Analysis

Network analysis provides a means of mapping the i­nterconnectivity


that exists between the different systems of Figures 4.1a,b or the
­intra-­connectivity of the subsystems that comprise these systems. Network
analysis is, however, more than just a graphical representation of connectiv-
ity as described in the book Networks: An Introduction by Newman (2010).
When used for mapping social systems, network analysis is referred
to as social network analysis (SNA). It can be used to show, for instance,
how members of a group and/or several organizations interact and con-
nect in a network (NRC 2009; Borgatti et al. 2009) and how the compo-
nents of a social network communicate, express their creativity, and reach
consensus (Perkin and Court 2005). In general, SNA network diagrams
are graphs that consist of multiple nodes (vertices) connected by links
(edges or ties) that define the social fabric. The nodes represent social
agents or actors (individuals, groups, organizations, or partners) and their
shapes, sizes, or colors can represent the importance or special attribute of
an agent in the network. The links represent how the agents or actors are
interconnected when addressing a specific issue at a given time.
Different graphical tools have been proposed to show the level of
influence of actors/agents on others. For instance, graphs that show edges
pointing inward toward or outward from each node are called directed
graphs. Symbols of different sizes can be used to indicate the strength
(attribute) of each node. Likewise, the same can be done with the links
by changing their thickness or color. Networks that have this additional
information are called valued networks (Borgatti et al. 2013) or weighted
networks (Newman 2010).
46  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

SNA maps can also be analyzed using mathematical tools. They can
be used to identify the strengths, weaknesses, and patterns of interaction
among system components, identify potential system attractors, make pre-
dictions, assess the network’s resilience, map assets and v­ ulnerabilities,
conduct simulations, and plan interventions that leverage or strengthen
existing network connections. Borgatti et al. (2009) review some of
the measures and properties than can be inferred from SNA maps at
the ­network level (cohesion, shape), the node level (different forms of
­centrality, betweenness), and the link level (cohesion and equivalence).
In turn, these measures and properties can help to identify the “underly-
ing ­processes that account for observed relationships among variables,”
which, ­according to Borgatti et al. (2009) can be processes of transmis-
sion, adaptation, binding, or exclusion.
SNA is a powerful method that can help to make decisions in the par-
ticipatory community-based management of resources across the FEW/
WELF nexus where there may be issues about (i) how information flows
between populations, households, and institutions at the community level
and (ii) how decisions about water, energy, land, and food resource allo-
cation are made in the community. SNA maps can help to show clearly
whether information flow, who controls what, who are the critical actors
and clusters of decision makers, where road blocks occur, and who is mar-
ginalized. As an example, Bodin et al. (2006) show how these properties
and measures of SNA maps can help users to better understand the role
that different components of social networks play in the adaptive man-
agement of natural resources around critical social features such as social
memory, heterogeneity, redundancy, learning, adaptive capacity, and trust.
Network analysis is not only used in mapping social systems ­(Newman
2010); nodes and links in network graphs may also represent connectivity
between nonsocial agents. Examples include mapping the different com-
ponents of the infrastructure systems discussed in Section 4.7 that contrib-
ute to providing a certain type of service under a certain context and at a
certain scale. In this case, the network maps may help to identify where
to intervene in the infrastructure to make a specific service more effi-
cient or accessible to a wider range of customers. A water distribution, an
­electricity grid (or microgrid), and transportation networks come to mind.
Network analysis can also be used to assess the critical infrastructure nec-
essary for a community to face adverse events and ­natural ­hazards, thus
increasing its overall resilience (Toole and McCown 2008; NRC 2009).
Finally, network analysis can be used to map the multiple ­economic pro-
cesses of production, distribution, and consumption of goods and services
at different scales.
A System Dynamics Approach to the Nexus   •  47

As reviewed by Freeman (2000) and Borgatti et al. (2013), the visual-


ization of networks has evolved over time from early hand-drawn images
in the 1930s to today’s computer-generated diagrams. Various tools and
software are available today that not only map the different components
of a network and their interactions, but also help identify the strength of
networks (NRC 2009). Borgatti et al. (2013) show several examples of
how SNA can be applied using the software UCINET, which is available
through Analytic Technologies based in Lexington, Kentucky (Borgatti
et al. 2002).
Examples of network analysis application in development and aid
studies abound in the literature, especially for policy decision making.
Among many case studies of the application of network analysis, some
well-documented ones can be found in papers by Kincaid (2000) on fam-
ily planning in Bangladesh; Moore et al. (2003) for floods in Mozambique;
Schiffer and Waale (2008) for the management of water basin resources
in Ghana; Dale (2011) for network identification in Timor-Leste; and
Blanchet and James (2012) for policy and health intervention in Ghana.
Zimmerman et al. (2016) proposed a network analysis of the FEW nexus
where the interconnections between the food, energy, and food sectors are
accounted for using a Leontief economic input–output model ­(Leontief
1986). In that analysis, the usage rate of one sector of the nexus was
assumed to be linearly related to how the usage rates of the other sectors
are related to the sector of interest. An example would be how much water
is needed per day knowing that so much water is needed to produce a ton
of food or a kilowatt hour (kWh) of energy.
Although network analysis focuses on the components of networks
and their patterns of interactions (i.e., the network structure), it pays less
attention on the nature and dynamic of those interactions, which can be
handled better both qualitatively and quantitatively by the system dynam-
ics tools discussed earlier in this chapter. As suggested by Schoenenberger
and Schenker-Wicki (2014), an ideal time to perform (social) network
analysis is during the first two steps of system dynamics modeling shown
in Figure 5.6, i.e., during problem familiarization and problem definition.
It can also play a role in step 8—for instance, when deciding who needs to
be involved in policy decision making.
In summary, network analysis and system dynamics are two methods
of analysis that, along with cross-impact analysis, mind maps, and prob-
lem trees, complement each other whether the goal is to map the landscape
in which the nexus unfolds and community development takes place or
decide on places to intervene in the landscape; the nexus; or the social,
natural, infrastructure, and economic systems.
48  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

5.6.3  From Cross-Impact to Network Analysis

It should be noted that cross-impact analysis and network analysis are


actually closely related. In fact, the double-causality table of cross-­impact
analysis (e.g., Table 3.5) becomes the so-called adjacency matrix in
­network analysis. For instance, Table 3.8 yields the following adjacency
matrix Eij (i = j = 4) for the enabling (E) environment:

0 1 3 3
3 0 2 3
Eij =  .
2 1 0 2
 1 0 1 0

In network analysis, this matrix translates into a valued-directed


graph that shows how the four sectors of the WELF nexus are con-
nected. The c­ entrality of each node, that is “the contribution the node
makes to the structure of the network” can be determined using different
­measures ­(Borgatti et al. 2013). For instance, using the degree of cen-
trality as a m
­ easure, the water, energy, land, and food sectors have out-
degree c­ entralities (sum of values of outgoing links) of 7, 8, 5, and 2,
respectively. They have indegree centralities (sum of values of ingoing
links) of 6, 2, 6, and 8, respectively. The outdegree centrality relates to
the degree of enabling influence of a sector of the nexus on the others, and
the indegree centrality indicates the enabling dependence of one sector
on the others.
For the constraining (C ) environment, Table 3.8 yields the following
adjacency matrix Cij (i = j = 4):

 0 0 −1 − 2 
 −1 0 −1 −1 
Cij =  .
 −1 0 0 −1 
 −1 0 0 0 

The outdegree centrality values are –3, –3, –2, and –1 and describe
the degree of constraining (risk) influence of the water, energy, and land
sectors on the others. The indegree centrality values are –3, 0, –2, and –4
for the four sectors, respectively. They indicate the constraining depen-
dence of each sector on the others.
Finally, both matrixes Eij and Cij can be added to determine the net
outdegree and indegree centrality for each sector of the nexus.
A System Dynamics Approach to the Nexus   •  49

