Ecological consequences of flow regulation by Run-

of-River hydropower on salmonids

by

Pascale Gibeau

M.Sc., Université de Montréal, 2007

B.Sc., Université de Montréal, 2003

Thesis Submitted in Partial Fulfillment of the

Requirements for the Degree of

Doctor of Philosophy

in the

Department of Biological Sciences

Faculty of Science

© Pascale Gibeau 2020

SIMON FRASER UNIVERSITY

Spring 2020

Copyright in this work rests with the author. Please ensure that any reproduction or re-use is done in accordance with the relevant national copyright legislation.

ii

Approval

Name: Pascale Gibeau

Degree: Doctor of Philosophy

Title: Ecological consequences of flow regulation by Run-of-River hydropower on salmonids

Examining Committee: Chair: John Reynolds Professor, Simon Fraser University

Wendy J. Palen Senior Supervisor Associate Professor, Simon Fraser University

Jonathan W. Moore Supervisor Professor, Simon Fraser University

Michael J. Bradford Supervisor Research Scientist, Fisheries and Oceans Canada

Andrew B. Cooper Supervisor Adjunct Professor, Simon Fraser University Principal Data Scientist (Seattle Children’s Hospital)

Josh Korman Internal Examiner Adjunct Professor, University of British Columbia, President at Ecometrics Research

Angela Strecker External Examiner Associate Professor Department of Environmental Sciences Western Washington University, Bellingham, USA

Date Defended/Approved: April 2, 2020

iii

Abstract

Streams are dynamic, disturbance-driven ecosystems, where flow plays a

dominant role structuring biological communities. Anthropogenic activities on streams

change natural patterns of flow and disturbance, which in turn alters the conditions to

which resident fishes are adapted, and their survival and fitness. Run-of-river (RoR)

hydropower projects are an example of an anthropogenic activity that may alter stream

ecosystems by temporarily diverting a proportion of stream flow to produce electricity.

RoR hydropower projects have increased considerably in number and importance in the

last three decades in both British Columbia and worldwide. Although there is a

perception that RoR hydropower has minimal effects on stream ecosystems due to the

small physical footprint of projects, we know surprisingly little about the impacts of RoR

hydropower on fish populations. In this thesis, I use a combination of published

research, empirical data, and models to evaluate a range of hypotheses regarding how

RoR hydropower may affect fish populations, concentrating on salmonid species whose

freshwater habitats often overlap with RoR projects. In Chapter 2, I synthesize the

impact pathways by which RoR hydropower may influence salmonid populations,

inferred from studies of reservoir-storage hydropower and salmonid ecology. In Chapter

3, I use empirical data to quantify increases in water temperature due to RoR flow

diversion and explore the possible consequences for resident fish growth with

bioenergetics models. In Chapter 4, I evaluate how stranding from anthropogenic flow

fluctuations affects the long-term population dynamics of coho salmon (Oncorhynchus

kisutch) in RoR-regulated rivers using a matrix model integrating both the strength and

timing of freshwater density dependence. Finally, in Chapter 5, I quantify the high level

of uncertainty in how much compensation habitat is required to offset chronic mortality

incurred by multiple life-stages of coho salmon. The global emergence of RoR

hydropower projects emphasizes the importance of understanding their effects on

aquatic ecosystems. Overall, our capacity to protect and restore threatened salmonid

populations rests upon our ability to not only better understand the pathways of impacts,

but also the effectiveness of both natural (density dependence) and human (habitat

compensation) interventions that can be used to offset anthropogenic mortality.

iv

Keywords: Run-of-River, Hydropower, anthropogenic disturbances, density

dependence, salmonids, compensation, flow fluctuations, matrix models

v

Dedication

To Maya and baby girl 2,

May this drop of water in the sea of

knowledge help make the world a

better place for you

vi

Acknowledgements

No PhD journey would be possible without the support, encouragement and love

of so many people. First, I would like to thank my supervisor Dr. Wendy Palen, for her

unlimited support and guidance, and for her willingness to welcome me in her lab under

unusual circumstances. I learnt a lot from you and am very grateful for all the knowledge

and wisdom you shared with me over the years. I would also like to thank my committee

members Dr. Mike Bradford, Dr. Jon Moore, and Dr. Andy Cooper. I also learnt a lot

from you and am very grateful for the time you spent helping me navigate the sometimes

murky and confusing waters of PhDs. Mike, I am indebted to you for believing in me and

giving me a chance in REM, and thank you for following me over to E20. I feel privileged

I could keep counting on your expertise and perspectives throughout. Jon, your

thoughtful guidance, insightful thinking, and unwavering support were always so

appreciated, as well as the amazingly quick feedback you always gave me! Andy, thank

you for sharing your knowledge and perspective, and your continuing involvement in my

journey despite your professional changes. I would also like to thank Dr. Murray

Rutherford and Dr. Ken Lertzman for the support they gave me over my first year in

REM. Finally, thank you to Dr. Josh Korman and Dr. Angela Strecker for acting as

internal and external examiners for my defense; I am looking forward to your input and

thoughts on my work. Thank you to Ecofish Research Ltd. and Adam Lewis for sharing

knowledge and expertise of the RoR hydropower industry in BC, and helping me access

and understand the data. Thank you also for supporting me financially through an

Industrial NSERC over the first three years of my PhD. Thank you to Matt Kennedy and

Innergex Renewables Inc. for giving me access to the empirical data that are central to

my Chapter 3. I am grateful for the trust you placed in me and your openness to share

with us, and the scientific world at large, the data you collected.

It sure takes a village to raise a PhD student (especially when that student also

becomes a mom along the way!). I am so grateful for the friendship, help, and support of

my past and present lab-mates (in no particular order, Robin Munshaw, Viorel Popescu,

Janie Dubman, Amanda Kissel, Rylee Murray, Dan Greenberg, Griffin Dare) and of the

amazing E20 community, old and new (I can’t name you all, but know you were all a big

part of the adventure). I benefitted greatly from the stimulation and knowledge of you all

and have great memories of time spent together at Stats-beerz, IDEAS, conferences

vii

and symposiums, or having beer to ponder the state of the world and our research.

Thank you to Kyle Wilson and Emma Hodgson who generously helped me over the past

year with the framing, modelling, and editing of my last two chapters; your help and

support were invaluable, and very much appreciated.

Finally, none of this would have been possible without the love and support of my

“village” both here and around the continent. I am very grateful for the love and support

(and almost daily, at least virtual, presence over the years!) of my friends and PhD

colleagues Ariane Cantin, Kara Pitman, and Michelle Jones – I could not have done this

without you. It was a privilege and an honor to have you by my side through the ups and

downs, both personal and professional, of the journey. Thank you to my amazing mama

group in Squamish who shared both laughs, tears, and high fives through our journeys

into motherhood. Thank you particularly to Tamsin, Katie, Angela, and Jaclyn for the

countless hours of childcare that let me concentrate on my work knowing that my little

girl was safe and happy. Thank you to my mother-in-law Kerry for her unwavering

support and love, and all the love and care she nurtures our girl with. Thank you for the

unlimited and unconditional love and support to my old group of girlfriends (Ariane,

Geneviève B., Mariane, Kim, Marie-Pierre, Geneviève P.) who are always there for me

even though we now live apart. And finally, of course I owe so much to my family, it is

hard to find the words. Thank you to my mom and dad, Henriette et François, for giving

me such a strong and rich environment to grow in, for taking me camping, in the forests

and raising me by a lake, for teaching me how to be a good person, to always be curious

about the world and to work hard. Thank you to my sister Élisabeth for being my best

friend from my youngest age, I can’t imagine life without you even though I moved so far

away (I know I know); and thanks to you and Stéphane for giving me the best niece and

nephew. Lucas et Raf, je vous aime de tout mon coeur, merci de votre amour, de votre

rire, de votre vivacité et curiosité qui nous donnent envie de nous battre pour vous offrir

un monde meilleur. Finally, thank you deeply to my husband Deon for all your love and

support through the years, thank you for always being there for me and making me

happy every day despite life’s challenges. You gave me the best little girl in the world

(and soon another one!); Maya I am so grateful for you and love you more than I can

possibly express. You help me concentrate everyday on what truly matters in this world,

and I believe that also made me a better scientist and write a better thesis.

viii

Table of Contents

Approval .......................................................................................................................... ii

Abstract .......................................................................................................................... iii

Dedication ....................................................................................................................... v

Acknowledgements ........................................................................................................ vi

Table of Contents .......................................................................................................... viii

List of Tables ................................................................................................................... x

List of Figures................................................................................................................. xi

General Introduction ................................................................................ 1

Run-of-River hydropower and salmonids: potential effects and perspective on future research .......................................................................... 5

2.1. Abstract ................................................................................................................. 5

2.2. Introduction ............................................................................................................ 5

2.3. Methods: Literature synthesis of RoR hydropower ................................................. 7

2.4. Characteristics of flow diversion for RoR hydropower and interactions with salmonids ........................................................................................................................ 8

2.5. Pathway 1: Reduction of flow in the bypassed reach ........................................... 11

2.6. Pathway 2: The presence of low-head dams ....................................................... 16

2.7. Pathway 3: Anthropogenic Flow Fluctuations ...................................................... 19

2.8. Avenues for Future Research .............................................................................. 21

Effects of flow diversion by Run-of-River hydropower on stream temperature and bioenergetics of salmonid fishes......................................... 34

3.1. Abstract ............................................................................................................... 34

3.2. Introduction .......................................................................................................... 35

3.3. Methods .............................................................................................................. 37

3.3.1. Study streams.............................................................................................. 37

3.3.2. Empirical Data ............................................................................................. 38

3.3.3. Temperature modeling ................................................................................. 38

3.3.4. Bioenergetics modeling ............................................................................... 39

3.4. Results ................................................................................................................ 42

3.4.1. Temperature changes in bypassed reaches ................................................ 42

3.4.2. Changes in trout consumption and growth in bypassed reaches ................. 43

3.5. Discussion ........................................................................................................... 44

3.6. Conclusion........................................................................................................... 47

Freshwater density dependence and potential consequences of flow fluctuations from Run-of-River hydropower on coho salmon (Oncorhynchus kisutch) .............................................................................................................. 58

4.1. Abstract ............................................................................................................... 58

4.2. Introduction .......................................................................................................... 59

ix

4.3. Methods .............................................................................................................. 61

4.3.1. Matrix model ................................................................................................ 61

4.3.2. Freshwater density-dependent survival ....................................................... 62

4.3.3. Scenarios of RoR-induced flow fluctuations ................................................. 64

4.3.4. Elasticity analyses ....................................................................................... 66

4.4. Results ................................................................................................................ 66

4.4.1. Probability of quasi-extinction ...................................................................... 67

4.4.2. Population size ............................................................................................ 68

4.4.3. Elasticity analyses ....................................................................................... 69

4.5. Discussion ........................................................................................................... 69

4.5.1. Implications for conservation ....................................................................... 71

4.6. Conclusion........................................................................................................... 72

How much is enough: the limitations of habitat compensation for offsetting chronic anthropogenic mortality of coho salmon (Oncorhynchus kisutch) .............................................................................................................. 83

5.1. Abstract ............................................................................................................... 83

5.2. Introduction .......................................................................................................... 84

5.3. Methods .............................................................................................................. 87

5.3.1. Scenarios of compensation.......................................................................... 87

5.3.2. Deterministic matrix model........................................................................... 89

5.4. Results ................................................................................................................ 91

5.4.1. Effects of chronic mortality without compensation ........................................ 91

5.4.2. Effects of habitat compensation ................................................................... 91

5.4.3. Productivity of compensation habitats .......................................................... 92

5.5. Discussion ........................................................................................................... 93

5.6. Conclusion........................................................................................................... 96

General conclusions ............................................................................ 103

References ................................................................................................................. 109

Appendix A ................................................................................................................ 129

Appendix B ................................................................................................................ 130

Appendix C ................................................................................................................ 132

Appendix D ................................................................................................................ 133

Appendix E ................................................................................................................ 135

Appendix F ................................................................................................................ 139

Appendix G ................................................................................................................ 140

x

List of Tables

Table 2-1. Empirical peer-reviewed papers studying RoR hydropower impacts on river ecosystems and salmonids. a= brown trout, b= European grayling, c= trout, d= rainbow trout, e= Atlantic salmon, f= high-head scheme, g= low-head scheme; **= dams were in cascade; MAD= mean annual discharge. ............................................................................................................... 26

Table 2-2. Relative vulnerability of salmonid species to potential impacts from RoR hydropower, based on chronic exposure and the potential for dams to create barrier for migrations. Information on life history, spawning season and length of freshwater residency come from the references indicated by footnotes. ............................................................................................... 29

Table 3-1. Mean proportional change in consumption (C) and mean difference in specific growth rate (SGR) between diverted (observed) and natural baseline (predicted) water temperatures (all years combined) for Age 1+ and 2+ trout experiencing the four food availability scenarios in Douglas and Fire creeks. (Standard deviations are in bracket) ............................ 49

Table 3-2. Average final fish weights (g) of Age 1+ and 2+ trout sampled in the bypassed and upstream reaches of Fire and Douglas creeks in the pre- (2006-2008) and post- (2010-2014) diversion periods. ........................... 50

Table 4-1. Matrix model structure and description of vital rates for spring and winter density-dependent mortality bottlenecks. ............................................... 74

Table 4-2. The 12 scenarios used in simulations with varying frequency and magnitude of RoR-induced flow fluctuations. White indicate low impact scenarios (annual survival >75%), grey indicate medium impact scenarios (annual survival between 45-60%), and dark grey indicate high impact scenarios (annual survival <36%). .......................................................... 75

Table 5-1. Deterministic matrix model and scenarios per life stage. ........................ 98

xi

List of Figures

Figure 2.1. Three main pathways of effects of RoR hydropower on salmonids and their associated main mechanisms of impacts. The number of empirical peer-reviewed papers identified in this review is shown in boxes between the pathways of effect (1-3) and their primary mechanisms. ................... 31

Figure 2.2. Generalized schematic of a run-of-river hydropower facility. A low-elevation dam impounds the river and reroutes water downstream through a penstock, to turn turbines in the powerhouse and produce electricity. The diversion of water leaves a reach of the river with reduced flows (bypassed reach), but all water is returned to the downstream reach of the river via the tailrace. The headpond has little water storage capacity and so does not act as a reservoir. Powerlines are part of the transmission corridor that connects the facility to a centralized power grid. ............................................................................................................... 32

Figure 2.3. Modeled natural upstream flows (black line) and flows passing through the bypassed reaches (dashed line) during an average runoff year below a low-head RoR hydropower dam in a) western Canada (high-head scheme in snowmelt-dominated Bull River, natural flow averaged from 2010-2013, diverted flow modelled with a hypothetical turbine flow of 9.9 cms; www.bchydro.com), and b) Yunnan Province, China (rain-dominated Gutan River, Lushui County, adapted from Kibler and Tullos 2013). Flow alterations in the bypassed reaches are most pronounced during naturally low to moderate flows, when RoR hydropower operations divert the greatest proportion of flow from the channel. The Bull River has a minimum flow requirement that varies between 0.25 and 2.0 cms, depending on the season. ...................................................................... 33

Figure 3.1. Location of the two high-gradient mountain streams regulated for RoR hydropower in British Columbia, Canada. The two bypassed reaches are highlighted in darker grey between the low-head dams and the powerhouses. ......................................................................................... 51

Figure 3.2. Conceptual diagram showing the methods used to estimate the change in water temperatures due to flow diversion by RoR hydropower alone. (a) The mean change in temperatures between the upstream reach and the future bypassed reach due to the longitudinal gradient in the stream (i.e. natural rate of warming) was estimated from temperature data before construction of the RoR facilities (2007-2008). (b) The mean natural rate of warming was added to the upstream temperatures for each year following flow diversion to predict the estimated temperatures in the bypassed reach would no flow diversion occur. ...................................... 52

Figure 3.3. a) Mean daily temperature (⁰C) in upstream and bypassed reaches, and b) mean daily flow (m3s-1) diverted for hydroelectricity production (plant flow, black) and recorded in bypassed reach, from January to December in Fire Creek (2010). Corresponding proportion of diverted flow is shown on the second y-axis in b. See Appendices B and D for other temperature and flow time series. .............................................................................. 53

xii

Figure 3.4. Difference in mean monthly temperature (⁰C) between the observed temperatures in the bypassed reaches and the natural (predicted) temperatures in a) Fire Creek and b) Douglas Creek from 2010 to 2014. Positive differences indicate that temperatures observed in the bypassed reaches following flow diversion were warmer than predicted based on the natural downstream rate of warming alone. The black line shows no difference between natural predicted and observed water temperatures. ............................................................................................................... 54

Figure 3.5. Increase in simulated consumption (C, % of baseline) of trout in the warmer temperatures of the bypassed reaches as compared to the predicted natural baseline temperatures for the Unlimited Consumption (UC) scenario (left panels), and the Increased Consumption (IC) scenario (right panels) in Douglas and Fire creeks during 2010 (dark grey), 2011 (black), and 2013 (light grey) for a) Age 1 fish, and b) Age 2 fish. Years are ordered from colder (2011) to warmer (2013) temperatures. Note the y-axis in a) displays much bigger values than in b). Boxes represent 25 to 75% of the ranked data and the horizontal line shows the median. Whiskers range from the extremities of the box to 1.5 interquantile range of the data. Outliers are shown by open circles. ..................................... 55

Figure 3.6. Change in specific growth rate (SGR, %) achieved by trout between the warmer temperatures of the bypassed reaches and the predicted natural baseline temperatures for the Unlimited Consumption (UC) scenario (left panels), and the Constant Consumption (CC), Increased Consumption (IC), and Reduced Consumption (RC) scenarios (right panels) in Douglas and Fire creeks during 2010 (dark grey), 2011 (black), and 2013 (light grey) for a) Age 1 fish, and b) Age 2 fish. Years are ordered from colder (2011) to warmer (2013) overall temperatures, based on upstream temperatures. ......................................................................................... 56

Figure 3.7. Frequency distributions of Age 1+ O.mykiss sampled in the bypassed reach of Douglas Creek from 2010 to 2014, a) weight (n= 105 fish, mean: 7.6g), and b) estimated specific growth rate (SGR, histogram) with predicted SGR of Age 1+ O. mykiss under reduced C, constant C, and increased C scenarios (dashed lines, right axis). SGR predicted by the three scenarios of food availability in the temperature recorded at the end of the bypassed reaches is lower than the empirical SGR observed at the beginning of the bypassed reaches. ....................................................... 57

Figure 4.1 Number of smolts produced under weak (α= 0.2), moderate (α= 0.5), and strong (α= 1) density dependence as derived with Beverton-Holt equations using data from 10 creeks (in light grey) in the US and Canadian Pacific Northwest. Carrying capacity (k) was fixed at 15,318 smolts over nine km of river.................................................................... 76

Figure 4.2. Mean frequency of flow fluctuations events per year (± SD for the creeks with multiple years of data, shown with error bars, top) and mean proportion of flow fluctuation events per season (±SD shown by error bars, bottom) induced by operations of RoR hydropower for eight regulated creeks in British Columbia, Canada. Creek names in the top panel are ordered by increasing watershed area (47-313 km2), and characteristics of the creeks and RoR dams are listed in Appendix G. ... 77

xiii

Figure 4.3. Probability of quasi-extinction after 45 years (z-axis) for populations simulated under 12 combinations of RoR-induced flow fluctuations of differing frequency (y-axis) and magnitude (x-axis). Rows show scenarios experiencing density dependence during Spring (a) or Winter (b), and columns indicate the strength of density dependence, ranging from Weak (left), Moderate (middle), and Strong (right). Probability of quasi-extinction for the baseline populations with weak, moderate or strong density dependence were 0.19, 0.13, and 0.11, respectively. ............................. 78

Figure 4.4. Probability of quasi-extinction (%) over time for a typical coho population subjected to a) medium impact scenario M-M, and b) high impact scenario H-H of RoR-induced flow fluctuations, under the three strengths (weak, moderate, and strong) and two timings (Spring, Winter) of density dependence. .......................................................................................... 79

Figure 4.5. Mean final population size (mean number of adults for the last six cohorts, years 40-45) for populations simulated under the 12 scenarios of RoR-induced flow fluctuations of differing frequency (y-axis) and magnitude (x-axis). Rows show scenarios experiencing density dependence during Spring (a) or Winter (b), and columns indicate the strength of density dependence, ranging from Weak (left), Moderate (middle), and Strong (right). Average final size for the baseline populations with weak, moderate or strong compensatory reserves were 359, 531, and 591 adults, respectively (thick black lines). Note that order of the x and y axis is different from that of Fig. 3.3. .............................................................. 80

Figure 4.6. Final population size (mean number ±SD of adults for years 40-45) for populations experiencing spring (black circles) and winter (white circles) bottlenecks simulated under weak, moderate, and high density dependence, for a) low impact L-L, b) medium impact M-M, and c) high impact H-H scenarios of RoR-induced flow fluctuations. Average final sizes for the baseline populations with weak, moderate or strong density dependence are represented with * (n=359, 531, and 591 adults, respectively). .......................................................................................... 81

Figure 4.7. Simulation-based elasticity values for lower level vital rates (R2adj= 0.982, p < 2.2 e-16). Fry survival in year 2 (φfryY2) and adult survival in the ocean

on year 3 (φoceY3) are not presented because they are fully collinear with

fry survival in year 1 (φfryY1) and smolt survival in the ocean on year 2

(φoceY2), respectively. .............................................................................. 82

Figure 5.1. Distribution of (a) the size (m2) of 27 side-channel habitats built in the PNW (from Morley et al. 2005; Roni et al. 2006; Rosenfeld et al. 2008), and (b) the productivity of 33 side-channel habitats (# of smolts per m2) built in the PNW (from Rosenfeld et al. 2008; Roni et al. 2010). ............. 99

Figure 5.2. Proportion of the final size of the baseline population represented by populations subjected to chronic anthropogenic mortality (2 to 20%) affecting a) Eggs, and b) Parr and Smolts/Adults as a function of the size of side-channel compensation habitat (m2). A relative population size of 1 implies that the population modelled had the same final population size as the baseline population (n= 808 adults). The grey box represents the 10th to the 90th percentile, and the vertical line indicates the average size of compensation habitat built in the PNW (based on 27 sites, from Morley et

xiv

al. 2005; Roni et al. 2006; Rosenfeld et al. 2008). Note that the Y-axis is different for the two panels. .................................................................. 100

Figure 5.3. Change in final number of adults for impacted populations compared to baseline abundances (%, colors) over a range of sizes of constructed side-channels (y-axis), annual chronic anthropogenic mortality (x-axis), and life stage (eggs, parr, smolts/adults) affected by the chronic mortality (panels), assuming mean productivity in constructed side-channels. 0 (white cells) indicates no change in final abundances of impacted populations compared to baseline, while positive values (cool shades) mean compensation increased the final number of adults in impacted populations and negative values (warm shades) mean final size of impacted populations decreased despite the amount of compensation habitat added. The black lines indicate offsetting equivalency for the combination of compensation habitat sizes and chronic mortality. ........ 101

Figure 5.4. Range in sizes of compensation habitat (m2) required to effectively offset annual chronic mortality ranging from 2 to 20% affecting, a) Eggs, b) Parr, and c) Smolts/Adults, assuming varying productivity (# of smolts produced per m2) for the constructed side-channels. Horizontal black lines indicate the 25th to 75th percentiles of productivity of compensation habitat, black triangles represent mean number of smolts contributed, and stars represent median number of smolts contributed by compensation habitat (based on data from 33 sites, from Rosenfeld et al. 2008; Roni et al. 2010). The grey box represents the mean (vertical line) and 10th to 90th percentiles (shaded) of sizes of compensation habitats built in the PNW (n= 27 sites, data from Morley et al. 2005; from Roni et al. 2006; Rosenfeld et al. 2008). ......................................................................... 102

1

General Introduction

Freshwater organisms are among the most endangered groups of organisms

worldwide even though freshwater ecosystems represent less than 1% of Earth’s

surface (Williams and Miller 1990; Dudgeon 2010; Tickner et al. 2020). Freshwater

ecosystems are in faster decline than terrestrial ecosystems, with up to a third of

freshwater species extinct or imperiled in North America alone (Jelks et al. 2008).

