The Water-Energy Nexus ‒ A Modern Case Study to Reassess Hydropower in the Niagara River

by

Samiha Tahseen

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Department of Civil Engineering University of Toronto

© Copyright by Samiha Tahseen 2017

ii

The Water-Energy Nexus ‒ A Modern Case Study to Reassess

Hydropower in the Niagara River

Samiha Tahseen

Doctor of Philosophy

Department of Civil Engineering University of Toronto

2017

Abstract The advent of variable renewable energy has created an urgent need for demand-based generation

and storage. At present, with batteries still awaiting a major technological breakthrough,

hydropower combined with pumped storage is suggested as a key response to demand variability.

Even overlooking the technical demands, developing this potential resource in a sustainable way

presents formidable challenges. While sustainability is a concern, the vulnerability of the resource

to changing climatic conditions poses a major threat. The present study proposes five modelling

approaches (and/or frameworks) as a foundation to a systems approach to hydropower and shows

how these tools address the key challenges. Overall, the research addresses the current demand for

dispatchable generation particularly in Ontario and proposes several alternatives including their

sustainability assessment.

Of the five models, the first two explore a variety of remuneration structures for pumped storage

in the context of Ontario. The work begins with an optimization approach that evaluates the

wholesale market for optimal profit. The tradeoff between hydropower and ecological targets is

explored using a Constraint Method. The results are compared with models based on contracted

iii

price and an integrated valuation approach that accounts for the socioeconomic attributes of

storage using representative applications.

While the first two approaches concentrate on the project economies, the third model evaluates the

potential for increased hydroelectric generation and assesses its vulnerability scenarios of climate

change. A 1D simulation model of the existing power system at Niagara is used to evaluate a

variety of innovative operating plans. One such scenario includes a revised approach to daily

operation with use of additional storage during the night and timed release during peak demand

hours.

The final section seeks to improve the existing frameworks for sustainability assessment and then

to use these improved metrics to evaluate various proposed generation options. The developed

decision support framework, applied to Niagara, allows quantitative evaluation based on survey

responses from key stakeholders. In contrast, the fifth and final approach uses the concept of

resilience within probabilistic graphical model to account for the inherent uncertainty associated

with climate projections.

Overall, the five approaches (optimization, integrated valuation, simulation, sustainability and

resilience assessment) facilitate rethinking the hydropower system with changing circumstances

and subsequent shift in priorities by the development, analysis, and interpretation of models. This

thesis contributes towards evaluating the merits of transitions between these approaches for future

modelling applications.

iv

Acknowledgments My PhD years has throughout been a learning experience for me, not only in the academic arena,

but also regarding personal growth and development. There are a number of people who have

supported, motivated and guided me throughout this journey. Here, I want to take the opportunity

to express my heartfelt thanks and my deepest appreciation for their profound contributions. First

and foremost, I would like to express my sincere gratitude to my supervisor Dr. Bryan Karney for

his unwavering support, patience, motivation, and mentorship throughout this work. While he

allowed incredible independence at every stage of the research, his engagement and enthusiasm

have made it a truly remarkable journey. While his breadth of knowledge has led to many

constructive discussions on my academic pursuits, his impeccable generosity and politeness have

offered me much to learn on a personal level. His meticulous revisions and unending efforts

towards details and clarity reflect on the studies presented herein, greatly improving their

articulation and organisation. Thank you for your support and counsel during the most difficult

period of my life. It has been an honor and a privilege working for you.

Besides my supervisor, there is one other person that I am greatly indebted to and i.e., Dr. Yu-Ling

Cheng. Her encouragement to apply for graduate studies, invaluable suggestions and last but not

the least introduction to my current supervisor have initiated my journey at University of Toronto.

While the opportunities of working with her have been a remarkable experience, my affiliation

with Centre for Global Engineering (CGEN), an institute directed by her, has allowed me to meet

some amazing peers and collaborators during my PhD years. In addition, I would also like to

extend my appreciation to my many tutors for their intriguing discussions, valuable guidance, and

insightful feedback. I further express my sincere gratitude to all my committee members for their

precious time in evaluating and subsequently raising valuable queries during the course of this

work. Their suggestions and feedback have greatly improved the quality of this thesis.

Finally, I must thank my parents for their unconditional love, support, motivation and the years of

sacrifices that provided the foundation for this work. I thank my mother for her unyielding care

and affection, her words of wisdom that keep me grounded and most importantly, being my

strength during the challenging times. I thank my sister for her companionship, and my friends for

being there for me. Lastly, I dedicate this thesis to my late father, Quazi Fariduddin. Without you

by my side, much of this achievement seems pointless; nonetheless I strive tirelessly to improve

v

myself hoping to make you proud. Thank you for being unconventional in your unconditional

support to my pursuits and dreams. Your perpetual belief in my limitless potential continues to

empower me every single day.

Table of Contents Acknowledgments.......................................................................................................................... iv

Table of Contents ........................................................................................................................... vi

List of Tables ................................................................................................................................ xii

List of Figures .............................................................................................................................. xiv

List of Appendices ..................................................................................................................... xviii

Introduction .................................................................................................................................1

1.1 Background ..........................................................................................................................1

1.2 Objectives ............................................................................................................................3

1.3 Research Overview and Organization .................................................................................5

1.4 Publications Related to Thesis .............................................................................................8

Section 1 A Review of Canada’s Hydropower Sector, Renewable Energy Policies and The Role of Storage ..........................................................................................................................11

Reviewing the Historical and Potential Contribution of Hydropower in Electricity Supply: The Ontario Case ......................................................................................................12

2.1 Background ........................................................................................................................12

2.2 Current Status of Hydropower in Ontario ..........................................................................14

2.2.1 Existing infrastructure and its development throughout the history ......................14

2.2.2 Hydropower for a low electricity rate ....................................................................18

2.2.3 Hydropower for low GHG emission ......................................................................18

2.2.4 Hydropower for renewable integration ..................................................................19

2.3 Potential Dispatchable Generation for Ontario ..................................................................19

2.3.1 Marmora pumped hydro project ............................................................................19

2.3.2 Increasing capacity at Niagara ...............................................................................20

2.4 Changing Policy and its Impact on Hydropower Development ........................................20

2.5 Conclusions and Policy Implications .................................................................................23

vii

Exploring the Multifaceted Role of Pumped Storage at Niagara ..............................................25

3.1 Background ........................................................................................................................25

3.2 Brief Literature Review .....................................................................................................28

3.3 Optimization Model Development ....................................................................................30

3.3.1 Formulating the context-specific optimization model ...........................................30

3.3.2 Analyzing input price data .....................................................................................34

3.4 Model Explorations ............................................................................................................35

3.4.1 Analysis of profit characteristics on a monthly basis ............................................35

3.4.2 Analysis of profit characteristics on a weekday-holiday basis ..............................36

3.4.3 Profit sensitivity to cycle length ............................................................................37

3.4.4 Evaluating potential improvement opportunities for SAB PGS ............................38

3.4.5 Profit sensitivity to energy price ............................................................................39

3.4.6 Trade-offs between power generation and scenic flow restrictions.......................40

3.5 Benefits and Possible Challenges for Pumped Storage .....................................................41

3.6 Conclusions and Recommendations ..................................................................................43

Assessing the Financial Incentives for Pumped Storage Development ....................................44

4.1 Introduction ........................................................................................................................44

4.2 Combined Wind and Pumped Storage System ..................................................................47

4.3 Methodology ......................................................................................................................48

4.3.1 Marginal cost based operation in the spot market .................................................49

4.3.2 Contracted fixed price per unit of electricity .........................................................50

4.3.3 Socioeconomic model ............................................................................................52

4.4 Analysis and Results ..........................................................................................................55

4.4.1 Marginal cost based operation in the spot market .................................................55

4.4.2 Contracted fixed price per unit of PHS electricity .................................................58

4.4.3 FIT with guarantees of origin (FIT_GO) ...............................................................60

viii

4.4.4 Socioeconomic cost-benefit model ........................................................................62

4.4.5 Model comparison and sensitivity analysis ...........................................................64

4.5 Limitation ...........................................................................................................................67

4.6 Conclusion .........................................................................................................................67

Section 2 Increased Hydropower Potential at Niagara: A scenario-based analysis ..............69

A Simulation Model on The Impact of The 1950 Treaty on the Generation Potential at Niagara ......................................................................................................................................70

5.1 Introduction ........................................................................................................................70

5.2 Study Area .........................................................................................................................71

5.3 Model Development...........................................................................................................73

5.3.1 Layout of key hydraulic components .....................................................................74

5.3.2 Operational characteristics .....................................................................................75

5.4 Model Calibration and Validation .....................................................................................77

5.5 Considering Additional Diversion for Enhanced Hydropower .........................................79

5.6 Critical Appraisal ...............................................................................................................82

5.7 Conclusion .........................................................................................................................83

Power Systems Vulnerability to Climate Change: An Analysis on the Niagara Hydropower System ..................................................................................................................84

6.1 Introduction ........................................................................................................................84

6.2 Study Area .........................................................................................................................87

6.3 Model and Scenario Development .....................................................................................88

6.3.1 Niagara River simulation .......................................................................................88

6.3.2 Potential scenarios .................................................................................................90

6.4 Results and Discussion ......................................................................................................95

6.4.1 Climate change impact on hydropower potential ..................................................95

6.4.2 Impact of lake storage on hydropower potential ....................................................99

6.4.3 Combined climate change and lake storage scenarios .........................................100

ix

6.5 Limitation .........................................................................................................................101

6.6 Conclusion and Discussion ..............................................................................................102

Section 3 Sustainability and Resilience Assessment of Hydropower Systems .....................104

Reviewing and Critiquing Published Approaches to the Sustainability Assessment of Hydropower.............................................................................................................................105

7.1 Introduction ......................................................................................................................105

7.2 Synergy Between Hydropower and Sustainability ..........................................................107

7.3 Existing Frameworks and Guidelines on Sustainable Development of Hydropower ......109

7.3.1 Low Impact Certification by LIHI .......................................................................110

7.3.2 Green Hydropower Certification by EAWAG ....................................................110

7.3.3 Hydropower Sustainability Assessment Protocol (HSAP) by IHA .....................111

7.3.4 Directions in Hydropower by World Bank ..........................................................112

7.3.5 Hydropower Implementing Agreement by IEA ..................................................113

7.3.6 Sustainable Energy Financing by EBRD .............................................................114

7.4 Selective Review of Sustainability Indicators .................................................................115

7.5 Limitations of the Existing Approaches ..........................................................................121

7.6 Conclusion .......................................................................................................................122

Opportunities for Increased Hydropower Diversion at Niagara: An sSWOT Analysis .........124

8.1 Introduction ......................................................................................................................124

8.2 Model Development.........................................................................................................126

8.3 Application of the sSWOT Model on Niagara ................................................................131

8.3.1 Identification of the sSWOT factors ....................................................................131

8.3.2 Application of the AHP and ANP ........................................................................142

8.4 Model Validation .............................................................................................................148

8.5 Conclusion .......................................................................................................................149

A Bayesian Evaluation of Reliability, Resiliency and Vulnerability of the Great Lakes to Climate Change .......................................................................................................................151

x

9.1 Introduction ......................................................................................................................151

9.2 Sustainability and Resilience ...........................................................................................153

9.3 Study Area .......................................................................................................................154

9.4 Methodology ....................................................................................................................155

9.4.1 Model preliminaries .............................................................................................156

9.4.2 Data management.................................................................................................156

9.4.3 BN structure .........................................................................................................159

9.4.4 Parameter learning ...............................................................................................160

9.4.5 Structural validation .............................................................................................161

9.4.6 Dynamic reliability, resilience and vulnerability .................................................163

9.5 Analysis of the Historical Data ........................................................................................164

9.6 Reliability .........................................................................................................................165

9.6.1 Critical level between 10th, 25th and 75th percentile values .................................166

9.6.2 Critical level between median and 90th percentile values ....................................167

9.6.3 Seasonal high and lows (spring and winter) ........................................................167

9.6.4 Flow conditions ....................................................................................................168

9.7 Resilience .........................................................................................................................169

9.7.1 Critical level between 10th, 25th and 75th percentile values .................................169

9.7.2 Critical level between median and 90th percentile values ....................................170

9.7.3 Seasonal high and lows ........................................................................................170

9.7.4 Flow conditions ....................................................................................................171

9.8 Vulnerability ....................................................................................................................171

9.8.1 Critical level between 10th, 25th and 75th percentile values .................................171

9.8.2 Critical level between median and 90th percentile values ....................................172

9.8.3 Seasonal high and lows ........................................................................................172

9.8.4 Flow conditions ....................................................................................................173

xi

9.9 Sensitivity Analysis .........................................................................................................173

9.10 Conclusion .......................................................................................................................174

Conclusions and Recommendations .......................................................................................176

10.1 Future Research ...............................................................................................................179

References ....................................................................................................................................181

Appendices A. Overview of Canada’s Electricity Sector ............................................................225

Appendices B. List of the Operating Rules Under Each Reservoir .............................................231

Appendices C. Pairwise Comparison Matrices for SWOT Sub-Factors Local Priorities ............233

Appendices D. Pairwise Comparison Matrices for the Priorities of the Alternative Strategies Based on the SWOT Sub-Factors ...........................................................................................237

Appendices E. The Bayesian Network Model .............................................................................243

Copyright Acknowledgements (if any) ........................................................................................245

xii

List of Tables Table 2.1: Prices per kWh under FIT 5.0 (2017) .......................................................................... 21

Table 4.1: Summary of input parameters ...................................................................................... 49

Table 4.2: Comparison among the pricing models ....................................................................... 65

Table 5.1: Existing hydropower infrastructure at Niagara ............................................................ 73

Table 5.2: Percentage error and standard deviation between simulated and overserved elevation

(2009) ............................................................................................................................................ 78

Table 6.1: Existing hydropower infrastructure at Niagara ............................................................ 87

Table 6.2: Required data for model simulation (Tahseen and Karney 2017) ............................... 89

Table 6.3: Simulated changes in the Great Lakes hydrology for 2°C rise in global temperature

under various climate scenarios .................................................................................................... 93

Table 7.1: Hydropower Sustainability Assessment Protocol topics (IHA 2010) ....................... 112

Table 7.2: List of environmental and social indicators under IEA framework (IEA 2000) ....... 114

Table 7.3: Environmental criteria for hydropower projects under EBRD (EBRD 2013) .......... 115

Table 7.4: List of hydropower sustainability indicators reported by researchers ....................... 118

Table 8.1: Saaty’s 1-9 scale for Analytical Hierarchical Process (AHP) preference (Saaty 1996)

..................................................................................................................................................... 129

Table 8.2: Random Consistency Index value (Saaty 1980) ........................................................ 130

Table 8.3: Pairwise comparison of SWOT factors by assuming there is no dependence ........... 144

Table 8.4: The inner dependence matrix of the SWOT factors with respect to strengths .......... 145

Table 8.5: The inner dependence matrix of the SWOT factors with respect to weaknesses ...... 145

Table 8.6: The inner dependence matrix of the SWOT factors with respect to threats .............. 145

xiii

Table 8.7: Pairwise comparison among sustainability parameters ............................................. 146

Table 8.8: Priority of the SWOT sub-factors .............................................................................. 147

Table 9.1: GCM-simulated temperature increase for the Great Lakes-St. Lawrence Basin:

Change from 2xCO2 to 1xCO2 .................................................................................................... 158

Table 9.2: GCM precipitation ratios for the Great Lakes-St. Lawrence Basin. (2xCO2 to 1xCO2)

..................................................................................................................................................... 158

Table 9.3: Reliability considering 10th, 25th and 50th (lower boundary), 75th and 90th (upper

boundary) percentile flows ......................................................................................................... 169

Table 9.4: Comparison between uniform and distribution-based discretization model under OSU

scenario ....................................................................................................................................... 174

Table A.1: Total electricity generation by provinces in 2013 (TWh) ......................................... 226

Table A.2: Ownership distribution (%) over generation assets in 2009 ..................................... 227

Table A.3: Existing SHP capacity in Canada (MW) .................................................................. 229

xiv

List of Figures Figure 2.1: Distribution of hydroelectric plants in Ontario based on plant size ........................... 15

Figure 2.2: Hydroelectric asset distribution on the basis of plant size and ownership ................. 16

Figure 2.3: Cumulative capacity contribution from large, medium and small stations based on

their year of commission ............................................................................................................... 18

Figure 3.1: Niagara Hydroelectric plants system .......................................................................... 27

Figure 3.2: Comparison between available power flow and the maximum capacity at SAB

Complex ........................................................................................................................................ 32

Figure 3.3: Estimation of monthly profit for PHS with and without flow restriction .................. 36

Figure 3.4: Variation in profit for PHS during weekdays and holidays ....................................... 37

Figure 3.5: Comparative analysis of economic return versus running time for pumped storage . 38

Figure 3.6: Impact of changing electricity rates on pumped storage profit .................................. 40

Figure 3.7: The trade-off surface between the economic gain and the environmental

consideration for pumped storage in July ..................................................................................... 41

Figure 4.1: The electricity price (HOEP) duration curves for four consecutive years (2012–2015)

....................................................................................................................................................... 56

Figure 4.2: Monthly profit (excluding capital costs) based on spot and ancillary service-based

(with 500 $/MW) market operation .............................................................................................. 57

Figure 4.3: Discounted payback period under the wholesale and ancillary service market (at CF =

15% and i = 5%) ........................................................................................................................... 58

Figure 4.4: Payback period with changing capacity factors considering contract price, on-peak

premium and different pumping costs .......................................................................................... 59

Figure 4.5: Payback period with varying compensation rates to the wind operators ................... 59

xv

Figure 4.6: Payback period with varying interest rates ................................................................. 60

Figure 4.7: FIT rates (i = 5%) under varying (a) CFs and wind contributions (15 yr return year)

(b) return period and wind contributions (25% capacity factor) ................................................... 61

Figure 4.8: Comparison between BEFITs and electricity production cost when replacing natural

gas-fired CC plants ....................................................................................................................... 63

Figure 4.9: Comparison between BEFITs and electricity production cost when replacing oil-fired

plants ............................................................................................................................................. 63

Figure 4.10: Comparison between BEFITs and electricity production cost when replacing oil-

fired plants .................................................................................................................................... 64

Figure 4.11: Tornado diagram showing impact of listed factors (on the left) on return period ... 66

Figure 4.12: Impact of socioeconomic factors on return period ................................................... 66

Figure 5.1: The Niagara River connecting Lake Erie and Lake Ontario ...................................... 72

Figure 5.2: Zoning of Lake Erie based on historical lake level data ............................................ 75

Figure 5.3: Comparison between simulated and observed water surface elevation at Ashland

Ave. in 2010 .................................................................................................................................. 78

Figure 5.4: Monthly variation in available power flow at the Niagara Complex under the baseline

and increased diversion scenarios ................................................................................................. 81

Figure 5.5: Increase in monthly power generation at the SAB Complex under the baseline and

increased diversion scenarios ........................................................................................................ 82

Figure 6.1: A systems for the Niagara River connecting Lake Erie and Lake Ontario ................ 88

Figure 6.2: Hourly variation in water level at Ashland Avenue gauge (September, 2010) .......... 90

Figure 6.3: Lake (a) Erie and (b) Ontario elevation under climate scenarios for 2050‒2060 ...... 96

xvi

Figure 6.4: Combined monthly available discharge at Niagara Hydropower Plant for 2050‒2060

....................................................................................................................................................... 97

Figure 6.5: Monthly power generation at SAB Complex by 2050-2060 under various climate

scenarios ........................................................................................................................................ 98

Figure 6.6: Available discharge at the SAB Complex under the flow regulation scenarios ........ 99

Figure 6.7: Hourly variation in power generation with lake regulation ..................................... 100

Figure 6.8: Hourly generation under the baseline, the climate and the combined climate and lake

storage (0.3 cm) scenario for July ............................................................................................... 101

Figure 8.1: Step-by-step evaluation process for the sSWOT model ........................................... 127

Figure 8.2: The hierarchy (a) and the network (b) representation of the sSWOT model. While (b)

allows interdependencies among SWOT factors, (a) permits downward influence only (Yüksel

and Dagˇdeviren 2007) ............................................................................................................... 128

Figure 8.3: The sSWOT model for Niagara................................................................................ 136

Figure 8.4: An example of a pairwise comparison of factors presented under the SWOT category

“Opportunity”. The respondent is asked to assign a value from 1 to 9 to one of the factors to

indicate the relative importance of that factor over another. ...................................................... 143

Figure 8.5: Inner dependence among SWOT factors (Yüksel and Dagˇdeviren 2007) .............. 144

Figure 9.1: The Niagara River basin ........................................................................................... 155

Figure 9.2: Variable discretization within the BN model ........................................................... 157

Figure 9.3: Probability ranges with different instantiations at Buffalo ...................................... 162

Figure 9.4: Step-by-step procedure for the BN model development .......................................... 162

Figure 9.5: Flow classification for the Niagara River (1950–2011) ........................................... 165

xvii

Figure 9.6: Reliability under the baseline and future climate scenarios when comparing (a) the

upper limits - 75th percentile, (b) the lower limits - 25th percentile and (c) 10th percentile ..... 167

Figure 9.7: Changing resilience under the climate scenarios considering winter highs and lows

..................................................................................................................................................... 171

Figure A.1: Installed electricity capacity by source (GW) ......................................................... 225

Figure A.2: Small hydropower capacities 2013-2016 in Canada (MW) .................................... 228

Figure E.1: Bayesian network for measuring reliability, resilience and vulnerability for Niagara

River…………………………………………………………………………………………….241

xviii

List of Appendices Appendices A. Overview of Canada’s Electricity Sector ............................................................225

A.1 Electricity Sector Overview .............................................................................................225

A.2 Small Hydropower Sector Overview and Potential .........................................................227

A.3 Renewable energy policy .................................................................................................229

A.3.1 Clean energy fund .............................................................................................229

A.3.2 Standard offer programs....................................................................................229

A.3.3 Feed-in Tariff (FIT) programs ..........................................................................229

A.3.4 Net metering ......................................................................................................229

A.3.5 Requests for proposal ........................................................................................230

A.4 Barriers to Small Hydropower Development ...................................................................230

Appendices B. List of the Operating Rules Under Each Reservoir .............................................231

Appendices C. Pairwise Comparison Matrices for SWOT Sub-Factors Local Priorities ............233

Appendices D. Pairwise Comparison Matrices for the Priorities of the Alternative Strategies Based on the SWOT Sub-Factors ...........................................................................................237

Appendices E. The Bayesian Network Model .............................................................................243

1

Introduction

1.1 Background While renewable resources are largely thought of as replacing carbon-based electrical energy, a

major concern lies with their intermittent nature. With the mounting integration of intermittent

renewables such as wind and solar, the ability to respond quickly to demand variability is now one

of the most sought-after grid attributes. At present, particularly with storage using batteries still

awaiting major technological breakthrough, hydropower is a key renewable source that offers an

effective means of permitting demand variability. Despite the perceived technical demand,

developing this potential resource in a sustainable manner offers a considerable challenge.

After its peak growth rate at the beginning of 1900, hydropower in Ontario has largely suffered

from a ‘been there, done that’ mindset where the general consensus is that the capacity for

economically feasible generation has reached a plateau. Although, the past few decades have seen

little research on sustainable development of the existing resources, with the realization of climate

change and the subsequent commitment to reduce emission, hydropower that previously lost its

eminence to nuclear and coal-based generation has become more important again. At present,

being the only commercially-proven, utility-scale storage technology, hydropower combined with

pumped hydroelectric storage (PHS) is suggested as a key response to demand variability (Rehman

et al. 2015). Hydro’s spinning reserve, quick-start and black-start capability provide flexibility and

protection to the overall grid (Zhang et al. 2015). Since its raw power comes from a renewable

source, it is able to reduce the electrical system’s reliance on fossil fuel. In Ontario, the future

decommissioning of the nuclear facilities, and the growing penetration of intermittent renewables

(wind and solar) lead to an increased demand for clean, dispatchable generation. At present, the

Feed-in Tariff (FIT) Program is expected to quadruple wind capacity by 2018 (Canadian Wind

Energy Association 2011), posing significant challenges in dealing with demand fluctuations. The

phase out of coal-fired electricity in 2014 (Miller and Carriveau 2017) and a proposed reduction

in natural gas usage by 2017 (Ontario Ministry of Energy 2009) elevate the role of hydropower in

Ontario as the most plausible dispatchable generation. In this context, the appraisal of appropriate

incentives for hydropower development raises all sorts of fascinating questions. Should tariffs be

based on marginal sites or the best resource sites? With the most suitable places already exploited

or far from the load centers, does it make sense to offer higher tariffs for capacity building at low

2

resource sites? If not, how does that impact Ontario’s commitment to increase its hydro capacity

from the current 8,400 MW to 9,000 MW by 2030 as mandated by Ontario’s Long Term Energy

Plan (Ontario Ministry of Energy 2010)? And is a greater price for energy politically feasible?

Despite its many benefits, hydropower is often criticized for typically high development costs and

its sometimes considerable environmental impacts. There also exists a continued conflict among

researchers regarding the profitability of PHS systems (German Advisory Council 2011;

Ingebretsen and Johansen 2014). Furthermore, the vulnerability of water resources under a

changing climate is considered a major challenge for hydropower generation interests (Vliet et al.

2016). Since hydropower taps into the energy of running water, issues like rising temperatures and

changing precipitation patterns may affect generation, while energy demands continue to increase

with economic development and a growing world population. While rehabilitation of ageing

infrastructure and construction of new components are widely viewed as solutions to water

shortage, it is easy to argue that sustainable use of these resources will require understanding and

broadening the current boundaries of water resources management.

A key tension associated with hydropower development is its impact on environmental and social

parameters in the form of biodiversity loss, disruptions to fish migration, potential land inundation,

human resettlement, and many others. To limit such impacts, detailed guidelines are now

prescribed as well as observation and experts’ opinion during project implementation are

documented. However, due to the variety of approaches, the list and priority of proposed indicators

varies between numerous published guidelines, scientific studies, and reports (Bakis 2004; IEA

2000; IHA 2010; Klimpt 2002; Rosso 2014; Supriyasilp 2009). Although there is an obvious

overlap in the obligatory range of considerations that must be pondered, there is as yet no

universally accepted standard for assessing sustainability of hydropower projects. Additionally,

several authors have adopted various approaches to define and assess power systems resilience in

terms of risk and vulnerability analysis (Maliszewski and Perrings 2012; Molyneaux et al. 2012).

The majority of these studies focus on structural resilience (withstanding a disruption to

distribution and transmission) as opposed to incremental changes in systems resilience

(withstanding then recovering) which is an equally important consideration for hydroelectric

systems.

3

The modelling approaches proposed in this thesis seek to address the abovementioned issues

currently faced by hydropower systems by mapping and better understanding the interaction

between the environmental, social, and economic spheres. The first section reviews the current

hydropower situation. The tools presented therein explore the payment options for water projects

in Ontario and run a comparative analysis based on capacity factors, contract price, avoided

externalities, wind energy exploitation and so on. The analysis here presents the perspective of

storage operators guiding their decisions to contract fixed payments or to act through price

arbitrage and reserve provision. Section 2 concentrates on evaluating the potential for increased

generation from the existing reservoir systems and assessing their vulnerability under the projected

climate. Here, the study sets up a comprehensive 1D simulation model for the existing power

system that, once validated, is used to assess innovative operating plans. The final section improves

the existing framework for sustainability and resilience assessment and then uses these improved

metrics to evaluate several proposed generation options. Overall, the approaches facilitate

rethinking the system with changing circumstances and subsequent shift in priorities by the

development, analysis, and interpretation of models.

1.2 Objectives The primary objective of this research is to evaluate the potential for increased hydropower

generation at Niagara through an integrated management approach. The specific objectives are as

follows:

1. Systems planning and evaluation should be preceded by acquiring extensive knowledge

about the system that is currently in place. The following chapter (chapter 2) seeks to

understand the contribution of hydropower at the current grid level and its future potential.

It provides a narrative for hydroelectric development in Ontario in the backdrop of

historical events and major energy transitions. The discussion also analyzes the evolving

energy policy landscape for its incentives towards hydroelectric development.

2. Despite the perceived technical demands, profitability remains a major obstacle for PHS

system as the current literature reports conflicted findings on their cost-effectiveness.

Chapter 3 illustrates a direct optimization approach to assess the Ontario wholesale

4

electricity market for pumped storage operations. The analysis seeks to inform the

discussion concerning the expected variation in profit due to seasons, long holidays,

changing reservoir size and cycle lengths. Since almost all real decision-making problems

are multi-objective in nature involving trade-offs among conflicting intentions, the

trajectory of these trade-offs and the subsequent shift in priorities are explored and

discussed.

3. The lack of progress in storage deployment is largely attributed to the absence of an

integrated valuation framework that justly and effectively rewards storage operators for the

range of services they provide to the grid. Chapter 4 examines the various financial

mechanisms designed to support PHS development in Ontario. As the published

approaches apply the pricing strategies on an individual basis, there have been a few

instances of comprehensive assessment of their merits. To this end, the author conducts a

comparative analysis among the alternative pricing structures with the objective to

determine the appropriate remuneration for pumped storage.

4. Considering that the world’s water resources face increasing challenges within the context

of both a growing population and a changing climate, there is a need to reassess the

resulting impact on water usage, ecology, energy and environment. With the realization

that hydropower provides an effective solution for reducing emission from the power

sector, chapter 5 and 6 seek to explore the potential for incremental generation from the

existing reservoir systems. Despite the extensive research on climate change impacts on

water resources, previous studies have provided limited attention on hydropower systems’

vulnerability to a changing pattern of runoff. The analysis here seeks to assess climate

change impact on hydropower generation potential, particularly that within the Great Lakes

system.

5. Energy systems are not only intrinsically interesting, since power is itself of importance,

but raise fascinating trade-offs with other areas as well, particularly since there are almost

invariably technical, economic and ecological dimensions to these considerations. Many

previous studies, however, set the system boundaries such that they omit some of the

relevant factors and neglect some key benefits offered by hydropower. These observations

motivate the development of a decision-support framework, elaborated in Chapter 8, that

5

explicitly incorporate sustainability and applies it to the strategic planning for the overall

development of a resource system.

6. Relatively steady hydrologic regimes are essential to the stability of river basins. While

historically anthropogenic activities were major drivers for river systems alteration, climate

change and its associated impacts could trigger important changes in watersheds. While

existing research focuses on characterizing flow regime alterations driven by climate

change, few approaches trace the resulting impacts on watershed resilience. Chapter 9

seeks to define a systematic approach to recognize the changing river basin resilience

through the application and development of a probabilistic graphical model.

1.3 Research Overview and Organization The research begins with an overview of Canada’s electricity sector and highlights the contribution

from various resources to the provincial grid. It summarizes the types of electricity market

(regulated or deregulated), variations in tariffs, and the diverse renewable energy policies enacted

by the provinces. The study then focuses exclusively on Ontario where it builds an inventory of

the existing hydropower fleet. The aggregated information is then analyzed for capacity-based

classification, annual generation, ownership distribution, age and the like. As the analysis points

towards a growing need for dispatchable generation in the province, the discussion concludes with

recommendations for increased hydropower capacity through pumped storage development and

incremental generation from the existing systems. The need for strong financial support and clear

regulatory framework for storage development motivate the formulation of an optimization model

that evaluates the Ontario spot market for pumped hydro operations. The model addresses the gap

in the literature by investigating the profit characteristics of PHS under varying operating and

market conditions and assessing the tradeoffs between power and environmental considerations.

While the aforementioned approach assumes marginal cost-based operations, the next chapter

compares among the pricing strategies including contracted fixed price, Feed-in Tariff (FIT) with

guarantees of origin, and finally a socioeconomic cost-benefit model that accounts for the social

costs and benefits of storage.

6

The second alternative, that of operating plans for incremental generation from the existing

reservoir systems, is assessed through systems modelling. One of the most crucial of the

hydropower sources for Ontario is still Niagara, which apart from hosting the only existing pumped

storage station, is complicated (and enriched) by being closely associated with a major tourist

draw. Recognizing the complex interactions among Niagara’s economy, environment, energy, and

policy, the site is chosen for the exploration. The study develops the Niagara Power Systems

Simulation (NPSS) model, ensuring adherence to the current regulatory regimes while providing

users the freedom to refine these values. The research makes a significant contribution using the

NPSS model for scenario-based explorations of planning options; the alternatives are subsequently

evaluated in terms of their contribution to increased generation and relative to potential climate

change impacts. An interesting possibility that the model explores is a revised daily management

with additional storage during the night and timed release during peak demand hours.

While the NPSS Model is used for technical analysis of the proposed generation options, the

economic, environmental and social aspects associated with these alternatives mandate equal

considerations. Here the research addresses the limitations of the published approaches that use

restricted system boundaries in assessing sustainability of hydropower projects. It elaborates the

existing decision support framework by explicitly incorporating sustainability and applies it for

assessing the increased generation options at Niagara. The new framework, called sustainability

SWOT (sSWOT), adopts a rather comprehensive approach that includes considerations of tourism,

navigation, avoided greenhouse gas (GHG) emission, resettlement, and so on. Following this, the

research moves to a more general formulation using System Integrity Evaluation (SIE) model that

based on the projections for climate variables predicts the resulting impact on hydrologic

conditions and translates the outcomes in terms of reliability, resilience and vulnerability of the

system. Finally, the concluding chapter synthesizes the results from the modelling applications to

develop generalized relationships among water-energy-economy and explores the benefits of

transitioning between these approaches.

The structure of the thesis is shaped by the writing and preparation of reports, conference and

journal papers. In particular, chapters 3 to 9, each relate to specific modelling applications, are

based on published, accepted, submitted or soon to-be-submitted manuscripts to various

established journals. Nevertheless, the connections between these are described throughout the

7

thesis. The thesis is categorized into three major sections with varying number of chapters under

each. Literature related to specific themes, pertinent to different modelling techniques, are

reviewed separately in each chapter. There is inevitably a small degree of overlap between the

chapters. The introductory section (Section 1) contains three chapters (2‒4) of which the foremost

informs about the existing systems and policies in place while the rest of each details a specific

modelling application pertaining to storage development. It begins with highlighting the

fragmented approach in almost all aspects of energy due to the provinces’ diverse resources and

electricity policies. The work is currently published as a regional contribution to “World Small

Hydro Development Report 2016” by International Center on Small Hydro Power (ICSHP) and

The United Nations Industrial Development Organization (UNIDO). Chapter 2 outlines Ontario’s

current regulatory framework for hydro development. While the discussion proposes alternatives

for increased generation, the approach towards achieving those are addressed in the following

chapters. Chapter 3 illustrates the interplay between economic and ecological dimensions of

hydropower systems through multi-objective optimization, followed by a comparative analysis

among the pricing mechanisms in the subsequent chapter. Chapter 3 is based on the manuscript

entitled “Exploring the Multifaceted Role of Pumped Storage at Niagara”, published in the Journal

of Water Resources Planning and Management and reproduced herein with permission from

ASCE. The final chapter in this section is soon to be submitted to an appropriate journal.

The second section elaborates the development and the potential applications of the NPSS Model.

Chapter 5 is based on the manuscript entitled “Increased Hydropower Potential at Niagara: A

Scenario Based Analysis”, submitted to the Journal of Water Resources Management. With the

current generation asset running below capacity and the potential to extend the existing

transmission at Niagara, the study here explores an increased power diversion scenario. The

subsequent chapter entitled “Power Systems Vulnerability to Climate Change: An Analysis on the

Niagara Hydropower System” extends the work in chapter 5 and is currently awaiting publication.

As implied by the title, the paper assesses the generation potentials at Niagara under changing

climate conditions.

The final theme of the thesis is sustainability and resilience assessment and comprises

contributions from three different chapters. It begins with a comprehensive literature review on

sustainability assessment of hydropower. The work is elaborated in chapter 7 under the title

8

“Reviewing and Critiquing Published Approaches to the Sustainability Assessment of

Hydropower”, published in the Journal of Renewable & Sustainable Energy Reviews. Chapter 8

is based on the paper entitled “Opportunities for Increased Hydropower Diversion at Niagara: An

sSWOT Analysis”, published in the Journal of Renewable Energy. It addresses the limitations of

the published approaches by extending the system boundaries to include environmental benefits

of avoided emission and externalities into the analysis. The final chapter in this section elaborates

the development and potential application of System Integrity Evaluation (SIE) model, a novel

risk assessment tool specifically designed to address the uncertainty of environmental and climate

projections. It is currently waiting to be submitted to a suitable journal.

Finally, chapter 10, the last chapter of the thesis, summarizes the contributions of the present

research, and discusses the potential improvements of the developed models, as well as possible

extensions to the work. The overall structure and layout are shown in the Table below.

1.4 Publications Related to Thesis As previously mentioned, the contributions of this research have been disseminated in published

format. The published works are listed below in chronological order.

Tahseen, S. Karney, B. (2015). Integrated sustainability strategies for the Great Lakes region. Engineering Dimensions, 36: 38-39. (Source article for Chapter 2)

Karney, B., Mandair, S. and Tahseen, S. (2016). World Small Hydro Development Report 2016.

International Center on Small Hydro Power (ICSHP) and The United Nations Industrial Development Organization (UNIDO). (Source report for Appendix A)

Tahseen, S. and Karney, B. W. (2016). Exploring the multifaceted role of pumped storage at

Niagara. Journal of Water Resource Planning and Management, 10.1061/(ASCE)WR.1943-5452.0000666, 05016007. (Source paper for Chapter 3)

Tahseen, S. and Karney, B. W. (2017). Reviewing and critiquing published approaches to the

sustainability assessment of hydropower. Journal of Renewable and Sustainable Energy Reviews, 67: 225-234. (Source paper for Chapter 7)

Tahseen, S. and Karney, B. W. (2017). Opportunities for increased hydropower diversion at

Niagara: An sSWOT analysis. Journal of Renewable Energy, 101: 757-770. (Source paper for Chapter 8)

9

Tahseen, S., Karney, B. W. and Drake, J. (2017). Increased hydropower potential at Niagara: A scenario based analysis. Journal of Water Resource Planning and Management, Submitted on 20.07.2016. (Source paper for Chapter 5)

Tahseen, S., and Karney, B. W. (2017). Power Systems Vulnerability to Climate Change: An

Analysis on the Niagara Hydropower System. Journal of Water Resource Planning and Management, Waiting for publication of the preceding work. (Source paper for Chapter 6)

Tahseen, S. and Karney, B. W. (2017). Reviewing the Historical and Potential Contribution of

Hydropower in Electricity Supply: The Ontario Case. Energy Policy, To be submitted. (Source paper for Chapter 2)

Tahseen, S. and Karney, B. W. (2017). A Bayesian Evaluation of Reliability, Resiliency and

Vulnerability of the Great Lakes to Climate Change. (Source paper for Chapter 9) Tahseen, S. and Karney, B. W. (2017). Assessing the Financial Incentives for Promoting Pumped

Storage Development. In preparation. (Source paper for Chapter 4)

I wrote the papers listed above, developed the model and performed the analysis presented in them.

The co-author in all cases is my PhD thesis supervisor, Prof. Bryan Karney with occasional

contributions from other professors (chapter 5) and peers (Appendix A). The co-authors primarily

critiqued ideas and insights as well as proofread and edited the manuscripts before submission.

The only exception is Appendix A, which is of secondary importance to this thesis, where the

exploration was jointly performed with equal contributions from other authors. I have received

permission and endorsement from them to include in this thesis all materials listed above.

10

Theme Paper/Chapter Title Current status Chapter 1: Background, Research Objectives and Thesis Organization

Section 1: A Review of Canada’s Renewable Energy Policies and the Role of Storage

Chapter 2 Overview of Canada’s electricity sector The current status of hydropower and Feed-in-Tariff (FIT) in Ontario

World Small Hydro Development Report 2016

Reviewing the Historical and Potential Contribution of Hydropower in Electricity Supply: The Ontario Case

Published

Almost ready for Submission

Chapter 3 Assessing Ontario spot market for pumped storage operation

Exploring the Multifaceted Role of Pumped Storage at Niagara Published

Chapter 4 Comparative analysis among the pricing strategies for storage

Assessing the Financial Incentives for Pumped Storage Development

In preparation

Section 2: Increased Hydropower Potential At Niagara: A Scenario-Based Analysis

Chapter 5 Development of the Niagara Power Systems Simulation (NPSS) Model and its application

A Simulation Model on The Impact of The 1950 Treaty on the Generation Potential at Niagara

Under review

Chapter 6 Climate change impact on the power systems

Power Systems Vulnerability to Climate Change: An Analysis on the Niagara Hydropower System

Waiting for Chapter 5 to be published

Section 3: Sustainability and Resilience Assessment of Hydropower Systems

Chapter 7 Brief literature review of existing sustainability assessment frameworks

Reviewing and Critiquing Published Approaches to the Sustainability Assessment of Hydropower

Published

Chapter 8 The development and application of sSWOT model for sustainable resource management

Opportunities for Increased Hydropower Diversion at Niagara: An sSWOT Analysis

Published

Chapter 9 Risk assessment using System Integrity Evaluation (SIE) model

A Bayesian Evaluation of Reliability, Resiliency and Vulnerability of the Great Lakes to Climate Change

Almost ready for Submission

Chapter 10 Synthesis of model applications Conclusions and Recommendations

11

Section 1 A Review of Canada’s Hydropower Sector, Renewable Energy

Policies and The Role of Storage Section 1 sets the context for the overall work by reviewing the current energy situation. The quest

for a carbon-free, reliable grid should begin with understanding the existing system in place to be

able to identify the potential alternatives, so as to direct analysis. The section presents an overview

of Canada’s diverse, provincially owned power sector, the electricity market operations and the

major renewable energy policies enacted by the provinces. Part of this work, published as a

contribution to “World Small Hydro Development Report 2016” by International Center on Small

Hydro Power (ICSHP) and The United Nations Industrial Development Organization (UNIDO),

recognizes a substantial potential for small hydropower (limited to 50 MW capacity) development

in Canada and identifies barriers to its development. Considering the secondary importance of this

Canada-wide discussion to the thesis, the detailed report is provided in the Appendix A.

Chapter 2 specifically highlights the potential and the need for dispatchable hydro in Ontario

followed by an exploration on different pricing options in the subsequent chapters. While currently

there exists a ‘been there, done that’ mindset towards hydro in Ontario where the general consensus

is that the capacity for economically viable developments has plateaued, the study here shows that

the current electricity market and its policies lack appropriate remuneration structures that justly

reward storage for the range of services it can provide to the grid. Thus, the final chapter in this

section (chapter 4) introduces an integrated valuation framework that accounts for the benefits

from replacing peak power plants, avoided GHG emission and the negative externalities. Based

on the analysis, a capacity factor-based tiered FIT with periodic revision and updates are

recommended.

12

Reviewing the Historical and Potential Contribution of Hydropower in Electricity Supply: The Ontario Case

Hydroelectricity has long powered the economic growth in Ontario, yet its contribution has

received little attention in the published literature. At present, the Feed-in Tariff (FIT) Program is

expected to quadruple wind capacity by 2018 (Canadian Wind Energy Association 2011), posing

significant challenges in dealing with demand fluctuations. The phase out of coal-fired electricity

(Miller and Carriveau 2017) and a proposed reduction in natural gas usage in Ontario (Ontario

Ministry of Energy 2009), elevates the role of hydropower as the most plausible dispatchable

generation. This chapter provides a narrative for the growth of hydroelectric power in Ontario in

the backdrop of historical events and major energy transitions. The discussion is targeted towards

a multidisciplinary audience interested in the changing energy policy landscape and its possible

influencing factors. It adopts a rather simplistic approach analyzing the present assets as well as

foregoing developments with the assertion that the past provides valuable lesson and guidance for

the future.

The work here builds a basic inventory of the existing hydropower fleet throughout the province

and analyzes it for capacity-based classification, grid contribution, plant age, ownership

distribution and so on. It highlights several important trends that suggest the need for systematic

rehabilitation or replacement of the aging hydro infrastructure considering its typical 100-year

service life. The analysis also confirms a substantial private sector investment in small and medium

plants. The discussion further explores the potentials for increased hydropower generation in

southern Ontario, complemented with a policy analysis that outlines the stimulus and possible

deterrents for such developments.

2.1 Background Power systems are fundamental components of the economy (Aliyu et al. 2015). Nevertheless,

energy security remains a challenging concept with limited diversity in fuel sources, intermittency

of most renewables, and the like (Brahim 2014; Brown et al. 2014; Gyamfi et al. 2015). In line

with the commitment to reduce emissions, there is an increasing focus on using a portfolio of low-

carbon technologies that includes wind and solar systems (Garrison and Webber 2011; Lopez and

13

Espiritu 2011; Mathiesen and Lund 2009; Pfenningeret al. 2014). With the mounting integration

of these intermittent renewables, generators with the ability to respond to demand variability have

become crucial for grid stability. At present, particularly with storage using batteries still awaiting

major technological breakthrough, hydropower with negligible emission and zero fossil fuel

dependency offers an effective means of permitting demand variability.

In Canada, regulatory and policy control over the electricity industry are primarily vested with the

provinces. This has led to a fragmented approach in almost all aspects of energy sector. For

Ontario, the power demand is projected to reach 176 TWh by 2030 (Ontario Ministry of Energy

2013a). To meet this growing demand, the province aggressively pursues increased renewable

deployment, use of natural gas and conservation (Ontario Ministry of Energy 2010). While the FIT

Program has resulted in tremendous growth in renewable industries, the development has so far

been limited to intermittent renewables (Wong et al. 2015; Yatchew and Baziliauskas, 2011).

Increased penetration from these intermittent sources along with the proposed refurbishments of

nuclear facilities (Ontario Ministry of Energy 2013a) may pose enormous challenges in dealing

with demand fluctuations. Moreover, phasing out coal-fired electricity in 2014 and a proposed

reduction in natural gas usage (Ontario Ministry of Energy 2009) (currently being used for

counteracting intermittency) elevates the role of hydropower as the most plausible source of

dispatchable generation.

Contemporary literature deals with hydroelectricity as a part of the overall generation fleet.

Researchers have written about the evolution of Ontario’s electricity system (Nelles, 1974;

Rosenbloom and Meadowcroft 2014) and related supply and demand models (Qudrat-Ullah 2013

and 2014; Zahedi et al. 2013). Others have examined specific policy issues, notably renewable

energy initiatives in Ontario (Heagle et al. 2011; Mabee et al. 2012; Pirnia et al. 2011; Rivard and

Yatchew 2016; Songsore and Buzzelli 2014; Stokes 2013; Winfield and Dolter 2014, Yatchew

and Baziliauskas 2011). Literary coverage of hydropower related studies includes hydraulic

modelling (Moeini et al. 2011; Ngo et al. 2007), analysis of multipurpose dams (Afshar et al. 2010;

Kamodkar and Regulwar 2013), sustainability evaluation of hydro projects (Kucukali and Baris

2009; Liu et al. 2013, Tahseen and Karney 2017: Chapter 9), hydropower policy development

(Ackere and Ochoa 2010; Koch 2002; Zhang et al. 2014) and so on. This paper diverges from the

tradition by tracing the chronological development of hydropower resources in Ontario and

14

analyzing its changing role in the backdrop of major energy transitions throughout the history. The

paper informs about the recent developments in the renewable energy policy landscape and

analyzes for its incentives towards hydroelectric development.

2.2 Current Status of Hydropower in Ontario

2.2.1 Existing infrastructure and its development throughout the history Along with its global counterparts, Ontario is striving to ensure the optimum utilization of its

existing hydroelectric resources. In 2014, hydropower generation exceeded 37 TWh which

represents 24% of the total generation (IESO 2017). Despite its limited contribution to the present

grid, hydroelectricity has powered Ontario’s economic growth since the beginning of the 20th

century (Rosenbloom and Meadowcroft 2014).

In the absence of international consensus regarding the classification of hydro systems, the analysis

here categorizes plants with or below 10 MW capacity as small stations. Systems between 10 –

100 MW are considered as medium while large stations represent more than 100 MW of installed

capacity. According to this definition, about half of Ontario’s hydro fleet is composed of small

plants. About 38% of these plants are medium followed by 16% large hydropower stations. The

infrastructure is dispersed among five zones, i.e., Central, Niagara or Southwest, Northeast,

Northwest and Ottawa/St. Lawrence (Ontario Power Generation 2016). The information on

specific plants are collected from Ontario Power Generation, Ontario Power Authority and various

company websites. The Central zone consists of about 46 relatively small run-of-the-river plants.

These are often credited for having minimal impact on the surrounding environment. The Niagara

zone operates a total of four conventional hydroelectric stations, two in St. Catharines and two on

the Niagara River, along with a pumped storage. These stations have a total capacity over 2.3 GW.

The Northeast part of the province hosts a total of 38 generating stations with a combined capacity

of just less than 2.3 GW. The Northwest group has approximately 22 stations with a combined

capacity of almost 800 MW. Ottawa has about 16 hydropower plants on the St. Lawrence River.

These plants, with a combined capacity of 2.6 GW, meet almost 8% of Ontario’s total electricity

demand (Ontario Power Generation 2015a). Figure 2.1 demonstrates the plant size distribution

among the Central, Niagara, Northeast, Northwest and Ottawa region. While the Central zone leads

15

in small hydro development, majority of the large plants are located in Ottawa (35%) and the

Northeast region (40%).

Figure 2.1: Distribution of hydroelectric plants in Ontario based on plant size

Since its separation from Ontario Hydro in 1998, the public electricity utility known as Ontario

Power Generation (OPG) has been vested with the responsibility of major hydropower fleet.

Besides, the sector enjoys a fair share of private investment. Figure 2.2 illustrates the ownership

distribution over Ontario’s hydroelectric assets. It confirms a substantial private sector

participation in the operation of small and medium plants. The capital intensive nature and the

long construction period associated with large-scale developments partly account for the private

sector’s rather modest participation in sizable projects. The ownership distribution also varies

among different regions. While the combination of Central and Northeast regions dominates with

63% private ownership, Niagara’s hydro infrastructure is entirely owned by OPG.

Figure 2.3 illustrates the chronological development of hydropower in Ontario. It began with a

number of small and medium plants that provided the necessary technical knowledge and

experience required for larger projects. Prior to 1900, the primary actors in Ontario electricity

system were privately-owned coal power plants (McKay 1983). The risk of potential strike among

coal mine workers and the dependence on imported coal leading to power shortage and sky-high

electricity prices, initially led the technological switch away from coal and transition to

hydroelectric power (Nelles 1974). A monumental step in this regard has been the establishment

of Hydro-Electric Power Commission of Ontario (later known as Ontario Hydro) in 1906 as a

0

15

30

45

Central Niagara Northeast Northwest Ottawa

Num

ber o

f pla

nts

Small hydro Medium hydro Large hydro

203 MW

2,338 MW

2,277 MW

792 MW

2,626 MW

Northwest Northeast

Central Ottawa

Niagara

16

publicly owned utility that would entrust government control on the electricity industry. The

regime change – from private to publicly owned electricity – was strongly promoted by Sir Adam

Beck (White, 1985). Under his direction, the Beck Commission explored the viability of a publicly

owned electricity system and based on the findings, recommended the establishment of the Hydro-

Electric Commission (Biggar 1920). Though initially it was vested with the responsibility of

developing grid infrastructure at Niagara, the commission’s changing role involved extending

control over generation assets.

Figure 2.2: Hydroelectric asset distribution on the basis of plant size and ownership

Since its initiation, growth in small hydropower capacity has remained fairly steady with a

substantial peak during 1920s. This period of rapid development is believed to be influenced by

the growing energy requirement during the wartime efforts (Evenden 2013). The growth in large

hydro came around 1930s, with a number of plants connecting to the grid between 1950 and 1970.

Rapid economic growth coupled with electrification of key industrial processes justified such a

grid expansion both in terms of capacity and transmission infrastructure. Though, the expansion

faced difficulties in the form of supply surplus during the Great Depression (1929‒1939), it did

not have lasting impact on the pace of development. Succeeding the peak development period of

1900‒1960, the growth experienced a temporary setback before being picked up again in 1990.

But this time, the capacity building was kept limited mainly to small and medium sized plants.

Several factors led to the temporary shunned growth in hydroelectric development post 1960. First,

the commission’s gross overestimation regarding the future electricity demand relegated hydro

0

10

20

30

40

Smallhydro

Mediumhydro

Largehydro

Smallhydro

Mediumhydro

Largehydro

Owned by OPG Private plants

Num

ber o

f pla

nts

Central Niagara Northeast Northwest Ottawa

17

development in favour of coal and nuclear projects (McKay 1983). Second, the successful

completion of Chalk River nuclear research project provided the technological expertise required

for the transition to nuclear. During 1971‒1983, the nuclear facility at Pickering extended the grid

capacity by 3,100 MW (Ontario Power Generation 2016). Bruce Nuclear Plant added another

6,272 MW between 1977 and 1987 (Bruce Power 2017). The remaining coal plants along with the

pumped storage at Niagara, known as Sir Adam Beck (SAB) PGS, provided the necessary

dispatchability required with the relatively invariable nuclear power. A high penetration of nuclear

power in combination with reduced growth in demand led to the decline in new hydropower

development. Third, 1973 oil crisis and the following rising fossil fuel prices concreted Ontario’s

preference for nuclear power (Rosenbloom and Meadowcroft 2014). Lastly, the best available sites

for hydroelectric development such as Niagara had already been exploited. However, the power

regime went through another shift with the onset of 1990s. Discussion on extending Ontario’s

renewable capacity gained momentum after the deregulation of the electricity system in 1998

(Swift and Stewart 2004). A rapid fall in demand growth due to a recession and a long-term

structural shift towards a service-based economy led to the preference for small projects

(Rosenbloom and Meadowcroft 2014). Interestingly, a number of hydroelectric projects, coming

online in the last 20 years or so, is believed to be influenced by the changing focus towards

intermittent renewables such as wind and solar.

At the beginning of 1900, small and medium hydro installed capacities were around 60 MW and

113 MW, respectively. By 1930, these amounts increased by about 60 MW (small) and 180 MW

(medium). The first large hydroelectric plant became online around the same time adding 500 MW

to the generation profile. The next few decades saw a gradual increase in capacity. The generation

experienced a massive boost in 1950s with the deployment of SAB Complex and RH Saunders

station. By the end of 1990, the capacity building in large hydro reached a plateau, while the

developments in small and medium stations continued. At present, though large plants represent

only 16% of Ontario’s hydro fleet, they provide 80% of the total hydropower generation (Figure

2.3). The combined average age of Ontario’s large hydropower fleet is close to 60 years which,

when aggregated over capacity, rises to 62. With the installation dating as early as the beginning

of the century, the average age for small hydropower stations varies between 82 to 85 years

(aggregated over capacity). The same for medium sized plants is around 67 years. The analysis

further confirms the critical role played by large hydroelectric plants in the overall generation.

18

Figure 2.3: Cumulative capacity contribution from large, medium and small stations based on

their year of commission

2.2.2 Hydropower for a low electricity rate Low and stable electricity prices have long influenced the growth of Ontario’s energy intensive

industries (pulp and paper, mining, chemicals, etc.) (Evenden 2013). Hydroelectric generation,

with zero reliance on fossil fuel, contributed to sustain this low rate. Till this date, it has continued

to be the lowest cost resource at $43‒55/MWh, closely followed by nuclear ($59‒68/MWh)

(Ontario Ministry of Energy 2013a). Biomass, the only renewable besides hydropower that can be

credited with flexible generation, is expensive (at $98‒156/MWh) (Ontario Ministry of Energy

2013a) and further limited by the available raw material. At present, the province largely depends

on gas-fired electricity for counteracting intermittency which, at $156‒166/MWh, is three times

the cost of hydroelectric generation (Ontario Ministry of Energy 2013a). Note that, the cost values

reported here are specific to Ontario and includes payments pursuant to contracts, regulated rates

or market clearing prices as negotiated by the generating facility.

2.2.3 Hydropower for low GHG emission Ontario’s Climate Change Action Plan (2007) sets ambitious GHG reduction targets: 15% and

80% below the 1990 level by 2020 and 2050 respectively (Ministry of Environment and Climate

Change 2013). While the electricity sector has responded with a 58% reduction in emission since

2005, nonetheless, it is still responsible for 14.5 Mt CO2 emission annually (Ministry of

0

20

40

60

80

100

1891

-19

09

1910

-19

29

1930

-19

49

1950

-19

69

1970

-19

89

1990

-20

14Cum

ulat

ive

capa

city

con

tribu

tion

(%)

Small hydro Medium hydro Large hydro

19

Environment and Climate Change 2014). The roadway for a further shift towards a low carbon

energy system is expected to experience quite a few challenges. First, with the nuclear covering

the baseload in Ontario, natural gas-fired electricity is expected to pick up the slack from declining

coal generation. However, there are increasing concerns over the risk from fugitive methane, a

GHG twenty times more potent than CO2 (Howarth et al. 2011). Second, capacity building in green

energy is facing oppositions in Ontario from various communities and First Nations groups

(Howlett and Ladurantaye 2011). With most traditional power sources either being non-

dispatchable or emitter, hydropower (especially pumped storage) presents the most plausible

means of offsetting GHG emissions while securing the flexibility and reliability of the grid.

2.2.4 Hydropower for renewable integration In response to the green development policies such as FIT, Ontario has experienced a substantial

growth in wind and solar industries. Moving forward, the province expects to reach its targeted

10,700 MW non-hydro based renewable capacity by 2021 (Ontario Ministry of Energy 2013a).

According to the Long Term Energy Plan (2013), the projected annual growth in renewables is

expected to contribute 64 TWh by the year 2020. However, integration of these intermittent

resources (wind and solar) poses enormous challenges in dealing with demand fluctuations.

Further, this study finds a substantial drop in wind generation from June–August (2006‒2013),

which conflicts with Ontario’s summer demand peak. By contrast, generation capacity from

hydroelectric resources remains relatively unaffected. Additionally, pumped storage can generate

electricity on demand while its pumping action can be used to absorb surplus generation. Other

benefits provided by hydropower include black-start and frequency regulation.

2.3 Potential Dispatchable Generation for Ontario

2.3.1 Marmora pumped hydro project There has been an ongoing discussion regarding the development of a pumped storage facility at

an abandoned mining property in Marmora. The project design takes advantage of the already

existing 20 m deep reservoir and plans to build a second one on the surrounding elevated land.

The project economics are reinforced by the long operating life of hydro and a short distance (8

km) to the closest transmission line (Northland Power 2014). With its 400 MW dispatchable

20

capacity, the proposed pumped hydro claims to produce power at one-fifth the cost of natural gas-

fired generation (Northland Power 2011). The plan is currently under consideration and if

approved, promises to inject nearly $700 million investment in Eastern Ontario (Northland Power

2011). Apart from a green, economic and flexible generation source, the project can be an

economic development platform for tourism and education.

2.3.2 Increasing capacity at Niagara Despite considerable undertakings, hydropower potential of the Great Lakes region is still partly

untapped primarily due to the policy constraint in the form of the 1950 Niagara River Treaty. It

establishes a minimum flow of 2,832 m3/s (100,000 ft3/s) over the falls during the daytime of the

tourist season. At all other times 1,416 m3/s (50,000 ft3/s) must go over the falls unless additional

water is necessary (Government of Canada 2015). The treaty identifies the unbroken crestline as

the most significant feature of the Niagara Falls and is aimed at ensuring this critical feature.

Nevertheless, the upper limit of 2,832 m3/s is not an absolute minimum since the crestline remains

unbroken even with the flow of 1,416 m3/s (Friesen and Day 1977). The additional flow during

the tourist season represents an annual cost of $52 million CAD in terms of the compromised

hydropower potential (Sedoff et al. 2014). From an environmental perspective, increased diversion

for power (by reducing flow over the falls) may decrease the continuous retreat of the escarpment

which, despite the diversions and the remedial works, still persists at a rate of 0.3 m per year

(Niagara Parks Commission 2015). Now, the expiration of the treaty in 2000 has opened the door

for renegotiation which may permit a greater allocation for power than currently allowed by the

existing terms of the treaty.

2.4 Changing Policy and its Impact on Hydropower Development Despite the economists’ preference for a direct measure such as carbon tax, a wide array of policies

and initiatives are designed to encourage renewable energy deployment. In 2006, Ontario adopted

its first FIT policy in the form of Renewable Energy Standard Offer Program (RESOP). It

guaranteed a sustained tariff for a period of 20 years. The qualifying projects, restricted to a 10

MW limit, were required to be connected to the distribution grid. RESOP successfully added 1,000

MW of renewable capacity with 53% solar, 37% wind and 3% water (MacDougall 2008) before

21

being discarded for FIT 1.0 in 2009. The new framework was divided into two streams: the FIT

and the microFIT. The projects exceeding 10 kW capacity was categorized as FIT while smaller

projects were contracted under a relatively simple microFIT. The projects were assigned priority

connections to existing transmission and distribution system with the commitment that further

infrastructural expansion would plan on accommodating increased number of FIT applications

(Ontario Ministry of Energy 2009). In 2012, the first two-year review of the program concluded

that the tariffs were too high for commercial proponents (Amin, 2002) and recommended a 15‒

20% reduction in price for certain non-hydro renewables. Since then the program price has been

updated several times with the latest one being effective from January 1, 2017. Table 2.1 lists the

latest tariffs for renewable generations under the Ontario FIT (Ontario Power Authority 2016a).