5.7  Participatory Model Building

As discussed in Chapter 4, appraisal of the landscape, the nexus, and the


community is done by an appraisal team consisting ideally of representa-
tives of the three groups of stakeholders shown in Figure 2.1. Following
participatory appraisal, the identification of critical issues, and the formu-
lation of mental models and dynamic hypotheses around these issues, the
appraisal team and/or a partnering team must transition into the modeling
and intervention decision-making phases, which will be discussed more
extensively in Chapter 7.
Like the appraisal, modeling requires the contribution and engage-
ment of several key stakeholders who need to make critical decisions
about what is being modeled, why it is modeled, who does the model-
ing, over what geographic scale and time frame the modeling will be
done, and how the modeling should be done, including data selection.
There has been a lot of discussion in the system dynamics literature
about stakeholder participation in the development of models and the
analysis of modeling results for a wide range of applications. This
stakeholder proactiveness has been framed in the broader context of
group decision making and negotiation around a problem faced by a
client. In our case, the client may be one or several of the three stake-
holder groups shown in Figure 2.1. They may have a stake in making
decisions about various aspects of the FEW/WELF nexus and commu-
nity development.
The concept of participatory system dynamics has been proposed
to actively involve members in the decision-making processes of sys-
tem dynamics modeling (Stave 2010), and various methods have been
­proposed to do so (Rouwette et al. 2002; Richardson and Andersen 2010;
Rouwette and Vennix 2015). Among these methods is group model build-
ing, which originated in the 1990s (Vennix 1996; Andersen et al. 1997a,b).
Other variants include the community-based system dynamics method
(Hovmand 2014) and the participatory system dynamics method (Stave
2010). Another variation is the “modeling as learning” methodology pro-
posed by Lane (1992) where “stakeholder ownership” combined with
“limited expert consultancy” is suggested as the norm.
The objective of all these participatory system dynamic methods is
to empower those who will benefit from the modeling by helping them to
make their “implicit mental models” explicit and providing them with a
better understanding of the structure that is behind the dynamic ­hypotheses
and their mental models (Andersen et al. 1997b). As ­ suggested by
50  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

­ ouwette and Vennix (2015), the empowerment takes place at three ­levels:
R
(i) changing mindsets and behaviors at the individual level; (ii) ­creating
consensus and communication at the group level; and (iii) doing things
differently at the organizational level.
It should be noted that this multilevel empowerment process is not
random. For instance, group model building follows specific ­standards,
scripts, group facilitation sessions, and iterative decision-making ­processes
that have been recommended by several authors based on real case
­studies. Guiding principles for group model building have been ­proposed,
for instance, by Luna-Reyes et al. (2006); Stave (2010); Richardson and
Andersen (2010); and Rouwette and Vennix (2015), among many others.
Many questions may be raised prior, during, and after sessions of partici-
patory system dynamics modeling (Andersen et al. 1997b). In addition to
those listed in Table 5.5, such questions may include:

• How many participants are included?


• Who are the participants, and how are they selected?
• What is the distribution of participants (community, government,
outsiders)?
• What complementary skills and knowledge should participants
bring to the group?
• What should be the role of each participant?
• How will be participants be trained in group model building?
• When and where should group modeling meetings take place?
• How will decisions in model building, evaluation, and change be
reached and disagreement/conflict handled?
• What criteria are used to evaluate the effectiveness, impact, and
efficiency of participatory system dynamics modeling?

As remarked by several authors (Andersen et al. 1997a,b; ­Vennix


1996; Luna-Reyes et al. 2006; Richardson and Andersen 2010;
­Hovmand 2014), in order to be functional, methods such as group model
­building require a high level of coordination and preparation from the
­decision-making team. This can only be done if the team members are
assigned specific roles in conversation facilitation, model building,
analyzing the results of the model, making decisions and recommenda-
tions, conveying the decisions to the community members, revising the
­decisions, dealing with disagreement and conflicts, and so on. Ideally,
the decision-making team needs to a come up with a shared reality upon
which they all agree (Vennix 1996).
A System Dynamics Approach to the Nexus   •  51

Participatory system dynamics modeling also requires a willingness


of all stakeholders involved to participate in that shared reality and be
engaged in the multistep methodology described above. This assumes that
all stakeholders are fully aware of the interactive nature of the method-
ology and its expectations and have also been trained to think using sys-
tems tools. The training is, of course, a very sensitive issue when dealing
with local communities where some individuals have limited levels of
­education. Hence, the challenge for the modeling team facilitators is to be
able to capture ideas from community members in various meetings and
­interviews and translate these ideas into the system lingo, which in turn is
shared with community members.
One major concern of those interested in using participatory sys-
tem dynamics methods has been whether these methods provide better
decision-making outcomes for the issues addressed in the modeling than
non-participatory methods. Berard (2010) addressed that specific concern
by comparing the value proposition of 16 different group model building
frameworks described in the literature. Likewise, Stave (2010) compared
four examples of participatory system dynamic modeling with different
levels of stakeholder participation in the development of models and the
analysis of modeling results. More recently, Rouwette and Vennix (2015)
provided an extensive literature review on the effectiveness of various
group model building methods. All of these reviews seem to show a com-
mon value proposition, which is that of engaging various stakeholders in
making better decisions in a collaborative way through a learning pro-
cess and increased communication. However, how to measure that value
proposition (i.e., measuring change, as proposed by Doyle et al. 2002)
remains a topic of debate in the literature, since it is mostly qualitative
(Rouwette 2012). Walters et al. (2017) give a very good example of how
group model building was applied to decision making for a water project
in Peru. The authors clearly demonstrate, in a semi-quantitative and sta-
tistical way, the added value of systems thinking and participatory system
dynamics modeling in defining factors that could contribute to project’s
increased success.
It should be noted that the scenario of well-functioning participatory
model building teams holding to a fixed, long-term shared reality is very
utopic. Even those who developed and used the group model building
method have voiced their concerns and have strong reservations about
it, seeing the method more as a craft than as a science, with too much
subjectivity and not enough rigor (Andersen et al. 1997b). Although
many attempts have been made, there has not been much success at
52  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

o­ perationalizing group model-building or other participatory system


dynamics ­modeling methods.
Another issue to keep in mind is that developing an objective
approach to participatory modeling and being able to measure the added
value of participatory system dynamics modeling may simply not be
possible due the intrinsic ill-defined and messy nature of the problems
for which the methods are used. This is particularly true in the ­uncertain
and complex landscape in which the FEW/WELF nexus unfolds and
community development takes place. It is also likely that in that context,
assumptions and preconditions used to reach consensus will not be ful-
filled, and unexpected situations (risks) will arise. Finally, the integrity
of any decisions made in that context is also limited by a lack of objec-
tive reality, belief systems, and the dynamic of the group participants
when faced with complexity and uncertainty (Vennix 1999). It must
be remembered that the three major groups of stakeholders shown in
­Figure 2.1 are more likely to have different opinions, cognition, biases,
and heuristics, and will have difficulty reaching some form of consensus,
let alone agreeing on a system model of the issues they are collectively
experiencing. The shared reality of a group should be interpreted as a
dynamic process that needs to be revisited on a regular basis through
reflection-in-action.

5.8  Concluding Remarks

System dynamics modeling offers a strong value proposition to model


the complexity and uncertainty associated with the different systems at
play when addressing the nexus. This is even more true if it is combined
with other methods of analysis, such as network analysis and cross-­impact
analysis, and integrates participatory methods. Hybrid methods combin-
ing system dynamics with methods such as agent-based modeling and
geographic information systems (not considered in this book) have the
potential to explore decision making in the spatial-temporal realm and
offer interesting applications.
Despite its many advantages, system dynamics should not be seen as
a panacea to model complex systems and, like all methods of analysis, it
needs to be handled with caution. Aside from all the recommendations and
tests suggested in the system dynamics literature (Table 5.4), most of the
risks associated with the method arise from the modelers who create the
models. As suggested by Turner et al. (2016), good modeling practices need
to be followed in order to get the most out of system dynamics ­modeling.
A System Dynamics Approach to the Nexus   •  53

According to these authors, bad modeling practices may include avoiding


rushing into policymaking based on trust in poor simulations with limited
appraisal data; not understanding the role that endogenous and exogenous
variables play; creating “unnecessarily large models”; making decisions
based on nonupdated models; developing models based on “flawed cogni-
tive maps or (linear) mental models”; and overcoming “deeply ingrained
perceptions [and biases],” among others. Models are only as good as the
modelers who create them.