Several threats challenge the survival and resilience of freshwater species, among which

irrigation, overharvest, invasion by exotic species, pollution, climate change, and

dams/water impoundment (Richter et al. 1997; Strayer and Dudgeon 2010; Tickner et al.

2020). Growing human societies in the past decades increased both the pressures on

stream ecosystems and the demands for energy to fuel development and booming

economies. For example, over 800,000 dams have been built worldwide since the

beginning of the 20th century, collectively influencing more than half of global runoff and

intercepting 25% of naturally transported sediment (Vörösmarty and Sahagian 2000;

Jackson et al. 2001; Pittock and Hartmann 2011). The global demand for energy is

expected to further increase 14-33% by 2035 (OECD 2013). In response, calls to

develop renewable electricity (e.g. wind, solar, biomass, small-scale hydropower) are

increasing rapidly because they are deemed more environmentally friendly than

traditional, large-scale industries like large dams and reservoirs or coal plants (Evans et

al. 2009).

Small-scale diversion run-of-river (RoR) hydropower projects are a small-scale

source of renewable energy increasingly relied upon in global energy policies, including

in British Columbia (BC, Canada). For example, BC experienced a shift away from

primarily relying on large dam-reservoirs for electricity production, to a system that

included the construction of many small production facilities operated by private

corporations (Jaccard et al. 2011). Worldwide, the contribution of small hydropower to

global power generation increased from 32,000 to 45,000 megawatts (MW) from 2001 to

2010 (Abbasi and Abbasi 2011). An average of 11 small dams are now built for every

large hydropower dam, contributing 11% of the global energy capacity of hydropower

2

(Couto and Olden 2018). Many regions developed new RoR hydropower facilities,

particularly in the southern hemisphere, e.g. in Indian Himalaya (Grumbine and Pandit

2013), China (Wang et al. 2010, Couto and Olden 2018), Thailand (Aroonrat and

Wongwises 2015), Africa (Chiyembekezo 2013), and Central and South America

(Anderson et al. 2006, Couto and Olden 2018), but also in Europe and the United States

(Couto and Olden 2018). Consequently, the effects of small hydropower dams on stream

ecosystems was listed as one of the 15 emerging issues of relevance to global

conservation in 2020 (Sutherland et al. 2020).

The flow of water across inland landscapes plays a dominant role regulating the

biological communities in the dynamic, disturbance-driven stream ecosystems. The

natural flow regime of rivers, characterised by the magnitude, frequency, duration,

timing, and rate of change of flow events, affects aquatic organisms over short- to long-

term timescales (Poff et al. 1997). RoR hydropower alters the natural flow regime of

regulated rivers by temporarily diverting, using a low-head dam, a proportion of stream

flow from the main river channel through a long tunnel leading to a powerhouse where

electricity is generated. The water is then returned to the river channel. Water storage is

limited to a headpond upstream of the low-head diversion dam, but the diversion of

water for RoR hydropower creates a bypassed river reach with reduced flow. RoR

hydropower typically refers to infrastructure with energy output up to 25 MW but more

generally below 10 MW, though the definition varies across political jurisdictions (Abbasi

and Abbasi 2011, Couto and Olden 2018). Although there is a perception that RoR

hydropower has minimal impacts on natural ecosystems due to the small physical

footprint of projects, we know surprisingly little about their impacts on stream

ecosystems (Abbasi and Abbasi 2011; Couto and Olden 2018). Effects of RoR

hydropower may be particularly important for anadromous and resident salmonid fishes,

whose habitats in high-gradient, mountainous streams often overlap with RoR

hydropower infrastructure. Salmonids have unique ecological, cultural, and economic

roles (Gende et al. 2012, Naiman et al. 2002), and are distributed in coastal and inland

rivers around the world. Native populations face many threats and have declined

dramatically in abundance over the last decades in their native ranges (Slaney et al.

1996; Gustafson et al. 2007; Crozier et al. 2019). The main causes of decline are low

marine survival, overharvesting, climate change, competition from hatchery

supplementation, and the degradation of freshwater habitat they rely on (Kareiva et al.

3

2000; Chittenden et al. 2010). The increasing importance of RoR hydropower in rivers

worldwide is likely to add pressures on salmonids by further altering their freshwater

habitats.

In this thesis, I combine published research, empirical data, and models to

evaluate a range of hypotheses exploring how RoR hydropower potentially affect

salmonids. In Chapter 2, I explore the impact pathways by which RoR hydropower may

influence salmonid populations by reviewing and synthesizing studies on larger

reservoir-storage hydropower and general salmonid ecology. I focus on three pathways

of effect by which RoR hydropower may influence salmonids and suggest avenues of

new research to fill in the main knowledge gaps identified. In Chapter 3, I explore how

the reduction in stream flow in bypassed reaches due to flow diversion by RoR

hydropower may influence temperature regimes and fish growth using empirical and

simulated data from two RoR regulated streams in BC. I quantify changes in water

temperatures and assess the potential impacts to growth of resident rainbow trout

(Oncorhynchus mykiss) using bioenergetics models under a range of food consumption

scenarios. In Chapter 4, I evaluate how fry stranding due to anthropogenic flow

fluctuations downstream of RoR infrastructures affects coho salmon (Oncorhynchus

kisutch) population dynamics. Specifically, I use a stochastic stage-structured matrix

model to simulate how the timing and magnitude of anthropogenic flow fluctuations

influenced population abundance and extinction risk of coho salmon. I also integrate how

the timing and strength of natural density-dependent survival bottlenecks experienced by

salmon in freshwaters moderate or amplify the potential for RoR-induced mortality to

scale to emergent population dynamics. Finally, in Chapter 5, I expand the consideration

of impacts to salmonid fishes in freshwater ecosystems beyond RoR hydropower alone,

and quantify the high level of uncertainty in how much compensation habitat is required

to offset a wide range of anthropogenic sources of mortality imposed on three life-stages

of coho salmon. To do so, I run simulations to determine the amount and productivity of

habitat compensation required to offset chronic anthropogenic mortality and avoid losses

of population productivity from anthropogenic sources like forestry practices, entrainment

in small and large dams, harvesting, ocean mortality etc. I also assess which life stages

have the best potential to react to the building of compensation habitat in freshwater and

affect emergent population dynamics.

4

The global emergence of RoR hydropower projects emphasizes the importance

of understanding their effects on aquatic ecosystems. Overall, to protect and restore

threatened salmonid populations we need to not only better understand the pathways of

impacts, but also the effectiveness of both natural (density dependence) and human

(habitat compensation) interventions that can offset anthropogenic mortality. As such,

protecting and restoring threatened salmonid populations require that we increase our

capacity to evaluate the efficacy of policies and practices of industry and regulators so

that we can both protect biodiversity and stream ecosystems, and support our growing

needs for energy production.

Contributions

My thesis was a collaborative work, and each chapter is either a published

manuscript (Chapter 2), in last rounds of revisions (Chapter 3), or prepared for

submission (Chapters 4 and 5). My thesis chapters are thus written using the first-person

plural; however, I assembled data from industry or literature sources, designed and

performed analyses (including writing R code), interpreted results, and wrote and revised

drafts. My work greatly benefited from feedback from co-authors, committee members,

and colleagues that are acknowledged in my thesis and in the acknowledgement section

of each chapter. The general introduction and conclusions represent my own views and

perspectives.

5

Run-of-River hydropower and salmonids: potential effects and perspective on future research1

2.1. Abstract

The spatial footprint of individual run-of-river (RoR) hydropower facilities is smaller than

reservoir-storage hydroelectric projects and their impacts to aquatic ecosystems are often

assumed to be negligible. However, these effects are poorly understood, especially for

salmonids whose freshwater habitat often overlaps with RoR hydropower potential. Flow

regulation for RoR hydropower is unique in how it influences the seasonality and magnitude of

flow diversion, and because low-head dams can be overtopped at high flows. Based on a

review of the primary literature, we identified three pathways of effects by which RoR

hydropower may influence salmonids: reduction of flow, presence of low-head dams

impounding rivers, and anthropogenic flow fluctuations. We synthesized empirical evidence of

effects of RoR hydropower on river ecosystems from 31 papers, of which only ten explicitly

considered salmonids. We identified key research gaps including impacts of extended low flow

periods, anthropogenic flow fluctuations, and cumulative effects of multiple RoR projects. Filling

these gaps is necessary to help manage and conserve salmonid populations in the face of the

growing global demand for small-scale hydropower.

2.2. Introduction

Rivers are dynamic, disturbance-driven ecosystems, where flow plays a

fundamental role in structuring aquatic and riparian communities (Resh et al. 1988; Poff et al.

1997a; Murchie et al. 2008). The natural flow regime (NFR) of rivers is defined by the

magnitude, frequency, duration, timing, and rate of change of flow events, each of which affect

stream-dwelling aquatic organisms over short-term to evolutionary timescales (Poff et al. 1997).

Many anthropogenic activities can alter the flow and disturbance regimes of streams, which in

turn may affect the survival and fitness of native species (Poff and Ward 1990; Reice et al.

1 A version of this chapter was published in the Canadian Journal of Fisheries and Aquatic Sciences in 2017

6

1990; Strayer and Dudgeon 2010). Impoundment of water by dams, either built for irrigation,

flood control, or hydroelectricity generation, is one of the greatest anthropogenic drivers of

change to NFRs. Over 800,000 dams have been built worldwide since the beginning of the 20th

century, collectively influencing more than half of global runoff and intercepting 25% of naturally

transported sediment (Vörösmarty and Sahagian 2000; Jackson et al. 2001; Pittock and

Hartmann 2011). Storage dams retain water for extended periods in reservoirs, and

subsequently release it at times that can be out of phase and frequency with NFRs (Rosenberg

et al. 1997; Murchie et al. 2008). Such deviations from NFRs cause flow alterations that have

well-documented consequences for river geomorphology, continental runoff, riparian

communities, and macroinvertebrate and fish populations (Nilsson et al. 2005; Murchie et al.

2008; Poff and Zimmerman 2010).

In recent decades, small-scale hydropower production, consisting primarily of small-

scale Run-of-River (RoR) hydropower, has emerged as an alternative to the construction of new

reservoir-storage or large dams because of their perceived lower economic, social, and

environmental costs (Postel et al. 1996; Abbasi and Abbasi 2011; Anderson et al. 2014). For

example, the contribution of small hydropower to global power generation nearly doubled from

2001 to 2010, increasing from 32,000 to 45,000 megawatts (MW) (Abbasi and Abbasi 2011), as

many regions have developed new small-scale RoR hydropower facilities (e.g., Canada: Cyr et

al. (2011); Sopinka et al. (2013); Indian Himalaya: Grumbine and Pandit (2013); China: Wang et

al. (2010); Thailand: Aroonrat and Wongwises (2015); Africa: Chiyembekezo (2013); Central

America: Anderson et al. (2006); Europe: Anderson et al. (2014); Spänhoff (2014). Although no

universally accepted definition exists, small-scale RoR hydropower (hereafter referred to as

RoR hydropower) typically refers to low-head diversion dams with energy output up to 25 MW

(Abbasi and Abbasi 2011). However, some countries like China and Canada consider facilities

with installed capacities less than 50 MW as small hydropower (Cyr et al. 2011; Kibler 2011).

Despite the increased development of RoR hydropower, there is a paucity of peer-reviewed

research into the effects of RoR hydropower on aquatic ecosystems in which they occur,

leading to knowledge gaps about the effects to aquatic species (but see (Abbasi and Abbasi

(2011) and Anderson et al. (2014) for recent reviews). Here we provide the most comprehensive

global synthesis of observed and potential effects from RoR hydropower on salmonid fishes.

We focus our review on salmonid fishes because of their unique ecological, cultural, and

economic importance (Naiman et al. 2002) as well as their near-global distribution in coastal

and inland rivers. Based on the unique characteristics of flow diversion created by RoR

7

hydropower, we synthesized empirical peer-reviewed literature and expanded upon information

from previous reviews to hypothesize three main pathways of effect (e.g. altered ecological

mechanisms, linking causes to effects) by which RoR hydropower operations could impact

salmonids: the reduction of flow in river reaches, the presence of low-head dams impounding

rivers, and the creation of anthropogenic flow fluctuations. We evaluated the evidence in

support of these pathways based on a comprehensive search of the peer-reviewed literature on

RoR hydropower (n= 31 empirical studies), as well as relevant literature from other forms of flow

regulation similar to RoR hydropower. Importantly, our synthesis excludes the effects of ROR

hydropower on riparian ecosystems, as such effects are expected to be very similar to those

produced by other industrial activities (e.g. terrestrial footprint of facilities, construction of roads

and powerlines, etc.) and reviewed elsewhere (e.g., Smith et al. (1991); Weltman-Fahs and

Taylor (2013); Abbasi and Abbasi (2011). The specific goals of our review are to: 1) categorize

the unique characteristics of flow diversion by RoR hydropower operation, 2) evaluate the

support from the peer-reviewed literature for three main pathways of effect on salmonid fishes,

and 3) identify critical knowledge gaps that can be used as priorities to guide future research.

We organize the findings of existing peer-reviewed studies into a framework under the NFR

paradigm, and maintain a broad geographic scope to maximize the extent of the synthesis. We

hope these unique characteristics of our review make its findings and conclusions applicable to

as many contexts as possible, and help focus new research on filling the highest priority

knowledge gaps.

2.3. Methods: Literature synthesis of RoR hydropower

To identify peer-review literature examining the effects of RoR hydropower on

salmonid fishes, we searched the Web of Science and Aquatic Sciences Fisheries Abstracts

databases through April 2016 for combinations of the keywords: "run of river", "small hydro",

"small hydropower", "water diversion" crossed with the keywords "salmonid", "trout", and "fish".

We also used the literature cited by each paper, as well as high-quality grey literature sources

(e.g. Robson et al. 2011), to identify additional peer-reviewed literature related to the effects of

RoR hydropower on salmonid fishes. We screened the peer-reviewed papers to include only

those focused on RoR hydropower technology (i.e., low-head dams producing hydroelectricity),

but information from low-head dams build for purposes other than hydropower (e.g. irrigation,

municipal uses) was used to substantiate pathways where possible. We identified 47 peer-

reviewed studies specific to the effects of RoR hydropower, of which 31 empirical studies

8

covered impacts of RoR hydropower on fishes, invertebrates, or river habitat (Table 2-1, Figure

2.1). Of these 31 empirical papers, 17 examined the effects of RoR hydropower on fish, but

only 10 specifically addressed salmonid fishes (Table 2-1). A total of 14 other papers were

related to economics, energy policy, geopolitics, and the geography of RoR hydropower (n= 9),

or covered related topics such as comparison between small and large dams or cumulative

impacts (n= 3), engineering (n= 1), and dam classification or removal (n= 1) (Appendix A). In

addition to these empirical papers, we identified two recent reviews, one focused on the history

of RoR hydropower technology that challenges the perception that ecological impacts are

minimal (Abbasi and Abbasi 2011), and a second that included a brief overview of two pathways

of impact by which RoR hydropower may affect the physical and ecological conditions of rivers,

but did not address impacts to fish or salmonids (Anderson et al. 2014). The papers included

here encompass a near global geography in order to compensate for the general paucity of

peer-reviewed studies on this topic, and to help inform our understanding of the potential effects

of RoR hydropower on river ecosystems using all available information.

2.4. Characteristics of flow diversion for RoR hydropower and interactions with salmonids

RoR hydropower projects can take on a variety of designs, however, the majority occur

in high-gradient, mountainous rivers where kinetic energy, produced from the drop of water over

a sharp elevation gradient, creates a hydraulic head for electricity production (Abbasi and

Abbasi 2011; Anderson et al. 2014). High-gradient RoR hydropower projects, called high-head

schemes (Anderson et al. 2014), are characterized by a low-elevation dam (henceforth called

low-head dam) that creates an upstream impoundment (headpond) with little storage capacity

(Figure 2.2). We found that 77% of the empirical papers we reviewed focused on such

hydropower schemes, while 17% provided no clear description of the RoR designs (see Table

2-1). Low-head dams create relatively small headponds with limited water storage capacity,

creating the expectation that flow regimes in rivers regulated by RoR hydropower will more

closely mimic NFR than reservoir-storage systems (Poff and Hart 2002; Shaw 2004; Kibler and

Tullos 2013). Low-head dams divert existing flow through intake structures into a pipe

(penstock) running parallel to the river for several kilometers until reaching a powerhouse where

turbines are rotated to generate electricity. The diversion of water from the dam to the

powerhouse results in a reach immediately downstream of the dam with lower than natural

flows, termed the bypassed reach (also referred to as the diversion reach in Canada, dewatered

9

reach in the USA, or depleted reach in the United Kingdom). Diverted water is then returned

back to the river channel after passing through the powerhouse. RoR hydropower operations

thus result in unique forms of flow regulation when compared to reservoir-storage hydropower

dams or RoR dams built for purposes other than hydropower (e.g., water diverted for irrigation

or municipal uses), which are consequently likely to influence river ecosystems in different

ways.

The high-gradient rivers where most RoR hydropower projects are located are

also often the well-oxygenated, clear, and cold river habitat favoured by many anadromous and

resident salmonids (subfamilies Salmoninae) (Shaw 2004; Wohl 2006). Salmonid fishes exhibit

a wide range of life history strategies, some of which are more likely than others to be sensitive

to the effects of RoR hydropower (Table 2-2). Overall, low-head dams have the potential to

impede the completion of salmonid life cycles through two primary mechanisms: blocking

movement and migration among important habitats, or by chronic exposure to reduced or more

variable flows within the footprint of RoR hydropower. For example, anadromous salmon

species with life histories requiring long-distance freshwater migrations through high order

streams, like steelhead (Onchorhynchus mykiss) and Atlantic (Salmo salar) salmon, may be

especially vulnerable to encountering barriers from dams along their long migrating journey.

Salmonids with resident life histories like rainbow (O. mykiss) and brook trout (Salvelinus

fontinalis) that spend the entirety of their lives in freshwater and do not undergo extensive

freshwater migrations, are more likely to be vulnerable to the effects of chronic exposure to RoR

hydropower. Salmonids like adfluvial and fluvial Bull Trout (S. confluents) that undergo both

extensive freshwater migrations and spend large portions of their lives in higher order streams

are likely to be among the most vulnerable salmonids to barriers to migration and chronic

exposure to RoR hydropower. In contrast, the short residence time in freshwater and shorter

freshwater migrations of pink (O. gorbuscha) and chum (O. keta) salmon make them potentially

the least vulnerable to impacts from RoR hydropower (Table 2-2). Other species like Chinook

salmon (O. tshawytscha) tend to spawn in larger riverine systems not directly affected by small

RoR low-head dams. Overall, the overlap between RoR hydropower operations and salmonid

habitats only provides a basis for possible negative consequences to salmonid fishes. Our

review of pathways below highlights the many knowledge gaps remaining about the relative

vulnerability of salmonids to flow diversion by RoR hydropower.