Table 2.1: Prices per kWh under FIT 5.0 (2017)

Energy

Group Category Project Size Tranche

FIT Price

(¢/kWh)*

Solar-PV

Rooftop

<=6 kW 31.1

> 6 kW <=10 kW 28.8

> 10 kW <=100 kW 22.3

> 100 kW <=500 kW 20.7

non-Rooftop <=10 kW 21

> 10 kW <=500 kW 19.2

Wind On-Shore <= 500 kW 12.5

Water power - <= 500 kW 24.1

Bioenergy

Renewable Biomass <= 500 kW 17.2

On-Farm Biogas <= 100 kW 25.8

> 100 kW <=250 kW 20

Biogas <= 500 kW 16.5

Landfill Gas <= 500 kW 16.8 Source: Ontario Power Authority 2016a

*price is without adders

22

While proponents of FIT focus on its ability to ensure price certainty, regulatory simplicity,

reduced unit cost and successful industrial development (Butler and Neuhoff 2008; Lipp 2007;

Mitchell et al. 2006; Stokes 2013), critics have accused it for increasing the cost to ratepayers,

diminishing competitiveness and impacting the economy by cutting capital to key sectors (Dewees

2013; Dio et al. 2015; McKitrick 2013; Stokes 2013). This section, however, limits its discussion

of FIT exclusively with respect to hydropower development in Ontario. The 2009 FIT offered 13.1

¢/kWh for 10 MW or less capacity projects while large schemes (≤ 50 MW) were guaranteed a

price of 12.2 ¢/kWh. A revision in 2016 augments this price to 24.1 ¢/kWh – an 84% increase

compared to FIT 1.0. The on-peak premium and reduction for off-peak generation – intended to

incentivize dispatchable generation – have remained consistent throughout multiple revisions since

2009. The FIT also offers remuneration for First Nations, community, municipality or public

sectors participation. While aboriginal participation gains 0.75‒1.5 ¢/kWh, others benefit from

0.5‒1.0 ¢/kWh on top of regular rate. These premiums encourage community and individuals, who

typically lack the financial and institutional capacity to take part in the competitive bidding

process, to become energy producers.

By contrast, a few undertakings of the current FIT may discourage further development in

hydropower. First, a revision in 2013 has effectively terminated FIT projects over 500 kW (Ontario

Ministry of Energy 2013b), while financing those through the Large Renewable Procurement

(LRP) and the Hydroelectric Standard Offer Program (HESOP) (Ontario Power Authority 2013a;

2014). Now, procurement under the HESOP follows two streams: municipal and expansion. While

the former requires a mandatory municipal collaboration, a prior application to the FIT 2010 and

a 50 MW capacity limit, expansion stream remunerates currently contracted facilities for

incremental capacity increase up to 40 MW. The HESOP offers a base price of 0.131‒0.141 $/kWh

with a 35% on-peak premium and a 10% decrease for off-peak generation (Ontario Power

Authority 2013a; b) which is substantially lower in comparison to the current FIT. In addition,

waterpower projects other than hydroelectric are restricted to connect to the distribution grid

(Ontario Power Authority, 2013a). In contrast, the LRP encourages to bid below the published FIT

prices that may lower profit margins for developers. Also, assessment criteria under the LRP are

increasingly demanding and require extensive pre-development activity (Lord and Tyler 2015).

Second, the current venture capital financing favours a relatively quick return on investment

(Rajan 2010). Thus, hydro which, followed by a relatively long development period guarantees a

23

stable return, is often overlooked in favour of other projects. Considering this, prior to 2013 the

OPG was formally denied to bid under the FIT with the exception of hydro projects. However, a

directive now allows the organization to design competitive procurement process for all

renewables (Ontario Ministry of Energy 2013b). Finally, when introduced in 2009, the FIT

projects had to adhere to stringent domestic content requirements: 25% for wind and 50% for solar

PV. In 2012, it was further increased to 50% for wind and 60% for solar before being completely

omitted in the following year to comply with World Trade Organization (WTO) rulings (Ontario

Ministry of Energy 2013c). This effectively eliminates the market barriers and unlike hydropower,

provides incremental incentives towards wind and solar development. All these have led to the

less than modest growth in hydroelectric projects under the FIT. Up to this writing, waterpower

comprises a mere 4% of the total contracts executed under FIT (Ontario Power Authority 2016b)

while the municipal stream has contracted a total of 35.25 MW capacity (Ontario Power Authority

2016c).

2.5 Conclusions and Policy Implications Addressing climate change will require the electricity system to transition towards very low carbon

emissions over this century (Caldeira et al. 2003; Hoffert et al. 1998; Lackner and Sachs 2005).

Negligible emission and zero fuel dependency – are the very attributes which make hydropower

so attractive towards such low-carbon economy. This chapter provides a narrative for the growth

of hydroelectric power in Ontario in the backdrop of historical events and major energy transitions

and emphasizes its role at the current grid level. The proportional contribution from Ontario’s large

hydroelectric plants is as high as 80% with respect to the total hydropower generation. The

provisional analysis here suggests that the combined average age of these stations is around 62

years. With a number of small plants operating since the early 1900s, their average age is close to

85 years. Considering the typical 100-year service life, many of these plants are about to expire

within the next 15‒20 years. However, there seems to be little consensus regarding how this aging

infrastructure would be rehabilitated or replaced within the current generation profile.

Ontario is currently reconsidering its electricity system’s future. With the aging nuclear reactors

and phasing out coal-fired electricity, the province has adopted aggressive renewable policies in

the form of FIT. Though being effective in large-scale, rapid renewable energy deployment (Butler

24

and Neuhoff 2008; Fouquet and Johansson 2008; Kwon 2015; Lipp 2007; Mendonça 2009;

Mendonça et al. 2009; Mitchell et al 2006; Sun and Nie 2015), FIT has been accused of trying to

pick “technology winners” by disproportionately awarding a few green technologies. Thus,

success of FIT program strongly depends on its design and implementation (Haas et al. 2004;

Couture and Gagnon 2010). In this context, deciding the appropriate incentives for hydropower

development raises all sorts of fascinating questions. Should the tariffs be based on marginal sites

or the best resource sites? With the most suitable places already exploited or far from the load

centers, does it make sense to offer higher tariffs for capacity building at low resource sites? If not,

how does that impact Ontario’s commitment to increase its hydro capacity from the current 8,400

MW to 9,000 MW by 2030? And is a greater price for energy politically feasible? The discussion

in this chapter weighs in with an historical perspective on energy transitions and advocates for an

active political debate over the rules and tariffs involving the experts for a greener, more flexible,

reliable and at the same time relatively inexpensive grid. It is taken as an axiom that all good

intentions and good directions must be subjected to excellent, thoughtful, and forward-looking

analyses and that such analyses must include technical, economic and social categories.

25

Exploring the Multifaceted Role of Pumped Storage at Niagara

An energy alternative discussed in Chapter 2 involves pumped storage development as a measure

for increased dispatchable generation in Ontario. However, such developments are often accused

of being unprofitable. While the long-term FIT contracts offer ways to finance these projects ‒ an

option discussed in the preceding chapter ‒ the generators can instead partake in the wholesale

electricity market where revenue is decided based on a competitive bidding process. This chapter

illustrates a direct optimization approach to evaluate the spot market for pumped storage operations

given well-forecasted flows and energy price. Sir Adam Beck Pumping Generating Station, located

on the Niagara River, is selected as the subject of the model application. The model is then used

for analyzing the impact of diurnal and seasonal price variations, possible improvements by

varying cycle length and reservoir size. With such restrictions in place, the analysis further

considers the trade-off between hydropower and ecological targets imposed by the 1950 Niagara

River Treaty.

The chapter is based on the paper entitled “Exploring the Multifaceted Role of Pumped Storage at

Niagara” by Samiha Tahseen and Bryan Karney, published (13.06.2016) in the Journal of Water

Resources Planning and Management and reproduced herein with permission from ASCE. The

objective of this chapter is to analyze profit characteristics under different operating and market

conditions, and reflect on the trade-offs among conflicting intentions.

3.1 Background Despite its typically high development costs and sometimes considerable environmental impacts,

hydropower has much to recommend it. Once installed, it has a desirable quick-start (coming

online within a short notice) and black-start (providing electrical supply during a total or partial

shutdown of the transmission system) capability (Evans et al. 2009; Sharma et al. 2015). It can

efficiently respond to peak load (Maxim 2014), its spinning reserve provides flexibility and

protection to the overall grid (Zhang et al. 2015) and allows leveraged investments in other

intermittent sources (Ayodele and Ogunjuyigbe 2015). Since its raw power comes from a

renewable source, hydro is able to reduce the electrical system’s reliance on fossil fuel. Pumped-

26

storage hydroelectricity (PSH) enhances power generation in that water can be pumped to a higher-

elevation reservoir and stored in the form of gravitational potential energy. Pumps are

predominately run using low-cost off-peak electricity, and the stored water later generates

electricity at peak price, usually during periods of high demand. Although the energy losses of the

pumping process make the plant a net energy consumer (IPCC 2011), the system often increases

revenue by selling high and buying low, and thus helping to balance the grid.

At present hydropower is experiencing a worldwide renaissance. The need for clean, affordable

energy and the increasing need to have a flexible component in the supply mix have driven interest

in hydroelectricity. Canada is the world’s third largest hydropower producer with 9.8% of total

production according to the BP Statistical Review of World Energy (British Petroleum Company

2015). In 2014 hydropower generation in Ontario alone exceeded 37 TWh (IESO 2015). The

Niagara River, along with its contribution to the tourism sector, acts as a key resource to this

generation. Presently river power provides nearly 8% of Ontario’s total electricity generated at the

Sir Adam Beck (SAB) complex. Along with two conventional power stations, the complex

currently hosts Ontario’s only pumped storage station, the SAB Pumping Generation Station

(PGS). With its limited capacity, PGS contributes to stabilizing the grid by producing power on

demand and also by storing surplus energy generated by nondispatchable and intermittent sources.

The SAB PGS’s ability to pump water is fundamental to water level control at the point of

crossover, a critical component in ensuring the appropriate performance (Maricic et al. 2009). One

of the roles of the plant is a little unconventional; that is, it is intended to store a volume of water

nearer to the two conventional hydro plants, thus enhancing their hydraulic capacity to improve

their responsiveness to peak power demands (Figure 3.1).

Being the only commercially proven, utility-scale energy storage technology, PSH has been

suggested as a key response to demand variability (Rehman et al. 2015). However, despite

perceived technical demand, profitability remains a major obstacle for PSH systems. Ingebretsen

and Johansen (2014) assessed six proposed PHSs in Norway and rejected the profitability of all of

them. However, this outcome contrasts with a report by the German Advisory Council (2011),

which suggested that those plants have a high return on investment. A comprehensive cost-benefit

model by Zafirakis et al. (2013) shows that both pumped hydro and compressed air energy storage

(CAES) can be cost-effective with the application of a “socially just” Feed-in Tariff (FIT). Such a

27

prognosis is interesting, since PGS, being a relatively well-known option, is not eligible for FIT

rates in Ontario. Salevid (2013) investigated the economic viability of restoring a currently

decommissioned Swedish pumped storage and established a correlation between price volatility

(energy price variability during on-peak and off-peak hours) and PSH profitability, concluding

that the feasibility of PSH depends on sustained highly volatile energy prices. The SAB PGS faces

similar challenges with the unit energy cost increasing by 72% between 2006 and 2008 (from

$47.1 to $81.2/MWh) (Ontario Power Generation 2010). Ontario FIT Program, expecting to

quadruple wind capacity by 2018, can impact PGS in two ways – increased availability of low-

cost, off-peak electricity to reduce pumping cost and low electricity prices to decrease the overall

profit opportunity (Linares et al. 2008).

Economic viability is a major consideration for any development and perhaps the strongest

motivation of investors. Deregulation of the electricity market has created a competitive, profit-

driven environment in which hydro producers, whose typical role was to balance the grid, face

new challenges with the ultimate goal of maximizing profits. With the development in storage

technologies such as flywheels, CAES, batteries, capacitors, and so forth, there is a need to assess

the role of PSH as the most likely to be a cost-effective storage option. Rangoni (2012) suggested

a case-by-case analysis to determine the most cost-efficient solution to grid flexibility, and

recommended investigating pumped storage feasibility with respect to the market's ability to

deliver profits. In this context, analyzing the profit characteristics of the SAB PGS under various

energy price scenarios is worth investigating. The expiration of the original 1950 Niagara River

Upper Niagara River

Storage reservoir

SAB PGS

forebay

SAB I and II

Niagara tunnel

Figure 3.1: Niagara Hydroelectric plants system

28

Water Diversion Treaty, which currently dictates limits on the available water for hydropower,

opens the possibility of using a greater allocation of water for power than currently allowed by the

terms of the existing treaty. Hence, the current study explores what portion of the hydropower

potential is compromised by the current terms of the treaty and the possibility of increased

generation through renegotiation. To analyze fully the role of Niagara’s potential PGS contribution

to meeting peak demands is outside the current scope of this research.

The idea of prescheduling pumping and generation using forecasted data on demand and

subsequent price is not new. Afshar (2012) and Bosona and Gebresenbet (2010) developed

optimization models for maximizing hydropower generation where the monthly values of the key

decision variables are generated for a year. However, aggregating the results on a monthly basis

may not be realistic because such coarse resolution aggregation fails to capture the impact of

changes in demand and price due to seasonal variations, long holidays, sudden weather changes

and interruptions to the power supply. Moreover, although most reservoirs tend to be used for

seasonal water storage, typical pumped storages are used for load leveling purposes. Thus, its

operation typically involves daily filling and subsequently discharging water. Given this, a model

that optimizes decision variables on a monthly basis holds few advantages over typical operational

models. In contrast, short-term models give more control over time, duration, and flow to

maximize the performance of PSH generators. Realizing this, Latorre et al. (2014), Mo et al.

(2013), and Prasad et al. (2012) proposed short-term hydro scheduling and discussed its challenges

and possible solutions. Considering the availability of reasonably accurate day-ahead energy price-

forecasting models (Aggarwal et al. 2008; Zareipour et al. 2006), the work described in this paper

can be used to optimize the daily operation schedule for maximizing benefits. The developed

model is then used for extensive analysis of profit characteristics, the impact of potential

constraints in the form of the 1950 Treaty and possible improvements by varying cycle length and

reservoir size.

3.2 Brief Literature Review In conventional hydropower optimization models, the objective function is nonlinear because the

product of the discharge and the head are required for decision making. Hydropower capacity of a

storage plant can be expressed as:

29

P = ηρgQ(t)Hn(t) = KQ(t)Hn(t) (3.1)

where K = ηρg; K = constant; η = overall efficiency to produce hydropower; ρ = water density; g

= gravitational acceleration; and Q = water release for power generation. The net head can be

written as:

Hn = H – Htail – Hloss (3.2)

where H = storage water level; Htail = tail water level; and Hloss = head loss at time t. If changes in

Htail and Hloss are insignificant compared with H, Hn = H can be approximated. Then the objective

function for maximizing hydropower energy can be formulated as:

E= K�Q(j)H(j)J

j=1

=KQH (3.3)

Yet even Equation 3.3 is a nonlinear product of the vector Q and H. Successive linear programming

(SLP) first appeared in Griffith and Stewart (1961). Although Palacios-Gomez et al. (1982)

reported a few rather unimpressive results, this approach remains highly recommended in reservoir

operations because of its easy implementation and tendency to converge to a global optimum.

Further effort by Kamodkar and Regulwar (2013) applied fully fuzzy linear programming (FFLP)

on a multipurpose reservoir to represent uncertainties in system parameters. Fleten and

Kristoffersen (2008) proposed a mixed-integer linear programming (MILP) model and

demonstrated its application with a Norwegian facility. Whereas many researchers have

approached reservoir operation through successful linearization of the objective function, Helset

et al. (2013) and Moeini et al. (2011) proposed a model based on stochastic dynamic programming

(SDP). Haguma et al. (2010) added consideration of climate induced flow variation whereas

Catalão et al. (2012) used nonlinear programming (NLP) for optimizing hydropower generation.

Clearly there is no general algorithm but rather a range of choices depending on reservoir-specific

system characteristics and the preferences of the modeler.

This study introduces a rather straightforward MILP model to investigate pumped storage

profitability at the SAB PGS. The model evaluates a daily operation schedule (pumping and

generation) to assess the available energy that can be offered on the market and at the same time

30

reduce cost associated with pumping operations. The objective is to examine the impact of

changing electricity rates, reservoir capacity, and treaty flow constraints on PGS profitability. The

exploratory nature of the study motivates the adoption of a simplified LP approach, rather than a

more complex nonlinear or dynamic programming approach. Kusakana (2015) studied the

technoeconomic feasibility of pumped storage and recommended it in conjunction with a stand-

alone hydrokinetic system. Similar studies by Caralis et al. (2012) and Steffen (2012) investigated

the potential of pumped hydro storage systems with increasing penetration of renewable resources.

This paper analyzes the role of PGS from an economic perspective and includes a tradeoff analysis

between financial and environmental considerations. It further explores the benefits and possible

challenges faced by PSH development in Ontario.

3.3 Optimization Model Development Based on the approaches discussed, the authors adopted a linearized optimization model for the

SAB PGS. The following sections discuss the model and the data used for the purpose.

3.3.1 Formulating the context-specific optimization model In Canada, regulatory and policy control over the electricity industry are primarily vested with the

provinces. The electricity system in Ontario is a hybrid between a market and a regulated entity

where generators submitting bids to the system operator are dispatched from the lowest bid until

the demand is satisfied (IESO 2015). Lately the annual demand curve has exhibited a dual peak

(summer and winter), where the highest demand situations usually occur during the summer (IESO

2015). The hourly average of the 5-minute energy market clearing price (MCP) is defined as the

hourly Ontario energy price (HOEP), and forms the basis for financial settlements. As intermittent

renewables are offered a guaranteed price through the FIT program, HOEP bears little to no

relation to the cost of building renewable capacity (Auditor General of Ontario 2011).

Since PSH benefits from price arbitrage, selection of pumping and generation durations is critical

for optimizing SAB PGS operation. Decision variables representing pumping and generating hours

(x; y) are required to be binary. In the model, these variables adopt a value of 0 or 1 for

nonoperating/operating stages. No time delay is considered for the transition from the pumping to

31

the generating sequence as suggested by Maricic et al. (2009). The model contains another set of

variables (u; v) representing inflow into the reservoir and outflow through the turbines.

Dispatchable sources such as the SAB PGS are primarily aimed at providing ancillary services and

are online when nondispatchable (nuclear) and intermittent sources (wind and solar) are exhausted.

Therefore, analyzing PGS’s contribution to peak power requires estimating residual demand —

generations from all nondispatchable and intermittent sources throughout the province subtracted

from total demand. Due to the resource-intensive nature of data collection and processing required

for such calculation, the HOEP is used as a suitable surrogate in the current model. There remains

a strong, positive, and statistically significant correlation (r = 0.6, P <0.001) between demand and

the HOEP, suggesting that a 36% variation in demand data can be explained by energy price, so

the approximation is reasonable given the exploratory purpose of this research. Such an

approximation permits reasonable estimation of profit for the facility. Now, in combination with

an electricity price forecasting model, pumping and generating hours and the corresponding flows

for the SAB PGS can be determined.

To represent the volume of water stored in the reservoir at start of hour i, the authors introduce a

storage function S and assume it to be empty (0) at the beginning of the operation. The model

operates to exhaust all the water stored within the same day. The upstream river flow is computed

using a United States Geological Survey (USGS) rating curve on 1-hour water level data from

National Oceanic and Atmospheric Administration (NOAA 2013) Station 9063020, located at the

mouth of the Niagara River. Water level data from 2007 to 2013, obtained in units of feet, are

converted to discharge (Q, cfs) using the following rating equation:

Q = 260.5 (H - 550.11)2.2 (3.4)

where, H = water level above IGLD1985, i.e., the international elevation reference for the Great

Lakes-St. Lawrence river system. The historical flow data is then averaged to get a reasonable

estimation of the hourly river flow for each month. The possibility of considerably large flow

variations is ignored due to the highly regulated nature of the upper Great Lakes. Not all of this

water can be used for hydropower generation since the power flow is subjected to the 1950 Treaty

restrictions which establish that during the period lasting from April 1 to September 15, no less

than 2,832 m3/s (100,000 ft3/s) must be going over the falls between 8:00 AM and 10:00 PM. The

32

same flow restrictions are effective between 8:00 AM and 8:00 PM from September 16 to October

31. At all other times, a minimum of 1,416 m3/s (50,000 ft3/s) should be maintained unless

additional water is necessary (Government of Canada 2015). Figure 3.2 shows the flow-handling

capacity at the SAB complex in comparison with the daily variation in available power flow during

representative months. The restrictions, which coincide with peak electricity demand in Ontario,

limit generation since the available power flow can at times be half of the SAB’s maximum

capacity. Now, the flow at the SAB PGS is calculated by deducting the discharge required for

reasonable generation (75% capacity) at the two conventional power plants (SAB I and II) from

Canada’s share of the treaty-specified available water. Note that the available flow at the SAB

complex is further limited by the diversion capacity of the existing tunnels and the power canal —

the impact of which is ignored considering the scope and exploratory nature of this study.

Figure 3.2: Comparison between available power flow and the maximum capacity at SAB

Complex

The mathematical expression for the model is as follows:

Objective function

The objective is to maximize the sum, of the revenue from power selling, minus the cost incurred

from pumping operations.

max �𝑦𝑦𝑖𝑖rii

vi - �𝑥𝑥𝑖𝑖cii

ui (3.5)

0

1000

2000

3000

4000

1 3 5 7 9 11 13 15 17 19 21 23

Ava

ilabl

e flo

w (m

3 /s)

Hour

Jan AprJune Nov Flow capacity at SAB

33

where, ri = revenue generated from turbine release; ci = cost incurred from pumping; ui, = volume

of water pumped; and vi = release through the turbine. Decision variables xi and yi represent the

decision to operate the pump and turbine at the ith hour, respectively. Considering the station

operating in a day-ahead electricity market, all the decision variables (xi; yi; ui; vi) are generated

on a 24-hour basis using the software package LINGO.

Constraints

The model must reflect a rather long list of constraints to replicate the existing system at the SAB

PGS. Beginning with typical reservoir constraints such as capacity, flow balance, and the like, the

list extends to limits that are specific to the system at Niagara. One such limit, the dichotomous

nature of pumping and generation functions, requires a constraint that ensures that only one of the

operations is running at a time (Equation 3.6). The last constraint built into the model is the flow

restrictions imposed by the 1950 Treaty followed by non-negativity constraints. A complete list of

constraints along with the mathematical expressions are provided below.

1. Pumping and generating cannot be done simultaneously.

xi + yi ≤ 1 (3.6)

2. Volume of water pumped into the reservoir is less than equal to the maximum pumping flow.

ui ≤ fxi (3.7)

where ui = volume of water pumped in hour i and f = maximum pumping flow per hour.

3. Volume of water released is less than equal to the maximum turbine flow.

vi ≤ hyi (3.8)

where vi = volume of water released in hour i and h = maximum turbine flow per hour.

4. Flow balance relationship.

Si+1 = Si + ui − vi (3.9)

where Si = volume of water in the storage reservoir at the start of i.

5. Volume released is less than equal to the water stored in the reservoir.

vi ≤ Si (3.10)

6. Water stored in the reservoir is less than or equal to the reservoir capacity.

Si ≤ C (3.11)

Where C = reservoir capacity.

7. Volume of water pumped into the reservoir is less than or equal to the water available.

34

𝑢𝑢𝑖𝑖 ≤ Available flow at SAB PGS (3.12)

8. Non-negativity constraints to the applied.

xi, yi, ui, vi, Si ≥ 0 (3.13)

3.3.2 Analyzing input price data For computing the cost and revenue component of the model, the authors use 2003‒2012 HOEP

data from the Independent Electricity System Operator (IESO 2015). These data are analyzed to

create four scenarios: (1) characteristic period in terms of peak demand, (2) possible electricity

price fluctuation for each month, (3) price variations on the basis of weekday or holiday, and (4)

increase in storage. For this purpose, the daily average electricity price is computed and this data

set was used to determine the 85th and 15th percentile average HOEP. A random selection is made

from the days with average HOEP above the 85th percentile value for extracting the hourly rate

from the original data set for each month. This procedure is repeated for the data set below the 15th

percentile value. These two separate energy price information for each month, extracted from the

85th and 15th percentile data sets, represent the high and low energy price, respectively, throughout

this paper. Such a procedure examines the impact of price volatility for two extreme cases in each

month so that the result is a range rather an exact number for profit. To investigate the effect of

weekday-holiday energy price variation on scheduling, the authors follow the same procedure

described previously; however, the percentile is now computed on data separated on the basis of

weekday and holiday. Thus, each month is associated with four different energy price data sets:

i.e., weekday high HOEP, weekday low HOEP, holiday high HOEP, and holiday low HOEP.

Based on the occurrence of consistently high energy prices, months with typically high electricity

demand are selected. The data is further analyzed for months with a large spread in energy price

data, which is captured by standard deviation. Such analysis is interesting since large fluctuations

in price provide room to increase profits by reducing costs associated with pumping, at the same

time maximizing revenue from generation. The authors then convert these data to price per unit

volume of water based on the operating conditions of the SAB PGS. Although the difference in

head between the upper and lower reservoir ultimately affects the potential energy of the stored

water, tracking the exact difference is awkward (Chang et al. 2013); thus, a fixed head is used to

35

ensure a representative price per unit volume of water. A relatively low pumped storage efficiency

(50%) is used to compensate for such approximation.

3.4 Model Explorations The developed model allows several characteristic of the SAB PGS to be explored, such as, daily

price variation for pumping and generating decisions, profit characteristics for weekdays and

holidays, profit sensitivity to cycle length, and so forth.

3.4.1 Analysis of profit characteristics on a monthly basis The task of choosing pumping, generating hours, and corresponding flows that best utilize the

price variation scheme is accomplished by the proposed model. When compared with the observed

values at the SAB PGS, the optimization model results in a 37% increased daily generation. While

the actual dispatch shows a rather large spread (between hours of 7 and 24) and is targeted to meet

the gap between demand and generation, running entirely on profit motive as in the case of the

optimization model, leads to maximum generation during hours of high energy price. The

deviations between the observed and the optimized values are further influenced by the assumption

of fixed generation at the run-of-the-river plants by the optimization approach.

August has the best scope of increased revenue generation. A typical day in this month can earn

an average profit of just over $17,500 CAD when allowed to run on full capacity, leading to a total

of approximately $543,000 CAD per month. However, the flow constraint imposed by the treaty

restricts the available flow for pumping. When incorporated into the model, it results in a reduced

profit of $10,700 CAD per day, leading to a substantial $211,000 CAD decrease in monthly profit.

The same holds true for the month of February, which is found to be the least profit-yielding month

according to the proposed model. Notably, certain days in February provide no incentive for

operating the reservoir from a financial return perspective. These cost savings results are hard to

compare with other jurisdictions since pumped storage feasibility, as suggested by Rangoni (2012),

needs to be investigated with respect to the market's ability to deliver profits. Also, being built in

1958, the SAB PGS mostly likely runs on a fixed price contract as opposed to marginal cost-based

operation. Here, the facility is used as a surrogate to investigate the profitability of pumped storage

36

in Ontario spot market. The wide disparity in daily profit with the high and low energy price

scenario for the same month indicates a highly volatile electricity market in Ontario. This finding

is consistent with Zareipour et al. (2007), who identified the provincial electricity market to be one

of the most volatile after comparing with similar markets worldwide.

When analyzed for the factors responsible for profit variation, both the average and the median

HOEP for February and August are found to be strikingly similar. The reason for the difference in

monthly profit lies in the interquartile range, which for August offers a greater scope for

optimization than February. The results obtained when running the model with the high and low

electricity price data set for each month are then averaged and aggregated to obtain the monthly

profit, as shown in Figure 3.3. In a comparison of annual income with and without the treaty flow

restriction, the compromised hydropower potential is found to be worth just over $6 million CAD.

The logic behind such aggregation is that the analysis here aims to provide a ballpark estimate

rather than a precise number. Since the profit relies on electricity rate which varies widely (‒$10

/MWh to over $200 /MWh) owing to hourly demand and supply (IESO 2015), there is little value

for such exact estimation.

Figure 3.3: Estimation of monthly profit for PHS with and without flow restriction

3.4.2 Analysis of profit characteristics on a weekday-holiday basis High and low energy price data separated on the basis of working days are used as input to the

model to evaluate profit sensitivity to these factors. The model predicts a reasonably higher profit

0

0.5

1

1.5

2

2.5

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Prof

it pe

r mon

th (m

illio

n $

CA

D)

Without flow restriction With flow restriction

37

during weekdays than during weekends (Figure 3.4). According to this revenue model, persistent

low electricity rates during the weekends of January and February promote little economic gain

for PGS operation. Another distinctive pattern is the reduced profit during the weekends of May,

June, and July, considering the typically high demands during these summer months. This can be

attributed to either generally high prices with little variation over a 24-hour period, leading to little

difference between the revenue and the pumping cost or increased outdoor activities that drive

down both demand and energy price. The latter case represents situations where the diurnal-scale

pumped storage contribution may be redundant; thus, the proposed model responds by reducing

generation. The result reinforces the need for price volatility on a 24-hour basis for pumped storage

feasibility.

Figure 3.4: Variation in profit for PHS during weekdays and holidays

3.4.3 Profit sensitivity to cycle length Since the change from pump to turbine operation (and vice versa) is achieved within a few minutes

(Maricic et al. 2009), the model assumes a seamless transition between these cycles. Frequent

changes in operating conditions are relatively common for pumps, but this is not a usual practice

in the case of turbines. Often these transitions are associated with vibrations and wear in the system

that lead to machine depreciation. Figure 3.5 compares average monthly profit for three different

cycle lengths. Although the model yields a significantly higher economic gain for the 1-hour cycle,

the difference in profit between the 2- and 3-hour cycles is negligible. Based on the results, the 1-

hour cycle may be chosen as an attractive alternative from the single perspective of profit

0

10000

20000

30000

40000

50000

60000

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Prof

it pe

r day

($ C

AD

)

profit per weekday (high HOEP) profit per weekday (low HOEP)profit per holiday (high HOEP) profit per holiday (low HOEP)

38

maximization. However, a 3-hour cycle is preferred over a 2-hour cycle in terms of both

maximizing economic gain and protecting the structural integrity of the current system.

Figure 3.5: Comparative analysis of economic return versus running time for pumped storage

3.4.4 Evaluating potential improvement opportunities for SAB PGS The third Niagara tunnel, completed in 2013 for $1.6 billion CAD, increases Canada’s diversion

capacity by 500 m3/s. With the tunnel in place, the PGS expansion plan now includes increasing

the reservoir footprint by raising the dyke elevation. The model assumes a 4 m increase in reservoir

height leading to a storage capacity of 32 Mm3 (instead of the current 20 Mm3). Analysis shows

no economic gain with such an increase in reservoir capacity. With the flow restrictions and the

24-hour reset constraint, the SAB PGS simply will not be able to utilize the excess capacity offered

by the increased reservoir footprint. Moreover, the scheme does not guarantee higher profit

throughout the year even when operated with a no-treaty restriction model. The hours with the best

profit opportunity having already been selected, the additional storage allows operation only

during periods that offer little difference between revenue and pumping cost. The outcome shows

additional revenue for months with relatively high electricity price variation without any

noticeable increase in revenue for the rest of the year. While these results hold true for diurnal

operation, a relatively longer reset condition (a few days or weekly storage) might be able to

benefit with such increase in reservoir footprint.

0

0.5

1

1.5

2

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Prof

it pe

r mon

th (m

illio

n $

CA

D)

1 hour cycle 2 hour cycle 3 hour cycle

39

3.4.5 Profit sensitivity to energy price

The analysis has heretofore explored pumped hydro operation based on historical electricity rates.

However, increased participation of pumped storage has the potential to influence the electricity

market. In Ontario the HOEP results from generators submitting bids to the Independent Electricity

System Operator (IESO), which dispatches generators starting with the lowest bid. When the

system is congested, some higher cost units may be necessary to release congestion. Pumped

hydro, because of its operational flexibility, can alter the electricity spot price by delaying the

participation of such higher cost units (Kanakasabapathy 2013). Depending on operational mode,

increased contribution from pumped hydro (through increased utilization of the SAB PGS and/or

with the completion of Marmora pumped storage project) can influence energy price in two ways.

First, in pumping periods the pumped hydro consumes electricity and therefore HOEP may rise.

Second, when the same facility generates, the marginal cost decreases (Sousa et al. 2014). Research

shows that the impacts of spinning reserve on hourly spot prices can be as large as 25% (Zhu et al.

2000). Considering the limited pumped storage capacity (a total of 574 MW with Marmora project)

in Ontario, the analysis here assumes a 2.5‒10% reduction in electricity rates during the peak

demand hours (9 to 15-hour and 18 to 21-hour). Due to the surplus generation from wind resources

in Ontario (Gallant 2015), this study ignores the impact of nighttime pumping operation on energy

price.

The model generated is next used for evaluating the impact of changing energy price on PGS profit

characteristics (Figure 3.6). According to the analysis, the changing electricity rate due to

increased grid participation by the SAB PGS can result in a 1‒24% reduction in profit depending

on particular months. The relatively large reduction is more consistent in winter months when the

energy price is generally low.

40

Figure 3.6: Impact of changing electricity rates on pumped storage profit

3.4.6 Trade-offs between power generation and scenic flow restrictions Almost all real decision-making problems are multi-objective in nature. These problems often

involve trade-offs among conflicting objectives. For PGS which serves hydropower generation as

a key purpose, the operator may wish to maximize power generation while adhering to the flow

restrictions imposed by the 1950 Treaty. However, these two objectives are typically conflicting

since the treaty restrictions, which specify a minimum flow over the Niagara Falls, also limit the

available flow. To this end, the authors use the Constraint Method which transforms

multidimensional problem into a series of one-dimensional problems. The technique involves

optimizing one objective while representing other objectives as constraints. Systematic repetition

with different constraints on the objectives generates the entire set of noninferior solutions

(Neufville 1990). The trade-off curve, often called a Pareto surface, elucidates the degree of

sacrifice of one benefit required for gain of another.

Here, generating profit for PGS is maximized and available pumping flow is constrained over a

range of target values. Figure 3.7 shows the resulting Pareto surface for July where Point A

represents the greatest possible profit with no flow restrictions and Point B represents profit under

current constraints. July is ideal for the trade-off analysis since the tourist flow requirement

coincides with high power demand. Whereas Points A and B are two major alternatives, Point C,

with 708 m3/s reduction in tourist flow, represents a compromise solution that results in a 2.3%

increase in profit. An additional 2% increase occurs when the entire tourist flow (which still

0.1

0.4

0.7

1

1.3

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Prof

it pe

r mon

th (m

illio

n $

CA

D)

Base case2.5% reduction5% reduction10% reduction

41

maintains 1,416 m3/s over the falls) is diverted for power generation purposes (Point D). The profit

is subjected to an increasing growth rate between Point A and D, which suggests a stronger conflict

between flow targets and economic gain when flow restriction is reduced below 1,416 m3/s. The

maximization of profit without consideration of environmental flow, however, represents an

extreme case. Visually assessing the trade-offs between multiple objectives helps in the selection

of policies that achieve a balance between different metrics of system performance.

Figure 3.7: The trade-off surface between the economic gain and the environmental

consideration for pumped storage in July

3.5 Benefits and Possible Challenges for Pumped Storage Hydro reservoirs are often used for storing electric energy generated by nondispatchable sources,

provided that the power plants are connected by a common grid, and that transmission capacity is

sufficient to allow load leveling. Apart from revenue considerations, the SAB PGS is operated for

several reasons. First, the SAB complex, being one of few carbon-free resources in Ontario with

black-start capability, energizes a portion of the grid without being dependent on an outside

electricity supply. It automatically adjusts output based on electronic signals to provide frequency

control and to maintain balance between demands. Second, the inherent nature of pumped hydro

operation allows it to serve as backup for intermittent sources by providing power when production

from these sources falls short of load. Third, regulated hydropower such as the SAB PGS can also

connect neighboring control areas for delivering electricity when such economic opportunities

arise. However, such interjurisdictional transactions are typically governed by transmission

0

0.5

1

1.5

2

0 500 1000 1500 2000 2500 3000

Mon

thly

pro

fit (m

illio

n $

CA

D)

Treaty flow restriction (m3/s)

A

D B C

42

capacity and energy prices in surrounding states/countries. On top of all of these factors, the SAB

PGS is unique in terms of its functionality. With its ability to rapidly move water in and out of the

reservoir, it maintains water elevation at the crossover, which is critical in ensuring appropriate

water diversion from the Niagara River (Maricic et al. 2009). It also complements the operations

of the two run-of-the-river hydropower plants at Niagara by maximizing available head.

Despite pumped storage’s potential, several technical, environmental, social and geopolitical

constraints have led to its under-utilization. Development and operation of hydro projects mandate

effective water resource management, which is complex and often requires consideration of a

broad range of social, economic, and environmental trade-offs. Being a transboundary water

system, the Niagara River faces more of those challenges in balancing various water needs. First,

effective water sharing among sovereign states, Canada and the US in the case of Niagara, requires

an agreement or contract between the parties. However, the tendency for the respective

governments to resist influence or control over assets challenges the very concept of shared

resources. Second, geographic and political issues surrounding the use of water resources are of

paramount interest when dealing with transboundary systems. Certainly the 1950 Treaty acts as a

major policy constraint for hydropower plants at Niagara. The expiration of the treaty in 2000,

which is currently being extended on an annual basis, opens the door for renegotiation with

opportunities for additional generation. However, such potential is seldom fully explored due to

complexity and negative public reaction against alteration of an age-old treaty. To make matters

even more complicated, neighbouring jurisdictions often have different priorities for conflicting

water uses. One possible example is the dismissal of the petition seeking hydrologic separation of

the Chicago Area Waterway from the Great Lake basin despite being identified as a potential

entryway for Asian Carp, an invasive species threatening the Great Lakes ecosystem. Third,

climate change consideration – frequently disregarded by policy makers – requires collaboration

for ensuring optimal use of water resources. Reflecting on such risks is indeed important for the

Great Lakes watershed given the sustained water level drop in the basin in late 1990s that has been

found to be related to El Niño events.

Water and energy systems are inextricably linked. This paper explores tradeoffs, the result of

which can be used for a possible renegotiation. With the current flow restrictions, the model

suggests how the selection of operating hours and flows would lead to profit maximization.

43

However, it refrains from making any direct suggestions based on the analysis, given that effective

water resource management requires consideration of a host of issues.

3.6 Conclusions and Recommendations Operational systems inherit various design, operational, and jurisdictional constraints that

complicate both the operation and redesign of components. The paper discusses such constraints

with respect to the transboundary river system at Niagara. Although pumped storage principally

operates for reasons of grid balancing, considerations such as relative cost, profitability, and long

term viability are also important. This exploratory study reflects on such perspectives and

formulates a revenue generation model for pumped storage operation. A 2.5–10% reduction in

electricity rate, due to increased pumped storage contribution, can result in a 1‒24% reduction in

profit depending on the month. Whereas increasing the reservoir footprint may bring little financial

gain, energy price variability and pumping-generation cycles appear to be the dominant factors in

PSH profitability. The conflict between flow targets and economic gain is quite strong for flow

restriction values between 0 and 1500 m3/s, but it is milder for increasing value of the treaty

restrictions. Energy revenues presented here are derived assuming operation of the facility in the

Ontario spot market, whereas capacity revenues such as for black-start and automatic generation

control are largely ignored. The profit values reported here are rough estimates only and depend

on the persistence of similar market and flow conditions.

The proposed model can be used by power authorities to evaluate the potential of pumped hydro

with respect to other emerging storage options such as CAES, batteries, and the like, once they

achieve the desired scalability. The authors expect this paper to contribute as a foundation for

further research on the role of pumped hydro in grid balancing. An interesting future extension of

this work may be the study of SAB PGS profitability due to the impact of changing time pattern

of price volatility with the integration of intermittent renewables.

44

Assessing the Financial Incentives for Pumped Storage Development

The economic viability of any project, whether existing or proposed, is imperative, regardless of

how indispensable a system may be at present, if its long-term sustainability is to be assessed.

Chapter 3 considered a marginal cost-based approach to evaluate pumped storage viability. The

limitation of the approach is that it disregards ancillary service-based revenues and their overall

impact on the investment; topics addressed in the current chapter. Considering that arbitrage, even

when combined with ancillary services valuation, may be a high-risk investment, various

supporting mechanisms are explored under which PHS projects might be developed. The analysis

further addresses some of the questions raised in chapter 2 by extending the discussion on

appropriate tariffs for hydropower.

Despite the increasingly large proportions of variable renewable energy (VRE) penetrations in

various electricity markets, there has been limited progress in grid-scale storage deployment.

While this can be attributed to the need for technological developments, it is exacerbated by the

absence of an integrated valuation framework that effectively and justly rewards storage operators

for the range of services they can provide to the grid. To this end, the author conducts a

comparative analysis among various financial mechanisms designed to support pumped storage,

such as contracted fixed price, marginal cost-based operations, and finally an integrated

socioeconomic cost-benefit model that better account for the social attributes (costs and benefits)

of storage.

This chapter is the basis of a planned paper titled “Analysis of Financial Incentives for Promoting

Pumped Storage Development” currently in final preparation. By analyzing the storage

remuneration structures and their underlying sensitivities, it addresses specific concerns and makes

recommendations for Ontario.

4.1 Introduction

The global effort to decarbonize electricity systems has led to widespread deployment of variable

renewable energy (VRE). However, increasing integration of these resources pose a considerable

45

threat to the grid (Barbour et al. 2016, Loisel et al. 2010, Wang et al. 2011a) as such generations

never mimic demand variability. The unpredictability associated with wind generation – currently

the fastest growing renewable energy (RE) ‒ if uncontrolled, may cause voltage and frequency

variations and affect power systems operations by inducing cyclic losses to conventional

generation units (Dursun and Alboyaci 2010; Georgilakis 2008). The challenges are largely

addressed by development and integration of energy storage systems (ESS). They not only

maximize the usage and benefits of VRE by reducing back-up from fossil fuel generators and

power curtailment measures but also provide ancillary services that are fundamental to network

reliability (Guittet et al. 2016). Balancing energy flows via ESS can improve power plant capacity

factors, provide flexibility to the grid through asset deferral and reduce grid congestion issues

(European Commission 2009). At present, pumped hydroelectric storage (PHS) is the only

commercially proven, grid-scale (>100 MW) storage technology that offers high roundtrip

efficiencies (75-82%), fast response-time (minutes to seconds), and a relatively long service life

(50-100 years) (Deane et al. 2010; StoRE 2014). It can operate in various possible modes that

stabilizes the baseload and the intermittent outputs, while its quick start capabilities make it

suitable for black-start as well as spinning and standing reserve (Bueno and Carta 2006; Kapsali

and Kaldellis 2010; Krajačić et al. 2013; Zeng et al. 2013).

Many authors address the use of PHS to permit increased VRE penetration. Anagnostopoulos and

Papantonis (2008), Canales et al. (2015), Caralis et al. (2010), Kaldellis et al. (2010),

Katsaprakakis et al. (2012) and Murage and Anderson (2015) presented algorithms for PHS,

designed to exploit wind energy surplus. In these studies, the authors show that PHS can have

excellent technical and economic performance while augmenting VRE penetration. Duic´ et al.

(2008) developed a methodology for assessing technical feasibility of integrated energy and

resource planning for island grid. The study asserted that PHS integration to the existing water

supply system could result in 25‒70% increased wind penetration. Benitez et al. (2008), Dursun

and Alboyaci (2010), Foley et al. (2015), NREL (2011) and Varkani et al. (2011) assessed the

potential of combined wind-pumped storage systems in meeting the electricity demand. A

national-scale energy system planning uses smart ESS for achieving a high share of wind and solar

in the supply system (Krajačić et al. 2011b; c), while a similar research also traces the reduction

in CO2 emissions (Cosi et al. 2012). Tuohy and O’Malley (2011) compared the Irish power systems

with and without pumped storage using a unit commitment model where PHS is shown to decrease

46

wind curtailment. Perez-Díaz and Jimenez (2016) assessed the impact of PHS on an isolated power

systems and reported a 2.5‒11% reduction in system scheduling costs. While these studies

emphasize the need for PHS for a secured, efficient and reliable grid, the high capital cost and the

absence of a valuation framework that remunerates storage for its range of services under the

current market structure discourage further development (Barbour et al. 2016; Zafirakis et al.

2013). While a study by Caralis et al. (2010) found the combined wind and PHS system to be cost-

competitive, the feasibility strongly depends on sustained highly volatile spot market price

(Salevid 2013) and without subsidies, rarely achieves economic sustainability (Locatelli et al.

2015; Melikoglu 2017).

To address this concern, contemporary literature has proposed and demonstrated application of

different supporting mechanisms for PHS (Kaldellis and Zafirakis 2007; Krajačić et al. 2011a;

Krajačić et al. 2013; Zafirakis et al. 2013). While Braun (2016), Malakar et al. (2014) and Tahseen

and Karney (2016) (Chapter 3) discussed short-term optimization for maximizing operational

profit considering intraday auction markets, Krajačić et al. (2011a; 2013) proposed a financial

scheme that rewards storage for discharging wind-originated surplus. Sandhya and Baker (2012)

and Díaz-González et al. (2012) analyzed the cost of different storage technologies that can

facilitate wind power integration. Kaldellis and Zafirakis (2007) and Zafirakis et al. (2013)

estimate “break-even” FITs by comparing social attributes (cost and benefits) of ESS with

electricity production costs. While the published approaches discuss application of these schemes,

neither runs a comparative analysis to assess their potential to recover costs.

To this end, this chapter analyzes the feasibility of alternative pricing strategies for a wind-based

PHS system in Ontario. The approaches discussed here are contracted fixed price, marginal cost-

based operations in the spot market, and finally a socioeconomic cost-benefit model that accounts

for the social costs and benefits offered by PHS. The use of appropriate supporting schemes for

storage is important as it transfers energy surplus (occurring when energy production is higher

than demand) from a period of excess to when there is a lack. With ample use of storage, VRE

could be used to meet peak demands, otherwise commonly satisfied by conventional thermal plants

that are associated with high cost and severe externalities. It further compensates for wind energy

curtailments by effectively deferring these amounts to peak hours. The current study presents the

47

perspective of PHS operators with the understanding that payment level acts as a floor basis for

decision-making to contract fixed payments or to act through price arbitrage and reserve provision.

The location of choice for the case study is perhaps unique, since Ontario experiences a relatively

low (≈12%) RE penetration. However, the hybrid market structure along with the rapidly

expanding wind capacity (Amor et al. 2014) makes it a challenging yet intriguing case. The paper

next describes the test system and the market and then elaborates on the financial models.

4.2 Combined Wind and Pumped Storage System Operation of a PHS is based on the principle that the system absorbs electrical energy when there

is a surplus (when demand is lower than supply), storing it in the form of gravitational potential

energy which is later released during periods when demand exceeds supply. Conventional PHS

uses two water reservoirs at different elevations where water is pumped from the lower to the upper

reservoir during off-peak hours. These charging-generating cycles are undertaken diurnally or

seasonally, although the latter requires a relatively large storage capacity. The energy losses in the

energy conversion process make it a net energy consumer (IPCC 2011), typically operating at an

efficiency in the 70–80% range (Deane et al. 2010; Letcher 2016; Levine 2003). For diurnal

operation, PHS usually takes advantage of arbitrage strategies based on price differences between

low (when energy excess normally appears) and peak demand periods (when prices increase

considerably). However, the high investment costs needed for the construction of reservoirs,

preferably close to consumption, are not always compensated by this profit margin (Steffen and

Weber 2016). This research investigates the economic implications of coordinated operation of

pumped storage and wind where both plants are further separately connected to the grid. The

arrangement allows PHS a choice between wind power and grid electricity for pumping, while

reducing curtailments for wind operators through opportunistic contribution to PHS charging.

The electricity system in Ontario is hybrid in the sense that energy is bought and sold in a semi-

competitive wholesale market, while planning and procurement are conducted through long-term

contracts. The wholesale market is used to guide dispatch decisions where generators submitting

bids to the system operator are dispatched from the lowest bid until the demand is satisfied (IESO

2017). The electricity price has three components—the market price, known as the Hourly Ontario

Electricity Price (HOEP), the Global Adjustment (GA) factor, and an “uplift charge” that recovers

48

the costs of ancillary services (IESO 2017). While HOEP is driven primarily by the marginal fuel

cost of production, the GA covers all the remaining fixed generation costs that are committed

through contracts or regulation. The market is jointly optimized for schedule and price where

twelve Market Clearing Prices (MCP), generated at every five minutes in each dispatch hour, are

averaged to calculate the HOEP. However, these prices are subjected to steady decline due to

Ontario’s aggressive renewable energy program in the form of Feed-in Tariff (FIT), while the GA

has gradually increased as more of the generation costs are recovered outside the market (Rivard

and Yatchew 2016). Similar effects have been observed in Germany where diminishing price

variations led to 70% reduced profit through arbitrage (Steffen and Weber 2016).

4.3 Methodology This section discusses different pricing models for PHS remuneration. Perhaps the most common

of these is the marginal cost (MC)-based operations, also known as arbitrage, where storage

operators optimize against an electricity wholesale price curve. Another scheme offers guaranteed

payments through long-term contracts. Examples of these are capacity development through FIT,

request for proposal (RfP) etc. The last approach considered here is a socio-economic cost benefit

model that determines “socially just” FITs by rewarding storage for social welfare attributes. Here,

the break-even FITs (BEFITs) that equate social costs and benefits are compared with electricity

production cost of ESS in order to investigate return on the investment. This section elaborates

application of these strategies to a hypothetical wind-based pumped storage station in Ontario. The

analysis uses reported values in the published literature for capital ($1,000‒2,000/kW) and O&M

costs ($3‒7.7/kW per year) (Dames and Moore 1981; Deane et al. 2010; United States Department

of Energy 2015) and adjusts them to 2015 CAD dollars. Table 4.1 provides a complete list of

variable names and their values. The study considers a diurnal (at least) cycling of PHS with the

underlying assumption that the daily peak demand requires mandatory pumped storage

contribution and on the process, replaces either natural gas or oil-based plants. In this regard,

energy surplus, deriving either exclusively or at a minimum permitted contribution from RE farms,

is used for charging the system. Greater details about the modelling approach can be found in

various cited references.

49

Table 4.1: Summary of input parameters

Parameter Values

Initial investment, Ccapital ($/kW) 1,200

Annual O&M, CO&M ($/kW) 5.5

Efficiency of PHS, 𝜂𝜂 (%) 75

Compensation rates, CRE ($/MWh) 51.2

Interest rate, i (%) 5

Average energy price, HOEPavg,pump ($/MWh) 15.3

Service life, 𝑛𝑛𝑚𝑚𝑚𝑚𝑚𝑚 25

Duration of premium, 𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 5

Service period prolongation factor, 𝜀𝜀 (%) 5

Electricity generation cost, 𝐶𝐶𝑡𝑡𝑡𝑡𝑡𝑡𝑚𝑚𝑡𝑡,𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝 ($/MWh) NG-75.2, oil-132.7

Fixed operating cost, CO&M,rest ($/MWh) NG-2.4, oil-2.3

Efficiency 𝜂𝜂𝑑𝑑 (%) NG-40, oil-35

Calorific value, Hu (kWhfuel/kgfuel) NG-50, oil-46

Average fuel price, Pfuel ($/kgfuel) NG-0.24, oil-0.39

Capital depreciation,𝐶𝐶𝑑𝑑𝑝𝑝𝑝𝑝𝑑𝑑𝑝𝑝𝑑𝑑𝑖𝑖𝑚𝑚𝑡𝑡𝑝𝑝 (%) 2

Net CO2 emission factor, 𝜀𝜀𝐶𝐶𝐶𝐶2 (KgCO2/MWh) NG-490, oil-735

Price of CO2 allowances, PCO2 ($/kgCO2) 0.05

Pollution related damages, bex ($/MWh) NG-19.7, oil-35.33

4.3.1 Marginal cost based operation in the spot market The idea of pre-scheduling PHS generation using forecasted data on demand and subsequent price

is not new. Since PHS benefits from price arbitrage in a spot market operation, selection of the

specific time for pumping and generation is critical to operation. Here, the author uses a profit

50

maximization model by Tahseen and Karney (2016) (Chapter 3) where decision variables

representing pumping and generating hours are binary, while inflow and outflow to/from the

reservoir are varied over allowable ranges. Time delays for the transition from pumping to

generating sequence are provisionally assumed to be negligible. To represent the volume of water

stored in the reservoir at start of hour i, the authors introduce a storage function S and assume it to

be empty (0) at the beginning of the operation. The model reflects a rather long list of constraints

beginning with typical reservoir limitations such as capacity, flow balance, and so on. Prior to

running the model, Ontario electricity price data for 2009 and 2015 were extracted from the IESO

(2017). This choice is motivated by the reported $7 decrease in HOEP between the corresponding

years (Rivard and Yatchew 2016). The hourly energy price data is then processed for days with

minimum and maximum standard deviations for each month. Following the identification, the

hourly data for the representative days is extracted from the original dataset and is used as input to

the model. Generally, in a spot market framework storages are offered added remunerations for

their ancillary services (Black and Veatch 2012; Paine et al. 2014). The Ontario ancillary services

market includes payments for black-start and frequency regulation. As a black-start facility, PHS

receives fixed monthly payments whereas regulation service has both fixed and variable cost

components (IESO 2016c).

Despite its popularity, arbitrage strategies based on wholesale price are often considered a high

risk investment largely because of imperfect long-term price prognosis and forecast (Connolly et

al. 2011; Weron and Misiorek 2008). Particularly in a thermal (nuclear) power-dominated market

like Ontario, the optimal dispatch (of nighttime pumping and peak generation) is often less

obvious, thus making arbitrage, even when combined with ancillary services valuation, a high-risk

investment.

4.3.2 Contracted fixed price per unit of electricity At present, supporting mechanisms that guarantee price through long-term contracts are used to

promote faster market integration of newer energy technologies. Being the most popular RE

supporting scheme, FIT now exists in many jurisdictions across Europe and North America. FIT

schemes offer a fixed price per unit (kWh or MWh) of renewable electricity delivered to the grid,

where the rates are usually determined based on technological maturity and local RE potential.

51

Despite the success of FITs, RE deployment is often limited due to the grid’s inability to balance

variable generation (Quansah et al. 2016; Solar Trade Association 2016). Thus, despite its

technical maturity, PHS justifies promotion through FIT mechanism.

As summarized in chapter 2, the Ontario FIT (version 5.0) offers 0.246 $/kWh for water projects

up to 500 kW (IESO 2016a) with 35% on-peak premium and a 10% decreased payments for off-

peak generations (IESO 2016b). Since the latest revisions introduce a 500 kW capacity restriction,

the analysis uses a scaled-down rates from FIT 2.1 (IESO 2010) that allowed larger projects. It

evaluates cost recovery period under the contract price of 98 $/MWh with premiums. The plant

primarily charges using wind-based electricity; however, the contribution of wholesale electricity

to pumping operations is also explored.

A variation of FIT by Krajačić et al. (2011a) and Krajačić et al. (2013), called FIT with guarantees

of origin (FIT_GO), offers additional remuneration to PHS operators for supporting RE power

through system charging. The FIT_GO rewards PHS discharge through long-term contracts when

the origin of supply is wind-based. In this context, FIT ($/MWh), i.e., the rate paid for electricity

produced by PHS, can provide top-up for the percentage of wind contributions during pumping

hours.

FIT = �𝐶𝐶𝑑𝑑𝑚𝑚𝑝𝑝𝑖𝑖𝑡𝑡𝑚𝑚𝑡𝑡.𝑅𝑅 + 𝐶𝐶𝐶𝐶&𝑀𝑀

𝐸𝐸� + 𝛼𝛼𝑅𝑅𝑅𝑅 ∙ �

𝐶𝐶𝑅𝑅𝑅𝑅𝜂𝜂� + (1 − 𝛼𝛼𝑅𝑅𝑅𝑅) ∙

HOEPavg,pump

𝜂𝜂

(4.1)

where Ccapital is initial investment in storage, CO&M is yearly PHS operation and maintenance costs,

R is annuity factor, 𝜂𝜂 is PHS efficiency, E is annual electricity generation by PHS, αRE is fraction

of wind generation during pumping and HOEPavg,pump is the corresponding average wholesale

market price. CRE represents the compensation offered to wind generators (for its contribution in

pumping) and is assumed to be 40% of the current FIT rates for wind power. Since system charging

utilizes wind power that with high VRE penetration would otherwise be curtailed, there is a

business case for having CRE lower than the market price (i.e., the FIT rates). Eventually, the

FIT_GO scheme is beneficial for both RE farms and PHS operators as it results in a lower charging

fee while allowing farms to receive compensation for curtailed RE power.

52

4.3.3 Socioeconomic model Realizing its potential benefits, recent publications have emphasized on assigning social attributes

to the analysis of storage (Sioshansi 2010; Sioshansi et al. 2009). To this end, the author uses an

integrated cost-benefit model by Kaldellis and Zafirakis (2007) and Zafirakis et al. (2013) to

determine the break-even FITs (BEFITs) by equating social costs and benefits. The BEFITs are

then compared with electricity production cost of PHS to investigate profit margin for the

respective investment. Here, the social costs correspond to support provided to storage while

benefits represent either avoided costs or direct social benefits derived from the operation using

renewable power (wind).

4.3.3.1 Determination of social supports/costs Initial cost subsidy One way of supporting the typically high development cost of PHS is to subsidize a portion of the

initial capital investment (Ccapital). The corresponding social support c1 (per MWh of electricity

delivered annually by PHS) can be expressed by the following equation:

𝑐𝑐1 = 𝛾𝛾

𝑛𝑛𝑚𝑚𝑚𝑚𝑚𝑚∙𝐶𝐶𝑑𝑑𝑚𝑚𝑝𝑝𝑖𝑖𝑡𝑡𝑚𝑚𝑡𝑡𝐸𝐸

(4.2)

where 𝛾𝛾 is initial cost subsidy (%) and nmax is service life of PHS, taken at a constant (and

conservative) value of 25 years. Since Ontario currently does not offer any subsidy to PHS

developments, the study performs a sensitivity analysis to understand its impact on investments.

Guaranteed power premium As grid reliability depends on the balance between demand and supply on a momentary basis,

dispatchable sources such as PHS are often valued over intermittent or relatively steady

conventional power. The Ontario FIT offers 35% on-peak premium for dispatchable generators

(IESO 2016b). The time-differentiated price is converted to a fixed annual premium 𝛿𝛿𝛿𝛿𝑁𝑁/𝑚𝑚 ($/MW

year) that can be applied for desired years of operation (nsubs). The net power premium c2 ($/MWh)

is calculated using the following equation:

𝑐𝑐2 = 𝛿𝛿𝛿𝛿𝑁𝑁/𝑚𝑚𝑁𝑁𝑡𝑡𝑡𝑡𝑚𝑚𝑑𝑑𝐸𝐸

∙𝑛𝑛𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑛𝑛𝑚𝑚𝑚𝑚𝑚𝑚

53

(4.3)

where 𝑁𝑁𝑡𝑡𝑡𝑡𝑚𝑚𝑑𝑑 is power output to the network in MW.

Tax credits Several researchers have recommended decreased tax rates and dedicated gratuitous loans for

incentivizing investments in RE and storage technologies (Kazempour et al. 2009; Lu et al. 2011).

At present, Ontario does not have any dedicated tax credits program for PHS.

4.3.3.2 Determination of social benefits Peak power station replacement by ESS Since the highest loads in a distribution system occur only during a fraction of a year, peak power

plants remain underutilized for majority of its service life. Replacing or delaying these more

expensive units, generally running on natural gas or oil, is a key benefit (or avoided cost) offered

by PHS (US Department of Energy 2013). Apart from their high operating costs, conventional

thermal plants operate at relatively low load factors and low efficiency. The resulting benefits (or

avoided costs) b1 is estimated by adding avoided operating cost of the replaced/delayed station

(𝐶𝐶𝐶𝐶&𝑀𝑀,𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝) and a percentage (𝜀𝜀 = 5%) of the constant cost reduction resulting from service period

prolongation of the peaking plants (Zafirakis et al. 2013).

𝑏𝑏1 = 𝐶𝐶𝐶𝐶&𝑀𝑀,𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝 + 𝜀𝜀 ∙ �𝐶𝐶𝑡𝑡𝑡𝑡𝑡𝑡𝑚𝑚𝑡𝑡,𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝 − 𝐶𝐶𝐶𝐶&𝑀𝑀,𝑝𝑝𝑝𝑝𝑚𝑚𝑝𝑝� (4.4)

Here, Ctotal,peak is total electricity generation cost of the replaced/delayed station and 𝜀𝜀 is service

period prolongation factor. Here, CO&M,peak has two components corresponding to the fuel cost

(Cfuel) and rest of the operating costs (CO&M,rest). Depending on the efficiency (𝜂𝜂𝑑𝑑) and calorific

value Hu (kWhfuel/kgfuel) of the fuel consumed, Cfuel can be estimated by the following equation:

𝐶𝐶𝑓𝑓𝑠𝑠𝑝𝑝𝑡𝑡 = 𝑃𝑃𝑓𝑓𝑠𝑠𝑝𝑝𝑡𝑡𝜂𝜂𝑑𝑑 ∙ 𝐻𝐻𝑠𝑠

(4.5)

where Pfuel ($/kgfuel) is average fuel price for the study period. Relevant data for Ontario are

collected from EIA (2015), Tidball et al. (2010), Ux Consulting Company (2017), World Nuclear

Association (2016) and Zafirakis et al. (2013) and then processed for further calculations.

54

Taxation of ESS Annual taxes paid by PHS on the basis of net cash flows is another source of social benefit. 𝑇𝑇(𝑗𝑗)

describes the taxes paid on previous year’s revenue (𝑅𝑅(𝑗𝑗−1)) accruing from the remuneration of

energy production and guaranteed power.