5.9  References

Abdelbari, H., and K. Shafi. 2017. “A Computational Intelligence-based Method


to ‘Learn’ Causal Loop Diagram-like Structures from Observed Data.” ­System
Dynamics Review 33, no. 1, pp. 3–33.
Adeniran, E.A., and O.A. Bamiro. 2015. “Development of a Strategic Planning
Model for a Municipal Water Supply Scheme using System Dynamics.”
­Journal of Water Resources and Protection 7, no. 15, pp. 1183–94.
Ahmad, S., and S.P. Simonovic. 2004. “Spatial System Dynamics: New Approach
for Simulation of Water Resources Systems.” Journal. of Computing in Civil
Engineering 18, no. 4, pp. 331–40.
Ahmad, S., and D. Prashar. 2010. “Evaluating Municipal Water Conservation
­Policies Using a Dynamic Simulation Model.” Water Resource Management
24, no. 13, pp. 3371–95.
Alcamo, J. 2015. “Systems Thinking for Advancing a Nexus Approach to Water,
Soil and Waste.” Keynote Address, Dresden Nexus Conference on Global
Change, Sustainable Development Goals and the Nexus Approach. http://
uni-kassel.de/einrichtungen/uploads/media/Alcamo-summary-­s ystems-
thinking-Dresden-Nexus-08-05-14-1.pdf (accessed October 21, 2016).
Amadei, B. 2015. A Systems Approach to Modeling Community Development
Projects. New York, NY: Momentum Press.
Andersen, D.F., J.A.M. Vennix, G.P. Richardson, and E.A.J.A. Rouwette. 1997a.
“Group Model Building: Problem Structuring, Policy Simulation and ­Decision
Support.” Journal of Operational Research Society 58, no. 5, pp. 691–94.
Andersen, D.F., G.P. Richardson, and J.A.M. Vennix. 1997b. “Group Model
­Building: Adding More Science to the Craft.” System Dynamics Review 13,
no. 2, pp. 187–91.
Ansari, N., and A. Seifi. 2013. “A System Dynamics Model for Analyzing Energy
Consumption and CO2 Emission in Iranian Cement Industry Under Various
Production and Export Scenarios.” Energy Policy 58, pp. 75–89.
Arcade, J., M. Godet, F. Meunier, and F. Roubelat. 2014. “Structural Analysis with
the MICMAC Method and the Actor’s Strategy with MACTOR Method.”
54  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

In Introduction to the Futures Methods Research Series. Futures Research


­Methodology, V3.0, Washington, DC: The Millennium Project.
Aslani, A., P. Helo, and M. Naaranoja. 2014. “Role of Renewable Energy Polices
in Energy Dependency in Finland: System Dynamics Approach.” Applied
Energy 113, pp. 758–65.
Atherton, J.T. 2013. A System Dynamics Approach to Water Resources and Food
Production in the Gambia [Master of Eng. Science thesis]. University of
Western Ontario, Electronic Thesis and Dissertation Repository, 1550. http://
ir.lib.uwo.ca/etd/1550 (accessed Febraury 24, 2008).
Barlas, Y. 1996. “Formal Aspects of Model Validity and Validation in System
Dynamics.” System Dynamics Review 12, no. 3, pp. 183–210.
Barney, G.O. 2002. “The Global 200 Report to the President and the Threshold
21 Model: Influences of Dana Meadows and System Dynamics.” System
­Dynamics Review 18, no. 2, pp. 123–36.
Bazilian, M., H. Rogner, M. Howells, S. Hermann, D. Arent, D. Gielen, P. Steduto,
A. Mueller, P. Komor, R.S.J. Tol, and K.K. Yumkella. 2011. “Considering the
Energy, Water and Food Nexus: Towards an Integrated Modeling Approach.”
Energy Policy 39, no. 12, pp. 7896–906.
Berard, C. 2010. “Group Model Building Using System Dynamics: An Analysis of
Methodological Frameworks.” The Electronic Journal of Business Research
Methods 8, no. 1, pp. 35–45.
Ben-Eli, M. 2016. “Martina Bustos: Project Transition.” A Sustainability Blueprint
for Community Development. New York, NY: The Sustainability Laboratory.
Benson, T., and S. Marlin. 2017. “The Habit-forming Guide to Becoming a
­Systems Thinker.” The Waters Foundation. Pittsburgh, PA: Systems Thinking
Group Publ.
Biswas, A.K., ed. 1976. Systems Approach to Water Management, 1–5. New York,
NY: McGraw Hill.
Blanchet, K., and P. James. 2012. “The Role of Social Networks in the Governance
of Health Systems: The Case of Eye Care Systems in Ghana.” Health Policy
and Planning 28, no. 2, pp. 1–14.
Bodin, O., B. Crona, and H. Ernstson. 2006. “Social Networks in Natural Resource
Management: What is there to Learn from a Structural Perspective?” Ecology
and Society 11, no. 2.
Borgatti, S.P., M.G. Everett, and L.C. Freeman. 2002. UCINET for Windows:
­Software for Social Network Analysis. Harvard, MA: Analytic Technologies.
Borgatti, S.P., A. Mehra, D.J. Brass, and G. Labianca. 2009. “Network Analysis in
the Social Sciences.” Science 323, no. 5916, pp. 892–95.
Borgatti, S.P., M.G. Everett, and J.C. Johnson. 2013. Analyzing Social Networks.
Thousand Oaks, CA: Sage Publications.
Borshchev, A., and A. Filippov. July 25–29, 2004. “From System Dynamics and
Discrete Event to Practical Agent-based Modeling: Reasons, Techniques,
Tools.” Proc. 22nd International Conference of the System Dynamics Society.
Oxford, England.
A System Dynamics Approach to the Nexus   •  55

Bossel, H. 2007a. System Zoo 1 Simulation Models. GmbH, Norderstedt, ­Germany:


Books on Demand.
Bossel, H. 2007b. System Zoo 2 Simulation Models. GmbH, Norderstedt, ­Germany:
Books on Demand.
Bossel, H. 2007c. System Zoo 3 Simulation Models. GmbH, Norderstedt, ­Germany:
Books on Demand.
Braun, W. 2002. The System Archetypes. University at Albany. http://albany.edu/
faculty/gpr/PAD724/724WebArticles/sys_archetypes.pdf (accessed September
2, 2014).
Breslin, P. 2004. “Thinking Outside Newton’s Box: Metaphors for Grassroots
Development.” Grassroots Development 25, no. 1, pp. 1–9.
Checkland, P., and J. Poulter. 2006. Learning for Action: Soft Systems ­Methodology
and Its Use for Practitioners, Teacher and Students. Chichester, UK:
John Wiley & Sons.
Cheng, L. 2010. “System Dynamics Model of Suzhou Water Resources Carry-
ing Capacity and Its Application.” Water Science and Engineering 3, no. 2,
pp. 144–55.
Chi, K.C., W.J. Nuttall, and D.M. Reiner. 2009. “Dynamics of the UK Natural Gas
Industry: System Dynamics Modeling and Long-term Energy ­Policy ­Analysis.”
Technological Forecasting and Social Change 76, no. 3, pp. 339–57.
Choucri, N., C. Electris, D. Goldsmith, D. Mistree, S.E. Madnick, J.B. Morrison,
M.D. Siegel, and M. Sweitzer-Hamilton. January 2006. ­“Understanding and
Modeling State Stability: Exploiting System Dynamics.” IEEAC paper #1278,
version 5, pp. 1–11.
Chung, G., J.H. Kim, and T.-W. Kim. 2008. “System Dynamics Modeling
Approach to Water Supply System.” KSCE Journal of Civil Engineering 12,
no. 4, pp. 275–80.
Creative Leaning Exchange. 2015. http:// clexchange.org/curriculum/roadmaps/
(accessed December 8, 2016).
Daher, B.T., and R.H. Mohtar. 2015. “Water-energy-food (WEF) nexus tool
2.0: Guiding Integrative Resource Planning and Decision-Making.” Water
International. http://dx.doi.org/10.1080/02508060.2015.1074148 (accessed
August 28, 2016).
Dale, P. 2011. “Ties that Bind: Studying Social Networks in Timor-Leste.”
People, Spaces, Deliberation, World Bank Blog. http://blogs.worldbank.
org/publicsphere/ties-bind-studying-social-networks-timor-leste (accessed
March 1, 2015).
Dale, L.L., N. Karali, D. Millstein, M. Carnall, S. Vicuña, N. Borchers, et al..
2015. “An Integrated Assessment of Water-Energy and Climate Change in
Sacramento, California: How Strong is the Nexus.” Climate Change 132,
pp. 223–35.
Davies, E.G.R., and S.P. Simonovic. 2010. “ANEMI: A New Model for Integrated
Assessment of Global Change.” Interdisciplinary Environmental Review 11,
nos. 2/3, pp. 127–61.
56  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Davies, E.G.R., and S.P. Simonovic. 2011. “Global Water Resources Modeling
with an Integrated Model of the Social-Economic-Environmental System.”
Advances in Water Resources 34, pp. 684–700.
Dawadi, S., and S. Ahmad. 2012. “Changing Climatic Conditions in the Colorado
River Basin: Implications for Water Resources Management.” Journal of
Hydrology 430, pp. 197–215.
De Savigny, D., and A. Taghreed., eds. 2009. Systems Thinking for Health System
Strengthening. WHO, Geneva, Switzerland: Alliance for Health Policy and
Systems Research.
Dent, E. B. 2001. “System Science Traditions: Differing Philosophical Assumptions.”
Systems: J. Transdisciplinary Systems Science 6, nos. 1–2, pp. 13–30.
Dhawan, R., M. O’Connor, and M. Borman. 2011. “The Effect of Qualitative and
Quantitative System Dynamics Training: An Experimental Investigation.”
System Dynamics Review 27, no. 3, pp. 313–27.
Downes, R.J., and S.R. Bishop. 2016. “Aid Allocation: A Complex Perspective.”
Chapter 16 in Global Dynamics: Approaches from Complexity Science, ed.
A. Wilson. Chichester, UK: Wiley.
Doyle, J.K., D.N. Ford, M.J. Radzicki, and W.S. Trees. 2002. “Mental ­Models
of Dynamic Systems.” In System Dynamics and Integrated Modeling, ed.
Y. Barlas from Encyclopedia of Life Support Systems (EOLSS), developed
under the auspices of the UNESCO, EOLSS Publishers, Oxford, UK.
http://eolss.net.
Feng, Y.Y., S.Q. Chen, and L.X. Zhang. 2013. System Dynamics Modeling for
Urban Consumption and CO2 Emissions: A case study of Beijing. China.
Ferroukhi, R., D. Nagpal, A. Lopez-Peña, T. Hodges, R.H. Mohtar, B. Daher,
S. Mohtar, and M. Keulertz. 2015. “Renewable Energy in the Water, Energy,
and Food Nexus.” United Arab Emirates: International Renewable Energy
Agency (IRENA) Policy Unit. Retrieved from www.irena.org/Publications
(accessed August 31, 2016).
Fisher, D.M. 2011. Modeling Dynamic Systems: Lessons for a First Course, 3rd ed.
ISEE Systems, Inc.
Flammini, A., M. Puri, L. Pluschke, and O. Dubois. 2014. “Walking the Nexus Talk:
Assessing the Water-Energy-Food Nexus in the Context of the Sustainable
Energy for All Initiative.” Environment and Natural Resources Working Paper
No. 58, FAO, Rome.
Food and Agriculture Organization of the United Nations (FAO). 2014. The
Water-Energy-Food Nexus: A New Approach in Support of Food Security and
Sustainable Agriculture. Rome: FAO.
Ford, A. 1996. “System Dynamics and the Electric Power Industry.” System
Dynamics Review 13, no. 1, pp. 57–85.
Ford, A. 2010. Modeling the Environment. Washington, DC: Island Press.
Forrester, J.W. 1961. Industrial Dynamics. Pegasus Communications, Waltham,
MA. A More Recent Reprint of this Book is Available through Martino
Publishing, Mansfield Center, CT (2013).
Forrester, J.W. 1969. Urban Dynamics. Portland, OR: Productivity Press.
A System Dynamics Approach to the Nexus   •  57

Forrester, J.W. 1971. World Dynamics. (second edition in 1973). Portland, OR:
Productivity Press.
Forrester, J., and P.M. Senge. 1980. “Tests for Building Confidence in System
Dynamics Models.” In System Dynamics. TIMS Studies in the Management
Sciences, eds. A.A. Legasto, J.W. Forrester and J.M. Lyneis, 14, 209–28.
Fowler, B., and E. Dunn. 2014. “Evaluating Systems and Systemic Change for
Inclusive Market Development: Literature Review and Synthesis.” Report
No. 3, Washington, DC: U.S. Agency for International Development.
Freeman, L.C. 2000. “Visualizing Social Networks.” Journal of Social ­Structure 1,
no. 1, p. 4.
Future Health Systems. 2014. “Workshop on Complex Adaptive Systems.”
­Baltimore, June. http://futurehealthsystems.org/news/2014/7/30/fhs-and-
steps-co-host-workshop-on-complex-adaptive-systems (accessed September
26, 2014).
Ghosh, A. 2017. Dynamic Systems for Everyone: Understanding How Our World
Works, 2nd ed. Cham, Switzerland: Springer.
Giraldo, D.P., M.J. Betancur, and S. Arango. 2008. “Food Security in Developing
Countries: A Systemic Perspective.” http://academia.edu/14539594/Food_
Security_in_Development_Countries_A_systemic_perspective (accessed
August 10, 2018).
Gober, P., E.A. Wentz, T. Lant, M.K. Tschudi, and C.W. Kirkwood. 2011. ­“WaterSim:
A Simulation Model for Urban Water Planning in Phoenix, ­Arizona, USA.”
Environment and Planning B: Planning and Design 38, ­no. 2, pp. 197–215.
Godet, M. 2001. Creating Futures: Scenario Planning and a Strategic ­Management
Tool. Washington, DC: Brookings Institution Press.
Goodman, M., and A. Kleiner. 1994. “The Archetype Family Tree.” In The
fifth ­Discipline Fieldbook: Strategies and Tools for Building a Learning
­Organization by Senge et al. 1994. New York, NY: Doubleday.
Guo, H.C., L. Liu, G.H. Huang, G.A. Fuller, R. Zou, and Y.Y. Yin. 2001. “A System
Dynamics Approach for Regional Environmental Planning and Management:
A Study for the Lake Erhai Basin.” Journal of Environmental Management
61, no. 1, pp. 93–11.
Hamilton, H.R. 1969. Systems Simulation for Regional Analysis: An Application
to River Basin Planning. Cambridge, MA: MIT Press.
Hannon, B., and M. Ruth. 2001a. Modeling Dynamic Biological Systems.
New York, NY: Springer.
Hannon, B., and M. Ruth. 2001b. Dynamic Modeling. New York, NY: Springer.
Hardin, G. 1968. “The Tragedy of the Commons.” Science 152, pp. 1243–48.
Hardin, G. 1994. “The Tragedy of the Unmanaged Commons.” Trends in Ecology
and Evolution 9, no. 5, p. 199.
Hargrove, J.L. 1998. Dynamic Modeling in the Health Science. New York, NY:
Springer.
Hermann, S. 2012. “Finding CLEWS. Exploring Sustainable Energy Develop-
ments: Looking at Climate-Land-Energy-water Interactions ­Methodology and
­Components.” Presentation at joint ICTP-IAEA ­Workshop on ­ Sustainable
58  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