Changes to natural hydrographs due to RoR hydropower operations are

expected to vary by reach (upstream, downstream, and bypassed) in rivers influenced by RoR

10

hydropower. The creation of headponds inundates riparian areas and fundamentally interrupts

the NFR in reaches immediately upstream of low-head dams by reducing flow variability,

velocity, and turbulence, and increasing the deposition of fine sediment (Csiki and Rhoads

2010; Butler and Wahl 2011). Such changes in physical habitats can lead to impacts on riverine

ecosystems by reducing water quality and altering the abundance, richness, and composition of

periphyton, invertebrate, and fish assemblages (Santucci et al. 2005; Mueller et al. 2011;

Anderson et al. 2014). In contrast, the flow regime in reaches downstream of RoR powerhouses

is expected to be the most similar to the NFR since water diverted for power generation is

returned to rivers after passing through turbines (Poff and Hart 2002; Kibler and Tullos 2013;

Senay et al. 2016). Nonetheless, the return of water at the tailrace may have a hydraulic impact

on benthic macroinvertebrate composition (Anderson et al. 2015) or fish as a result of dissolved

gas super-saturation (Weitkamp and Katz 1980). RoR hydropower operations are expected to

cause the greatest changes to the NFR in bypassed reaches, where a proportion of flow is

removed for power production. The amount of flow removed can vary widely depending on

national, regional, or local regulations (e.g. up to 100% in some systems in China, Kibler and

Tullos 2013, or the Czech Republic, Kubečka et al. 1997) or the time of year. The degree of flow

alteration in bypassed reaches can be especially pronounced during seasonal periods of low to

moderate natural flows, when the amount of water diverted by RoR hydropower operations

translates into the highest proportion of flow removed from the channels. For example, up to

97% of the natural flow was diverted for RoR hydropower during fall, winter, early spring, and

late summer months in some snow-dominated watersheds of western Canada or during dry fall,

mid-summer, and winter months in rainfall-runoff dominated mountainous areas of China

(Figure 2.3). Diverting the highest proportion of flow during periods of low flow results in large

changes to the frequency, duration, timing, and magnitude of low flows in bypassed reaches

compared to natural conditions (Ovidio et al. 2008; Kibler and Tullos 2013). In contrast, during

seasonal periods of high natural flows (and episodic high flow events), bypassed reaches

downstream of RoR hydropower dams more closely match the NFR because headponds have

limited water storage capacity (Poff and Hart 2002) and withdrawals for power production are

proportionally smaller. For example, the proportion of flow removed during late spring and early

summer months only equals 6-30% of discharge during high flows in the snow-dominated

watersheds of Northwestern Canada or rainfall-runoff dominated mountainous areas of China

(Figure 2.3). Seasonal alterations to the NFR by RoR hydropower operations, in addition to the

physical barrier presented by low-head dams, in turn affect physical and geomorphic

characteristics of rivers that are important for salmonid fishes.

11

A large amount of literature documents changes in water quality, habitat quantity, and

geomorphology in rivers downstream of reservoir-storage hydropower systems (reviewed by

Poff and Zimmerman 2010). Though evidence in the peer-reviewed literature specific to RoR

hydropower is more limited (n= 31), it suggests water quality, habitat quantity, and

geomorphology can also be affected by RoR hydropower operations (Kubečka et al. 1997;

Baker et al. 2011; Nislow and Armstrong 2012; Bilotta et al. 2016). Changes to NFR following

diversion of flow can affect water quality mainly through changes to temperature regimes, pH,

and dissolved oxygen (Valero 2012). Changes to NFR also alter channel hydraulic, sediment

transport, and geomorphology downstream of low-head dams. Such changes to water quality

and physical habitat in turn diminish habitat quality, quantity, and diversity for fishes (Baker et al.

2011; Fuller et al. 2016). The diversion of flow for RoR hydropower, and subsequent changes to

water quality and physical habitats in bypassed reaches, therefore constitutes the first pathway

of effect by which RoR hydropower has the potential to influence salmonids. We further

consider two additional pathways of effect based on other characteristics of RoR hydropower:

the presence of low-head dams (Pathway 2), and anthropogenic flow fluctuations due to the

diversion of water (Pathway 3). The three pathways of effect are described below, where we

draw upon the empirical evidence from the peer-review literature specific to RoR hydropower to

support hypothesized mechanisms of impact for salmonid fishes.

2.5. Pathway 1: Reduction of flow in the bypassed reach

The effects of flow diversion on water quality and habitat quantity in bypassed

reaches were reported in 24 studies specifically focussed on RoR hydropower (Figure 2.2),

although only 10 directly report consequences for salmonids (Table 2-1). Consequences for fish

ranged from shifts in species assemblages and age composition, to declines in biomass and

density. For example, in the bypassed reaches of 20 out of 23 RoR hydropower systems

surveyed in the Czech Republic, water diversion by RoR hydropower dams was found to cause

a shift in fish assemblages from large- to small-bodied species, as well as a decline in individual

weight and biomass (Kubečka et al. 1997). Those systems that diverted more than half the

average annual discharge saw decreases in fish biomass of over 60% (Kubečka et al. 1997).

Similarly, low flows generated by RoR hydropower were associated with a decrease in adult

trout densities in 7 out of 11 bypassed reaches studied in France (Sabaton et al. 2008), as well

12

as declines of 42-53% in the biomass of brown trout and shifts in size and age composition from

adults to juveniles over four years in a bypassed reach in Belgium (Ovidio et al. 2008). The

operation of RoR hydropower plants also lead to small decreases in species richness in several

rivers in the UK (Bilotta et al. 2016). Below we outline the main mechanisms by which declines

in flow may lead to declines in salmonid biomass, density, and changes in size structure in

bypassed reaches of rivers regulated by RoR hydropower.

The diversion of flow for the production of electricity has led to documented

changes to water quality parameters including dissolved oxygen, pH, conductivity, invertebrate

communities, and temperature regimes in bypassed reaches of rivers regulated by RoR

hydropower. In general, changes to water quality parameters were small, and frequently studies

offered opposite conclusions as to the direction and magnitude of effects, especially for water

temperature and macroinvertebrates communities. For example, Valero (2012) recorded short-

term increases in pH as well as declines in dissolved oxygen and conductivity in the bypassed

reach of a RoR hydropower project in Spain. However, slightly lower average pH and variable

conductivity were noted in bypassed reaches in China (Zhou et al. 2009; Wu et al. 2010a; Wu et

al. 2012). Empirical evidence of the effects of flow reduction on water temperature in bypassed

reaches is also mixed. The factors influencing stream temperatures are a complex mix of

external drivers and internal stream dynamics (Poole and Berman 2001), and reducing flow in a

river reach has the potential to lead to warmer and more variable stream temperature regimes.

Studies conducted in bypassed reaches of rivers with RoR hydropower dams report slight

increases in water temperature in Spain (Valero 2012), USA (McManamay et al. 2015), and

China (Fu et al. 2008; Zhou et al. 2008; Zhou et al. 2009; Wu et al. 2010a; Wu et al. 2012), but

no significant differences between bypassed and upstream reaches in Portugal (Jesus et al.

2004). Low flows during dry and hot seasons, however, triggered markedly warmer

temperatures in bypassed reaches (~1-3℃) compared to upstream reaches in Costa Rica

(Anderson et al. 2006) and the Czech Republic (Kubečka et al. 1997), and compared to water

temperature immediately downstream in China (~5-6C, Wu et al. 2010b). Finally, primary

producers like diatoms were adversely affected in bypassed reaches in China (Wu et al. 2010b;

Wu et al. 2012), while one study points out that significant differences in algal community

composition between upstream and bypassed sites appeared only after two years of operation

(Wu et al. 2009). In addition, macroinvertebrates and zooplankton were also affected, with for

example decreases in species richness and abundance of benthic macroinvertebrates of up to

38% and 54%, respectively, in Sweden (Englund and Malmqvist 1996), decreases in biomass,

13

density and richness of macroinvertebrates in Portugal (Jesus et al. 2004) and China (Fu et al.

2008), and decreases in density of zooplankton during low flow months in China (Zhou et al.

2008). Conversely, other studies in bypassed reaches of RoR hydropower systems documented

increases in benthic diatom richness, likely due to a relaxation of predation pressure from

macroinvertebrates and creation of habitat more favorable to algal growth at lower flows (Wu et

al. 2010b). In other cases, no conclusive evidence of effects, positive or negative, on

macroinvertebrate biomass or density have been observed in bypassed reaches (Kubečka et al.

1997; Sabaton et al. 2008). Despite such empirical evidence of changes to water quality

parameters in bypassed reaches of RoR hydropower, neither the factors influencing these

changes nor consequences for fish have been clearly identified in the studies reviewed. For

example, of all the studies reporting on changes to water temperatures, pH or conductivity, only

three (Anderson et al. 2006; Kubečka et al. 1997; Valero 2012) report on the size of the rivers

(small to medium), the length of bypassed reaches (0.05 to 4 km), or the amount of flow

diverted (59-100%) (Table 2-1). Details about designs and river characteristics are equally

sparse in the studies reporting changes to invertebrate assemblages (Table 2-1). Such limited

details about ecological context and RoR hydropower design on rivers where studies were

conducted limit the opportunity to suggest why some studies would note increases in water

quality parameters while others do not. We can nonetheless hypothesize that changes to water

quality, temperature, and invertebrates are likely to affect fish habitat and food supply.

The magnitude of changes to water quality parameters noted in peer-reviewed

literature specific to RoR hydropower are generally small in magnitude, and so their effects on

salmonids remains unclear. Of all water quality parameters discussed above, alterations to

temperature have the highest likelihood of significantly impacting cold-water species like

salmonids (McManamay et al. 2015) since even small changes to water temperature regimes

(e.g. 0.6°C) have the potential to directly affect metabolism and growth of poikilotherm fishes

like salmonids (Beakes et al. 2014). In general, higher temperatures increase metabolic rates

and potential for growth up to the thermal optimum, beyond which increases in water

temperature are physiologically detrimental to fish and can lead to death (Brett 1995). Below the

thermal optimum, increases in growth with temperature are only possible when food or other

resources are not limiting (Bryant 2009; Taylor and Walters 2010). If food consumption

decreases while water temperatures increase, fish experience higher metabolic costs that may

lead to slower growth and later maturation (e.g., van Poorten and McAdam 2010), or declines in

total biomass (Beakes et al. 2014). The most important characteristic of RoR hydropower for

14

temperature regimes in bypassed reaches may be the seasonality of flow diversion, and its

consequence for water temperature, and ultimately, salmonids. For example, diverting

proportionally more flow during periods of naturally low flows may accelerate the timing and

increase the magnitude of warming in bypassed reaches of RoR hydropower systems during

spring and summer. These periods of accentuated low flow in spring and summer also

correspond to important times for spawning, incubating or rearing salmonids (Table 2-2). As

found by studies conducted in reservoir-storage systems and in unregulated rivers, increased

temperatures coinciding with low flows have the potential to change the timing of spawning

migrations and interspecific interactions (e.g., Freeman et al. 2001; Bendall et al. 2012; Malcolm

et al. 2012), reduce survival of smolts before and during migrations (Nislow and Armstrong

2012), and make fish more vulnerable to pathogens (Crozier et al. 2008; Mantua et al. 2010).

Additionally, these extended periods of low flow can reduce water quality, limit movement of

nutrients and sediment, and increase competition and predation (Lake 2000; Bradford and

Heinonen 2008; Walters and Post 2011). In contrast, during winter months, reduced flows may

decrease water temperature and increase the occurrence of frazil ice (i.e. ice anchored to the

stream bottom) and freeze-thaw cycles, potentially leading to mortality of salmonid eggs by

reducing oxygen concentrations, and of fry by damaging gill tissues (Bradford 1994; Bradford

and Heinonen 2008). Though the consequences of low flows and changes to water temperature

regimes observed in reservoir-storage systems and unregulated rivers may also manifest in

bypassed reaches of rivers regulated by RoR hydropower, the magnitude of the differences in

temperature experienced in bypassed reaches following flow diversion is poorly documented

and so should be established more clearly and rigorously by future research and monitoring. As

the River Continuum (Vannote et al. 1980) and Serial Discontinuity concepts (Ward and

Stanford 1983) predict, natural gradients in abiotic parameters like temperature exist from

headwaters to mouths of river systems. Control-impact comparisons between upstream and

bypassed reaches can thus potentially confound water quality changes due to diversion of flow

with the natural longitudinal changes through watershed networks. Comparisons to conditions

before diversion of flow, or in reaches of similar order and network position, are thus needed to

rigorously quantify the magnitude of changes to water temperature and other water quality

parameters after flow diversion, and understand their consequences for metabolism, growth,

and population dynamics of salmonids.

The second mechanism by which a reduction in flow from RoR hydropower operations

could impact fish is through a reduction in habitat quantity and diversity (Anderson et al. 2006;

15

Baker et al. 2011; Mueller et al. 2011). Overall, the literature demonstrates that the diversion of

flow in streams reduces the depth and velocity of water in bypassed reaches, which generally

lead to a decline in habitat quantity and quality, and in turn, a decline in fish biomass and

density. For example, declines in the number of habitats due to reductions in flow was linked to

a shift towards small-bodied fish species observed in several rivers regulated by RoR

hydropower in the Czech Republic, in part because of increases in predation on larger-bodied

fish species (Kubečka et al. 1997). Declines in European grayling (Thymallus thymallus)

biomass of 76% and declines in proportions of adults in the population were associated with

reductions in preferred deep and fast water habitats (Ovidio et al. 2008), a pattern also

observed with reductions of rainbow and brown trout abundances in a bypassed reach in Chile

(Habit et al. 2007). Lower densities of brown trout were found when greater amounts of water

were diverted from bypassed reaches in a long-term study (Gouraud et al. 2008). Most of these

systems appear to be in rather small, headwater rivers, with long bypassed reaches that

diverted a high proportion of water during most of the year (Table 2-1). However, reduced flows

in bypassed reaches were found to temper some negative effects of natural flooding events by

reducing the mortality of emerging rainbow and brown trout fry in similar rivers in France (Capra

et al. 2003; Gouraud et al. 2008). The potential for negative impacts of natural high spring or

summer discharges on trout fry is further supported by studies on the influence of hydrological

and biotic processes on the population dynamics of brown trout in France (Cattanéo et al.

2002), and on steelhead population dynamics in snowmelt-driven rivers of western Canada

(Smith 2000). Anthropogenically-created lower flows can also reduce the metabolic costs of

foraging and maintaining position for fish, thus increasing the amount of energy available for

growth, as has been shown in reaches downstream of large reservoir-storage hydropower

(Cleary et al. 2012). However, other research conducted in reservoir-storage systems supports

the idea that reduced habitat availability due to lower flows downstream of reservoir-storage

dams result in short-term increases in fish densities, subsequently leading to increases in

competition among and within salmonid life stages, and increased susceptibility to disease

(Bradford 1994; Nislow and Armstrong 2012). Geomorphically, lower flows can also reduce the

frequency, diversity, and quantity of microhabitats including gravel bars, side channels, and

pools which are important for salmonid spawning, overwintering and rearing, as well as for

refugia during extreme flow events like floods or droughts (Sedell et al. 1990; Bonneau and

Scarnecchia 1998; Walters and Post 2008). In many countries and jurisdictions, mitigation

measures for RoR hydropower operation, such as requiring minimum instream flows, are

mandated to minimize the potential impacts of reduced flows on fish and fish habitat in RoR

16

hydropower systems (e.g., France, Gouraud et al. 2008; Portugal, Santos et al. 2006; Sweden,

Renöfält et al. 2010; for more information, see review by Anderson et al. 2014). In some cases,

river systems with a highly modified NFR, but observing the required minimum flows (3-12% of

natural flow) in reaches downstream of reservoir-storage and low-head dams, were able to

maintain highly productive trout and invertebrate populations (Jowett and Biggs 2006).

Unfortunately, only half of the empirical papers on effects of RoR hydropower that we reviewed

mentioned the presence or absence of mitigation measures (n= 16), and none of them

discussed their results in light of the mitigation measures used, which prevents us from drawing

stronger inferences about their effectiveness. Ultimately, the consequences of flow reductions in

bypassed reaches downstream of RoR dams are likely to be highly dependent on the systems

in which they occur. Bypassed reaches in fast flow, flashy systems may benefit from a decline in

discharge and velocity because it would increase the amount of habitat available to fish during

medium to high flows. On the other hand, the opposite may be true for lower gradient, more

meandering rivers where a reduction in flow may limit the amount of habitat available to fish.

2.6. Pathway 2: The presence of low-head dams

A total of 13 peer-reviewed papers reported on the consequences of low-head dams in

the context of RoR hydropower, ranging from fish entrainment in intake structures, to barriers to

migration and habitat fragmentation, and effects on geomorphology and sediment transport

(Figure 2.2,Table 2-1). Despite the limited peer-reviewed literature dedicated specifically to RoR

hydropower, the potential for fish entrainment in infrastructures generating RoR hydropower is

clear. Similar to reservoir-storage hydropower systems (Skalski et al. 2002), the intake

structures of RoR hydropower projects may entrain fish in the penstock or turbines, leading to

injury and mortality (Kubečka et al. 1997). Mortality rates following entrainment in intake

structures of RoR hydropower dams increase with fish size (Kubečka et al. 1997), hydraulic

head, the number of blades, and varies by turbine type (on average, 100% in Pelton turbines, 5-

90% in Francis turbines, and 5-20% in Kaplan turbines; (Larinier 2008). Turbine mortality is

inversely related to RoR hydropower plant size, because smaller capacity plants often contain

small turbines that rotate faster than those in larger plants (Larinier 2008). The magnitude and

seasonality of flow diversion by RoR hydropower may also affect fish entrainment in RoR

hydropower intake structures. For example, fry in reaches upstream of RoR hydropower intakes

may be at higher risk of entrainment during periods of naturally high flows that often coincide

with when fry emerge (Table 2-2). Research conducted in reaches immediately downstream of

17

reservoir-storage dams has documented increased mortality from predation due to

disorientation or loss of equilibrium following entrainment (Čada 2001; Barnthouse 2013). The

potential for similar sub-lethal effects of entrainment through the intake structures specific to

RoR hydropower warrant investigation, especially given their long penstocks which may

increase negative consequences of entrainment for fish compared to reservoir-storage systems.

Finally, entrainment of invertebrates through RoR hydropower intake structures is also likely and

may be accentuated at low flows, given empirical evidence noted from RoR dams built to supply

water for municipal uses where up to 100% of drifting shrimp larvae were entrained at low flows

(Benstead et al. 1999). Overall, the potential for fish entrainment through RoR hydropower

structures is substantial and additional investigations specific to RoR operations are warranted.

The low-head dams associated with RoR hydropower can act as barriers to upstream

and downstream movements of fishes (Santucci et al. 2005; Habit et al. 2007; Santos et al.

2012) and fragment river networks (Vannote et al. 1980). Many RoR hydropower dams have

fish passage structures to restore upstream, but not downstream, migrations, however their

efficiency in passing fish is highly variable (Kubečka et al. 1997; Santos et al. 2012). For

example, the presence of 15 consecutive RoR dams in a single river in the USA blocked

upstream migrations of up to a third of local fish species, leading to species being present only

in the upper or lower sections, but not in the central regions of the river where most of the dams

were concentrated (Santucci et al. 2005). Another study in France found that only 16 out of 30

(53%) fish passage structures at RoR hydropower dams allowed migrating individuals to move

upstream without delays (Larinier 2008), and a study in Portugal found that 8 out of 18 (44%)

fish passage structures allowed fish passage between bypassed and upstream reaches (Santos

et al. 2006). In contrast, downstream migrations are expected to be less affected than upstream

migrations since small-scale RoR dams are generally low-elevation and passable at high flows,

provided that entrainment does not occur and passage over the dam is not traumatic for fish

(Anderson et al. 2006; Boubée and Williams 2006; Larinier 2008). For example, neither the

relative abundance, richness, nor diversity of fishes differed between upstream and bypassed

reaches in 18 RoR hydropower systems studied in Portugal, despite the fact that 55% of the

dams had unsuitable fish passage structures, suggesting that downstream migrations of

migratory species at high flows might be occurring (Santos et al. 2006). Furthermore, indicators

of community composition like species richness and fish abundance were not statistically

different between control and impacted sites in several rivers regulated by RoR hydropower in

the UK (Bilotta et al. 2016). However, large uncertainties remain regarding impacts of low-head

18

dams on downstream fish movements that do not coincide seasonally with high flows.

Nevertheless, the artificial disruption of longitudinal connectivity by RoR dams is akin to

terrestrial habitat fragmentation, and can affect the composition of fish assemblages by favoring

more generalist species (e.g., Santucci et al. 2005; Anderson et al. 2006) or disfavoring small or

benthic species unable to pass dams (Habit et al. 2007). RoR hydropower dams are often

located in high gradient streams that tend to support relatively small salmonid populations, and

habitat fragmentation as a result of impassable barriers may also increase the potential for

genetic drift or reduce population viability. Small populations are generally at higher risk of

adverse consequences from genetic drift, including inbreeding depression and increased

vulnerability to environmental stress and stochasticity (Altukhow et al. 2000; Heggenes and

Røed 2006;Lucas et al. 2009). Overall, the potential for habitat fragmentation due to the

presence of low-head RoR dams represents a substantial threat to salmonid populations, which

all undertake migrations between spawning, rearing and overwintering habitats.

Finally, RoR hydropower dams may act as discontinuities in the geomorphology

of rivers, affecting the natural transport of sediment and organic matter in streams (Fuller et al.