𝑇𝑇(𝑗𝑗) = 𝜑𝜑𝑠𝑠𝑠𝑠(𝑗𝑗) ∙ �𝑅𝑅(𝑗𝑗−1) − 𝐶𝐶𝑝𝑝𝑠𝑠𝑚𝑚𝑝𝑝(𝑗𝑗−1) − 𝐶𝐶𝐶𝐶&𝑀𝑀(𝑗𝑗−1) − 𝐶𝐶𝑑𝑑𝑝𝑝𝑝𝑝𝑑𝑑𝑝𝑝𝑑𝑑𝑖𝑖𝑚𝑚𝑡𝑡𝑝𝑝(𝑗𝑗−1)� (4.6)

where, 𝜑𝜑𝑠𝑠𝑠𝑠 is a law-defined tax-coefficient on previous year’s net cash flow excluding pumping

cost (𝐶𝐶𝑝𝑝𝑠𝑠𝑚𝑚𝑝𝑝), estimated based on varying contributions from wholesale electricity and wind energy

and capital depreciation (𝐶𝐶𝑑𝑑𝑝𝑝𝑝𝑝𝑑𝑑𝑝𝑝𝑑𝑑𝑖𝑖𝑚𝑚𝑡𝑡𝑝𝑝), assumed to be 2% of the initial investment. 𝑇𝑇𝑛𝑛𝑝𝑝𝑡𝑡(𝑗𝑗) is the net

benefit from taxation after deducting taxes paid by the replaced peak power plants. Due to high

levels of uncertainty in the latter, the analysis uses a 25% net taxation coefficient (𝛿𝛿),

complemented with a sensitivity analysis in the later sections.

𝑇𝑇𝑛𝑛𝑝𝑝𝑡𝑡(𝑗𝑗) = 𝛿𝛿𝜑𝜑𝑠𝑠𝑠𝑠(𝑗𝑗) ∙ �𝑅𝑅(𝑗𝑗−1) − 𝐶𝐶𝑝𝑝𝑠𝑠𝑚𝑚𝑝𝑝(𝑗𝑗−1) − 𝐶𝐶𝐶𝐶&𝑀𝑀(𝑗𝑗−1) − 𝐶𝐶𝑑𝑑𝑝𝑝𝑝𝑝𝑑𝑑𝑝𝑝𝑑𝑑𝑖𝑖𝑚𝑚𝑡𝑡𝑝𝑝(𝑗𝑗−1)� (4.7)

Finally, the total benefit b2 ($/MWh) is estimated by dividing the net tax gain by annual generation.

Avoided carbon dioxide allowances PHS may avoid emission costs by replacing or delaying fossil fuel generations provided that the

system is charged with renewable power (Silva and Hendrick 2016). The avoided carbon cost (b3)

due to the recovery of wind curtailments is estimated by multiplying net CO2 emission coefficient

(𝜀𝜀𝐶𝐶𝐶𝐶2) (Schlömer et al. 2014) of the replaced thermal stations, considering also any emissions

derived from PHS operation, and the price of CO2 allowances (PCO2).

𝑏𝑏3 = 𝜀𝜀𝐶𝐶𝐶𝐶2 ∙ 𝑃𝑃𝐶𝐶𝐶𝐶2 (4.8)

A carbon price of 50 $/tCO2 is considered according to Canada’s Federal Carbon Price Plan for

2022 (Carbon Tax Center 2017) and is further varied for sensitivity analysis.

Avoided negative externalities The study accounts for the net social benefit from avoided negative externalities of electricity

production. The net external cost bex ($/MWh) i.e., the negative externalities attributed to the use

55

of conventional power minus that of PHS and wind energy (for charging), is obtained from

reported values in the published literature (European Commission 2005; 2008; Georgakellos 2012;

Markandya 2012). Since the values in the aforementioned studies are specific to Europe, they are

prorated (≈10% reduction) considering Ontario’s low population density. The studies apply the so-

called ExternE methodology that traces the damage caused by harmful by-products of electricity

generation and converts them into monetary values. The approach begins with modelling emission

dispersion and estimating their impacts on health, building materials, crops biodiversity and so on

(climate impact is considered in the earlier section).

Finally, the total social benefits per MWh of electricity produced by wind energy-based PHS is

estimated by the following equation:

𝐵𝐵𝑡𝑡𝑡𝑡𝑡𝑡𝑚𝑚𝑡𝑡 = (𝑏𝑏1 + 𝑏𝑏3 + 𝑏𝑏𝑝𝑝𝑚𝑚) ∙ 𝑘𝑘𝑤𝑤 + 𝑏𝑏2 (4.9)

where 𝑘𝑘𝑤𝑤 represents annualized contribution from wind farms in systems charging. Thus, the FIT

rates that ensure balance between the social costs and the benefits (BEFITs) are estimated as:

BEFITs = 𝐵𝐵𝑡𝑡𝑡𝑡𝑡𝑡𝑚𝑚𝑡𝑡 − 𝑐𝑐1 − 𝑐𝑐2 (4.10)

The BEFITs are then compared with electricity production cost of PHS to determine the required

support for overall profitability.

4.4 Analysis and Results

4.4.1 Marginal cost based operation in the spot market This section presents the outcome of PHS operation in the wholesale and ancillary service-based

market. From arbitrage perspective, the worst-case scenario is a relatively stable price, regardless

of high or low, that offers little variations on a 24-hour basis. The optimization model focuses on

such critical cases by running each time with the minimum, maximum, 25th, 50th and 75th percentile

HOEP for each month of 2009 and 2015. Figure 4.1 illustrates the price duration curves for 2012‒

2015. It indicates relatively high peak demand prices, exceeding 250 $/MWh, during which the

peak power plants or energy imports are called to cover the load. The off-peak prices, ranging

between -100–50 $/MWh, correspond to late night and early morning hours when PHS is typically

56

charged. Furthermore, there is a decline in HOEP from 2012 to 2015 as more of the (renewable)

generation are now being procured outside the market (Rivard and Yatchew 2016).

Figure 4.1: The electricity price (HOEP) duration curves for four consecutive years (2012–2015)

Following the dispatch optimization model, the minimum daily profit (from January to December)

in 2009 ranges from $0‒9,700 CAD which may rise up to $315,300 CAD (Figure 4.2). Similar

analysis for the year 2015 results in a slightly lower minimum profit ($30‒7,000 CAD). However,

interestingly the maximum daily profit shows a 2.5‒74% increase compared to that of 2009 (Figure

4.2). These results confirm increasing profit opportunities through arbitrage with the rising price

volatility resulting from increased VRE penetration in Ontario. The optimization model though

provides little scope for capacity factor (CF) variations, the analysis confirms an average 3‒4 hour

daily generation (CF≈15%). The profits are then averaged over month to get an approximate

annual income in the wholesale market. The information gap in the Ontario ancillary market

framework led to the assumption of a range of possible values between 500‒4,500 $/MW for fixed

monthly payments. As expected, operating in the ancillary market achieves greater revenue due to

black-start and regulation service payments (on top of marginal cost-based settlements). With the

500 $/MW ancillary payment, the profit increases by 4‒76% on a monthly basis.

0

20

40

60

80

100

120

-100 -50 0 50 100 150 200 250

Ann

ual p

roba

blity

(%)

HOEP ($/MWh)

2015 2014 2013 2012

57

Figure 4.2: Monthly profit (excluding capital costs) based on spot and ancillary service-based

(with 500 $/MW) market operation

Figure 4.3 illustrates the payback on PHS investment following its operation in the wholesale and

ancillary services market. Here, the discounted payback period is calculated with due consideration

for initial capital, annual O&M (values reported in the methodology section) and system charging

cost. Interestingly, the return period increases with increasing wind contributions in system

charging. A 30% increased wind utilization leads to 1.2−11.4 yr increase in payback period under

varying ancillary payment structures. In general, FIT rates are closely tied to wholesale market

price, offering somewhat higher rates for incentivizing development in relatively new, green

technologies. The compensation rates for wind curtailments ‒ assumed to be 40% of the current

FIT rates ‒ allow wind operators to be rewarded for the lost opportunity (who would otherwise

face curtailment for free), thus resulting in a lower system charging cost. This makes wind-

generated electricity a preferred option for system charging over their wholesale alternative. The

analysis here draws attention to the fact that Ontario’s energy market makes quite the reverse case

where PHS becomes increasingly cost ineffective with growing wind energy contributions. The

analysis further suggests that the payments offered in the combined market operations are barely

enough for a reasonable return on PHS investment. While a fixed ancillary payment of 2,000

$/MW on top of the spot market profit yields a 28−47 yr return period (in 2015) under changing

wind contributions, any rate below 900 $/MW rarely achieves profitability at CF below 15%.

0

50

100

150

200

250

300

350

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Dai

ly p

rofit

(x10

3 ) ($

CA

D)

2009_min 2015_min 2009_max 2015_max

58

Figure 4.3: Discounted payback period under the wholesale and ancillary service market (at CF =

15% and i = 5%)

4.4.2 Contracted fixed price per unit of PHS electricity Figure 4.4‒4.6 apply 98 $/MWh fixed contract price with 35% peak premium (in line with Ontario

FIT) for PHS electricity and demonstrate the resulting impact on payback period with varying

wind energy contributions for charging. The discounted payback period diminishes with increasing

capacity factor (Figure 4.4) suggesting that an increased grid penetration leads to a more

economical system. The return period is also influenced by pumping cost, estimated based on

varying contributions from grid electricity (using wholesale market price) and wind energy (using

compensation rates for wind). The wholesale market price in Ontario being substantially low, there

is a 1−10 yr increase in return period (higher values corresponding to low CFs) with 30% increased

wind exploitation. With the pumping cost exclusively based on HOEP, the return period ranges

between 10‒17.4 yr with CF varying between 20‒35%. Similar results for a (full) wind electricity

charged system are between 13‒38.7 yr (Figure 4.4). The grey line representing the zero pumping

cost in Figure 4.4 is for comparison purposes, in particular with respect to demand response which

is a major competitor to storage. Also, this investigates the possibility of a wind-based PHS system

in a decentralized grid where both the assets are operated to complement each other.

0

20

40

60

80

0 1000 2000 3000 4000 5000

Payb

ack

perio

d (y

r)

Ancillary payments ($/MW capacity)

0% wind-2009 0% wind-2015 30% wind-200930% wind-2015 60% wind-2009 60% wind-2015

59

Figure 4.4: Payback period with changing capacity factors considering contract price, on-peak

premium and different pumping costs

Figure 4.5 demonstrates the impact of changing compensation rates on payback period under a

constant (30%) wind energy contribution (for charging). The increase in compensation rates

contribute to a larger pumping cost, thus resulting in a higher return period as observed in this

analysis. For example, a 15% increase in these rates yields 0.8‒12.4 yr increase in return period

under varying PHS contributions to the grid. While the rates have substantial effect on project

economies at a relatively low capacity factor, the 2 yr difference in return period with CF = 35%

suggests that the impacts might be trivial for large systems.

Figure 4.5: Payback period with varying compensation rates to the wind operators

0

20

40

60

10 15 20 25 30 35 40

Payb

ack

perio

d (y

r)

Capacity factor (%)

FIT + premiumFIT + premium - 0% windFIT + premium - 50% windFIT + premium - 100% wind

0

15

30

45

60

75

30 40 50 60 70 80

Payb

ack

perio

d (y

r)

Compensation rates (%)

CF = 15

CF = 20

CF = 35 CF = 25

60

Figure 4.6 illustrates the effect of changing interest rates on PHS economies. Here, the interest

rates each time are varied between 3‒7% with changing wind energy contributions to charging. As

expected, increasing interest rates lead to a rise in payback period. While the effect may not be

substantial at a low wind penetration scenario, the return period increases rather rapidly for an

expensive system. For example, the return period increases by 23.3 yr with a 4% increase in

interest rates for a wind-based PHS system, whereas the same with a low system charging cost

(with 30% wind energy exploitation) increases by 4 yr only.

Figure 4.6: Payback period with varying interest rates

4.4.3 FIT with guarantees of origin (FIT_GO) While previous analysis (Figure 4.4‒4.6) set the contract price for water projects at 98 $/MWh,

Figure 4.7(a, b) focuses on determining these rates based on FIT with guarantees of origin

(FIT_GO). Figure 4.7(a) estimates the changing FIT_GO rates with varying capacity factors that

ensure a 15 yr payback period. A 20% increase in capacity factor leads to a 40‒57% decline in the

rates where the lowest decrease corresponds to high wind penetration scenario. While increasing

capacity factors diminish the contract rates, the growing RE contributions to pumping has the

opposite effect; a 50% rise in wind penetration increases the FIT rates by 6.4‒12.4% under various

CFs. Furthermore, the estimated FIT_GO rates are below 98 $/MWh in only 4 out of 40 cases

(combination of 8 different wind exploitation rates and 5 CFs) analyzed here ‒ that too, for three

cases without pumping cost considerations and the other at a maximum CF of 35% ‒ suggesting

that peak premiums are crucial for a reasonable return on PHS investment.

0

10

20

30

40

50

2 3 4 5 6 7 8

Payb

ack

perio

d (y

r)

Interest rate (%)

FIT + premium FIT + premium - 0% windFIT + premium - 50% wind FIT + premium - 100% wind

61

Figure 4.7: FIT rates (i = 5%) under varying (a) CFs and wind contributions (15 yr return year)

(b) return period and wind contributions (25% capacity factor)

Next, the author estimates FIT_GO prices for a range of possible payback periods at a constant CF

(25%) and interest rate (5%) (Figure 4.7b). As expected, the longer return periods allow spreading

out the cash flows, thus resulting in lower contract prices. A 5 yr increase in return period leads to

a 6.9‒16% decline in the rates under varying degrees (0‒100%) of RE utilization. With

remuneration below 100 $/MWh, no system covers the investment within 20 yr of its operation

(when operating at CF = 25%) and requires a minimum of 110‒207 $/MWh for various wind

electricity charged systems. The estimated remunerations vary between 95‒181, 118‒205, 137‒

223, and 192‒278 $/MWh for 0, 30, 50 and 100% wind energy exploitations respectively, thus

suggesting a 9‒15.4% increase in FIT rates for each 20% increase in wind contributions.

0

50

100

150

200

250

5 10 15 20 25 30 35 40

FIT_

GO

rate

s ($/

MW

h)

Capacity factor (%)

0% wind

w/o pump

100% wind50% wind

0

50

100

150

200

250

300

0 10 20 30 40

FIT_

GO

rate

s ($/

MW

h)

Discounted payback period (yr)

(a)

(b)

w/o pump

100% wind

50% wind 0% wind

62

4.4.4 Socioeconomic cost-benefit model The analysis here investigates the impact of replacing different peak power stations: natural gas-

fired combined cycle (CC) and oil-based power plants. Figure 4.8‒4.9 compare wind-based PHS

system with natural gas and petroleum-based power respectively. The BEFITs are equated with

the corresponding electricity production cost of PHS, in relation to the variation in wind energy

contribution and CF. The break-even point, i.e., the intersection between the BEFITs and the

electricity production cost defines the value of appropriate support mechanisms at respective

annual PHS contribution that allows the system to be cost-effective. Both sets of curves (BEFIT

and production cost) follow asymptotical pattern with reduced electricity production cost at higher

values of CF. The analysis explores a variety of annualized wind contributions in system charging

(kw) where rising kw leads to increasing social benefit (BEFIT) in terms of peak power

replacement, avoided carbon emission and negative externalities. The cost curves converge at 60–

68 $/MWh suggesting that PHS operating at a high capacity factor would experience little increase

in production cost with increasing wind exploitation (kw) for charging. When PHS replaces gas-

fired CC plants to be fully charged with wind electricity (Figure 4.8), the critical annual

contribution (or CF) from PHS must be above 34%, i.e., 2980 h of annual generation to the grid.

When wind energy exploitation is reduced to 80%, the break-even point occurs at CF = 37% as a

result of notable reduction in the BEFIT curve, dropping from a maximum of 70.3 $/MWh for kw

= 100% to 55.2 $/MW h for kw = 80%. While increasing wind contributions result in rising prices,

there is potential to reduce support with increasing annual PHS contribution to the grid. At kw

below 80%, PHS may not be cost-effective considering the low social benefits derived from peak

power displacement, avoided emission and negative externalities.

Similar results for PHS replacing or delaying oil-fired diesel plants are demonstrated by Figure

4.9. In case of 100% wind energy exploitation for pumping, the critical capacity for a cost-effective

system is around 17% at a FIT price of 149 $/MWh. With reduced wind exploitation rate, these

supports decrease to 118 $/MWh (at kw = 80% and CF = 19%) and 87 $/MWh (at kw = 60% and

CF = 23%). At wind contributions below 30%, PHS proves to be cost-ineffective with the

supporting scheme due to the reduction in BEFIT derived from the reduction of social benefits. As

electricity production cost curves converge at a relatively high value of CF, there should be an

optimum point defined by minimum electricity production cost and maximum BEFIT. This would

63

correspond to a marginally cost-effective configuration, producing a profit in the order of 25

$/MWh for kw = 60% and CF= 35%.

Figure 4.8: Comparison between BEFITs and electricity production cost when replacing natural

gas-fired CC plants

Figure 4.9: Comparison between BEFITs and electricity production cost when replacing oil-fired

plants

0

50

100

150

200

0 10 20 30 40

BEF

ITs o

r CPH

S ($

/MW

h)

Capacity factor (%)

Natural Gas

BEFIT (kw=60%) Production cost (kw=60%)BEFIT (kw=80%) Production cost (kw=80%)BEFIT (kw=100%) Production cost (kw=100%)

0

50

100

150

200

0 10 20 30 40

BEF

ITs o

r CPH

S ($

/MW

h)

Capacity factor (%)

Petroleum

BEFIT (kw=60%) Production cost (kw=60%)BEFIT (kw=80%) Production cost (kw=80%)BEFIT (kw=100%) Production cost (kw=100%)

64

Figure 4.10 illustrates the impact of state-subsidy on BEFIT and electricity production cost curves.

Increase in state-subsidy (γ) not only reduces production cost from 163 $/MWh to 146 $/MWh

(for kw = 60% and CF = 15%) but also decreases the BEFIT. For example, 15% subsidy on initial

investment leads to a 9.3 $/MWh reduction in the BEFITS for both CC and oil-based plants. With

the subsidy having a diminishing effect on both the systems cost and the BEFIT, little variations

are realized in the critical capacity ensuring cost effectiveness of the system. A 15% subsidy results

in a mere 1% difference in critical capacity (at kw = 60%) when compared with an unsubsidized

system. The point of maximum profit (around 28 $/MWh) under kw = 60% and γ = 15% extends

the capacity factor beyond 35% which may be too long for annual peak demand duration.

Figure 4.10: Comparison between BEFITs and electricity production cost when replacing oil-

fired plants

4.4.5 Model comparison and sensitivity analysis In this section, different pricing models are compared with respect to their rates and effectiveness

in cost recovery (return period). Table 4.2 lists the values given 2,200 hour annual generation, 5%

interest rate, 60% wind exploitation and 51.2 $/MWh wind electricity charge. The price range

provided under the wholesale market corresponds to the variation in HOEP during peak demand

hours (9‒18 h) which is also the revenue PHS accrues for generating during these hours. The

resulting return period when operating in the wholesale and ancillary services market (with 1,500

$/MW fixed payments) is 2 times higher in comparison to the other schemes assessed in this study.

While the analysis of the FIT and the socioeconomic model converges to a contract payment

0

30

60

90

120

150

180

0 10 20 30 40BEF

ITs o

r CPH

S ($

/MW

h)

Capacity factor (%)

Petroleum

BEFIT (γ=0%) Production cost (γ=0%)BEFIT (γ=10%) Production cost (γ=10%)

65

around 95 $/MWh with a 35% peak premium, the cost-benefit model suggests increasing the

remuneration with rising wind exploitations (for system charging) and decreasing capacity factors.

For example, the model justifies a rate of 155 $/MWh at 15% capacity factor and 100% wind

exploitation. On the contrary, the FIT_GO mechanism achieves a similar return period (20 yr) with

126 $/MWh contract price which is 28 and 36% higher than the FIT and the cost-benefit model,

respectively. The increased rates are due to the absence of peak premium in the FIT_GO that

results in a similar payback period as the aforementioned models.

Table 4.2: Comparison among the pricing models

Wholesale market FIT contract FIT_GO Cost-benefit

model (gas) Cost-benefit model (oil)

Price ($/MWh) at CF= 25% and

kw = 60%

-138‒1891 + ancillary payments

98 + 35% peak premium

126

Too little BEFIT to justify system

cost

92 + 35% peak premium

Return period (yr) 43.3 16.9 20 19.8

Figure 4.11 presents a Tornado diagram that compares the relative importance of model variables

(such as capacity factor, contract price, interest rate, and the like) with respect to their impact on

payback period. Here, each variable is varied between a probable range: contract price between

74‒122 $/MWh, interest rate from 3‒8%, wind electricity price between 51.2‒96 $/MWh, and

wind contributions from 10‒60%. The return period at the base case, i.e., the middle of these ranges

is 16.9 yr and the impact of changing variable conditions are represented by the bars. The analysis

suggests that low capacity factor has the highest sensitivity ranking followed by reduced contract

price and are both crucial for a reasonable return. The last two factors in the Tornado diagram are

wind power contribution and its price as they have little influence on PHS investment.

66

Figure 4.11: Tornado diagram showing impact of listed factors (on the left) on return period

Figure 4.12 illustrates the return period’s sensitivity to service life, avoided negative externalities,

fuel (petroleum) and carbon price, and net tax coefficient under the socioeconomic cost-benefit

model. The parameters in this case are varied between 10‒100 yr for service life, 20‒50 $/MWh

for avoided negative externalities, 0.25‒0.5 $/Kg for fuel price, 30‒110 $/tCO2 for carbon price,

and 0‒35% for tax coefficient. The return period which at the base case was 19.8 yr shows the

maximum deviations with possible changes in petroleum price, closely followed by price of carbon

and avoided externalities.

Figure 4.12: Impact of socioeconomic factors on return period

0 10 20 30 40 50 60 70 80

Wind price ($/MWh)

Wind contribution (%)

Interest rate (%)

Contract price ($/MWh)

Capacity factor (%)

Return period (yr)

0 5 10 15 20 25 30 35 40 45 50

Net tax coefficient (%)

Service life (yr)

Carbon price ($/tCO2)

Avoided externalities ($/MWh)

Fuel price ($/kg)

Return period (yr)

0.25 0.5

20 50

10

110

100

35 0

30

67

4.5 Limitation The study has several limitations. First, it excludes the much desired ancillary benefits of PHS

(quick-start and spinning reserve) from the socioeconomic cost-benefit calculation. One of the

crucial services provided by pumped storage includes black-start which restores a power station

or a part of an electric grid without relying on external transmission network. While the cost

associated with avoiding a potential blackout in Ontario would be substantial, accounting for its

value within the socioeconomic model would at best be dubious. Also, the study does not include

the cost associated with fugitive methane, often associated with natural gas extraction. Second, in

the absence of Ontario-specific data on power generation externalities, the research uses relevant

information on Europe that are not tailored to Ontario. There is considerable uncertainty associated

with such prognosis. Third, the payback under the dispatch optimization model is estimated with

due consideration for ancillary benefits. In reality, the optimal dispatch decisions may reduce the

potential for full-capacity ancillary market operations. Lastly, the benefits of storage are evaluated

in terms of replacing fossil fuel-based plants, avoided carbon emission and negative externalities.

However, the reported FIT rates in the analysis are likely to change depending on the source of

replaced power and in the case of nuclear, might be even lower.

4.6 Conclusion The projected growth in world energy consumption coupled with increasing demand for low-

carbon renewable sources has brought increasing awareness of the need for efficient energy storage

systems. However, at present market fragmentation combined with unfavourable regulation do

little to promote storage development and integration. In particular, there are few incentives and

little financial support for PHS operation in the Ontario electricity market that is yet to experience

complete deregulation and transparency. Strong financial support, stable economic and political

environment are crucial for further development of PHS systems. The chapter provisionally

explores various supporting mechanisms under which wind-based PHS projects could be

developed, integrated and supported by renewable energy sources in Ontario. When analyzed for

marginal cost-based operation, a 50% increase in return period is realized compared to guaranteed

price schemes. Considering that the Ontario wholesale market price is too low to justify a

68

reasonable return on PHS investment ‒ a reality confirmed by this study ‒ various fixed contract

schemes are explored that are believed to be effective in driving growth in storage capacities.

One of the major findings confirms that profitability is highly sensitive to capacity factors and

remuneration rates (per MWh) offered under the contract. A tiered remuneration is thus

recommended based on annual PHS contribution to the grid. Also note that, the estimated return

period (16.9 yr) with the modified FIT rates is no longer plausible under the current market

framework as the capacity building is restricted to a 500 kW limit (Ontario Ministry of Energy

2013). This study modifies FIT_GO, originally designed for isolated grids with traceable

electricity sources, for a conventional grid where such separations may not be realistic. The

contract price under FIT_GO varies between 97‒190 $/MWh under varying PHS contribution

(capacity factors), thus conforming to the need for a tiered FIT approach. A capacity based pricing

that is subjected to periodic revision would perhaps split the risk between consumers and investors,

thus creating a balanced market for PHS investment.

The benefits of storage are evaluated in terms of replacing the peak power plants, avoided emission

and externalities. The estimated benefits (in monetary values) are then assigned to PHS through

socioeconomic cost-benefit model to determine the value of its service. The operating costs of the

replaced peak power stations, mainly oil and natural gas in Ontario, are found to be significant

closely followed by avoided emission. Although increased wind energy exploitation (for charging)

implies relatively higher electricity production cost, increase in the BEFITs (i.e., benefit minus

cost) is far more significant leading to higher marginal profit for PHS. The high cost and severe

environmental damage in the case of oil-fired plants result in a considerably high BEFIT, thus

justifying the cost of PHS configuration. In contrast, the low operating cost and reduced emission

potential of CC plants lead to a lower BEFIT, thus making PHS to be cost-effective only at high

wind exploitations. However, such conclusions may be revised with consideration of fugitive

methane release, often associated with natural gas extraction. The pricing is also highly sensitive

to fuel price and carbon tax, expected to reflect the true cost of carbon. The study extends the

understanding of pumped storage economics using various supporting mechanisms in order to

enrich the electricity planning debate with a quantified data point.

69

Section 2 Increased Hydropower Potential at Niagara: A scenario-based

analysis While the previous section is devoted to the analysis of pumped storage economies, section 2

focuses on the second alternative (proposed in Chapter 2), i.e., incremental generation from the

existing power systems at Niagara. It details the development of a HEC-ResSim representation of

the existing power system and discusses the application of this model to explore a variety of

possible future scenarios under multiple chapters. These possibilities include considering the

sensitivity of the system to climate change, reducing tourist flows, and exploring the possibility of

revised daily management using additional storage to augment operational flows.

A few factors motivate the choice of location: First, the Canadian hydropower assets at Niagara,

known as Sir Adam Beck Complex, currently run below capacity suggesting the potential for

increased utilization. Second, the presence of interties between Ontario and New York at Niagara

Falls allows the export of excess domestic capacity should they not be required or contracted to

meet Ontario’s requirements. There is also potential to increase this capacity at a relatively low

cost with the completion of 230 kV lines between Allanburg and Middleport. In contrast, with the

best available sites for hydropower in Ontario already been exploited, further development has to

locate far from the load centers. The losses along long transmission lines serve little economic gain

for these projects to be justified in the long run.

70

A Simulation Model on The Impact of The 1950 Treaty on the Generation Potential at Niagara

Stretched along the border between Canada and the US and regulated by a 1950 Treaty, the Niagara

River currently provides almost 5,000 MW of renewable power. Chapter 5 details the development

of the Niagara Power System Simulation (NPSS) Model which ensures adherence to the current

regulatory regimes while providing users the freedom to refine these values. The model is then

used for scenario-based explorations that increase power diversions by somewhat relaxing the

treaty flow restrictions. Such an arrangement is shown to have potential to increase monthly hydro

discharges by 16% relative to the current baseline, and thus to permit an additional 1,050 GWh

annual generation capacity on the Canadian side alone. This exploratory study makes no pretense

of dictating future policy developments, but rather simply considers what possibly might be at

stake through a creative reassessment of historical constraints.

This chapter is based on the paper entitled “Increased Hydropower Potential at Niagara: A

Scenario-based Analysis” by Samiha Tahseen and Bryan Karney, submitted to the Journal of

Water Resources Management.

5.1 Introduction Development and operation of reservoirs often require complex water management since reservoir

operations must simultaneously meet varied objectives including flood control, power generation,

recreational uses, downstream environmental quality and safety, not to mention structural

integrity. These needs often conflict, a reality that often highly constrains operation.

The Niagara River is a case in point – it is not only an international boundary but hosts a world-

renowned waterfall as well as major hydropower developments. Apart from their significance to

tourism, the head difference along the river provides for hydropower installations on both sides of

the Canadian-United States border. The Niagara River produces almost 5,000 MW of renewable

power shared by both jurisdictions. Moreover, it is a strategically important international waterway

contributing to local growth and tourism. As a multipurpose resource, balancing the competing

demands among recreational, commercial, and industrial uses is an on-going challenge.

71

Regulations limiting water diversion for industrial and/or power generation purposes were first

introduced in 1909. At present, the 1950 Niagara River Water Diversion Treaty between Canada

and the United States defines the scenic minimum flow over the Niagara Falls and, after allocations

for navigation, domestic and sanitary purposes, the balance of Lake Erie outflow is equally divided

between Canada and the US (Government of Canada 2015).

Various published approaches address complex water allocation problems (Lowry et al. 2007;

Huaizhi et al. 2008; Zhao et al. 2014; Castelletti et al. 2014; Zoltay et al. 2007; Bosona and

Gebresenbet 2010; Latorre 2014). While optimization techniques have been widely used (Faber

and Harou 2007; Kumar and Reddy 2007; Li et al. 2014; Taghian et al. 2014; Wu et al. 2016),

simulation models allow for more detailed representation of reservoir systems, and thus a more

detailed prediction and exploration of both existing and possible system behaviour (Fisher 1995;

McMahon et al. 2009; Seifollahi-aghmiuni et al. 2016; Zgrzywa et al. 2008). Simulation is of

course not a new approach for exploring the behaviour of important and sensitive systems.

Through simulation, Thomas and Fiering (1962) and Hufschmidt and Fiering (1966) contributed

insights into multi-reservoir systems. Bekele and Knapp (2012) developed a model for examining

the potential of increased water supply and navigation usage on lake levels during drought

conditions. Eichert and Davis (1976), Hickey et al. (2003) and Stefanovic and Kreymborg (2004)

simulated flood control options whereas Teasley et al. (2004) used HEC-ResSim to investigate the

river restoration potential. Fagot et al. (2012) and Lara et al. (2014) applied the HEC-ResSim to

evaluate water management plans, while Osroosh (2012) considered scenarios for the Dez

reservoir. The current paper develops a simulation model for the lower Great Lakes, extending

from Lake Erie to Lake Ontario and uses this model to explore the potential of increased power

diversion and several climate-related changes to operation.

5.2 Study Area The Great Lakes comprise a series of interconnected freshwater lakes located on the Canada–

United States border with lakes Superior, Michigan, Huron, Erie, and Ontario collectively

containing 21% of the world's surface fresh water (US EPA 2015). The system contains three

major artificial diversions: (1) the Long Lac and Ogoki diversions (141.6 m3/s) into Lake Superior

from the Albany River system, (2) the Chicago diversion (90 m3/s) from Lake Michigan into the

72

Mississippi River basin, and (3) the New York State Barge Canal (22 m3/s) diverting to the Hudson

River basin. The lower lakes (Erie and Ontario) in the Great Lake basin are connected by the

Niagara River and the Welland Canal. The 58 km long Niagara River carries an average discharge

of 5,660 m3/s (Kirkham 2010) and an additional small portion of Lake Erie outflow is diverted

through the shipping canal. Figure 5.1 summarizes the study area and its key features to the current

work.

Figure 5.1: The Niagara River connecting Lake Erie and Lake Ontario

The relatively steady outflow and the elevation drop between these lakes have long supported a

valuable hydropower asset. The hydroelectric infrastructure at the Canadian side is known as the

Sir Adam Beck (SAB) Complex which hosts the only pumped storage station (SAB Pumped

Generation Station) in Ontario, along with two run-of-the-river plants (SAB I and SAB II).

Interestingly, the arrangement is such that the water discharged from the SAB PGS increases the

head pond elevation at the SAB I and SAB II (Tahseen and Karney 2016: Chapter 3). The river

N

73

flow must first satisfy the requirements of the 1950 Treaty with the Canadian portion of the residual

directed towards the SAB stations using three large tunnels and a power canal. The diversion takes

place at the Grass Island Pool (GIP), upstream of the International Niagara Control Structure

(INCS). The GIP is an in-river reservoir created by the INCS, and is shared between Canada and

the US. Despite the separation, generation at the DeCew Falls hydroelectric station (DeCew Falls

1 and 2) influences the available flow at the SAB Complex. The DeCew stations draw from the

Welland Canal flow which is naturally included in Canada’s share of the available water

(Government of Canada 2015). The US hydropower infrastructure is known as the Sir Robert

Moses Plant. Table 5.1 summarizes the hydropower plants along with their installed capacity. The

water levels along the upper Niagara River are measured in real time by the New York Power

Authority (NYPA) and the Ontario Power Generation (OPG) at key locations along the river

(Crissman et al. 1993).

Table 5.1: Existing hydropower infrastructure at Niagara

Ownership Plant name Installed capacity (MW)

Canada

Sir Adam Beck I 488

Sir Adam Beck II 1,694

Sir Adam Beck PGS 174

DeCew Falls 1 23

DeCew Falls 2 144

US Robert Moses 2,275

Lewiston 240

5.3 Model Development The model for the lower Great Lakes (from Lake Erie to Ontario) is developed using the HEC-

ResSim, a simulation software by US Army Corps of Engineers (USACE - HEC 2015a). The

HEC-ResSim uses a rule-based approach to mimic decision-making processes to meet flood

control, power generation, water supply, and environmental quality requirements (Klipsch and

Evans 2006) and generally uses hydrologic routing (Sherman 1932; Cunge 1969; Dooge et al.

1982; Wilson 1990) along with storage-outflow relationships. Notable applications of the HEC-

ResSim program includes reservoir regulation on the Columbia River system (Modini 2010), a

74

reservoir operation plan for the West Point Dam (Fagot et al. 2011), development of alternative

hydrologic index for the Russian River watershed (USACE 2012) and many others.

5.3.1 Layout of key hydraulic components A georeferenced map (ESRI 2015) with required shape files was initially imported as background

to the system. The lakes, the SAB PGS and the GIP are modelled as reservoirs. Each of these

reservoirs comprises a pool and a dam where the pool’s hydraulic behaviour is defined by

elevation-storage-area relationships and with various outlets (gate, spillway, pumping, etc.). A

separate reservoir at the forebay of the SAB I and the SAB II (known as crossover) allows the

water from the SAB PGS to also be utilized at SAB I and SAB II.

River reaches connect the reservoirs and complete the flow network. While Muskingum-Cunge

routing method is often the preferred choice, the simpler Muskingum option is reasonably chosen

when data is scarce. Basic reach data, including length, cross-section and slope, are extracted from

the river bathymetry data using the ArcGIS 10.2 and the HEC-GeoRAS (NOAA 2015a; USACE

- HEC 2015b). Manning’s n values specific to the Niagara River are obtained from Lal (1995)

while standard values are used for the tunnels and the canal (Chow 1959). Junctions are chosen to

coincide with gauge stations and stream confluences. Local flows are introduced at three locations

along the river: at the headwater junction (inflow to Lake Erie), 45 m3/s inflow prior to the GIP,

and 35 m3/s allocated to the Welland Canal (Harvey 2004). Hourly water level data for National

Oceanic and Atmospheric Administration (NOAA) station no. 9044036 are used to compute the

flow to Lake Erie according to a suitable rating curve.

The Buffalo (Lake Erie) and Olcott (Lake Ontario) stations are chosen to represent the respective

pool elevations. The water level data at Buffalo, Olcott, Ashland Ave. and NY Intake are extracted

using NOAA’s web-based platform, and the measured flow at Queenston are obtained from the

Environment Canada’s website. These data support model validation through comparisons with

simulated outcomes. The rating curves for Buffalo and Ashland Ave. (Harvey 2004; LimnoTech

2010) are used for stage-discharge conversions. While the data at Ashland Ave. represent the flow

over the falls, the Queenston gauging station measures the combined flow over the falls plus any

outflows from the hydropower projects on both sides of the river. Groundwater exchanges, whether

75

inflows or outflows, are represented by lateral flows. The combined effect of evaporation, rainfall

and runoff on the lake is captured by the Net Basin Supply (NBS), a unique parameter typically

measured by the change in storage level on a monthly basis. The NBS data for Lake Erie and

Ontario are extracted from the NOAA Great Lakes Environmental Research Laboratory (NOAA

GLERL 2015). Diverted outlets in the model represent the power canal and the tunnels at Niagara.

Unlike the New York State Barge Canal which removes water from the river system, flow in the

canals and the tunnels is returned downstream.

5.3.2 Operational characteristics Simulation is initiated by dividing each reservoir pool or storage element into zones. Zone

boundaries are set based on historical lake water level where average values corresponding to the

time of the year are set as guide curve or target elevation. Figure 5.2 demonstrates the hourly

minimum, maximum and average lake level for each month, estimated using the data from 1990

to 2006, representing the dead water, the flood control and the conservation zone respectively.

Then, prioritized operating rules are set under each zone to constrain the reservoir to meet the

specified guide curve elevation.

Figure 5.2: Zoning of Lake Erie based on historical lake level data

For example, a rule for Lake Erie ensures a minimum navigational flow of 80 m3/s to the Welland

Canal. Other operating rules for the lake include minimum and maximum limits on releases and

hydropower schedules. The GIP reservoir is subject to operating rules that best represent the two

172

173

174

175

176

177

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Wat

er le

vel (

m)

Dead zone Conservation zone Flood control zone

76

regulatory regimes on the Niagara flow diversion – the 1950 Treaty and the 1993 INBC Directive.

The treaty, which specifies a minimum flow over the falls, is modelled as a mandatory time-

dependent release. The incorporation of the INBC directive is more complex: it maintains a long-

term mean water level of 171.16 m at the GIP, while allowing ±0.46 m daily and ±0.91 m monthly

deviations. This constraint is represented with a specific rule that limits the rate at which this level

is changed. The change in water level is also measured using two state variables that aggregate

deviations (from 171.16 m) over 24 hour and 720 hour period, respectively. The variables are then

used in a conditional rule block (if-then-else block) to initiate specific releases. The GIP operation

also includes a Jython (a Java-based implementation of the Python programming language)

scripting to ensure a uniform flow distribution between Canada and the US. The scripted rule

computes the available flow for power production by subtracting the treaty flow requirements from

the GIP inflow. The model attributes the first 141.6 m3/s of the available flow to Canada to

compensate for the amount diverted into Lake Superior through Long Lac and Ogoki. The

remainder of the available flow minus the amount used for power generation at the DeCew plants

is divided evenly between the two countries (to be complete, the Robert Moses is attributed an

additional flow equivalent to the discharge at the DeCew stations). Nevertheless, these allocations

are further limited by maximum diversion capacity of the tunnels and the canal, i.e., 2,300 m3/s

and 2,800 m3/s for Canada and the US respectively.

The operations at SAB PGS, critical for maintaining sufficient water level at the downstream

locations (Maricic et al. 2009), is achieved through an if-else conditional block that ceases

pumping operation (inflow into the PGS reservoir) when the downstream elevation falls below a

pre-set limit. A parallel reservoir system is defined among the PGS, the crossover and the GIP to

influence releases from each of these reservoirs in order to operate towards a user-defined storage

balance. This approach strives to achieve a minimum storage in the flood control zone at the

crossover and the GIP. In the conservation zone, the system keeps both these reservoirs at

maximum storage with a compromise in water level at the PGS. Another conditional rule invokes

the alternative nature of pumping and generation at the pumped storage. Appendix B provides a

list of the operating rules under each reservoir. Initial pool elevation and releases at the beginning

of the simulation period are provided in the model as lookback data. In addition, power plant

generation data from Independent Electricity System Operator (IESO) inform the hydropower

generation schedule.

77

5.4 Model Calibration and Validation The model uses an hourly timestep and is calibrated using recorded water level data from 2007–

2008, and validated with reported data from the subsequent three years (2009–2011). The model

results are compared against the gauge data from the NOAA (2015b) and the Environment Canada

(2015). Figure 5.3 (a, b) demonstrates the 2010 simulated water level to the observed values at

Ashland Ave. Similar comparisons are performed at Lake Erie, Lake Ontario and the GIP (Table

5.2). The percentage error and standard deviation of differences between the simulated and the

actual water level are reported in monthly mean values based on hourly data. The statistical Root

Mean Squared Error (RMSE) is calculated for the validation period for further comparisons. While

RMSE values for water level at Ashland Avenue show the largest deviations (1.54 m), such values

at other stations show variations in the range between 0.14–0.76 m. Comparisons between

simulated and measured flow at Queenston and Buffalo gauges show a maximum error of 8.7%

and 10% respectively. However, it should be emphasized that neither the calibration or validation

data use actual power diversions, but only feasible diversions based on the summarized

stipulations, while the water levels are based on actually recorded values, values reflect actual

diversions and not theoretically possible power extractions.

(a) Hourly variation in water level (June‒July, 2010)

95.5

96.5

97.5

98.5

99.5

100.5

01-J

un

06-J

un

11-J

un

16-J

un

21-J

un

26-J

un

01-J

ul

06-J

ul

11-J

ul

16-J

ul

21-J

ul

26-J

ul

31-J

ul

Wat

er le

vel (

m)

Observed Simulated

78

(a) Variation in water level on a monthly basis

Figure 5.3: Comparison between simulated and observed water surface elevation at Ashland

Ave. in 2010

Table 5.2: Percentage error and standard deviation between simulated and overserved elevation

(2009)

Months

Lake Erie Lake Ontario GIP

% error Std Dev (m) % error Std Dev (m) % error Std Dev (m)

Jan 0.09 0.17 0.20 0.02 0.01 0.07

Feb 0.20 0.15 0.15 0.03 0.00 0.08

Mar 0.25 0.14 0.24 0.04 0.00 0.07

Apr 0.31 0.15 0.28 0.02 0.04 0.10

May 0.35 0.12 0.28 0.02 0.05 0.10

Jun 0.43 0.06 0.22 0.02 0.06 0.10

Jul 0.48 0.09 0.15 0.03 0.10 0.09

Aug 0.53 0.09 0.12 0.03 0.09 0.11

Sep 0.57 0.18 0.01 0.05 0.07 0.11

Oct 0.55 0.22 0.18 0.04 0.04 0.12

Nov 0.53 0.13 0.29 0.03 0.01 0.04

Dec 0.50 0.33 0.31 0.03 0.00 0.09

95

96

97

98

99

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Wat

er le

vel (

m)

Model result Observed data

79

The difference between the model output and the observed data at the beginning of the simulation

period in Figure 5.3(b) is of little consequence and can be explained by the standard iterative

methods which require a minimum number of steps for the network to be stabilized and

equilibrated as the flow process is simulated. Also, the hydrologic model does not account for ice

cover, wind set-up associated with real wind speed and direction, flow retardation due to ice and

aquatic vegetation, storm activity or any additional flows for emergency response or clearing ice

jams, all of which might partly account for the discrepancies between the observed and simulated

values in Figure 5.3(a, b). A previous model (Clites and Lee 1998) also cited ice retardation

(winter) and storm activity (fall) as key sources of inaccuracies in estimating lake levels and flows.

The model’s under-prediction of daily lows in Figure 5.3(a) is also likely influenced by operational

limitations in delivering energy to the provincial grid that do not reflect actual diversions or for

lower-than-normal power demand leading to reduced diversions.

5.5 Considering Additional Diversion for Enhanced Hydropower The simulated model is evaluated for several hypothetical scenarios, including a reduced flow

restriction scenario; that is, the model can be used to explore the outcome of lowering the treaty

specified flow over the falls. At present, the discharge over the falls is regulated by the 1950

Niagara River Water Diversion Treaty which specifies a minimum of 2,832 m3/s (100,000 ft3/s)

over the falls from April 1 to September 15 between 8:00 AM and 10:00 PM. The same flow

restrictions are in effect between 8:00 AM and 8:00 PM from September 16 to October 31. At all

other times, a minimum of 1,416 m3/s (50,000 ft3/s) must generally be allocated (Government of

Canada 2015).

The treaty stipulates an unbroken crestline as the most striking visual feature for achieving the

“impression of volume” deemed necessary for the scenic spectacle of Niagara. Nevertheless, the

upper limit of 2,832 m3/s is not an absolute value, since the crestline remains unbroken with a flow

half this value (Friesen and Day 1977). Previously researchers have raised questions regarding the

rationale behind the treaty, and stated the need for a more thorough review of the existing

relationship between river flow and the scenic beauty of the falls. Furthermore, the continuous

retreat of the Niagara escarpment at the crest, currently at a rate of about 0.3 m per year, may

benefit from a reduced flow over the falls. Considering that the additional 1,416 m3/s during the

80

tourist season represents an annual cost of $103 million CAD in terms of compromised

hydropower potential (Sedoff et al. 2014), the current authors investigate the prospect of greater

diversion for power than currently allowed by the existing treaty. Such considerations, though

controversial, can help address the emerging challenges that might be readily addressed by a treaty

revision. For example, the expected eventual decommissioning of the nuclear facilities and the

reality of the growing penetration of intermittent renewables in Ontario both elevate the

desirability of using hydro to mitigate demand variability from intermitted sources. Moreover,

generation from the additional 1,416 m3/s (50,000 ft3/s) during the tourist season would come from

a GHG free resource, potentially reducing atmospheric loads (Sedoff et al. 2014). Third, the

completion of the third Niagara Tunnel now offers for the first time the technical possibility of a

greater flow to the SAB Complex (Sedoff et al. 2014).

Assessing the impact on power from a possible treaty flow relaxation involves introducing a

second operation set into the model; each time increasing the power diversion by 200, 400, 600

and 830 m3/s during the day time of the tourist season when the treaty doubles the flow

requirements from 1,416 m3/s to 2,832 m3/s. The limit on diversion is kept consistent with the

findings of Friesen and Day (1977) which suggest that a minimum of 845 m3/s additional diversion

might be permissible as long as confirmatory studies, and possibly some remedial work, are

undertaken. The additional diversion scenarios reduce the tourist flow requirements (over the falls)

from 2,832 m3/s to a minimum of 2,002 m3/s between 8:00 AM and 10:00 PM from 1 April to 15

September and 8:00 AM and 8:00 PM from 16 September to the end of October (based on the

original tourist flow hours).

Because other constraints are also active in the Niagara Complex, the results show that not all of

a flow diversion at the falls immediately translates to increased power production. In fact, the

calibrated model shows that only about 150 of the entire 200 m3/s reduced falls-flow (increased

power diversion) is realized at the downstream hydropower stations. With 400 m3/s reduction in

scenic flow, these stations experience about a 350 m3/s increased discharge. This suggests that

about 15–25% of the benefits offered by additional power diversion are lost owing to other

operational constraints. Figure 5.4 illustrates the effect of these diversions on available discharge

at the Robert Moses and the SAB stations. The model suggests a maximum of 16% increase in

monthly available flow from April to August, while the same for September and October is 15%

81

and 13.5% respectively. The slight reductions in available power flow for September and October

is attributed to the reduced tourist flow hours (8:00 AM to 8:00 PM) when compared to its

preceding months (8:00 AM to 10:00 PM). The Sir Robert Moses station experiences a 7–29%

increase in monthly accessible power flow under varying diversion scenarios (from 200–830 m3/s),

while the SAB Complex experiences a 5–20% escalation. The uneven percentage increase in

station discharge is attributed to the 500 m3/s of excess tunnel capacity on the US side.

Figure 5.4: Monthly variation in available power flow at the Niagara Complex under the baseline

and increased diversion scenarios

For this part of the analysis, the authors focus on generation at the SAB Complex, while assuming

similar outcomes, if not less, for its corresponding station on the US side of the border. The SAB

station may benefit from a 1,100 MWh increased daily generation with a mere 7% reduction in

scenic flow over the falls. The 830 m3/s increase in power diversion, as suggested by Friesen and

Day (1977), holds the potential for a 5,000 MWh increased daily generation, which translates to

an additional 163 GWh for a typical summer month. Figure 5.5 shows a 5–20% increase in monthly

power generation under various diversion scenarios. The relative peak in the increase rates for July

and August is mostly due to the improved flow conditions. Interestingly, this period happens to

coincide with the peak demand period in Ontario. The additional generations from April to October

add up to 230 GWh with the lowest diversion of 200 m3/s, and 1,050 GWh under the maximum

power diversion scenario. Similar or more such enhanced generation can be expected at Robert

Moses plant under the scenarios.

1500

1800

2100

2400

2700

3000

Apr May Jun Jul Aug Sep Oct

Pow

er fl

ow (m

3 /s)

Baseline 200 m3/s 400 m3/s 600 m3/s 830 m3/s

82

Figure 5.5: Increase in monthly power generation at the SAB Complex under the baseline and

increased diversion scenarios

5.6 Critical Appraisal The study must be understood to embody several crucial limitations. First, due to its politically

sensitive nature, access to the specific hydrologic and power systems data on Niagara is restricted.

The information used for model assembly is obtained from various publicly available documents,

web resources etc. Unfortunately, specific releases from each hydropower plant is not part of the

public record. The model provisionally diverts all available flow to generation ‒ an obvious

distortion since such decisions are often guided by demand and available generation for a particular

duration. It is in fact common practice for hydroelectric stations either to store or to release surplus

flow. There is clearly potential to improve the simulation of baseline conditions, and consequently

to better predict responses under various scenarios. Second, the model uses hydrologic routing

(Muskingum-Cunge) rather than hydrodynamic approaches. This choice of river routing is

considered a trade-off between a number of criteria including the scale of river catchment to be

modeled, available data and required accuracy. Though, Muskingum-Cung provides reasonable

results for moderate flows propagating through mild to steep sloping watercourses (Maidment and

Fread 1993), the method sometimes produces unrealistic initial negative dips in the computed

hydrograph (Baláž et al. 2010). Considering the intensive data requirement for a large-scale basin

such as Niagara (Arora et al. 2001), hydrologic approach is preferred here over a more accurate

hydrodynamic model. Thirdly, the model obtains a fixed value for local flow contributions at the

0

5

10

15

20

Apr May Jun Jul Aug Sep Oct

% In

crea

se in

gen

erat

ion

200 m³/s 400 m³/s 600 m³/s 830 m³/s

83

GIP and the Welland Canal from previous studies on Niagara. Such values are clearly sensitive to

changing evaporation and precipitation patterns. However, the river flows are quite insensitive to

these values, and negligible error is expected from this source.

5.7 Conclusion Water resource management is a complex issue that requires comprehensive investigation in

social, economic, ecological, technical and policy fields. The work here first develops a simulation

model for the existing power systems at Niagara with river operation strictly adhering to the 1950

treaty. However, the complexity of legal and political discussion and the possibility of negative

public reaction associated with any alteration of an age-old treaty may have led its potentials to be

mostly unexplored. The primary contribution of this work is to evaluate the potential for increased

hydropower generation with a possible renegotiation of the treaty and to assess the impact of the

system constraints in reducing the potential benefits. When evaluated for relaxed flow restrictions,

the available monthly discharge increases by 16%; however, interestingly, the model at times is

constrained by the non-treaty issues in the system in the form of plant and tunnel diversion

capacity, losses etc.

Apart from power, increased the diversion of water from the Niagara River may reduce the falls

recession (erosion) rate and decrease misting (Case 2004), while having some impact on the

shoreline, on aquatic ecosystems, and obviously on the visual experience of the falls itself. This

exploratory study recommends further research that would account for such changes more

holistically (economic-environmental-social), and necessarily coupled with new discussions

between Canada and the US. The countries could initiate a joint exploration of the treaty rational

in today’s light. The International joint commission (IJC), a cooperation between Canada and the

US for protecting the transboundary water, could facilitate and perhaps guide these discussions.

This research neither promotes nor forecasts likely developments, but rather attempts to enable,

and perhaps enrich, these discussions through quantitative evaluation of possible hydropower

benefits through a creative reassessment of historical constraints.

84

Power Systems Vulnerability to Climate Change: An Analysis on the Niagara Hydropower System

Considering that the world’s water resources face increasing challenges within the context of a

changing climate, securing supply and equitable allocation of water pose novel challenges. While

the published approaches assess the impacts on runoff, water quality, agriculture, management

plans, and so on, limited attention is vested on power systems’ vulnerability to a changing runoff.

Chapter 6 extends the work of the preceding chapter to simulate power generation under varying

climate scenarios, developed based on the aggregated climate projections for the Great Lakes. The

outcome suggests that, apart from a sustained low water level, the changing precipitation and

evaporation pattern result in a 4‒36% reduction in annualized available discharge at the

downstream hydropower plants under the IPCC SRES A1FI and A21 emission scenario. The

analysis is further extended to evaluate an ambitious lake regulation plan with additional storage

during nighttime and timed release during peak demand hours. This scheme provides opportunity

for flow shifting, which augments hourly generation by 0.1‒30% during peak demand hours (while

reducing during off-peak hours). When aggregated on an annual basis, the regulation results in a

390‒570 GWh added generation for Canada.

The chapter is based on the completed manuscript entitled “Power Systems Vulnerability to

Climate Change: An Analysis on the Niagara Hydropower System” by Samiha Tahseen and Bryan

Karney, prepared for an intended follow-up submission to the Journal of Water Resources

Management. Since the current study builds directly on the NPSS model elaborated in chapter 5,

the submission is currently awaiting publication of the preceding work.

6.1 Introduction While details of any future climate are uncertain, the vulnerability of natural systems to changing

climate patterns is regarded as one of the major challenges in the coming years (Bates et al. 2008;

Bou-Zeid and El-Fadel 2002). One of the chief areas to be affected is the water resource systems;

securing supply and equitable allocation of water under the changing climate and hydrologic

conditions pose novel challenges (Arnell et al. 2011; Olmstead et al. 2016; Poff et al. 2015;

Schindler 2001). Rising temperatures lead to higher evaporation (USGCRP 2009, Melkonyan

85

2015), in effect either heightening water demand or shrinking available supplies (Doll 2002).

Further complications arise from the variability of precipitation; a possible increase in intensity

and a greater fraction of this precipitation occurring as rain can result in long dry spells

(Christensen and Lettenmaier 2007; Shamir et al. 2015) followed by a sudden deluge (Islam and

Gan 2014; Peterson et al. 1997). Since the traditional power generating facilities depend on water

(either as an energy source for hydropower plants or as coolant for thermal plants), climate change

and its resulting impacts on water resources will affect generation, while energy demands continue

to increase with economic development and a growing world population (Davies and Simonovic

2011; Vliet et al. 2016).

Being dubbed as ‘inland seas’, the Great Lakes and their connecting waterways comprise the

world’s largest surface freshwater system (Kling et al. 2003; US EPA 2015). The lakes encompass

a surface area of 245,000 km2 having over 14,000 km of shoreline, and a drainage area of 746,000

km2 (NOAA GLERL 2016; US Fish and Wildlife Service 2015). Within the Great Lakes system,

water flows from Lake Superior via the St. Marys River into Lake Michigan-Huron. The relatively

deeper and cooler upper lakes are connected with Lake Erie through the St. Clair River, Lake St.

Clair, and the Detroit River. Lake Erie flows over the Niagara Falls and into Lake Ontario before

flowing through the St. Lawrence River into the Atlantic Ocean. While the outflows from Lake

Superior and Lake Ontario are regulated, the middle lakes (i.e., Michigan, Huron and Erie) rely

exclusively on the connecting rivers. Being the most upstream, Lake Superior regulation somewhat

influences the entire system, while that of Lake Ontario has essentially no influence on the upper

lakes owing to the elevation difference at the Niagara Falls (Indiana Department of Natural

Resources 2015). While the regulations reduce natural variability in Lake Superior and Lake

Ontario, the unregulated lakes experienced extremely high water levels in 1929, 1952, 1973, 1986,

and 1997, as well as low levels in 1926, 1934, 1964, and 2003 (Wilcox et al. 2007). The abrupt

and sustained water level drops in the late 1990s, believed to be related to the El Niño event

(NOAA 2014), further suggest the inability of the current regulation to alleviate lake level

extremes. A possible solution for reducing climate-induced variability may involve increased

regulation, an option that is briefly discussed later in this chapter.

Recently published climate studies predict a warmer temperature and changing precipitation

pattern for the Great Lakes including a higher risk of more intense drought and flooding (Kahl and

86

Stirratt 2012). These sustained changes pose economic threats to numerous industries that rely on

lakes water supply. This chapter primarily focuses on the possible impacts on hydroelectricity

while considering potential solutions to reduce climate induced variability. Being dispatchable,

historically dominant and renewable, hydropower comprises a significant part of the combined

generation assets, thus playing a critical role in the electricity markets of New York and Ontario

(the chief jurisdictions relying on Lake Erie outflow). Nevertheless, the generation capacity at

these plants is threatened by the variability in natural flow conditions. While high water levels may

cause local flooding, thus risking the physical integrity of the associated infrastructure, the

shortage of water supply is the greatest risk to hydroelectric generation interests over the long term

(International Joint Commission 2012). The cost of replacing the lost hydroelectric generation at

these plants has been estimated to reach $2.83 billion CAD through 2050 (Shlozberg et al. 2014).

The literature on potential impacts of climate change on water resources tend to fall into three

broad categories. First, the majority of studies investigate the impact on streamflow or runoff

(Akhtar et al. 2008; Arnell 2004; Barnett and Pierce 2009; Chen et al. 2012; Christensen and

Lettenmaier 2007; Devkota and Gyawali 2015; Fujihara et al. 2008; Knapp et al. 2005; Koutroulis

et al. 2013; Kusangaya et al. 2014; Mimikou et al. 2004; Nkomozepi and Chung 2014) with a few

deliberating on potential adaptation strategies (Bou-Zeid and El-Fadel 2002; Wang and Zhang

2015). A second group explores approaches or techniques for incorporating climate change into

water management issues (Charlton and Arnell 2011; Cohen et al. 2006; Georgakakos et al. 2012;

Poff et al. 2015; Purkey et al. 2007). The third group studies the impacts of changing runoff on

agriculture, water delivery, quality, and the like (Barnett and Pierce 2009; Biemans et al. 2013;

Doll 2002; Islam and Gan 2014; Piao et al. 2010; Tsanis and Apostolaki 2009). All the studies

converge to a common conclusion that climate change would almost certainly have alarming

consequences for streamflow variability. While a number of studies have described the impacts on

runoff, water budget and quality, agriculture and management plan, limited attention is vested on

power systems’ vulnerability to a changing runoff. A recent study by Vliet et al. (2016) reports a

global reduction in usable capacity by 61‒74% for hydropower plants and 81‒86% for

thermoelectric plants for 2040‒2069. In this chapter, the authors use a systems model for the lower

Great Lakes, extending from Lake Erie to Lake Ontario (Tahseen and Karney 2017: Chapter 5) to

investigate the power systems vulnerability to the changing runoff conditions. A comprehensive

literature review is conducted to aggregate the existing climate projections for the Great Lakes, the

87

knowledge of which aids in developing several possible future climate scenarios. The chapter

further explores a unique operating plan that involves additional lake storage during nighttime and

timed release during peak demand hours.

6.2 Study Area The Great Lakes are a series of interconnected freshwater lakes located on the Canada-United

States border. The lower lakes in the Great Lake basin are connected by the Niagara River and the

Welland Canal. The river is about 58 km long, and carries an average 5,660 m3/s from Lake Erie

to Lake Ontario (Kirkham 2010). The relatively steady outflow and the natural drop in elevation

between these lakes have long been a valuable asset for hydropower development. The existing

hydroelectric infrastructure at the Canadian side is known as the Sir Adam Beck (SAB) Complex.

It hosts the only pumped storage station in Ontario (known as the SAB Pumped Generation

Station), along with two conventional run-of-the-river plants (SAB I and SAB II). The

hydroelectric asset on the US side is known as the Sir Robert Moses Plant. Figure 6.1 shows the

study area along with the location of these plants. The river flow, after meeting the requirements

of the 1950 Niagara River Water Diversion Treaty, are directed towards the power stations using

underground tunnels and a power canal. The diversion takes place at the Grass Island Pool (GIP),

and is assisted by the International Niagara Control Structure (INCS). The GIP is an in-river

reservoir created by the INCS, and is shared equally between Canada and the US. Table 6.1

provides a list of the hydropower plants along with their installed capacity.

Table 6.1: Existing hydropower infrastructure at Niagara

Ownership Plant name Installed capacity (MW)

Canada

Sir Adam Beck I 488

Sir Adam Beck II 1,694

Sir Adam Beck PGS 174

DeCew Falls 1 23

DeCew Falls 2 144

US Robert Moses 2,275

Lewiston 240 Source: http://www.niagarafrontier.com/riverdiversion.html

88

Figure 6.1: A systems for the Niagara River connecting Lake Erie and Lake Ontario

6.3 Model and Scenario Development

6.3.1 Niagara River simulation The paper employs a model developed for the lower Great Lakes (from Lake Erie to Lake Ontario)

using HEC-ResSim by Tahseen and Karney (2017) (Chapter 5). The data used for model

simulation are provided in Table 6.2.

The model is calibrated using the water level data from 2007‒2008, and validated with the

subsequent three years’ data (2009‒2011). The model output is compared against the gauge data

from NOAA (2015b) and the Environment Canada (2015). The percentage error and the standard

deviation of differences between the simulated and the actual water level are estimated at four

Canal 625 m3/s

619.5 m3/s

Tunnel I, II & III 1700 m3/s 1888 m3/s

Generate &

pump

SAB #2 La

ke O

ntar

io

Grass Island Pool

SAB PGS

Niagara River 2832 m3/s or 1416 m3/s

SAB #1

N

195 m3/s Welland

Canal

845 m3/s hr

58 m3/s DeCew #1

193 m3/s

DeCew #2

Falls

Lake

Erie

89

gauges that results in a maximum 1.95% and 1.57 m respectively. The statistical Root Mean

Squared Error (RMSE) is calculated for the validation period for further comparisons. While the

RMSE values show the largest deviations at 1.54 m, in most cases it varies between 0.14‒0.76 m.