energy Development, October 1–5, 2012. Trieste, Italy. http://indico.ictp.it/


event/a11197/session/40/contribution/30/material/0/0.pdf (accessed ­November
15, 2017).
Hermann, S., M. Welsch, R.E. Segerstrom, M.I. Howells, C. Young, T. A ­ lfstad,
H-H. Rogner, and P. Steduto. 2012. “Climate, Land, Energy and Water (CLEW)
Interlinkages in Burkina Faso: An Analysis of Agricultural ­Intensification and
Bioenergy Production.” Natural Resources Forum, 36, pp. 245–62.
Hjorth, P., and A. Bagheri. 2006. “Navigating Towards Sustainable Development:
A System Dynamics Approach.” Future 38, pp. 74–92.
Hoff, H. 2011. Understanding the Nexus. Background paper for the Bonn 2011
Conference: The Water, Energy and Food Security Nexus. Stockholm
­Environment Institute, Stockholm, Sweden.
Holling, C.S., and G.K. Meffe. 1996. “Command and Control and Pathology
of Natural Resources Management.” Conservation Biology 10, no. 2,
pp. 328–37.
Homer, J., and R. Oliva. 2001. “Maps and Models in System Dynamics:
A Response to Coyle.” System Dynamics Review 17, no. 4, pp. 347–56.
Homer, J.B., and G.B. Hirsch. 2006. “System Dynamics Modeling for Public
Health: Background and Opportunities.” American J. Public Health 96, no. 3,
pp. 452–58.
Hosseinichimeh, N., R. MacDonald, A. Hyder, A. Ebrahimvandi, L. Porter,
R. Reno, J. Maurer, D.L. Andersen, G. Richardson, J. Hawley, and D.F.
Andersen. 2017. “Group Model Building Techniques for Rapid ­Elicitation of
Parameter Values, Effect Sizes, and Data Sources.” System Dynamics Review
33, no. 1, pp. 71–84.
Houghton, J., M. Siegel, A. Wirsch, A. Moulton, S. Madnick, and D. Goldsmith.
2014. “A Survey of Methods for Data Inclusion in System Dynamics ­Models:
Methods, Tools and Applications.” Working paper CISL 2014-03, Sloan
School of Management, MIT, Cambridge, MA.
Hovmand, P.S. 2014. Community Based System Dynamics. New York, NY:
Springer.
Howells, M., S. Hermann, M. Welsch, M. Bazilian, R. Segerström, T. Alfstad,
D. Gielen, H. Rogner, G. Fischer, H. Van Velthuizen, and D. Wiberg. 2013.
­“Integrated Analysis of Climate Change, Land-use, Energy and Water Strate-
gies.” Nature Climate Change 3, no. 7, pp. 621–26.
Hughes, B.B. 1999. International Futures: Choices in the Face of Uncertainty.
Boulder, CO: Westview Press.
Hughes, B.B., and E.E. Hillebrand. 2006. Exploring and Shaping International
Futures. Boulder, CO: Paradigm Publishers.
Hughes, B.B. 2015. “International Futures (IFs) Training Manual.” Working paper
2015.04.02. Pardee Center for International Futures, Josef Korbel School of
International Studies, University of Denver, Denver, CO.
Hughes, B.B. 2016. “International Futures (IFs) and Integrated, Long-term
­Forecasting of Global Transformations.” Futures 81, pp. 98–118.
A System Dynamics Approach to the Nexus   •  59

Jaeger, A., and T. Henrichs. 2008. Modeling Environmental Change in Europe:


Towards a Model Inventory (SEIS/Zforward). EEA Technical report 11,
Copenhagen: European Environment Agency.
Janssen, M.A., and W.J.M. Martens. 1997. “Modelling Malaria as a Complex
Adaptive System.” Artificial Life 3, no. 3, pp. 213–36.
Jeon, C., and J. Shin. 2014. “Long-term Renewable Energy Technology ­Valuation
Using System Dynamics and Monte-Carlo simulation: Photovoltaic
­Technology Case.” Energy 66, pp. 447–57.
Jones, A.P., J.B. Homer, D.L. Murphy, J.D.K. Essien, B. Milstein, and
D.A. Seville. 2006. “Understanding Diabetes Population Dynamics through
Simulation Modeling and Experimentation.” American Journal of Public
Health 96, no. 3, pp. 488–94.
Khan, S., L. Yufeng, and A. Ahmad. 2007. “System Dynamics Modeling for Water
Savings and Conjunctive Water Management.” Proc. Asian Simulation and
Modeling Conference, Chiang Mai, Thailand.
Kiani, B., S. Mirzamohammadi, and S.H. Hosseini. 2010. “A Survey on the Role
of System Dynamics Methodology on Fossil Fuel Resources Analysis.”
­International Business Research 3, no. 3, pp. 84–93.
Kim, D.H. 2000. Systems Archetypes I and II. Charleston, NC: Nabu Press.
Kim, D.H., and V. Anderson. 2007. System Archetypes Basics: From Story to
Structure. Waltham, MA: Pegasus Communications.
Kincaid, L. 2000. “Social Networks, Ideation, and Contraceptive Behavior in
­Bangladesh: A Longitudinal Analysis.” Social Science and Medicine 50,
no. 2, pp. 215–31.
Lane, D.C. 1992. “Modelling as Learning: A Consultancy Methodology
for Enhancing Learning in Management Teams.” European Journal of
­Operational Research 59, no. 1, pp. 64–84.
Lebel, J. 2003. Health: An Ecosystem Approach. Ottawa, Canada: International
Development Research Center.
Leontief, W. 1986. Input-Output Economies. New York, NY: Oxford University
Press.
Luna-Reyes, L.F., and D. Andersen. 2003. “Collecting and Analyzing Qualitative
Data for System Dynamics: Methods and Models.” System Dynamics Review
19, no. 4, pp. 271–96.
Luna‐Reyes, L.F., I.J. Martinez‐Moyano, T.A. Pardo, A.M. Cresswell, D.F. ­Andersen,
and G.P. Richardson. 2006. “Anatomy of a Group ­Model-Building Interven-
tion: Building Dynamic Theory from Case Study Research.” System Dynamics
Review 22, no. 4, pp. 291–320.
Mavotras, G., ed. 2010. Foreign Aid for Development: Issues, Challenges, and the
New Agenda. WIDER Studies in Development Economics. New York, NY:
Oxford University Press.
McCormick, R., and L.A. Kapustka. 2016. “The Answer is 42…What is the
­Q uestion?” Journal of Environmental Studies and Sciences 6, no. 1,
pp. 208–13.
60  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