2016), and in turn, the quality and quantity of fish habitat. Low-head RoR hydropower dams and

the presence of headponds create conditions that are likely to interrupt the NFR and natural

longitudinal connectivity of rivers, but the fact that low-head dams often become overtopped at

high flows may compound the magnitude of the geomorphologic disruptions. Evidence suggests

that RoR dams and headponds temporarily store sediments, and therefore alter the timing and

size of sediment delivered to the bypassed and downstream reaches compared to NFR (Kibler

and Tullos 2013; Fuller et al. 2016). RoR hydropower dams may also intercept different forms of

organic matter including coarse woody debris. Coarse woody debris is a key component of fish

habitat as it promotes the formation of pools, limits erosion, and provides cover and refugia

(Sedell et al. 1990; Mossop and Bradford 2004). Comparisons between bypassed and upstream

reaches of low-head RoR dams found higher levels of fine sediment and significantly slower

water velocity in 13 bypassed reaches, as well as significant differences between upstream and

bypassed reaches in 32 of 41 hydraulic variables (Baker et al. 2011). Based on their modelling

results, Baker et al. (2011) also concluded that small, low-gradient streams with smaller-sized

substrate were more susceptible to fine-sediment accumulation than large streams. However,

RoR dams that are regularly overtopped at high flows are expected to experience fewer

discontinuities in the morphological and sediment size distribution of stream channels (Kondolf

1997; Kibler and Tullos 2013; Csiki and Rhoads 2013). The proportion of flow diverted is also of

19

importance for the transport of sediments. For example, Morris (1992) found that the diversion

of less than 2 m3s-1 (< 8% of annual peak flood) of water for hydroelectricity production did not

affect the transport of spawning size gravel in a salmonid stream, given the considerably higher

discharge left in the bypassed reach. Beyond the uncertainties regarding how sediment

transport is affected downstream of RoR hydropower dams, higher fine sediment deposition as

well as coarsening of substrates are expected to generate negative effects for salmonids. For

example, new research shows that grain-size coarsening downstream of powerhouses could

degrade salmonid habitat in RoR hydropower systems, particularly in streams with naturally

high sediment supply rates (Fuller et al. 2016). Based on literature from manipulative

experiments, higher fine sediment embeddedness will often induce a shift in invertebrate

assemblages from drifting to burrowing taxa, resulting in reduced food supply for resident

salmonids, and higher feeding costs (Suttle et al. 2004). Higher embeddedness can also clog

spawning gravel, reducing the survival of overwintering eggs and alevins (Kondolf 1997), and

affect primary production by decreasing diatom richness and diversity (Wu et al. 2009). For

example, experimental increases in deposited fine sediments in shallow riffles led to up to 90%

lower rainbow trout survival attributed to the decline in overall riverine habitat complexity and

increased predation (Harvey et al. 2009). We conclude that fish entrainment, habitat

fragmentation, and alteration to geomorphology and sediment transport induced by the

presence of low-head dams all have the potential to negatively affect salmonids. The fact that

RoR hydropower dams may be overtopped at high flows has the potential to mediate the

severity of some of these impacts, but the absence of published research makes such

determinations impossible at present.

2.7. Pathway 3: Anthropogenic Flow Fluctuations

We found only one peer-reviewed paper (Almodόvar and Nicola 1999) that evaluated the

potential for flow fluctuations downstream of RoR hydropower dams to impact fishes (Figure

2.2, Table 2-1). However, the operation of RoR hydropower dams has the potential to create

anthropogenic variations in flow in both bypassed and downstream reaches of rivers where they

operate. Rapid fluctuations in flow downstream of dams and powerhouses in RoR hydropower

systems may occur as a result of intentional increases or decreases in the proportion of stream

flow diverted to turbines to optimize electricity production, or because of emergency shutdowns

or operational malfunctions that unexpectedly halts the diversion of water. Because headponds

have no water storage capacity, changes in the amount of water diverted for RoR electricity

20

generation are quickly propagated downstream as flow fluctuations in both the bypassed and

downstream reaches. However, the potential effects of flow fluctuations on river ecosystems

and salmonids differ in bypassed and downstream reaches. As the amount of stream flow

diverted to RoR turbines decreases, the amount of water rerouted to the bypassed reach

increases rapidly, while flow released at the tailrace of the powerhouse is reduced. The

reduction of flow at the tailrace causes a temporary decrease in flow in the downstream reach,

lasting until the rerouted water travels through the bypassed reach and reaches the

powerhouse. The drop in flow downstream of powerhouses depends on site-specific

characteristics including channel confinement, substrate type, and bathymetry, as well as

characteristics of the RoR hydropower facilities, including the length of the penstock and the

presence of mitigating structures like turbine bypass valves (Hunter 1992 in Bell et al. 2008).

Rapid anthropogenic fluctuations of flow in bypassed and downstream reaches have the

potential to create negative consequences for fish ranging from unintentional downstream

displacement and increased stress and mortality, to interference with natural cues used to

indicate flood or a drought risk and mismatches in behavioral strategies (Lytle and Poff 2004).

We found only a single study from the peer-reviewed literature specific to RoR hydropower,

which reported that fluctuations in flow downstream of a small RoR diversion dam in Spain

(diverting 85 to 100% of the seasonal discharge) lead to an average change in stage of 30 cm

over minutes (Almodόvar and Nicola 1999). These frequent fluctuations in flow caused declines

of brown trout density (-50%) and biomass (-43%), displacement of 0+ trout, and no change to

macroinvertebrates in the first year following flow diversion (Almodόvar and Nicola 1999). In

reservoir-storage systems, rapid increases in flow have also been associated with the

downstream displacement of juvenile trout, increased energetic costs for fry, and scouring of

redds (Harby and Halleraker 2001; Nislow and Armstrong 2012). Conversely, rapid decreases in

flow can lead to a decrease in stage in downstream reaches that can strand fish on rapidly

dewatered channel margins, or trap fish in disconnected side channels and isolated pools. Data

from an unregulated montane stream in the USA Pacific Northwest show that natural

fluctuations in stage rarely exceeded 5 cm per hour (Hunter 1992 in Bell et al. 2008), while

frequent declines in river stage of 80-90 cm over 10 minutes were reported downstream of a

reservoir-storage dam in Norway (Hvidsten 1985). Fish stranding or isolation in side channels

may in turn lead to negative effects on survival, biomass, density, or fitness, as reported in

studies from reservoir-storage systems (Young et al. 2011; Nagrodski et al. 2012; Senay et al.

2016). Given the limited research that has specifically sought to understand the consequences

21

of anthropogenic flow fluctuations in RoR hydropower systems, large uncertainties remain about

how their effects differ from those documented downstream of reservoir-storage systems where

most peer-reviewed research has occurred.

RoR hydropower operations produce artificial fluctuations in flow that are likely to pose

risks to salmonids that differ from those posed by large dams and reservoir-storage systems.

For example, periods of natural low flow are often when RoR hydropower operations divert the

highest proportion of flow, or create frequent fluctuations from practices called ‘flow cycling’, by

which water is temporarily stored in penstocks to generate power for short periods of time

(Hunter 1992 in Bell et al. 2008). These periods often coincide with the presence of newly

emerged salmonid fry. Salmonid fry are likely to be the most vulnerable life history stage to the

negative effects of anthropogenic flow fluctuations as they have limited swimming capacities

and inhabit low-velocity, shallow habitats that are highly susceptible to dewatering (Bell et al.

2008; Korman et al. 2011). Flow fluctuations during natural low flow periods may also reduce

the availability of salmonid rearing habitats, as well as interfere with spawning of anadromous

and resident salmonids downstream of powerhouses (Nagrodski et al. 2012). Besides their

timing, the frequency and magnitude of anthropogenic flow fluctuations from RoR operations

have the potential to generate impacts to salmonids. For example, even small, but repeated,

flow fluctuations were found to decrease the feeding time, growth, and survival of fish (Young et

al. 2011). In addition, small but frequent anthropogenic fluctuations in flow have also been found

to reduce fish density and biomass when compared to infrequent but large fluctuations

generated by storage-reservoir hydropower in Canada (Senay et al. 2016). These findings may

be particularly relevant for RoR hydropower systems which can create more frequent pulses in

flow compared to reservoir storage hydropower due to the lack of water storage in headponds,

and wider magnitude fluctuations during low flow seasons. Overall, how seasonal low flow

periods and anthropogenic flow fluctuations interact to influence salmonid survival and growth in

bypassed and downstream reaches of rivers regulated by RoR hydropower remains largely

unknown. However, the potential for fluctuating flows to perturb fish habitat, induce fry mortality,

and increase stress and activity costs for salmonids suggest that anthropogenic flow fluctuations

are likely to be important for salmonid populations.

2.8. Avenues for Future Research

As highlighted by our literature search, published empirical research on the impact of

RoR hydropower on the ecology of river ecosystems (n= 31 peer-reviewed studies) and

22

salmonids specifically (n= 10 peer-reviewed studies) is very limited. Despite this limited

literature, we found that RoR hydropower operations alter NFR in unique ways by changing the

timing, and increasing the magnitude, frequency, and duration of low flow periods in bypassed

reaches downstream of RoR hydropower dams. Moreover, the low-head dams of RoR

hydropower facilities create unique conditions upstream and downstream since they may be

overtopped by water at high flows. Though in need of additional study, this phenomenon may

allow for more flushing of sediment and downstream passage of fish compared to reservoir-

storage dams, thus potentially reducing the potential for discontinuities in habitat and impacts on

stream geomorphology. RoR hydropower dams also divert a considerable proportion (up to 97%

of incoming flow, Figure 2.3) of water away from the natural river channel for the production of

electricity. We found a clear pattern where the magnitude of flow diversion by RoR hydropower

varies by season and is proportionally higher when flows in rivers are naturally low. Moreover,

flow fluctuations in bypassed and downstream reaches of RoR hydropower systems typically

occur unexpectedly, and differ in timing, frequency, magnitude, and duration from those

occurring downstream of reservoir-storage systems. All these pathways of effects from RoR

hydropower have the potential to alter habitat quantity and quality, and to negatively affect

salmonid survival, growth, and fitness. We reiterate the need to standardize data and report

details on project designs and environmental characteristics of rivers where the studies are

conducted to allow for more powerful quantitative comparisons between RoR hydropower

schemes and studies in the future (Anderson et al. 2014, Bilotta et al. 2016). For example, very

few papers reported ecological details like climactic zones (only 6 of the 31 empirical papers),

location in watershed (n= 9 out of the 31 empirical papers), or river size (only 17 of the 31

empirical papers reported mean annual discharge for example), which precluded us from

explicitly categorizing the studies and the range of effects based on these criteria. Our review

also highlights the lack of empirical knowledge about the effects of RoR hydropower on

salmonid species other than brown trout, which were the focus of over 90% of the empirical

papers on salmonids we reviewed. In addition to the specific knowledge gaps identified in our

review of the literature in each section above, below we identify three areas where future

research is most needed. We chose to highlight these three research priorities because we

believe they represent uncertainties related to those processes with the greatest potential to

affect resident and anadromous fish populations. Addressing these knowledge gaps will help

reduce the currently large uncertainties surrounding impacts of RoR hydropower on river

ecosystems and salmonids, and in turn reduce conflict between proponents and opponents of

additional RoR hydropower energy development.

23

(1) Effects of amplified, short- and long-term low flow periods. The effects of

natural and anthropogenically-induced high flows on salmonids have received more attention

than the effects of low flows (Lake 2000; Jowett et al. 2005). With the exception of research on

environmental flows (e.g., Poff and Matthews 2013), most research has focused on large

magnitude changes, particularly for large dams and reservoir-storage systems (Poff and

Zimmerman 2010). Our literature review revealed that RoR hydropower operations artificially

generate longer and more frequent periods of low flow compared to natural flow regimes. The

semi-predictable seasonal low flows resulting from RoR hydropower operation also represent a

unique opportunity to study how increases in the duration, frequency, and magnitude of low flow

periods affect fish and river ecosystems in isolation of other factors (Niemi et al. 1990; Matthews

and Marsh-Matthews 2003; Waples et al. 2009).

(2) Effects of anthropogenic flow fluctuations. Other reviews have highlighted

important knowledge gaps about how the magnitude, frequency, and timing of anthropogenic

flow fluctuations downstream of reservoir-storage hydropower systems affect the growth,

survival, and reproductive success of fish (Young et al. 2011; Nagrodski et al. 2012). These

knowledge gaps are also relevant to rivers regulated by RoR hydropower, although we argue

the gaps are even larger. We found very little peer-reviewed literature that examined how RoR

hydropower fluctuations in flow deviate from the NFR or how these deviations impact salmonids

in the bypassed and downstream reaches (n= 1). In addition to direct impacts, research is

needed to quantify potential indirect impacts of anthropogenic flow fluctuations on salmonids, as

deviations from NFR are expected to affect food availability for fish by altering the composition

and biomass of invertebrate communities. Beyond individual-level consequences, there is also a

need to understand how short-term and individual level impacts of flow fluctuations scale up to

the population level. Individual mortality or reductions in reproductive potential due to fish

stranding can translate into reductions in population growth rates if they occur frequently, but

density-dependent growth and survival of salmonid fishes may partially or completely

compensate for early life-history mortality caused by RoR operations (Shuter 1990; Harby and

Halleraker 2001). The consequences of anthropogenic flow fluctuations on individual fishes and

populations need to be established in RoR hydropower systems before appropriate mitigation

measures can be developed to compensate for potential losses.

(3) Cumulative impacts of multiple RoR hydropower projects. Multiple RoR

hydropower projects may be developed within individual watersheds to achieve energy

production targets. For example, 15 low-elevation dams exist over 160 km in a Illinois river, USA

24

(Santucci et al. 2005), and over 260 RoR hydropower dams are currently under consideration in

28 of the 32 major river valleys in the Indian Himalayas (1 RoR hydropower dam per 32 km of

river, Grumbine and Pandit 2013). Multiple RoR hydropower projects in the same watershed

may interact in additive, synergistic, or antagonistic ways to affect river ecosystems (e.g.

increasing turbidity and decreasing invertebrate quality, Santucci et al. 2005) and salmonid

populations (e.g. to alter or delay upstream fish migrations, Larinier 2008; Lucas et al. 2009).

Some studies suggest that landscape-scale impacts per unit of energy may be greater for RoR

hydropower dams than for large hydropower dams (Bakken et al. 2012; Kibler and Tullos 2013),

while others reach the opposite conclusion (Taylor 2010). However, these conclusions are

limited because they have not explicitly considered whether impacts from multiple RoR dams

accumulate in non-linear ways (Larinier 2008), as observed for populations migrating past

multiple large-scale hydropower dams in the lower Columbia River (Elder et al. 2016). In

addition, the impacts of RoR hydropower to salmonid populations are likely to depend on the

regional context in which individual RoR hydropower projects are built (Habit et al. 2007). Flow

diversion by RoR hydropower in watersheds with existing impacts from forestry and other

industries will likely have different, and possibly compounding, effects on salmonids in contrast

to watersheds with largely intact upland and riparian forests (Reice et al. 1990; Bunn and

Arthington 2002). The disparity between the regional spatial scale relevant to the survival and

persistence of most salmonid populations and the often reach-specific scale at which

environmental impacts of individual RoR hydropower projects are typically assessed needs to

be addressed.

The rapid adoption of RoR hydropower around the world presents both

opportunities for increasing renewable energy production, and challenges in predicting impacts

to salmonids and river ecosystems. The spatial and temporal scales of anthropogenic

perturbations from RoR hydropower are likely to differ from the historical conditions under which

salmonids evolved. These perturbations may alter life-history diversity, population size, and

connectivity among sub-populations, weakening their ability to dampen variability in the

dynamics of regional population complexes (i.e., the portfolio effect; Tilman et al. 1998; Yeakel

et al. 2014) and leading to reduced population resilience in the face of additional environmental

change (Waples et al. 2008; Waples et al. 2009; Moore et al. 2015). The inherent characteristics

of RoR hydropower infrastructures and operations, such as small non-storage headponds, low-

head dams, bypassed reaches with reduced flows, and anthropogenic flow fluctuations,

uniquely alter NFR. The limited peer-reviewed literature we found specific to RoR hydropower

25

strongly suggests that impacts to salmonids and other fishes are probable under certain

conditions, which is supported by better studied impacts of other forms of flow regulation.

However, we caution that the knowledge developed from other types of flow regulation may not

be directly applicable to RoR hydropower (Senay et al. 2016). As such, research specific to how

RoR hydropower operations alter the NFR and how these alterations impact fish populations

and river ecosystems is needed to confirm the pathways and mechanisms we outline in this

paper. A suite of small-scale studies examining individual-level impacts on salmonids as well as

long-term and large-scale experimental studies would be useful to identify effective mitigation

and management strategies for river ecosystems where RoR hydropower dams have the

potential to impact salmonids. These strategies may be as simple as protocols dictating the rate

at which water levels may be diverted or fluctuate, or as complex as full-watershed analyses to

determine how to best maintain connectivity of salmonid populations and manage cumulative

impacts across multiple dams and river networks. As the global need for RoR hydropower

increases, concerted efforts to study the impacts of RoR hydropower dams on salmonids is

crucial to ensure that evidence-based decisions are made for sustainable management of

salmonid populations, and the river ecosystems in which they occur.

Acknowledgements

The authors thank M.J. Bradford, T. Hatfield, V. Popescu, R. Munshaw, A. Kissel, R.

Murray, two anonymous reviewers and the handling editor for comments on earlier versions that

greatly improved the manuscript, and R. Munshaw for assistance with figures. This work was

funded by a Natural Science and Engineering Research Council (NSERC) Industrial

Postgraduate Scholarship in association with Ecofish Research Ltd. to P.G., an NSERC

Discovery Grant, and funding from the Gordon and Betty Moore and Wilburforce Foundations to

WJP.

26

Table 2-1. Empirical peer-reviewed papers studying RoR hydropower impacts on river ecosystems and salmonids. a= brown trout, b= European grayling, c= trout, d= rainbow trout, e= Atlantic salmon, f= high-head scheme, g= low-head scheme; **= dams were in cascade; MAD= mean annual discharge.

Papers Taxa/ characteristics

Mechanism(s) addressed

Type of dam(s) Number of dams

Height of dam(m)

Length of bypassed reach (km)

MAD (cms) Flow diverted

Pathway 1: Flow diversion

Bilotta et al. 2016

Fisha,e Reduction in habitat quantity and quality

hydropowerf,g

(54 kW) 23 11 0.22 --- ---

Capra et al. 2003

Fisha Reduction in habitat quantity and quality

Hydropowerf 1 --- --- 2.4 max of 4.4 cms

Gouraud et al. 2008

Fisha Reduction in habitat quantity and quality

Hydropowerf 3 7.1

4.5 8.55 --

McManamay et al. 2015

Fish Changes to water quality

Hydropower and others

22** -- -- -- --

Ovidio et al. 2008

Fisha,b Reduction in habitat quantity and quality

Hydropowerf

(900 MW) 1 -- 1.2 1.78 --

Sabaton et al. 2008

Fisha Changes to water quality

Hydropowerf 7 (on 5 streams)

-- 4.7 32.3 --

Mueller et al. 2011

Periphyton, aquatic macrophytes,

macroinvertebrates and fish

Changes to water quality, reduction in habitat quantity and

diversity

Hydropower 5 2.6 -- 5.7 --

Anderson et al. 2015

Benthic macroinvertebrates

Changes to water quality

Hydropowerf (30 kW)

1 3 ~ 0.1 1.17 1.05 cms

Englund and Malmqvist 1996

Macroinvertebrates

Changes to water quality

Hydropower and others

51 -- -- 10 86-99.8% reduction

Fu et al. 2008 Macroinvertebrates Changes to water quality

Hydropowerf 5** -- -- -- --

Jesus et al. 2004

Macroinvertebrates Changes to water quality

Hydropowerf 1 -- -- 0.93 90% in rainy season, 92%

27

in drought season

Wu et al 2010a Benthic diatoms Changes to water quality

Hydropowerf 23** -- -- -- --

Wu et al 2010b Benthic algal community

Changes to water quality

Hydropowerf 1 ~ 5 -- --

Wu et al. 2012 Benthic diatoms Changes to water quality

Hydropowerf 23** -- -- -- --

Zhou et al. 2008 Zooplankton Changes to water quality

Hydropowerf 1 -- -- -- --

Zhou et al. 2009 Zooplankton Changes to water quality

Hydropowerf 6** -- -- -- --

Valero 2012 Water quality Changes to water quality

Hydropowerf

(11.8 MW) 1 11.05 3.4 14.4 59% of water

flow

Pathway 1: Flow diversion and Pathway 2: Presence of low-head dams

Anderson et al. 2006

Fish Changes to water quality, reduction in habitat quantity and diversity, barrier to

migration

Hydropowerf (18 MW)

2 < 10 3 4.75 90-95% of MAD

Habit et al. 2007 Fisha,d Reduction in habitat quantity and quality, barrier to migration

Hydropowerf

(130 MW) 2 -- 14 25.8 up to 90% of

MAD

Jowett and Biggs 2006

Fishc mitigation measures Hydropower and others

1 -- -- 450 96.5-97.3%

Santos et al. 2006

Fisha Reduction in habitat quantity and quality, barrier to migration

Hydropowerf (3.97 MW)

18 5.8 --

--

--

Kubecka et al. 1997

Benthic and fish communitiesa,b,c

Changes to water quality, reduction in habitat quantity and

diversity, fish entrainment, and

barrier to migration

Hydropowerf (< 10 MW)

23 all but one < 2

0.89 1.185 up to 100% of discharge

Wu et al. 2009 Benthic algal community

Changes to water quality, discontinuities

Hydropowerf 1 1.5 > 3 -- --

28

in geomorphology

Pathway 1: Flow Diversion and Pathway 3: Flow fluctuations

Almodόvar and Nicola 1999

Fisha Changes to water quality, flow fluctuations

Hydropower (0.7 MW)

-- -- -- 0.40 (summer)

to 2 (winter)

--

Pathway 2: Presence of low-head dams

Boubée and Williams 2006

Fish entrainment Hydropowerf (4.5 MW)

1 3.5 -- -- --

Larinier 2008 Fish entrainment and barrier to migration

Hydropowerf,g (< 10 MW)

(several schemes)

-- -- varied --

Santos et al. 2012

Fish Barrier to migration Hydropower (< 10 MW)

37 -- -- -- --

Santucci et al. 2005

Fish, macroinvertebrates

Barrier to migration Hydropower and other uses

15** 0.85 NA 662 --

Fuller et al. 2016

Geomorphology discontinuities in geomorphology and sediment transport

Hydropowerf

1 -- 1 1.6 --

Kibler and Tullos 2013

Geomorphology discontinuities in geomorphology and sediment transport

Hydropowerf (19 MW)

31 8.8 6.8 3.1 up to 100%

Morris 1992 Spawning gravels discontinuities in geomorphology and sediment transport

Hydropowerf (2.3 MW)

1 -- ~ 6 -- 2

29

Table 2-2. Relative vulnerability of salmonid species to potential impacts from RoR hydropower, based on chronic exposure and the potential for dams to create barrier for migrations. Information on life history, spawning season and length of freshwater residency come from the references indicated by footnotes.