Nonetheless, the hydrologic model does not account for ice cover, wind set-up associated with real

wind speed and direction, flow retardation due to ice and aquatic vegetation, storm activity or any

additional flows for emergency response or clearing ice jams, all of which might partly account

for the discrepancies between the observed and simulated values in Figure 6.2. A previous model

(Clites and Lee 1998) also cited ice retardation (winter) and storm activity (fall) as key sources of

inaccuracies in estimating lake levels and flows. The model’s underprediction of daily lows in

Figure 6.2 can also be influenced by operational limitations that barely allow the exact amount to

be diverted, and low power demand leading to reduced diversion. The model simulation is

described in detail in Tahseen and Karney (2017) (Chapter 5) for interested readers.

Table 6.2: Required data for model simulation (Tahseen and Karney 2017)

Data type Equation/parameter Source

Geo-referenced map ESRI 2015

River bathymetry cross-section and slope NOAA 2015a

Roughness coefficient Manning’s n Lal 1995

Hydrologic data local flows

hourly water level

Harvey 2004

NOAA 2015b

Tunnel, canal data flow capacity

Manning’s n

Harvey 2004

HATCH 2015

Chow 1959

Rating curve Q = 338.14 (Z – 549.87)2

Q = 260.5 (Z – 550.11)2.2

Q = 33.75 (Z – 91.42)2 + 728.74

Clites and Lee 1998

LimnoTech 2010

Net Basin Supply (NBS) evaporation, rainfall and runoff NOAA GLERL 2015

Power plant capacity Capacity, head, efficiency OPG 2015c

90

Figure 6.2: Hourly variation in water level at Ashland Avenue gauge (September, 2010)

6.3.2 Potential scenarios In this chapter, the simulated hydrologic model is evaluated for potential climate change and lake

regulation scenarios. The following section elaborates the scenario development.

6.3.2.1 Climate change projections for the Great Lakes The existing climate studies predict a changing precipitation and evapotranspiration pattern for the

Great Lakes-St. Lawrence river basin (Tsanis et al. 2011). Croley (1990) developed a hydrologic

model for the Great Lakes. The developed model used the results from equilibrium-response

General Circulation Models (GCM) to develop four climate change scenarios: Canadian Climate

Centre GCM (CCC GCM) (Boer et al. 1992; McFarlane et al. 1992), Goddard Institute for Space

Studies (GISS) (Hansen et al. 1983), Geophysical Fluid Dynamics Laboratory (GFDL) (Manabe

and Wetherald 1987), and Oregon State University (OSU) (Ghan et al. 1982). To further test the

sensitivity of the model, transposition climate scenarios were developed by changing the mean and

the variability of temperature and precipitation (Croley et al. 1995; Mortsch and Quinn 1996).

However, more recent assessments use transient GCM models (CGCM1, HadCM2, CGCM2,

HadCM3) which are dynamic ocean models coupled to an atmosphere with changing CO2

concentration (Croley 2003; Mortsch et al. 2005). These contemporary studies indicate a less

significant warming compared to the previous equilibrium models which did not include the

cooling effect from aerosols (Sousounis and Bisantz 2000). While Lofgren et al. (2002) and

Mortsch et al. (2000) apply the transient models with IPCC IS92a scenario, more recently IPCC

SRES emission scenarios have been used.

95.5

96.5

97.5

98.5

99.5

1-Se

p

3-Se

p

5-Se

p

7-Se

p

9-Se

p

11-S

ep

13-S

ep

15-S

ep

17-S

ep

19-S

ep

21-S

ep

23-S

ep

26-S

ep

28-S

ep

30-S

ep

Wat

er le

vel (

m)

Simulated Observed

91

6.3.2.1.1 Air temperature Trend. The annual mean temperatures have increased by 0.7–0.9˚C between 1895 and 1999 for the

southern part of the Great Lakes basin (Mortsch et al. 2000). The warming is greater in minimum

temperature than in maximum (Zhang et al. 2000) with the most significant increases occurring in

winter and spring (Bolsenga and Norton 1993; Magnuson et al. 1997). This gradual warming has

resulted in a 71% reduction in the Great Lakes ice cover between 1973 and 2010 (Wang et al.

2012).

Projection. Alarmingly, the trend is expected to continue with substantial increase in minimum

annual temperatures (Kling et al. 2003; Taylor et al. 2006). The CCC GCM simulated a 3.5˚C

increase in global mean temperature for a doubling of pre-industrial CO2 level (560 ppm)

(Magnuson et al. 1997). The outcome is consistent with other GCM simulations by Boer et al.

(1992) and McFarlane et al. (1992). With scenarios for the Laurentian Great Lakes, increases in

mean air temperature range from 2‒5˚C in summer and 4‒8˚C in winter (Magnuson et al. 1997).

The transient model based on IPCC SRES emission scenarios by Mortsch et al. (2005) indicate a

1.5–6.5°C increase in mean annual temperature for the Great Lakes-St. Lawrence basin by 2050

(at the time of 2°C warming above pre-industrial level). Gula and Peltier (2012) used the WRF

Regional Climate Model and reported a 2–3°C increase in air temperature for southern Great Lake

basin by 2050‒2060 under the SRES A1B and A2 emission scenario (both scenarios having similar

projections for 2050-2060). Projected changes for 2090‒2100 period are quite similar to the 2050‒

2060 under the A1B scenario. However, the projections for the A2 scenario show a 5–6°C increase.

Effect. The warmer climates are likely to decrease the spatial extent of ice cover on the Great

Lakes; especially the small lakes in the south will no longer freeze every year (Magnuson et al.

1997; Mortsch and Quinn 1996). The rising temperature also affects evapotranspiration which

plays an important role in determining water availability. All the studies recognize a rising trend

in lake evaporation due to increased lake surface temperatures, lack of ice cover, and wind speed

(Kahl and Stirratt 2012; Mortsch and Quinn 1996; Schindler 2001). Historical data (1970-1990)

on the lakes near Kenora in northwestern Ontario illustrate the relationship between temperature

and evaporation in boreal lakes and streams. During this period, evaporation increased by an

average of 35 mm/1°C increase in annual air temperature (Schindler et al. 1990; Schindler et al.

1996). Following similar trends, a 3°C temperature increase by 2050‒2060 under the A1B and A2

92

emission scenario leads to a 105 mm (11−15%) increase in lake evaporation. The recent estimates

(from transient models) by Croley (2003) and Mortsch et al. (2006) show a 12‒23% increase in

mean annual evaporation along with 5‒26% reductions in lake outflow for 2050.

6.3.2.1.2 Precipitation Trend. Analyses to determine precipitation trends in the Canadian Great Lakes-St. Lawrence basin

indicate that total precipitation has increased between 1895 and 1995 (Mortsch et al. 2000). While

Magnuson et al. (1997) found precipitation to be increasing at a rate of 2.1% per decade

(corresponding to 21‒24% increase per century), Zhang et al. (2000) reported a 5‒35% increase

over southern Canada between 1900 and 1998.

Projection. All projections expect this trend to continue throughout this century. Under the CCC

GCM scenario, precipitation over the Great Lakes is expected to change from ‒20 to +10% in

summer and ‒10 to +20% in winter (Magnuson et al. 1997; Mortsch and Quinn 1996). In contrast,

Gula and Peltier (2011) reports a 0‒8% increase in total precipitation (corresponding to significant

increase in rainfall and small decrease in snowfall) in 2050–2060 under the SRES A1B and A2

scenario. For 2090–2100, rainfall shows an overall increase by 10–20% and snowfall decreases by

40–50% in the southern part of the domain under the A2 scenario. Projections by Mortsch et al.

(2005) under various SRES-based scenarios show 1–15% increase in mean annual precipitation

by 2050.

Table 6.3 compares the outcomes from these models/scenarios. The model outputs (changes in

annual runoff, outflow, water level, and surface-water temperature) are generated on an individual

lake basis. Since the data used for model (baseline) simulation in Chapter 5 (Tahseen and Karney

2017) are available in the form of individual hydrologic parameters (precipitation, evaporation and

runoff) (NOAA GLERL 2015), they are adjusted based on the discussion above to represent the

changes in future climate.

93

Table 6.3: Simulated changes in the Great Lakes hydrology for 2°C rise in global temperature

under various climate scenarios

Scenario

Basin

runoff

(m3/s)

Over lake

precipitation

(m3/s)

Over lake

evaporation

(m3/s)

Transient (SRES) -5‒(-26)% 1‒13% 8‒26%

Transient (IS92a) -28‒4% -2‒11% 9‒24%

Transposition -25‒(2)% +3‒45% 49‒69%

CCC GCM -32% 0% +32%

GISS -24% +4% +27%

GFLD -23% 0% +44%

OSU -11% +6% +26% Source: Mortsch et al. (2006), Croley (2003), Croley et al. (1995), Croley (1993) Croley (1990)

Considering the uncertainty associated with these projections, the analysis considers a variety of

possible future scenarios discussed below:

Scenario 1: It considers a 3% increase in annual precipitation with a 105 mm increase in

annual evaporation (while keeping the runoff constant) following the WRF model projections for

2050–2060 under the IPCC SRES A1B and A2 emission scenario.

Scenario 2: The second scenario is based on the transient HadCM3 A1FI scenario for 2050

where mean annual precipitation is increased by 6% and 9%, evaporation by 18% and 26% while

lake outflows are reduced by 22% and 21% for Lake Erie and Lake Ontario, respectively.

Scenario 3: This scenario modifies mean annual precipitation by -1% and 5% and

evaporation by 12% and 23% from the baseline for Lake Erie and Lake Ontario, respectively. Lake

Erie outflow is reduced by 26% following the 2050 CGCM2 (transient) A21 emission scenario.

Scenario 4: The final scenario is based on the outcomes from the earlier equilibrium

models. Here, evaporation is increased by 32% (average of four GCM scenarios) while keeping

precipitation unchanged (as per GFDL). To account for the changing upstream conditions, mean

94

annual outflow from Lake Michigan-Huron (through Detroit River) is reduced between 20‒33%

as per the CCC and the OSU projections.

These changes are calculated on historical lake averages (1900‒2011) obtained from NOAA

GLERL (2015). The increase in evaporation is uniformly distributed throughout the year, while

the same for precipitation has a somewhat increasing trend during summer and fall.

6.3.2.2 Increased storage in Lake Erie At present, the available flow at the Niagara Hydropower Plant is subjected to the 1950 Niagara

River Treaty stipulations which establish that during the period lasting from April 1 to September

15, no less than 2,832 m3/s (100,000 ft3/s) must be going over the falls between 8:00 AM and 10:00

PM. The same flow restrictions are effective between 8:00 AM and 8:00 PM from September 16

to October 31. At all other times, a minimum of 1,416 m3/s (50,000 ft3/s) should be maintained

unless additional water is necessary (Government of Canada 2015). With the treaty restrictions in

place, an interesting possibility is to use Lake Erie as a storage reservoir. This scenario considers

a hypothetical control structure downstream of Lake Erie in order to retain flow, while minimizing

the resulting hydrological impacts. The flow structure allows a few centimeters of upstream lake

storage during the night, and releases the volume during peak demand hours for power generation

purposes. If plausible, this unique arrangement might lead to a stronger electrical grid while saving

billions of dollars otherwise needed in canal and reservoir construction.

In the model, Lake Erie outflow is partially retained from 0‒7 hour and again from 20‒23 hour

throughout the year unless the downstream water level is critically low. The resulting increase in

lake level is constrained by a maximum allowable limit of 0.3 cm from the baseline. The enormity

of the lake’s surface area (25,744 km²) allows 77 Mm3 additional volume to be stored with such

an infinitesimal increase in lake level. Since Lake Erie level is subjected to a natural variation of

13 cm during the monsoon (Ohio Department of Natural Resources 2014) – about forty times

higher than the proposed maximum increase – the regulation can at times reduce the seasonal

variations and minimize impacts on the shoreline. The arrangement is expected to benefit from the

lower treaty restriction (1,416 m3/s) which allows greater diversions for power during nighttime

and utilize this excess capacity when the demand peaks. Scenarios with further increase in storage

95

are left unexplored as they at times would violate the treaty flow restrictions. The generations at

the DeCew stations are left unchanged (from the baseline) in order to maximize the impact of the

regulation at the SAB Complex.

6.4 Results and Discussion

6.4.1 Climate change impact on hydropower potential Figure 6.3 (a, b) demonstrates Lake Erie and Lake Ontario elevations for 2050‒2060 under four

emission scenarios. Analysis of the model output aggregates the data on a monthly basis and

compares it with that of the baseline. The simulation results are found to be consistent with Croley

(1990, 1993), Hartmann (1990), Kling et al. (2003), Lofgren et al. (2002), Mortsch and Quinn

(1996), Mortsch et al. (2000) and Mortsch (2003) which reported that mean annual water depths

at the Great Lakes would decline below the historic levels. The 2050‒2060 SRES A21 emission

scenario (warm and dry) results in a 0.25‒1.2 m reduction in mean monthly water level at Lake

Erie. In contrast, reduced lake outflow under the equilibrium model (scenario 4) yields a maximum

1.4 m decrease in Lake Erie elevation. The decline in the lake level ranges from 0.24–1.1 m under

the SRES A1FI scenario (warm and wet) which describes a fossil fuel intensive growth. While the

changes in pool elevation are quite subtle in spring, increased evaporation during winter and fall

leads to a substantial drop in water level. Similar outcome under the A1B and A2 emission scenario

(scenario 1) appears to be quite optimistic (0.04‒0.2 m) since it disregards any impact on the

upstream lake conditions. The large decrease in lake elevations under the equilibrium model

relative to its transient counterpart (scenario 1, 2 and 3), as projected by this model, is rather logical

since the earlier GCMs (equilibrium models) do not account for the cooling effect of aerosols.

The impact on Lake Ontario is less severe where water level declines by a maximum of 0.28 m

under scenario 2, 3 and 4. These results are also consistent with Hebb and Mortsch (2005) and

Mortsch et al. (2006) which project a reduced impact on the lower lake. The relatively minor

changes in Lake Ontario elevations can be attributed to its large storage capacity (1,640 km³) and

the existing regulations (represented in the model as a controlled outlet) which mitigate short-term

variability. Interestingly, the simulation results show a slight increase (0.02 m) in Lake Ontario

elevations during summer and fall under the A1B and A2 emission scenario. While the increased

precipitation might cause such temporary upsurges, they are further influenced by regulation and

96

lack of consideration for the upstream lake conditions. These changing lake (Erie and Ontario)

conditions, though trivial, may exert significant impacts on the terrestrial and aquatic ecosystems

by modifying or eliminating wetlands (Branfireun et al. 1999; Devito et al. 1999; Lemmen and

Warren 2004; Mortsch 1998) or cause supply, odour, and taste problems in communities with

shallow water intakes (Nicholls 1999; Schindler 1998). The above discussion and comparisons

serve as a means of validating the model. In the following passages, the analysis focuses on its

core objectives, i.e., to evaluate the power systems performance under varying operating and

climate conditions.

Figure 6.3: Lake (a) Erie and (b) Ontario elevation under climate scenarios for 2050‒2060

172

173

174

175

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Lake

Erie

ele

vatio

n (m

)

Baseline SRES A1B & A2A1FI Equilibrium_33%A21

74.3

74.6

74.9

75.2

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Lake

Ont

ario

ele

vatio

n (m

)

Baseline SRES A1B & A2A1FI Equilibrium_33%A21

A1B & A2 A1FI

Equilibrium

A21

A1B & A1

Equilibrium

A1FI

97

The changing climate, apart from a sustained low water level, reduces the available discharge at

downstream hydropower plants in Niagara. Figure 6.4 shows the combined monthly discharge at

Niagara Hydropower Plant for 2050‒2060 under various climate scenarios. The A1B and A2

emission scenario that respectively describe a balanced and fragmented technological growth,

results in a maximum 4.3% reduction (137,000 m3/s) in mean monthly power flows. The

reductions under the A1FI and the A21 emission scenario show a comparable outcome ranging

between 4‒36% of the baseline generation. As expected, the maximum reductions are realized for

the GCM scenarios where monthly power discharge varies between 4.5‒40%. Also, these

reductions are not uniformly distributed over the year, but rather show a fairly large declination

during winter. In all the cases, winter power generation is found to be most impacted as the

simulation suggests a 4.3‒33% drop in available discharge for December.

Figure 6.4: Combined monthly available discharge at Niagara Hydropower Plant for 2050‒2060

Figure 6.5 demonstrates the impact on hydropower generation at the Canadian side under the

aforementioned climate scenarios. Monthly power generations for 2050‒2060 drop by 2‒20% and

3‒26% under the A1FI and the A21 emission scenario. These reductions are less pronounced

during spring and summer, even so that there is a slight increase (≈1%) in power production during

August under the A1B and A2 emission scenario. These represent conditions where the combined

effect of increased evaporation and precipitation leads to a slightly favourable outcome for power

generation. The combined annual production could potentially reduce by 1,100‒1,400 GWh over

the next 40 years under the A1FI and the A21 emission scenario. In contrast, the equilibrium GCM

0

1

2

3

4

Jan Apr Jul Sep Nov Dec Jan Apr Jul Sep Nov Dec

No inflow reduction Reduced Lake Erie inflow

Pow

er fl

ow (M

m3 /s

)

Baseline A1B & A2 Equilibrium_33% A1FI A21

98

scenario predicts a 6.5‒30% drop in monthly generation totaling up to a 2,000 GWh reductions in

annualized value. The findings resonate with studies by Buttle et al. (2004), Lofgren et al. (2002),

Mortsch et al. (2006), Shlozberg et al. (2014) and Wilcox et al. (2007) which have made similar

claims about the reduced hydropower generation potential in a warmer climate. Lofgren et al.

(2002) used transient CGCM1 and HadCM2 under the IS92a emission scenario with a hydrologic

model to quantify the changing generation along the St. Lawrence River. The hydropower needs

at these facilities could be satisfied less than 16% and 2% of the time for 2030 and 2050

respectively. Shlozberg et al. (2014) valued the losses from decreased power production in the

Erie-Ontario sub-region at $951M CAD through 2030 and $2.83B CAD through 2050. Apart from

the SAB and the Robert Moses station, the estimate includes RG&E (45 MW) and Varick station

(8 MW). It applies and updates the values reported by Buttle et al. (2004) with recent pricing data.

While the analysis by Buttle et al. (2004) was based on the earlier transient models (CCC-GCM1)

under the IS92a scenario, the reported reduction (25‒35%) in hydropower capacity is fairly close

to the value stated in this study. The study by Mortsch et al. (2006), the most contiguous to the

present appraisal, reported a 17% and 14% reductions in hydropower generation on the St.

Lawrence River for 2050 under the CGCM2 A21 and the HadCM3 A1F1 scenario. While these

reductions represent the impact on the Moses-Saunders Power Project and the Beauharnois-Les

Cèdres Complex, the outcomes correspond with the present-day analysis.

Figure 6.5: Monthly power generation at SAB Complex by 2050-2060 under various climate

scenarios

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Gen

erat

ion

(TW

h)

Baseline A1B & A2 Equilibrium_33% A1FI A21

99

6.4.2 Impact of lake storage on hydropower potential The analysis here considers a regulation plan that increases Lake Erie level by 0.1–0.3 cm during

specific hours of the day and evaluates the resulting impact on generation potential at SAB

Complex. The regulation essentially prioritizes on-peak power generation while compromising

production during low demand hours through the means of flow shifting. Figure 6.6 shows the

available discharge at the SAB Complex under the flow regulation scenarios. Surprisingly, a 350‒

600 m3/s increase in available power flow is observed on an annual basis. In the absence of

adequate tunnel and/or canal capacity, a portion of the available power flow that was previously

directed either towards Robert Moses plants or over the falls (in the baseline scenario), now

becomes available at the SAB Complex. This represents a 390‒570 GWh increase in power

generation potential at the SAB Complex. However, the generation growth is not proportional to

increased regulation, rather achieves a maximum potential with 0.2 cm increase in lake level and

then decreases. With increased regulation, the previously underutilized tunnels/canal may reach a

point where they are insufficient to handle all the water offered by 0.3 cm additional lake storage.

Figure 6.6: Available discharge at the SAB Complex under the flow regulation scenarios

The impact of lake regulation can most readily be realized by analysing hourly power generation

data. Figure 6.7 compares the diurnal generation at the SAB Complex under the lake regulation

scenarios with that of the baseline. A typical day in July is selected for demonstration as Ontario’s

high power demand coincides with the 1950 Treaty requirements during this month. The flow

shifting achieved through the proposed regulation results in a 0.1‒30% increased (hourly)

800

1000

1200

1400

1600

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Pow

er fl

ow (x

103

m3 /s

)

Baseline 0.1 cm 0.2 cm 0.3 cm

100

generation during peak hours, while simultaneously decreasing production during the rest of the

day. The generation spikes in Figure 6.7 represent the beginning and ending of the treaty-specified

tourist flow hours during which flow over the Niagara Falls is reduced from 2,832 m3/s to 1,416

m3/s, resulting in a substantial increase in power diversion.

Figure 6.7: Hourly variation in power generation with lake regulation

6.4.3 Combined climate change and lake storage scenarios Here, the authors combine the climate change with the lake storage scenarios to investigate the

potential of lake regulation as mitigation against climate induced variability. The discussion will

examine in detail one such scenario, i.e., the combination of the A1FI emission scenario with

increased lake storage. Figure 6.8 illustrate the average hourly generation at SAB Complex for

July under the baseline, the future climate and the combination of climate and lake storage. Lake

regulation, i.e., storing water during nighttime for timed release during peak demand hours, not

only shifts generation from off-peak to peak hours, but also increases it by 2‒3% annually. Similar

to the outcome in the previous section, the potential for such increase (generation) reaches a

maximum with 0.2 cm additional lake storage followed by a slow decline.

600

900

1200

1500

18000:

00

2:00

4:00

6:00

8:00

10:0

0

12:0

0

14:0

0

16:0

0

18:0

0

20:0

0

22:0

0

Gen

erat

ion

(MW

h)

Baseline 0.1 cm 0.2 cm 0.3 cm

101

Figure 6.8: Hourly generation under the baseline, the climate and the combined climate and lake

storage (0.3 cm) scenario for July

While increased evaporation and changing precipitation pattern under a varying climate results in

an elevation drop as high as 1 m as outlined in the earlier sections, lake regulations can partly assist

in reducing this variability. Though the systems response under the combined scenario, as

illustrated by Figure 6.9, results in greater fluctuations on a short-term (24 hr), it may ameliorate

long-term lake level lows. While the analysis here considers a daily drawdown irrespective of the

lake level, it is worthwhile to note that such (lake) regulation can be accompanied with guidelines

conditional upon maintaining a critical lake elevation.

6.5 Limitation The study has several crucial limitations. First, due to the restrictive access to specific hydrologic

and power systems data, information required for building the model is obtained from various

publicly available documents, web resources etc. Specifically, the model’s assumption of all

available flow to be used for power generation purposes may not be realistic, since such decisions

are often guided by demand and available generation from other non-dispatchable and intermittent

sources for a particular duration. Consequently, it is a common practice for many hydroelectric

stations either to store or to release surplus flow through spillways to preserve the upstream

conditions. Also, the model is simulated using hourly water level data at three major gauges along

the river. There is a potential to improve the model’s ability to simulate the baseline condition, and

0

500

1000

1500

2000

0:00

2:00

4:00

6:00

8:00

10:0

0

12:0

0

14:0

0

16:0

0

18:0

0

20:0

0

22:0

0

Hou

rly g

ener

atio

n (M

Wh)

Baseline Climate change Climate + lake storage

102

consequently its responses under different scenarios with information on rest of the stations.

Second, the model uses hydrologic routing (Muskingum-Cunge) rather than hydrodynamic

approaches. Though, Muskingum-Cung provides reasonably accurate results for moderate flows

propagating through mild to steep sloping watercourses (Maidment and Fread 1993), the method

sometimes produces unrealistic initial negative dips in the computed hydrograph (Baláž et al.

2010). Considering the intensive data requirement for a large-scale basin such as Niagara (Arora

et al. 2001), hydrologic approach is preferred here over a more accurate hydrodynamic model.

Third, the constant values used in the model for local flow contributions at the GIP and the Welland

Canal are vulnerable to changes with the changing evaporation and precipitation patterns.

However, the river has no more than a 0.8% contribution from these factors; thus resulting in

negligible error from this source. Lastly, the outcome of this research should be used with

reservation as there are considerable uncertainties associated with the projections. The major

predicaments are sourced from the future emissions (captured by various scenarios), uncertainty

regarding the climate’s response to it and due to the biases introduced by dynamic downscaling.

6.6 Conclusion and Discussion Considering the increases in various stressors – whether climate, population growth or economic

development – securing the supply and equitable allocation of water to support human well-being

while sustaining healthy, functioning ecosystems is one of the major challenges of the twenty-first

century. Awareness of increasing water scarcity has driven efforts to model water resources for

improved insight into infrastructure and management strategies. Despite the increasing research

efforts, there are still considerable uncertainties regarding the impact on water resources, and how

the systems dependent (water supply, power etc.) on these resources will respond to changes. The

current chapter contributes to the existing knowledge base by evaluating the impacts of a changing

climate in terms of hydropower generation potential. It explores a wide variety of "what if"

scenarios to investigate the likely impact on power generation at Niagara Hydropower Plant. The

analysis is further extended to investigate an ambitious lake regulation plan that may reduce natural

variability, while increasing generation potential with storage infrastructure development. When

evaluated for possible climate change scenarios, the model predicts a 4‒33% decrease in combined

annual power flow at the downstream hydroelectric stations. For Canada, monthly power

generation potential reduces by a maximum of 8% over the next 40 years, while in long run it

103

ranges from 9‒30% for doubling CO2 concentration scenarios. The proposed regulation plan bears

the potential to increase peak-hour generations by 0.1‒30%, while simultaneously decreasing

production during the rest of the day.

Previously published work has predicted a substantial increase in evapotranspiration in the Great

Lakes catchment despite increases in precipitation. The current regulation plan at Lake Superior

and Lake Ontario can only partially alleviate lake level extremes and do not affect the long-term

trends. The uncertainty associated with the future of the Great Lakes may call for increased

regulation – a potential scenario that is briefly assessed in this study. The primary contribution of

this chapter is to develop a simulation model to investigate the impact on power generation

potential under a warming climate. This exploratory study makes no pretense to forecast likely or

advisable developments, but rather recommends further research ideally coupled to an opening of

new discussions between Canada and the US.

104

Section 3 Sustainability and Resilience Assessment of Hydropower

Systems While the forgoing sections explore the techno-economic viability of increased hydropower

generation in Ontario and assess its sensitivity to climate change, the third and final segment of

the thesis is devoted to sustainability evaluation. Sustainable development, sometimes still referred

in the classic way as “meeting the needs of the present without compromising the ability of future

generations to meet their own needs” (WCED 1987), is an increasingly important consideration in

resource management. Despite being recognized as a fundamental part of any decision making

process, sustainability is too often viewed as an isolated agenda competing with other priorities.

Considering a wide spectrum of possible interpretations of ‘sustainable development’, this section

begins with establishing a clear definition of the goal, so as to direct the development of a

sustainability assessment tool for our purpose. To this end, Chapter 7 provides a systematic review

of the existing literature on sustainable hydropower development. The limitations in the published

approaches that neglect the role of hydroelectric resources in stabilizing the electrical grid and

leveraging investments in other intermittent renewables motivate the development of

Sustainability SWOT (sSWOT) model. It can be easily interpreted and ranked, and encourages

stakeholders’ participation – which makes it particularly palatable for strategy/policy

prioritization. The model framework and its detailed application for evaluating increased

hydroelectric generation options at Niagara are illustrated in the following chapter. Chapter 9

elaborates the development and application of System Integrity Evaluation (SIE) model, a novel

risk assessment tool specifically designed to address the uncertainty of environmental and climate

projections.

105

Reviewing and Critiquing Published Approaches to the Sustainability Assessment of Hydropower

Chapter 7 reviews a series of published economic, environmental and social indicators that are

used to characterize hydroelectric resource. It summarizes the current state-of-the-art in the field

of hydropower sustainability assessment, so as to evaluate the schemes proposed in chapter 2. The

discussion here argues that present studies sometimes set system boundaries too narrowly so that

they omit key factors associated with hydropower. In particular, the role that hydroelectric

resources can play to stabilize the overall electrical grid, and thus to leverage investments in other

intermittent renewables, is only rarely accounted for in the current sustainability assessments.

Based on a broad literature review, the authors articulate two key recommendations: first, that such

assessments should reflect on policy issues as well as environmental challenges with respect to

existing hydropower potential within the current framework; second, that system boundaries

should be extended not only to allow broad hydrological, ecological and geological assessments,

but also to reasonably estimate hydro’s potential benefits to the functioning of the overall electrical

grid.

This chapter is based on the paper entitled “Reviewing and Critiquing Published Approaches to

the Sustainability Assessment of Hydropower” by Samiha Tahseen and Bryan Karney, published

(15.09.2016) in the Journal of Renewable & Sustainable Energy Reviews. It outlines the definition

of sustainability and the existing approaches towards achieving it for hydroelectric systems.

7.1 Introduction Robust electrical supply systems clearly play a crucial role toward achieving human well-being

and can act as a foundation of economic growth and prosperity. Yet with increasing concern over

global climate change and the health ramifications of using carbon-based fuels, a progressively

greater use of renewable resources is seen as having distinct advantages over non-renewables

(Koljonen et al. 2009; Ballester and Furió 2015). Several studies have surveyed the environmental

and climatic effects of the fossil fuel energy systems (Pereira and Pereira 2014; Singh et al. 2012)

and the benefits of a transition to a lower-pollution energy system (Pan et al. 2014; Wakiyama et

al. 2014; Zhang et al. 2013), including both hydropower and natural gas (Zhang et al. 2014a; Zhang

106

et al. 2014b). Some studies have focused on energy-based carbon emissions (Ma et al. 2011; Wang

and Yang 2015; Zhang 2003; Wang et al. 2012), while others have traced and modelled various

mechanisms that lead to environmental effects (Zhang et al. 2013; Wang and Feng 2015). One

obvious conclusion is that hydropower, when used well, has the potential to reduce the pollutants

and carbon emission, and thus to improve environmental and climatic health. Of course, when

used poorly hydropower can cause devastation to human and natural systems, as many well-known

historical disasters attest (Paolini and Vacis 2000).

Apart from its doubtless advantages, even nominally successful hydro projects are often associated

with negative environmental consequences in the form of biodiversity loss, disruptions to fish

migration, potentially large-scale land inundation, the disruption of human resettlement, and many

others. Although early consideration and adoption of mitigation measures can limit such impacts,

it is often impossible to completely eliminate or fully control the adverse influences on local

ecosystems. The modern evaluation of the impact of engineering projects ideally considers –

indeed, is often mandated to consider – anticipated economic, social and ecological impacts

through all stages of the development. While standard environmental impact assessments may

have been enough in the past, more detailed guidelines are now prescribed by many international

(financial) institutions such as World Bank (2013), International Hydropower Association (IHA),

International Energy Agency (IEA), European Bank for Reconstruction and Development

(EBRD), and others. These guidelines establish a set of recommendation for impact assessment,

suggesting ways to ameliorate adverse effects and criteria for the application of mitigation

measures. The process in turn provides a basis for comparison between hydropower and other

electricity generation sources. The need to seek a wide range of opinions during project

implementation is reflected in most of the published mandates. Due to the variety of contexts and

perspectives, the list and priority of proposed indicators naturally varies between the numerous

published guidelines, studies, and reports. Although there is, as yet, no universally accepted

standard for assessing the sustainability of hydropower projects, there is an obvious and important

overlap in the obligations to consider.

All power developments, be they coal, nuclear, gas, wind or solar, offer certain benefits while

possessing inherent drawbacks. Thus, when performing a sustainability assessment of a project,

it is imperative to evaluate as much as possible the system as a whole, not just its individual

107

components. But this creates ambiguity regarding the scale at which sustainability should be

examined since, realistically, no power system, nor indeed any large scale human activity, is

absolutely sustainable. This paper argues that while existing frameworks are quite comprehensive

on a project scale, there are merits of choosing a scale that is sufficiently broad to take into account

all key impacts, including those occurring at the grid level. To this end, the study first summarizes

existing knowledge and compares several different sustainability approaches or guidelines.

7.2 Synergy Between Hydropower and Sustainability Sustainable development is a concept that at its core is both revolutionary and somewhat intuitive,

yet incredibly difficult to comprehensively define and perhaps even more difficult to fully

operationalize. In the absence of a collective, pragmatic and operational interpretation, the existing

literature encompasses a variety of approaches, frameworks and models for evaluating sustainable

practices (Altieri 1987; WCED 1987; Douglass 1984; Norton 2005; Ott 2003; Seghezzo 2009;

Thompson 1992 and 2010; Werkheiser and Piso 2015; Williams and Millington 2004). The

simplest of these approaches is perhaps the concept of the triple bottom line that recognizes that

sustainability rests upon three pillars encompassing economic, environment and social domains.

There is debate about who originated this approach, though it is contended by some to have been

first used by Altieri (1987) in relation to agricultural production. Despite being a little ambiguous,

the most widely cited definition of sustainability is now decades old, stated as “development that

meets the needs of the present without compromising the ability of future generations to meet their

own needs” (WCED 1987). By contrast, Thompson’s (1992) approach distinguishes between

systemic and goal-directed sustainability. As a systems approach, sustainability inquires about

whether a system, with defined boundaries, external inputs and self-healing capacity, can continue

over a specified time scale without major degradation to its context or to itself – informally,

without falling apart. Another way of assessing sustainability is to use Life Cycle Analysis (LCA),

which traces relevant input-output data for the purpose of estimating resource consumption and

emission from the system throughout phases of design, construction and operation.

The International Energy Agency (IEA) predicts the growth in electricity demand at an annual rate

of 2.5% sustained until 2030 and that will require and commensurate energy investment up to $26

trillion (IEA 2009). Indeed, the overall context for the assessment of energy sustainability is

108

dominantly one of growth. Using traditional fossil fuel to meet the growing power demand is both

unsustainable (since these resources cannot be replenished in a reasonable time frame), and is

tending to serious climate repercussions (Wang et al. 2014; Wang et al. 2016; Chiari and Zecca

2011). In this context, implementation of hydro can have the enormous benefit of reducing fossil

fuel-based generation (Tahseen and Karney 2016: Chapter 3; Wagner et al. 2015; Li et al. 2015).

From a life cycle perspective, the CO2 produced during construction and operational phase of

hydro projects is seldom comparable in scale to that associated with the use of non-renewables.

For example, decomposition of flooded biomass in hydro reservoir can cause emission of 4–8 gram

CO2 eq./KWh (Gagnon and van de Vate 1997; Meier 2002; Tremblay et al. 2006; van de Vate

2002) which is 36 to 167 times lower compared to that of fossil fuel-based generation (Tremblay

et al. 2006; van de Vate 2002). Overall greenhouse gas emission (GHG) from a typical hydro plant

ranges from 2 – 18 kt CO2 eq./TWh throughout its life cycle while that of fossil fuel run plants

ranges from 389 to 1272 kt CO2 eq./TWh (Fritsche 1992; Gagnon et al. 2002; IEA 1998; Uchiyama

1996; Zhang et al. 2007). Studies have also shown that development of even half of the world's

economically feasible hydropower potential could reduce GHG emissions by about 13%, and the

impact on avoided SO2 and NOx emissions would be even greater (Bates et al. 2008; Swingland

2003).

Apart from using a renewable power source, hydroelectricity usually includes a capacity to store

energy and thus can provide flexibility to the operation of the grid (Rehman et al. 2015; Maxim

2014; Zhang et al. 2015). If leveraged well, this storage/reserve function can allow greater

integration of intermittent renewables, particularly wind and solar resources (Ayodele and

Ogunjuyigbe 2015; Caralis et al. 2012; Steffen 2012; Kusakana 2015). On a community level,

hydropower projects are often multi-purpose in nature; serving various needs including power,

flood control, water supply, and recreational benefits (Capik et al. 2012; Evans et al. 2009;

Kaygusuz 2009). Investments in infrastructure (access roads, dams, and canals), communications,

and skill building in large projects can support regional economic development. On the negative

side, though, these projects often inevitably alter many environmental and social parameters due

to the conversion of portions of terrestrial ecosystems into aquatic ones, whether through

resettlement, restriction of navigation, modifications of local land use, loss of biodiversity, or

changes in aquatic sediment composition and distribution (Nautiyal et al. 2011; Ribeiro et al. 2011;

Sarkar and Karagöz 1995; Sternberg 2008; Williams and Porter 2006; Yuksel 2010a; Xingang et

109

al. 2012). Interestingly, often these effects vary from one case to the other irrespective of the

project scale. Consequently, it is invariably challenging to make generalized comments regarding

the impact of hydropower development on surrounding environment (Frey and Linke 2002). Even

projects having the same installed capacity may have quite different environmental consequences

depending on specifics of design, hydrology, geology, available field conditions, and the specific

fluvial parameters.

Certainly one of the key attractions of hydropower is its occasional abundance and that it is both

renewable and dispatchable; but perhaps just as obviously, the environmental and social

consequences of hydro’s development can range from daunting to devastating. It is also obvious

that preference for hydroelectric projects, or indeed for almost any type of power project, can

seldom be judged solely on its own attributes, but depends also on local system (grid)

requirements, on the performance and availability of other local options and on whether the

associated impacts and risks can be limited.

7.3 Existing Frameworks and Guidelines on Sustainable Development of Hydropower

At present, nearly all countries mandate an assessment of the expected impacts of any new

hydroelectric development prior to construction. Nevertheless, historically many developers have

apparently perceived such requirements as mere formalities, a rather onerous but necessary steps

to obtain regulatory approval with requirements to be as frequently ignored during implementation

(Kumar and Katoch 2014). Such occurrences are perhaps even more common in developing

countries (Bell and Russell 2002). Several international and financial institutions provide monetary

assistance (including low interest loans) for large-scale hydropower development. In most cases,

the funding is conditional on reasonable performance under agreed frameworks. The current paper

reviews the major institution-specific guidelines on sustainability assessment of hydropower

projects starting with the frameworks proposed by various international organizations, certification

bodies and funding agencies. Latter sections aggregate other indicators extracted from research

studies, project documents and expert recommendations.

110

7.3.1 Low Impact Certification by LIHI Originating in 2000, the Low Impact Hydropower Institute (LIHI) is a US-based non-profit

organization dedicated to certifying hydropower projects with reduced environmental impacts.

Their stated purpose is to protect the river ecosystem and enable low impact projects to access

renewable energy markets. The approach assesses hydroelectric schemes based on a number of

criteria developed in eight categories: river flows, water quality, upstream fish passage,

downstream fish passage and protection, watershed protection, threatened and endangered species

protection, recreation, and cultural resource protection (LIHI 2014). Each criterion is evaluated on

a pass-fail basis, and satisfactory performance in all the criteria is required for certification.

This certification emphasizes the ecological impact with little focus on social and economic aspect

of hydropower development. So far, a total of 121 projects have been certified by LIHI with 9

under review and 16 pending applications (LIHI 2016a). Until recently, the certification program

was strictly limited to run-of-river plants and did not extend to pumped storage or projects that

require construction of new dam or diversions. However, through a subsequent revision in 2016,

the guidelines were extended to cover facilities with limited storage capacity (LIHI 2016b). The

revision proposes no new criteria but offers an extended list of alternative standards to ensure

compliance.

7.3.2 Green Hydropower Certification by EAWAG Following a successful pilot certification, the Swiss Federal Institute of Aquatic Science and

Technology (EAWAG) presented Green Hydropower Certification Scheme. It sets out the

technical basis of a uniform and scientific certification process for hydropower plants. The

program claims that following its stated procedure can ensure design and operation of a facility

that safeguard basic features of the ecological integrity of the river system (Bratrich and Truffer

2001). The process involves forming an environmental management matrix that accounts for direct

impact on the river ecosystem and its riverine landscape. It places five management criteria in

different columns describing the operational issues or aspects of construction related to

hydropower development, i.e., minimum flow regulations, hydro-peaking, reservoir management,

bedload management and power plant design. Environmental dimensions, such as hydrological

character, connectivity of river system, solid material and morphology, landscape and biotopes,

111

etc., are placed in separate rows and are expected to cover the most important aspects relevant to

ensuring ecological viability (Bratrich and Truffer 2001). The elements within the matrix are

assigned with specific requirements and are designed to be universally applicable to all kinds of

hydropower plants. As the concepts and criteria are independent of Swiss law requirements, the

scheme should be applicable in principle to other countries with minor modification. However,

this should not be used for licensing purposes or as a substitute to environmental impact analysis

as warned by EAWAG. Similar to LIHI, the certification scheme concentrates on ecological issues

with little consideration on key economic and social indicators associated with hydroelectric

development.

7.3.3 Hydropower Sustainability Assessment Protocol (HSAP) by IHA IHA’s so-called HSAP protocol is the outcome of a collective effort by Hydropower Sustainability

Assessment Forum that is launched in 2008 by the International Hydropower Association (IHA)

along with its key strategic partners. Recognizing the inconsistencies in the existing approaches,

the framework aims to develop an enhanced sustainability assessment tool to measure and guide

performance, and to streamline the approaches for hydropower projects. It encourages or seeks a

considerably high level of convergence amongst the diverse views from its members which

included representatives of governments from both developed and developing countries,

commercial and development banks, social and environmental NGOs, and the hydropower sector.

The 2010 Protocol updates 2006 version and comprises a set of four stand-alone assessment

frameworks: an early stage tool for assessing risk and opportunities during initial planning

followed by three detailed schemes for preparation, implementation and operation stage (IHA

2010). The proposed indicators further reflect on four different sustainability perspectives ‒

economic, environmental, social and technical. The detailed guideline is summarized here in Table

7.1. A 5-level scale system is used to determine the status of each criterion where level 5 describes

the proven best practice, level 3 stands for basic good practice while level 1 represents significant

gaps relative to accepted practices.

From a sustainability viewpoint, the protocol is unique with its handling of the issues from both a

triple bottom approach and life cycle perspective. Interestingly, the protocol does not provide any

specification regarding the requirements for acceptable performance on the criteria, rather relies

112

on the institution’s policies and positions to decide on such critical issues. Since its introduction,

HSAP has been implemented in developing several large-scale hydropower projects including

China’s Three Gorges. Realizing the complex nature of the project, sustainability issues were

evaluated using a triple bottom line approach, with the LCA consideration and from a systems

perspective (i.e., reservoir, dam, power plant, transmission, the location of the project and the

surrounding area) (Liu et al. 2013).

Table 7.1: Hydropower Sustainability Assessment Protocol topics (IHA 2010)

Technical Environmental Social Economical Integrated Siting and

design Downstream

flows Project affected

communities and livelihoods

Economic viability

Demonstrated need and strategic

fit Hydrological

resource Erosion and

sedimentation Resettlement Financial

viability Communications and consultation

Reservoir planning, filling and management

Water quality Indigenous peoples

Project benefits Governance

Infrastructure safety

Biodiversity and invasive species

Cultural heritage Procurement Integrated project management

Asset reliability and efficiency

Waste, noise and air quality

Public health Environmental and social issues

management

7.3.4 Directions in Hydropower by World Bank Recognizing its multidimensional role in poverty alleviation and sustainable development, the

World Bank emphasizes on exploiting the maximum strategic value of hydropower resources in

an environmentally and socially responsible manner. Directions in Hydropower (World Bank

2013) summarizes the key issues in scaling-up hydropower projects along with setting priorities

for the organization in lending and nonlending activities. Unlike the sustainability frameworks

reviewed here, it discusses key challenges and both policy and regulatory issues regarding

hydropower development. The directive highlights the bank’s two track approach towards

hydropower scale-up: first, through direct investment and second, by strengthening sectoral

foundations by providing technical assistance, knowledge sharing, initiating policy dialogue and

113

several other roles. In 2014, the World Bank endorsed the Hydropower Sustainability Assessment

Protocol (HSAP) as a tool for guiding hydroelectric development in client countries (Liden and

Lyon 2014). The relatively slow adoption of the protocol by low- and middle-income countries

where much of the remaining hydropower potential exists, motivates the bank to raise awareness

about the HSAP, particularly through sector-level engagement. A pilot assessment was carried out

on a World Bank-supported Vietnamese hydropower storage plant to learn about the practicalities

of the tool. While the protocol findings are conducive to management action, the manuals are

reported to be complex and insufficient (Liden and Lyon 2014). Given the Protocol documents’

extreme site-specificity, the bank further emphasizes the use of accredited assessors for quality

assurance.

7.3.5 Hydropower Implementing Agreement by IEA The Hydropower Implementing Agreement is a collaborative program under International Energy

Agency (IEA) which aims “to improve technical and institutional aspects of the existing

hydropower industry and increase future development in an environmentally and socially

responsible manner” (IEA 2006). The outcome of this agreement results in several technical

reports that provide a comprehensive overview in its entirety including trends in hydroelectric

development, comparative analysis with other generation sources, ethical considerations,

financing options, methods for education and training in hydropower etc. The first phase of the

program reviews the processes and conditions which make hydroelectric projects environmentally

and socially acceptable, identifies international best practices, and proposes a set of

recommendations (IEA 2000). This document reflects the points of view of academic specialists

and professionals from varied backgrounds and organizations from IEA member countries. Table

7.2 highlights the key issues discussed in the report. Following the identification of the indicators,

their representativeness is verified with knowledge gathered from sixty case studies around the

globe and rigorous experts’ examination.

The authors of the current paper found the work of the task force under its Hydropower

Implementing Agreement to be quite comprehensive, highlighting issues such as restructuring of

the electricity market, reliability and backup benefits of hydroelectric resources, and credits it for

avoided emission and impact on human health from LCA and environmental impacts perspective.

114

Table 7.2: List of environmental and social indicators under IEA framework (IEA 2000)

Categories Key Issue

Biophysical environment

Reservoir impoundment

Loss of biological diversity

Sedimentation characteristics

Water quality

Hydrologic Regime

Barriers for fish migration and river

navigation

Socioeconomic

environment

Resettlement and rehabilitation

Health and safety Impacts

Vulnerable community groups

Land Use and cultural heritage

Sharing development

benefits

Benefits due to power generation

Benefits due to dam function

Improvement of Infrastructure

Development of Regional Industries

7.3.6 Sustainable Energy Financing by EBRD The European Bank for Reconstruction and Development (EBRD), operating primarily in Central

and Eastern Europe, finances hydropower projects under its Sustainable Energy Financing

Facilities (SEFFs) initiative. EBRD uses eight environmental criteria for assessing hydropower

projects: environmental flow, water quality, fish passage and protection, watershed protection,

threatened and endangered species, recreation, cultural and community issues (EBRD 2013).

Mitigation of negative impacts under these criteria is vital to complete licensing procedure and

secure funding from EBRD. Table 7.3 provides a list of criteria prescribed by EBRD.

Due to its attractive simplicity and clarity, several researchers have used the guideline for

evaluating hydropower projects. Kucukali (2014) assessed the environmental risks associated with

115

a small plant on the basis of EBRD standards. Each of the five criteria was scored on a scale of 1

to 3 on the basis of documented evidence, measured data, and observations during the operation

stage of the plant. A similar study by Schmalz and Thürmer (2012) found a 300 kW hydropower

plant in Germany to be of low environmental risk. From a sustainability viewpoint, the existing

framework by EBRD seems to be missing a few critical environmental indicators (such as loss of

biodiversity, impact on flora and fauna), and also lacks considerations of key social, economic and

technical issues. Nevertheless, other impacts being negligible, this sort of simple scoring system

can offer a first glimpse at project EIA during the initial planning stage.

Table 7.3: Environmental criteria for hydropower projects under EBRD (EBRD 2013)

Criteria Requirement Environmental flow

Maintains a minimum river flow accounting for seasonal fluctuations.

Water quality Does not contribute to deterioration of upstream or downstream water quality.

Fish passage Has minimal impact on local fish populations and provides effective fish passage.

River basin Does not negatively impact environmental conditions in the river basin or integrity of the existing ecosystem.

Endangered species

Not constructed on a protected river and neither negatively impacts any endangered species.

Recreation Accommodates recreational activities.

Cultural Issues Protects archaeological, paleontological, historical, religious and unique natural values.

Community Issues

Does not stop or limit local communities’ ability to provide a livelihood.

7.4 Selective Review of Sustainability Indicators A broad list of indicators for assessing hydropower projects is now summarized based on a range

of research articles. The selection process was guided by a systematic review where relevant

materials were collected by conducting multiple search operations with designated keywords such

as hydropower, sustainability, power systems etc. on major academic databases for scientific and

technical research. The entries found throughout this process were first screened for relevance and

their bibliographies were further scanned for related scientific resources that were missed during

116

the initial search operations. To the best of the author’s knowledge, literature on sustainable

development of hydropower dates as far back as 1995. Thirty-five documents were found to be

dedicated, either partially or entirely, to this issue. While a few of the selected studies primarily

rank generation assets with respect to a set of criteria (Maxim 2014; Evans et al. 2009; Afgan et

al. 2000; Onat and Bayar 2010; Afgan and Carvalho 2002; Carrera and Mack 2010; Scannapieco

et al. 2014), the majority focused exclusively on hydroelectric power (Rosso et al. 2014;

Supriyasilp et al. 2009; Vučijak et al. 2013; Morimoto 2013; Ji et al. 2015; Stevović et al. 2015;

Chen et al. 2015; Pang et al. 2015; Bakis and Demirbas 2004; Kaunda et al. 2012; Klimpt et al.

2002; Capik et al. 2012; Kentel and Alp 2013; Kaygusuz 2002; McNally et al. 2009; Yuksel 2010b;

Balat 2007; Sparkes 2014; Jager et al. 2015; Zhang et al. 2015). Table 7.4 provides a chronological

list of research studies along with the reported technical, ecological, economic and social

parameters.

Due to the complexity in quantifying social impacts, the current literature identifies it as one of

the most challenging aspects of hydropower development. Indeed, the International Association

for Hydro-Environment Engineering and Research (IAHR) has recently identified this topic as a

real need in hydropower assessments (IAHR 2005). Involuntary resettlement, relocation and

rehabilitation of local indigenous community, potential conflict and increased incidence of

waterborne diseases are widely identified indicators in this category. Benefits of hydroelectric

development reported in these studies typically include job creation, investment in infrastructure,

irrigation, recreation, tourism, navigation etc. The environmental indicators cited by most

researchers usually touch on inundation, loss of biodiversity, fish migration, land use, reservoir

sedimentation and water quality. A handful of studies have also included increased water

temperature, lower dissolved oxygen, loss of soil fertility, soil erosion, increased salinity or lesser

recognized fact of seismic activity to the list. Whereas GHG emission is considered a major

criterion when comparing hydropower with other generation resources, avoided emissions is rarely

identified as an environmental benefit in studies that solely focus on hydroelectric generation.

Other more rarely mentioned environmental indicators include its impact on fisheries or other

projects in the vicinity, aesthetics and methane emission as a result of decomposition of buried

organic matter in hydro reservoirs. While social and environmental issues associated with

hydropower development are widely discussed in the existing literature, economic and technical

parameters have received less attention. A limited number of studies have highlighted technical

117

issues such as efficiency, difficulty in construction, flexibility and the influence of the estimated

development period. A similar practice of exclusion is observed for plant life, unit electricity cost,

and presence of other infrastructure (access roads, transmission networks, etc.) which substantially

affect project economies. The review process also suggests a contrast among researchers regarding

the hierarchy of certain indicators. Job creation and complementary benefits of hydropower

(irrigation, recreation, tourism etc.) are at times considered as social indicators while a few have

placed them under economic category.

The review also sheds light on the typical methods used for assessing hydropower sustainability.

Multicriteria analysis is found to be the most common among published approaches (Maxim 2014;

Evans et al. 2009; Afgan and Carvalho 2002; Rosso et al. 2014; Supriyasilp et al. 2009; Vučijak

et al. 2013; Morimoto 2013; Ji et al. 2015) with variations such as the weighted sum method

(WSM), the weighted product method (WPM), the preference ranking organization method for

enrichment evaluation (PROMETHEE), the elimination and choice translating reality

(ELECTRE), the technique for order preference by similarity to ideal solution (TOPSIS), Analytic

Hierarchy Process (AHP) being widely used. The methodology is at times complemented with

stakeholders’ analysis (Carrera and Mack 2010; Rosso et al. 2014) or life cycle assessment

(Scannapieco et al. 2014). Recent literature includes the application of relatively novel

methodologies such as fuzzy mathematical functions (Stevović et al. 2015), information network

analysis (INA) (Chen et al. 2015) and emergy analysis (Pang et al. 2015) for evaluating non-

technical and ecological impacts of hydropower construction. Many researchers provide a rather

general (Kaygusuz 2009; Bakis and Demirbas 2004; Kaunda et al. 2012; Klimpt et al. 2002) or

project/country-specific narrative (Sarkar and Karagöz 1995; Capik et al. 2012; Kentel and Alp

2013; Kaygusuz 2002; McNally et al. 2009; Yuksel 2010b; Balat 2007; Sparkes 2014) that

highlights the key issues. A few of the papers are review articles that further extend the scope by

suggesting spatial design principles (Jager et al. 2015) or performing a systematic assessment of

hydropower externalities (Zhang et al. 2015).

118

Table 7.4: List of hydropower sustainability indicators reported by researchers

Reference article Social Environmental Economic Technical

Sarkar and Karagöz (1995)

Resettlement, job creation, waterborne diseases, traffic, immigration, colonization

Biodiversity loss, fish migration, deforestation

Recreation, tourism navigation

Afgan et al. (2000)

Job creation, capital produced, diversity and vitality

Emission of CO2, NOx, SO2, waste Efficiency, investment per unit power, GNP per KWh

Afgan and Carvalho (2002)

CO2 Emission, land use Efficiency, installation cost per kWh, electricity cost

Kaygusuz (2002)

Relocation, waterborne diseases, colonization, migration, job creation, investment, traffic, recreation, tourism, navigation

Inundation, species extinction, landslides, biodiversity loss, fish migration, erosion, soil fertility & salinity

Drought & flood protection, affordable power, construction cost

Klimpt et al. (2002)

Public participation, resettlement, irrigation, heritage sites, shared benefits, human health

Hydrologic regime, biodiversity loss, fish migration, water quality, sedimentation, flood, avoided emission, seismic activity

Efficiency, flexibility, demand response, storage, black-start

Bakis and Demirbas (2004)

Displacement, employment opportunities, living standards

Sedimentation, topographical & hydrological conditions, flooding, species extinction, ecosystem

Unit electricity cost, capital & maintenance cost, irrigation, water supply benefits

Balat (2007) Visual impact, irrigation No pollution, flooding, fishery, avoided GHG emission, air quality

Health cost, cheap power, life span, maintenance

Evans at al. (2009)

Public acceptance, displacement, flood protection, irrigation, recreation

GHG emission, inundation, land use, water consumption, siltation

Price per KWh Availability, efficiency technological limitations

Kaygusuz (2009)

Resettlement, recreation, drought & flood protection, navigation, waterborne diseases

Climate benefits, flood, DO, pH, hydrologic regimes, aquatic habitat, sedimentation, water temperatures, macro-invertebrate

Affordable power, job creation, expensive mitigation, maintenance

119

Supriyasilp et al. (2009)

Safety, social conflict, land use, water resource problem, legal obstacle, infrastructure

Flow pattern and amount, habitat loss, land use, river bank collapse, sedimentation, dust and noise

Project cost, IRR, NPV, cost per KWh

Feasibility, construction, slope, alignment, flow, accessibility, installed capacity, development period, transmission

Onat and Bayar (2010)

Public perception, resettlement, human health, employment, agriculture, tourism

Land use, water consumption, CO2 emission, air pollution, water quality

Unit price of energy, capital & operating cost, resource availability

Efficiency

Carrera and Mack (2010)

Innovative ability, shared benefits, health concerns, functional & aesthetic Impact, conflict & catastrophic potential, public participation & perception, traffic

Land use, waste disposal Reserve capacity, flexibility to incorporate other technologies

Kaunda et al. (2012)

Involuntary resettlement, loss of livelihood and cultural identity

Inundation, air & water pollution, biodiversity loss, land use, sedimentation, aquatic weed, CO2 & methane emission, climate benefits

Agriculture, power, mining, tourism

Capik et al. (2012)

Water supply, waterborne disease, flood control, irrigation, navigation, recreation, accessibility, improved living, relocation, job opportunities, tourism, land use

No emission, acid rain, waste, inundation, air quality, erosion, flooding, aquatic life, climate change, hydrologic regime, fish migration, sedimentation, water table

Cheap power, agricultural loss, market fluctuation, construction, O&M cost, efficiency, plant life, safety, employment, development period

Vučijak et al. (2013)

Biological indicators, morphological condition, water quality, terrestrial habitat

Kentel and Alp (2013)

Investment, aesthetic impact, impact on locals, water supply, irrigation, fishing

Dust, air pollution, noise, erosion, landslide, debris, aquatic life, pH, fish passage, sedimentation, diversion, suspended solids, deforestation

Affordable power, reduced dependency on imported energy

Scannapieco et al. (2014)

Public acceptance, employment, traffic

Global warming potential, water, land use, underground resources, waste,

Capital, O&M costs, energy demand

120

biodiversity, ecosystem impact, emission, landscape, hazards

Rosso et al. (2014)

Compensation fees, enterprise activities, marginal area, local employment, stakeholder preferences

Protected areas, hydrological risks, environmental flow, river discharge, water quality, mitigation

Operation & investment costs, incentives, IRR compensation fees, annual benefit, payback period

River length, efficiency, productivity, discharge, intake height, typology, head, structure volume

Maxim (2014)

Job creation, human health, social acceptability, external supply risk

Land use, environmental costs Levelized cost of electricity (LCOE)

Demand response, efficiency, capacity factor

Zhang et al. (2014)

Soil erosion, pollution, fish & human habitat, inundation, land productivity, sedimentation, water quality, ecological alteration, emission

Chen et al. (2015)

Food web impact, sedimentation, discharge, heavy metal pollution

Pang et al. (2015)

Natural flow disruption, land use, aquatic life, biodiversity loss

High initial investment, low maintenance

Morimoto (2013)

Resettlement, loss of agricultural productivity, community cohesion, psychological distress, human health

Submergence, biodiversity loss, aquatic life, endangered and rare species, soil erosion

Average generation cost (capital + operating + resettlement + opportunity) per kWh

Yuksel (2010b)

Flood protection, navigation, recreation, accessibility, living conditions, livelihood, water uses

Reduced GHG emission, air quality, waste, avoided depletion in non-renewables, increased productivity

Life span, reliable service, O&M cost, proven technology, regional development, efficiency, employment

Sparkes (2014)

Resettlement, livelihood, public participation, navigation, irrigation, infrastructure development, water supply, human health

Biodiversity, aquatic life, wildlife, flooding

Impact on local economy

121

7.5 Limitations of the Existing Approaches While the guidelines created by international/financial institutions at times have limited scope or

complex and insufficient manuals that act as a barrier to their adoption, the approaches published

in academic papers also have limitations. The existing literature tends to evaluate hydropower

sustainability from three different perspectives: the triple bottom approach, LCA and from a

systems perspective. Despite their apparent comprehensive nature, these traditional approaches

often miss key challenges faced by hydroelectric development. First, the growth of hydropower is

largely influenced by policies and incentives provided by governing bodies. In the absence of a

carbon tax sufficient to internalize the externalities of carbon-based electricity generations, future

hydropower development will largely be at the mercy of policy initiatives. These policies, typically

imposed for protecting environmental integrity or enhancing local conditions, exhibit a large

spatial variation and can change abruptly depending on particular interests of the power regime.

The situation is usually more complex for hydropower systems located at transboundary rivers.

Typically, water sharing at these plants is guided by some form of international agreement.

However, the jurisdictions, often having different priorities for conflicting water uses, tend to resist

influence or take control over these resources. Sustainable development of remaining hydropower

resources would require favourable local policies as well as international collaboration to avoid

potential conflicts. Second, hydropower potential is sensitive to climate change because of its

dependency on runoff (Kaunda et al. 2012). Several studies have projected the impending changes

in runoff pattern as a result of global warming (increase in some regions while decrease in others)

(Milly et al. 2005; Hamududu and Killingtveit 2012). This changing climate can reduce the usable

capacity of hydropower plants by 61‒74% in the next 25‒50 years (Vliet et al. 2016). The likely

impact of these events may require hydropower system to adjust operation or adapt through new

measures (Martin et al. 2010). Third, the consequences of extreme weather events, such as floods,

ice or hailstorms or droughts, negatively impact generation by effecting water quality, quantity

and damaging plant or transmission infrastructure (Kaunda et al. 2012; Aragon 2010; Elakanda

2010; Gondwe 2010; Nair 2010). Risk of flooding and sedimentation is also likely to increase with

changes in local hydrology as a result of climate related extreme weather events (World Bank

2007). Despite their functionality, current approaches rarely address these long-term

environmental (climate) challenges within their frameworks that may substantially affect future

122

hydropower generation potential. Of course, the challenges of a less predictable future are real and

formidable.

From a systems perspective, the majority of the research have set boundaries that include project

location and extend into the surrounding areas. Due to such restricted considerations, traditional

life-cycle assessment omits some of the relevant factors and neglects some key benefits offered by

hydropower. First, the value of irrigation, flood control, avoided emission from fossil fuel

generation, and recreation provided by multipurpose reservoirs are sometimes ignored. Second,

traditional system boundaries exclude the power grid which in turn disregards the role of pumped

or hydro storage in system stabilization through ancillary services. At present, particularly with

storage using batteries still awaiting major technological breakthrough, hydropower is really the

one renewable source that offers an effective means of permitting demand variability. Moreover,

reservoir-based hydropower is rarely credited in the existing literature for leveraging investment

in intermittent renewables (Jaramillo et al. 2004 and 2010; Matevosyan et al. 2009). Despite being

addressed by IEA (2000), indicators that reflect on the stability, flexibility and resilience offered

by hydroelectric resources are rarely incorporated in the present matrices. A more detailed and

complete analysis will require extending the system boundaries to include the grid, thus allowing

inclusion of previously omitted parameters. Also, this line of research may also benefit from

application of novel methods and approaches that integrate the disparate and currently

unconnected aspects/dimensions of hydropower, environment and policy.