McDonell, G., M. Hefferman, and A. Faulkner. 2004. “Using System Dynamics


to Analyze Health System Performance within the WHO Framework.” http://
systemdynamics.org/conferences/2004/SDS_2004/PAPERS/337MCDON.
pdf (accessed December 16, 2016).
Meadows, D.H., J. Randers, and D. Meadows. 2004. Limits to Growth: The
30-year Update. White River Junction, VT: Chelsea Green Publishing.
Meadows, D. 2008. Thinking in Systems. White River Junction, VT: Chelsea
Green Publishing.
Metlay, D., and D. Sarewitz. 2012. “Decision Strategies for Addressing Complex,
‘Messy’ Problems.” The Bridge 42, no. 3, pp. 6–16.
Mirchi, A., K. Madani, D. Watkins, and S. Ahmad. 2012. “Synthesis of S ­ ystem
Dynamics Tools for Holistic Conceptualization of Water Resources
­Problems.” Water Resources Management 26, no. 9, pp. 2412–42.
Monasterolo, I., R. Pasqualino, and E. Mollona. n.d. “The Role of System Dynam-
ics Modeling to Understand Food Chain Complexity and Address Challenges
for Sustainability Policies.” In Proceedings of the SYDIC and the Food and
Agriculture Organization of the United Nations (FAO)“Meeting Urban Food
Needs” project (forthcoming).
Moore, S., E. Eng, and M. Daniel. 2003. “International NGOs and the Role
of Network Centrality in Humanitarian Aid Operations: A Case Study of
Coordination During the 2000 Mozambique Floods.” Disaster 27, no. 4,
pp. 305–18.
Moumouni, Y., S. Ahmad, and R.J. Baker. 2014. “A System Dynamics Model for
Energy Planning in Niger.” Int. J. of Energy and Power Eng 3, no. 6, pp. 308–22.
Naill, R.F. 1992. “A System Dynamics Model for National Energy Policy Plan-
ning.” System Dynamics Review 8, no. 10, pp. 1–19.
National Research Council (NRC). 2009. Applications of Social Network Analysis
for Building Community Disaster Resilience. Washington, DC: The National
Academies Press.
Nava Guerrero, G., P. Schwarz, and J. Slinger. 2016. “A Recent Overview of the Inte-
gration of System Dynamics and Agent-based Modelling and Simulation.” In
Proceedings of the 34th International Conference of the System Dynamics Soci-
ety, July 17–21, 2016 Delft, Netherlands. [P1153] System Dynamics Society.
Neely, K. 2015. “Complex Adaptive Systems as a Valid Framework for
­Understanding Community Level Development.” Development in Practice
25, no. 6, 785–97, doi:10.1080/099614524.2015.1060949
Newell, B., D. Marsh, and D. Sharma. 2011. “Enhancing the Resilience of
the ­Australian Electric Market; Taking a Systems Approach in Policy
­Development.” Ecology and Society 16, no. 2, Article 5.
Newman, J., M.A. Velasco, L. Martín, and A.M. Fantini. 2003. “A System
Dynamics Approach to Monitoring and Evaluation at the Country Level.”
Presented at the 5th Biennial World Bank Conference on Evaluation
and Development, Washington DC. http://csdnet.dyson.cornell.edu/papers/­
newman.pdf (accessed August 29, 2014).
A System Dynamics Approach to the Nexus   •  61

Newman, J.L. 2010. “Finding System Dynamics: An Exploration in International


Development.” In Tracing connections: Voices of Systems Thinkers, eds.
J. Richmond, L. Stuntz, K. Richmond, and J. Egner, 143–65. Lebanon, NH:
ISEE Systems Inc.
Newman, M.E.J. 2010. Networks: An Introduction. Oxford, UK: Oxford
University Press.
Pasqualino, R., A.W. Jones, I. Monasterolo, and A. Phillips. 2015. “Understanding
Global Systems Today: A Calibration of the World3-03 Model Between 1995
and 2012.” Sustainability 7, pp. 9864–89.
Perkin, E., and J. Court. 2005. “Networks and Policy Processes in International
Development: A Literature Review.” Overseas Development Institute,
­Working Paper 252, London, UK: Research and Policy in Development
­Programme (RAPID).
Pidd, M. 2004. Computer Simulation in Management Science, 5th ed. Chichester,
UK: John Wiley & Sons.
Porter, J. 2011. “UK Collaborative on Development Sciences Report: Complexity,
Science, and International Development.” https://www2.warwick.ac.uk/fac/
cross_fac/complexity/people/students/dtc/students2007/porter/international_
development_and_complexity_science.pdf (accessed November 5, 2017).
Pruyt, E. 2013. Small System Dynamics Models for Big Issues: Triple Jump Towards
Real-World Dynamic Complexity. Version 1.0, Delft, The ­Netherlands: TU
Delft Library.
Qudrat-Ullah, H. 2016. The Physics of Stocks and Flows of Energy Systems.
Springer Briefs in Complexity, Chapter 3. http:// bookmetrix.com/detail/chap-
ter/3322126b-d968-437f-8139-d3891c2cc53b#downloads (accessed November
21, 2017).
Ramalingam, B. 2014. Aid on the Edge of Chaos: Rethinking International
­Cooperation in a Complex World. New York, NY: Oxford University Press.
Reynolds, J.F., D.M. Stafford-Smith, and E. Lambin. 2003. “Do Humans
Cause Deserts? A Land Problem Through the Lens of a New Framework:
The Dahlem Desertification Paradigm.” Proceedings of VIIth ­International
Rangelands Congress, eds. N. Allsopp, A.R. Palmer, S.J. Milton,
K.P. Kerkman, G.I.H. Kerley, C.R. Hurt, and C.J. Brown. South Africa: Durban.
Richardson, G.P. 1999. Feedback thought in Social Science and Systems Theory.
Waltham, MA: Pegasus Communications.
Richardson, G.P., and D.F. Andersen. 2010. “Systems Thinking, Mapping, and
Modeling in Group Decision and Negotiation.” In Handbook for Group
­Decision and Negotiation, eds. C. Eden and D.N. Kilgour, 313–24. Springer.
Richmond, B. 1994. “System Dynamics/Systems thinking: Let’s Just Get on With
It.” International System Dynamics Conference, Sterling, Scotland. http://
iseesystems.com/resources/Articles/SDSTletsjustgetonwithit.pdf (accessed
August 28, 2014).
Richmond, B. 1997. “The ‘Thinking’ in Systems Thinking: How Can We Make It
Easier to Master?” The Systems Thinker 8, no. 2, pp. 1–5.
62  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Richmond, B. 2004. An Introduction to Systems Thinking, STELLA Software.


­Lebanon, NH: ISEE Systems, Inc.
Richmond, J., L. Stuntz, K. Richmond, and J. Egner. 2010. Tracing Connections:
Voices of Systems Thinkers. Lebanon, NH: isee Systems, Inc.
Rihani, S., and R. Geyer. 2001. “Complexity: An Appropriate Framework for
Development?” Progress in Development Studies 1, no. 3, pp. 237–45.
Rihani, S. 2002. Complex Systems Theory and Development Practice:
­Understanding Nonlinear Realities. London, UK: Zen Books.
Ritchie-Dunham, J.L. 1997. “Initiating Management Dialog Using a Summary
Presentation that Integrates the Findings from Multiple System Dynamics
Analytical Tools.” Proc. Int. System Dynamics Conference, Istanbul, Ref#087.
Rittel, H., and M. Webber. 1973. “Dilemmas in a General Theory of Planning.”
Policy Sci. 4, pp. 155–69.
Robalino-Lopez, A., A. Mena-Nieto, and J.E. Garcia-Ramos. 2014. “System
Dynamics Modeling for Renewable Energy and CO2 Emissions: A Case
Study of Ecuador.” Energy for Sustainable Development 20, pp. 11–20.
Robinson, W.A. 2001. Modeling Dynamic Climate Systems. New York, NY
Springer.
Rotmans, J., and B. deVries., eds. 1997. Perspectives on Global Change: The
­TARGETS Approach. Cambridge, UK: Cambridge University Press.
Rouwette, E., J.A.M. Vennix, and T. van Mullekon. 2002. “Group Model Building: A
Review of Assessment Studies.” System Dynamics Review 18, no. 1, pp. 5–45.
Rouwette, E. 2012. “Does Group Modeling Work? Evidence from and Comments on
the Paper by N. Videira, R. Lopes, P. Antunes, R. Santos, and J.L. Casanova.”
Systems Research and Behavioral Science 29, pp. 620–23.
Rouwette, E., and J.A.M. Vennix. 2015. “Group Model Building.” Encyclopedia
of Complexity and Systems Science. doi:10.1007/978-3-642-27737-5_264-3
Rwashana, A.S., and D.W. Williams. 2008. “System Dynamics Modeling in
Healthcare: The Uganda Immunization System.” Int. J. Computing and ICT
Research 1, no. 1, pp. 85–98.
Schiffer, E., and D. Waale. 2008. “Tracing Power and Influence in Networks:
­Net-Map as a Tool for Research and Strategic Network Planning.” Discus-
sion paper 00772, Washington, DC: International Food Policy Research
Institute.
Schoenenberger, K., and A. Schenker-Wicki. 2014. “Can System Dynamics Learn
from Social Network Analysis?” Proc. International Conference of the ­System
Dynamics Society, Delft.
Schwaninger, M. 1997. “Integrative Systems Methodology: Heuristic for
Requisite Variety.” International Transactions in Operational Research 44,
no. 4, pp. 109–23.
Schwaninger, M. 2009. “The Role of System Dynamics Within the Systems
Movement.” In Vol. 1 of Encyclopedia of Complexity and Systems Science,
ed. R.A. Myers. Springer.
Scoones, I., M. Leach, A. Smith, S. Stagl, A. Stirling, and J. Thompson. 2007.
“Dynamic Systems and the Challenge of Sustainability.” STEPS Working
paper 1. Brighton, UK: STEPS Center.
A System Dynamics Approach to the Nexus   •  63