Common name Scientific name Life history Spawning season

Typical length of freshwater residencya

Relative vulnerability

Barrier to migrationb

Chronic exposurec

rainbow troutd,e Oncorhynchus mykiss

Resident Spring life High Very high

steelheadd,e Oncorhynchus mykiss

Anadromous Spring 1-3 years Very high High

brown troutd,e Salmo trutta Resident Late fall life High Very high

sea-run brown troutd,e

Salmo trutta Anadromous Late fall 1-3 years Very high High

cutthroat troutf Oncorhynchus clarki spp.

Resident Late winter, spring

life High Very high

sea-run cutthroat troutf

Oncorhynchus clarki spp.

Anadromous Late winter, spring

2-3 years Very high High

Stream resident Summer, fall life High

Very high

bull troutg

Salvelinus confluentus

Fluvial/adfluvial migrant

Summer, fall life Very High Very high

Anadromous Summer, fall 2-3 years Very High High

Dolly Vardenh

Salvelinus malma malma

Resident Fall life High Very High

Anadromous Fall 2-4 years Very High High

atlantic salmone Salmo salar Anadromous Fall 2-3 years Very High High

kokaneed Oncorhynchus nerka

Resident Fall life Moderate Low

30

sockeye salmond Oncorhynchus nerka

Anadromous Fall 1-2 years High Low

masu salmond

Oncorhynchus masou

Resident Fall life High Very High

Anadromous Fall 1-3 years Very High High

coho salmond Oncorhynchus kisutch

Anadromous Fall 1-2 years Very High High

Chinook salmond Oncorhynchus tshawytscha

Anadromous Fall 1-2 years Moderate Moderate

pink salmond Oncorhynchus gorbuscha

Anadromous Fall Days-weeks Low Low

chum salmond Oncorhynchus keta Anadromous Fall Days-weeks Low Low aDefined as time in freshwater post emergence from gravel bBased on extent of freshwater migration and size of river systems spawning and rearing typically occurs in cBased on portion of life cycle spent in freshwater and size of river systems spawning and rearing typically occurs in dGroot and Margolis 1995 eMcPail 2007 fCOSEWIC 2007 gCOSEWIC 2006 hCOSEWIC 2010

31

Figure 2.1. Three main pathways of effects of RoR hydropower on salmonids and their associated main mechanisms of impacts. The number of empirical peer-reviewed papers identified in this review is shown in boxes between the pathways of effect (1-3) and their primary mechanisms.

32

Figure 2.2. Generalized schematic of a run-of-river hydropower facility. A low-elevation dam impounds the river and reroutes water downstream through a penstock, to turn turbines in the powerhouse and produce electricity. The diversion of water leaves a reach of the river with reduced flows (bypassed reach), but all water is returned to the downstream reach of the river via the tailrace. The headpond has little water storage capacity and so does not act as a reservoir. Powerlines are part of the transmission corridor that connects the facility to a centralized power grid.

33

Figure 2.3. Modeled natural upstream flows (black line) and flows passing through the bypassed reaches (dashed line) during an average runoff year below a low-head RoR hydropower dam in a) western Canada (high-head scheme in snowmelt-dominated Bull River, natural flow averaged from 2010-2013, diverted flow modelled with a hypothetical turbine flow of 9.9 cms; www.bchydro.com), and b) Yunnan Province, China (rain-dominated Gutan River, Lushui County, adapted from Kibler and Tullos 2013). Flow alterations in the bypassed reaches are most pronounced during naturally low to moderate flows, when RoR hydropower operations divert the greatest proportion of flow from the channel. The Bull River has a minimum flow requirement that varies between 0.25 and 2.0 cms, depending on the season.

34

Effects of flow diversion by Run-of-River hydropower on stream temperature and bioenergetics of salmonid fishes2

3.1. Abstract

Many anthropogenic disturbances impact stream ecosystems by changing flow

and temperature regimes. For example, an emerging industry like Run-of-River (RoR)

hydropower reduces streamflow in bypassed reaches, with largely unknown consequences for

water temperatures and fish growth. We used empirical and simulated data from two RoR

regulated streams in British Columbia (Canada) to quantify changes in water temperatures and

assess the potential impacts to resident rainbow trout (Oncorhynchus mykiss) growth using

bioenergetics models under a range of consumption scenarios. We found increases in mean

monthly water temperature due to flow diversion of 0.5-0.8⁰C (0.17-0.19⁰C/km). Bioenergetics

models using those temperatures predicted large increases in annual O. mykiss growth

(compared to natural temperatures) if consumption was unlimited (+200-450%), increases (+15-

42%) if consumption scaled with higher metabolic demand, and small reductions (-5-7%) if

consumption remained constant. If food availability was reduced by 25%, annual growth

declined by 45%. Empirical estimates of annual growth of fish sampled at the top of the

bypassed reaches indicate modest reductions in annual growth less severe than those modeled

by our Reduced Consumption Scenario. Our results highlight that increases in water

temperature induced by RoR flow diversion are large enough to have consequences for O.

mykiss growth, but the impacts depend on how and when RoR affects food supply and

consumption.

2 A version of this chapter is in review at River Research and Applications (April 2020)

35

3.2. Introduction

Many anthropogenic disturbances affect lotic ecosystems by modifying flow and

temperature regimes. For example, the presence of dams and reservoirs on rivers often

reduces the frequency and magnitude of naturally-occurring high and low flows that are known

to structure riverine ecosystems (Poff et al. 1997). Modifications to flow regimes also influence

water temperatures (Poole and Berman 2001), with cascading effects to multiple trophic levels

including changes to food consumption, growth, and population dynamics of aquatic organisms

(Petersen and Kitchell 2001). Modifications to natural flow and temperature regimes have

consequences for biotic and abiotic processes in stream ecosystems, including primary

production (Englund and Malmqvist 1996), maintenance and creation of habitats (Poff et al.

1997), timing of life-history events (Bradford 1994), growth opportunities for fishes (Harvey et al.

2006), and adaptation to climate change (Ayllόn et al. 2019). A better understanding of the

consequences of altered flow and temperature regimes on abiotic conditions and stream-

dwelling organisms stands to improve the effectiveness of riverine management.

Small-scale hydropower facilities like Run-of-River (RoR) dams are an example

of an anthropogenic activity that has the potential to influence stream habitats and ecosystems

(Anderson et al. 2014; Couto and Olden 2018). The importance of small-scale hydropower has

increased around the world in recent decades, with global power generation increasing from

32,000 to 45,000 MW between 2001 and 2010 (Abbasi and Abbasi 2011). As of 2018, at least

82,891 small hydropower plants were operating in 150 countries, corresponding to 11 small

dams for every large hydropower plant, with a potential for this number to at least triple in

coming years (Couto and Olden 2018). Although there is no single accepted definition, small-

scale hydropower typically refers to various types of infrastructure that divert flow (without

storage) from small streams to produce electricity, generally below 10 MW (Couto and Olden

2018). Small-scale off-channel RoR hydropower (hereafter called RoR hydropower) is one form

of small-scale hydropower that diverts a proportion of stream flow for several kilometers through

a tunnel to a powerhouse where electricity is generated, before returning to the channel. Water

diverted for RoR hydropower creates a bypassed reach with reduced flow. Flows diverted by

RoR dams can represent a large proportion of total river flow, with up to 97% of natural stream

flow diverted at times (Gibeau et al. 2017). Impacts of RoR hydropower on stream ecosystems

remain poorly understood, despite the perception that RoR facilities have lower environmental,

economic, and social costs than large hydroelectric reservoirs (Couto and Olden 2018). The

36

limited literature on RoR hydropower impacts to stream ecosystems suggests the potential for

direct effects to fish migration, biomass, density, and species composition (Kubečka et al. 1997;

Bilotta et al. 2016), as well as negative effects on water quality and invertebrates (Valero 2012).

We hypothesize further that reductions in natural stream flow in bypassed reaches has the

potential to increase water temperatures by reducing water velocity and depth, as well as the

magnitude of hyporheic flow (Poole and Berman 2001). Warmer temperatures in turn can

induce direct mortality (Farrell et al. 2008), increase metabolism, or even increase growth

potential of stream fishes, but the consequences will likely vary based on other ecological

conditions in rivers, especially food availability (Railsback and Rose 1999). Warmer stream

temperatures from river diversion may create opportunities for increased growth in fishes if the

increase in metabolic demand is matched by an increase in food availability and consumption

(Petersen and Paukert 2005). Alternatively, if food consumption remains unchanged or

decreases in warmer temperatures, a higher proportion of energy intake will be required to meet

higher metabolic demands, thus leaving less energy for growth (Railsback and Rose 1999).

Bioenergetics models provide a unique way to contextualize the effects of

anthropogenic flow diversions within the spatial and temporal environment that fishes

experience, and consist of a set of size- and species-specific energy balance equations, where

energy gained is equated with energy spent for metabolism and growth (Hansen et al. 1993;

Petersen and Kitchell 2001). They synthesize the complex and often nonlinear relationships

between fish consumption or growth and prey availability, temperature, and other environmental

conditions (e.g. Railsback and Rose 1999; Taylor and Walters 2010). Bioenergetics models are

often heralded as useful "deductive engines" (Hartman and Kitchell 2008) for exploring

hypothesized scenarios based on empirical and simulated data, and allow for relative

comparisons between baseline and altered conditions (Chipps and Wahl 2008; Sweka and

Hartman 2008). Thus they present a useful tool for assessing consequences of flow diversion

by RoR hydropower on stream fishes.

Here we used empirical and simulated data from two streams regulated by RoR

hydropower in southwestern British Columbia (Canada) to quantify changes in water

temperatures for five years after flow diversion and assess the potential for such changes to

affect fish consumption and growth using a bioenergetics framework. Our specific objectives

were to: 1) quantify changes in water temperatures created by flow diversion from RoR

hydropower in bypassed reaches, and 2) explore how altered temperature regimes may

influence growth and consumption of fishes across a range of ecological scenarios. We

37

modelled consumption and growth of resident rainbow trout (Oncorhynchus mykiss) as their

habitats in high-gradient, mountainous streams often overlap with RoR hydropower

infrastructure (Gibeau et al. 2017). Given the paucity of data on primary and secondary

production in the streams, we calculated the consequences of four levels of food availability as

the basis for bioenergetics simulations, where RoR hydropower may affect food availability and

fish consumption through increases in water temperatures and/or declines in stream flow. Few

studies thus far have related fish growth or consumption to flow regulation from hydropower

facilities (Almodόvar and Nicola 1999; Petersen and Paukert 2005), and none have focused

specifically on the bioenergetic consequences of flow regulation by RoR hydropower. A better

understanding of how flow diversion affects water temperatures and trout metabolism in

bypassed reaches is particularly important given the increasing importance of RoR hydropower

worldwide. These results may also be useful for bounding the consequences of other

disturbances like anthropogenic climate change, that also change flow and temperature regimes

(Kerkhoven and Yew Gan 2011).

3.3. Methods

3.3.1. Study streams

Douglas and Fire creeks are two high-gradient streams each draining a

catchment of 104 km2 and regulated by RoR dams in southwestern British Columbia, Canada

(Figure 3.1). Their total capacity for hydroelectric production is 13.5 and 11.5 MW and their

dams are located at 346 and 359m above sea level, respectively. Both streams are in temperate

climates that experience over 1,000 mm of precipitation per year and are primarily fed by

snowmelt and glaciers. Their riparian zones are made of second-growth forested vegetation

above the dam but are impacted by logging or roads downstream of the dams, though RoR

construction activities only affected zones immediately around the dams and powerhouses.

Both streams host populations of resident O. mykiss in their bypassed reaches (measuring 3

km, 4.3 km, respectively). Mean annual discharge (2010-2013) was 6.5 m3s-1 in Douglas Creek,

and 5.2 m3s-1in Fire Creek, and the annual mean flow diverted to the turbines for hydropower

production varied between 3.8-4.7 m3s-1and 4.0-5.2 m3s-1, respectively, corresponding to 72% of

daily mean stream flow in Fire Creek, and 68% in Douglas Creek (Appendix B). Requirements

for minimum flows in bypassed reaches change seasonally, between 0.45-0.9 m3s-1in Douglas

Creek and 0.5-2.0 m3s-1in Fire Creek.

38

3.3.2. Empirical Data

We used data provided by Ecofish Research and Innergex Renewable Energy

Inc. that carried out monitoring of water temperatures and fish populations in the two creeks for

two years prior to dam construction (2007-2008), and five years after construction (2010-2014).

Time series of hourly water temperatures were generated by temperature loggers in two

locations in each stream (+/-0.1⁰C, Onset Tidbit), upstream of the planned RoR dam locations,

and immediately upstream of the planned RoR powerhouses (i.e. at the end of the future

bypassed reach, 3-4.3 km downstream of the dams). Two to three temperature loggers were

installed at each location from 2006 to 2014 to mitigate loss of data during high flow events, and

hourly temperatures were averaged across all available loggers for each location separately.

Flow was recorded hourly from 2010 to 2014 at the dam intake and in the bypassed reaches.

O. mykiss, the only fish species present, were sampled annually in September.

Abundance was estimated (2010-2014) at six locations in the first 500m upstream of the

headpond and six locations in the first 500m downstream of the dam (i.e.at the top of bypassed

reach) using multiple-pass depletion electrofishing. Electrofishing sites were chosen to be in

high quality habitat with non-embedded boulder, cobble and/or gravel substrate, and averaged

depths of 0.5m and sizes of 100m2. Weight (g) and length (cm) of each fish were recorded, and

scales were analyzed from usually n=10 individuals per suspected age class (0+, 1+, 2+, 3+) to

confirm age classification(for a total of n=144 and 171 Age 1+ and 2+ in Douglas and Fire

creeks, respectively, over all years) . As sample sizes of collected scales were too small in each

year class, year, and creek to calculate individual growth from scales, we estimated annual

growth of all Age 1+ and 2+ fish caught in the bypass reaches of the two creeks by subtracting

the mean weight of 0+ trout (or Age 1+) in prior year (2010-2014) from the individual weights at

sampling the subsequent year (we used overall mean for 2010 since no data from 2009 were

available).

3.3.3. Temperature modeling

To estimate changes in water temperature due to flow diversion for RoR

hydropower in the two creeks, we predicted the daily temperature of water in the bypassed

reaches from 2010 to 2014 under natural conditions (i.e., without flow diversion) and compared

it to the observed temperature recorded at the end of the bypassed reaches (Figure 3.2). To

estimate predicted natural (unregulated) water temperatures without flow regulation, we needed

39

to isolate changes in water temperature due to flow diversion from the warming of waters that

naturally occur as water flows down watersheds in mountainous landscapes (Vannote and

Sweeney 1980). To do so, we used the temperature data in the upstream and (future) bypassed

reaches from before construction of the RoR dams (2007-2008) to calculate a natural rate of

warming with distance downstream. We computed that natural rate of downstream warming

(∆warming) for each creek and day by subtracting the daily mean temperatures in the (future)

bypassed reaches (TBYP) from those in the upstream reach (TUS), and then averaged over both

years (Figure 3.2a):

∆𝑤𝑎𝑟𝑚𝑖𝑛𝑔= 𝑇𝐵𝑌𝑃 − 𝑇𝑈𝑆

For the five years following flow diversion (2010-2014), we then predicted unregulated

daily temperatures in the bypassed reaches (TBYP_pred) by adding the estimated natural rate of

downstream warming to daily temperatures recorded in the upstream reaches (Figure 3.2b):

𝑇𝐵𝑌𝑃_𝑝𝑟𝑒𝑑 = 𝑇𝑈𝑆 + ∆𝑤𝑎𝑟𝑚𝑖𝑛𝑔

Finally, we compared our predicted unregulated temperatures of the bypassed reaches

to the observed water temperatures recorded in the bypassed reaches from 2010 to 2014

(TBYP_obs) to calculate changes in water temperatures due to RoR flow diversion alone (∆T_DIV).

∆𝑇_𝐷𝐼𝑉 = 𝑇𝐵𝑌𝑃_𝑜𝑏𝑠 − 𝑇𝐵𝑌𝑃_𝑝𝑟𝑒𝑑

Daily changes in water temperatures due to flow diversion were averaged per month to

minimize daily variations and emphasize seasonal trends, and divided by the length of the

bypassed reaches (i.e. 3.0 km in Douglas Creek and 4.3 km in Fire Creek) to compare warming

rates per km.

3.3.4. Bioenergetics modeling

We used the unregulated (predicted) and diverted (observed) daily stream

temperatures in bioenergetics simulations to compare consumption and growth of two age

classes of rainbow trout following flow diversion to what might have occurred without flow

diversion (baseline). We modeled Age 1+ and Age 2+ fish separately due to ontogenetic

changes in mass-specific metabolic demands (Jenkins and Keeley 2010). Empirical estimates

of annual growth could not be directly related to temperatures in the bypassed reaches because

40

fish sampling occurred several kilometers upstream (i.e. immediately downstream of the RoR

dam) of where temperature was recorded in the bypassed reaches (i.e. at the end of the

bypassed reach, immediately upstream from the powerhouse). The very small number of fish

captured before flow diversion and the limited amount of trout caught on each year in each

reach made a formal before-after control-impact (BACI) analysis impossible to perform. Instead,

to perform the bioenergetics simulations, we pooled all individuals with confirmed ages (e.g.

scales) from the upstream and bypassed reaches from 2010-2014 (Age 0+, n= 41 and 30; Age

1+, n= 93 and 104; Age 2+, n= 51 and 67, in Douglas and Fire creeks, respectively) and fit a

log-normal probability distribution to each age class to create an empirically-based, simulated

size distribution for 500 randomly selected rainbow trout. For Age 1+ trout, we then matched

final weights drawn from the 500 random fish to initial weights of the same percentile from log-

normal probability distribution of Age 0+ fish to estimate annual individual fish growth. We

repeated that computation for Age 2+ fish by matching final weights drawn from the 500 random

fish to initial weights of the same percentile from the log-normal probability distribution of Age 1+

fish. Finally, we created a baseline for each age class in each creek by simulating consumption

in the unregulated (predicted) temperatures using the same sample of simulated fish weights.

We developed four food availability scenarios that represent a range of hypothetical

consequences of declines in flow and increases in temperatures on trout consumption and

growth (Appendix C). Scenario 1- Unlimited Consumption (UC) assumed that flow diversion

resulted in large increases to the availability of prey for each fish as well as the capacity for

rainbow trout to forage. Under this scenario, trout consumption was at its maximum achievable

(i.e. within physiological limits) and growth was maximum for the experienced temperatures. We

hypothesized with Scenario 2 - Constant Consumption (CC) that reduced flow and increased

water temperatures due to RoR diversion did not change food availability for trout, nor the

capacity of trout to access prey. In this scenario, we assumed that rainbow trout metabolized at

higher rates due to the higher water temperatures but ate the same amount of food per fish as

in unregulated streams, resulting in a higher proportion of energy devoted to increased

metabolic costs, and less energy allocated to growth. We hypothesized in Scenario 3 -

Increased Consumption (IC) that flow diversion increased either the availability of food for

rainbow trout, or the capacity of trout to forage for food. We parameterized Scenario 3 such that

trout experiencing warmer temperatures in the bypassed reaches ate at the same proportion of

their maximum capacity for a given size as in the natural temperatures, but that maximum

capacity scaled with warmer temperatures (i.e. with fixed pCmax). Scenario 4 - Reduced

41

Consumption (RC) represents a hypothesis where reduced flows and increased water

temperatures result in decreased annual fish consumption and growth. We imposed a 25%

reduction in annual consumption in warmer bypassed reaches, similar to empirical results

reporting that a decline in flow of 75-80% led to declines in invertebrate drift of 80-90% and

reduced caloric intake by 27% (Harvey et al., 2006). In summary, Scenario 1 represents the

upper limit of trout growth under optimal conditions, while Scenario 2 assumes that changes in

flow and temperature would not affect food consumption. Scenarios 2, 3 and 4 are all possible

depending on site-specific characteristics of different regulated streams. These four contrasting

scenarios allowed us to bound the magnitude of potential changes in annual growth and

consumption of rainbow trout induced by flow diversion from small RoR hydropower dams.

We simulated trout growth and consumption in unregulated (predicted baseline)

and diverted (observed) temperatures of bypassed reaches for three years after flow diversion,

with 2010 representing an average year, 2011 a generally cold year, and 2013 a warm year,

based on the upstream water temperatures (annual means of 7.1℃, 5.9℃ and 8.0℃,

respectively; Appendix D). We ran bioenergetics simulations in R using formulas from the

Wisconsin Fish Bioenergetics Model 4.0 (Hanson et al. 1997). We used parameters developed

for rainbow trout (Rand et al. 1993; Railsback and Rose 1999), and used a predator energy

density of 5861 J/g of wet weight and a prey energy density of 2659 J/g of wet weight (van

Poorten and Walters 2010). We used the estimated initial and final fish weights under baseline

(predicted) temperatures in each creek and year to compute the consumption and pCmax that

each age class would have achieved in absence of flow regulation. We then used these

estimates of consumption and pCmax to parameterize the simulations in the warmer

temperatures of bypassed reaches for Scenario 2 (fixed at the same value as in the baseline

unregulated temperatures) and Scenario 4 (fixed at 75% of the values in the baseline

unregulated temperatures) (Appendix C). Scenarios 1 and 3 were set by fixing the proportion of

maximum consumption (pCmax) to either 1 (Scenario 1), or to the value observed in the

baseline unregulated temperatures (Scenario 3). The scenarios thus allowed us to compare

growth, consumption and pCmax estimated under the warmer water temperatures following flow

diversion to those estimated under baseline conditions in the absence of flow diversion. All

simulations ran on a daily time step for 365 days starting on Sept 15 each year, which was the

mean date of fish sampling in Fire and Douglas Creeks. Annual growth of O.mykiss was

estimated for the baselines and scenarios for each age class, creek, and year by computing the

specific growth rate (SGR, Budy and Luecke 2014) as:

42

[ln (𝐹𝑊) − ln (𝐼𝑊) ] × 100,

where FW was the final weight of fish (in g) at the end of the 365 days of simulations and

IW was the initial weight of fish (in g). We estimated differences in SGR (in %) per fish between

each scenario and its respective baseline (i.e. predicted unregulated temperatures) by

subtracting the baseline SGR from the simulated SGR. Differences in consumption were

computed as proportional change from the respective baselines:

(𝐶𝑠𝑐𝑒𝑛𝑎𝑟𝑖𝑜−𝐶𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒)

𝐶𝑏𝑎𝑠𝑒𝑙𝑖𝑛𝑒× 100, where C is the annual consumption per fish (in g).