7.6 Conclusion Hydroelectric projects, when built in the right places and following proper guidelines and with

adequate mitigation measures, can bring multiple benefits to the community. Apart from providing

electricity access, hydropower serves twofold purposes in climate change mitigation - as a

renewable resource producing power at minimal GHG emission and as a backup facility to move

the electrical grid to a low-carbon future. However, in the absence of a unified consensus or with

poor safeguards and project execution, developing this potential resource in a sustainable way

offers both considerable challenge and tangible risk. While environmental and social issues

associated with hydropower development are often right cited with criticisms, here the core of the

debate is not whether the environment is impacted, as much as to what is the degree of negative

123

impacts that can be allowed while being sustainable. All power sources are problematic in a variety

of ways, and yet the use of power brings dramatic and measurable benefits to human communities.

The requirement is to determine holistic and defendable measures that evaluate pros and cons and

that are open to external scrutiny and debate. Power and controversy are inevitable linked in human

systems.

In a world with growing electricity demand, intermittent renewable sources pose a major

predicament in terms of its impact on grid stability. The key balancing question, i.e., how much

wind and/or solar can be incorporated without compromising flexibility – requires rigorous

consideration of available dispatchable sources of coal, natural gas and hydropower. Now, the use

of hydrocarbon-based fuel entails considerable financial risks with the growing adoption of carbon

pricing mechanism. In a carbon constrained economy, hydroelectric projects can potentially be

financed through carbon offsetting which may balance its typically high installation cost. This

research seeks to summarize the existing state of play relating to the how sustainability is assessed

for hydropower. The indicators used for these analyses are aggregated and summarized. The

discussion briefly documents research methods and points out some limitations in the existing

approaches. Thus, this study attempts to advance the current sustainability assessment. Two

recommendations are also put forward: first, that such assessments should reflect on major

environmental challenges and broad policy issues with respect to existing hydropower potential

and, second, that system boundaries should be extended to allow reasonable estimation of hydro

benefits on the overall grid.

124

Opportunities for Increased Hydropower Diversion at Niagara: An sSWOT Analysis

While the forgoing chapter analyzes the literature on sustainable hydropower development, the

concerns raised in the review process are addressed in chapter 9 that elaborates an existing tool to

explicitly incorporate sustainability. The discussion here proposes the improved decision-support

framework under the Sustainability SWOT (sSWOT) model that can be used for analyzing

resource systems and shows how the new framework applies it to the strategic planning for

increased hydropower generation at Niagara. The analysis sheds light on the current economic,

environmental, social and political dynamics, presenting and analyzing a holistic perspective of

various stakeholders. The Analytical Hierarchy Process (AHP) and Analytical Network Process

(ANP) are both used to identify a priority sequence among potential decision alternatives. The

analysis shows that renegotiation of the 1950 Treaty is a preferred option over the current flow

restrictions.

This chapter is based on the paper entitled “Opportunities for Increased Hydropower Diversion at

Niagara: An sSWOT Analysis” by Samiha Tahseen and Bryan Karney, published (18.09.2016) in

the Journal of Renewable Energy. The proposed framework with features such as easy

interpretation and ranking can be reasonably applied to any decision making problem with simple

modifications in its structure.

8.1 Introduction The Niagara River, an integral part of the Great Lake Basin, not only transports vast quantities of

water but hosts a world-renowned waterfall, a spectacle that itself attracts 12-14 million tourists

each year. The river water, diverted according to the 1950 Niagara River Water Diversion Treaty,

renewably powers generation facilities on both sides of the Canadian-United States border.

However, balancing the competing demands between recreational, commercial, and industrial uses

within this river system has proven to be a challenge since at least the nineteenth-century.

Integrated Water Resource Management (IWRM) is defined as “a process which promotes the

coordinated development and management of water, land and related resources in order to

125

maximize the resultant economic and social welfare in an equitable manner without compromising

the sustainability of vital ecosystems” (GWP 2012). IWRM moves from broad policy goals to

selecting, implementing, and subsequently evaluating and revising suitable strategies. Resource

management of transboundary water faces a specific set of challenges as conflicts on water

resource allocation and benefit sharing are inevitably complex. Due to this nature, strategic action

plans involving transboundary water should follow the simplest approaches, drawing from

engineering, economics, ecosystem and social studies, but with the overall goal of being easily

understandable to policy makers.

The existing literature references a number of decision support tools. The so-called SWOT

approach (for Strengths, Weaknesses, Opportunities and Threats) originated from business

literature (Kotler 1988). The DPSIR approach (Driving forces, Pressures, States, Impacts, and

Responses) is a causal framework for describing interactions between society and environment

(Canu et al. 2011; Ma et al. 2012). A closely-related but sophisticated variant called PESTLE

(Political, Economic, Social, Technological, and Environmental) provides a multidimensional

purview of the whole environment to track proposal-specific considerations (Collen et al. 2014).

Another document (Skondras and Karavitis 2015) introduced the CSDA, a combination of SWOT

and DPSIR, for issues vexed with environmental and climatic uncertainties or economic

instabilities. While the DPSIR framework lacks economic considerations – a major concern in

resource management – the PESTLE focuses mainly on the external environment (Makos 2011).

This paper uses the SWOT for its simplicity and ability to analyze both internal and external

environmental factors, relatively easier but effective methodology, graphic representation and easy

interpretation. The SWOT has been successfully applied in resource planning (Baycheva-Merger

and Wolfslehner 2016; Chen et al. 2014; Jaber et al. 2015), ecosystem management (Bull et al.

2016; Grošelj and Stirn 2015; Viegas et al. 2014) and strategy prioritization at the industrial and

policy level (Michailidis et al. 2015; Rauch et al. 2015; Shahabi et al. 2014). Management of

transboundary water system such as Niagara requires participation from a vast array of

stakeholders. The nature of the analysis necessitates addressing the conflicting objectives of

consumptive use, navigation, hydropower, tourism, erosion etc. The SWOT’s ability to engage

and involve stakeholders can be particularly beneficial for analyzing these systems.

126

Appropriate strategic action requires systems to be able to adapt and survive in a changeable

environment. Since the conventional SWOT often fails to comprehensively appraise the decision

situation, Kurttila (2000) developed a hybrid method using the Analytical Hierarchy Process

(AHP) to accommodate the assessment of alternatives. However, recent studies (Stewart et al.

2002; Kajanus et al. 2004; Shrestha et al. 2004; Leskinen et al. 2006; Masozera et al. 2006) with

the AHP application do not account for the possible dependencies among the underlying factors –

a limitation which application of the Analytical Network Process (ANP) addresses.

Sustainable development, sometimes still referred in the classic way as “meeting the needs of the

present without compromising the ability of future generations to meet their own needs” (WCED

1987), has become an increasingly important consideration in resource management. Such

considerations are motivated by obvious evidence of ecosystem degradation, natural resource

depletion and global climate change. At present, we still need to find ways to incorporate

sustainability into long-term resource planning in order to mitigate adverse impacts and to promote

resilience. Here, the authors develop a framework based on the concept of a “sustainability SWOT”

(or sSWOT) (Metzger et al. 2012) which provides a specific sustainability dimension to the

familiar SWOT considerations. The sSWOT analysis connects long-term environmental and social

challenges with economic priorities and can communicate new policy insights. It is designed to

drive action and collaboration on environmental challenges, creating risks and opportunities which

otherwise may go unnoticed. The paper applies the AHP and ANP within the sSWOT framework

for the purpose of assessing the potential for increased hydropower diversion at Niagara, which is

an option that opens up with the expiration of the bilateral treaty between the two neighboring

countries, the US and Canada. This exploratory study attempts to integrate the disparate and

currently unconnected aspects/dimensions of water, energy, tourism and policy to promote

sustainable development of the incredible resource system at Niagara.

8.2 Model Development A nine-step sequential evaluation process is used to analyze resource systems from sustainability

perspective using the sSWOT (Figure 8.1). First, specific objectives for resource planning are set,

which in this case is to increase the hydropower potential at Niagara. Unlike the traditional SWOT

which initiates with internal factors (strengths and weaknesses), the sSWOT begins with

127

synthesizing information on the future environmental challenges and the changing policy

landscape (step 2). Next, the analysis recognizes the threats and opportunities with respect to the

observed challenges. Throughout the process, the benefits and potential risks changes may pose

are considered. After identifying the external factors (opportunities and threats), the strengths and

weaknesses of the existing system are articulated (step 4). The elements under each SWOT factor,

called the SWOT sub-factors, are categorized into economic, environmental and social parameters

in step 5. On the basis of the SWOT sub-factors, alternative strategies are proposed (step 6). Next,

stakeholders’ surveys are conducted to weigh relative importance among sub-factors and

alternatives. In the final stages, survey data are analyzed with the application of AHP and ANP to

obtain ranking of alternatives.

Figure 8.1: Step-by-step evaluation process for the sSWOT model

128

The application of AHP within the SWOT framework allows quantitative evaluation of the

identified factors (Saaty 1988 and 1990) and it has been widely applied in natural resource

planning, industrial and corporate strategy assessment (Chatzimouratidis and Pilavachi 2008;

Ghodsypour and O'Brien 1988; Wind 1987). However, a basic simplifying assumption limits the

possibility of incorporating interdependencies among the factors. Many decision-making

sequences simply cannot be accurately structured hierarchically because they involve interaction

among various factors (Saaty 1996; Saaty and Takizawa 1986). The application of Analytical

Network Process (ANP), introduced by (Saaty 1996), allows an assessment of the relative

importance of interdependencies (Saaty 2004). The method is widely used by academic

community and consulting industry (Ergu et al. 2014; Köne and Büke 2007; Shahabi et al. 2014).

Figure 8.2 summarizes the hierarchy and the network representation of the SWOT model.

(a) (b)

(a) (b)

The algorithm for ANP application within the sSWOT is composed of the following steps:

Step 1: The problem is first decomposed into a rational network system.

SWOT

SWOT sub-factors

SWOT

Alternatives

SWOT sub-factors

Goal Goal

Alternatives

Figure 8.2: The hierarchy (a) and the network (b) representation of the sSWOT model. While

(b) allows interdependencies among SWOT factors, (a) permits downward influence only

(Yüksel and Dagˇdeviren 2007)

129

Step 2: The SWOT factors (Strength-Weakness-Opportunity-Threat) are pairwise compared with

respect to the control criteria assuming no dependence among the factors. Experts respond to

questions to extract the relative importance of strength over weakness with respect to the planning

objective, so that the SWOT elements can be evaluated with respect to their contribution to the

upper level criteria. The relative importance is determined using Saaty’s 1-9 scale (Table 8.1)

where a score of 1 represents equal importance, while 9 stands for extreme importance of one

element (row in the matrix) over the other (element in column). This scale is developed on the

basis that positive integers are intrinsic to human’s ability to make comparisons (Saaty 2008). For

inverse comparison, a reciprocal value is assigned, i.e., aij = 1/ aij where aij denotes the importance

of the ith element.

Table 8.1: Saaty’s 1-9 scale for Analytical Hierarchical Process (AHP) preference (Saaty 1996)

Intensity of

importance Definition Explanation

1 Equal importance Two activities contribute equally to the objective

3 Moderate importance Experience slightly favor one over another

5 Strong importance Experience strongly favor one over another

7 Very strong importance Activity strongly favored and its dominance

demonstrated in practice

9 Absolute importance Importance of one over another affirmed on

highest possible order

2,4,6,8 Intermediate values Used to represent compromise between priorities

listed above

Now, these values are arranged in a matrix framework which is then used to derive a local priority

vector (w1). If A is a non-negative, primitive matrix, then one of its eigenvalues λmax is positive and

greater than or equal to (in absolute value) all other eigenvalues, and there is a positive eigenvector

w corresponding to that eigenvalue, and that eigenvalue is a simple root of the following

characteristic equation (Alonso and Lamata 2006):

A × w = λmax × w (8.1)

130

where A is the matrix, w is the eigenvector and λmax is the largest eigenvalue of A. (There are several

well-known methods for calculating or approximating the eigenvectors). The obtained roots are

summed and normalized to obtain the final eigenvector elements. After determining the

importance degrees of the SWOT factors, the procedure involves calculating Consistency Index

(CI) and Consistency Ratio (CR) using the following formula:

CI = (λmax – n)/(n-1) (8.2)

CR = CI/RI (8.3)

where RI is the Random Consistency Index, which can be directly obtained from Table 8.2 and n

is the order of the matrix. If CR ≤ 0.1, the calculation of relative importance criteria is considered

acceptable. Otherwise, the process has to be repeated due to inconsistency (Zoran et al. 1980).

Table 8.2: Random Consistency Index value (Saaty 1980)

n 1 2 3 4 5 6 7 8 9 100

RI 0 0 0.52 0.89 1.11 1.25 1.35 1.40 1.45 1.49

Step 3: The inner dependence matrix among the SWOT factors is calculated using the same 1-9

scale. Queries such as, “What is the relative importance of strengths when compared with threats

on controlling weaknesses?” may arise when determining inner dependencies. The resulting

priority vectors following the step described above are called w2. The interdependent priorities of

the SWOT factors (i.e., wfactors) are calculated by matrix multiplication between w1 and w2.

Step 4: The SWOT sub-factors are compared pairwise with respect to their contribution to the

objective following the same procedure described in step 2. The SWOT sub-factors under

Strengths, Weaknesses, Opportunities and Threats are further categorized according to the

sustainability parameters, i.e., economic, environment and social. To limit the size of the matrix,

the elements under each of these parameters are compared pairwise in a matrix framework, and

then multiplied by the relative importance of the sustainability parameters. The resulting local

importance degrees of the SWOT sub-factors are denoted as wsub-factors (local).

131

Step 5: The local importance values are next multiplied with wfactors to convert them into global

importance degrees of the SWOT sub-factors, i.e., wsub-factors (global) = wfactors × wsub-factors (local).

Step 6: The importance degrees of the alternative strategies with respect to each SWOT sub-factor

using Saaty’s 1-9 scale are obtained following the same procedure stated above. The resulting

matrix is called w4.

Step 7: Finally, the overall priorities of the alternative strategies are determined by the relationship,

walternative = w4 × wsub-factors (global).

Step 8: The Supermatrix is then a partitioned matrix where each matrix segment represents a

relationship between two clusters. With a four level network, i.e., goal, SWOT factors, SWOT

sub-factors and alternatives, the supermatrix representation should be as follows:

𝑊𝑊 =

goalSWOT factors

SWOT sub − factorsalternatives

�

0 0 0 0𝑤𝑤1 𝑤𝑤2 0 00 𝑤𝑤3 0 00 0 𝑤𝑤4 I

� (8.4)

where w1 is a vector that represents the impact of the goal on the SWOT factors, w2 matrix signifies

the inner dependence of the SWOT factors, w3 denotes the impact of the SWOT factors on the

SWOT sub-factors and w4 characterizes the impact of each SWOT sub-factor on the alternative

strategies.

8.3 Application of the sSWOT Model on Niagara To apply the sSWOT model to Niagara, two steps are taken. First, the major environmental

challenges for the resource system are considered while sketching out the internal (strengths and

weaknesses) and external (opportunities and threats) factors. Following this, the AHP and ANP

are used to identify and evaluate potential decision options.

8.3.1 Identification of the sSWOT factors The analysis of Niagara as a resource system begins with identifying the key environmental and

policy trends. The existing climate models for the Laurentian Great Lakes project an increase in

132

mean air temperature ranging from 2 ‒ 5˚C in summer and 4 ‒ 8˚C in winter (Magnuson et al.

1997). Gula and Peltier (2012) found that the temperatures in the watershed are expected to

increase by 2‒3°C by 2050 under IPCC SRES A2 emission scenario which assumes a continuous

increase in greenhouse gas (GHG) concentration. During the same period, precipitation is expected

to increase by 3 - 10% (Gula and Peltier 2012). This changing climate is likely to impact the overall

system (ecology, tourism, navigation, power systems etc.) at Niagara (Kling et al. 2003; Shlozberg

and Dorling 2014), and is therefore considered as potential threat. Meanwhile, Canada’s energy

policy landscape is quickly shifting in response to this changing climate and growing concern over

energy security. First, Ontario’s Feed-in Tariff (FIT) Program has induced growth in intermittent

renewables (Wong et al. 2015; Yatchew and Baziliauskas 2011), a reality that pose enormous

challenges in dealing with power fluctuations. Second, the recent phaseout of coal-fired electricity

and a proposed reduction in natural gas use by 2017 (Ontario Ministry of Energy 2009) limit the

options for dispatchable generation in Ontario. Third, Ontario’s major nuclear generators need to

be refurbished (Ontario Ministry of Energy 2013a) or replaced. Fourth, though Ontario’s

electricity sector can be credited with a 58% reduction in GHG emission since 2005, it is still

responsible for 14.5 Mt CO2e emission annually (Ontario Ministry of Environment and Climate

Change 2014). Particularly with the information that the current initiatives to reduce GHG

emissions will deliver only 60% reductions needed to reach the 2020 target (15% below 1990

levels) (Ontario Ministry of Environment and Climate Change 2014), the need for green and

reliable sources peaks. Finally, Ontario may perhaps adopt a carbon pricing mechanism as part of

their commitment to Western Climate Initiative (WCI). Inspired by the carbon market in British

Columbia, the province is progressively moving towards a similar direction. With such policy in

place, reliance on fossil fuel may result in rising electricity prices, which is unfavorable to the local

business.

After identifying the major environmental and policy challenges, the selection of factors under

strength, weakness, opportunity and threat are guided by several field trips as well as review of the

existing literature, published documents, articles, stakeholders’ reports, online resources etc.

Experts’ opinion was sought (via emails or face-to-face discussions) at several stages of this

selection process. Based on the suggestions, the original list was narrowed down to the most

important and relevant factors. The procedure is repeated until a general consensus has been

133

reached. Figure 8.3 illustrates the sSWOT model for Niagara. A brief description of the itemized

SWOT sub-factors are provided below:

8.3.1.1 Opportunities

8.3.1.1.1 Economic Expiration of the 1950 Niagara River Treaty

The Niagara River carries on average about 5,660 m3/s (Lee at al. 1988). The present day flow

control strictly adheres to the 1950 Treaty to preserve the scenic spectacle of Niagara, and

consequently limiting the flow for hydropower generation. The treaty establishes that no less than

2,832 m3/s (100,000 ft3/s) must go over the falls between 8:00 AM and 10:00 PM from April 1 to

September 15. The same flow restrictions are in effect between 8:00 AM and 8:00 PM from

September 16 to October 31. At all other times, a minimum of 1,416 m3/s (50,000 ft3/s) must be

maintained over the falls (Government of Canada 2015). Now, the treaty has expired in 2000 and

is currently being extended year by year. The expiration opens the door for renegotiation with the

key question being, of course, the possible risks and rewards of such a renewed negotiation.

Potential for generation with third Niagara tunnel

The situation at Niagara is involving in many ways, not least of which through engineered actions

and decisions. The new third Niagara tunnel, inaugurated in 2013, allows an additional 500 m3/s

of water to be used on the Canadian site at the Sir Adam Beck (SAB) complex (Ontario Power

Generation 2015b). However, interestingly, the current flow restriction, imposed by the treaty,

would restrict this tunnel to be utilized to its full potential (Sedoff et al. 2014). Further exploitation

of this extended intake capacity would require greater diversion from the treaty.

Demand mitigation in the absence of nuclear power

Two major nuclear facilities, providing significant portion of Ontario’s baseload will be offline by

2021 (Boland 2013). Meanwhile, the power consumption is likely to increase with Ontario’s vision

of increased electric vehicle usage by 2020 (Canadian Electricity Association 2013). With the

retirement of the nuclear plants and greater infiltration of wind and solar resources, the province

can greatly benefit from the increased capacity at the SAB hydropower facilities.

134

Profit opportunities during summer

The average summer temperature in Ontario is projected to increase by 3.8°C, while the maximum

extreme humidex, an index number combining the effect of heat and humidity, is estimated to

increase from 48°C (in 2000-2009) to 57°C in 2040-2049 (SENES Consultants Ltd 2011). This

changing context is likely to increase peak power demand, particularly during the summer ‒ a

condition that would benefit from an extension of dispatchable hydro capacity at Niagara.

Flow alteration as a tourist attraction

Niagara has experienced a gradual decline in the same-day visitors (Ontario Ministry of Tourism

2008; Niagara Falls Review 2011). Though the factors responsible for this reduction is yet

unknown, a changing flow conditions – one that involves enhancing Niagara’s appeal while

protecting its present beauty and ecological balance – may possibly be interest to visitors.

8.3.1.1.2 Environmental Erosion control through flow diversion

The flows in the Niagara River erode the falls. Despite rehabilitation and flow control, the

approaching water causes the escarpment to retreat about 0.3 m per year (Niagara Parks

Commission 2015). Though it contradicts with the general conception about the untainted beauty

of Niagara, remedial works was an important step towards protecting the spectacle of the falls.

Additional flow diversion, apart from extending Niagara’s hydropower potential, may reduce its

current erosion rate.

Improving the excessive misting condition

While the falls’ beauty is greatly enhanced by the formation of some mist, there have been reports

of excessive misting obstructing the view of the falls (Niagara Parks Commission 2004). These

over-misting events are on the rise, and a cause of alarm for the growing tourism industry at

Niagara. Considering that there is a positive correlation between the mist plume height and the

Niagara River flow rate (Case 2004), these excessive misting conditions may improve with

additional power diversion.

135

Reduced emission from power generation

Extending Niagara’s hydropower capacity, if plausible, will contribute towards offsetting

emissions from the power sector. The compromised hydropower potential, worth 1.6 million

MWh, if generated through alternative sources such as coal, leads to an emission of 3.36 million

metric tonnes of CO2 (Sedoff et al. 2014). The expiration of the treaty presents the opportunity to

bring into consideration the previously neglected issues of climate change and GHG emission.

8.3.1.1.3 Social Employment in energy and tourism industry

Additional diversion that augments Niagara’s hydropower capacity can lower electricity prices in

the domestic market or bring revenues by exporting to the neighboring jurisdictions. In addition,

the variation of flow over the falls can add a different flavour to the familiar splendor of the falls,

which may attract tourists leading to increased revenue generation.

Policy debate revisiting the treaty

The current treaty ensures a minimum flow of 2,832 m3/s over the falls during the tourist season,

and a reduced flow of 1,416 m3/s at all other times. A research by Friesen and Day (1977) claimed

that the upper limit of 2,832 m3/s is not the absolute minimum to achieve the scenic spectacle of

Niagara, and further diversions might be possible without adversely affecting the falls. These

varied opinions may lead to a potential public debate involving various stakeholders.

8.3.1.2 Threats

8.3.1.2.1 Economic Reduced power potential under the climate change

The vulnerability of water resource systems to the changing climate raises a new threat in securing

supply and equitable allocation of water. While a rising temperature and precipitation variability

in the Great Lakes region may cause local flooding, the shortage of water supply is a greater risk

long-term to hydroelectric generation (International Joint Commission 2012). The annual costs of

replacing this capacity with nuclear or fossil fuel plants are estimated to be US $160 million in

1988 for New York (Crissman 1989), and CAD $1 billion for Ontario (Melo 1989).

136

Drawing from the broader environmental and policy challenges, strength, weakness, opportunity and threat for the Niagara region are identified.

Environmental challenges and big trends

Increasing air temperature due to climate change Green development in Ontario Phasing out coal and limiting gas generation Refurbishment of the existing nuclear power stations Rapid growth of renewables under FIT

Threats Economic Reduced power potential under climate change (T1) Long payback period for third Niagara tunnel (T2) Unfavourable outcome from renegotiation (T3) Declining tourists (T4) Cost associated with the renegotiation (T5) Environmental Increase in excessive misting (T6) Erosion of the Niagara escarpment (T7) Social Stringent travelling requirements (T8) Decreasing the appeal of Niagara Falls (T9)

Strengths Economic Installed capacity with no fuel dependency (S1) Pumped storage benefits (S2) Protection of the installed equipment (S3) Revenue from tourism (S4) Environmental Renewable raw material lowering pollution (S5) Regulating water level (S6) Social Employment opportunities (S7) Improved health conditions (S8)

Weaknesses Economic High unit energy cost at Beck PGS (W1) High investment cost (W2) Turbine refurbishment (W3) Environmental Disrupting the natural environment (W4) Methane emission from flooded biomass (W5) Social Resettlement (W6) Restrict navigation (W7)

Opportunities

Economic Expiration of the 1950 Treaty (O1) Potential for generation with third tunnel (O2) Demand mitigation in absence of nuclear plant (O3) Profit opportunities during summer (O4) Flow alteration as a tourist attraction (O5) Environmental Erosion control through flow diversion (O6) Improving the excessive misting conditions (O7) Reduced emission from power generation (O8) Social Employment in energy and tourism industry (O9) Policy debate revisiting the treaty (O10)

Figure 8.3: The sSWOT model for Niagara

137

Long payback period for the third Niagara tunnel

The third Niagara tunnel increases the diversion capacity (to the power stations) at a substantial

cost of CAD $1.6 billion (OPG 2015). The treaty restriction, currently limiting the potential use

of this tunnel (Sedoff et al. 2014), may prolong its cost recovery period.

Unfavorable outcome from treaty renegotiation

Another obvious threat is that any revision attempt on the age old treaty bears the risk of losing

Canada’s strategic hold on water to the US. Perhaps, both countries have yet to quantify the

maximum potential economic value, and so cannot use it to negotiate a fair deal.

Declining tourists

The 12-14 million annual visitors to the falls is an important source of revenue for the city of

Niagara. The economic crisis combined with market volatility have the potential to adversely

impact this tourism sector (UNWTO World Tourism Barometer 2009). Ritchie et al. (2010)

reported a 7% decline in overnight trips in Canada due to the recession in 2009. Repetition of such

events, excessive misting or some combination of changing taste and the lack of novelty in

Niagara’s scenic beauty can be potential threats to the existing tourism industry.

Cost associated with a renegotiation

Attempt to renegotiate the treaty should be backed by sound research that addresses how and to

what extent additional flow diversion will impact the beauty of the falls, people’s perception of it

and the probable effect of diversion on the tourism sector considering the impact on shoreline,

aquatic ecosystem and the importance of low power prices in the region etc. All this may require

millions of dollars in research and development.

8.3.1.2.2 Environmental Increase in excessive misting at Niagara

The Niagara Parks Commission (NPC) has reported an increase in excessive misting events at the

falls - 68 in 2003 compared to 29 in 1996 (The NY Times 2006). The temperature difference

between the river water and the surrounding air, considered as one of the reasons for excessive

misting (Case 2004), is likely to increase under a warming climate.

138

Erosion of the Niagara escarpment

The continuous erosion in the last 560 years has led to the recession of the Horseshoe fall by 1 -

1.5 m per year (Niagara Parks Commission 2015). Though the rate has been reduced by the flow

control and the remedial works, the current erosion continues at a rate of 0.3 m per year.

8.3.1.2.3 Social Stringent travelling requirements discouraging tourists

Invasive forms of security procedures are associated with higher perceived threat to personal

dignity, thereby invalidating the purpose of these measures on a personal level (Alards-Tomalin et

al. 2014). For example, enhanced security measures in the US post-9/11 resulted in a sharp decline

in short-term visitors (Bonham eta al. 2006).

Decreasing the appeal of Niagara Falls

The advent of the electronic age comes with the tradeoff that natural beauty is perhaps

uninteresting to some, or worse, seen as unneeded by others. There may be a decreasing appeal for

Niagara’s natural beauty compounded by the increased security in the area.

8.3.1.3 Strengths

8.3.1.3.1 Economic Installed capacity with no fuel dependency

Presently the Niagara River provides the driving force for about 3,000 MW generation capacity at

the SAB Complex (OPG 2015). Along with two conventional power stations, the facility hosts a

pumped storage known as the Beck PGS. These plants have relatively low operation and

maintenance cost. Unlike coal or gas generation, their operating costs are not susceptible to fuel

price increase.

Pumped storage benefits

The Beck PGS absorbs surplus energy by pumping during low-cost off-peak hours. It then

generates electricity when the demand is high, in the process delaying some more expensive

generators to be online. With its operating reserve and black-start capability, the facility provides

the much required flexibility and reliability to the grid. With emerging storage options (such as

139

batteries, capacitors etc.) still at their technological infancy, pumped storage is the only renewable

that provides grid-scale storage.

Protection of the installed equipment

Apart from its contribution to the grid, the Beck PGS controls water level at the power canal in

order to ensure the appropriate diversion (Maricic et al 2009). The action further protects the

runners from cavitation and associated depreciation.

Revenue from tourism

Tourism is responsible for 57% annual occupancy at the Niagara Falls accommodation market

(City of Niagara Falls 2014) and generates 11% Hotel Room Night Occupancy (RNO) in Ontario

(Ontario Ministry of Tourism 2009). Tourism is a dynamic market capable of positively

influencing the economy in many ways such as creating new opportunities for existing businesses,

feeding behind-the-scene support services and supply industries, fueling construction of

commercial, residential, and infrastructure projects and the like.

8.3.1.3.2 Environmental Renewable raw material lowering pollution

Acknowledging the imminent threat posed by global climate change, the present world is

conscious about keeping its carbon emission in check. Nonetheless, most of our traditional power

sources are associated with considerable CO2 emissions. Given this, Niagara’s hydropower

continues to be a key generation asset while limiting emission from the power sector.

Water level regulation

The Beck PGS assists in maintaining water elevation at the canal, which is critical for ensuring

appropriate diversion from the river. This ensures that water levels stay within an allowable range,

while maintaining a minimum flow over the falls.

8.3.1.3.3 Social Social welfare by creating employment opportunities

Niagara’s thriving tourism and power industries continue to bolster the local economy by creating

numerous jobs.

140

Expected improvement in health condition

Researchers from MIT’s Laboratory for Aviation and the Environment found in a study that air

pollution causes about 200,000 early deaths in the US each year (MIT News 2013). When

simulated for pollution by sector, electricity generation accounted for 52,000 premature deaths

annually. Being a carbon free resource, Niagara hydropower system has long contributed towards

offsetting emission from the power sector, thereby reducing adverse health impact.

8.3.1.4 Weaknesses

8.3.1.4.1 Economic High unit energy cost at Beck PGS

Despite the perceived technical demand, profitability remains a major obstacle for pumped storage

operations (Ingebretsen and Johansen 2014; Tahseen and Karney 2016: Chapter 3). The Beck PGS

faces similar challenges with the unit energy cost increasing by 72% between 2006 and 2008 (OPG

2010). Some of the factors contributing to this high energy price are lower efficiency, relatively

small head and plant size, expensive refurbishments etc.

High investment cost

Compared to other renewable options, hydroelectric developments are associated with higher

initial cost. The installation cost for modern wind turbine varies between $1,500 -$3,000 per kW

(IRENA 2015), while the same for utility-scale solar PV ranges from $2,450 – $6,260 per kW

(Feldman et al. 2012). In contrast, large hydropower plants (>10 MW) with an investment cost of

$1,750 - $6,250 per kW (Lako et al. 2010) makes it one of the most expensive renewable options.

Turbine refurbishment

Refurbishment is an important way of boosting hydropower output from aging plants. Though,

refurbishment may cost as low as one-third of the cost of new development, it can still lead to a

substantial amount depending on the project. For example, Lewiston pump-generating plant at

Niagara is currently undergoing renovation with a substantial cost of $460 million (NYPA 2015).

141

8.3.1.4.2 Environmental Disrupting the natural environment

Hydropower developments may be associated with inundation, loss of agricultural land, damage

to the natural environment, disturb ecosystem and fish population, etc. Further exploration of

hydropower potential at Niagara should limit such negative impacts.

Methane emission from flooded biomass

Flooded biomass at hydro reservoirs can be a potential source of GHG emission during its

operational phase. The emission from a typical plant ranges from 2–15 kt CO2 eq./ TWh

throughout its life cycle (Gagnon et al. 2002).

8.3.1.4.3 Social Resettlement due to inundation

Relocating the local population is one of the most challenging aspect of hydropower development.

Since Niagara’s hydropower potential had mostly been developed, the incremental increase in

capacity or diversion such as the options analyzed here, has a remote possibility to cause such

disruption.

Restrict navigation

Unplanned hydropower development has led to severe drought, dried lakes and rivers in many

parts of the world. Further development of Niagara’s resources must maintain the flow required

for recreation and navigation.

8.3.1.5 Alternatives Current flow restriction

This alternative involves continuing with the current flow restrictions.

Renegotiation for greater flow diversion

A number of developments following the treaty ratification command for a renewed consideration

of its rationale. The 1950 Treaty identifies the unbroken crestline as the most significant feature

for achieving the “impression of volume” necessary for the scenic spectacle of Niagara (Friesen

and Day 1977). The remedial works following the treaty ensure this feature at 1,416 m3/s. The

142

additional 1,416 m3/s over the falls during the tourist season (making it a total of 2,832 m3/s)

represents 1.6 million MWh capacity for Ontario which translates into an annual cost of $52

million CAD. Moreover, foregoing this generation could result in 3.36 Mt CO2 emission when

using other carbon-based fuel (Sedoff et al. 2014).

Renegotiation considerations, though controversial, are indeed important with respect to Ontario’s

current policy perspective. The first two years of FIT program has brought in about 1,500 MW of

wind capacity (Yatchew and Baziliauskas 2011) and is projected to provide about 10% of the

supply by 2030 (Marshall 2013). With increased participation from intermittent renewables and

policies discouraging the use of carbon-based fuel, the need for non-emitting, dispatchable

generation peaks. Hydropower, apart from being clean and renewable, offers an effective means

of permitting demand variability. The generation capacity also remains relatively unaffected

(unless there is a large variation in seasonal flows). Hydro development at Niagara, therefore, may

result in a more resilient system capable of absorbing shocks and accidents. The treaty rationale

also needs to be examined in light of OPG’s recent $1.6 billion investment into the diversion

tunnel. Renegotiating the treaty may permit additional hydropower generation along with reducing

both the erosion rate and heavy misting at Niagara without compromising the beauty of the falls.

8.3.2 Application of the AHP and ANP Fifteen (15) experts in the field of energy and environment participated in a questionnaire. The

focus group members were not meant to be statistically representative, but were selected based on

their expertise and knowledge about the issues at hand. Most participants have attended at least a

presentation or discussion session, while the rest have studied the power systems. Two (2)

specialists in climate change and sustainable development were interviewed (making a total of 17

respondents) to reinforce the sustainability considerations. These participants made tradeoffs

among the identified SWOT factors and sub-factors by determining which is more important and

by how much on a scale from 1– 9. Figure 8.4 demonstrates an example pairwise comparison under

the “Opportunity” category. The survey responses were then aggregated and analyzed through

sSWOT-ANP model. The Appendices present the responses in a series of pairwise comparison

matrices. Here, each cell is in a form: xi, …., xj:m,…; (x̄, σ) where xi stands for value of the

responses chosen by only one respondent and xj:m represents response xj chosen by m

respondents. Also, x̄ and σ represent the average and standard deviation of the response values for

143

each pairwise comparison. To illustrate with an example, the comparison between ‘expiration of

the treaty (O1)’and ‘potential generation with the third Niagara Tunnel (O2)’ are presented as 7,

0.5, 0.17, 0.14, 0.1, 1:2, 3:2, 0.33:2, 0.25:2, 0.2:4; (1.1,1.8). It suggests that values 7, 0.5, 0.17,

0.14 and 0.1 are chosen only once; values 1, 3 and 0.33 are chosen by two whereas 0.2 is chosen

by four respondents. Also, the average and standard deviation of all responses for this pairwise

comparison are 1.1 and 1.8 respectively.

Treaty expiration

9 8 7 6 5 4 3 2 1 2 3 4 5 6 7 8 9 Potential generation with third tunnel

9 8 7 6 5 4 3 2 1 2 3 4 5 6 7 8 9 Replacement for nuclear plants

9 8 7 6 5 4 3 2 1 2 3 4 5 6 7 8 9 Profit opportunities during summer

9 8 7 6 5 4 3 2 1 2 3 4 5 6 7 8 9 Flow alteration attracting tourists

Figure 8.4: An example of a pairwise comparison of factors presented under the SWOT category

“Opportunity”. The respondent is asked to assign a value from 1 to 9 to one of the factors to

indicate the relative importance of that factor over another.

Following the articulated steps, the application of ANP begins with converting the identified

elements into a network in which the aim of choosing the best strategy is placed as the goal. Based

on the survey data, pairwise comparison matrix for the SWOT factors (strength – weakness –

opportunity – threat) is formulated assuming no dependencies. The comparison results are shown

in Table 8.3. The matrix is analyzed and the following eigenvector (w1) is obtained as described

in step 2.

𝑤𝑤1 = �

SWOT

� = �

0.480.160.220.14

�

144

Table 8.3: Pairwise comparison of SWOT factors by assuming there is no dependence

SWOT factors S W O T

Strengths (S) 1 3.3 2.9 2.5

Weaknesses (W) 1 1.1 0.9

Opportunities (O) 1 2.8

Threats (T) 1

CR = 0.07

In step 3, the inner dependence among the SWOT factors is determined assuming strength to be

dependent on weakness, opportunity and threat. While weakness relies on strength and threat, the

latter influences both strength and weakness (Figure 8.5). Based on these relations, pairwise

comparison matrices are formed (Table 8.4‒8.6). The inner dependence matrix of the SWOT

factor (w2) shows the computed relative importance weights.

𝑤𝑤2 = �

1 0.71 1 0.770.32 1 0 0.230.44 0 1 00.24 0.29 0 1

�

Next, the interdependent priorities of the SWOT factors are calculated as follows:

𝑤𝑤𝑓𝑓𝑚𝑚𝑑𝑑𝑡𝑡𝑡𝑡𝑑𝑑𝑠𝑠 = �

1 0.71 1 0.770.32 1 0 0.230.44 0 1 00.24 0.29 0 1

� × �

0.480.160.220.14

� = �

0.460.170.220.15

�

Strength depends on opportunity, weakness and threat, while threat depends on strength and

weakness. Weakness depends on strength and threat, whereas opportunity depends on strength.

Opportunity Strength Weakness

Threat

Figure 8.5: Inner dependence among SWOT factors (Yüksel and Dagˇdeviren 2007)

145

Table 8.4: The inner dependence matrix of the SWOT factors with respect to strengths

Strengths (S) W O T

Weaknesses (W) 1 1.1 0.9

Opportunities (O) 1 2.8

Threats (T) 1

CR = 0.15

Table 8.5: The inner dependence matrix of the SWOT factors with respect to weaknesses

Weaknesses (W) S T

Strengths (S) 1 2.5

Threats (T) 1

CR = 0.00

Table 8.6: The inner dependence matrix of the SWOT factors with respect to threats

Threats (T) S W

Strengths (S) 1 3.3

Weaknesses (W) 1

CR = 0.00

In the next step, priorities of the sustainability parameters (wsus), i.e., economic, environment and

social, are determined using pairwise comparisons (Table 8.7) and then multiplied with the relative

weights of the SWOT sub-factors. The results are denoted by wsub-factors (local) and represented as

local priority factor in Table 8.8. The data formulating the pairwise comparison matrices (among

the SWOT sub-factors) are detailed in the Appendix C. The overall priorities of the SWOT sub-

factors (wsub-factors (global)) are determined by multiplying the interdependent priorities of the SWOT

factors (wfactors) with the local priorities of the SWOT sub-factors (step 5). The resulting overall

priorities of the SWOT sub-factors (wsub-factors (global)) are shown in Table 8.8.

𝑤𝑤𝑠𝑠𝑠𝑠𝑠𝑠 = �EcoEnvSoc

� = �0.390.440.17

�

146

Table 8.7: Pairwise comparison among sustainability parameters

Sustainability parameters Eco Env Soc

Economic 1 1.1 1.9

Environmental 1 3.3

Social 1

CR = 0.04

The importance degrees of alternative strategies (increased flow diversion and continuing with the

current flow restrictions) with respect to each SWOT sub-factor are calculated following step 6.

The details of the pairwise comparison matrices are provided in the Appendix D. The comparisons

involve queries such as what is the importance of the increased power diversion (or continuing

with the current flow restriction) with respect to the installed generation capacity. The resulting

eigenvectors are shown by w4.

𝑤𝑤4 =

0.8 0.8 0.5 0.6 0.8 0.5 0.75 0.75 0.75 0.5 0.5 0.25 0.5 0.5 0.5 0.8 0.8

0.2 0.2 0.5 0.3 0.2 0.5 0.25 0.25 0.25 0.5 0.5 0.75 0.5 0.5 0.5 0.2 0.2

0.8 0.8 0.76 0.8 0.8 0.8 0.67 0.5 0.8 0.7 0.6 0.7 0.2 0.7 0.8 0.5 0.7

0.2 0.2 0.24 0.2 0.2 0.2 0.33 0.5 0.2 0.3 0.4 0.3 0.8 0.2 0.2 0.5 0.3

Finally, the overall priorities of the alternative strategies are calculated following step 7:

𝑤𝑤𝑚𝑚𝑡𝑡𝑡𝑡𝑝𝑝𝑑𝑑𝑛𝑛𝑚𝑚𝑡𝑡𝑖𝑖𝑎𝑎𝑝𝑝 = �Increased diversionCurrent restriction � = �0.675

0.325�

For comparison purposes, the model is analyzed with the hierarchical model without considering

the interdependence among the SWOT factors. When dependencies are ignored within the AHP

framework, factor priorities change from 0.46, 0.17, 0.22 and 0.15 to 0.48, 0.16, 0.22 and 0.14

respectively for the strength, weakness, opportunity and threat. Though, the priority order of the

strategies remains the same, the priorities under the AHP analysis change as follows:

𝑤𝑤𝑚𝑚𝑡𝑡𝑡𝑡𝑝𝑝𝑑𝑑𝑛𝑛𝑚𝑚𝑡𝑡𝑖𝑖𝑎𝑎𝑝𝑝 = �Increased diversionCurrent restriction � = �0.678

0.322�

147

Table 8.8: Priority of the SWOT sub-factors

SWOT factor SWOT sub-factor

Local priority factor

Global priority factor

Strength

Installed capacity with no fuel dependency 0.074 0.034

Pumped storage benefits 0.089 0.041

Protection of the installed equipment 0.194 0.089

Revenue from tourism sector 0.034 0.016

Renewable raw material lowering pollution 0.295 0.136

Regulating water level 0.147 0.068

Employment opportunities 0.102 0.047

Expected improvement in health condition 0.064 0.029

Weakness

High unit energy cost at SAB PGS 0.190 0.033

High investment cost 0.111 0.019

Turbine refurbishment 0.090 0.016

Disrupting the natural environment 0.241 0.042

Methane emission from flooded biomass 0.201 0.035

Resettlement 0.124 0.021

Restrict navigation 0.043 0.007

Opportunity

Expiration of the 1950 Treaty 0.091 0.020

Potential for generation with third tunnel 0.127 0.027

Demand mitigation in the absence of nuclear 0.081 0.017

Profit opportunities during summer 0.055 0.012

Flow alteration attracting return tourists 0.039 0.008

Erosion control through flow diversion 0.238 0.051

Control of misting with flow modification 0.096 0.021

Reduced emission from power generation 0.109 0.023

Employment in energy and tourism industry 0.127 0.027

Policy debate revisiting the treaty 0.040 0.009

Threat Reduced power potential under changing climate 0.140 0.021

Long payback period for third Niagara tunnel 0.080 0.012

148

Unfavourable outcome from renegotiation 0.119 0.018

Declining tourists 0.031 0.005

Cost associated with the renegotiation 0.023 0.003

Increase in over misty days at Niagara 0.241 0.036

Erosion of the Niagara escarpment 0.201 0.030

Stringent travelling requirements 0.113 0.017

Decreasing appeal of Niagara falls 0.054 0.008

8.4 Model Validation As the validity of the theoretical base of the ANP model is yet fully established, the proposed

methodology is subjected to similar shortcomings present in all studies that apply the technique

(Catron et al. 2013; Grošelj and Stirn 2015; Rauch et al. 2015; Shahabi et al. 2014). A number of

issues arises when validating the ANP model. First, the priority values are determined by pairwise

comparisons drawn from the experts’ judgment. However, assigning numerical measurements to

elements in a decision-making problem can often be challenging. Moreover, it is not always

possible to reproduce similar results each time since the data used in pairwise comparison matrices

may change depending on the selection and subjective views of the experts (Yüksel and

Dagˇdeviren 2007). However, this limitation is fundamental to all decision-making problems.

Second, the absence of previously analyzed models using past data provides little opportunity for

comparisons and validation. The comparison matrices are defined under the present conditions,

thus making it is possible to achieve different results at different points in time.

In the absence of specific criteria, the analysis attempts to validate it in three ways. First, the result

obtained using the ANP model is compared with that of AHP. A relatively small difference in

results is noted when these two methods are implemented. However, such differences are expected

as AHP allows only unidirectional hierarchical relationship among the SWOT factors, while ANP

accounts for the complex interactions among the decision attributes. While the ANP modelling

approach is more realistic and justified in this case, the method is less prominent in the literature

(Görener 2012; Othman et al. 2011). Hence, there might be challenges associated with

communicating the ANP framework to decision-makers in an engineering and public policy

149

context. However, lately more and more researchers (Azimi et al. 2011; Azizi and Maleki 2014;

Dubromelle et al. 2010; Foroughi et al. 2012; Fouladgar et al. 2011; Ostrega et al. 2011; Shahabi

et al. 2014; Stavrovsky et al. 2013; Wang et al. 2011b) are adopting the methodology and

demonstrating its potential use. Considering the superiority of the technique in representing

realistic decision situations, the authors join these researchers in advancing the application of the

ANP approach. Another parameter that verifies the validity of the model is the Consistency Ratio,

which stands for reliability with respect to the comparison matrix. Ideally, the Consistency Ratio

should be less than 10%, and reevaluated when above 20% (Saaty 1977; Margles et al. 2010). The

Consistency Ratio of the pairwise comparison matrices used in this study are well below 20% in

all cases. Finally, the authors repeated the analysis using the responses from three (3) seasoned

experts (out of the total seventeen) and compared it with the overall outcome. The differences in

the results are considerably low (0.01), thus providing further evidence of the model’s stability.

8.5 Conclusion The research proposes an improved decision-support framework that can be used for analyzing

resource systems from sustainability perspective and applies it for assessing the hydropower

potential at Niagara. It sheds light on Niagara’s current economic, environmental, social and

political dynamics, presenting and analyzing a holistic perspective of various stakeholders. The

sustainability framework (sSWOT) used here can be easily interpreted and ranked, and encourages

stakeholders’ participation – which makes it particularly palatable for policy prioritization. It also

permits assigning priority values to each sustainability parameters depending on the policy

objectives and interest of the analysis.

Based on survey responses, the analysis found that renegotiation of the 1950 Niagara River Treaty

may be favourable given the climate change and future energy scenarios. These findings call for

careful studies before any renegotiation attempt and illustrate the need for more flexible treaty

arrangements to permit periodic adjustments for addressing future challenges. A limitation of the

analysis is its reliance on authentic, representative survey data. Hence, choosing the appropriate

stakeholders who has adequate knowledge and understanding of the subject matter is of great

importance. In an academic context, the authors had to be content with the individuals with prior

knowledge and interest on the topic. However, real application would require adequate

150

representation of all the stakeholders – the local people, the industries, the conservation and power

authorities, etc. Despite these limitations, the approach demonstrated here can be reasonably

applied to any decision making problem with simple modifications in its structure.

151

A Bayesian Evaluation of Reliability, Resiliency and Vulnerability of the Great Lakes to Climate Change

While sustainable development is a concern, the vulnerability of water resources under a changing

climate is a major challenge for hydropower generation interests. Chapter 9 introduces a systematic

approach to recognize the changing resilience for Niagara River basin through the combined

application of Bayesian Network (BN) and systems performance criteria. Here, the BN is used to

model the complex interactions between hydro-climatic variables and predict consequences of a

projected altered climate on hydrologic conditions. The model outputs are analyzed and compared

with the baseline where the changing systems characteristics represent reliability, resilience and

vulnerably. The analysis on the Niagara watershed divulges considerable degradation under the

projected climate where the systems reliability and resilience reduce to a maximum of 0.99 and

0.5 respectively. Interestingly though, sections of the river on rare occasions are found to benefit

from a reduced vulnerability under specific climate scenarios.

The chapter will soon form the basis of a journal submission tentatively entitled “A Bayesian

Evaluation of Reliability, Resiliency and Vulnerability of the Great Lakes to Climate Change” by

Samiha Tahseen and Bryan Karney. Currently, this work is being finalized for a submission to a

suitable journal.

9.1 Introduction While the specific implications of global warming remains uncertain, the vulnerability of water

resource systems to changing climate is regarded as one of the major challenges of the twenty-first

century (Cosens and Williams 2012; Nemec et al. 2014). To address this vulnerability, there is a

growing interest in understanding the impact of climate stressors on water resources. The existing

climate studies project a changing precipitation pattern and an annual increase in global mean

temperature by the end of 21st century (IPCC 2013, US EPA 2016). Consequently, the resulting

impact on water resources varies depending on geographic location and the state of the watershed

itself. As a measure of its capacity to absorb and recover from these stresses (Folke et al. 2010;

Randhir 2014; Wilson and Browning 2012), watershed resilience is now a much discussed topic

in river basin conservation and management (Davidson et al. 2012, Soundharajan et al. 2016).

152

In general, resilience is the positive ability of a system to recover from the consequences of a

failure. Folke et al. (2005), Holling (1973) defined resilience slightly differently: ‘the extent to

which a system can absorb recurrent natural and human perturbations and continue to maintain

essential function without slowly degrading or even unexpectedly flipping into a less desirable

state’. Application of this concept to river basin has led to a process that assesses resilience by

investigating the system implications of natural variability in temperature, precipitation,

streamflow (Huang et al. 2009; Qi et al. 2016; Sun and Feng 2013; Wu et al. 2012; Yang et al.

2012), social and ecosystem responses, and drivers and feedbacks within the watershed (Davidson

et al. 2012; Merritt et al. 2015; Nemec et al. 2014; Randhir 2014). These studies largely apply

statistical methods (Liu et al. 2012; Zhang et al. 2014; Zhao et al. 2015), terrestrial or process-

based simulation models (Coe et al. 2002; Levine et al. 2015; Tahseen and Karney 2017: Chapter

5) for detecting changes across basins. While emerging approaches indicate the importance of

assessing and actively managing watershed resilience, application of these techniques to

multipurpose river systems continue to be a challenge because of the highly dynamic interaction

between the climate-hydro variables. Moreover, approaches to environmental management in the

past have largely focused on steady-state assessments (Milly et al. 2008), interpreting change as a

slow and gradual process, and thus disregarding interactions across scales (Folke et al. 2005).

Realizing this, recent studies emphasize the importance of explicitly incorporating the time

dimension into the definition of resilience (Haimes 2006, 2009a; b). Francis and Bekera (2014)

proposed a dynamic resilience framework while Botter et al. (2013) derived an index embedding

climate and landscape that characterizes erratic flow regimes. Qi et al. (2016) used a convex model

to measure the changing watershed resilience where annual discharge represents alterations in

long-term hydrological processes.

Accounting for one-fifth of the surface freshwater on the earth, the Great Lakes system is expected

suffer serious consequences from a warming climate. Present studies predict a warmer temperature

and changing precipitation pattern including a higher risk of more intense drought and flooding

(Kahl and Stirratt 2012). These sustained changes pose economic threats to numerous industries

that rely on lakes water supply, including tourism and hydroelectric generation (Magnuson et al.

1997; Tsanis et al. 2011). This chapter traces the spatial and temporal variations in the Niagara

river basin resilience under potential climate change scenarios in order to identify changing

direction/orientation in a way that should be useful to local conservation authorities. Probabilistic

153

approaches are particularly suitable for climate studies as they can explicitly incorporate the

inherent uncertainty associated with future climate projections (Cinar and Kayakutlu, 2010; Dessai

and Hulme 2004; Harris et al. 2014). In order to describe the probabilistic nature of risk, this

chapter applies the concept of a Bayesian Network (BN), an adaptive, flexible framework that can

test a wide range of system configurations against potential scenarios (Baroud and Barker 2014;

Mensah and Duenas-Osorio 2010).

9.2 Sustainability and Resilience In the context of natural hazards, the terms ‘sustainable planning’, and ‘resilience planning’ are

often used interchangeably. While the term ‘sustainable development’ has many definitions (Berke

1995; Berke and Conroy 2000; Campbell 1996; Kemp and Parto 2005; Lele 1991; May et al.

1996), the more recent versions include “the ability to prevent new risk creation and the reduction

of existing risk” (United Nations 2014). Resilience, as defined earlier, is the ability to recover after

a disaster (Manyena 1996; Mileti and Peek 2002; Paton et al. 2003). The more recent literature

describes resilience as an ‘adaptive capacity’ to acclimate/adjust to the demands, challenges and

changes encountered during and after a disaster (Klein et al. 2003; Norris et al. 2008; Paton and

Johnston 2006). The adaptation does not necessarily suggest ‘bouncing back’ to their former state

but rather evolving to deal with the changing circumstances. The definitions suggest that

sustainability and resilience are not one but rather interdependent phenomena. A system can only

be sustainable if it holds some degree of resilience. This is particularly reflected in the definition

by the UN Commission on Sustainable Development (Godschalk 2002) which suggests that

“Sustainable development… cannot be successful without enabling… to be resilient to natural

hazards”. In contrast, a resilient system could possibly exist in an unsustainable environment.

Following a disaster, there are two typical timeframes. First is the relatively short duration

immediately after a disaster. The second timeframe is much longer and encompasses the recovery

period which may extend up to years (Schwab et al. 1998). In this latter period, the boundary

between resilience and sustainability may overlap. When hazardous events occur repeatedly

leading to compounding effects, short-term adaptations may fall short of addressing the situation.

Such events require a set of adaptive measures employed over the long term which is more in line

with the concept of sustainable development (Saunders and Becker 2015). While improved

154

resilience enhances the likelihood of sustaining development into the future, the scope of this

research is limited to the changing systems resilience immediately following an anomaly/disaster.

9.3 Study Area The Great Lakes comprise a series of interconnected freshwater lakes located on the Canada–

United States border. The relatively deeper and colder upper lakes connect to Lake Erie through

St. Clair River, Lake St. Clair, and the Detroit River. Lake Erie flows over Niagara Falls and into

Lake Ontario before flowing through the St. Lawrence River into the Atlantic Ocean. The lower

lakes (Erie and Ontario) are connected by the Niagara River. The river is approximately 58 km

long, and carries an average of 5,660 m3/s from Lake Erie to Lake Ontario (Kirkham 2010). While

the current 5,000 MW hydropower potential at Niagara provides significant economic value, the

lack of regulation at Lake Michigan-Huron and Lake Erie puts this generation at risk from a

changing climate. Considering the valuable tourism industry and its strategic importance as an

international waterway, this study selects the Niagara River basin for the proposed resilience

assessment. Figure 9.1 shows the study area along with key locations of interest to this work.

Within the Great Lakes region, annual mean temperatures have increased by 0.7–0.9˚C between

1895 and 1999 (Mortsch et al. 2000). Alarmingly, the trend is expected to continue with a

substantial increase in mean annual temperatures (Kling et al. 2003; Taylor et al. 2006). These

increases may range from 2‒5˚C in summer and 4‒8˚C in winter under a doubling CO2

concentration scenario (Magnuson et al. 1997). Kling et al. (2003) predicted a warming of 3‒7°C

in winter and 3‒11°C in summer by the end of this century. Precipitation trends for the Great Lakes

- St. Lawrence basin indicate that the amounts have increased between 1895 and 1995 (Mortsch et

al. 2000). Projections generally expect this trend to continue throughout this century. Croley

(1990) developed a hydrologic model for the Great Lakes linking the climate variables with that

of the lakes. In all possible climate scenarios considered to date, water resources are expected to

decline in the basin (Mortsch and Quinn 1996; Mortsch 2003). These changing lake conditions

may exert significant impacts on the terrestrial and aquatic ecosystems, modify or eliminate

wetlands (Branfireun et al. 1999; Devito et al. 1999; Lemmen and Warren 2004; Mortsch 1998)

or cause supply, odour, and taste problems in communities with shallow water intakes (Nicholls

1999; Schindler 1998). The present study formulates a probabilistic model for predicting

155

hydrologic changes under future climate scenarios based on the understanding of regional climate

projections for the Great lakes.

Figure 9.1: The Niagara River basin

9.4 Methodology This section elaborates the BN model development and the application of the Reliability-

Resilience-Vulnerability (RRV) criteria by Hashimoto et al. (1982).

Ashland Ave. stn.

American Falls stn. Niagara

Intake stn.

Buffalo stn.

Olcott stn.

Niagara on the Lake Golf stn. Niagara on

the Lake stn.

Buffalo International Airport stn.

Niagara Falls stn. Niagara Falls

NPCSH stn.

Ontario Hydro stn.

Chippawa stn.

Niagara Falls International Airport stn.

Niagara Falls 5.7E stn.

Legends: Weather stations Hydrologic stations

New York

Ontario

156

9.4.1 Model preliminaries The first step in model building, that of specifying the study’s objective or purpose, is to assess

the risk associated with changing hydrologic conditions under a varying climate. In general, the

climate system exhibits such complex, interdependent behaviour that makes it fundamentally

impossible to include all the climate processes regardless of how complex the model is (Tebaldi

and Knutti 2007). Hence, choices have to be made on which variables and processes to include

and how to parametrize them depending on the scope of the work. Literature review and interviews

with one or more domain experts are typically required at this stage in order to identify the

important variables required to meet the core objective of the BN model (Lawson et al. 2016,

Constantinou et al. 2015). This study relies on the recent literature on the Great Lakes which is

used in conjunction with suggestions from two domain experts for variable selection. The primary

(natural) factors affecting the lake levels are found to be precipitation, evaporation from the lake

surface and wind (Fisheries and Ocean Canada 2016; International Joint Commission 2016).

9.4.2 Data management The key objective of data management is to link variables to model nodes. The process begins by

identifying potential data sources in order to gather information for training the model, validation

and drawing inferences about the long-term changes in hydrologic variables.

The data required for the study were collected in two stages. First, historical climate and hydrologic

data were obtained for the Niagara watershed. Next, the author gathered information on climate

projections for the study area which is then used to develop scenarios to be modelled.

Meteorological data were collected from National Oceanic and Atmospheric Administration

(NOAA) sites for nine strategic locations along the Niagara River at Canada-US border (Figure

9.1). The stations were spatially distributed throughout the watershed but aligning closely with the

river. The dataset obtained from NOAA National Centers for Information and Technology (2016)

includes daily information on temperature, precipitation, snow depth, snow fall and occasionally

wind speed for the years spanning from 1950 to 2015. The full climate time series further includes

wind direction data at Lake Erie, collected from NOAA National Data Buoy Center (NDBC)

(2016), with the evidence going as far back as 1980. The hydrologic data used in the model are

elevations, flows, lake surface temperatures and ice concentrations. The flow and water level

157

information were extracted and further processed for five different gauges (Buffalo, NY Intake,

American Falls, Ashland Ave. and Olcott as shown in Figure 9.1) from NOAA Center for

Operational Oceanographic Products and Services (2016). Lake surface temperatures were

collected from NDBC (2016) and Sharma et al. (2015), while lake ice concentration data were

obtained from NOAA Great Lakes Ice Atlas (2016). Next, the author performed normality test

(Shapiro and Walk 1965) in order to categorize the nodal variables. Upon failing normality

assumption, the model divides the data into ‘N’ categories of uniform bin sizes (where N is 2 or

3). The categorizations for a selection of variables are shown in Figure 9.2.

Temperature at Buffalo Elevation at Ashland Ave.

-19.01‒ -0.37 33.69 -0.37 ‒ 18.17 33.25 18.17 ‒ 36.81 33.07

95.86 ‒ 98.33 34.18 98.33 ‒ 100.78 32.99 100.78 ‒ 103.25 32.83

Figure 9.2: Variable discretization within the BN model (The ranges on the left represent the data intervals (bin sizes) and the number 33.7, and its visualization, indicates

the percentage of data points in that interval)

As mentioned, the second stage of the data management elaborates four different climate scenarios

to be assessed through the model. In this study, these were the Canadian Climate Centre GCM

(CCC GCM2) (Boer et al. 1992; McFarlane et al. 1992), the Goddard Institute for Space Studies

(GISS) (Hansen et al. 1983), Geophysical Fluid Dynamics Laboratory (GFDL) (Manabe and

Wetherald 1987), and Oregon State University (OSU) (Ghan et al. 1982). These scenarios

represent changes in meteorological conditions (temperature and precipitation) under a doubling

CO2 concentration scenario (Mortsch and Quinn 1996) and are briefly discussed in Table 9.1 and

Table 9.2. Based on these projections, four annual climate time series (CCC-GCM2, GFDL, GISS

and OSU) were developed to represent the changing temperature and precipitation pattern under

2xCO2 scenario. Finally, the data were processed for missing values and possible inaccuracies

before moving to the subsequent step in model development which involves constructing a

Bayesian network structure.