Senge, P. 1994. The Fifth Discipline: The Art & Practice of the Learning
­Organization. New York, NY: Doubleday.
Senge, P.M., A. Kleiner, C. Roberts, R.B. Ross, and B.J. Smith. 1997. The
Fifth Discipline Fieldbook: Strategies and Tools for Building a Learning
­Organization. New York, NY: Doubleday.
Simonovic, S.P., and H. Fahmy. 1999. “A New Modelling Approach for
Water Resources Policy Analysis.” Water Resources Research 35, no. 1,
pp. 295–304.
Simonovic, S.P. 2003. “Assessment of Water Resources through System ­Dynamics
Simulation: From global Issues to Regional Solutions.” Proc. 36th Hawaii
International Conference on Systems Science. Washington, DC: Computer
Society Press.
Simonovic, S.P., and V. Rajasekaram. 2004. “Integrated Analyses of Canada’s
Water Resources: A System Dynamics Approach.” Canadian Water Resources
J. 29, no. 4, pp. 223–50.
Simonovic, S.P. 2008. “Managing Water Resources: Methods and Tools for a
­Systems Approach.” vodoprivreda 0350-0519, no. 40, pp. 157–65. Paris:
UNESCO Publishing.
Sohofi, S.A., A. Melkonyan, C.K. Karl, and K. Krumme. 2016. “System
­Archetypes in the Conceptualization Phase of Water-Energy-Food-Nexus
Modeling.” Proceedings of 34th International Conference of the System
Dynamics ­Society, Delft, The Netherlands. http://systemdynamics.org/­
conferences/2016/proceed/papers/P1197.pdf (accessed September 12, 2016).
Stahl, C., and A. Cimorelli. 2006. “A Demonstration of the Necessity and
­Feasibility of Using a Clumsy Decision Analytic Approach on Wicked
­Environmental Problems.” Integrated Environmental Assessment and Man-
agement 9, no. 1, pp. 17–30.
Stave, K. 2010. “Participatory System Dynamics Modeling for Sustainable
­Environmental Management: Observations from Four Cases.” Sustainability
2, pp. 2762–84.
Sterman, J. 2000. Business Dynamics: Systems Thinking and Modeling for a
­Complex World. New York, NY: Irwin, McGraw Hill.
Sterman, J. 2006. “Learning from Evidence in a Complex World.” Am. J. Public
Health 96, no. 3, pp. 505–14.
Stockholm Energy International (SEI). 2013. “LEAP: Long Range Alternative Plan-
ning Systems.” https://sei-international.org/leap (accessed December 24, 2017).
Stroh, D.P. 2015. Systems Thinking for Social Change. White River Junction, VT:
Chelsea Green Publishing.
Sustainable Development Solutions Network (SDSN). 2015. Indicators and
a Monitoring Framework for the Sustainable Development Goals. http://
unsdsn.org/resources/publications/indicators/ (accessed February 19, 2016).
Sweeney, L.B. 2001. When a Butterfly Sneezes: A Guide for Helping Kids Explore
Interconnections in Our World Through Favorite Stories. Waltham, MA:
­Pegasus Communications.
System Dynamics Society. 2016. “The Field of System Dynamics.” http://­
systemdynamics.org/what-is-s/ (accessed December 8, 2016).
64  •   A SYSTEMS APPROACH TO MODELING THE WELF NEXUS

Takahashi, S. 2012. “Water Cycle Analysis System Dynamics Model for D ­ esigning
Optimal Water Reclamation Scheduling.” Proc. System Dynamics Society
Conference, St. Gallen, Switzerland.
Toole, C.L., and A.W. McCown. 2008. “Interdependent Energy Infrastructure
Simulation System.” In Handbook of Science and Technology for Homeland
Security, ed. J.G. Voeller. John Wiley and Sons.
Turner, B.L., H.M. Menendez, R. Gates, L.O. Tedeschi, and A.S. Atzori. 2016.
“System Dynamics Modeling for Agricultural and Natural Resource
­Management Issues: Review of Some Past Case and Forecasting Future
Roles.” Resources 5, no. 4, p. 40.
Ullah, M.A. 2012. Enhancing the Understanding of Corruption through
­System Dynamics Modelling: A Case Study analysis of Pakistan [Doctoral
­Dissertation]. New Zealand: University of Auckland.
Umpleby, S.A., and E.B. Dent. 1999. “The Origins and Purposes of Several
­Traditions in Systems Theory and Cybernetics.” Cybernetics and Systems 30,
pp. 79–103.
U.S. Agency for International Development (USAID). 2011a. Systems Thinking
in conflict Assessment: Concepts and Application. Washington, DC: USAID.
U.S. Agency for International Development (USAID). 2011b. USAID Complexity
Event. Washington, DC: USAID.
U.S. Agency for International Development (USAID). 2014. Local Systems: A
Framework for Supporting Sustained Development. Washington, DC: USAID.
Vaughan, M. 2013. The Thinking Effect. Boston, MA: Nicholas Brealey ­Publishing.
Vennix, J.A.M. 1996. Group Model Building: Facilitating Team Learning using
System Dynamics. New York, NY: Wiley.
Vennix, J.A.M. 1999. “Group Model-Building: Tacking Messy Problems.” System
Dynamics Review 15, no. 4, pp. 379–401.
Wakeford, J., C. Kelly. and S.M. Lagrange. 2015. Mitigating Risks and
­Vulnerabilities in the Energy-Food-Water Nexus in Developing Countries.
Sustainability Institute, South Africa: Stellenbosch.
Walters, J.P., B. Greiner, E. O’Morrow, and B. Amadei. 2017. “Fostering ­Systems
Thinking within Engineers Without Borders Student Teams using Group
Model Building.” Int. J. of Engineering Education 33, no. 1, pp. 1–14.
Waters Foundation. 2016. http://watersfoundation.org/systems-thinking/­overview/
(accessed December 9, 2016)
Watzlawick, P., J.H. Weakland, and R. Fisch. 1974. Change: Principles of
­Problem Formation and Problem Resolution. New York, NY: W.W. Norton,
­Republished in 2011 by W.W. Norton & Company, Inc.
Welsch, M., S. Hermann, M. Howells, H.H. Rogner, C. Young, I. Ramma, M.
Bazilian, G. Fischer, T. Alfstad, D. Gielen, and D. Le Blanc. 2014. “Adding
Value with CLEWS: Modeling the Energy System and Its Interdependencies
for Mauritius.” Applied Energy 113, pp. 1434–45.
Williams, R. December 2, 2008. “Bucking the System.” The Broker. http://the-
brokeronline.eu/Articles/Bucking-the-system (accessed December 11, 2016).
A System Dynamics Approach to the Nexus   •  65