3.4. Results

3.4.1. Temperature changes in bypassed reaches

Mean daily water temperatures in Douglas and Fire creeks ranged from 0-15℃,

and streams were on average 0.78±0.2℃ (Douglas Creek) and 0.59±0.4℃ (Fire Creek) warmer

in the bypassed reaches than in the upstream reaches before flow diversion, but up to 3⁰C

warmer in the years following flow diversion (mean of 1.26 ±0.4℃ in Douglas Creek and

1.4±0.4℃ in Fire Creek, Figure 3.3, Appendix D). The mean natural rate of warming between

upstream and bypassed reaches was 0.74℃ in 2007 and 0.82℃ in 2008 for Douglas Creek

(0.25-0.27℃/km), and 0.55℃ in 2007 and 0.63℃ in 2008 for Fire Creek (0.13-0.15℃/km). The

mean daily flow diverted for hydroelectricity peaked between May and August and flows in the

bypassed reaches during that period varied widely (Figure 3.3, Appendix B). Flows in bypassed

reaches were often at or near the legislated minimum flows in winter, early spring, and fall

months, punctuated by periodic high flow events (Figure 3.3, Appendix B).

Using the temperature data from 2007-2014, we predicted that flow diversion by

RoR hydropower increased monthly mean water temperatures in bypassed reaches by 0.5℃ (±

0.2℃) for the length of the bypassed reach (0.17℃/km) in Douglas Creek and 0.8℃ ±0.3℃

(0.19℃/km) in Fire Creek, above natural rates of warming in each stream (Figure 3.4). Water

temperatures were almost always warmer after flow diversion than would be expected due to

the natural rate of warming alone, except in February 2014 in Douglas Creek when the monthly

predicted water temperatures were warmer by 0.2℃. We observed larger increases in water

temperatures in bypassed reaches for late spring and early summer months in all five years

(Apr – Jul mean 0.6℃ Douglas Creek, 1.1℃ Fire Creek, including a maximum of 1.4℃ in May

43

2011 in Fire Creek). Months with little flow diversion exhibited minimal differences between

natural and water temperatures of bypassed reaches (e.g. January and February 2013 in Fire

Creek).

3.4.2. Changes in trout consumption and growth in bypassed reaches

Differences in food consumption arising from the four bioenergetics scenarios led

to differences in annual growth for trout in warmer temperatures of the bypassed reaches when

compared to that realized in predicted baseline (unregulated) temperatures. For scenarios

where consumption increased (Scenarios 1 and 3), trout growth increased with warmer

temperatures in all years in both creeks for Age 1+ and Age 2+ trout (Figure 3.5). Age 1+ trout

consumed between 15-40% more in the warmer waters of the bypassed reaches under

Scenario 3-Increased C depending on the creek and the year (Figure 3.5a) and showed 15-40%

higher SGR (Figure 3.6a). Age 2+ trout exhibited similar trends to Age 1+ trout under Scenario

3-Increased C, with higher consumption (12-35%, Figure 3.5b) and annual SGR (9-36%, Figure

3.6b). When Age 1+ and Age 2+ trout could eat at will in Scenario 1-Unlimited C, consumption

(100-5000%, Figure 3.5) and annual SGR (200-450%, Figure 3.6) increased by large amounts

in warmer waters of the bypassed reaches as compared to the baseline temperatures (Table

3-1). When consumption was constant in the warmer temperatures of the bypassed reaches

under Scenario 2-Constant C, annual SGR of Age 1+ and Age 2+ trout declined slightly (-5 to -

7%, Table 3-1). When we imposed a 25% decline in consumption in Scenario 4-Reduced C, the

average reduction in annual SGR was -45% for Age 1+ and Age 2+ trout (Table 3-1).

Final weights of fish sampled in the bypassed reaches of Douglas and Fire creeks after

flow diversion averaged, respectively, 7.3g and 7.8g for Age 1+ trout and 18.0g and 20.0g for

Age 2+ trout (Table 3-2). The limited empirical data available indicate a negligible decline in

mean final weights between pre- and post-flow diversion periods in both creeks and reaches,

and for most age classes of trout (Table 3-2). Empirical-based estimates of mean annual growth

for Age 1+ trout sampled in the bypassed reaches from 2010 to 2014 (5.8 and 6.2 g/yr in

Douglas and Fire creeks respectively) fell between those predicted by Scenario 2-Constant C

(6.3 and 5.7 g/yr, Douglas and Fire creeks, respectively) and Scenario 4-Reduced C (3.8 and

3.4 g/yr, Douglas and Fire creeks, respectively) (Appendix E), while empirical estimates of SGR

were generally higher (mean of 237%) than those simulated for the three scenarios (131-192%,

Figure 3.7). For Age 2+ trout, empirical estimates of mean annual growth of fish in the top of

bypassed reaches (9.7 and 9.3 g/yr in Douglas and Fire creeks, respectively) and SGR (mean

44

of 92% in Douglas Creek and 87% in Fire Creek) were most similar to those simulated in

Scenario 2-Constant C (11.9 g/yr and 88%, and 10.5 g/yr and 92%, Douglas and Fire creeks,

respectively; Appendix E). Food consumption had the largest influence on annual growth during

late spring and early summer, coinciding with a period of rapid growth (53-63% of annual growth

occurred between June to September for Age 1+, 39-48% for Age 2+ trout) in the unregulated

temperatures of the baseline scenario. Fish growth in the warmer temperatures of the bypassed

reaches was similarly higher during June to September, for all scenarios except Scenario 4-

Reduced C, when growth was 6-11% lower. The simulated proportions of maximum

consumption (pCmax) across all scenarios varied between 0.31-0.53 for the baseline

simulations, 0.29-0.50 for Scenario 2-Constant C, and 0.26-0.45 for Scenario 4-Reduced C.

Annual patterns in fish growth among scenarios indicated very little growth over the winter and

early spring months, followed by rapid growth during May through October for all scenarios

except for Scenario 1-Unlimited C, where growth was rapid and linear throughout the year given

the unlimited consumption (Appendix F).

3.5. Discussion

We found that flow diversion by RoR hydropower in two high-gradient mountain

streams in Western Canada increased mean stream temperatures in bypassed reaches, with

potential consequences for the bioenergetics of rainbow trout. Our estimated increases in water

temperatures due to flow diversion by RoR hydropower (mean 0.5-0.8℃, max 1.4℃) were

similar in magnitude to those observed in other high-gradient mountain streams as a result of

wildfires (up to 0.6℃, Beakes et al. 2014) or clear-cut logging (0.7℃ in December, 3.2℃ in

August, Holtby 1988), and would correspond to approximately 50-80 years of anthropogenic

climate warming in streams (Isaak et al. 2016). Increases in temperatures were similar in both

creeks, but varied among months and years by a mean of 0.17±0.07℃/km. Data (unpubl) from

the reach downstream of the powerhouses show that water temperature in the two creeks were

generally colder by 0.36 ±0.21℃ for most of the year, but not for the summer months in Fire

Creek where the temperatures were slightly warmer from June to August (0.17 ±0.17℃). We do

not report further on effects of flow diversion in the downstream reaches because predicted

water temperature increases in the bypassed reaches were larger and warmer than in the

downstream reaches and we lack data on fish populations in the downstream reach to connect

changes to downstream water temperatures with fish bioenergetics. The limited number of

studies conducted in rivers regulated by RoR dams report increases in temperature in bypassed

45

reaches (Jesus et al. 2004; Anderson et al. 2006; Valero 2012), but did not isolate specific

effects of flow diversion by comparing temperatures in bypassed reaches to those before flow

diversion or in nearby non-regulated reaches (Gibeau et al. 2017), making a direct comparison

difficult. Our approach to estimating changes in water temperature due to flow diversion has the

advantage of using data (e.g. minimal longitudinal and lateral sampling) required by many

governmental permitting agencies. Despite the simplicity of the approach, our results suggest

that flow diversion for RoR hydropower can increase water temperatures in regulated reaches,

which in turn has the potential to change the bioenergetics of stream fishes.

Modelling fish growth is complex, as it is the result of multiple biotic factors and

feedbacks with environmental conditions (e.g. Crozier et al. 2010; Chittaro et al. 2014; Finstad

et al. 2011). While temperature is widely regarded as an important control on fish growth (Brett

1995), there are additional site-specific factors created by flow diversion that could affect trout

growth. In particular, flow diversion by RoR can interrupt natural processes such as sediment

transport and deposition, leading to changes in sediment size distributions downstream and

possible decreases to the productivity of salmon populations if spawning or rearing habitats are

modified or reduced (Fuller et al. 2016). Similarly, by reducing discharge in bypassed reaches,

flow diversion by RoR hydropower can reduce the area available as habitat (reduced wetted

width) and result in higher densities of resident fishes such as O. mykiss. While we did not

model higher trout densities, the combination of higher densities and elevated temperatures

could exacerbate our findings of reductions in growth when food is limited (Crozier et al. 2010),

suggesting our results may be conservative. The lack of recaptured individuals forced us to

compare cohort-based growth, which also has limitations in terms of size-selective habitat use

or mortality, as smaller individuals tend to be removed from the best habitats, and emigrate or

die at a higher rate than larger individuals (Keeley 2001). Lastly, our modelling did not consider

spatial variability in stream temperatures or diurnal variation (e.g. Coulter et al. 2016). For

example, increased variability in water temperatures over the summer was linked to a small

reduction in the growth of juvenile Atlantic salmon (Salmo salar, Imholt et al. 2011). If fish

regularly employ small-scale movements to optimize environmental temperatures, the effects of

warmer temperatures on O. mykiss metabolism in the bypassed reaches may be tempered.

However, juvenile rainbow trout are territorial (McPhail 2007) and such behaviours may limit

temperature-related movements until they experience extremely high water temperatures.

Additionally, the annual growth and consumption we estimated following flow diversion might

have differed if the increases in temperatures were closer to O.mykiss optima (e.g. Crozier et al.

46

2008; Ayllόn et al. 2019).. Similarly, our Scenario 3 Increased Consumption could overestimate

the benefits of higher consumption in terms of fish growth, as higher consumption does not

necessarily mean higher growth efficiency (Finstad et al. 2011; Allen et al. 2016). When more

detailed data are available, Net Energy Intake (NEI) and drift-foraging models, that integrate fish

swimming costs and foraging efficiency as a result of changing water velocity and depth

(Rosenfeld et al. 2013), could improve on the current approach. Additionally, data on trout diet

and foraging behavior collected before and after flow diversion would have improved the

parameterization of the bioenergetics models and our understanding of the effects of flow

diversion on fish consumption and growth. Unfortunately, such data were not required by

governmental permitting agencies or collected by proponents but their collection in future

studies would greatly help test the trends our modelling suggests. Despite these limitations, our

results align with other studies suggesting that changes in food availability and consumption are

more likely to affect growth of resident trout than the direct effects of warmer temperatures on

metabolism (Leach et al. 2012; Sweka and Hartman 2008) though we recognize that both

intrinsically combine to act as constraints on growth (Bacon et al. 2005).

Our findings suggest that the impacts of RoR hydropower on trout growth will

likely be influenced not only by temperature changes but also prey availability and consumption.

The very high and unrealistic growth rates experienced by fish in our unlimited C scenario show

the large potential for growth that fish would have in ideal conditions and with unlimited food.

Flow diversion for RoR hydropower has the potential to influence food availability for trout by

changing the amount of drifting terrestrial invertebrates or aquatic primary and secondary

production. Drifting invertebrates are the primary prey for juvenile stream-dwelling salmonids,

and terrestrial invertebrates in particular are important for trout in headwater streams (Sweka

and Hartman 2008). Feeding during warmer spring months with high food availability is

important to O.mykiss (Thompson and Hayes 2010), and terrestrial invertebrates have been

shown to make up 50-78% of trout diets between May and August (Kawaguchi and Nakano

2001). A larger proportion of stream flow is diverted for RoR hydropower during spring and

summer months (Appendix B), which may result in a decline in the availability of terrestrial

invertebrates through the reduction in stream wetted width. Our bioenergetics simulations

support this possibility, as we found a decrease of 6-12% in the annual proportion of O. mykiss

growth occurring during June to September in simulations that reduced food consumption by

25% (Scenario 4) compared to the baseline scenarios. Flow diversion for RoR hydropower may

also reduce primary and secondary aquatic production in bypassed reaches through reductions

47

in the surface area of productive habitat (Englund and Malmqvist 1996), decreases in

macroinvertebrate density (Wu et al. 2012), and increases in deposition of fine sediments (Wu

et al. 2009). However, if declines in stream wetted width are associated with reduced velocity

and depth, higher invertebrate density may lead to increased availability to trout, but empirical

data are needed to evaluate this hypothesis. Overall, empirical evidence from peer-reviewed

literature about how flow diversion affects invertebrate drift points towards negative impacts on

species richness, abundance, biomass, and density of benthic macroinvertebrates in regulated

reaches (Harvey et al. 2006; Bradford and Heinonen 2008). These empirical studies support our

Scenario 4, where trout growth was reduced by 45% when we reduced food consumption by

25%. Reduced food availability could be particularly detrimental for larger fish that may struggle

in warmer waters to keep up with their increased metabolism demands more than smaller fish

(Myrvold and Kennedy 2015a). Overall, our results align with the findings that food availability

and consumption act as key biological processes that can either moderate or amplify the effects

of warmer stream temperatures for stream fish metabolism and growth (Railsback and Rose

1999; Finstad et al. 2011).

3.6. Conclusion

Our results suggest that flow diversion by RoR hydropower, and subsequent increases

in water temperature, could have negative effects on O. mykiss by affecting their growth in

bypassed reaches if food availability does not compensate for increases in mean temperature.

Alternatively, if food availability and fish consumption increase along with temperature, and

temperature remains below O.mykiss optima, warmer waters could increase the growth

potential for fish (Myrvold and Kennedy 2018). We found that the small magnitude, but large

proportion, of water diverted for RoR hydropower increased mean temperatures by as much as

0.5-0.8℃ in bypassed reaches of two high-gradient mountain rivers. Our simulations indicate

that rainbow trout grew more in the warmer temperatures of the bypassed reaches as long as

food consumption was increased, but their growth declined in scenarios when they ate equally

or less than in baseline temperatures. The very limited empirical fish data collected from

Douglas and Fire creeks suggest that trout populations were relatively small even prior to flow

regulation, when compared to other populations in British Columbia (Scott and Crossman 1973).

This finding highlights the fact that time series of observed trout growth for multiple life-cycles

(minimum of 20 years) as well as information on diet, foraging and invertebrate communities are

needed to confirm the predictions from our bioenergetics simulations. Such time series will also

48

be important to disentangle the effects of RoR hydropower from natural variation, given the

naturally high variability in individual growth and population abundance typically observed for

salmonids (Bisson et al. 2008). Additional research could also separate changes in individual

growth trajectories from other population demographic processes that can occur simultaneously,

like changes in population density. Our study also points to the importance of better

understanding how aquatic and terrestrial invertebrates respond to flow diversion by RoR

hydropower, since they play important roles mediating the effects of warmer water on fish

metabolism and growth. Understanding how flow diversion for RoR hydropower affects water

temperature, invertebrate communities, and fish growth and metabolism is crucial to effective

management of salmonids populations in regulated rivers. The total number of trout populations

potentially affected by RoR hydropower infrastructure may be small relative to the large number

of trout populations in the Pacific Northwest, but the importance of quantifying risks and

uncertainties remains high to adequately protect salmonids populations in those regulated

rivers.

Acknowledgements

We thank M.J. Bradford, J.W. Moore, J. Leach, M. Kennedy, and A.J. Lewis for

comments that greatly improved the manuscript, K.J. Pitman for assistance with figures, and D.

Deslauriers for providing R codes. Sampling design and data collection were performed by

Ecofish Research Ltd. and data were kindly provided by Innergex Renewable Energy Inc. This

work was funded by a Natural Science and Engineering Research Council (NSERC) Industrial

Postgraduate Scholarship in association with Ecofish Research Ltd. to PG., and grants from

NSERC Discovery Grant, the Gordon and Betty Moore, and Wilburforce Foundations to WJP.

49

Table 3-1. Mean proportional change in consumption (C) and mean difference in specific growth rate (SGR) between diverted (observed) and natural baseline (predicted) water temperatures (all years combined) for Age 1+ and 2+ trout experiencing the four food availability scenarios in Douglas and Fire creeks. (Standard deviations are in bracket)

Age Scenario

Douglas Creek Fire Creek

Change in C (mean %)

Change in SGR (mean %)

Change in C (mean %)

Change in SGR (mean %)

1+

Unlimited Consumption (UC)

+1811 (±643) +333 (±39) +2028 (±696) +345 (±37)

Constant Consumption (CC)

-- -5 (±0.7) -- -7 (±0.7)

Increased Consumption (IC)

+18 (±1.5) +17 (±2) +31 (±5) +28 (±5)

Reduced Consumption (RC)

- 25 (fixed) -44 (±1) - 25 (fixed) -46 (±1)

2+

Unlimited Consumption (UC)

+1050 (±388) +276 (±40) +1129 (±290) +285 (±28)

Constant Consumption (CC)

-- -5 (±1.7) -- -7 (±0.5)

Increased Consumption (IC)

+14 (±1) +13 (±1.3) +25 (±4) +22 (±4)

Reduced Consumption (RC)

- 25 (fixed) -44 (±2) - 25 (fixed) -46 (±1)

50

Table 3-2. Average final fish weights (g) of Age 1+ and 2+ trout sampled in the bypassed and upstream reaches of Fire and Douglas creeks in the pre- (2006-2008) and post- (2010-2014) diversion periods.

Creek Reach Age Period n Weight (ave,

g)

Fire

Bypassed

0+ Pre -- --

Post 14 1.5 (1.0)

1+ Pre 6 8.5 (2.0)

Post 55 7.8 (2.8)

2+ Pre 15 21.3 (7.8)

Post 43 20 (7.1)

3+ Pre -- --

Post 11 36.5 (15.5)

4+ Pre -- --

Post 5 64.5 (14.3)

Upstream

0+ Pre -- --

Post 16 1.6 (1.1)

1+ Pre 8 10.6 (3.2)

Post 49 7.3 (3.2)

2+ Pre 14 20.4 (7.2)

Post 24 18.0 (4.8)

3+ Pre -- --

Post 4 38.1 (8.5)

4+ Pre -- --

Post 1 56.7

Douglas

Bypassed

0+ Pre -- --

Post 25 1.4 (0.6)

1+ Pre -- --

Post 69 8.6 (4.9)

2+ Pre 3 35.8 (7.8)

Post 38 22.6 (10.8)

3+ Pre 5 47.2 (8.8)

Post 17 42.1 (14.4)

4+ Pre 2 48.6 (19.5)

Post 3 59.2 (18.2)

Upstream

0+ Pre -- --

Post 16 1.6 (0.8)

1+ Pre 4 9.15 (0.3)

Post 25 8.3 (3.8)

2+ Pre 7 20.8 (4.4)

Post 13 16.7 (8.4)

3+ Pre 5 54.4 (10.3)

Post 5 78.8 (53.8)

4+ Pre 2 65.3 (7.6)

Post 1 58.3

51

Figure 3.1. Location of the two high-gradient mountain streams regulated for RoR hydropower in British Columbia, Canada. The two bypassed reaches are highlighted in darker grey between the low-head dams and the powerhouses.

52

Figure 3.2. Conceptual diagram showing the methods used to estimate the change in water temperatures due to flow diversion by RoR hydropower alone. (a) The mean change in temperatures between the upstream reach and the future bypassed reach due to the longitudinal gradient in the stream (i.e. natural rate of warming) was estimated from temperature data before construction of the RoR facilities (2007-2008). (b) The mean natural rate of warming was added to the upstream temperatures for each year following flow diversion to predict the estimated temperatures in the bypassed reach would no flow diversion occur.

53

Figure 3.3. a) Mean daily temperature (⁰C) in upstream and bypassed reaches, and b) mean daily flow (m3s-1) diverted for hydroelectricity production (plant flow, black) and recorded in bypassed reach, from January to December in Fire Creek (2010). Corresponding proportion of diverted flow is shown on the second y-axis in b. See Appendices B and D for other temperature and flow time series.

0

2

4

6

8

10

12

14 Upstream reach

Bypassed reach

Tem

pera

ture

(C

)a)

Ave.

daily

flo

w (

cm

s)

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

0

5

10

15

20

25

30

35

Month

b)

0

20

40

60

80

100

Pro

port

ion o

f div

ert

ed f

low

(%

)

Plant f low

Bypassed flow

% flow diverted

54

Figure 3.4. Difference in mean monthly temperature (⁰C) between the observed temperatures in the bypassed reaches and the natural (predicted) temperatures in a) Fire Creek and b) Douglas Creek from 2010 to 2014. Positive differences indicate that temperatures observed in the bypassed reaches following flow diversion were warmer than predicted based on the natural downstream rate of warming alone. The black line shows no difference between natural predicted and observed water temperatures.