158

Table 9.1: GCM-simulated temperature increase for the Great Lakes-St. Lawrence Basin:

Change from 2xCO2 to 1xCO2

GCM Winter Spring Summer Autumn CCC GCM2

Greatest increase of all seasons (4.0 – 9.1 °C)

SW part shows sharp increase (3.3 – 8.3 °C)

SW part shows sharp increase (3.9 – 6.2 °C)

Smallest increase of all seasons (2.7 – 4.7 °C)

GISS Warmest in N part (4.5 – 6.6 °C)

Warmest in S part (3.8 – 4.8 °C)

Steady increase in temp. (2.7 – 3.8 °C)

Increase in temp. as move NE-SW (3.0 – 6.0 °C)

GFDL Sharp increase as move N (5.0 – 8.7 °C)

Similar rise as in winter (4.4 – 8.0 °C)

Very sharp rise as move E-W (5.6 – 8.6 °C)

Season with smallest rise as move E-W (5.6 –7.0 °C)

OSU Greatest increase of all seasons (3.4 – 4.2 °C)

Warming gradually as move E-W (2.9 – 3.5 °C)

Temp. increase as move E-W (3.0 – 4.0 °C)

Very gradual increase as move S (2.6-3.3 °C)

Table 9.2: GCM precipitation ratios for the Great Lakes-St. Lawrence Basin. (2xCO2 to 1xCO2)

GCM Winter Spring Summer Autumn CCC GCM2

Wetter in N and NW parts; drier in SW parts (0.9 – 1.2)

Sharp rise in precip. as move N (0.9-1.4)

Generally drier than normal except for NE (0.8-1.1)

Sharp drop in precip. as move S; increase in N (0.7-1.3)

GISS Progressively wetter as move N (1.0 – 1.2)

Wetter as move NE (1.0 – 1.1)

Increase in precip. as move N (1.0 – 1.3)

Sharp decrease in precip. as move NW-SE (0.7-1.2)

GFDL Sharp rise in precip. throughout basin (1.1 – 1.3)

Precip. increases as move NW-SE (0.95-1.2)

Sharp decrease in precip. throughout basin (0.7-0.9)

Precip. increase in SE part (0.8-1.1)

OSU Precip. increases as move SE-NW (1.0 – 1.2)

Precip. decreases as move NE-SW (0.9-1.1)

Decrease in N part; increase in S portion (0.9-1.1)

Sharp increase as move SE-NW (1.0-1.3)

159

9.4.3 BN structure BN is a probabilistic graphical modelling method where a system is represented with a joint

probability distribution compacted with the notion of conditional independence (Li et al. 2016;

Sagrado et al. 2014). Later, this model can be used to understand the dynamics within the system

and compute posterior probabilities of unobserved variables conditioned on variables that have

been observed (Herring 2014; Hines and Landis 2014; Hosseini and Barker 2016). Any belief

about uncertainty of some event A is assumed to be provisional upon experience or data gained to

date. This is called the prior probability, written P(A). This prior probability is then updated by

new experience or data B to provide a revised belief about the uncertainty of A that we call the

posterior probability, written P(A|B); that is, the refined probability of A given the occurrence of

B. The formula to determine P(A|B) based on Bayes’ theorem is:

𝑃𝑃(𝐴𝐴|𝐵𝐵) = 𝑃𝑃(𝐵𝐵|𝐴𝐴) 𝑃𝑃(𝐴𝐴)

𝑃𝑃(𝐵𝐵)

(9.1)

In BN, each variable is represented with a node and the influence of a variable on others is

demonstrated with directed edges/arcs which may represent causal, influential, or correlated

relationships (Constantinou et al. 2015; Pearl 2003). When using the model to make predictions,

the variables predicted are called the outcome variables and the remainder are called decision

variables. The full joint probability distribution of a BN consisting of n variables X1, X2,…….,Xn

is as follows:

𝑃𝑃(𝑋𝑋1, 𝑋𝑋2, … . ,𝑋𝑋𝑛𝑛) = 𝑃𝑃�𝑋𝑋1�𝑋𝑋2,𝑋𝑋3, … ,𝑋𝑋𝑛𝑛�𝑃𝑃�𝑋𝑋2�𝑋𝑋3, … ,𝑋𝑋𝑛𝑛�…𝑃𝑃(𝑋𝑋𝑛𝑛−1|𝑋𝑋𝑛𝑛) 𝑃𝑃(𝑋𝑋𝑛𝑛)

= �𝑃𝑃(𝑋𝑋𝑖𝑖|𝑋𝑋𝑖𝑖+1 …𝑋𝑋𝑛𝑛)𝑛𝑛

𝑖𝑖=1

(9.2)

Provisionally assuming conditional independence of the variables, the joint probability distribution

of a BN can be written using parent nodes of each node. For example, if we know that node X1 has

exactly two parents, X2 and X3, then Eq. 9.2 can be compactly written like this:

𝑃𝑃(𝑋𝑋1, 𝑋𝑋2, … . ,𝑋𝑋𝑛𝑛) = �𝑃𝑃(𝑋𝑋𝑖𝑖|Parents(𝑋𝑋𝑖𝑖))𝑛𝑛

𝑖𝑖=1

(9.3)

160

This substantially reduces the size of the conditional probability tables (CPTs) of different nodes.

When evidence received about possible states of a set of variables, the marginal and conditional

probabilities can be computed by marginalizing over the joint.

The structure and the relationships in BNs can rely on both expert knowledge and relevant

statistical data (Lawson et al. 2016; Peter et al. 2009), and can incorporate both quantitative and

qualitative information in a conditional probability format, i.e., variables could be Boolean

(yes/no), qualitative (low/medium/high), or continuous. The open modelling architecture that

allows easy updating and handling missing data makes BN a preferred choice for modelling in this

study. Also, the BN structure constructed at the initial stage was quite different from the final

version as a result of subsequent iterations. However, the conceptual flow of the network has

remained effectively unchanged. The BN model for Niagara River hydro-climatic interactions is

provided in Appendix E.

9.4.4 Parameter learning The process of determining CPT entries for each node of the BN model is called parameter

learning. Here, the hydro-climatic data series are used for calculating the prior probabilities.

However, because of the real-world limitations, there are nodes or individual parameter with

missing values. Some publications provide data-driven techniques for dealing with missing data

(e.g., Little and Rubin 2002):

(1) Restrict parameter learning only to cases with complete data.

(2) Use imputation-based approaches where missing values are filled with the most

probable value, based on the values of known cases, and then the CPTs are calculated considering

a full dataset (Enders 2006).

(3) Use likelihood-based approaches where missing values are inferred from existing

model and data (Constantinou et al. 2015). The Expected Maximization (EM) algorithm, an

iterative method for approximating values of missing data (Lauritzen 1995), is commonly used for

this purpose and is widely accepted as a standard tool in BNs. This study applies the algorithm to

learn the CPTs as well as to develop the model structure by searching for connections between

parameters, while establishing recognized relationships (such as air temperature affecting lake

surface temperature) from the data. To reduce complexity and computing resource requirements,

161

direct links among meteorological variables were restricted. Other restrictions set during model

development include the assumption that there are no “directed edges” from hydrologic variables

to climate variables and that climate data are expected to influence the nearest upstream river

conditions.

9.4.5 Structural validation Sensitivity analysis (SA) is a simple yet useful technique for model validation. In SA, outcomes

of a set variables are recalculated using alternative assumptions. The process provides insights

regarding which nodes, and under what states, have the greatest impact on a selected outcome

variable. Here, SA is performed to assess the model structure, the CPTs and the overall robustness

of the BN model and to identify possible irrationalities in both the BN structure and the underlying

CPTs (Coupe 2016). The analysis can further validate the influence of instantiated climate

variables, since different nodal instantiations lead to different sensitivity scores. Figure 9.3 is

generated by analyzing the impact of changing meteorological conditions at Buffalo based on the

hydro-climate time series. The influence of four climate variables, i.e., maximum and minimum

temperature, lake ice coverage and lake surface temperature are assessed where each of these

parameters are instantiated with long-term maximum and minimum values. For instance, the graph

indicates that if recorded minimum temperature at Buffalo is set below the 33th percentile value,

the probability of Buffalo elevations remaining between 172.8 – 173.8 m is 0.331. However, if the

same value is set to be above the 66th percentile value, the probability reduces to 0.327. It further

suggests that the changing temperatures and increased lake ice coverage are likely to have the

greatest impact on water level at this location. The relatively small differences in probabilities with

changing climate variables deem reasonable considering that these changes were kept limited to

their historical ranges.

162

Figure 9.3: Probability ranges with different instantiations at Buffalo

Once validated, the previously discussed climate scenarios were run through the model where the

maximum probability indicates the most probable range for the respective hydrologic conditions.

Figure 9.4 illustrates the step-by-step procedure for the Bayesian model development.

Figure 9.4: Step-by-step procedure for the BN model development

0.326

0.327

0.330

0.328

0.330

0.331

0.326

0.328

Tmax

Tmin

Erie_ice

Erie_Temp

Min Max

Determine the model objectives

Identify core variables

Variables categorization Data collection

Define the BN structure

Predictive model validation

Satisfied with model performance

EM algorithm

Perform decision analysis

yes No

Mode objectives

Data management

BN structure

Parameter learning

Validation

163

9.4.6 Dynamic reliability, resilience and vulnerability After predicting future hydrologic conditions and its associated probabilities, it is important to

incorporate temporal dynamics into the analysis. A comprehensive approach for evaluating

systems performance through the application of Reliability, Resilience, and Vulnerability (RRV)

metrics is provided by Hashimoto et al. (1982) where reliability stands for the probability of

failure, resilience indicates the time required to recover from a failure condition and vulnerability

is a measure of the degree or extent of the unsatisfactory conditions. The metrics assesses the

system's ability to (i) anticipate and absorb potential disruptions; (ii) develop adaptive means to

accommodate changes within the system; and (iii) establish response behaviours aimed at either

building the capacity to withstand the disruption or recover as quickly as possible after an impact.

The indicators are discussed in the following section.

Reliability is defined as the probability of system being in satisfactory state at any given time

(Hashimoto et al. 1982), where failure implies deviation from long-term averages. In other words,

reliability is a measure of the likelihood that a stream does not violate the established standard for

a given constituent at a given time. Denote the system state by random variable Xt at time t, where

t takes on discrete values 1,2…..,n. Then, the possible Xt values can be partitioned into two sets:

S, the set of all satisfactory outputs, and F, the set of all unsatisfactory outputs. The reliability of

the system can be expressed as (Hashimoto et al. 1982):

Reliability = Probability (Xt ∈ S) (9.4)

Defining a state variable Z, where, if Xt ∈ S; Zt = 1 else Xt ∈ F and Zt = 0. Then,

Reliability = ∑ 𝑍𝑍𝑡𝑡𝑇𝑇𝑡𝑡=1

𝑇𝑇

(9.5)

In this research, the RRV model has a daily temporal resolution. Hence, the value for T in this

analysis is 365 days.

Resiliency is described as how quickly a system is likely to recover or bounce back from failure

once failure has occurred (Hashimoto et al. 1982). Resilience, therefore, strongly depends on the

assimilative capacity of the stream and can be stated as (Hashimoto et al. 1982; Mondal et al. 2010;

Weeraratne et al. 1986):

164

Resiliency = Probability (Xt ∈ F and Xt+1 ∈ S) (9.6)

Let Wt be an indicator for the transition from unsatisfactory to satisfactory state such that

𝑊𝑊𝑡𝑡 = �1, 𝑋𝑋𝑡𝑡 ∈ 𝑆𝑆 and 𝑋𝑋𝑡𝑡+1 ∈ 𝐹𝐹0, otherwise � (9.7)

Then, resiliency can be defined with respect to Wt as:

Resiliency = ∑ 𝑊𝑊𝑡𝑡𝑇𝑇−1𝑡𝑡=1

𝑇𝑇 − ∑ 𝑍𝑍𝑡𝑡𝑇𝑇𝑡𝑡=1

(9.8)

Lastly, vulnerability provides a measure of potential damage caused by a system failure

(Hashimoto et al. 1982). For watershed health, this would estimate the impact of violating a

constituent standard flow or water level. In reality, few systems can be made so redundant to avoid

failure altogether. However, even when the probability of failure is low, attention should be paid

to the associated damage in order to minimize the effect. The vulnerability is an important criterion

to describe the severity of failures for water systems which is defined by Hashimoto et al. (1982)

as:

Vulnerability = ∑ 𝑒𝑒(𝑗𝑗)ℎ(𝑗𝑗)𝑗𝑗∈𝐹𝐹 (9.9)

where h(j) is the most damage that can be incurred during the jth failure instance and e(j) is the

probability that h(j) is indeed the most damage that could have been incurred.

9.5 Analysis of the Historical Data The historical elevation and flow conditions across the river are analyzed in terms of magnitude,

frequency, duration, timing and date using Indicators of Hydrologic Alteration (IHA) (Poff et al.

1997; The Nature Conservancy 2009). Non-parametric statistics were preferred because of the

skewed (non-normal) nature of the hydrologic datasets. The hydrograph in Figure 9.5 categorizes

the flow into high, low and extremely low where all flows exceeding 75% are considered high,

below 50% are low and below 10% are classified as extremely low. IHA further calculates median

per month, daily, weekly, monthly, 90-day minimum and maximum, seasonal highs and lows when

given relevant information for the watershed. The river experiences higher than normal flows

during spring (Figure 9.5) which according to the Environment and Climate Change Canada

165

(2016) is influenced by snowmelt. The low flow conditions are likely to occur in late winter

(January and February) when the river has considerable ice coverage and precipitation gets stored

in the form of ice and snow. The stream flow is more uniform during spring and summer, yet

shows an erratic regime with high and low flow pulses during winter. Similar explorations with

water level suggest a large seasonal variation ranging from 0.4–3 m at different gauges across the

river basin. Ashland Ave. shows the greatest variability as a result of upstream power operations.

Both Lake Erie and Lake Ontario experience relatively large elevation variations with a maximum

of 1 m and 0.8 m respectively. The analysis further estimates the 10th, 25th, 50th, and 90th percentile

elevations and flows for each month using the historical data. These ranges are set as constituent

standards for respective hydrologic conditions at each station.

Figure 9.5: Flow classification for the Niagara River (1950–2011)

9.6 Reliability The reliability values are analyzed separately for water level and flows. The results are reported

varying the allowable range (or standard) between 10th and 90th percentile long-term values. Since

the output from the BN model also provides a range for each predicted hydrologic conditions,

these constituent standards (based on the historical data) allow comparison of the upper and lower

boundaries.

166

9.6.1 Critical level between 10th, 25th and 75th percentile values The analysis here compares the systems reliability between the baseline and the climate scenarios

(CCC-GCM2, GFDL, GISS and OSU) at Buffalo, Niagara Intake, American Falls, Ashland Ave.

and Olcott gauges. Figure 9.6 (a, b, c) shows the estimated reliability considering the monthly

allowable ranges between 10th and 75th percentile long-term elevations at these stations (from

IHA), meaning that for a reliable system, the river is conditioned to maintain a critical elevation

between these ranges. When the 75th percentile elevations are compared with those in the base year

(1995) (Figure 9.6a), the reliability ranges between 0.79–0.92 with the maximum at Ashland Ave.

station. Under the climate scenarios, the upstream river stations (Buffalo, Niagara Intake and

American Falls) exceed these upper bounds in all instances for a given year. The large water level

fluctuations at Ashland Ave. in the baseline establish a relatively relaxed standard (due to increased

interquartile space) that results in an improved reliability under the climate scenarios (Figure 9.6a).

Elevations at Olcott station, representing Lake Ontario in the model, surpass the high water

benchmarks during most of the year. Comparison between the model output and the baseline with

respect to the lower bounds (Figure 9.6c) suggests that the river elevations are likely to decrease

below the historical 10th percentile values under future climate scenarios.

The least affected are the lakes – Erie (represented by Buffalo) and Ontario – where their relatively

large size and volume play an important role in mitigating large elevation fluctuations, thus making

them comparatively stable. Being the shallowest and smallest (by volume) of the Great Lakes,

Lake Erie is more prone to violate the percentile ranges compared to Lake Ontario (Figure 9.6) –

a fact confirmed by this analysis. The systems reliability shows reasonable variations with the

changing percentiles values as the constituent standards. For example, if the lower bounds are

increased from the 10th to 25th percentile monthly elevations, the systems reliability declines under

the baseline and all four climate scenarios (Figure 9.6b). While the reliability ranges from 0.44–

0.98 in the baseline, similar results under the climate scenarios demonstrate a persisting low water

level at all gauges except Olcott. Lake Ontario, by the virtue of its size, mostly mitigate the low

water conditions, however lake elevations are likely to drop below the benchmark during summer

and spring.

167

Figure 9.6: Reliability under the baseline and future climate scenarios when comparing (a) the

upper limits - 75th percentile, (b) the lower limits - 25th percentile and (c) 10th percentile

9.6.2 Critical level between median and 90th percentile values In this section, the results are analyzed considering a more conservative lower bounds of median

(50th percentile value) and upper bounds of 90th percentile elevations for each month. This means

the daily water level, if predicted to exceed this range, the system reliability would be

compromised. While the baseline reliability ranges between 0.87–0.98 with respect to the upper

bounds, the upstream river gauges are more likely to exceed the limit in all instances under the

climate scenarios. While elevation overshoots are likely at Lake Ontario during fall and winter and

at Ashland Ave. during late summer, these stations are relatively more stable under a varying

climate with a reliability score of 0.34 and 0.75 respectively. The outcome is rather alarming when

the model outputs are compared with the monthly lower bounds (median); predicted elevations at

all stations except Lake Ontario go beyond the respective medians in all instances for a given year.

Also, the trivial differences in reliability under CCC-GCM2, GFDL, GISS and OSU scenarios

indicate a relatively high degree of convergence among these projections. The null score with

respect to both the boundaries (upper and lower) at Buffalo and American Falls further indicates

the need for increased data discretization (increased number of bins) – a measure that is found to

be quite time and resource intensive.

9.6.3 Seasonal high and lows (spring and winter) Here, the authors compare the predicted water level from the BN model with seasonal highs and

lows. The water year starts from the beginning of April and two distinct seasons representing the

0.00.20.40.60.81.0

Buf

falo

Nia

gara

Inta

keA

mer

ican

Falls

Ash

land

Ave

.

Olc

ott

Baseline CCC-GCM2GFDL GISSOSU

0.00.20.40.60.81.0

Buf

falo

Nia

gara

Inta

keA

mer

ican

Falls

Ash

land

Ave

.

Olc

ott

BaselineCCC-GCM2GFDL

0.00.20.40.60.81.0

Buf

falo

Nia

gara

Inta

keA

mer

ican

Falls

Ash

land

Ave

.

Olc

ott

Baseline CCC-GCM2GFDL GISSOSU

168

high (spring) and low flow (winter) conditions are considered. The analysis estimates the changing

reliability when the predicted outcomes are compared with the seasonal highs and lows in spring

(April to July). The elevations in the base year rarely reach the seasonal extremes, resulting in a

relatively high system reliability between 0.98–1; however, the number decreases substantially

under all climate scenarios. The analysis suggests that lake elevations, represented by Buffalo and

Olcott stations, are likely to exceed the high water benchmark in spring, while that of Ashland

Ave. may decrease below the minimum. Since elevations at Ashland Ave. are primarily influenced

by the upstream power plants, successful planning and operations bear the potential of reducing

variability at this location. Again, the null values at Niagara Intake and American Falls highlight

the need for increased data discretization.

Similar analysis is performed for the winter that lasts from December to March. The reliability in

the baseline ranges from 0.88–1, however declines under the climate scenarios. While Niagara

intake and Ashland Ave. remain below the critically high level, these stations are likely to

experience low winter elevations under a varying climate. In contrast, lake elevations may rise

under CCC-GCM2, GFDL and GISS scenarios. Interestingly, the increase in Lake Ontario

elevations in a few instances allow the water levels to transition to a satisfactory state, thus

resulting in an improved reliability compared to the base year.

9.6.4 Flow conditions Here, the analysis focuses on the changing flow characteristics at the Niagara River under the

CCC-GCM2, GFDL, GISS and OSU scenarios. The flow data used for the model development

and the comparative analysis are measured at Buffalo station. The predicted flows from the model

are tested for its compliance to the historical 10th, 25th and 50th percentiles as the lower bounds and

75th and 90th percentiles as the upper bounds (Table 9.3). Under the baseline, the systems reliability

varies between 0.56–0.99 where the lowest estimate results from a strictly conservative analysis

where the monthly median flows were set as critical minimum. These high reliability values in the

base year suggest that the flow regime is fairly uniform with occasional large-scale variations.

However, increased temperature coupled with a changing precipitation pattern under the climate

scenarios results in a reduced reliability for the river basin. The stream flow is likely to exceed the

long-term values under the climate change. However, at times during late spring and early summer,

the river is likely to experience a low flow conditions.

169

Table 9.3: Reliability considering 10th, 25th and 50th (lower boundary), 75th and 90th (upper

boundary) percentile flows

Boundary Baseline CCC-GCM2 GFDL GISS OSU

10% 0.99 0.75 0.76 0.75 0.75 25% 0.98 0.5 0.5 0.5 0.5 50% 0.56 0 0 0 0 75% 0.90 0 0 0 0

90% 0.97 0 0 0 0

9.7 Resilience

9.7.1 Critical level between 10th, 25th and 75th percentile values The changing resilience for the watershed is reported on the basis of the 10th, 25th (lower limit)

and 75th (upper limit) percentile long-term elevations under the baseline and the climate scenarios.

Resilience typically varies between 0 and 1, where 0 stands for low resiliency and 1 represents a

highly resilient system. When compared with the 75th percentile elevation, the watershed in the

base year is moderately resilient with values ranging between 0.13–0.5. While deviations are

relatively quickly mitigated at Niagara Intake and Ashland Ave., most conceivably through the

partial control due to power operations, low water levels at Lake Ontario are often long, persisting

events (in the baseline). Interestingly, quite the opposite takes place under all the climate scenarios

where Olcott shows a greater resiliency among other stations. This can be explained by the spatial

variation in the climate projections that at times increases precipitation over Lake Ontario, thus

allowing relatively faster rebounds. In general, the overall resiliency plummets under all climate

scenarios with most significant reduction under the CCC-GCM2 scenario. Comparison concerning

the 10th and 25th percentiles values suggests that the upstream river, which is more resilient to

changing elevations under the baseline, fails to mitigate low water conditions under the climate

scenarios. Notably, with the more conservative lower boundary (25th percentile), there is a

substantial decrease in Lake Ontario resilience in the base year such that the system may gain from

a varying climate conditions under GISS and OSU scenarios.

170

9.7.2 Critical level between median and 90th percentile values When analyzed with respect to the upper boundary of 90th percentile monthly elevation, the system

resilience under the baseline ranges from 0.57–1 indicating that high water levels are relatively

quickly mitigated across the river basin. However, these values decrease to 0.04–0.37 across

various gauges when compared to a conservative lower boundary of median. Under all four climate

scenarios, the systems resilience declines throughout the river basin suggesting the possibility of

high and low flow pulses. Notably, the downstream section of the river is more resilient to such

changes. The analysis further emphasizes the need for improved data discretization for hydrologic

variables.

9.7.3 Seasonal high and lows In the base year, the system is quite resilient with the exceptions of Grass Island Pool (GIP) when

compared to spring highs, and American Falls when considering seasonal lows. The decreasing

resilience at Buffalo and Olcott under CCC-GCM2, GFDL, GISS and OSU scenarios indicate the

need for control to alleviate critically high level at these locations. An opposite response is

recorded at Ashland Ave. where elevations may decrease below the long-term lows, and resulting

in a reduced resilience.

Similar analysis when performed for the winter, the baseline resilience varies between 0.07–1 with

relatively low estimates at American Falls (considering winter highs) and Olcott (with respect to

seasonal lows) (Figure 9.7). The declining resilience under the climate scenarios suggests that the

river’s assimilative capacity may be insufficient to mitigate low water level adjacent to the Niagara

Falls. A reverse can be seen at Buffalo as the elevations are more likely to overshoot during winter.

Interestingly, the analysis suggests a substantial gain in terms of improved resilience at Lake

Ontario under all climate scenarios (CCC-GCM2, GFDL and GISS and OSU).

171

Figure 9.7: Changing resilience under the climate scenarios considering winter highs and lows

9.7.4 Flow conditions The discussion here informs about the impact on flow conditions and the resulting changes in

systems resilience due to climate change. Under the baseline, the systems resilience varies between

0.49–0.83 with relatively lower estimates resulting from more conservative boundaries. Under the

projected climate scenarios, the resilience reduces close to zero when compared to the long-term

upper and lower limits. These outcomes suggest that the river is likely to experience large-scale

flow variations with a changing climate. The flow predictions are further compared with seasonal

highs and lows in winter and spring. In both the seasons, the flow remains above the seasonal lows,

however exceeds the critically high values.

9.8 Vulnerability

9.8.1 Critical level between 10th, 25th and 75th percentile values Vulnerability measures the severity associated with failure even though the probability of such

events are low. Here, the damage with regards to each failure event is estimated by deviations from

the historical values. To assess vulnerability, the most severe of these damages is multiplied with

the probability of such event in the sojourn and aggregated over the year. Interestingly, when

predicted elevations from the model are compared with the long-term values, Lake Erie, Niagara

Intake and Ashland Ave. are found to be less vulnerable under all climate scenarios. The reason

behind such contradictory outcome, as both the reliability and resilience decline under these

0.0

0.2

0.4

0.6

0.8

1.0

Buffa

lo

Nia

gara

Inta

ke

Am

eric

anFa

lls

Ash

land

Ave

.

Olc

ott

Baseline_high Baseline_low Climate_high Climate_low

172

scenarios, is that the increased number of unsatisfactory events (elevations exceeding or falling

below the historical range) reduces the probability of occurrence for the most severe events in the

sojourn (which is probability of the most severe event divided by probability of all unsatisfactory

events). This, when multiplied with the associated damage, results in a reduced overall

vulnerability. For example, the predicted elevations at Buffalo under the CCC-GCM2 scenario

exceed the upper bounds at every time step with severity varying between 0.23–0.53. Similar

analysis for the base year shows that the events with undesirable elevations occur sporadically

within a year. Though the maximum severity in the latter case is only 0.32, the overall vulnerability

is greater when aggregated over the year. Also, the increased vulnerability at Ashland Ave. in the

baseline may not be representative, since elevations at this r station vary substantially as a result

of upstream power operations – an effect not represented in the BN model outputs. Consequently,

the difference between the instantaneous observations and the relevant constituent standards

(based on the historical data) are greater in the baseline, resulting in a higher severity than that

under the climate scenarios. The portion of the river close to the American Falls and Lake Ontario

are more vulnerable under a changing climate as elevations occasionally drop below the historical

monthly lows. The worst affected is Lake Ontario under the OSU projections where vulnerability

increases by 2.3 unit despite a low probability of failure.

9.8.2 Critical level between median and 90th percentile values The vulnerability increases at the downstream locations under future climate projections with

respect to the 90th and 50th percentile elevations. What this means is elevations at these stations

may exceed or drop below the respective long-term values. Notably, low elevations (below

median) are more common in summer whereas a reverse is seen during winter. There is a

substantial increase in vulnerability for Lake Ontario under GFDL and OSU projections, as these

scenarios predict increased temperature along with reduced precipitation for the south-east part of

the Great Lakes basin in summer. Interestingly, the upstream river section benefits from a reduced

vulnerability under the climate scenarios which results from an increased failure events reducing

the probability of the most severe failure.

9.8.3 Seasonal high and lows This section discusses the changing vulnerability with regards to the spring highs and lows. The

upstream section of the river (Buffalo, NY Intake and American Falls) is increasingly vulnerable

173

under the climate scenarios due to elevations exceeding the long-term seasonal trends.

Interestingly, Lake Ontario elevations, which in previous sections are reported to exceed the

seasonal highs, are found relatively less vulnerable, suggesting that the potential damage

associated with such rising water level can be trivial/insignificant. The increased vulnerabilities at

the immediate upstream (NY Intake) and downstream (American Falls) of the Niagara Falls with

respect to the seasonal boundaries suggest the possibility of large elevation fluctuations. The

results at these locations can also benefit from improved data discretization. The increased

vulnerability at Ashland Ave. under the climate scenarios reinforces previous assessments that

reject this location to be the highly vulnerable under the baseline. Similar analysis when performed

for winter results in large elevation variations and subsequent increase in vulnerability at Lake

Ontario under all climate scenarios, particularly OSU. However, a few locations such as Buffalo

and GIP may benefit from a slightly reduced vulnerability under these scenarios.

9.8.4 Flow conditions The systems vulnerability under changing flow conditions shows a rather misleading outcome as

the inherent calculations multiply the probability of the most severe event in the sojourn by the

maximum damage. Though, the damage associated with the most severe event increases

substantially under all climate scenarios (for example, maximum severity under CCC-GCM2 is

1498, while the same for the baseline is 1014), the overall vulnerability reduces when aggregated

over the year. The alarming outcome that is perhaps not represented by the misleading

vulnerability numbers is that both the occurrences and the severity where the system violates the

constituent standard increases under a varying climate.

9.9 Sensitivity Analysis In this section, the authors evaluate the sensitivity of the hydro-climatic model by repeating the

analysis using a nonuniform bin size for the hydrologic variables. The water level and flow data

are discretized based on their continuous distribution where the top and bottom intervals each has

15% of the total instances (leaving 70% in the middle interval). The exact climate evidences are

introduced to the model and the results are analyzed for reliability, resilience and vulnerability.

When compared, the two models with uniform and distribution-based discretization show quite

similar results with an exception at Ashland Avenue. While the reliability values show a maximum

174

3% variation at Lake Ontario, the differences in other locations are nonexistent (Table 9.4).

Comparison between resiliencies indicates an improved system performance in the latter case

(with uneven bin sizes) with a maximum discrepancy of 0.02 at Olcott station. However, the

stations adjacent to the falls (American Falls and Ashland Avenue) become more vulnerable with

large deviations from the historical normals. This can be explained by the fact that the class

intervals under nonuniform discretization are pushed further apart, thus resulting in greater

differences between the intervals and the normals. Note that, Ashland Avenue under the

nonuniform discretization is found to be more reliable and resistant to deviations, most plausibly

due to improved predictions that result in fewer cases in the top and bottom intervals.

Table 9.4: Comparison between uniform and distribution-based discretization model under OSU

scenario

Reliability (low) Buffalo Niagara Intake

American Falls

Ashland Ave. Olcott

Reliability 0 0 0 0 0.55

Reliability_nonuniform 0 0 0 0.99 0.58

Resilience 0 0 0 0 0.06

Resilience_nonuniform 0 0 0 1 0.01

Vulnerability 0.13 0.2 1.04 0.41 2.33

Vulnerability_nonuniform 0.13 0.2 2.52 1.40 0.05

9.10 Conclusion The severe perturbations that might occur, or are occurring, within watersheds across the globe as

a result of climate change and anthropogenic disturbances provide at least two distinct challenges.

First, understanding the possible impact and assessing the resulting change in resilience require

process-based modelling with clear structural relationships between hydrologic and climate

models. However, the structural complexity and outcome uncertainties associated with these

models have led to their limited use. Second, although the watersheds are often closely linked such

that the impacts cascade downstream, there remains a disconnect between the observed trends and

near-term decisions by resource managers and policy makers. In the absence of a process-level

understanding of key relationships (Clark et al. 2015; Fatichi et al. 2016), this study provides an

approach that can shed light on the changing watershed resilience in the wake of climate change.

175

The analysis presented here combines the Bayesian Belief Net (BBN) with risk-based performance

indicators for estimating the changing reliability, resilience and vulnerability of the Niagara river

basin. Though the methodology offers certain benefits such as predicting several variables at once,

easy updating, handling missing data, etc. that dominates the decision of choosing BN in

combination with RRV metric as the preferred approach, such practice is not free from limitations.

First, being a probabilistic model, BN does not provide any statistical significance. Second, when

the model structure is unknown, the relationships among different variables are explained using

structural learning methods based on conditional Independence. As such, the model does not infer

causal relationships and may over-simplify the real-world phenomena. This shortcoming may

enlarge the uncertainty of a simulated outcome when combining climate change and hydrological

responses into one integrated model. Third, a typical problem faced during model development is

the intensive resource and time required for data collection, processing and structural learning. The

run-time also increases in proportion to the increased data discretization. Fourth, the model here

does not consider the effect of changing soil moisture, as there is a high degree of uncertainty

regarding how this parameter will change under a varying climate. Because of the upstream

hydropower plant, elevations at Ashland Ave. are more sensitive to plant operations than to a

changing climate. Hence, results at this location should be carefully examined and verified to

ensure their representativeness. The other gauges are chosen such that they are either upstream or

further downstream from the power station to avoid/minimize impact of the discharge decisions.

Finally, due to lack of adequate information on changing wind speed, ice coverage, lake

temperature etc., the model did not provide evidences for these variables under the climate

scenarios. However, the current model allows the authors to introduce new evidences once they

become available.

176

Conclusions and Recommendations Being a pivotal part of the economy, the North American electrical power system has been

historically guided by cost control and increasing demand. This thesis takes a step back in order to

gain a comprehensive view of the past developments as well as existing assets for better alignment

to the changing needs and priorities. This work fundamentally recommends increased

hydroelectric generation for a flexible and reliable grid and consider the technical, economic and

ecological dimensions associated with such development. Overall, this work explores some of the

complexities of these systems, and addresses emerging issues from a systems perspective by

analyzing innovative approaches and their trade-offs using a variety of techniques such as

optimization, simulation, decision support tool, investment and probabilistic graphical modelling.

The proposed tools were envisioned to address the changing needs of Ontario power system. The

issues covered in them include the unexplored potential, climate risk and the resulting changes in

systems resiliency, unfavourable market, disconnection between the needs and the incentives,

stakeholders’ participation and sustainable development. Nonetheless, the suggested tools cannot

address all the issues related to hydropower, rather they attempt to shift perspectives and to address

important issues systemically and sustainably.

A common dilemma faced by modellers is to determine an appropriate scale that captures sufficient

detail while limiting resource exploitation. The problem with selecting such a scale is, if too fine,

the available resources are exhausted whereas too coarse a scale leads to missing details in the

analysis. To address this universal modelling tension, the thesis first seeks a somewhat panoramic

view of the existing assets, followed by discussions on Ontario’s evolving energy policy landscape

(chapter 2) and the market conditions (chapter 3 and 4). Much attention is paid to the payment

structures for hydropower schemes, analyzed through the application of financial models and

optimization technique. Starting from this extended perspective, the focus then narrows down to

location-specific models targeted towards increasing dispatchable hydro capacity close to the load

centers (chapter 5 and 6). These central chapters set up a comprehensive 1D simulation model for

the Niagara power system and use it to explore innovative approaches that at times challenge the

traditional realm of policy. The shift from optimization (chapter 3) is motivated by the detailed

representation of systems under simulation technique that allows a more refined prediction of

system behavior. Comparison between the (optimization and simulation) results show a moderate

deviation since the optimization model assumes a fixed generation at the run-of-the-river plants,

177

while those are rather accurately modelled using the simulation process. Instead of terminating the

exploration at this point, as is traditionally done, the study establishes a more general formulation

in chapter 8 and 9, accounting for the issues that were previously overlooked due to its strong focus

on economic and technical considerations. While the sSWOT framework draws from the previous

models (chapter 3, 4, 5 and 6) and extends the analysis to include the environmental and social

considerations, the probabilistic graphical model in chapter 9 incorporates the inherent uncertainty

associated with climate projections into the analysis of systems resilience. The projected elevations

by the simulation model show a relatively high degree of convergence – 78 and 100% for Lake

Erie and Lake Ontario, respectively – with that from the Bayesian (probabilistic graphical) model.

The research here argues that there are merits of transitions that begin with an extended outlook,

followed by a specific focus, and finally adding layers of complexities to it in order to capture

holistic view. The range of modelling techniques used here, starting from optimization, simulation

to decision support tools and probabilistic graphical models, each contributes towards extending

the system boundaries, thus allowing progressively more elements to be included into the analysis.

Thus, the thesis advocates for alternating between these (broad-narrow-broad) perspectives within

a systems approach for future modelling applications.

The thesis uses a systems approach for the development of models and tools taking into account

the complex interactions between economy, energy, environment and policy and tracing the

trajectory of these trade-offs with changing circumstances and subsequent shift in priorities.

Specific contributions pertinent to each of the chapters are described below.

1. Chapter 2 provides a narrative for the growth of hydroelectric power in Ontario in the

backdrop of historical events and major energy transitions. The discussion explores the

potentials for increased hydropower generation in the province and the impacts of related

policies which may provide useful insights to researchers, energy developers and policy

makers. The analysis also reveals a lack of consensus regarding the replacement or

rehabilitation of the aging hydro infrastructure in Ontario. Subsequent chapters are devoted

to evaluating the technical, ecological and economic viability of the proposed alternatives.

2. Chapter 3 illustrates a direct optimization approach for a pumped storage operating in the

Ontario wholesale market given well-forecasted flows and energy price. The diurnal price

178

variations and pumping-generating cycle are found to be the dominant factors in PHS

profitability. The analysis further suggests that Ontario’s wholesale market price does not

justify storage investments purely from an economic perspective, as the HOEP bears little

to no relation to the cost of building new capacities. The proposed model can be used by

power authorities to compare between the emerging storage options.

3. Considering the limited incentives under marginal cost-based operation, chapter 4 analyzes

various supporting mechanisms under which PHS projects could be developed, integrated

and supported by renewable generation in Ontario. One of the major findings here is that

the project profitability is highly sensitive to capacity factors and payments (per MWh)

offered under the fixed contract. In this context, the study recommends a tiered

remuneration approach based on annual PHS contribution to the grid. The study shows that

a capacity based pricing subjected to periodic revision splits the risk between consumers

and investors, thus creating a balanced market for PHS investment.

4. Chapters 5 and 6 set up a comprehensive model for the Niagara power system and use it to

explore a variety of possible future scenarios. It contributes to the existing knowledge base

by evaluating the impacts of climate change on hydropower generation potential and

extending it to investigate ambitious regulation plans involving reduced tourist flows and

revised daily management using lake storage. The relaxed flow restrictions and increased

lake storage scenario augment/shift the generation by a maximum of 16% and 30%

respectively. Nevertheless, the generation potential reduces by a maximum of 8% over the

next 40 years and 9‒30% in long run under a doubling CO2 concentration scenario.

5. Chapter 7 compares different definitions of sustainability and argues that the conflicting

objectives associated with sustainable hydropower development call for integrated

approaches. It summarizes the existing state of play relating to how sustainability is

assessed. The discussion briefly documents research methods and points out limitations in

the existing approaches. Two recommendations advance the current sustainability

assessment: first, that such assessments should reflect on major environmental challenges

and broad policy issues with respect to existing hydropower potential and, second, that

179

system boundaries should be extended to allow reasonable estimation of hydro benefits on

the overall grid.

6. Recognizing the need for an integrated approach, chapter 8 proposes an improved decision-

support framework for analyzing resource systems from sustainability perspective and

applies it for assessing the hydropower potential at Niagara. The proposed model is

designed to drive action and collaboration on environmental challenges, creating risks and

opportunities which otherwise may go unnoticed. Based on the survey responses,

renegotiation of the treaty is favoured over the current restrictions given the climate change

and future energy scenarios. These findings call for a detailed investigation addressing the

need for more flexible treaty arrangements to permit periodic adjustments.

7. The final development uses a probabilistic graphical approach that allows considerations

of uncertainty associated with climate projections. It introduces a systematic approach to

estimate the changing watershed resilience through the combined application of Bayesian

Network (BN) and systems performance criteria. The analysis predicts a considerable

reduction in reliability and resiliency for the Niagara river basin under a changing climate.

Interestingly though, some sections of the river rarely benefit from increased resilience and

reduced vulnerability under these scenarios.

10.1 Future Research The thesis has proposed various approaches and tools and applied them to evaluate increased

hydropower generation in Ontario. In order to further validate the suggested approaches, future

research should apply the tools to other examples and case studies. The challenges encountered

while application, limitations of the approaches, and persisting questions have also inspired

potential extensions to the present research.

1. The profit values reported in chapter 3 are rough estimates and are dependent on the

persistence of similar flow conditions. Future research should investigate their variability

and the resulting changes in the profit characteristics.

180

2. The socioeconomic cost-benefit model presented in chapter 4 suggests pumped storage to

be cost-ineffective at low wind exploitations when replacing natural gas combined cycle

plant. However, such conclusions may be revised with consideration of fugitive methane

release, often associated with natural gas extraction.

3. The hydrologic routing (Muskingum-Cung) used in the Niagara Power System Simulation

(NPSS) model, elaborated in chapter 5 and 6, could be traded with a hydrodynamic

approach given the required data, computation time and resources. In this approach, the

governing equations are discretized and solved through numerical algorithm, primarily

finite difference method. Hydrodynamic modelling accomplished through skillful

development combined with understanding of the physical system is shown to be accurate

for a wide range of coastal processes. Future studies can extend the analysis by integrating

a water balance model that traces the resulting changes in runoff under a varying climate.

4. The data processing for the BN model in chapter 9 was found to be rather time consuming

since the information was extracted from different sources. Future work can seek to

standardize data integration in order to build more interoperable databases. The current

study neither accounts for soil moisture nor does it provide evidences for the change in ice

coverage and lake surface temperatures. Given the progress in regionals climate models

and the resulting projections, the model can be updated to include new variables and

evidences under the scenarios. An interesting extension of the work would be assessing the

impact of wind setup, i.e., vertical rise in the still-water level caused by wind stresses.

Given Ontario’s recent changes in precipitation patterns and the subsequent increase in

lake levels, there is a need to reassess the resulting impact on hydropower potential.

In addition to applying and validating the suggested approaches and tools, future research should

explore the benefits of juggling between a relatively broad and narrow perspective while planning

and reviewing the existing system design through the lens of models.

181

References Ackere, A., and Ochoa, P. (2010). Managing a Hydro-Energy Reservoir: A Policy Approach.

Energy Policy 38(11): 7299–7311.Ackoff, R. L. (1961). Progress in Operations Research, vol. 1, Wiley, New York, USA.

Afgan, N. H., and Carvalho, M. G. (2002). Multi-Criteria Assessment of New and Renewable Energy Power Plants. Energy 27: 739-755.

Afgan, N. H., Carvalho, M. G., and Hovanov, N. V. (2000). Energy System Assessment with Sustainability Indicators. Energy Policy 28: 603–612.

Afshar, A., Shafii, M., and Haddad, O. B. (2010). Optimizing Multi-Reservoir Operation Rules: An Improved HBMO Approach. Journal of Hydroinformatics 13: 121–139.

Afshar, M. H. (2012). Large Scale Reservoir Operation by Constrained Particle Swarm Optimization Algorithms. Journal of Hydro-environment Research 6: 75–87.

Aggarwal, S. K., Saini, L. M., and Kumar, A. (2008). Electricity Price Forecasting in Ontario Electricity Market Using Wavelet Transform in Artificial Neural Network Based Model. International Journal of Control, Automation, and Systems 6(5): 639–650.

Akhtar, M., Ahmad, N., and Booij, M. J. (2008). The Impact of Climate Change on the Water Resources of Hindukush-Karakorum-Himalaya Region Under Different Glacier Coverage Scenarios. Journal of Hydrology 355(1-4): 148–163.

Alards-Tomalin, D., Ansons, T. L., Reich, T. C., Sakamoto, Y., Davie, R., Leboe-McGowan, J. P., and Leboe-McGowa, L. C. (2014). Airport Security Measures and Their Influence on Enplanement Intentions: Responses from Leisure Travelers Attending a Canadian University. J. Air Transp. Manag. 37: 60‒68.

Aliyu, A. S., Dada, J. O., and Adam, I. K. (2015). Current Status and Future Prospects of Renewable Energy in Nigeria. Renewable and Sustainable Energy Reviews 48: 336–346.

Alonso, J. A., and Lamata, M. T. (2006). Consistency in the Analytic Hierarchy Process: A New Approach. International Journal of Uncertainty 14: 445–449.

Altieri, M. A. (1987). Agroecology: The Scientific Basis of Alternative Agriculture. Westview Press.

Amin, F. (2012). Ontario’s Feed-in Tariff Program Two-Year Review Report, Ontario Ministry of Energy.

Amor, M. Ben, Villemeur, E. B. de, Pellat, M., and Pineau, P. (2014). Influence of Wind Power on Hourly Electricity Prices and GHG (greenhouse gas) Emissions: Evidence That Congestion Matters from Ontario Zonal Data. Energy 66: 458–469.

182

Anagnostopoulos, J. S. Ã., and Papantonis, D. E. (2008). Simulation and Size Optimization of a Pumped – Storage Power Plant for the Recovery of Wind-Farms Rejected Energy. Renewable Energy 33: 1685–1694.

Aragon, J. (2010). Vulnerability of Hydroelectric Plants in Tropical Basins with High Slope in Cost Lica. The 6th International Conference on Hydropower.

Arnell, N. W. (2004). Climate Change and Global Water Resources: SRES Emissions and Socio-Economic Scenarios. Global Environmental Change 14(1): 31–52.

Arnell, N. W., van Vuuren, D. P., and Isaac, M. (2011). The Implications of Climate Policy for the Impacts of Climate Change on Global Water Resources. Global Environmental Change 21(2): 592–603.

Arora, V. K., Seglenieks, F., Kouwen, N., and Soulis, E. (2001). Scaling Aspects of River Flow Routing. Hydrological Processes 15(3): 461–477.

Auditor General of Ontario (2011). 2011 Annual Report, Queen’s Printer for Ontario, Toronto.

Ayodele, T. R. and Ogunjuyigbe, A. S. O. (2015). Mitigation of Wind Power Intermittency: Storage Technology Approach. Renewable and Sustainable Energy Reviews 44: 447–456.

Azimi, R., Yazdani-Chamzini, A., Fouladgar, M. M., Zavadskas, E. K. and Basiri, M. H. (2011). Ranking the Strategies of Mining Sector Through ANP and TOPSIS in a SWOT Framework. Journal of Business Economics and Management 12(4): 670–689.

Azizi, A. and Maleki, R. (2014). Comparative Study of AHP and ANP on Multi-Automotive Suppliers with Multi-Criteria. International MultiConference of Engineers and Computer Scientists, Hong Kong.

Bakis, R., and Demirbas, A. (2004). Sustainable Development of Small Hydropower Plants (SHPs). Energy Sources 26: 1105–1118.

Balat, H. (2007). A Renewable Perspective for Sustainable Energy Development in Turkey: The Case of Small Hydropower Plants. Renewable and Sustainable Energy Reviews 11: 2152–2165.

Baláž, M., Danáčová, M., and Szolgay, J. (2010). On the Use of the Muskingum Method for the Simulation of Flood Wave Movements. Slovak J. of Civil Engineering 18(3): 14–20.

Ballester, C., and Furió, D. (2015). Effects of Renewables on the Stylized Facts of Electricity Prices. Renewable and Sustainable Energy Reviews 52: 1596–1609.

Barbour, E., Wilson, I. A. G., Radcliffe, J., Ding, Y., and Li, Y. (2016). A Review of Pumped Hydro Energy Storage Development in Significant International Electricity Markets. Renewable and Sustainable Energy Reviews 61: 421–432.

183

Barnett, T. P., and Pierce, D. W. (2009). Sustainable Water Deliveries from The Colorado River in a Changing Climate. Proceedings of the National Academy of Sciences of the United States of America 106(18): 7334–7338.

Baroud, H. and Barker, K. (2014). Bayesian Kernel Methods for Critical Infrastructure Resilience Modeling. Proceedings of Second International Conference on Vulnerability and Risk Analysis and Management, Liverpool, UK.

Bates, B., Kundzewicz, Z. W., and Shaohong W. (2008). Climate Change and Water, Intergovernmental Panel on Climate Change (IPCC), IPCC Secretariat, Geneva.

Baycheva-Merger, T., and Wolfslehner, B. (2016). Evaluating the Implementation of the Pan-European Criteria and Indicators for Sustainable Forest Management – A SWOT Analysis. Ecological Indicators 60: 1192–1199.

BC Hydro (2011). Standing Offer Program: Report on the SOP 2-Year Review, Government of British Columbia.

Bekele, E., and Knapp, H. (2012). Evolutionary Computation for Simulating Reservoir Release Rates of Lake Shelbyville and Carlyle Lake. World Environmental and Water Resources Congress, E. D. Loucks, ed., New Mexico, USA.

Bell, R. G., and Russell, C. (2002). Environmental Policy for Developing Countries. Issues in Science and Technology 18(3).

Benitez, L. E., Benitez, P. C. and Kooten, G. C. (2008). The Economics of Wind Power with Energy Storage. Energy Economics 30: 1973–1989.

Berke, P. R. (1995). Natural-Hazard Reduction and Sustainable Development: A Global Assessment. Journal of Planning Literature 9(4): 370–382.

Berke, P. R., and Conroy, M. M. (2000). Are We Planning for Sustainable Development? Journal of the American Planning Association 66 (1): 21–33.

Biemans, H., Speelman, L. H., Ludwig, F., Moors, E. J., Wiltshire, A. J., Kumar, P., Gerten, D., and Kabat, P. (2013). Future Water Resources for Food Production in Five South Asian River Basins and Potential for Adaptation ‒ A Modeling Study. Science of the Total Environment 468-469: S117–S131.

Biggar, E. B. (1920). Hydro-Electric Development in Ontario: A History of Water-Power Administration under the Hydro-Electric Power Commission of Ontario, Ryerson Press, Toronto.

Black and Veatch. (2012). Cost Report Cost and Performance Data for Power Generation.

Boer, G. J., McFarlane, N. A., and Lazare, M. (1992). Greenhouse Gas Induced Climate Change Simulated with The CCC Second-Generation General Circulation Model. Journal of Climate 5(10): 1045–1077.

184

Boland, B. (2013). <http://wpweb2.tepper.cmu.edu/ceic/seminarPDFs/Optimizing_Ontario_Generation_ April_24_ 2013.pdf> (Jul. 06, 2015).

Bolsenga, S. J., and Norton, D. C. (1993). Great Lakes Air Temperature Trends for Land Stations, 1901-1987. Journal of Great Lakes Research 19(2): 379–388.

Bonham, C., Edmonds, C. and Mark, J. (2006). The Impact of 9/11 and other Terrible Global Events on Tourism in The United States and Hawaii. Journal of Travel Research 45: 99‒110.

Bosona, T. G., and Gebresenbet, G. (2010). Modeling Hydropower Plant System to Improve Its Reservoir Operation. Journal of Water Resources and Environmental Engineering: 2(4): 87–94.

Botter, G., Basso, S., Rodriguez-Iturbe, I., and Rinaldo, A. (2013). Resilience of River Flow Regimes. Proceedings of the National Academy of Sciences 110(32): 12925–12930.

Bou-Zeid, E., and El-Fadel, M. (2002). Climate Change and Water Resources in Lebanon and the Middle East. Journal of water resources planning and Management 128(5): 343–355.

Brahim, S. P. (2014). Renewable Energy and Energy Security in The Philippines. Energy Procedia 52: 480–486.

Branfireun, B. A., Roulet, N. T., Kelly, A. I., and Rudd, J. W. M. (1999). In situ Sulphate Stimulation of Mercury Methylation in a Boreal Peatland: Toward a Link Between Acid Rain and Methylmercury Contamination in Remote Environments. Global Biogeochemical Cycles 13(3): 743–750.

Bratrich, C., and Truffer, B. (2001). Green Electricity Certification for Hydropower Plants - Concept, Procedure, Criteria, Green Power Publications.

Braun, S. (2016). Hydropower Storage Optimization Considering Spot and Intraday Auction Market. Energy Procedia 87: 36–44.

Bruce Power (2017). <http://www.brucepower.com/about-us/history/> (Jun. 01, 2017).

British Petroleum Company. (2015). BP Statistical Review of World Energy. British Petroleum Co., London.

Brown, M. A., Wang, Y., Sovacool, B. K., and D’Agostino, A. L. (2014). Forty Years of Energy Security Trends: A Comparative Assessment of 22 Industrialized Countries. Energy Research & Social Science 4: 64–77.

Bueno, C., and Carta, J. A. (2006). Wind Powered Pumped Hydro Storage Systems, a Means of Increasing the Penetration of Renewable Energy in the Canary Islands. Renewable and Sustainable Energy Reviews 10: 312–340.

185

Bull, J. W., Jobstvogt, N., Böhnke-Henrichs, A., Mascarenhas, A., Sitas, N., Baulcomb, C., Lambini, C. K., Rawlins, M., Baral, H., Zähringer, J., Carter-Silk, E., Balzan, M. V., Kenter, J. Häyhä, O. T., Petz, K., and Koss, R. (2016). Strengths, Weaknesses, Opportunities and Threats: A SWOT Analysis of the Ecosystem Services Framework. Ecosyst. Services 17: 99–111.

Butler, L., and Neuhoff, K. (2008). Comparison of Feed-In Tariff, Quota and Auction Mechanisms to Support Wind Power Development. Renewable Energy 33: 1854–1867.

Buttle, J., Muir, T. and Frain, J. (2004). Economic Impacts of Climate Change on the Canadian Great Lakes Hydro–Electric Power Producers: A Supply Analysis. Canadian Water Resources Journal 29: 89-110.

Caldeira, K., Jain, A. K., and Hoffert, M. I. (2003). Climate Sensitivity Uncertainty and The Need for Energy Without CO2 Emission. Science 299: 2052–2054.

Campbell, S. (1996). Green Cities, Growing Cities, Just Cities? Urban Planning and the Contradictions of Sustainable Development. Journal of the American Planning Association 62(3): 296–312.

Canadian Electricity Association. (2013). <http://www.electricity.ca/media/Presentations/Electric_Vehicle_and_Infrastructure_Summit_Feb2013.pdf> (Jul. 03, 2015).

Canadian Electricity Association. (2014a). Canada’s Electricity Industry, <http://www.electricity.ca/media/Electricity101/Electricity101.pdf> (Sept. 08, 2015).

Canadian Electricity Association. (2014b). Key Canadian Electricity Statistics, <http://www.electricity.ca/media/Electricity101/KeyCanadianElectricityStatistics10June2014.pdf> (Sep. 14, 2015).

Canales, F. A., Beluco, A., and Mendes, C. A. B. (2015). A Comparative Study of a Wind Hydro Hybrid System with Water Storage Capacity: Conventional Reservoir or Pumped Storage Plant? Journal of Energy Storage 4: 96–105.

Canu, D. M., Campostrini, P., Riva, S. D., Pastres, R., Pizzo, L., Rossetto, L. and Solidoro, C. (2011). Addressing Sustainability of Clam Farming in the Venice Lagoon. Ecology and Society 16(3): 26.

Capik, M., Yılmaz, A. O., and Cavusoglu, İ. (2012). Hydropower for Sustainable Energy Development in Turkey: The Small Hydropower Case of the Eastern Black Sea Region. Renewable and Sustainable Energy Reviews 16(8): 6160–6172.

Caralis, G., Papantonis, D., and Zervos, A. (2012). The Role of Pumped Storage Systems Towards the Large Scale Wind Integration in the Greek Power Supply System. Renewable and Sustainable Energy Reviews 16: 2558–2565.

186

Caralis, G., Rados, K., and Zervos, A. (2010). On the Market of Wind with Hydro-Pumped Storage Systems in Autonomous Greek Islands. Renewable and Sustainable Energy Reviews 14(8): 2221–2226.

Carbon Tax Center (2017). <https://www.carbontax.org/where-carbon-is-taxed/british-columbia/> (Dec. 08, 2016).

Carrera, D. G., and Mack, A. (2010). Sustainability Assessment of Energy Technologies via Social Indicators: Results of a Survey Among European Energy Experts. Energy Policy 38: 1030–1039.

Case, P. (2004). Plume Movement at Niagara Falls: A Bridge to Other Natural Systems. M.A. Thesis at The State University of New York.

Castelletti, A., Yajima, H., Giuliani, M., Soncini-Sessa, R., and Weber, E. (2014). Planning the Optimal Operation of a Multioutlet Water Reservoir with Water Quality and Quantity Targets. Water Resources Planning and Management, 10.1061/(ASCE)WR.1943-5452.0000348.

Catalão, J. P. S., Pousinho, H. M. I., and Contreras, J. (2012). Optimal Hydro Scheduling and Offering Strategies Considering Price Uncertainty and Risk Management. Energy 37: 237–244.

Catron, J., Stainback, G. A., Dwivedi, P., and Lhotka, J. M. (2013). Bioenergy Development in Kentucky: A SWOT-ANP Analysis. Forest Policy and Economics 28: 38–43.

Chang, M. K., Eichman, J. D., Mueller, F., and Samuelsen, S. (2013). Buffering Intermittent Renewable Power with Hydroelectric Generation: A Case Study in California. Applied Energy 112: 1–11.

Chang, S. E., and Shinozuka, M. (2004). Measuring Improvements in the Disaster Resilience of Communities. Earthquake Spectra 20(3): 739–755.

Charlton, M. B., and Arnell, N. W. (2011). Adapting to Climate Change Impacts on Water Resources in England - An Assessment of Draft Water Resources Management Plans. Global Environmental Change 21(1): 238–248.

Chatzimouratidis, A. I., and Pilavachi, P. A. (2008). Multicriteria Evaluation of Power Plants Impact on the Living Standard Using the Analytic Hierarchy Process. Energy Policy 36: 1074–1089.

Chen, H., Xu, C. Y., and Guo, S. (2012). Comparison and Evaluation of Multiple GCMs, Statistical Downscaling and Hydrological Models in the Study of Climate Change Impacts on Runoff. Journal of Hydrology 434-435: 36–45.

Chen, S., Chen, B., and Fath, B. D. (2015). Assessing the Cumulative Environmental Impact of Hydropower Construction on River Systems Based on Energy Network Model. Renewable and Sustainable Energy Reviews 42: 78–92.

187

Chen, W. M., Kim, H., and Yamaguchi, H. (2014). Renewable Energy in Eastern Asia: Renewable Energy Policy Review and Comparative SWOT Analysis for Promoting Renewable Energy in Japan, South Korea, and Taiwan. Energy Policy 74: 319–329.

Chiari, L., and Zecca, A. (2011). Constraints of Fossil Fuels Depletion on Global Warming Projections. Energy Policy 39: 5026–5034.

Chow, V. T. (1959). Open-Channel Hydraulics, McGraw-Hill, New York.

Christensen, N. S., and Lettenmaier, D. P. (2007). A Multimodel Ensemble Approach to Assessment of Climate Change Impacts on the Hydrology and Water Resources of the Colorado River Basin. Hydrol. Earth Syst. Sci 11: 1417–1434.

Cinar, D., and Kayakutlu, G. (2010). Scenario Analysis Using Bayesian Networks: A Case Study in Energy Sector. Knowledge-Based Systems 23(3): 267–276.

City of Niagara Falls. (2014). <http://www.niagarafalls.ca/city-hall/business> (Oct. 26, 2014).

Clark, M. P., Fan, Y., Lawrence, D. M., Adam, J. C., Bolster, D., Gochis, D. J., Hooper, R. P., Kumar, M., Leung, L. R., Mackay, D. S., and Maxwell, R. M. (2015). Improving the Representation of Hydrologic Processes in Earth System Models. Water Resources Research: 5929–5956.

Canadian Wind Energy Association (2011). The Economic Impacts of the Wind Energy Sector in Ontario 2011-2018. <http://canwea.ca/pdf/economic_impacts_wind_energy_ontario2011-2018.pdf> (May 31, 2017).

Clites, A. H., and Lee, D. H. (1998). MIDLAKES: A Coordinated Hydrologic Response Model for the Middle Great Lakes, Great Lakes Environmental Research Laboratory, Michigan, USA.

Coe, M. T., Costa, M. H., Botta, A., and Birkett, C. (2002). Long-Term Simulations of Discharge and Floods in the Amazon Basin. Journal of Geophysical Research Atmospheres 107(20): 1–17.

Cohen, S., Neilsen, D., Smith, S., Neale, T., Taylor, B., Barton, M., Merritt, W., Alila, Y., Shepherd, P., McNeill, R., Tansey, J., Carmichael, J., and Langsdale, S. (2006). Learning with Local Help: Expanding the Dialogue on Climate Change and Water Management in the Okanagan Region, British Columbia, Canada. Climatic Change 75(3): 331–358.

Collen, Z., Blanchard, R. E., Eames, P. C., Juma, A. M., Chitawo, M. A., and Gondwe, K. T. (2014). Overview of the Malawi Energy Situation and A PESTLE Analysis for Sustainable Development of Renewable Energy. Renew. and Sustainable Energy Reviews 38: 335–347.

Connolly, D., Lund, H., Finn, P., Mathiesen, B. V, and Leahy, M. (2011). Practical Operation Strategies for Pumped Hydroelectric Energy Storage (PHES) Utilising Electricity Price Arbitrage. Energy Policy 39(7): 4189–4196.

188

Constantinou, A. C., Fenton, N., Marsh, W., and Radlinski, L. (2015). From Complex Questionnaire and Interviewing Data to Intelligent Bayesian Network Models for Medical Decision Support. Artificial Intelligence in Medicine 67: 75–93.

Cosens, B. A., and Williams, M. K. (2012). Resilience and Water Governance: Adaptive Governance in the Columbia River Basin. Ecology & Society 17(4): 128–141.

Cosi, B., Markovska, N., and Kraja, G. (2012). Environmental and Economic Aspects of Higher RES Penetration into Macedonian Power System. Applied Thermal Engineering 43: 158–162.

Coupe, V. M. H. (2016). Sensitivity Analysis: An Aid for Belief-Network Quantification. The Knowledge Engineering Review 15(3): 215–232.

Couture, T., and Gagnon, Y. (2010). An Analysis of Feed-In Tariff Remuneration Models: Implications for Renewable Energy Investment. Energy Policy 38(2): 955–965.

Crissman, R. D. (1989). Impacts on Electricity Generation in New York State. First U.S.-Canada Symposium on Impacts of Climate Change on the Great Lake Basin, Sept. 27-29, Illinois.

Crissman, R. D., Chiu, C., Yu, W., Mizumura, K. and Corbu, I. (1993). Uncertainties in Flow Modeling and Forecasting for Niagara River. J. Hydraul. Eng. 119(11): 1231-1250.

Croley, T. E. (1990). Laurentian Great Lakes Double-CO2 Climate Change Hydrological Impacts. Climatic Change 17: 27–47.

Croley, T. E., II, Quinn, F. H., Kunkel, K. and Changnon, S. J. (1995). Potential Great Lakes Hydrology and Lake Level Impacts Resulting from Global Warming. The Sixth Conference on Global Change Studies, Boston, Massachusetts.

Croley, T. E., II. (1993). CCC GCM 2×CO2 Hydrological Impacts on the Great Lakes, Climate, Climate Change, Water Level Forecasting and Frequency Analysis, International Joint Commission (IJC) Levels Reference Study.

Croley, T. E., II. (2003). Great Lakes Climate Change Hydrologic Impact Assessment. IJC Lake Ontario – St. Lawrence River Regulation Study. NOAA Technical MemoranduGLERL-126.