Williams, R., and H. Britt. 2014. “Systems Thinking for Monitoring: Attending to
Interrelationships, Perspectives, and Boundaries.” Discussion note, version
1.0. Washington, DC: U.S. Agency for International Development.
Winz, I., and G. Brierley. 2007. “The Use of System Dynamics Simulation
in ­Integrated Water Resources Management.” Proceedings of the 25th
­International Conference of the System Dynamics Society, Boston. http://­
systemdynamics.org/conferences/2007/ (accessed December 16, 2016).
Winz, I., G. Brierley, and S. Trowsdale. 2009. “The Use of System Dynamics
Simulation in Water Resources Management.” Water Resources Management
23, no. 7, pp. 1301–23.
Wolfe, M.L., K.C.Ting, N. Scott, A. Sharpley, J.W. Jones, and L. Verma. 2016.
“Engineering Solutions for Food-Energy-Water Systems: It Is More than
Engineering.” J. Environ. Stud. Sci., 6, pp. 17282.
Wolstenholme, E.F. 1990. System Enquiry: A System Dynamics Approach.
­Chichester, UK: John Wiley & Sons.
Wolstenholme, E.F. 1999. “Qualitative vs. Quantitative Modeling: The Evolving
Balance.” J. of the Operational Research Society 50, no. 4, pp. 422–28.
World Business Council for Sustainable Development (WBCSD). 2009. Water,
Energy and Climate Change: A Contribution from the Business Community.
Washington DC.
Wu, B. 2016. “International Development Aid: A Complex System.” Chapter 14
in Global Dynamics: Approaches from Complexity Science, ed. A. Wilson.
Chichester, UK: Wiley.
Xu, Z.X., K. Takeuchi, H. Ishidaira, and X.W. Zhang. 2002. “Sustainability
­Analysis for Yellow River Resources using the System Dynamics Approach.”
Water Resources Management 16, pp. 239–61.
Yates, D., J. Sieber, D. Purkey, and A. Huber-Lee. 2005. “WEAP21- A Demand-,
Priority-, and Preference-Driven Water Planning Model.” Water International
30, no. 4, pp. 487–500.
Zhang, Q., C. Prouty, J.B. Zimmerman, and J.R. Mihelcic. 2016. “More Than
­Target 6.3: A Systems Approach to Rethinking Sustainable Development
Goals in a Resource-Scarce World.” Engineering 2, pp. 481–89.
Zhou J. 2014. “From Disease to Epidemic: A Look at the Spanish Flu, Ebola, and
HIV.” http://systemsandus.com/2014/08/02/epidemics/ (accessed December 4,
2014).
Zimmerman, R., Q. Zhu, and C. Dimitri. 2016. “Promoting Resilience for Food,
Energy, and Water Interdependencies.” J. Environ. Stud. Sci 6, no. 1, pp. 50–61.
Zhuang, Y. 2014. A System Dynamics Approach to Integrated Water and Energy
Resources Management. [Graduate Theses and Dissertations]. http://­
scholarcommons.usf.edu/etd/5164 (accessed November 20, 2017).
Index

A D
Agricultural land submodule, 104, Decision making
105–107 external environment, 164–165
Agro-food and land resources human dimension of, 158–159
management, 11 internal environment, 162–164
Archetypes, 27–33 intervention decision invariance,
165
B methods of, 159–162
Backward scenario planning, trade-offs, benefits, and
130–136 synergies, 166–168
Dynamic scenario analysis, 180
C Dynamic scenario planning,
Causal loop diagrams, 22–24, 128–130
32–33
Climate-land-energy-water E
strategy (CLEWS) framework, Energy balance submodule, 117
149 Energy demand submodule, 116,
Combining methods of analysis 118
cross-impact analysis, 44–45 Energy resources management,
from cross-impact to network 10–11
analysis, 48 Energy resources module
network analysis, 45–47 energy balance submodule, 117
Commercial land submodule, 104, energy demand submodule, 116,
108, 109 118
Community capacity assessment variables, parameters, and
submodule, 91–95, 199–200 multipliers, 208
Complexity vs. usefulness model, External environment, decision
72 making, 164–165
Constrained optimization,
151–153 F
Cross-impact analysis, 44–45 FEW/WELF nexus
Cross-impact balance (CIB) agro-food and land resources
analysis, 138–140 management, 11
212  •   Index

energy resources management, Intervention decision invariance,


10–11 165
global formulations, 3–8
hard systems approaches to, 3–13 L
other related frameworks, 12–13 Land resources module
sector-specific formulations, agricultural land submodule,
8–13 104, 105–107
water resources management, industrial/commercial land
9–10 submodule, 104, 108, 109
Food resources module, 95 land zoning submodule, 98, 99
variables, parameters, and overview of, 95, 98
multipliers, 201 residential–rural land submodule,
Formulating system-level 101–103
intervention dynamics soil/vegetation submodule, 98,
backward vs. forward scenario 100
planning, 130–136 system dynamics representation
cross-impact balance analysis, of, 96–97
138–140 variables, parameters, and
dynamic scenario planning, multipliers, 201–205
128–130 water demand submodule, 111,
scenarios formulation guidelines, 114
141–147 wild land submodule, 108, 110
stakeholder dynamics in scenario Land zoning submodule, 98, 99
planning, 136–138
Forward scenario planning, M
130–136 Model WELF-02, 73–76
Model WELF-03, 76–79
G Model WELF-04, 79–83
Groundwater, 111 Model WELF-05, 83–85
Multicriteria decision analysis,
H 150–151
Human dimension of decision Multiobjective decision analysis
making, 158–159 combining optimization and
system dynamics, 155–156
I constrained optimization,
IISD. See International Institute 151–153
for International Development examples of, 156–157
Industrial land submodule, 104, multicriteria decision analysis,
108, 109 150–151
Integrated approach, 177–181 multiobjective optimization,
Internal environment, decision 154–155
making, 162–164 single-objective optimization,
International Institute for 153–154
International Development Multiobjective optimization,
(IISD), 149 154–155
Index   •   213

N Synergy, in decision Making,


Network analysis, 45–47 166–168
System-and complexity-aware
P approach, 179–180
Participatory model building, System dynamics
49–52 archetypes, 27–33
Population and capacity module causal loop diagrams, 22–24,
community capacity assessment 32–33
submodule, 91–95 characteristics of, 18–21
population submodule, 87–91 combining methods of analysis,
Population submodule 44–48
demography, 87–89 components of, 21–22
employment, 91 cross-impact analysis, 44–45
health, 89 definition of, 18
variables, parameters, and hard systems approaches, 3–13
multipliers, 197–199 network analysis, 45–47
workforce, 91 overview of, 1–3
participatory model building,
Q 49–52
Qualitative system dynamics stock-and-flow diagrams, 24–27
modeling, 36–41 systems thinking, 14–18
Quantitative system dynamics System dynamics modeling
modeling, 41–44 complexity vs. usefulness model,
72
R modeling steps, 33–35
Reflection-before-action, 169 overview of, 67–73
Reflection-in-action, 169 qualitative system dynamics
Residential–rural land submodule, modeling, 36–41
101–103 quantitative system dynamics
modeling, 41–44
S setting expectations, 35–36
Scenarios formulation guidelines, System-level intervention
141–147 dynamics
Sensitivity analysis, 118–119 decision making, 158–168
Single-objective optimization, evaluating and selecting,
153–154 147–150
Soil/vegetation submodule, 98, goal-seeking dynamics, 126–128
100 from modeling to addressing
Soil water, 111 issues, 124–126
Spring water, 111 multiobjective decision analysis,
Stakeholder dynamics in scenario 150–157
planning, 136–138 from selecting to implementing,
Stock-and-flow diagrams, 24–27 168–169
Surface water, 111 Systems thinker, 14–16
214  •   Index

T Water resources management,


Trade-offs, in decision making, 9–10
166–168 Water resources module
variables, parameters, and
V multipliers, 205–207
Value proposition, 16–18 wastewater submodule, 114, 115,
116
W water balance submodule, 111,
Wastewater submodule, 114, 115, 112–113
116 WEF Nexus Tool 2.0, 148–149
Water balance submodule, 111, WELF-G model, 85–86
112–113 Wild land submodule, 108, 110

You might also like