55

Figure 3.5. Increase in simulated consumption (C, % of baseline) of trout in the warmer temperatures of the bypassed reaches as compared to the predicted natural baseline temperatures for the Unlimited Consumption (UC) scenario (left panels), and the Increased Consumption (IC) scenario (right panels) in Douglas and Fire creeks during 2010 (dark grey), 2011 (black), and 2013 (light grey) for a) Age 1 fish, and b) Age 2 fish. Years are ordered from colder (2011) to warmer (2013) temperatures. Note the y-axis in a) displays much bigger values than in b). Boxes represent 25 to 75% of the ranked data and the horizontal line shows the median. Whiskers range from the extremities of the box to 1.5 interquantile range of the data. Outliers are shown by open circles.

56

Figure 3.6. Change in specific growth rate (SGR, %) achieved by trout between the warmer temperatures of the bypassed reaches and the predicted natural baseline temperatures for the Unlimited Consumption (UC) scenario (left panels), and the Constant Consumption (CC), Increased Consumption (IC), and Reduced Consumption (RC) scenarios (right panels) in Douglas and Fire creeks during 2010 (dark grey), 2011 (black), and 2013 (light grey) for a) Age 1 fish, and b) Age 2 fish. Years are ordered from colder (2011) to warmer (2013) overall temperatures, based on upstream temperatures.

57

Figure 3.7. Frequency distributions of Age 1+ O.mykiss sampled in the bypassed reach of Douglas Creek from 2010 to 2014, a) weight (n= 105 fish, mean: 7.6g), and b) estimated specific growth rate (SGR, histogram) with predicted SGR of Age 1+ O. mykiss under reduced C, constant C, and increased C scenarios (dashed lines, right axis). SGR predicted by the three scenarios of food availability in the temperature recorded at the end of the bypassed reaches is lower than the empirical SGR observed at the beginning of the bypassed reaches.

58

Freshwater density dependence and potential consequences of flow fluctuations from Run-of-River hydropower on coho salmon (Oncorhynchus kisutch)

4.1. Abstract

Predicting whether anthropogenic sources of mortality have negative consequences at

the level of population dynamics is challenged by mechanisms like density-dependent growth

and survival that can either amplify or offset the loss of individuals from anthropogenic

disturbances. Run-of-River (RoR) hydropower is a growing industry that can cause frequent

mortality of salmonid fry through rapid reductions in flow leading to stranding on dewatered

shores. However, whether such individual-level impacts reduce population growth rates and

increase local extinction risk is difficult to predict. We evaluated how the timing and magnitude

of anthropogenic flow fluctuations influenced population abundance and extinction risk of coho

salmon (Oncorhynchus kisutch), which spend up to 1.5 years in many streams regulated by

RoR hydropower. We additionally assessed how the timing (spring, winter) and strength (weak,

moderate, high) of natural density-dependent bottlenecks experienced by salmon in freshwaters

moderates or amplifies the potential for RoR-induced mortality to scale to emergent population

dynamics. We built a stochastic stage-structured matrix model to compare population sizes and

the 45-year probability of quasi-extinction under 12 scenarios that varied the frequency (0-20

events per year) and magnitude (1-10% of mortality per event) of RoR-induced flow fluctuations,

as well as the timing and strength of density-dependent bottlenecks occurring during the first

year in freshwater. Results indicate that even mild flow fluctuations by RoR hydropower can

have some impact on coho population dynamics, especially if density dependence is weak or

occurs shortly after fry emerge in spring. Under strong winter density dependence, the potential

for population-level effects was lessened, but populations still declined by 13-42% when RoR-

induced mortality was severe (5-10%) or frequent (10-20 events/year). We conclude that strong

density-dependent survival bottlenecks could partially mitigate the loss of fry from anthropogenic

flow fluctuations, especially if bottlenecks occur late in freshwater residency. However, even

with strong density dependence in winter, our models predict declining populations by up to

59

70% under strong and very frequent flow fluctuations, which should serve to caution those

tasked with regulating flows in streams affected by RoR hydropower.

4.2. Introduction

Many anthropogenic disturbances impact species by increasing mortality of individuals,

but whether individual mortality ultimately scales up to affect species’ abundance or persistence

remains difficult to predict (e.g. Fodrie et al. 2014). Mechanisms of population regulation like

density-dependent growth and survival evolved to buffer populations against natural fluctuations

in environmental conditions and catastrophic events (Rose et al. 2001). Density dependence

may compensate for some level of individual mortality and limit population-level consequences

of disturbances (Hodgson et al. 2017). Density-dependent survival, where survival rates change

as a function of the number of individuals per unit area or volume in a population, is

compensatory when population growth rate slows at high population densities (Power 1997;

Rose et al. 2001). The presence of compensatory density dependence allows for sustainable

harvest of exploited populations, by removing a proportion of the population that would have

died during natural bottlenecks, and is central to the management of many species, from

kangaroos (McCarthy 1996) to palm trees (Freckleton et al. 2003), but especially marine fish

populations (Schaefer 1954; Beverton and Holt 1957; Ricker 1992; Hilborn and Walters 1992;

Myers 2002). Higher compensatory reserve, defined as the reproductive rate in excess of what

is needed to maintain growth and allowing a population to sustain stresses (Goodyear 1977),

leads to a larger capacity for populations to maintain productivity despite additional mortality

induced by anthropogenic removal (e.g., fishing) or disturbances (e.g. climate change, habitat

degradation).

Anadromous salmon have complex life cycles with life stages in both freshwater and

marine ecosystems that expose them to multiple anthropogenic risks, and more than half of the

Pacific salmon populations in the contiguous USA are currently threatened with extinction

(Walters et al. 2013; Gustafson et al. 2007; Crozier et al. 2019). The main causes of declines

are destruction and degradation of freshwater habitats that salmon rely on for spawning and

rearing, overfishing, and deteriorating ocean conditions due to climate change (Kareiva et al.

2000; Honea et al. 2016; Cunningham et al. 2018). Most anadromous salmon experience

density-dependent bottlenecks during their freshwater residency, though the precise timing of

60

such bottlenecks remains debated (Zabel et al. 2006; Einum et al. 2006; Walters et al. 2013).

For some populations, a freshwater survival bottleneck occurs during spring, immediately

following fry emergence, since territorial fry compete for limited territories and resources

(Chapman 1962;Ward and Slaney 1993 Nislow et al. 2004). For other populations, there is

evidence for a survival bottleneck during winter where habitats that allow parr to survive

conditions like high-flow storm events are limited (Walters et al. 2013). Similarly, for resident

salmonids evidence supports the presence of density-dependent survival bottlenecks in either

the spring (Salmo trutta, Mortensen 1977, Capra et al. 2003) or winter (Oncorhynchus mykiss,

Mitro and Zale 2002). Disturbances occurring before density-dependent bottlenecks are less

likely to affect emergent population dynamics (Greene and Beechie 2004). Thus, additional

anthropogenic mortality will be more efficiently offset if it happens before density-dependent

bottlenecks.

Run-of-river (RoR) hydropower, whereby small streams are dammed to generate

electricity without water storage, is an emerging industry that has the potential to cause mortality

for salmonids fishes (Gibeau et al. 2017). RoR hydropower has been gaining in importance

globally, with its contribution to global power generation nearly doubling from 2001 to 2010, as it

increased from 32,000 to 45,000 MW (Abbasi and Abbasi 2011). RoR hydropower projects

typically have small physical footprints with low-head dams that temporarily divert water through

tunnels into powerhouses, where water is run through turbines to create electricity before being

returning to the river (Anderson et al. 2014). RoR hydropower may generate a range of impacts

to stream ecosystems (Gibeau et al. 2017), including anthropogenic flow fluctuations (or pulsed

flows, Young et al. 2011) that are often created by the operation of RoR power plants. Flow

fluctuations occur when changes in electricity production lead to the sudden rerouting of water

from the dam into the bypassed river reach. This creates rapid increases in flow in bypassed

reaches, and a concurrent, sometimes sharp, decline in flow in reaches just downstream of the

powerhouse. The increase in flow in the bypassed reach has the potential to displace juvenile

fish, increase energy costs, or scour spawning redds (Harby and Halleraker 2001; Nislow and

Armstrong 2012). Simultaneously, the decline in flow in downstream reaches may strand fish on

dewatered river beds or isolate fish in disconnected side pools (Nagrodski et al. 2012). Fry are

the most vulnerable life stages because they are poor swimmers and take refuge along shallow

stream margins that dewater rapidly (Hunter 1992; Halleraker et al. 2003; Young et al. 2011).

Coho salmon (Oncorhynchus kisutch) are expected to be particularly vulnerable to operations

from RoR hydropower due to their long residence time in freshwater and high spatial overlap

61

with many RoR facilities in the Pacific Northwest. Coho fry are more vulnerable to stranding

than rainbow trout (Oncorhynchus mykiss) fry or other anadromous salmon species following

anthropogenic flow fluctuations from studies in both large reservoirs and laboratory

experiments, because of their tendency to bury deep into the substrate during the day, which

made them less likely to detect declines in water levels and avoid stranding (Hunter 1992;

Bradford et al. 1995). While many studies noted lethal and sub-lethal impacts on individuals in

regulated rivers or downstream of reservoirs (Young et al. 2011; Nagrodski et al. 2012; Dibble et

al. 2015), quantitative explorations of population-level effects are few (Sauterleute et al. 2016),

especially in smaller streams regulated by RoR dams.

Here we evaluated the potential for anthropogenic flow fluctuations induced by RoR

hydropower to impact population dynamics of coho salmon using simulation models integrating

density dependence. Specifically, we explored the degree to which the strength and timing of

density-dependent survival bottlenecks experienced by salmon during the first year of

freshwater residency could mitigate the additional mortality experienced by fry from

anthropogenic flow fluctuations. We built a stage-structured stochastic matrix model,

parameterized with data from the literature, to simulate potential impacts of RoR dams, across a

range of flow fluctuations, on the probability of quasi-extinction and population size of coho

salmon under weak, moderate, or strong density dependence assumptions. Understanding

better how natural density dependence can compensate for mortality induced by human actions

is crucial to guide effective management strategies for salmonid populations inhabiting

regulated rivers.

4.3. Methods

4.3.1. Matrix model

We built a stochastic stage-structured matrix model (Morris and Doak 2002) describing

the 3-year coho salmon life-cycle (Nickelson and Lawson 1998; Bradford et al. 2000). Our

model included density-dependent survival in freshwater and the effects of simulated

anthropogenic flow fluctuations of different frequency and magnitude on population abundance

and dynamics. All analyses were performed in the R language (version 3.5.1, R Core Team

2013). Adult coho spawn in natal freshwater streams in late fall, eggs incubate over winter, and

fry hatch and emerge in the spring. Juveniles (termed parr) typically remain in freshwaters until

the following spring, before migrating to the Pacific Ocean as smolts (at 1.5 years old), where

62

individuals spend up to 18 months before returning to spawn (Groot et al. 1995). In some

populations, juveniles may spend an extra year in freshwaters before migrating to the ocean

(Holtby and Healey 1990), or return early as to spawn as two years-old (Bennett et al. 2015),

but those are considered exceptional. We modelled a typical coho population including three life

stages (Eggs/fry/parr, parr/smolts, adults) in a stage-structured stochastic matrix model with an

annual time step, beginning with spawning in late fall (Table 4-1). As a result, the first life stage

(F13), comprised both egg and fry transitions until the end of the first fall. We assumed a mean

sex-ratio, Pfem, of 0.452 (σ ± 0.009) female, to reflect the recorded higher mortality of adult

females than males in the ocean (mean of two creeks, Holtby and Healey 1990; Spidle et al.

1998), and mean fecundity, Neggs, of 2597 eggs (σ ± 386) per female (mean of 16 creeks,

Shapovalov and Taft 1954; Crone and Bond 1976). The survival to emergence, φemerg, was

estimated at 0.223 (σ ± 0.032) based on studies in five creeks of the Pacific Northwest (Bradford

1995; Hall 2008). The second life stage (a21) included parr to smolt and encompassed survival

from the end of the first fall, through leaving freshwaters as smolts, and for the first six months

of young adults in the ocean. Finally, the last stage (a32) comprised three-year-old adults in the

ocean until spawning in freshwater. The geometric mean of recent data from three wild coho

populations from the Pacific Northwest suggests an overall ocean survival of 6.04% (σ ±2.27%)

prior to harvesting (Korman and Tompkins 2014; Korman et al. 2019), which reflects the high

ocean mortality suffered by coho populations in recent years. When generating stochastic

annual estimates of ocean mortality, we included a temporal autocorrelation of 0.6 from the

previous year to reflect the long-term trends observed in ocean survival due to large-scale

climatic events like El Niño cycles or the Pacific Decadal Oscillations (Bradford et al. 2005). The

ocean survival rate was then split for each brood year between the six months of ocean survival

in Year 2 (φoceY2) and 12 months in the ocean of Year 3 (φoceY3). We added stochasticity to Pfem,

φemerg, and ocean survival by randomly drawing each vital rate at each time step from a beta

distribution parameterized with the respective mean and standard deviations (Morris and Doak

2002). Stochasticity was added to Neggs using a normal distribution N(μ, σ), with the mean and

standard deviation described above.

4.3.2. Freshwater density-dependent survival

We estimated the transition probability from fry to smolt (f(φspr) or f(φwin)) using a density-

dependent survival function derived from data on 10 coho populations in creeks similar in size to

those regulated by RoR-dams in the Pacific Northwest (i.e. mean length = 9 km, mean

63

drainage area = 32 km2), where out-migrating smolts and returning adults were counted over

several years (Bradford 1995; Korman and Tompkins 2014). To estimate the number of fry

produced in each of the 10 creeks, we multiplied the number of returning adults (per km of

stream to account for the varying length of streams) by the mean values of sex ratio, fecundity,

and survival to fry emergence, as reported above. We regressed the density of fry (Dfry) against

the density of smolts (Dsmolt) to estimate the strength of density-dependent survival for each of

the 10 streams using the Beverton-Holt relationship:

Dsmolt(t+1) = α × Dfry(𝑡)

1+[(α

k )×Dfry(𝑡)]

(eq. 1)

Where α is the number of smolts per fry at the origin and k is carrying capacity for smolts

in the stream (smolts per km).

We limited the parameter α to a maximum of 1 to avoid unrealistic overcompensation at

low fry abundance, given that the density-dependent function only describes survival between

fry to smolt stages. We categorized the strength of density dependence as weak (α= 0.2),

moderate (α= 0.5), and strong (α= 1.0) based on the relations parameterized for each 10 creeks

while holding k constant for all streams (Figure 4.1). Weak and strong density dependence

corresponded to the lowest (0.2) and highest (1.0) values of α observed among the 10 creeks,

respectively, and we set moderate density dependence to the average α of all 10 creeks (0.5).

Carrying capacity for smolts (k) was set to 15,318 individuals, which corresponded to the mean

k over all 10 creeks (1702 smolts/km), scaled to simulate populations inhabiting a hypothetical

stream with an average 9 km spawning and rearing reach.

We developed the model to compare both how the strength of density dependence

(weak, moderate, or strong) and the timing of the bottleneck (spring or winter) impacted the

capacity of populations to compensate for anthropogenic mortality. We varied timing of the

density-dependent bottleneck by either attributing all density-dependent survival to Spring or

Winter and assumed the remainder of freshwater survival to be density-independent similar to

other studies (Zabel et al. 2006). Thus, fry-to-smolt survival was partitioned into two transition

rates (F13, a21) given that the life-cycle model was based on yearly time steps. We added

variation in smolt survival (smoltE) following the density-dependent bottleneck by adding a

truncated normally-distributed error term, N(0, σ), that allowed smolt survival to vary within the

range of residuals standardized per k for each creek (σ= 0.074) from the density-dependent

function (i.e. Beverton-Holt relationships for each 10 creeks). The normal distribution was

64

truncated between -0.27 and 0.27, which corresponded to the mean ratio between the residuals

from the 10 fits of the Beverton-Holt relationship standardized by each creek’s respective

carrying capacity.

4.3.3. Scenarios of RoR-induced flow fluctuations

Flow fluctuations created by operations of RoR facilities vary in frequency, magnitude,

and timing throughout the year, but limited data exist in the peer reviewed literature to

quantitatively describe their characteristics (Gibeau et al. 2017). As a result, we developed

scenarios by varying the timing and frequency of anthropogenic flow fluctuations using limited

data from a selection of monitoring reports submitted to government agencies regulating the

RoR industry in British Columbia (BC), Canada (Appendix G). The definition of flow fluctuation

events followed slightly different definitions in each creek, but generally, an event was defined

by river water levels dropping faster than regulated site-specific minimum (generally vertical

height of 2.5 cm per hr) for periods longer than 10 minutes, after at least 24-hours of stable

water levels. Given these criteria, the events reported by plant managers are likely

underestimates of all anthropogenic flow fluctuations created by flow regulation. We found that

between 0 and 21 events occurred per creek every year (overall mean of 9 ± 5 events per year,

Figure 4.2). Most events occurred in the fall (mean of 37%), followed by summer (mean of

29%), spring (mean of 20%), and winter (mean of 14%; Figure 4.2).

There are very few published estimates linking records of stranded fish to population-

level estimates of fry mortality in streams regulated by RoR dams or larger reservoir-storage

systems, and most studies were in systems subjected to daily flow fluctuations. For example,

Almodόvar and Nicola (1999) attributed rapid declines in density (-50%) of brown trout over one

year to drastic daily flow fluctuations (at times, water depth changed by 0.3m over a few

minutes) downstream of a small RoR dam that lead to the loss of suitable habitat. Other studies

below larger reservoir-storage dams point to the potential for similarly high mortality of juveniles

after flow events. Bauersfeld (1978) estimated 1.5% of chinook (Oncorhynchus tshawytscha)

and coho salmon fry mortality per flow fluctuation event downstream of a large dam and

concluded that a total loss of 59% of the salmon fry population could have occurred over a

single season. Similarly, a study in a reservoir-storage system with large daily fluctuations in

flow reported an 80% decline in juvenile Atlantic salmon densities of over four years (Ugedal et

al. 2008), and stranding mortality rates between 5-90% were used for a study with Atlantic

65

salmon in Norway (Sauterleute et al. 2016). Thus, based on these few studies, we estimate that

a single flow fluctuation event could lead to 1 – 10% mortality of fry.

Given the uncertainty in the estimates of fry mortality per event, we created 12 scenarios

with varying frequency and magnitude of RoR-induced flow fluctuation to encompass a wide

range of possible consequences for coho populations (Table 4-2). The 12 flow fluctuation

scenarios were combinations of four different annual frequencies, low (n= 1-5 events), mid (n=

6-10 events), high (n= 11-15 events) and very high (n= 16-20 events), paired with three

magnitudes of fry mortality per event, high (mean of 10%), mid (mean of 5%) and low (mean of

1%). Within each frequency category (low, mid, high, very high), we randomly drew the number

of events from a uniform distribution for each year, and distributed the events seasonally using

the proportions calculated from observed data in BC (i.e. with probabilities that of 20%, 29%,

37%, and 14% of events would occur in the Spring, Summer, Fall, and Winter, respectively,

Figure 4.2). The survival of juveniles during RoR-induced flow fluctuations was randomly drawn

from a beta distribution using a mean survival of 99 ±1% for the low, 95±1% for the mid, and

90±1% for the high magnitude scenarios. We classified the 12 scenarios into three groups (low,

mid, high impacts) based on the minimum survival possible if frequency of flow fluctuations was

maximum and magnitude average on each year (e.g. minimum annual survival would be 0.995

=0.95 for the low frequency-low magnitude LL scenario). Low impact scenarios included those

that had annual survival of fry or parr to anthropogenic flow fluctuations above 75%, while

medium impact scenarios had annual survival between 45 to 60%, and high impact scenarios

had annual survival below 36% (including survival as low as 12% for the worst-case scenario of

very high frequency and high magnitude of flow fluctuations) (Table 4-2). Given the lack of

precise population-level mortality rates, our scenarios thus model a wide range of mortality that

span from small impacts of minor and rare fluctuations (L-L scenario), to worst-case scenarios

with very frequent and damaging flow fluctuations still within the range of the extremes reported

in the literature cited above (VH-H scenario).

We compared populations simulated under the 12 scenarios of RoR-induced flow

fluctuations to baseline populations that were parameterized in the same way (i.e. with the same

strength and timing of density dependence) but not subjected to the influence of additional

mortality from RoR hydropower. We included an initial burn-in period of 10 years without

stochasticity to allow simulated populations to reach equilibrium, and then ran each simulation

for 45 additional years (~15 generations, of 3-year cohorts). We started all simulations with an

initial population size of 500 adults over a hypothetical 9 km river and derived the starting

66

number of fry and smolts by using the stable age distribution of the deterministic model without

density dependence. We repeated each combination of strength and timing of density

dependence for 10,000 iterations, for a total of 75 combinations (12 scenarios x 3 strengths x 2

timings of density dependence, and 1 baseline model x 3 density dependence strengths). Each

scenario was then compared to the baseline with the same strength of density dependence

(which were identical between spring and winter bottlenecks, given all mortality from fry to smolt

was attributed either in the spring or the winter). We compared baseline and RoR-impacted

populations through mean final population size and the 45-year probability of quasi-extinction.

To capture the variation from cohort to cohort, the mean final population size was the average

number of adults over the last six cohorts (years 40 to 45) of the simulations. The probability of

quasi-extinction was calculated as the number of times that the number of adults fell below a

threshold of 92 individuals for three consecutive cohorts over the 10,000 iterations. This

threshold corresponded to the number of adults that smolts at 10% carrying capacity would

produce assuming an average ocean survival of 6.04% (similar to Crozier et al. (2008)).

4.3.4. Elasticity analyses

Following Wisdom et al. (2000), we performed a simulation-based elasticity analysis by

simulating 10,000 matrices with vital rates drawn at random from uniform distributions bounded

by the 95% confidence interval for each parameter, and calculated the deterministic growth rate

(λ) without density dependence. In this simplified deterministic model, the density-dependent fry

survival (ƒ(φspr) or ƒ(φwin)) was replaced by deterministic fry survival in year 1 (φfryY1) and year 2

(φfryY2), that averaged 7.5%, the mean fry-to-smolt survival across the 10 creeks used to derive

the Beverton-Holt density-dependent function. We decomposed variation in lambda with a

general linear model with Gaussian distribution, to predict the contribution of each vital rate to

the variation observed in lambda (given by the standardized regression coefficients associated

to each vital rate in the model). By comparing the standardized regression coefficients, we infer

the contribution of each vital rate to a proportional change in lambda (Wisdom et al. 2000).