Cunge, J. A. (1969). On the Subject of a Flood Propagation Computation Method (Muskingum Method). Journal of Hydraulic Research 7(2): 205–230.

Dames and Moore. (1981). An Assessment of Hydroelectric Pumped Storage, The U.S. Army Engineer, Washington D.C.

Davidson, E. A., Araújo, A. C., Artaxo, P., Balch, J. K., Brown, I. F., C. Bustamante, M. M., Coe, M. T., DeFries, R. S., Keller, M., Longo, M., Munger, J. W., Schroeder, W., Soares-Filho, B. S., Souza, C. M., and Wofsy, S. C. (2012). The Amazon Basin in Transition. Nature 481(7381): 321–328.

189

Davies, E. G. R., and Simonovic, S. P. (2011). Global Water Resources Modeling with an Integrated Model of the Social-Economic-Environmental System. Advances in Water Resources 34(6): 684–700.

Deane, J. P., Gallacho´ir, B. P. O., and Mckeogh, E. J. (2010). Techno-Economic Review of Existing and New Pumped Hydro Energy Storage Plant. Renewable and Sustainable Energy Reviews 14: 1293–1302.

Dessai, S., and Hulme, M. (2004). Does Climate Adaptation Policy Need Probabilities? Climate Policy 4: 107–128.

Devito, K. J., Hill, A. R., and Dillon, P. J. (1999). Episodic Sulphate Export from Wetlands in Acidified Headwater Catchments: Prediction at the Landscape Scale. Biogeochemistry 44(2): 187–203.

Devkota, L. P., and Gyawali, D. R. (2015). Impacts of Climate Change on Hydrological Regime and Water Resources Management of the Koshi River Basin, Nepal. Journal of Hydrology: Regional Studies 4: 502–515.

Dewees, D. N. (2013). Renewable Electricity Policy in Ontario. <https://www.economics.utoronto.ca/public/workingPapers/tecipa-478.pdf> (June 3, 2015).

Dio, V. D., Favuzza, S., Cascia, D. L., Massaro, F., and Zizzo, G. (2015). Critical Assessment of Support for the Evolution of Photovoltaics and Feed-In Tariff(s) in Italy. Sustainable Energy Technologies and Assessments 9: 95–104.

Díaz-González, F., Sumper, A., Gomis-Bellmunt, O., and Villafáfila-Robles, R. (2012). A Review of Energy Storage Technologies for Wind Power Applications. Renewable and Sustainable Energy Reviews 16(4): 2154–2171.

Doll, P. (2002). Impact of Climate Change and Variability on Irrigation Requirements: A Global Perspective. Climatic Change 54(3): 269–293.

Dooge, J. C. I., Strupczewski, W. G. and Napiorkowski, J. J. (1982). Hydrodynamic Derivation of Storage Parameters for Muskingum Model. Journal of Hydrology 54: 371–387.

Douglass, G. K. (1984). The Meanings of Agricultural Sustainability, The Westview Press.

Dubromelle, Y., Louati, T., Ounnar, F. and Pujo, P. (2010). AHP/ANP a Decision Making Service in PROSIS Model. IFAC Proceedings Volumes 43(4): 138–143.

Duic´, N., Krajac, G., and Carvalho, M. da G. (2008). RenewIslands Methodology for Sustainable Energy and Resource Planning for Islands. Renewable and Sustainable Energy Reviews 12: 1032–1062.

Dursun, B., and Alboyaci, B. (2010). The Contribution of Wind-Hydro Pumped Storage Systems in Meeting Turkey ’S Electric Energy Demand. Renewable and Sustainable Energy Reviews 14(7): 1979–1988.

190

EIA (2015). Weighted Average Cost of Fossil Fuels for the Electric Power Industry, 2005 through 2015, <http://www.eia.gov/electricity/annual/html/epa_07_04.html> (Dec. 11, 2016).

Eichert, B. S. and Davis, D. W. (1976). Sizing Flood Control Reservoir Systems by Systems Analysis. U.S. Army Corps Eng., Tech. Paper 44.

Elakanda, S. (2010). Impacts of Climate Change on Hydropower Generation in Sri Lanka. The 6th International Hydropower Conference on Hydropower.

Enders, C. K. (2006). A Primer on the Use of Modern Missing-Data Methods in Psychosomatic Medicine Research. Psychosomatic Medicine 68(3): 427–436.

Environment and Climate Change Canada (2016). <https://www.ec.gc.ca/eau-water/default.asp?lang=En&n=45BBB7B8-1> (Jul. 15, 2016).

Environment Canada (2015). <http://wateroffice.ec.gc.ca/search/searchResult_e.html> (Aug. 31, 2015).

Ergu, D., Kou, G., Shi, Y., and Shi, Y. (2014). Analytic Network Process in Risk Assessment and Decision Analysis. Comput. Oper. Res. 42: 58-74.

ESRI (2015). Free Geospatial Data of the Niagara Region, <https://www.brocku.ca/maplibrary/digital/Niagara/free/basemap.zip> (Aug. 18, 2015).

European Bank for Reconstruction and Development (EBRD). (2013). Mid-Size Sustainable Energy Financing Facility.

European Commission (2005). ExternE externalities of Energy, <http://www.externe.info/externe_2006/brussels/methup05a.pdf> (Dec. 15, 2016).

European Commission (2008). Cost Assessment for Sustainable Energy Systems, <http://www.feem-project.net/cases/links_databases.php> (Dec. 15, 2016).

European Commission. (2009). <http://iet.jrc.ec.europa.eu/remea/2009-technology-map-european-strategic-energy-technology-plan-set-plan> (Dec. 15, 2016).

Evans, A., Strezov, V., and Evans, T. J. (2009). Assessment of Sustainability Indicators for Renewable Energy Technologies. Renew. and Sustainable Energy Reviews 13: 1082–1088.

Evenden, M. (2013). World War as A Factor in Energy Transitions: The Case of Canadian Hydroelectricity. RCC Perspectives: Transformations in Environment and Society: 91–94.

Faber, B. A., and Harou, J. J. (2007). Multi-Objective Optimization of Reservoir Systems Using HEC-Resprm. The World Environmental and Water Resources Congress, Florida, USA.

Fagot, K. Bartel, J. M., Ashley, A., and Schmidt, M. (2011). Development of a Restricted Reservoir Operating Plan at West Point Dam, Georgia. The 31st Annual USSD Conference San Diego, California, USA.

191

Fagot, K., Ashley, A., Hathorn, J., Klipsch, J., Hu, H., and Eggers, D. (2012). Scripting of Rules in Hec-Ressim for the ACF and ACT Basins. The United States Society on Dams, USA.

Fatichi, S., Vivoni, E. R., Ogden, F. L., Ivanov, V. Y., Mirus, B., Gochis, D., Downer, C. W., Camporese, M., Davison, J. H., Ebel, B., Jones, N., Kim, J., Mascaro, G., Niswonger, R., Restrepo, P., Rigon, R., Shen, C., Sulis, M., and Tarboton, D. (2016). An Overview of Current Applications, Challenges, and Future Trends in Distributed Process-Based Models in Hydrology. Journal of Hydrology 537: 45–60.

Feldman, D., Barbose, G., Margolis, R., Wiser, R., Darghouth, N., and Goodrich, A. (2012). Photovoltaic (PV) pricing trends: Historical, recent, and near-term projections, US Department of Energy.

Fisher, W. F. (1995). Toward Sustainable Development?: Struggling Over India's Narmada River, M. E. Sharpe, New York, USA.

Fisheries and Ocean Canada (2016). Fluctuations in Lake Levels – Types, <http://www.waterlevels.gc.ca/C&A/fluctuations_e.html#TC14> (Jul. 15, 2016).

Fleten, S. E., and Kristoffersen, T. K. (2008). Short-Term Hydropower Production Planning by Stochastic Programming. Computers & Operations Research 35: 2656–2671.

Foley, A. M., Leahy, P. G., Li, K., McKeogh, E. J., and Morrison, A. P. (2015). A Long-Term Analysis of Pumped Hydro Storage to Firm Wind Power. Applied Energy 137, 638–648.

Folke, C., Carpenter, S. R., Walker, B., Scheffer, M., Chapin, T., and Rockström, J. (2010). Resilience Thinking: Integrating Resilience, Adaptability and Transformability. Ecology and Society 15(4): 20.

Folke, C., Hahn, T., Olsson, P., and Norberg, J. (2005). Adaptive Governance of Social-Ecological Systems. Annual Review of Environment and Resources 30(1): 441–473.

Foroughi, A., Rasoulian, M. and Esfahani, M. J. (2012). Prioritize Strategies of University by Using SWOT Analysis and ANP Method. American Journal of Scientific Research 46: 83–91.

Fouladgar, M. M., Yakhchali, S. H., Chamzini, A. Y. and Basiri, M. H. (2011). Evaluating the Strategies of Iranian Mining Sector Using an Integrated Model. International Conference on Financial Management and Economics: 58–63.

Fouquet, D., and Johansson, T. B. (2008). European Renewable Energy Policy at Crossroads ‒ Focus on Electricity Support Mechanisms. Energy Policy 36: 4079–4092.

Francis, R., and Bekera, B. (2014). A Metric and Frameworks for Resilience Analysis of Engineered and Infrastructure Systems. Reliability Engineering and System Safety 121: 90–103.

Frey, G. W., and Linke, D. M. (2002). Hydropower as a Renewable and Sustainable Energy Resource Meeting Global Energy Challenges in a Reasonable Way. Energy Policy 30: 1261–

192

1265.

Friesen, B. F., and Day, J. C. (1977). Hydroelectric Power and Scenic Provisions of The 1950 Niagara Treaty. Journal of the American Water Resources Association 13: 1175–1190.

Fritsche, U. (1992). TEMIS – A Computerized Tool for Energy and Environmental Fuel and Life Cycle Analysis ‒ Current Status and Perspectives. Expert Workshop on Lifecycle Analysis of Energy Systems Methods and Experience, Paris, France.

Fujihara, Y., Tanaka, K., Watanabe, T., Nagano, T., and Kojiri, T. (2008). Assessing the Impacts of Climate Change on the Water Resources of the Seyhan River Basin in Turkey: Use of Dynamically Downscaled Data for Hydrologic Simulations. Journal of Hydrology 353(1-2): 33–48.

Gagnon, L., Belanger, C., and Uchiyama, Y. (2002). Life-Cycle Assessment of Electricity Generation Options: The Status of Research in 2001. Energy Policy 30(14): 1267–1278.

Gagnon, L., van de Vate J. F. (1997). Greenhouse Gas Emissions from Hydropower. Energy Policy 25: 7–13.

Gallant, P. (2015). Ontario Surplus Power Exports Cost Consumers in January, <http://www.windconcernsontario.ca/ontario-surplus-power-exports-cost-consumers-in-january/> (Aug. 8, 2015).

Garrison, J. B., and Webber, M. E. (2011). An Integrated Energy Storage Scheme for A Dispatchable Solar and Wind Powered Energy System. Journal of Renew. and Sustainable Energy 3.

Georgakakos, A. P., Yao, H., Kistenmacher, M., Georgakakos, K. P., Graham, N. E., Cheng, F. Y., Spencer, C., and Shamir, E. (2012). Value of Adaptive Water Resources Management in Northern California under Climatic Variability and Change: Reservoir Management. Journal of Hydrology 412-413: 34–46.

Georgakellos, D. A. (2012). Climate Change External Cost Appraisal of Electricity Generation Systems from a Life Cycle Perspective: The Case of Greece. Journal of Cleaner Production 32: 124–140.

Georgilakis, P. S. (2008). Technical Challenges Associated with the Integration of Wind Power into Power Systems. Renewable and Sustainable Energy Reviews 12: 852–863.

Ghan, S. J. (1982). A Documentation of the OSU Two-level Atmospheric GCM, CRI Report 35, Oregon State University.

Ghodsypour, S. H., and O'Brien, C. (1988). A Decision Support System for Supplier Selection Using an Integrated Analytic Hierarchy Process and Linear Programming. International Journal of Production Economics 56: 199–212.

Global Water Partnership. (GWP) (2012). The Handbook for Integrated Water Resources Management in Transboundary Basins of Rivers, Lakes and Aquifers.

193

Godschalk, D. R. (2002). Urban Hazard Mitigation: Creating Resilient Cities. Urban Hazards Forum.

Gondwe, M. (2010). Aspects of Climate Change: Impacts on Generation—The Case of Malawi’s Runoff-River Hydropower Schemes. The 6th International Conference on Hydropower.

Görener, A. (2012). Comparing AHP and ANP: An Application of Strategic Decisions Making in a Manufacturing Company. International Journal of Business and Social Science 3(11): 194–208.

Government of Canada (2014). Small Hydropower, <https://www.nrcan.gc.ca/energy/renewable-electricity/small-hydropower/7363>

Government of Canada (2015). <http://www.treaty-accord.gc.ca/text-texte.aspx?id=100418> (Jun. 8, 2013).

Government of Canada (2008). Economic Scan of Canada’s Energy Sector. Ottawa: Government of Canada, <www.ec.gc.ca/Publications/default.asp?lang=En&xml=73EC7D79-5F34-4C4C-ADBE-66CAAFF5DAC2> (Oct. 01, 2015).

Government of Canada (2015). Treaty Between Canada and the United States of America Concerning the Diversion of the Niagara River, <http://www.treaty-accord.gc.ca/text-texte.aspx?id=100418> (Jun. 8, 2013).

Griffith, R. E., and Stewart, R. A. (1961). A Nonlinear Programming Technique for Optimization of Continuous Processing Systems. Management Science 7(3): 379–392.

Grošelj, P. and Stirn, L. Z. (2015). The Environmental Management Problem of Pohorje, Slovenia: A New Group Approach Within ANP – SWOT Framework. Journal of Environmental Management 161: 106–112.

Guittet, M., Capezzali, M., Gaudard, L., and Romerio, F. (2016). Study of the Drivers and Asset Management of Pumped-Storage Power Plants Historical and Geographical Perspective. Energy 111: 560–579.

Gula, J., and Peltier, W. R. (2012). Dynamical Downscaling Over the Great Lakes Basin of North America Using The WRF Regional Climate Model: The Impact of the Great Lakes System On Regional Greenhouse Warming. Journal of Climate 25: 7723–7742.

Gyamfi, S., Modjinou, M., and Djordjevic, S. (2015). Improving Electricity Supply Security in Ghana — The Potential of Renewable Energy. Renewable and Sustainable Energy Reviews 43: 1035–1045.

Haas, R., Eichhammer, W., Huber, C., Langniss, O., Lorenzoni, A., Madlener, R., Menanteau, P., Morthorst, P.E., Martins, A., Oniszk, A., Schleich, J., Smith, A., Vass, Z., and Verbruggen, A. (2004). How to Promote Renewable Energy Systems Successfully and Effectively. Energy Policy 32: 833–839.

194

Haguma, D., Leconte, R., Krau, S., Côté, P. and Brissette, F. (2010). Water Resources Optimization Method in the Context of Climate Change. Water Resources Planning and Management, 10.1061/(ASCE)WR.1943-5452.0000445.

Haimes, Y. Y. (2006). On the Definition of Vulnerabilities in Measuring Risks to Infrastructures. Risk Analysis 26(2): 293–296.

Haimes, Y. Y. (2009a). On the Definition of Resilience in Systems. Risk Analysis 29(4): 498–501.

Haimes, Y. Y. (2009b). On the Complex Definition of Risk: A Systems-Based Approach. Risk Analysis 29(12): 1647–1654.

Hamududu, B., and Killingtveit, A. (2012). Assessing Climate Change Impacts on Global Hydropower. Energies 5(2): 305–322.

Hansen, J., Russell, G., Rind, D., Stone, P., Lacis, A., Lebedeff, S., Ruedy, R., and Travis, L. (1983). Efficient Three-Dimensional Global Models for Climate Studies: Models I and II. Monthly Weather Review 111(4): 609–662.

Harris, C. N. P., Quinn, A. D., and Bridgeman, J. (2014). The Use of Probabilistic Weather Generator Information for Climate Change Adaptation in the UK Water Sector. Meteorological Applications 21(2): 129–140.

Hartmann, H. C. (1988). Potential Variation of Great Lakes Water Levles: A Hydrologic Response Analysis, NOAA GLERL Technical Memorandum–68, Ann Arbor, Michigan.

Hartmann, H. C. (1990). Climate Change Impacts on Laurentian Great Lakes Levels. Climatic Change 17(1): 49–67.

Harvey, W. B. (2004). Economic Feasibility and Means for Financing Study of the Beck 3 Generating Station, Klohn Crippen Consultants Ltd. on the direction of Ministry of Energy.

Hashimoto, T., Stedinger, J. R., and Loucks, D. P. (1982). Reliability, Resiliency, and Vulnerability Criteria. Water Resources Research 18(1): 14–20.

HATCH. (2015). <https://www.hatch.com/Projects/Infrastructure/Niagara-Tunnel> (Jan. 5, 2016).

Hebb, A. J. and Morstch, L, D. (2005). Climate Change Scenarios – Lake Michigan-Huron, St.Clair and Erie. Huron-Erie Corridor / Lake St. Clair Research Needs Workshop, Windsor, Ontario.

Heagle, A. L. B., Naterer, G. F., and Pope, K. (2011). Small Wind Turbine Energy Policies for Residential and Small Business Usage in Ontario, Canada. Energy Policy 39: 1988–1999.

Helset, A., Gjelsvik, A., Mo, B., and Linnet, Ú. (2013). A Model for Optimal Scheduling of Hydro Thermal Systems Including Pumped-Storage and Wind Power. IET Generation, Transmission & Distribution 7(12): 1426–1434.

195

Herring, C. E. (2014). Evaluating Non-Indigenous Species Management in a Bayesian Networks Derived Framework, Padilla Bay, WA. Master’s Thesis at Western Washington University.

Hickey, J., Bond, M., Patton, T., Richardson, K., and Pugner, P. (2003). Reservoir Simulations of Synthetic Rain Floods for the Sacramento and San Joaquin River Basins. J. Water Resour. Plann. Manage., 10.1061/(ASCE)0733-9496(2003)129:6(443), 443–457.

Hines, E. E., and Landis, and W. G. (2014). Regional Risk Assessment of the Puyallup River Watershed and the Evaluation of Low Impact Development in Meeting Management Goals. Integrated Environmental Assessment and Management 10(2): 269–278.

Hoffert, M. I., Caldeira, K., Jain, A. K., Haites, E. F., Harvey, L. D. D., Potter, S. D., Schlesinger, M. E., Schneider, S. H., Watts, R. G., Wigley, T. M. L., and Wuebbles, D. J. (1998). Energy Implications of Future Stabilization of Atmospheric CO2 Content. Nature 395: 881–884.

Holling, C. S. (1973). Resilience and Stability of Ecological Systems. Annual Review of Ecology, Evolution and Systematics 4: 1–23.

Hosseini, S., and Barker, K. (2016). Modeling Infrastructure Resilience using Bayesian Networks: A Case Study of Inland Waterway Ports. Computers and Industrial Engineering 93: 252–266.

Howarth, R. W., Santoro, R., and Ingraffea, A. (2011). Methane and the Greenhouse-Gas Footprint of Natural Gas from Shale Formations. Climatic Change 106: 679–690.

Howlett, K., and Ladurantaye, S. (2011). How Mcguinty’s Green-Energy Policy Cost Him a Majority in Ontario. <http://www.theglobeandmail.com/news/politics/how-mcguintys-green-energy-policy-cost-him-a-majority-in-ontario/article556454/> (Mar. 16, 2015).

Huaizhi, S. U., Zhiping, W. E. N., Jiang, H. U., Zhongru, W. U., and Chongshi, G. U. (2008). A Decision Support System for Monitoring Dam Behavior. The 11th ASCE Aerospace Division International Conference, California, USA.

Huang, Y., Cai, J., Yin, H., and Cai, M. (2009). Correlation of Precipitation to Temperature Variation in the Huanghe River (Yellow River) basin during 1957-2006. Journal of Hydrology 372(1): 1–8.

Hufschmidt, M. M., and Fiering, M. B. (1966). Simulation Techniques for Design of Water-Resource System, Harvard University Press, Cambridge.

Hydro Québec (2015). <http://www.hydroquebec.com/projects/carte_production.html> (Oct. 01, 2015).

IEA Small Hydro (n.d.). <http://www.small-hydro.com/Past-Contributors-Pages/Canada.aspx> (Sept. 08, 2015).

Independent Electricity System Operator (IESO) (2010). <http://fit.powerauthority.on.ca/Storage/11128_FIT_Price_Schedule_August_13_2010.pdf> (Nov. 24, 2015).

196

Independent Electricity System Operator (IESO) (2015). <http://www.ieso.ca/> (Jul. 31, 2015).

Independent Electricity System Operator (IESO) (2016a). <http://fit.powerauthority.on.ca/fit-program/fit-program-pricing/fit-price-schedule> (Nov. 22, 2015).

Independent Electricity System Operator (IESO) (2016b). <http://fit.powerauthority.on.ca/fit-program/eligibility-requirements/waterpower> (Nov. 22, 2015).

Independent Electricity System Operator (IESO) (2016c). <http://www.ieso.ca/Pages/Participate/Markets-and-Programs/Ancillary-Services-Market.aspx> (Nov. 20, 2015).

Independent Electricity System Operator (IESO) (2017). <http://www.ieso.ca/> (Nov. 22, 2015).

Indiana Department of Natural Resources (2015). <http://www.in.gov/dnr/water/3660.htm> (Jan. 9, 2015).

Ingebretsen, E., and Johansen, T. H. G. (2014). The Profitability of Pumped Hydro Storage in Norway. Master’s Thesis, Norwegian School of Economics, Bergen, Norway.

Intergovernmental Panel on Climate Change (IPCC) (2011). Special Report on Renewable Energy Sources and Climate Change Mitigation, <http://srren.ipcc-wg3.de/> (Aug. 07, 2015).

International Association for Hydro-Environment Engineering and Research (IAHR) (2016). <http://iahr.informz.net/informzdataservice/onlineversion/ind/bWFpbGluZ2luc3RhbmNlaWQ9NTMzNjk0MSZzdWJzY3JpYmVyaWQ9Nzc2MDcwNzE4> (Jun. 17, 2016)

International Energy Agency (IEA) (1998). Benign Energy? The Environmental Implications of Renewables, <http://www.iea.org/tech/pubs/> (Jun. 10, 2015).

International Energy Agency (IEA) (2000). Hydropower and the Environment: Present Context and Guidelines for Future Action.

International Energy Agency (IEA) (2006). Hydropower Good Practices: Environmental Mitigation Measures.

International Energy Agency (IEA) (2009). World Energy Outlook 2009.

International Hydropower Association (IHA) (2010). Hydropower Sustainability Assessment Protocol.

International Joint Commission (2012). <http://www.ijc.org/files/publications/Lake_Superior Regulation_Full Report.pdf> (Mar. 13, 2015).

International Joint Commission (2012). Lake Superior Regulation: Addressing Uncertainty in Upper Great Lakes Water Levels, <http://www.ijc.org/files/publications/Lake_Superior _Regulation_Full_Report.pdf> (Mar. 3, 2015).

International Joint Commission (2016). Influences on Water Levels and Flows, <http://ijc.org/en_/ilsbc/FAQ_3> (Sep. 5, 2016).

197

International Renewable Energy Agency (IRENA) (2015). <http://www.irena.org/menu/index.aspx?mnu=Subcat&PriMenuID=36&CatID=141&SubcatID=494> (Jan. 04, 2016).

IPCC (2013). Climate Change 2013: The Physical Science Basis. Contribution of the Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press.

Islam, Z., and Gan, T. Y. (2014). Effects of Climate Change on the Surface-Water Management of the South Saskatchewan River Basin. J. Water Resour. Plann. Manage. 140(3): 332–342.

Jaber, J. O., Elkarmi, F., Alasis, E., and Kostas, A. (2015). Employment of Renewable Energy in Jordan: Current Status, SWOT and Problem Analysis. Renewable and Sustainable Energy Reviews 49: 490–499.

Jager, H. I., Efroymson, R. A., Opperman, J. J., and Kelly, M. R. (2015). Spatial Design Principles for Sustainable Hydropower Development in River Basins. Renewable and Sustainable Energy Reviews 45: 808–816.

Jaramillo, O. A., Borja, M. A., Huacuz, J. M. (2004). Using Hydropower to Complement Wind Energy: A Hybrid System to Provide Firm Power. Renewable Energy 29(11): 1887–1909.

Jaramillo, O. A., Rodríguez-Hernández, O., and Fuentes-Toledo, A. (2010). Stand-Alone and Hybrid Wind Energy Systems, Woodhead Publishing, Elsevier.

Ji, Y., Huang, G. H., and Sun, W. (2015). Risk Assessment of Hydropower Stations Through an Integrated Fuzzy Entropy-Weight Multiple Criteria Decision Making Method: A Case Study of the Xiangxi River. Expert Systems with Applications 42(12): 5380–5389.

Kahl, K. J., and Stirratt, H. (2012). What Could Changing Great Lakes Water Levels Mean for Our Coastal Communities? A Case for Climate-Adapted Planning Approaches, <http://www. nature.org/ourinitiatives/regions/northamerica/areas/greatlakes/explore/great-lakes-lake-levels-case-study.pdf> (Dec. 21, 2015).

Kajanus, M., Kangas, J., and Kurttila, M. (2004). The Use of Value Focused Thinking and the A’WOT Hybrid Method in Tourism Management. Tourism Management 25: 499–506.

Kaldellis, J. K. Ã., and Zafirakis, D. (2007). Optimum Energy Storage Techniques for the Improvement of Renewable Energy Sources-Based Electricity Generation Economic Efficiency. Energy 32: 2295–2305.

Kaldellis, J. K., Kapsali, M., and Kavadias, K. A. (2010). Energy Balance Analysis of Wind-Based Pumped Hydro Storage Systems in Remote Island Electrical Networks. Applied Energy 87(8): 2427–2437.

Kamodkar, R. U., and Regulwar, D. G. (2013). Multipurpose Reservoir Operating Policies: A Fully Fuzzy Linear Programming Approach. Journal of Agricultural Science and Technology 15: 1261–1274.

198

Kanakasabapathy P. (2013). Economic Impact of Pumped Storage Power Plant on Social Welfare of Electricity Market. Electrical Power and Energy Systems 45: 187–193.

Kapsali, M., and Kaldellis, J. K. (2010). Combining Hydro and Variable Wind Power Generation by Means of Pumped-Storage Under Economically Viable Terms. Applied Energy 87(11): 3475–3485.

.

Katsaprakakis, D. A., Christakis, D. G., Pavlopoylos, K., Stamataki, S., Dimitrelou, I., Stefanakis, I., and Spanos, P. (2012). Introduction of a Wind Powered Pumped Storage System in the Isolated Insular Power System of Karpathos – Kasos. Applied Energy 97: 38–48.

Kaunda, C. S., Kimambo, C. Z., and Nielsen, T. K. (2012). Hydropower in the Context of Sustainable Energy Supply: A Review of Technologies and Challenges. ISRN Renew Energy 2012: 1–15.

Kaygusuz, K. (2002). Sustainable Development of Hydroelectric Power. Energy Sources 24: 803–815.

Kaygusuz, K. (2009). The Role of Hydropower for Sustainable Energy Development. Energy Sources, Part B Econ Planning, Policy 4:365–376.

Kazempour, S. J., Moghaddam, M. P., Haghifam, M. R., and Yousefi, G. R. (2009). Electric Energy Storage Systems in a Market-Based Economy: Comparison of Emerging and Traditional Technologies. Renewable Energy 34(12): 2630–2639.

Kemp, R. and Parto, S. (2005). Governance for Sustainable Development: Moving from Theory to Practice (2005). International Journal of Sustainable Development 8(1/2): 12–30.

Kentel, E., and Alp, E. (2013). Hydropower in Turkey: Economical, Social and Environmental Aspects and Legal Challenges. Environmental Science and Policy 31: 34–43.

Kirkham, A. (2010). A Brief History of Hydroelectric Development at Niagara Falls, <http://www.nxtbook.com/nxtbooks/naylor/CDAQ0410/index.php?startid=12> (Nov. 18, 2015).

Klein, R. J. T., Nicholls, R. J., and Thomalla, F. (2003). Resilience to Natural Hazards: How Useful is This Concept? Journal of Environmental Hazards 5: 35–45.

Klimpt, J. É., Rivero, C., Puranen, H., and Koch, F. (2002). Recommendations for Sustainable Hydroelectric Development. Energy Policy 30: 1305–1312.

Kling, G. W., Hayhoe, K., Johnson, L. B., Magnuson, J. J., Polasky, S., Robinson, S. K., Shuter, B. J., Wander, M. M., Wuebbles, D. J., and Zak, D. R. (2003). Confronting Climate Change in the Great Lakes Region: Impacts on Our Communities and Ecosystems, Union of Concerned Scientists and the Ecological Society of America, Washington, D.C.

199

Klipsch, J. D., and Evans, T. A. (2006). Reservoir Operations Modeling with HEC-Ressim. The 3rd Federal Interagency Hydrologic Modeling Conference, Nevada, USA.

Knapp, V., Singh, J., Xie, H., Lian, Y., and Demissie, M. (2005). Potential Effects of Climate Change and Environmental Variability on the Resources of the Benguela Current Large Marine Ecosystem. World Water and Environmental Resources Congress: 1–35.

Koch, F. H. (2002). Hydropower – The Politics of Water and Energy: Introduction and Overview. Energy Policy 30, 1207–1213.

Koljonen, T., Flyktman, M., Lehtilä, A., Pahkala, K., Peltola, E., and Savolainen, I. (2009). The Role of CCS and Renewables in Tackling Climate Change. Energy Procedia 1:4323–4330.

Köne, A. Ç., and Büke, T. (2007). An Analytical Network Process (ANP) Evaluation of Alternative Fuels for Electricity Generation in Turkey. Energy Policy 35: 5220–5228.

Kotler, P. (1988). Marketing Management: Analysis, Planning, Implementation, and Control, Prentice-Hall College Div.

Koutroulis, A. G., Tsanis, I. K., Daliakopoulos, I. N., and Jacob, D. (2013). Impact of Climate Change on Water Resources Status: A Case Study for Crete Island, Greece. Journal of Hydrology 479: 146–158.

Krajačić, G., Duic, N., Tsikalakis, A., Zoulias, M., Caralis, G., Panteri, E., Krajac, G., and Grac, M. (2011a). Feed-in tariffs for Promotion of Energy Storage Technologies. Energy Policy 39: 1410–1425.

Krajačić, G., Duic, N., and Carvalho, M. G. (2011b). How to Achieve a 100 % RES Electricity Supply for Portugal? Applied Energy 88: 508–517.

Krajačić, G., Duic´, N., Zmijarevic, Z., Mathiesen, B. V., Vucini, A. A., and Carvalho, M. G. (2011c). Planning for a 100 % Independent Energy System based on Smart Energy Storage for Integration of Renewables and CO2 Emissions Reduction. Applied Thermal Engineering 31: 2073–2083.

Krajačić, G., Loncˇar, D. ˇen, Duic´ a, N., Zeljko, M., Lacal, R. A., Loisel, R., and Raguzin, I. (2013). Analysis of Financial Mechanisms in Support to New Pumped Hydropower Storage Projects in Croatia. Applied Energy 101: 161–171.

Kucukali, S. (2014). Environmental Risk Assessment of Small Hydropower (SHP) Plants: A Case Study for Tefen SHP Plant on Filyos River. Energy for Sustainable Development 19: 102–110.

Kucukali, S., and Baris, K. (2009). Assessment of Small Hydropower (SHP) Development in Turkey: Laws, regulations and EU policy Perspective. Energy Policy 37(10): 3872–3879.

Kumar, D. N. and Reddy, M. J. (2007). Multipurpose Reservoir Operation Using Particle Swarm Optimization. J. Water Resour. Plann. Manage., 10.1061/(ASCE)0733-9496(2007)133:3(192), 192-201.

200

Kumar, D., and Katoch, S. S. (2014). Sustainability Indicators for Run of the River (RoR) Hydropower Projects in Hydro Rich Regions of India. Renewable and Sustainable Energy Reviews 35: 101–108.

Kurttila, M., Pesonen, M., Kangas, J., and Kajanus, M. (2000). Utilizing the Analytic Hierarchy Process (AHP) in SWOT Analysis-A Hybrid Method and Its Application to a Forest-Certification Case. For. Policy and Economics 1: 41–52.

Kusakana, K. (2015). Feasibility Analysis of River Off-Grid Hydrokinetic Systems with Pumped Hydro Storage in Rural Applications. Energy Conversion and Management 96: 352-362.

Kusangaya, S., Warburton, M. L., Archer van Garderen, E., and Jewitt, G. P. W. (2014). Impacts of Climate Change on Water Resources in Southern Africa: A Review. Physics and Chemistry of the Earth 67-69: 47–54.

Kwon, T. (2015). Rent and Rent-Seeking in Renewable Energy Support Policies: Feed-in Tariff Vs. Renewable Portfolio Standard. Renewable and Sustainable Energy Reviews 44: 676–681.

Lackner, K. S., and Sachs, J. D. (2005). A Robust Strategy for Sustainable Energy. Brookings Papers on Economic Activity 2: 215–284.

Lako, P., Simbolotti, G., and Tosato, G. (2010). Hydropower, International Energy Agency.

Lal, A. M. W. (1995). Calibration of Riverbed Roughness. Journal of Hydraulic Engineering 121: 664–671.

Lara, P. G., Lopez, J. D., Luz, G. M., and Bonumá, N. B. (2014). Reservoir Operation Employing HEC-Ressim: Case Study of Tucuruí Dam, Brazil. The 6Th International Conference on Flood Management, São Paulo, Brazil.

Latorre, J. M., Cerisola, S., Ramos, A., Perea, A., and Bellido, R. (2014). Coordinated Hydropower Plant Simulation for Multireservoir Systems. Water Resources Planning and Management, 10.1061/(ASCE)WR.1943-5452.0000306.

Lauritzen, S. L. (1995). The EM Algorithm for Graphical Association Models with Missing Data. Computational Statistics & Data Analysis 19: 191–201.

Lawson, A., Goldstein, M., and Dent, C. J. (2016). Bayesian Framework for Power Network Planning Under Uncertainty. Sustainable Energy, Grids and Networks 7: 47–57.

Lee, D. H., Quinn, F. H., and Clites, A. H. (1988). Effect of The Niagara River Chippawa Grass Island Pool On Water Levels of Lake Erie, St. Clair, And Michigan-Huron. J. Great Lakes Res. 24: 936–948.

Lele, S. M. (1991). Sustainable Development: A Critical Review. World Development 19(6): 607–621.

201

Lemmen, D. S., and Warren, F. J. (2004). Climate Change Impacts and Adaptation: A Canadian Perspective, Natural Resources Canada, Ontario.

Leskinen, L. A., Leskinen, P., Kurttila, M., Kangas, J., and Kajanus, M. (2006). Adapting Modern Strategic Decision Support Tools in The Participatory Strategy Process-A Case Study of a Forest Research Station. For. Policy and Economics 8: 267–278.

Letcher, T. M. (2016). Storing Energy: with Special Reference to Renewable Energy Sources, Elsevier, Netherlands.

Levine, J. G. (2003). Pumped Hydroelectric Energy Storage and Spatial Diversity of wind Resources as Methods of Improving Utilization of Renewable Energy Sources. MASc Thesis, University of Colorado.

Levine, N. M., Zhang, K., Longo, M., Baccini, A., Phillips, O. L., Lewis, S. L., Alvarez-Dávila, E., Segalin de Andrade, A. C., Brienen, R. J. W., Erwin, T. L., Feldpausch, T. R., Monteagudo Mendoza, A. L., Nuñez Vargas, P., Prieto, A., Silva-Espejo, J. E., Malhi, Y., and Moorcroft, P. R. (2015). Ecosystem Heterogeneity Determines the Ecological Resilience of the Amazon to Climate Change. Proceedings of the National Academy of Sciences 113(3): 793–797.

Li, X., Chen, G., and Zhu, H. (2016). Quantitative Risk Analysis on Leakage Failure of Submarine Oil and Gas Pipelines Using Bayesian Network. Process Safety and Environmental Protection, Institution of Chemical Engineers 103: 163–173.

Li, X., Wei, J., Fu, X., Li, T., and Wang, G. (2014). Knowledge-Based Approach for Reservoir System Optimization. Journal of Water Resource Planning and Management, 10.1061/(ASCE)WR.1943-5452.0000379, 04014001.

Li, Y., Li, Y., Ji, P., and Yang, J. (2015). The Status Quo Analysis and Policy Suggestions on Promoting China’s Hydropower Development. Renewable and Sustainable Energy Reviews 51: 1071–1079.

Liden, R., and Lyon, K. (2014). The Hydropower Sustainability Assessment Protocol for Use by World Bank Clients.

LimnoTech (2010). Water Quality Model Development and Calibration Report for Buffalo River, Scajaquada Creek, Niagara River, and Black Rock Canal, <http://bsacsoimprovements. org/wp-content/uploads/2012/05/BSA-LTCP-Appendix-4-5.pdf> (Jan. 4, 2015).

Linares, P., Santos, F. J., and Ventosa, M. (2008). Coordination of Carbon Reduction and Renewable Energy Support Policies. Climate Policy 8: 377–394.

Lipp, J. (2007). Lessons for Effective Renewable Electricity Policy from Denmark, Germany and The United Kingdom. Energy Policy 35: 5481–5495.

Little R. J. A, Rubin D. B. (2002). Statistical Analysis with Missing Data, 2nd Ed., Wiley, New Jersey.

202

Liu, F., Chen, S., Dong, P., and Peng, J. (2012). Spatial and Temporal Variability of Water Discharge in the Yellow River Basin over the Past 60 Years. Journal of Geographical Sciences 22(6): 1013–1033.

Liu, J., Zuo, J., Sun, Z., Zillante, G., and Chen, X. (2013). Sustainability in Hydropower Development – A Case Study. Renewable and Sustainable Energy Reviews 19: 230–237.

Locatelli, G., Palerma, E., and Mancini, M. (2015). Assessing the Economics of Large Energy Storage Plants with an Optimisation Methodology. Energy 83: 15–28.

Lofgren, B. M., Quinn, F. H., Clites, A. H., Assel, R. A., Eberhardt, A. J., and Luukkonen, C. L. (2002). Evaluation of Potential Impacts on Great Lakes Water Resources Based on Climate Scenarios of Two GCMs. Journal of Great Lakes Research 28(4): 537–554.

Loisel, R., Mercier, A., Gatzen, C., Elms, N., and Petric, H. (2010). Valuation Framework for Large Scale Electricity Storage in a Case with Wind Curtailment. Energy Policy 38(11): 7323–7337.

Lopez, N., and Espiritu, J. F. (2011). An Approach to Hybrid Power Systems Integration Considering Different Renewable Energy Technologies. Procedia Computer Science 6: 463–468.

Lord, A., and Tyler, S. (2015). Ontario 2015 Renewable Energy Outlook: Procurement Update. <https://www.dlapiper.com/en/canada/insights/publications/2015/02/ontario-2015-renewable-energy-outlook-procuremen__/> (Sep. 14, 2016).

Low Impact Hydropower Institute (LIHI) (2014). Revised LIHI Certification Criteria.

Low Impact Hydropower Institute (LIHI) (2016a). <http://lowimpacthydro.org/> (May 20, 2016)

Low Impact Hydropower Institute (LIHI) (2016b). Low Impact Hydropower Certification Handbook.

Lowry, T. S., Tidwell, V. C., Pierce, S. A., Cain, W., and Dulay, M. M. (2007). Multi-platform Decision Support System. The World Environmental and Water Resources Congress, Florida, USA.

Lu, X., Tchou, J., Mcelroy, M. B., and Nielsen, C. P. (2011). The Impact of Production Tax Credits on the Profitable Production of Electricity from Wind in the US. Energy Policy 39: 4207–4214.

Ma, C., Ju, M. T., Zhang, X. C., and Li, H. Y. (2011). Energy Consumption and Carbon Emissions in A Coastal City in China. Procedia Environmental Sciences 4: 1–9.

Ma, C., Zhang, X., Zhang, G., Ju, M., Zhou, B., and Li, X. (2012). Application of DPSIR Framework in Environmental Impact Assessment for Port Planning. China Environ. Sci. 32: 107–111.

203

Mabee, W. E., Mannion, J., and Carpenter, T. (2012). Comparing the Feed-in Tariff Incentives for Renewable Electricity in Ontario and Germany. Energy Policy 40: 480–489.

MacDougall, J. (2008). Ontario’s Renewable Energy Standard Offer Program, Lessons from Large Scale Distribution Connected Electricity Procurement Program, <http://www.conference-on-integration.com/pres/16_MacDougall.pdf> (Mar. 19, 2015).

Magnuson, J. J., Webster, K. E., Assel, R. A., Bowser, C. J., Dillon, P. J., Eaton, J. G., Evans, H. E., Fee, E. J., Hall, R. I., Mortsch, L. R., Schindler, D. W., and Quinn, F. H. (1997). Potential Effects of Climate Changes On Aquatic Systems: Laurentian Great Lakes and Precambrian Shield Area. Hydrological Processes 11: 825–871.

Maidment, D. R., and Fread, D. L. (1993). Handbook of Hydrology, McGraw-Hill, New York.

Makos, J. (2011). <http://pestleanalysis.com/pestle-and-swot-analysis/> (Aug. 6, 2015)

Malakar, T., Goswami, S. K., and Sinha, A. K. (2014). Optimum Scheduling of Micro Grid Connected Wind-Pumped Storage Hydro Plant in a Frequency Based Pricing Environment. International Journal of Electrical Power and Energy Systems 54: 341–351.

Maliszewski, P. J. and Perrings, C. (2012). Factors in the Resilience of Electrical Power Distribution Infrastructures. Applied Geography 32: 668–679.

Manabe, S., and Wetherald, T. (1987). Large-Scale Changes of Soil Wetness Induced by an Increase in Atmospheric Carbon Dioxide. Journal of Atmospheric Sciences 44(8): 1211–1235.

Manyena, S. B. (2006). The Concept of Resilience Revisited. Disasters 30(4): 433–450.

Margles, S. W., Masozera, M., Rugyerinyange, L., and Kaplin, B. A. (2010). Participatory Planning: Using SWOT-AHP Analysis in Buffer Zone Management Planning. J. Sustainable Forestry 29: 613–637.

Maricic, M., Haber, D., and Pejovic, S. (2009). Niagara Pump Generating Station Proven Functionality Unique in Canada. The IEEE Electrical Power & Energy Conference, Quebec.

Markandya (2012). Externalities from Electricity Generation and Renewable Energy. Methodology and Application in Europe and Spain. Información Comercial Española Cuad Econ 1(83): 85–100.

Marshall, C. (2013). <http://www.scientificamerican.com/article/ontario-phases-out-coal-fired-power/> (Jul. 15, 2015).

Martin, O., Lillehammer, L., and Hveling, O. (2010). Hydropower Development and Curbing Climate Gas Emissions: A Win-Win Opportunity. The 6th International Conference on Hydropower.

204

Masozera, M. K., Alavalapati, J. R. R., Jacobson, S. K., and Shresta, R. K. (2006). Assessing the Suitability of Community-Based Management for the Nyungwe Forest Reserve, Rwanda. For. Policy and Economics 8: 206–216.

Matevosyan, J., Olsson, M., and Söder, L. (2009). Hydropower Planning Coordinated with Wind Power in Areas with Congestion Problems for Trading on the Spot and the Regulating Market. Electric Power Systems Research 79:39–48.

Mathiesen, B. V., and Lund, H. (2009). Comparative Analyses of Seven Technologies to Facilitate the Integration of Fluctuating Renewable Energy Sources. IET Renewable Power Generation 3(2): 190-204.

Maxim, A. (2014). Sustainability Assessment of Electricity Generation Technologies Using Weighted Multi-Criteria Decision Analysis. Energy Policy 65: 284–297.

May, P. J., Burby, R. J., Ericksen, N. J., Handmer, J. W., Dixon, J. E., Michaels, S. et al. (1996). Environmental Management and Governance: Intergovernmental Approaches to Hazards and Sustainability, Routledge, London.

McFarlane, N. A., Boer, G. J., Blanchet, J. P., and Lazare, M. (1992). The Canadian Climate Centre Second-Generation General Circulation Model and its Equilibrium Climate. Journal of Climate 5: 1013–1044.

McKay, P. (1983). Electric Empire: The Inside Story of Ontario Hydro, Between the Lines, Ontario.

McKitrick, R. R. (2013). Environmental and Economic Consequences of Ontario’s Green Energy Act, Fraser Institute.

McMahon, G. and Farmer, M. (2009). Rule-Based Storage Accounting for Multipurpose Reservoir Systems. J. Water Resour. Plann. Manage., 10.1061/(ASCE)0733-9496(2009)135:4(286), 286–297.

McNally, A., Magee, D., and Wolf, A. T. (2009). Hydropower and Sustainability: Resilience and Vulnerability in China’s Powersheds. Journal of Environmental Management 90: 286–293.

Meier, P. J. (2002). Life-Cycle Assessment of Electricity Generation Systems and Applications for Climate Change Policy Analysis. PhD Thesis, University of Wisconsin–Madison.

Melikoglu, M. (2017). Pumped Hydroelectric Energy Storage: Analysing Global Development and Assessing Potential Applications in Turkey Based on Vision 2023 Hydroelectricity Wind and Solar Energy Targets. Renewable and Sustainable Energy Reviews 72: 146–153.

Melkonyan, A. (2015). Climate Change Impact on Water Resources and Crop Production in Armenia. Agricultural Water Management 161: 86–101.

Melo, O. T. (1989). Electric Supply and Demand in Ontario. First U.S.-Canada Symposium on Impacts of Climate Change on the Great Lakes Basin, Sept. 27-29, Illinois.

205

Mendonça, M. (2009). Feed-in Tariffs: Accelerating the Deployment of Renewable Energy, Routledge, UK.

Mendonça, M., Jacobs, D., and Sovacool, B. K. (2009). Powering the Green Economy: The Feed-in Tariff Handbook, Routledge, UK.

Mensah, A. F., and Duenas-Osorio, L. (2010). Efficient Resilience Assessment Framework for Electric Power Systems Affected by Hurricane Events. Journal of Structural Engineering 142(8): 1–10.

Merritt, W., Rao, K. V, Patch, B., Reddy, V. R., Syme, G. J., and Sreedevi, P. D. (2015). Exploring Implications of Climate, Land Use, and Policy Intervention Scenarios on Water Resources, Livelihoods, and Resilience. Integrated Assessment of Scale Impacts of Watershed Intervention, Elsevier Inc.

Metzger, E., Pino, S. P. D., Prowitt, S., Goodward, J., and Perera, A. (2012). <http://www.wri.org/> (Dec. 19, 2013).

Michailidis, A., Papadaki-Klavdianou, A., Apostolidou, I., Lorite, I. J., Pereira, F. A., Mirko, H., Buhagiar, J., Shilev, S., Michaelidis, E., Loizou, E., Chatzitheodoridis, F., Restoy, R. C., and Lopez, A. L. (2015). Exploring Treated Wastewater Issues Related to Agriculture in Europe, Employing a Quantitative SWOT Analysis. Procedia Economics and Finance 33: 367–375.

Mileti, D. S., and Peek, L. A. (2002). Understanding Individual and Social Characteristics in the Promotion of Household Disaster Preparedness. T. Dietz, P.C. Stern (Eds.), New Tools for Environmental Protection: Education, Information and Voluntary Measures, National Academy Press, Washington, DC.

Miller, L. and Carriveau, R. (2017). Balancing the Carbon and Water Footprints of the Ontario Energy Mix. Energy 125: 562–568.

Milly, P. C. D., Betancourt, J., Falkenmark, M., Hirsch, R. M., Kundzewicz, Z. W., Lettenmaier, D. P., and Stouffer, R. J. (2008). Stationarity is Dead: Whither Water Management? Science 319 (5863): 573–574.

Milly, P. C. D., Dunne, K. A., and Vecchia, A. V. (2005). Global Pattern of Trends in Streamflow and Water Availability in a Changing Climate. Nature 438: 347–350.

Mimikou, M. A., Baltas, E., and Varanou, E. (2004). Climate Change Impacts on Water Resources: Quantity and Quality Aspects, ASCE Library, Virginia, the US.

Ministry of Tourism (2008). The Niagara Region’s Tourism Opportunities: The U.S. and Ontario Markets, Ontario.

MIT News. (2013). <http://news.mit.edu/2013/study-air-pollution-causes-200000-early-deaths-each-year-in-the-us-0829> (Aug. 05, 2015).

206

Mitchell, C., Bauknecht, D., and Connor, P. M. (2006). Effectiveness Through Risk Reduction: A Comparison of the Renewable Obligation in England and Wales and the Feed-in System in Germany. Energy Policy 34, 297–305.

Mo, L., Lu, P., Wang, C., and Zhou, Z. (2013). Short-Term Hydro Generation Scheduling of Three Gorges–Gezhouba Cascaded Hydropower Plants Using Hybrid MACS-ADE Approach. Energy Conversion and Management 76: 260–273.

Modini, C. (2010). Using HEC-ResSim for Columbia River Treaty Flood Control. The 2nd Joint Federal Interagency Conference, Las Vegas, USA.

Moeini, R., Afshar, A., and Afshar, M. H. (2011). Fuzzy Rule-Based Model for Hydropower Reservoirs Operation. International Journal of Electrical Power and Energy Systems 33(2): 171–178.

Molyneaux, L. Wagner, L., Froome, C. and Foster, J. (2012). Resilience and Electricity Systems: A Comparative Analysis. Energy Policy 47: 188–201.

Morimoto, R. (2013). Incorporating Socio-Environmental Considerations into Project Assessment Models Using Multi-Criteria Analysis: A Case Study of Sri Lankan Hydropower Projects. Energy Policy 59: 643–653.

Mortsch, L. D., Alden, M. and Klaassen, J. (2005). Development of Climate Change Scenarios for Impact and Adaptation Studies in The Great Lakes-St. Lawrence Basin. International Joint Commission.

Mortsch L. D., Alden, M. and Scheraga, J. D. (2003). Climate Change and Water Quality in the Great Lakes region, International Joint Commission.

Mortsch, L. D. (1998). Assessing the Impact of Climate Changes on the Great Lakes Shoreline Wetlands. Climatic Change 40: 391–416.

Mortsch, L. D., and Quinn, F. H. (1996). Climate Change Scenarios for Great Lakes Basin Ecosystem Studies. Limnology and Oceanography 41(5): 903–911.

Mortsch, L. D., Croley, T. and Fay, D. (2006). Impact of Climate Change on Hydro-Electric Generation in the Great Lakes. C-CIARN-Water, Hydropower and Climate Change Workshop, Inn at the Forks, Winnipeg, Manitoba.

Mortsch, L. D., Hengeveld, H., Lister, M., Wenger, L., Lofgren, B., Quinn, F., and Slivitzky, M. (2000). Climate Change Impacts on the Hydrology of the Great Lakes-St. Lawrence System. Canadian Water Resources Journal 25(2): 153–179.

Mortsch. (2003). Climate Change and Water Quality in the Great Lakes Region, <http://www.ijc. org/rel/pdf/climate_change_2003_part3.pdf> (Jul. 30, 2015).

Murage, M. W., and Anderson, C. L. (2014). Contribution of Pumped Hydro Storage to Integration of Wind Power In Kenya: An Optimal Control Approach. Renewable Energy 63: 698–707.

207

Nair, K. (2010). An Analysis of the Issues Associated with Hydropower Generation in Kerala-India, Under Changing Climate. The 6th International Conference on Hydropower.

National Energy Board (2013). <https://www.neb-one.gc.ca/nrg/ntgrtd/ftr/2013/index-eng.html#smmr> (Feb. 14, 2015).

National Energy Board (2015). <http://www.neb.gc.ca/bts/whwr/index-eng.html> (Sep. 14, 2015).

National Oceanic and Atmospheric Administration (NOAA) (2013). <http://tidesandcurrents.noaa.gov/waterlevels.html?id=9063007> (Dec. 27, 2013).

National Oceanic and Atmospheric Administration (NOAA) (2015a). Great Lakes Bathymetry, <http://maps.ngdc.noaa.gov/viewers/wcs-client/> (Aug. 30, 2015).

National Oceanic and Atmospheric Administration (NOAA) (2015b). Water levels – Station Selection, <http://tidesandcurrents.noaa.gov/stations.html?type=Water+Levels> (Aug. 31, 2015).

National Renewable Energy Laboratory (NREL) (2011). Integration of Wind and Hydropower Systems, < https://www.nrel.gov/docs/fy12osti/50182.pdf> (Aug. 27, 2017).

National Renewable Energy Laboratory (NREL) (2013). Integrated Canada-U.S. Power Sector Modeling with the Regional Energy Deployment System (ReEDS).

Natural Resources Canada. (2007). Emerging Hydropower Technologies R&D in Canada: A Strategy for 2007-2011, <http://www.nrcan.gc.ca/energy/publications/sciences-technology/renewable/small-hydropower/6489> (Sep. 14, 2015).

Natural Resources Canada. (2011). Clean Energy Fund Program, <www.nrcan.gc.ca/eneene/science/ceffep-eng.php> (Feb. 14, 2015).

Nautiyal, H., Singal, S. K., Varun, and Sharma, A. (2011). Small Hydropower for Sustainable Energy Development in India. Renewable and Sustainable Energy Reviews 15: 2021–2027.

Nelles, H. V. (1974). The Politics of Development: Forests, Mines, and Hydro-Electric Power in Ontario, 1849-1941, Second Ed. McGill-Queen’s University Press, Montreal.

Nemec, K. T., Chan, J., Hoffman, C., Spanbauer, T. L., Hamm, J. A., Allen, C. R., Hefley, T., Pan, D., and Shrestha, P. (2014). Assessing Resilience in Stressed Watersheds. Ecology and Society 19(1): 34.

Neufville, R. D. (1990). Applied Systems Analysis: Engineering Planning and Technology Management. McGraw-Hill, the United States.

New York Power Authority (NYPA) (2015). <http://www.nypa.gov/Press/2015/111215.html> (Dec. 17, 2015).

208

Ngo, L. L., Madsen, H., and Rosbjerg, D. (2007). Simulation and Optimization Modelling Approach for Operation of the Hoa Binh Reservoir, Vietnam. Journal of Hydrology 336: 269–281.

Niagara Falls Review (2011). <http://www.niagarafallsreview.ca/2011/08/10/tourism-flat-in-falls> (Jan. 21, 2016).

Niagara Parks Commission (2004). Niagara Falls Mist Dispersion Study.

Niagara Parks Commission (2015). <http://www.niagaraparks.com/about-niagara-falls/geology-facts-figures.html> (Jan. 08, 2015).

Nicholls, K. H. (1999). Effects of Temperature and Other Factors on Summer Phosphorus in the Inner Bay of Quinte, Lake Ontario: Implications for Climate Warming. Journal of Great Lakes Research 25(2): 250–262.

Nkomozepi, T., and Chung, S. O. (2014). The Effects of Climate Change on the Water Resources of the Geumho River Basin, Republic of Korea. Journal of Hydro-environment Research 8(4): 358–366.

NOAA Center for Operational Oceanographic Products and Services (2016). <https://tidesandcurrents.noaa.gov/> (Jul. 18, 2016).

NOAA Great Lakes Environmental Research Laboratory (NOAA GLERL) (2015). Hydrology & hydraulics data, <http://www.glerl.noaa.gov/data/pgs/hydrology.html> (Aug. 30, 2015).

NOAA Great Lakes Environmental Research Laboratory (NOAA GLERL). (2016). <http://www. glerl.noaa.gov/pr/ourlakes/> (Feb. 2, 2016).

NOAA Great Lakes Ice Atlas. (2016). <https://www.glerl.noaa.gov//data/ice/atlas/daily_ice_cover/daily_grids/dailygrids.html> (Apr. 30, 2016).

NOAA National Centers for Information and Technology. (2016). <https://www.ncei.noaa.gov/> (Jun. 20, 2016).

NOAA National Data Buoy Center (NDBC). (2016). <http://www.ndbc.noaa.gov/> (Jun. 20, 2016).

Norris, F. H., Stevens, S. P., Pfefferbaum, B., Wyche, K. F., Pfefferbaum, R. L. (2008). Community Resilience as a Metaphor, Theory, Set of Capacities, and Strategy for Disaster Readiness. American Journal of Community Psychology 41(1-2): 127–150.

Northland Power (2011). Marmora Pumped Storage Project.

Northland Power (2014). Marmora Pumped Storage Hydro.

Norton, B. G. (2005). Sustainability: A Philosophy of Adaptive Ecosystem Management, University of Chicago Press.

209

OECD. (2004). OECD Economic Survey of Canada 2004: Electricity Sector, <www.oecd.org/dataoecd/52/0/33847613.pdf> (Sep. 14, 2015).

Ohio Department of Natural Resources. (2014). <https://www.ohiodnr.gov/tabid/3774/Default.aspx.> (Mar. 11, 2014).

Olmstead, S. M., Fisher-Vanden, K. A., and Rimsaite, R. (2016). Climate Change and Water Resources: Some Adaptation Tools and Their Limits. Journal of Water Resources Planning and Management 142(6): 1–2.

Onat, N., Bayar, H. (2010). The Sustainability Indicators of Power Production Systems. Renewable and Sustainable Energy Reviews 14: 3108–3115.

Ontario Ministry of Energy (2009). Green Energy and Green Economy Act, <http://www.e laws.gov.on.ca/html/source/statutes/english/2009/elaws_src_s09012_e.htm> (Mar. 16, 2015).

Ontario Ministry of Energy (2010). Ontario’s Long-Term Energy Plan.

Ontario Ministry of Energy (2013a). Achieving Balance: Ontario’s Long-Term Energy Plan, Government of Ontario.

Ontario Ministry of Energy (2013b). Ontario Working with Communities to Secure Clean Energy Future, <http://news.ontario.ca/mei/en/2013/05/ontario-working-with-communities-to-secure-clean-energy-future.html> (Mar. 12, 2015).

Ontario Ministry of Energy (2013c). Ontario Lowering Future Energy Costs, <http://news.ontario.ca/mei/en/2013/12/ontario-lowering-future-energy-costs.html> (Mar. 12, 2015).

Ontario Ministry of Environment and Climate Change (2013). Greenhouse Gas Emissions Reductions.

Ontario Ministry of Environment and Climate Change (2014). Ontario’s Climate Change Update, <https://dr6j45jk9xcmk.cloudfront.net/documents/3618/climate-change-report-2014.pdf> (Jan. 29, 2015)

Ontario Ministry of Tourism (2009). <http://www.mtc.gov.on.ca/> (Jan. 06, 2015).

Ontario Power Authority (2013a). HESOP Municipal Stream Program Rules.

Ontario Power Authority (2013b). HESOP Expansion Stream Program Rules.

Ontario Power Authority (2014). LRP Draft, <http://www.powerauthority.on.ca/sites/default/files/page/Draft-LRP-I-RFP-r2.pdf> (Sep. 12, 2016).

Ontario Power Authority (2016a). FIT Price Schedule, <http://fit.powerauthority.on.ca/what-feed-tariff-program> (Sep. 12, 2016).

210

Ontario Power Authority (2016b). Active FIT Contracts, <http://www.powerauthority.on.ca/sites/default/files/Active-FIT-contracts-2015-01-30.pdf> (Sep. 15, 2016).

Ontario Power Authority (2016c). HESOP Municipal Stream contract offers, <http://www.powerauthority.on.ca/sites/default/files/HESOP-Contract-List.pdf > (Sep. 12, 2016).

Ontario Power Generation (2016). < http://www.opg.com/generating-power/Pages/generating-power.aspx> (Jun. 01, 2016).

Ontario Power Generation (2015a). Eastern Operations, <http://www.opg.com/generating-power/hydro/ottawa-st-lawrence/Pages/ottawa-st-lawrence.aspx> (Mar. 21, 2015).

Ontario Power Generation (OPG) (2015b). <http://www.opg.com/generating-power/hydro/projects/niagara-tunnel-project/Pages/the-beck-complex.aspx> (Jun. 06, 2015).

Ontario Power Generation (OPG) (2010). <http://www.opg.com/about/reg/filings/paymentamounts/files/Exhibit%20L%20%20Interrogatory%20Responses/Issue%2006.02.pdf> (Jan 12, 2012).

Osroosh, M. (2012). HEC-Ressim Model for Simulating Utilization from Reservoir in the Form of Multiple – Dam. International Journal of Structronics & Mechatronics, 1–6.

Ostrega, A., De Felice, F. and Petrillo, A. (2011). ANP-SWOT Approach to Minimize Environmental Impacts Due Mining Activities. International Symposium on the Analytic Hierarchy Process, Italy.

Othman M. R., Wozny, G. and Repke, J. (2011). Selection of Sustainable Chemical Process Design Using ANP: A Biodiesel Case Study. International Symposium on the Analytic Hierarchy Process, Italy.

Ott, K. (2003). The Case for Strong Sustainability. Greifswald’s Environmental Ethics: 59–64.

Paine, N., Homans, F. R., Pollak, M., Bielicki, J. M., and Wilson, E. J. (2014). Why Market Rules Matter: Optimizing Pumped Hydroelectric Storage When Compensation Rules Differ. Energy Economics 46: 10–19.

Palacios-Gomez, F., Lasdon, L., and Engquist, M. (1982). Nonlinear Optimization by Successive Linear Programming. Manage. Sci. 28(10): 1106-1120.

Pan, X., Teng, F., Wang, X., and Wang, G. (2014). Energy Transition Within a Carbon Constrained World: How Allocation Schemes Influence the Development of Energy System in the Future. Energy Procedia 61: 8–11.

Pang, M., Zhang, L., Ulgiati, S., and Wang, C. (2015). Ecological Impacts of Small Hydropower in China: Insights from an Emergy Analysis of a Case Plant. Energy Policy 76: 112–122.

211

Paolini, M., and Vacis, G. (2000). The Story of Vajont. Bordighera Press.

Paton, D., and Johnston, D. M. (2006). Disaster Resilience: An Integrated Approach, Charles C Thomas Pub Ltd, Springfield, IL.

Paton, D., Violanti, J. M., Smith, L. (2003). Promoting Capabilities to Manage Post-Traumatic Stress, Charles C Thomas Pub Ltd, Springfield, IL.

Pearl, J. (2003). Causality: Models, Reasoning, and Inference. Econometric Theory 19: 675–685.

Pereira, A. M., and Pereira, R. M. (2014). On the Environmental, Economic and Budgetary Impacts of Fossil Fuel Prices: A Dynamic General Equilibrium Analysis of the Portuguese Case. Energy Economics 42: 248–261.

Perez-Díaz, J. I., and Jimenez, J. (2016). Contribution of a Pumped-Storage Hydropower Plant to Reduce the Scheduling Costs of an Isolated Power System with High Wind Power Penetration. Energy 109: 92–104.

Peter, C., De Lange, W., Musango, J. K., April, K., and Potgieter, A. (2009). Applying Bayesian Modelling to Assess Climate Change Effects on Biofuel Production. Climate Research, 40(2): 249–260.

Peterson, G., Alessandro, G. D. L., Hellmann, J. J., Janssen, M. A., Kinzig, A., Malcom, J. R., O’Brien, K., Pope, S. E., Rothman, D. S., Shevliakova, E., and Tinch, R. R. T. (1997). Uncertainty, Climate Change, and Adaptive Management Atmospheric Change. Conservation Ecology 1(2): 1-7.

Pfenninger, S., Gauché, P., Lilliestam, J., Damerau, K., Wagner, F., and Patt, A. (2014). Potential for Concentrating Solar Power to Provide Baseload and Dispatchable Power. Natural Climate Change 4: 689–692.

Piao, S., Ciais, P., Huang, Y., Shen, Z., Peng, S., Li, J., Zhou, L., Liu, H., Ma, Y., Ding, Y., Friedlingstein, P., Liu, C., Tan, K., Yu, Y., Zhang, T., and Fang, J. (2010). The Impacts of Climate Change on Water Resources and Agriculture in China. Nature 467(7311): 43–51.

Pirnia, M., Nathwani, J., and Fuller, D. (2011). Ontario Feed-in-Tariffs: System Planning Implications and Impacts on Social Welfare. The Electricity Journal 24: 18–28.

Poff, N. L., Allan, J. D., Bain, M. B., Karr, J. R., Prestegaard, K. L., Richter, B. D., Sparks, R. E., and Stromberg, J. C. (1997). Natural Flow Regime. BioScience 47(11): 769–784.

Poff, N. L., Brown, C. M., Grantham, T. E., Matthews, J. H., Palmer, M. a., Spence, C. M., Wilby, R. L., Haasnoot, M., Mendoza, G. F., Dominique, K. C., and Baeza, A. (2015). Sustainable Water Management under Future Uncertainty with Eco-Engineering Decision Scaling. Nature Climate Change 6: 1–23.

Prasad, S. A., Umamahesh, N. V., and Viswanath, G. K. (2012). Short‐Term Real‐Time Reservoir Operation for Irrigation. Water Resources Planning and Management, 10.1061/ (ASCE)WR.1943-5452.0000234.

212

Purkey, D. R., Huber-Lee, A., Yates, D. N., Hanemann, M., and Herrod-Julius, S. (2007). Integrating a Climate Change Assessment Tool into Stakeholder-Driven Water Management Decision-Making Processes in California. Water Resources Management 21(1): 315–329.

Qi, M., Feng, M., Sun, T., and Yang, W. (2016). Resilience Changes in Watershed Systems: A New Perspective to Quantify Long-Term Hydrological Shifts Under Perturbations. Journal of Hydrology 539: 281–289.

Quansah, D. A., Adaramola, M. S., and Mensah, L. D. (2016). Solar Photovoltaics in Sub-Saharan Africa – Addressing Barriers, Unlocking Potential. Energy Procedia 106: 97–110.

Qudrat-Ullah, H. (2013). Understanding the Dynamics of Electricity Generation Capacity in Canada: A System Dynamics Approach. Energy 59: 285–294.

Qudrat-Ullah, H. (2014). Green Power in Ontario: A Dynamic Model-Based Analysis. Energy 77: 859–870.

Quinn, F. and Lofgren. B. M. (2000). The Influence of Potential Greenhouse Warming on Great Lakes Hydrology, Water Levels, and Water Management. The 15th Conference on Hydrology, January 9-14.

Rajan, A. T. (2010). Venture Capital and Efficiency of Portfolio Companies. IIMB Management Review 22(4): 186–197.

Randhir, T. (2014). Resilience of Watershed Systems to Climate Change. Journal of Earth Science & Climatic Change 5(6): 1–2.

Rangoni, B. (2012). A Contribution on Electricity Storage: The Case of Hydro-Pumped Storage Appraisal and Commissioning in Italy and Spain. Utilities Policy 23: 31–39.

Rauch, P., Wolfsmayr, U. J., Borz, S. A., Triplat, M., Krajnc, N., Kolck, M., Oberwimmer, R., Ketikidis, C., Vasiljevic, A., Stauder, M., Mühlberg, C., Derczeni, R., Oravec, M., Krissakova, I., and Handlos, M. (2015). SWOT Analysis and Strategy Development for Forest Fuel Supply Chains in South East Europe. For. Policy and Economics 61: 87–94.

Rehman, S., Al-Hadhrami, L. M., and Alam, M. M. (2015). Pumped Hydro Energy Storage System: A Technological Review. Renewable and Sustainable Energy Review 44: 586–598.

Ribeiro, F., Ferreira, P., Araújo, M. (2011). The Inclusion of Social Aspects in Power Planning. Renewable and Sustainable Energy Reviews 15(9): 4361–4369.

Ritchie, J. R. B., Molinar, C. M. A., and Frechtling, D. C. (2010). Impacts of the World Recession and Economic Crisis on Tourism: North America. Journal of Travel Research 49: 5‒15.

Rivard, B., and Yatchew, A. (2016). Integration of Renewables into the Ontario Electricity System. The Energy Journal 37: 221–42.

213

Rosenbloom, D., and Meadowcroft, J. (2014). The Journey Towards Decarbonization: Exploring Socio-Technical Transitions in the Electricity Sector in the Province of Ontario (1885-2013) And Potential Low-Carbon Pathways. Energy Policy 65: 670–679.

Rosso, M., Bottero, M., Pomarico, S., Ferlita S. L., and Comino, E. (2014). Integrating Multicriteria Evaluation and Stakeholders Analysis for Assessing Hydropower Projects. Energy Policy 67: 870–881.

Saaty, T. L. (1977). A Scaling Method for Priorities in Hierarchical Structures. Journal of Mathematical Psychology 15: 234-281.

Saaty, T. L. (1980). The Analytic Hierarchy Process, McGraw-Hill, New York.

Saaty, T. L. (1988). What is Analytic Hierarchy Process? Mathematical Models for Decision Support 48: 109–121.

Saaty, T. L. (1990). How to Make a Decision: The Analytic Hierarchy Process. European Journal of Operational Research 48:9–26.

Saaty, T. L. (1996). Decision Making with Dependence and Feedback: The Analytic Network Process, RWS Publications.

Saaty, T. L. (2004). Decision making—the Analytic Hierarchy and Network Processes (AHP/ANP). Journal of Systems Science and Systems Engineering 13: 1–35.

Saaty, T. L. (2008). Relative Measurement and Its Generalization in Decision Making: Why Pairwise Comparisons Are Central in Mathematics for The Measurement of Intangible Factors the Analytic Hierarchy/Network Process. Revista de la Real Academia de Ciencias Exactas, Fisicas y Naturales. Serie A. Matematicas 102(2): 251–318.

Saaty, T. L., and Takizawa, M. (1986). Dependence and Independence: From Linear Hierarchies to Nonlinear Networks. Eur. J. Oper. Res. 26: 229–237.

Sagrado, J., Sanchez, J. A., Rodriguez, F., and Berenguel, M. (2014). Bayesian Networks for Greenhouse Temperature Control. Journal of Applied Logic 17: 25–35.

Salevid, K. (2013). Market Requirements for Pumped Storage Profitability, <https://uu.diva-portal.org/smash/get/diva2:661286/FULLTEXT01.pdf> (Jan 03, 2017).

Sarkar, A. U., Karagöz, S. (1995). Sustainable Development of Hydroelectric Power. Energy 20: 977–981.

Saunders, W. S. A., and Becker, J. S. (2015). A Discussion of Resilience and Sustainability: Land Use Planning Recovery from the Canterbury Earthquake Sequence, New Zealand. International Journal of Disaster Risk Reduction 14: 73–81.

Scannapieco, D., Naddeo, V., Belgiorno, V. (2014). Sustainable Power Plants: A Support Tool for the Analysis of Alternatives. Land Use Policy 36: 478–484.

214

Schindler, D. W. (1998). A Dim Future for Boreal Waters and Landscapes. BioScience 48(3): 157–164.

Schindler, D. W. (2001). The Cumulative Effects of Climate Warming and Other Human Stresses on Canadian Freshwaters in the New Millennium. Can J. Fish. Aquat. Sci. 58: 18–29.

Schindler, D. W., Bayley, S. E., Parker, B. R., Beaty, K. G., Cruikshank, D. R., Fee, E. J., Schindler, E. U., and Stainton, M. P. (1996). The Effects of Climatic Warming on the Properties of Boreal Lakes and Streams at the Experimental Lakes Area, Northwestern Ontario. Limnology and Oceanography 41(5): 1004–1017.

Schindler, D. W., Beaty, K. G., Fee, E. J., Cruikshank, D. R., DeBruyn, E. D., Findlay, D. L., Linsey, G. A., Shearer, J. A., Stainton, M. P., and Turner, M. A. (1990). Effects of Climatic Warming on Lakes of the Central Boreal Forest. Science 250(4983): 967–970.

Schlömer, S., Bruckner, T., Fulton, L., Hertwich, E., McKinnon, A., Perczyk, D., Roy, J., Schaeffer, R., Sims, R., Smith, P. and Wiser, R. (2014). Annex III: Technology-Specific Cost and Performance Parameters. In: Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Edenhofer, O., Pichs-Madruga, R., Sokona, Y., Farahani, E., Kadner, S., Seyboth, K., Adler, A., Baum, I., Brunner, S., Eickemeier, P., Kriemann, B., Savolainen, J., Schlömer, S., von Stechow, C., Zwickel T. and Minx J. C. (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Schmalz, M., and Thürmer, K. (2012). Long Term Investigations at the Small Hydropower Plant Döbritschen/ Germany. Proc Hidro Energia Conference.

Schwab, J., Topping, K. C., Eadie, C. C., Deyle, R. E., and Smith, R. A. (1998). Planning for Post-Disaster Recovery and Reconstruction. American Planning Association, Chicago, Illinois.

Sedoff, A., Schott, S., and Karney, B. (2014). Sustainable Power and Scenic Beauty: The Niagara River Water Diversion Treaty and Its Relevance Today. Energy Policy 66: 526–536.

Seghezzo, L. (2009). The Five Dimensions of Sustainability. Environmental Politics 18: 539–556.

Seifollahi-Aghmiuni, S., Bozorg Haddad, O., and Mariño, M. (2016). Generalized Mathematical Simulation Formulation for Reservoir Systems. J. Water Resour. Plann. Manage., 10.1061/(ASCE)WR.1943-5452.0000618, 04016004.

SENES Consultants Ltd. (2011). Toronto’s Future Weather & Climate Driver Study.

Shahabi, R. S., Basiri, M. H., Kahag, M. R., and Zonouzi, S. A. (2014). An ANP–SWOT Approach for Interdependency Analysis and Prioritizing the Iran’s Steel Scrap Industry Strategies. Resources Policy 42: 18–26.

Shamir, E., Megdal, S. B., Carrillo, C., Castro, C. L., Chang, H. I., Chief, K., Corkhill, F. E., Eden, S., Georgakakos, K. P., Nelson, K. M., and Prietto, J. (2015). Climate Change and Water Resources Management in the Upper Santa Cruz River, Arizona. Journal of Hydrology 521: 18–33.

215

Shapiro, S. S. and Wilk, M. B. (1965). An Analysis of Variance Test for Normality. Biometrika 52(3–4): 591–611.

Sharma, R. N., Chand, N., Sharma, V. and Yadav, D. (2015). Decision Support System for Operation, Scheduling and Optimization of Hydropower Plant in Jammu and Kashmir Region. Renewable and Sustainable Energy Review 43: 1099–1113.

Sharma, S., Gray, D. K., Read, J. S., O’Reilly, C. M., Schneider, P., Qudrat, A., Gries, C., Stefanoff, S., Hampton, S. E., Hook, S., Lenters, J. D., Livingstone, D. M., McIntyre, P. B., Adrian, R., Allan, M. G., Anneville, O., Arvola, L., Austin, J., Bailey, J., Baron, J. S., Brookes, J., Chen, Y., Daly, R., Dokulil, M., Dong, B., Ewing, K., Eyto, E. De, Hamilton, D., Havens, K., Haydon, S., Hetzenauer, H., Heneberry, J., Hetherington, A. L., Higgins, S. N., Hixson, E., Izmest’eva, L. R., Jones, B. M., Kangur, K., Kasprzak, P., Köster, O., Kraemer, B. M., Kumagai, M., Kuusisto, E., Leshkevich, G., May, L., MacIntyre, S., Müller-Navarra, D., Naumenko, M., Noges, P., Noges, T., Niederhauser, P., North, R. P., Paterson, A. M., Plisnier, P.-D., Rigosi, A., Rimmer, A., Rogora, M., Rudstam, L., Rusak, J. A, Salmaso, N., Samal, N. R., Schindler, D. E., Schladow, G., Schmidt, S. R., Schultz, T., Silow, E. A, Straile, D., Teubner, K., Verburg, P., Voutilainen, A., Watkinson, A., Weyhenmeyer, G. A, Williamson, C. E., and Woo, K. H. (2015). A Global Database of Lake Surface Temperatures Collected by in Situ and Satellite Methods from 1985–2009. Scientific Data 2: 1–6.

Sherman, L. K. (1932). Stream-Flow from Rainfall by the Unit-Graph Method. Engineering News Record, 108, 501–505.

Shlozberg, R., Dorling, R., and Sprio, P. (2014). Low Water Blues Steering Committee, <https:// mowatcentre.ca/wp-content/uploads/publications/89_low_water_blues.pdf> (May 3, 2015).

Shrestha, R. K., Alavalapati, J. R. R., and Kalmbacher, R. S. (2004). Exploring the Potential for Silvopasture Adoption in South-Central Florida: An Application of SWOT-AHP Method. Agricultural Systems 81: 185–199.

Silva, G. D. O., and Hendrick, P. (2016). Pumped Hydro Energy Storage in Buildings. Applied Energy 179: 1242–1250.

Singh, B., Strømman A. H., and Hertwich, E. G. (2012). Scenarios for the Environmental Impact of Fossil Fuel Power: Co-Benefits and Trade-Offs of Carbon Capture and Storage. Energy 45: 762–770.

Sioshansi, R. (2010). Welfare Impacts of Electricity Storage and the Implications of Ownership Structure. International Association for Energy Economics 31(2): 173–198.

Sioshansi, R., Denholm, P., Jenkin, T., and Weiss, J. (2009). Estimating the Value of Electricity Storage in PJM: Arbitrage and Some Welfare Effects. Energy Economics 31(2): 269–277.

Skondras, N.A., and Karavitis, C. A. (2015). Evaluation and Comparison of DPSIR Framework and the Combined SWOT – DPSIR Analysis (CSDA) Approach: Towards Embracing Complexity. Global NEST Journal 17.

216

Solar Trade Association (2016). Solar + storage = opportunities, <http://www.solar-trade.org.uk/solar-storage-opportunities/> (Jan 03, 2017).

Songsore, E., and Buzzelli, M. (2014). Social Responses to Wind Energy Development in Ontario: The Influence of Health Risk Perceptions and Associated Concerns. Energy Policy 69: 285–296.

Soundharajan, B., Adeloye, A. J., and Remesan, R. (2016). Evaluating the Variability in Surface Water Reservoir Planning Characteristics During Climate Change Impacts Assessment. Journal of Hydrology 538: 625–639.

Sousa, J. A. M., Teixeira, F., and Faias, S. (2014). Impact of a Price-Maker Pumped Storage Hydro Unit on the Integration of Wind Energy in Power Systems. Energy 69: 3–11.

Sousounis, P. J. and Bisanz, J. M. (eds.) 2000. Preparing for Climate Change: The Potential Consequences of Climate Variability and Change, US EPA, Ann Arbor, MI.

Sparkes, S. (2014). Sustainable Hydropower Development: Theun-Hinboun Expansion Project Case Study, Laos. Water Resources and Rural Development 4: 54–66.

Standing Committee on Resource Stewardship (2013). Review of the Potential for Expanded Hydroelectric Energy Production in Northern Alberta, Legislative Assembly of Alberta.

Statistics Canada (2009). Catalogue 57-206-XIB. Ottawa, <http://publications.gc.ca/Collection/Statcan/index-e.html.> (Oct. 01, 2009).

Statistics Canada (2012a). Table 127-0009 Installed Generating Capacity, by Class of Electricity Producer, <www.statcan.gc.ca> (Feb. 14, 2015).

Statistics Canada (2012b). Table 127-0002 Electric Power Generation, by Class of Electricity Producer, <www.statcan.gc.ca> (Sep. 14, 2015).

Stavrovsky, E., Krasilnikova, M., and Pryadko, S. (2013). AHP and ANP as Particular Cases of Markov Chains. IFAC Proceedings Volumes 46(9): 531–536.

Stefanovic, L. S., and Kreymborg, L. R. (2004). Red River CWMS Watershed Modeling. The World Water and Environmental Resources Congress, Utah, USA, 1–10.

Steffen, B. (2012). Prospects for Pumped-Hydro Storage in Germany. Energy Policy 45: 420–429.

Steffen, B., and Weber, C. (2016). Optimal Operation of Pumped-Hydro Storage Plants with Continuous Time-Varying Power Prices. European Journal of Operational Research 252: 308–321.

Sternberg, R. (2008). Hydropower: Dimensions of Social and Environmental Coexistence. Renewable and Sustainable Energy Reviews 12(6): 1588–1621.

Stevović, S., Milovanovic, Z., and Stamatovic, M. (2015). Sustainable Model of Hydro Power Development—Drina River Case Study. Renewable and Sustainable Energy Reviews 50: 363–371.

217

Stewart, R. A., Mohamed, S., and Daet, R. (2002). Strategic Implementation of IT/IS Projects in Construction: A Case Study. Automation in Construction 11: 681–694.

Stokes, L. C. (2013). The Politics of Renewable Energy Policies: The Case of Feed-In Tariffs in Ontario, Canada. Energy Policy 56: 490–500.

StoRE. (2014). <www.store-project.eu/documents/results/en_GB/final-publishable-report> (Jan. 10, 2015).

Sun, P., and Nie, P. (2015). A Comparative Study of Feed-in Tariff and Renewable Portfolio Standard Policy in Renewable Energy Industry. Renewable Energy 74: 255–262.

Sun, T., and Feng, L. (2013). Multistage Analysis of Hydrological Alterations in the Yellow River, China. River research and applications 29: 991–1003.

Sundararagavan, S., and Baker, E. (2012). Evaluating Energy Storage Technologies for Wind Power Integration.” Solar Energy 86(9): 2707–2717.

Supriyasilp, T., Pongput, K., and Boonyasirikul, T. (2009). Hydropower Development Priority Using MCDM Method. Energy Policy 37: 1866–1875.

Swift, J., and Stewart, K. (2004). Hydro: The Decline and Fall of Ontario’s Electric Empire, Between the Lines, Ontario.

Swingland, I. (2003). Capturing Carbon and Conserving Biodiversity: The Market Approach. Routledge Publisher.

Taghian, M., Rosbjerg, D., Haghighi, A., and Madsen, H. (2014). Optimization of Conventional Rule Curves Coupled with Hedging Rules for Reservoir Operation. J. Water Resour. Plann. Manage., 10.1061/(ASCE)WR.1943-5452.0000355, 693-698.

Tahseen, S. and Karney, B. (2016). Exploring the Multifaceted Role of Pumped Storage at Niagara. J. Water Resour. Plann. Manage., 10.1061/(ASCE)WR.1943-5452.0000666, 05016007.

Tahseen, S. and Karney, B. (2017). Increased Hydropower Potential at Niagara: A Scenario-based Analysis. Journal of Water Resources Planning and Management. (Under review)

Tahseen, S., and Karney, B. (2017). Reviewing and Critiquing Published Approaches to the Sustainability Assessment of Hydropower. Renewable and Sustainable Energy Review 67: 225-234.

Taylor, M. E., Gray, P. A, and Schiefer, K. (2006). Helping Canadians Adapt to Climate Change in the Great Lakes Coastal Zone. Great Lakes Geographer 13(1): 14–25.

Teasley, R. L., Mckinney, D. C., and Patino–Gomez, C. (2004). Modeling the Forgotten River Segment of the Rio Grande/Barvo Basin. The University of Texas at Austin, Center for Research in Water Resources, Texas, USA.

218

Tebaldi, C., and Knutti, R. (2007). The Use of the Multi-Model Ensemble in Probabilistic Climate Projections. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 365(1857): 2053–2075.

The Conference Board of Canada (2011). Canada’s Electricity Infrastructure: Building a Case for Investment, <http://www.electricity.ca/media/ReportsPublications/11257_ElectricityInfras tructure%5B1%5D.pdf> (Oct. 01, 2015).

The German Advisory Council on the Environment (SRU) (2011). <http://www.umweltrat.de/SharedDocs/Downloads/EN/02_Special_Reports/2011_10_Special_Report_Pathways_renewables.pdf?__blob=publicationFile> (Aug. 08, 2015).

The Nature Conservancy (2009). Indicators of Hydrologic Alteration 7.1 User’s Manual.

The NY Times (2006). <http://www.nytimes.com/2006/07/18/science/18NIAG.html?_r=0> (Mar. 03, 2013).

Thomas, H. A., and Fiering, M. B. (1962). Mathematical Synthesis of Streamflow Sequence for The Analysis of River Basins by Simulation. Design of Water Resources System, Harvard University Press, Cambridge.

Thompson, P. B. (1992). The Varieties of Sustainability. Agriculture and Human Values 9: 11–19.

Thompson, P. B. (2010). The Agrarian Vision: Sustainability and Environmental Ethics, The University Press of Kentucky.

Tidball, R., Bluestein, J., Rodriguez, N., Knoke, S., Tidball, R., Bluestein, J., and Rodriguez, N. (2010). Cost and Performance Assumptions for Modeling Electricity Generation Technologies, <www.nrel.gov/docs/fy11osti/48595.pdf> (Dec. 13, 2016).

Tremblay, A., Varfalvy, L., Roehm, C., and Garneau, M. (2006). The Issue of Greenhouse Gases from Hydroelectric Reservoirs: From Boreal to Tropical Regions. Environmental Protection.

Tsanis, I. K., and Apostolaki, M. G. (2009). Estimating Groundwater Withdrawal in Poorly Gauged Agricultural Basins. Water Resources Management 23(6): 1097–1123.

Tsanis, I. K., Koutroulis, A. G., Daliakopoulos, I. N., and Jacob, D. (2011). Severe Climate-Induced Water Shortage and Extremes in Crete. Climatic Change 106(4): 667–677.

Tuohy, A., and O’Malley, M. (2011). Pumped Storage in Systems with Very High Wind Penetration. Energy Policy 39(4): 1965–1974.

U.S. Army Corps of Engineers: Hydrologic Engineering Center (USACE-HEC) (2015a). HEC-ResSim, <http://www.hec.usace.army.mil/software/hec-ressim/> (Aug. 30, 2015).

U.S. Army Corps of Engineers: Hydrologic Engineering Center (USACE-HEC) (2015b). HEC-GeoRAS, <http://www.hec.usace.army.mil/software/hec-georas/> (Aug. 30, 2015).

219

Uchiyama, Y. (1996). Life Cycle Analysis of Electricity Generation and Supply Systems, Net Energy Analysis and Greenhouse Gas Emissions. Symposium of Electricity, Health and The Environment: Comparative Assessment in Support of Decision Making.

United Nations (2014). Suggested Elements for the Post-2015 Framework for Disaster Risk Reduction. Third United Nations World Conference on Disaster Risk Reduction Preparatory Committee, Geneva.

United States Department of Energy (2015). Pumped Storage and Potential Hydropower from Conduits.<https://energy.gov/sites/prod/files/2015/06/f22/pumped-storage-potential-hydropower-from-conduits-final.pdf> (Dec. 13, 2016).

UNWTO World Tourism Barometer (2009). <http://www.unwto.org/facts/eng/pdf/barometer/UNWTO_Barom09 _1_en.pdf> (Jan. 23, 2014).

US Army Corps of Engineers (USACE) (2012). Determination of a Hydrologic Index for the Russian River Watershed Using HEC-Ressim, San Francisco, CA.

US Environmental Protection Agency (US EPA) (2015). Great Lakes, <http://www.epa.gov/greatlakes/basicinfo.html> (Aug. 31, 2015)

US Environmental Protection Agency (US EPA) (2016). <https://www.epa.gov/climate-change-science/future-climate-change> (Jul. 03, 2016).

US Fish and Wildlife Service (2015). <http://www.fws.gov/midwest/fisheries/Library/fact-ecoteam.pdf> (Jun. 5, 2015).

United States Global Change Research Program (USGCRP) (2009). Global Climate Change Impacts in the United States, Cambridge University Press, NY, USA.

Ux Consulting Company. (2017). <https://www.uxc.com/p/prices/UxCPrices.aspx> (Jan. 03, 2017).

Van de Vate, J. F. (2002). Full-Energy-Chain Greenhouse-Gas Emissions: A Comparison Between Nuclear Power, Hydropower, Solar Power and Wind Power. International Journal of Risk Assessment and Management 3(1): 59‒74.

Van Vliet M. T. H., Wiberg, D., Leduc, S., and Riahi, K. (2016). Power-Generation System Vulnerability and Adaptation to Changes in Climate and Water Resources. Nature Climate Chang 6: 375–380.

Varkani, A. K., Daraeepour, A., and Monsef, H. (2011). A New Self-Scheduling Strategy for Integrated Operation of Wind and Pumped-Storage Power Plants in Power Markets. Applied Energy 88(12): 5002–5012.

Viegas, M. D. C., Moniz, A. B., and Santos, P. T. (2014). Artisanal Fishermen Contribution for The Integrated and Sustainable Coastal Management – Application of Strategic SWOT Analysis. Procedia - Social and Behavioral Sciences 120: 257–267.

220

Vliet, M. T. H., Wiberg, D., Leduc, S., and Riahi, K. (2016). Power-Generation System Vulnerability and Adaptation to Changes in Climate and Water Resources. Nature Climate Change 6: 375–380.

Vučijak, B., Kupusović, T., MidŽić-Kurtagić, S., and Ćerić A. (2013). Applicability of Multicriteria Decision Aid to Sustainable Hydropower. Applied Energy 101: 261–267.

Wagner, B., Hauer, C., Schoder, A., and Habersack, H. (2015). A Review of Hydropower in Austria: Past, Present and Future Development. Renewable and Sustainable Energy Reviews 50: 304–314.

Wakiyama, T., Zusman, E., and Monogan, J. E. (2014). Can a Low-Carbon-Energy Transition Be Sustained in Post-Fukushima Japan? Assessing the Varying Impacts of Exogenous Shocks. Energy Policy 73: 654–666.

Wang, G. Q., and Zhang, J. Y. (2015). Variation of Water Resources in the Huang-Huai-Hai Areas and Adaptive Strategies to Climate Change. Quaternary International 380-381: 180–186.

Wang, J., Bai, X., Hu, H., Clites, A., Colton, M., and Lofgren, B. (2012). Temporal and Spatial Variability of Great Lakes Ice Cover, 1973-2010. Journal of Climate 25(4): 1318–1329.

Wang, J., Botterud, A., Bessa, R., Keko, H., Carvalho, L., Issicaba, D., Sumaili, J., and Miranda, V. (2011a). Wind Power Forecasting Uncertainty and Unit Commitment. Applied Energy 88(11): 4014–4023.

Wang, W., Du, X. and Lu, Z. (2011b). Analysis of the Cumulative Effect of Pollution in the Atmospheric Environment Management Based on the Method of ANP Embedding into SWOT. International Conference on Remote Sensing, Environment and Transportation Engineering: 65–68.

Wang, Z., and Feng, C. (2015). A Performance Evaluation of the Energy, Environmental, And Economic Efficiency and Productivity in China: An Application of Global Data Envelopment Analysis. Applied Energy 147: 617–626.

Wang, Z., Han, B., and Lu, M. (2016). Measurement of Energy Rebound Effect in Households: Evidence from Residential Electricity Consumption in Beijing, China. Renewable and Sustainable Energy Reviews 58: 852–861.

Wang, Z., Wang, C., and Yin, J. (2014). Strategies for Addressing Climate Change on the Industrial Level: Affecting Factors to CO2 Emissions of Energy-Intensive Industries in China. Natural Hazards 75: 303–317.

Wang, Z., and Yang, L. (2015). Delinking Indicators on Regional Industry Development and Carbon Emissions: Beijing-Tianjin-Hebei Economic Band Case. Ecological Indicators 48: 41–48.

Wang, Z., Yin, F., Zhang, Y., and Zhang, X. (2012). An Empirical Research on the Influencing Factors of Regional CO2 Emissions: Evidence from Beijing City, China. Applied Energy 100: 277–284.

221

Weeraratne, J. R., Logan, L., and Unny, T. E. (1986). Performance Evaluation of Alternate Policies on Reservoir System Operation. Canadian Journal of Civil Engineering 13: 203–212.

Werkheiser, I., and Piso, Z. (2015). People Work to Sustain Systems: A Framework for Understanding Sustainability. Water Resource Planning and Management 141(12): A4015002.

Weron, R., and Misiorek, A. (2008). Forecasting Spot Electricity Prices: A Comparison of Parametric and Semiparametric Time Series Models. International Journal of Forecasting 24: 744–763.

White, R. (1985). Ontario 1610–1985: A Political and Economic History, Dundurn Press, Toronto.

Wilcox, D. A, Thompson, T. A, Booth, R. K., and Nicholas, J. R. (2007). Lake-Level Variability and Water Availability in the Great Lakes, U.S. Geological Survey, Reston, Virginia.

Williams, A., and Porter, S. (2006). Comparison of Hydropower Options for Developing Countries with Regard to the Environmental, Social and Economic Aspects. International conference on renewable energy for developing countries.

Williams, C. C., and Millington, A. C. (2004). The Diverse and Contested Meanings of Sustainable Development. The Geographical Journal 170: 99–104.

Wilson, E. M. (1990). Engineering Hydrology, Macmillan, Indianapolis, Indiana.

Wilson, M. A., and Browning, C. J. (2012). Investing in Natural Infrastructure: Capacity in the Face of a Changing Climate. Ecological Restoration 30(2): 96–98.

Wind, Y. (1987). An Analytic Hierarchy Process Based Approach to The Design and Evaluation of a Marketing Driven Business and Corporate Strategy. Mathematical Modeling 9: 285– 290.

Winfield, M., and Dolter, B. (2014). Energy, Economic and Environmental Discourses and Their Policy Impact: The Case of Ontario’s Green Energy and Green Economy Act. Energy Policy 68: 423–435.

Wong, S. L., Ngadi, N., Abdullah, T. A. T., and Inuwa, I. M. (2015). Recent Advances of Feed-in Tariff in Malaysia. Renewable and Sustainable Energy Reviews 41: 42–52.

World Bank (2015). <http://data.worldbank.org/indicator/EG.ELC.ACCS.ZS> (Sept. 08, 2015).

World Bank. (2007). Estimating Global Climate Change Impacts on Hydropower Projects: Application in India, Sri Lanka and Vietnam.

World Bank. (2013). Directions in Hydropower.

World Commission on Environment and Development (WCED) (1987). Our Common Future (The Brundtland Report), Oxford University Press.

222

World Nuclear Association (2016). <http://www.world-nuclear.org/information-library/facts-and-figures/heat-values-of-various-fuels.aspx> (Dec. 16, 2016).

World Small Hydropower Development Report (2013). United Nations Industrial Development Organization & International Center on Small Hydro Power, <www.smallhydroworld.org> (Sep. 08, 2015).

Wu, C. S., Yang, S. L., and Lei, Y. (2012). Quantifying the Anthropogenic and Climatic Impacts on Water Discharge and Sediment Load in the Pearl River (Zhujiang), China (1954-2009). Journal of Hydrology 452-453: 190–204.

Wu, X., Cheng, C., Zeng, Y., and Lund, J. (2016). Centralized Versus Distributed Cooperative Operating Rules for Multiple Cascaded Hydropower Reservoirs. Journal of Water Resource Planning and Management, 10.1061/(ASCE)WR.1943-5452.0000685, 05016008.

Xingang, Z., Lu, L., Xiaomeng, L., Jieyu, W., and Pingkuo, L. (2012). A Critical-Analysis on the Development of China Hydropower. Renewable Energy 44: 1–6.

Yang, Z., Yan, Y., and Liu, Q. (2012). Assessment of the Flow Regime Alterations in the Lower Yellow River, China. Ecological Informatics 10: 56–64.

Yatchew, A., and Baziliauskas, A. (2011). Ontario Feed-in-Tariff Programs. Energy Policy 39(7): 3885–3893.

Yuksel, I. (2010a). As a Renewable Energy Hydropower for Sustainable Development in Turkey. Renewable and Sustainable Energy Reviews 14: 3213–3219.

Yuksel, I. (2010b). Hydropower for Sustainable Water and Energy Development. Renewable and Sustainable Energy Reviews 14(1): 462–469.

Yüksel, İ., and Dagˇdeviren, M. (2007). Using the Analytic Network Process (ANP) in a SWOT Analysis - A Case Study for a Textile Firm. Information Sciences 177: 3364–3382.

Zafirakis, D., Chalvatzis, K. J., Baiocchi, G., and Daskalakis, G. (2013). Modeling of Financial Incentives for Investments in Energy Storage Systems That Promote the Large-Scale Integration of Wind Energy. Applied Energy 105: 138-154.

Zahedi, G., Azizi, S., Bahadori, A., Elkamel, A., and Alwia, S.R.W. (2013). Electricity Demand Estimation Using an Adaptive Neuro-Fuzzy Network: A Case Study from The Ontario Province – Canada. Energy 49: 323–328.

Zareipour, H., Bhattacharya, K., and Cañizares, C. A. (2007). Electricity Market Price Volatility: The Case of Ontario. Energy Policy 35: 4739–4748.

Zareipour, H., Cañizares, C. A., Bhattacharya, K., and Thompson, J. (2006). Application of Public-Domain Market Information to Forecast Ontario’s Wholesale Electricity Prices. IEEE Transactions on Power Systems 21(4): 1707 – 1717.

223

Zeng, M., Zhang, K., and Liu, D. (2013). Overall Review of Pumped-Hydro Energy Storage in China: Status Quo, Operation Mechanism and Policy Barriers. Renewable and Sustainable Energy Reviews 17: 35–43.

Zgrzywa, A. Choroś, K., and Siemiński, A. (2008). New Trends in Multimedia and Network Information systems, IOS Press, Amsterdam, Netherlands.

Zhang, J., Xu, L., and Li, X. (2015). Review on the Externalities of Hydropower: A Comparison Between Large and Small Hydropower Projects in Tibet Based on the CO2 Equivalent. Renewable and Sustainable Energy Reviews 50: 176–185.

Zhang, J., Xu, L., Yu, B., and Li, X. (2014). Environmentally Feasible Potential for Hydropower Development Regarding Environmental Constraints. Energy Policy 73: 552–562.

Zhang, Q., Karney, B., MacLean, H. L., and Feng, J. (2007). Life-Cycle Inventory of Energy Use and Greenhouse Gas Emissions for Two Hydropower Projects in China. Journal of Infrastructure Systems 13:271–279.

Zhang, S., Andrews-Speed, P., and Perera, P. (2015). The Evolving Policy Regime for Pumped Storage Hydroelectricity in China: A Key Support for Low-Carbon Energy. Applied Energy 150: 15-24.

Zhang, X., Cao, L., and Caldeira, K. (2013). Energy Switching Threshold for Climatic Benefits. AGU Fall Meeting Abstract 1: 1095.

Zhang, X., Chen, W., Ma, C., and Zhan, S. (2013). Modeling Particulate Matter Emissions During Mineral Loading Process Under Weak Wind Simulation. Science of the Total Environment 449: 168–173.

Zhang, X., Fan, D., Wang, H., and Yang, Z. (2014). Water Discharge Variability of Changjiang (Yangtze) and Huanghe (Yellow) Rivers and its Response to Climatic Changes. Chinese Journal of Oceanology and Limnology 32(6): 1392–1405.

Zhang, X., Myhrvold, N. P., and Caldeira, K. (2014a). Key Factors for Assessing Climate Benefits of Natural Gas Versus Coal Electricity Generation. Environmental Research Letters 9: 114022.

Zhang, X., Vincent, L. a., Hogg, W. D., and Niitsoo, A. (2000). Temperature and Precipitation Trends in Canada During the 20th Century. Atmosphere-Ocean 38(3): 395–429.

Zhang, Z. (2003). Why Did the Energy Intensity Fall in China’s Industrial Sector in the 1990s? The Relative Importance of Structural Change and Intensity Change. Energy Economics 25: 625–638.

Zhao, T., Zhao, J., Lund, J., and Yang, D. (2014). Optimal Hedging Rules for Reservoir Flood Operation from Forecast Uncertainties. Journal of Water Resources Planning and Management, 10.1061/ (ASCE)WR.1943-5452.0000234.

224

Zhu, J., Jordan G., and Ihara, S. (2000). The Market for Spinning Reserve and Its Impacts on Energy Prices. IEEE Power Engineering Society Winter Meeting 2: 1202–1207.

Zoltay, V. (2007). Water Resources Management: Optimizing Within a Watershed Context. World Environmental and Water Resources Congress, Florida, USA.

Zoran, D., Saša, M., and Dragi, P. (2011). Application of the AHP Method for Selection of a Transportation System in Mine Planning. Underground Mining Engineering 19: 93–99.

225

Appendices A. Overview of Canada’s Electricity Sector

A.1 Electricity Sector Overview Canada’s energy sector has an installed capacity of 127.8 GW, providing electricity to the entire

population (Canadian Electricity Association 2014a; World Bank 2015). The major sources of

energy across Canada are hydro, fossil fuels, and nuclear which make up 59.3, 24.1, and 10.5% of

the total generation, respectively; the balance comes from a combination of wind, solar, and tidal

(Figure A.1). The energy mix varies substantially, with British Columbia, Manitoba, Quebec,

Newfoundland and Labrador’s energy generated predominantly from hydroelectric sources.

Alberta, Saskatchewan, Nova Scotia and New Brunswick depend heavily on fossil fuel. Among

other renewable sources, wind accounts for 4.2% of the total installed capacity, but has shown

rapid growth over the past decades (The Conference Board of Canada 2011). Table A.1 illustrates

the mix of generation within major Canadian provinces.

Source: Canadian Electricity Association (2014)

Canada has a predominately north-south transmission network that connects most strongly to the

United States (NREL 2013). The transmission grid, apart from facilitating interprovincial trade,

plays a key role in exporting electricity to the U.S. market. Overall, Canada exports between 7 and

9% of its power generation and has traditionally been a net electricity exporter (Government of

Canada 2008).

75.8

30.8

13.4

7.8

Hydropower

Fossil fuel

Nuclear

Others

Figure A.10.1: Installed electricity capacity by source (GW)

226

Table A.1: Total electricity generation by provinces in 2013 (TWh)

Sources B.C. Alta. Sask. Man. Ont. Que. N.S. N.B. N.L. P.E.I.

Hydro 58.2 2.2 4.5 35.4 36.7 205 1.1 3.3 40.7 0 Nuclear 0 0 0 0 93.1 0 0 3.9 0 0 Conventional Steam 4.5 44.9 17.2 0.1 7.1 0.9 8.5 4.4 1 0 Internal Combustion 0.1 0.1 ~0 ~0 0.8 0.3 0 0 0.1 0 Combustion Turbine 1.2 13.7 0.7 ~0 9.38 0.4 0.5 1.9 0.3 0 Wind 0 2.3 0.7 0.4 3.3 0.7 0.1 0.6 0.1 0.5 Solar 0 0 0 0 0.24 0 0 0 0 0 Total 64.1 63.6 23.1 35.9 150 207 10.5 14 42.1 0.9

Source: Canadian Electricity Association (2014b)

In Canada, regulatory and policy control over the electricity industry are primarily vested

provincially. Provincial governments have ownership over generation assets, especially hydro,

nuclear, and conventional steam plants. Generation and transmission are often provided through a

public entity (BC, Quebec, Manitoba) or produced by a competitive, bidding process as is found

in Alberta and Ontario (Government of Canada. 2008). The private sector nevertheless, in all

provinces, owns an important share of the generation capacity. Table 2.2 reflects the ownership

distribution among various generation sources. The national transmission grid is a collection of

relatively loosely-connected provincial grids that are linked together through varying levels of

intertie capacity. British Columbia, Manitoba, Ontario, and Quebec have the largest external

connections to the regional U.S. markets. The system operator coordinates power flows in real

time, and the entity that acts as system operator depends on the provincial market structure.

Ontario, Alberta and New Brunswick have Independent System Operator (ISO); in most other

provinces the operator also owns transmission assets (The Conference Board of Canada 2011).

At the federal level, the stated plan has been to develop a green energy sector that, apart from

having employment benefits, will help to meet the emission targets. The government projects to

double non-hydro renewable sources as well as the retirement of coal-fired power plants (National

Energy Board 2013). With a recent change in Federal government in Oct. 2015, it will be

interesting to see how national goals will evolve. Ontario has already eliminated coal generation

and other provinces (Alberta and Saskatchewan in particular) face pending federal regulations.

Several provinces pursue demand-side management programs, and are leaning towards smart grid

227

investments to support the behavioural shifts. Some have taken steps in that direction by installing

smart meters (Ontario Ministry of Energy 2013a).

Canadians enjoy some of the lowest residential energy prices among Organization for Economic

Co-operation and Development (OECD) countries (OECD 2004). Each province has its own

electricity policy and regulatory agency, leading to disparate electricity tariffs. Quebec, BC,

Manitoba and Newfoundland and Labrador produce 56% of the Canadian electricity almost

exclusively from hydropower plants (Statistics Canada 2012b). Given the low operational cost of

their generation portfolio, these provinces have the lowest electricity rates in Canada. The lack of

hydropower potential for Alberta, Ontario and New Brunswick led to a reliance on thermal

generation (fossil fuel and nuclear), leading to higher production costs.

Table A.2: Ownership distribution (%) over generation assets in 2009

Government Investor Industry Hydro 87 6 7 Wind 7 91 2 Nuclear 62 38 0 Combustion turbine 36 50 7 Conventional steam 56 37 7

Source: Statistics Canada (2012a)

Provinces have separate regulation entities for reviewing and approving plans. In a majority of

provinces, utilities are operating as regulated monopolies with the exception of Ontario and

Alberta which have at least partly deregulated their electric industry over the last decade. A few

key responsibilities are still handled by the federal government such as issuing permits for inter-

provincial and international power lines, assessment for major hydroelectric developments etc.

(National Energy Board 2015). The federal government retains some oversight and permit

responsibilities on issues relating to fisheries.

A.2 Small Hydropower Sector Overview and Potential Natural Resources Canada (2007) defines Small Hydropower (SHP) as 50 MW of generating

capacity (Government of Canada 2014). However, in the absence of international convention, a

228

10‒15 MW limit can also be seen. Installed capacity of small hydropower in Canada is 3,400 MW

(up to 50 MW) while the potential is estimated to be 5,650 MW indicating that 60% has been

developed. Between the 2013 and 2016 World Small Hydropower Reports installed capacity has

increased for only 1% while estimated potential has decreased by approximately 25% (Figure A.2).

The 3,400 MW of SHP capacity (up to 50 MW) accounts for 4.5% of Canada’s total hydro capacity

(Natural Resources Canada 2007). There are an estimated 5,500 sites throughout Canada that are

technically feasible for small hydropower, totaling for a potential capacity of 11,000 MW (IEA

Small Hydro n.d.). But only 10-15% of this total is economically feasible (IEA Small Hydro n.d.).

Table 2.3 presents the existing SHP by province (up to 50 MW).

Figure A.10.2: Small hydropower capacities 2013-2016 in Canada (MW)

Source: Natural Resources Canada (2007); IEA Small Hydro (n.d.); World Small Hydropower Development Report

2013

As each province has a unique strategy, the plans for future development of small hydro varies

across jurisdictions. In British Columbia, the Standing Offer Program targets small producers of

electricity (less than 15 MW) and encourages them to sell electricity to BC Hydro, the publicly

owned utility (BC Hydro 2011). Québec, by contrast, does not have specific published plans to

develop small hydro. Since the province produces more electricity than it needs, their focus is to

refurbish their aging infrastructure (Hydro Québec 2015). Similarly, with most small hydro

installation in Ontario dating as early as 1990s, major investments are expected in replacing the

aging assets.

Ontario’s Feed-in Tariff (FIT) provides 0.246 CA$/kWh for small hydropower development under

500 kW. This price is subjected additional remuneration for aboriginal participation and on-peak

3372

7500

3400

5650

InstalledCapacity

(MW)

PotentialCapacity

(MW)

2016 2013

229

generation (Ontario Power Authority 2016a). Standing Offer Program in BC offers 0.09139

CA$/kWh (BC Hydro 2011).

Table A.3: Existing SHP capacity in Canada (MW)

B.C. Alta. Sask. Man Ont. Que. N.S. N.B. N.L. Yuk. NWT

Existing 568 424 7 23 498 345 469 43 803 127 67 Source: Statistics Canada (2009)

A.3 Renewable energy policy Due to the provincial dominance over electricity sector, there is a large variation in incentives

provided for clean, renewable development across different provinces. The schemes are also

subjected to frequent amendments and adjustments. A brief description of some of these renewable

policy measures are discussed below:

A.3.1 Clean energy fund Canada’s Economic Action Plan includes the Clean Energy Fund, a five-year, CA$ 795 million

program to support clean energy technology research (Natural Resources Canada 2011).

A.3.2 Standard offer programs The qualifying projects are subjected to a capacity restriction and required to connect to the

distribution. The program usually guarantees a sustained tariff for a period of 20 years.

A.3.3 Feed-in Tariff (FIT) programs The program assures priority grid connection and long term stable prices (40 years for hydropower

and 20 years for others) for electricity generated from renewable resources, subjected to capacity

restrictions. At present, FIT program is available only in Ontario. Within the first two years, it has

extended Ontario’s renewable capacity by 4,600 MW (Amin 2012).

A.3.4 Net metering Net metering allows generators to send the excess electricity, after their own use, to the grid. The

credit received in return can be applied against future electricity use or at times, can be subjected

to annual monetary returns. Net metering programs are available in almost every province across

Canada.

230

A.3.5 Requests for proposal A request for proposal (RFP) usually involves a specific target announced by the government that

needs to be executed by the monopoly utility in particular jurisdiction. The proponents bid

according to a fixed delivery schedule and are eligible to get a defined tariff rates which may or

may not be subjected to escalation.

A.4 Barriers to Small Hydropower Development Alberta has the most underdeveloped hydropower resources in the country, and as such they are

currently facing barriers to development that are not unique to the province. Firstly, they recognize

that the success of a hydro development depends on transboundary cooperation, between upstream

and downstream jurisdictions (Standing Committee on Resource Stewardship 2013). The

fragmented approach in almost all aspects of energy sector due to provinces’ own electricity policy

and renewable energy targets, has in some cases led to its underperformance. Another common

consideration for hydropower development throughout Canada is Aboriginal rights. Native

communities throughout the country are diverse, but often their livelihood depends on water

courses, therefore consultation of these stakeholders is important (Standing Committee on

Resource Stewardship 2013). A crucial step in this regard has been including the First Nations as

hydropower project partners. From a more technical perspective, however, ice formation can be a

particular change. The installation of hydropower generators can cause unpredictable ice

formations that cause damage to other infrastructure, such as transport passages and bridges

(Standing Committee on Resource Stewardship 2013).

Overall, the situation in Canada, as in most jurisdictions, is dynamic and hard to forecast in detail.

Many anticipate that a hydro renaissance is possible with hydro resources playing a larger role in

the quest for a more renewable, sustainable, stable and economical power system.

231

Appendices B. List of the Operating Rules Under Each Reservoir Table C.1: Operation set for baseline simulation

Reservoir Operating rules Description

Lake Erie

Flood control zone Historical maximum water level as a function of time

Navigation requirement Minimum 80 m3/s release through Welland Canal outlet

Power rule_DeCew_I Daily power requirements at DeCew I station

Power rule_DeCew_II Daily power requirements at DeCew II station

Welland_canal_max Maximum monthly flow through Welland Canal

Riverflow_max Elevation dependent maximum release through Niagara

Riverflow_min Elevation dependent minimum release through Niagara

Conservation zone

Same as flood control zone

Historical mean water level as a function of time

Inactive Historical minimum water level as a function of time

GIP

Flood control zone Set at 172.07 m

Treaty restriction_min Hourly minimum flow as per treaty guideline

GIP_elev

“If” logic sets a range of flow values (min to max) for 0

≤ daily_elev ≤ 0.46 and 0 ≤ monthly_elev ≤ 0.91.

Daily_elev and Monthly_elev are water level deviations

from 171.16 m, aggregated over a 24 hr and 720 hr

period respectively.

Elev_rate of change_day Maximum of 0.46 m elevation change in 24 hr

Elev_rate of change_mon Elevation change limited to 0.91 m in 720 hr

Treaty restriction_max Hourly flow as per treaty guideline through river outlet

Powerflow_tunnel_1 Scripted rule ensuring uniform flow distribution between

Canada and the US

Powerflow_tunnel_2 Same as before

Powerflow_tunnel_3 Same as before

232

Powerflow_tunnel_ canal Same as before

Powerflow_US tunnel Same as before

Conservation zone

Same as flood control

zone

Set 171.16 m

SAB PGS

Flood control zone Set at 189.6 m

Nighttime_pumping Pumping allowed only from 0 – 7 hr

Alternative pump_gen “If” logic sets generation to zero when pumping

Crossover_elev If crossover elevation <= 165, cease pumping or else

maximize

PGS tandem Tandem operation of PGS reservoir

Power rule_PGS Daily power requirements at PGS

Conservation zone

Same as flood control zone

Set at 185 m

Crossover

Flood control zone

Power rule_SAB_I

Set at 166 m

Daily power requirements at SAB I

Power rule_SAB_II Daily power requirements at SAB II

Conservation zone

Same as flood control zone

Set at 165 m

Inactive Set at 164.8 m

Lake

Ontario

Flood control zone

St. Lawrenceflow_min

Historical maximum water level as a function of time

Elevation dependent minimum release through river

St. Lawrenceflow_max Elevation dependent maximum release through river

Conservation

Same as flood control zone

Historical mean water level as a function of time

Inactive Historical minimum water level as a function of time

233

Appendices C. Pairwise Comparison Matrices for SWOT Sub-Factors Local Priorities

Eco O1 O2 O3 O4 O5

O1 1 7,0.5,0.17,0.14,0.1,1:2,3:2,0.

33:2,0.25:2,0.2:4; (1.1,1.8)

3,5,7,0.5,0.25,0.17,0.11,0.33:

6,0.2:2,0.14:2; (1.1,2)

1,5,7, 0.5,0.25,0.17,0.11,3:3,

0.33:5,0.2:2; (1.5,2)

9,0.5,0.25,0.17, 1:4, 3:2,4:2,

0.33:3, 0.2:2; (1.7,2.3)

O2 1 2,0.5,0.2,0.13,0.1, 1:6,

5:2,9:2,0.33:2; (2.2,3)

7,8,0.5,0.33,0.2,1:4,3:5,5:3;

(2.9,2.4)

4,5,6,7,0.33,0.2,1:2,2:4,3:5;

(2.8,1.9)

O3 1 2,4,7,0.5,0.3,0.1,1:4,3:5,0.2:2

; (2.2,1.6)

2,5,0.3,0.14,1:2,3:6,4:2,0.2:2;

(2.2,1.6)

O4 1 2,5,7,9,0.5,0.2,1:6,3:2,0.33:3;

(2.2,2.6)

O5 1

CR = 0.035

Env O6 O7 O8

O6 1 7,0.5,0.2,0.14,1:2,3:3,4:3,5:3,6:2; (3.4,2.2) 5,0.33,0.25,1:2,2:2,3:5,0.5:2,0.14:3; (1.6,1.5)

O7 1 1,7,0.5,0.14,0.11,3:3,0.33:4,0.2:4; (1.2,1.9)

O8 1

CR = 0.085

234

Soc O9 O10

O9 1 7,9,0.5,0.17,1:2,3:4,4:2,5:3,0.33:2; (3.2,2.5)

O10 1

CR = 0.00

Eco T1 T2 T3 T4 T5

T1 1 3,5,0.5,6:2; (4.1,2.36) 0.33,0.25,0.17,0.5:2;

(0.4,0.15)

2,5;2, 6:2; (4.8,1.52) 5,6:2,7:2; (6.2,0.89)

T2 1 2,3,9,0.14,1:4,5:2,0.33:4,0.

2:3;(1.8,2.4)

2,7,1:2,3:4,5:2,0.5:3,0.33:2,

0.2:2; (2.1,2.06)

2,3,5,8,0.5,0.2,0.17,1:3,4:3,

7:2,0.33:2; (2.9,2.7)

T3 1 4,5,7,0.33,0.2,0.17,1:2,3:5,

9:2,0.5:2; (3.1,2.9)

2,4,7,8,9,0.5,0.33,0.14,1:5,

3:2,5:2; (3.1,2.8)

T4 1 7,0.2,0.14,0.11,1:2,3:3,4:2,

0.5:2,0.33:2,0.25:2;

(1.7,1.9)

T5 1

CR = 0.16

Env T6 T7

T6 1 5,6,3:2,0.33:4,0.25:2,0.2:5,0.14:2; (1.2,1.9)

T7 1

CR = 0.00

235

Soc T8 T9

T8 1 5,0.5,0.17,1:4,3:2,9:2,0.33:2,0.25:4; (2.1,2.9)

T9 1

CR = 0.00

Eco S1 S2 S3 S4

S1 1 0.17,1:5,2:2,0.5:2,0.33:3,0.2:2,0.14

3:2; (0.7,0.6)

3,0.5,0.25,0.13,0.33:4,0.2:5,0.14:3;

(0.4,0.7)

1,4,5,2:2,3:4,9:2,0.33:4,0.2:2;

(2.7,2.8)

S2 1 2,4,5,0.5,0.11,1:2,0.33:3,0.2:6;

(1,1.5)

1,4,5,6,7,9,0.33,0.14,2:3,3:6;

(3.3,2.3)

S3 1 2,6,0.5,1:2,3:4,4:2,5:3,7:2; (3.7,2)

S4 1

CR = 0.05

Env S5 S6

S5 1 2,4,0.5,0.11,1:5,3:4,5:2,0.33:2; (2,1.6)

S6 1

CR = 0.00

236

Soc S7 S8

S7 1 2,3,0.25,0.14,1:3,5:3,0.5:4,0.33:3; (1.6,1.8)

S8 1

CR = 0.00

Eco W1 W2 W3

W1 1 5,9,0.14,1:3,3:4,0.5:2,0.33:2,0.2:3

; (1.8,2.3)

4,9,1:5,2:3,3:3,0.33:2,0.2:2;

(2,2.1)

W2 1 2,4,0.5,0.14,1:7,3:2,0.33:3;

(1.3,1.1)

W3 1

CR = 0.002

Env W4 W5

W4 1 6,0.5,0.25,1:5,3:2,0.33:4,0.2:3; (1.2,1.50)

W5 1

CR = 0.00

Soc W6 W7

W6 1 3,4,7,9,0.25,0.14,1:3,2:2,5:3,0.5:2; (2.9,2.6)

W7 1

CR = 0.00

237

Appendices D. Pairwise Comparison Matrices for the Priorities of the Alternative Strategies Based on the SWOT Sub-Factors

Flow alteration

attracting tourist (O5)

Increased

diversion

Current restriction

Increased diversion 1 5,7,0.5,0.2,1:2,2:2,3:2,

9:3,0.33:4; (3.1,3.4)

Current restriction 1

Expiration of the

treaty (O1)

Increased

diversion

Current restriction

Increased

diversion

1 4,6,0.33,1:4,2:2,5:3,

9:3,0.14:2; (3.6,3.2)

Current restriction 1

Potential generation

with third tunnel (O2)

Increased

diversion

Current restriction

Increased diversion 1 4,8,1:3,3:4,5:4,9:2;

(4.3,2.7)

Current restriction 1

Demand mitigation

without nuclear (O3)

Increased

diversion

Current restriction

Increased diversion 1 1,9,3:4,5:5,7:4,

0.5:2; (4.5,2.5)

Current restriction 1

Increased profit

opportunities (O4)

Increased

diversion

Current restriction

Increased diversion 1 2,4,7,8,9,0.33,0.11,1:2,

3:6,5:2; (3.6,2.6)

Current restriction 1

Erosion control

(06)

Increased

diversion

Current restriction

Increased diversion 1 1,2,4,7,0.5,0.25,3:2,5:3,

6:2,9:4; (4.9,3.03)

Current restriction 1

238

Control of misting

(O7)

Increased

diversion

Current restriction

Increased

diversion

1 1,2,8,0.33,3:3,4:2,5:2,

7:2,9:2,0.5:2; (4.2,2.9)

Current restriction 1

Reduced emission

(O8)

Increased

diversion

Current

restriction

Increased diversion 1 3,4,6,7,5:3;

(5,1.29)

Current restriction 1

Employment in energy and

tourism (O9)

Increased

diversion

Current restriction

Increased diversion 1 2,3,4,1:3; (2,1.26)

Current restriction 1

Policy debate revisiting

the treaty (010)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

Payback period

for tunnel (T2)

Increased

diversion

Current restriction

Increased

diversion

1 6,9,0.25,0.17,0.11,1:3,2:2,

3:3,5:2,0.33:2; (2.5,2.5)

Current restriction 1

Reduced power

potential (T1)

Increased

diversion

Current restriction

Increased

diversion

1 2,3:2,4:2; (3.2, 0.84)

Current restriction 1

239

Unfavourable outcome

from renegotiation (T3)

Increased

diversion

Current restriction

Increased diversion 1 5,7,0.25,0.2,0.14,0.11,

1:5,3:2,0.33:4;(1.5,1.9)

Current restriction 1

declining tourists

(T4)

Increased

diversion

Current restriction

Increased diversion 1 2,5,6,9,0.2,0.11,1:6,

3:2,0.33:3;(2.1,3.33)

Current restriction 1

Increased misting

(T6)

Increased

diversion

Current restriction

Increased

diversion

1 1,2,4,9,0.2,0.11,3:5,5:2,

7:2,0.33:2; (3.3,2.6)

Current restriction 1

Cost associated with

renegotiation (T5)

Increased

diversion

Current

restriction

Increased diversion 1 0.25,0.2:2;

(0.22,0.03)

Current restriction 1

Erosion of

escarpment (T7)

Increased

diversion

Current restriction

Increased diversion 1 5,6,0.17,3:3,4:2,7:2,

8:2,9:3,0.33; (3.1,5)

Current restriction 1

Stringent travelling

requirements (T8)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

240

Decreasing appeal

of Niagara (T9)

Increased

diversion

Current restriction

Increased diversion 1 0.33,0.25,0.2,1:6,2:2,

3:2,5:2,7:2; (2.4,2.3)

Current restriction 1

With zero fuel

dependency (S1)

Increased

diversion

Current restriction

Increased

diversion

1 2,3:3,4:3,5:3,6:2,

7:3,9:2; (5.2,2.1)

Current restriction 1

Pumped storage

benefits (S2)

Increased

diversion

Current restriction

Increased

diversion

1 1,2,4,9,0.33,0.2,3:3,

5:2,6:3,7:3; (4.4,2.6)

Current restriction 1

Protection of

equipment (S3)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

Revenue from

tourism (S4)

Increased

diversion

Current restriction

Increased diversion 1 3,7,0.14,1:4,2:2,5:2,

0.33:3,0.25:3; (1.8,2.1)

Current restriction 1

Renewable raw material

lowering pollution (S5)

Increased

diversion

Current restriction

Increased diversion 1 7,1:6,3:2,4:3,5:2,

8:3; (3.8,2.7)

Current restriction 1

241

Employment

opportunities (S7)

Increased

diversion

Current

restriction

Increased diversion 1 2:2,4:2;(3,1.15)

Current restriction 1

Regulating water level

(S6)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

Expected improvement

in health condition (S8)

Increased

diversion

Current restriction

Increased diversion 1 1,2,4,5,3:2;(3,1.41)

Current restriction 1

High unit energy cost

at Beck PGS (W1)

Increased

diversion

Current

restriction

Increased diversion 1 3

Current restriction 1

High investment

cost (W2)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

Turbine

refurbishment

(W3)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

242

Disrupting

environment (W4)

Increased

diversion

Current

restriction

Increased diversion 1 1,0.33,0.25,0.2:4;

(0.3,0.3)

Current restriction 1

Methane emission from

flooded biomass (W5)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

Resettlement of

local (W6)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

Restrict navigation

(W7)

Increased

diversion

Current

restriction

Increased diversion 1 1

Current restriction 1

243

Appendices E. The Bayesian Network Model

Figure E.1: Bayesian network for measuring reliability, resilience and vulnerability for Niagara River

Legend: TMAX: Maximum temperature TMIN: Minimum temperature Temp: Lake surface temperature PRCP: Precipitation WDIR: Wind direction AWND: Average wind speed SNOWF: Snow fall SNWD: Snow depth

244

245

Copyright Acknowledgements (if any)