4.4. Results

Overall, under all scenarios of flow fluctuations from ROR hydropower, the probability of

quasi-extinction increased, and population sizes declined over 45 years as compared to the

respective baselines, for all strengths and timings of density dependence. The magnitude of the

67

response at the population level was more pronounced as the pressure from anthropogenic flow

fluctuations increased (either by more frequent and/or more severe fluctuations). Density-

dependent survival bottlenecks compensated for some of the increased mortality from

anthropogenic flow fluctuations, especially when density dependence was strong and/or

occurred in the winter. We detail below the results from all scenarios.

4.4.1. Probability of quasi-extinction

Quasi-extinction probabilities in the absence of RoR-induced flow fluctuations were

highest at 0.19 for the baseline population when density dependence was weak but declined to

0.13 and 0.11 when density dependence was modeled as moderate or strong (Figure 4.3).

When we modeled survival from RoR-induced flow fluctuations under weak density

dependence, the probability of quasi-extinction increased especially for medium impact (means

of 0.5 for spring bottleneck and 0.38 for winter bottleneck) and high impact scenarios (means of

0.96 and 0.86 for spring and winter bottlenecks, respectively) (Figure 4.3). The probability of

quasi-extinction also increased compared to baseline when mortality from RoR-induced flow

fluctuations was included with either moderate or strong density dependence, but primarily

when the bottleneck occurred in spring or when the magnitude or frequency of mortality events

was high (Figure 4.3). For example, the probability of quasi-extinction increased by a factor of 5

in high impact scenarios when density dependence was moderate (mean of 0.68) or strong

(mean of 0.54) in the spring, and by a factor of 2-3 when density dependence was moderate

(mean of 0.43) or strong (mean of 0.22) during winter. Under strong density dependence in the

winter, the probability of quasi-extinction remained at near baseline levels (~0.2) for all

scenarios, except for the most extreme scenario (very high frequency/high magnitude VH-H

scenario) that had a probability of quasi-extinction of 0.36 (Figure 4.3). Overall, probability of

quasi-extinction declined by 14-18% when the bottleneck was in the winter as opposed to spring

under weak density dependence across all scenarios, and for moderate and strong density

dependence in the low and medium impact scenarios. However, as the pressure from

anthropogenic flow fluctuations increased in the high impact scenarios, the probability of quasi-

extinction declined by an average of 57% when the bottleneck was in the winter as opposed to

spring, and by 41% under moderate density dependence. Probability of quasi-extinction

increased steadily over time for all strengths and timings of density dependence under the

medium impact scenarios (Figure 4.4a). However, when flow fluctuations were frequent and of

high magnitude, high probabilities of quasi-extinction were reached more quickly over time for

68

the weak density dependence in Spring and Winter, and when density dependence was

moderate in the Spring (Figure 4.4b).

4.4.2. Population size

Flow fluctuations from RoR hydropower reduced modeled population sizes across all

scenarios including the full range of density dependence assumptions, but the largest declines

occurred when either density dependence was weak (mean of 58% decline from baseline

across all scenarios and both timings of bottleneck), when the natural density-dependent

bottleneck was in the spring (mean of 53% decline from baseline across all scenarios and all

strengths of bottleneck), or when RoR-induced mortality was high (mean of 78% decline from

baseline for high impact scenarios across all strengths and timings of bottleneck) (Figure 4.5).

For example, we observed a 90% population decline in the high impact scenarios in the spring

compared to the baseline. Mortality from flow fluctuations had a particularly large influence on

the population size when density dependence was weak, as final number of spawners declined

from the baseline by an average of 68% for medium-, and 99% for high-impact scenarios with a

spring bottleneck, and 49% and 94% declines, respectively, for the weak density dependence

with a winter bottleneck. However, population sizes were consistently higher for all scenarios

when density dependence was strong as compared to weak or moderate strengths (Figure 4.5).

When we modeled high survival from anthropogenic flow fluctuations (survival per event ~0.99),

the timing of density dependence did not influence population size regardless of our

assumptions about the strength of density dependence (e.g. low impact L-L scenario, Figure

4.6.a). However, when we increased the frequency and magnitude of flow fluctuations (e.g.

medium and high impact scenarios), the decline in population size with a winter bottleneck was

less severe than with a spring bottleneck when density dependence was moderate or strong

(Figure 4.6.b and c). However, even under strong density dependence in the winter, population

sizes still declined from baseline abundances by 13% to 42%, for medium and high impact

scenarios, respectively. Comparatively, the average decline in population size from the baseline

was 46% for medium impact scenarios and 88% for high impact scenarios for moderate density

dependence in the spring, and 20% (medium impact) and 63% (high impact) declines, for

moderate density dependence in the winter.

69

4.4.3. Elasticity analyses

The elasticity analysis showed that not all rates contributed equally to variation in

deterministic population growth rates (λ). Vital rates with the highest simulated elasticities in our

model included the survival of fry to smolt and ocean survival (standardized regression

coefficients: 0.15, 0.16, respectively), suggesting they have the largest influence on λ (Figure

4.7). The next most important vital rates were fecundity (Feggs) and the survival of fry at

emergence (standardized regression coefficients: 0.051, 0.049, respectively), followed by the

proportion of females (Pfem) that showed very little influence on population growth

(standardized regression coefficient: 0.0007).

4.5. Discussion

Our results demonstrate that mortality induced by flow fluctuations from the operation of

RoR hydropower dams have the potential to affect emergent population dynamics of coho

salmon populations. Both the strength and timing of density-dependent survival during the first

year in freshwaters were critical to predicting the impacts of anthropogenic flow fluctuations for

population dynamics of coho. When density dependence was weak in the spring or the winter,

even rare and low mortality flow fluctuations substantially increased the probability of quasi-

extinction and decreased population sizes 45 years later. However, populations experiencing

moderate to strong density dependence had more capacity to attenuate the loss of fry from

RoR-induced flow fluctuations, especially if density dependence occurred in the winter. Even

when we modeled strong density dependence occurring in the winter, the final population size

for the most extreme scenario of RoR-induced mortality was 70% smaller than baseline

populations. In turn, smaller populations would have higher vulnerability to other anthropogenic

disturbances, as well as demographic stochasticity and extinction (Lande et al. 2003). The

limited data we found in the grey literature about RoR-induced flow fluctuations (Figure 4.2)

suggests that most rivers regulated by RoR hydropower in the Pacific Northwest experience

annual frequencies of flow fluctuation events that fall into our low to mid-impact scenarios

(overall mean of 9 ± 5 events per year). These results suggest that if populations experience

moderate to strong density dependence, at least some of the mortality induced by

anthropogenic flow fluctuations can be absorbed by the life history, especially if density

dependence operates primarily in the winter. However, these empirical data on RoR flow

fluctuations also illustrate that high to very high frequency of events is possible (maximum

70

frequency of 21 events experienced in a year), and our modelling suggests that such scenarios

could lead to smaller population sizes and elevated extinction risk for coho salmon, regardless

of the form of density dependence. Our results suggest that, if survival of individuals is very low

during each RoR-induced flow fluctuation event, such mortality would overwhelm even the

strongest and most favourably-timed density-dependent bottleneck, a result aligned with

findings from a study looking at impacts of climate change on Chinook salmon density-

dependent survival in freshwater (Crozier et al. 2008a).

Beyond fry mortality induced by anthropogenic flow fluctuations, there are several

additional mechanisms or limitations that could further impact salmon in rivers regulated by RoR

hydropower that make our estimates conservative. For example, RoR-induced flow fluctuations

could influence life stages other than fry and parr in ways not included in our simulations thus

increasing negative pressures on populations (e.g. stranding of smolts, Sauterleute et al. 2016).

Other impacts of anthropogenic flow fluctuations include nest abandonment, redd scouring,

home range reductions, and loss of habitat connectivity downstream of reservoirs (Geist et al.

2008; Stradmeyer et al. 2008; Korman and Campana 2009). Though our models assumed the

same rates of survival over time, timing of flow fluctuations within diurnal or seasonal cycles

may also influence the severity of consequences for fish populations. For example, simulations

performed by Sauterleute et al. (2016) showed that population abundance of Atlantic salmon

(Salmo salar) was most reduced downstream of a hydropeaking reservoir in Norway when flow

fluctuations occurred during winter daylight hours. Flow fluctuations in winter or spring may have

particularly severe consequences for fish populations as negative effects of high water flows in

March was linked to poor recruitment for fall-spawning brown trout (Sabaton et al. 2008;

Gouraud et al. 2008; Capra et al. 2003) and high flows are known to cause of egg mortality

(Ward and Slaney 1993; Battin et al. 2007).

Other limitations include our choice of the Beverton-Holt curve to describe the shape of

the density-dependent relationship between fry and smolt. Though the Beverton-Holt convex

curve is commonly used to quantify freshwater density dependence for anadromous salmonids

(Bradford et al. 2000; Zabel et al. 2006; Crozier et al. 2008b; Thorson et al. 2014), a linear or

concave shape to the density-dependent function may have further mitigated or magnified the

population-level consequences of anthropogenic stressors (Ratner et al. 1997). The intensity of

the mortality from a flow fluctuation event may also be density-dependent, as more fry may be

pushed to shallower waters at high densities thus increasingly their vulnerability to stranding.

Our models did not include other forms of density dependence beyond survival, like density-

71

dependent growth, which often co-occurs with density-dependent survival (Myrvold and

Kennedy 2015b). Declines in population size following density-dependent mortality may be

moderated by density-dependent growth in the remaining individuals (Keeley 2001), ultimately

leading to higher survival of the bigger individuals (Hurst 2007). For example, higher parr sizes

due to warmer stream temperatures after intensive logging were linked to higher overwinter

survival in Carnation Creek, Canada, though interestingly, higher parr abundances did not lead

to higher adult abundances or other population-level impacts (Tschaplinski and Pike 2017). Our

models also depend largely on the choices we made about the level of fry mortality to attribute

to each flow fluctuation event. We simulated a wide range of mortality in the absence of robust,

reliable data relating observed mortality from flow fluctuation events to population density

estimates. An important next step to clarify our modelling results and conclusions would thus be

to collect such data through empirical research and derive field-based mortality rates. Despite

having to make simplifying assumptions about these systems, our results support other findings

highlighting that understanding both the strength and forms of density dependence in wild

populations is crucial to estimating and mitigating anthropogenic impacts (Sharma et al. 2005;

Hodgson et al. 2017).

4.5.1. Implications for conservation

Our results highlight that anthropogenic disturbances will have more severe

consequences for populations experiencing weak density dependence, especially if the density-

dependent bottleneck occurs before most of the human-imposed mortality (Hodgson et al.

2017). For example, populations of resident trout, with typically lower fecundity and weaker

density-dependent bottlenecks than anadromous salmonids, may suffer more from the mortality

incurred from anthropogenic disturbances (Walters et al. 2013b). The presence, strength and

timing of density dependence are also likely to be largely site-specific, suggesting different

populations of same species may experience very different types and strengths of density-

dependent bottlenecks (Wainwright and Waples 1998; Rose et al. 2001). This spatial variation

in resilience to anthropogenic disturbances may increase the odds of meta-population survival

by allowing strongly regulated populations to re-colonize vulnerable populations (Nickelson and

Lawson 1998). The timing and strength of density dependence, as well as which life stages are

affected, also have important implications for habitat restoration (Greene and Beechie 2004),

which appears to be a management action crucial to recover and maintain coho populations in

heavily-altered systems (Nickelson and Lawson 1998; Oosterhout et al. 2005). For example, our

72

results support that increasing carrying capacity in freshwater, and particularly restoring

overwintering habitats (e.g. Solazzi et al. 2000), may be particularly important to increase

compensatory reserves for coho populations in streams regulated by RoR hydropower. Such

actions would lead to increased mean abundance and help prevent populations from falling

below the quasi-extinction threshold, similarly to what was observed for Chinook salmon in the

Snake River Basin (Zabel et al. 2006).

Our results highlight the importance of assessing anthropogenic impacts cumulatively to

properly understand the dynamics of populations and threats to species’ survival (Hodgson et

al. 2017). The relatively high probability of quasi-extinction we calculated for our baseline

populations (18%), especially among populations with weak density dependence, integrate the

background cumulative impacts from diffuse anthropogenic sources currently experienced by

coho populations in the Pacific Northwest, and is similar to that experienced by Chinook salmon

(estimated extinction risk of 0.309 for the baseline population, Zabel et al. 2006). Increased

variability in extreme flow events and river hydrology following climate change may further

jeopardize survival of coho salmon in freshwaters (Ohlberger et al. 2018). Moreover, lower

ocean survival due to climate change could further exacerbate the capacity of coho populations

to withstand additional anthropogenic disturbances, like RoR-induced flow fluctuations, as seen

with water diversion and climate change for the Chinook salmon in the Lemhi River Basin in the

USA (Walters et al. 2013a).

4.6. Conclusion

Our results show that the mortality induced by flow fluctuations from operations of RoR

hydropower can affect emergent population dynamics of coho salmon populations, particularly if

natural density dependence is weak or if the flow fluctuations are frequent and strong

overlapping with a natural density-dependent survival bottleneck in the spring. The crucial

importance of density dependence for partially offsetting the negative consequences of

anthropogenic disturbances highlights the need to quantify the presence, timing, and strength of

density-dependent survival bottlenecks for salmonids during their freshwater life residency (e.g.

Barnthouse et al. 1984). Our study also highlights the need to empirically quantify the mortality

rates leading to population-level impacts of anthropogenic flow fluctuations, which our models

suggest may be important for salmonids residing in streams regulated by RoR hydropower. This

task is particularly difficult given the considerable inter-annual variation in abundance seen for

most salmonids populations (e.g. Beamish et al. 1999; Jenkins et al. 1999; Bradford et al.

73

2000). The mitigation potential of density dependence and the high inter-annual variation in

population abundance further supports the need for long-term empirical datasets on regulated

rivers (Rose et al. 2001; Sabaton et al. 2008; Lobon-Cervia 2009). Such long-term data sets are

generally lacking, as noted by a comprehensive study in of RoR facilities in BC, where authors

concluded that the lack of monitoring data made it impossible to either confirm or quantify the

risks from flow fluctuations by RoR hydropower (Connors et al. 2014). Nonetheless, the fact that

even mild anthropogenic flow fluctuations can influence population dynamics of coho salmon

should be taken seriously for the sound management of flow fluctuations on regulated rivers and

conservation of threatened salmon populations.

Acknowledgements

We thank E. E Hodgson, K. Wilson, D.A Greenberg, A. Cantin, J.W. Moore, and M. J.

Bradford for help with conceptualizing the analysis and for comments that greatly improved the

manuscript. This work was funded by grants from NSERC Discovery Grant, the Gordon and

Betty Moore, and Wilburforce Foundations to WJP.

74

Table 4-1. Matrix model structure and description of vital rates for spring and winter density-dependent mortality bottlenecks.

Time t+1 Time t

Eggs/Fry Parr/Smolt Adults

Eggs/Fry 0 0 F13

Parr/Smolt a21 0 0

Adults 0 a32 0

Transition rate aij Parameter equations

Spring

Fecundity (Eggs to fry)

F13 Pfem * Feggs * φem * φRoR_sp * (ƒ(φspr) +-smoltE) * φRoR_sum * φRoR_fall

Parr to smolt a21 φRoR_wi * φoceY2

Smolt to adult a32 φoceY3

Winter

Fecundity (Eggs to fry)

F13 Pfem * Feggs * φem * φRoR_sp * φRoR_sum * φRoR_fall

Parr to smolt a21 (ƒ(φwin) +-smoltE1) * φRoR_wi * φoceY2

Smolt to adult a32 φoceY3

Pfem= proportion of females; Feggs= # of eggs per female; φem= survival from hatching to emergence; f(φf_spr)= density-dependent survival of fry through spring bottleneck; f(φf_win)= density-dependent survival of fry through winter bottleneck; φoceY2= survival of smolt through 1st summer and fall in ocean; φoceY3= survival of adults in ocean (Year 3); φRoR_sp, φRoR_sum, φRoR_fall, φRoR_wi= survival of fry and parr after RoR-induced flow fluctuations

1The parameter smoltE was offset by a year for baselines and scenarios with winter bottlenecks, so that winter and Spring baselines and scenarios could be directly comparable.

75

Table 4-2. The 12 scenarios used in simulations with varying frequency and magnitude of RoR-induced flow fluctuations. White indicate low impact scenarios (annual survival >75%), grey indicate medium impact scenarios (annual survival between 45-60%), and dark grey indicate high impact scenarios (annual survival <36%).

Frequency (# events per year)

Magnitude (mortality per event)

Low Mid High

Low L-L L-M L-H

Mid M-L M-M M-H

High H-L H-M H-H

Very high VH-L VH-M VH-H

76

Figure 4.1 Number of smolts produced under weak (α= 0.2), moderate (α= 0.5), and strong (α= 1) density dependence as derived with Beverton-Holt equations using data from 10 creeks (in light grey) in the US and Canadian Pacific Northwest. Carrying capacity (k) was fixed at 15,318 smolts over nine km of river.

77

Figure 4.2. Mean frequency of flow fluctuations events per year (± SD for the creeks with multiple years of data, shown with error bars, top) and mean proportion of flow fluctuation events per season (±SD shown by error bars, bottom) induced by operations of RoR hydropower for eight regulated creeks in British Columbia, Canada. Creek names in the top panel are ordered by increasing watershed area (47-313 km2), and characteristics of the creeks and RoR dams are listed in Appendix G.

78

Figure 4.3. Probability of quasi-extinction after 45 years (z-axis) for populations simulated under 12 combinations of RoR-induced flow fluctuations of differing frequency (y-axis) and magnitude (x-axis). Rows show scenarios experiencing density dependence during Spring (a) or Winter (b), and columns indicate the strength of density dependence, ranging from Weak (left), Moderate (middle), and Strong (right). Probability of quasi-extinction for the baseline populations with weak, moderate or strong density dependence were 0.19, 0.13, and 0.11, respectively.

79

Figure 4.4. Probability of quasi-extinction (%) over time for a typical coho population subjected to a) medium impact scenario M-M, and b) high impact scenario H-H of RoR-induced flow fluctuations, under the three strengths (weak, moderate, and strong) and two timings (Spring, Winter) of density dependence.

80

Figure 4.5. Mean final population size (mean number of adults for the last six cohorts, years 40-45) for populations simulated under the 12 scenarios of RoR-induced flow fluctuations of differing frequency (y-axis) and magnitude (x-axis). Rows show scenarios experiencing density dependence during Spring (a) or Winter (b), and columns indicate the strength of density dependence, ranging from Weak (left), Moderate (middle), and Strong (right). Average final size for the baseline populations with weak, moderate or strong compensatory reserves were 359, 531, and 591 adults, respectively (thick black lines). Note that order of the x and y axis is different from that of Fig. 3.3.

81

Figure 4.6. Final population size (mean number ±SD of adults for years 40-45) for populations experiencing spring (black circles) and winter (white circles) bottlenecks simulated under weak, moderate, and high density dependence, for a) low impact L-L, b) medium impact M-M, and c) high impact H-H scenarios of RoR-induced flow fluctuations. Average final sizes for the baseline populations with weak, moderate or strong density dependence are represented with * (n=359, 531, and 591 adults, respectively).

82

Figure 4.7. Simulation-based elasticity values for lower level vital rates (R2adj= 0.982, p < 2.2 e-16). Fry survival in year 2 (φfryY2) and adult survival in the ocean on

year 3 (φoceY3) are not presented because they are fully collinear with fry

survival in year 1 (φfryY1) and smolt survival in the ocean on year 2 (φoceY2),

respectively.

PfemFeggs em fryY1 oceY2

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Vital rates

Ela

sti

cit

y

83

How much is enough: the limitations of habitat compensation for offsetting chronic anthropogenic mortality of coho salmon (Oncorhynchus kisutch)

5.1. Abstract

Substantial effort and resources are devoted to rehabilitating freshwater habitats to

improve survival and recovery of endangered salmon populations. Mitigation strategies aim to

offset population losses from human activities by creating or restoring habitats (e.g.

compensation), and in some jurisdictions are bound by a requirement to achieve “No Net Loss”

of population productivity. However, properly devising compensation schemes is challenging,

given large uncertainties in quantifying population losses from development activities as well as

gains (if any) from compensation projects. Anadromous Pacific salmon are ecologically,

culturally, and economically important in the Pacific Northwest (PNW) of the US and Canada,

and face numerous threats from climate change, over-harvesting, and degradation of freshwater

habitats. Here we used a matrix population model of coho salmon (Oncorhynchus kisutch) to

determine the amount of habitat compensation required to offset chronic mortality caused by

development activities (2-20% per year). We ran simulations that imposed chronic mortality on

three different life stages (egg, parr, smolt/adult), to mimic impacts from development, and

tested if smolts produced in constructed side-channels offset these losses. We show that the

typical size of a constructed side-channel built in the PNW is sufficient to compensate for

relatively low levels of chronic mortality for parr or smolt/adult (2-7% per year), but is insufficient

if mortality is >10% per year. When we assumed lower than ideal productivity of constructed

side-channels (e.g.; 25th percentile or median, rather than mean), we found that constructed

channels would need to be 2.5-4.5 times larger to maintain population sizes for all levels of

chronic mortality. We conclude that compensation habitats in freshwater have the potential to

mitigate chronic mortality, especially to early life stages, but often require larger areas than are

typical built in the PNW when we incorporate more realistic or risk-adverse assumptions about

smolt productivity in constructed side-channels. Underestimating the size of compensation

habitat needed to maintain population sizes in the face of development activities is crucial for

the ability of managers and regulators to mitigate effects of disturbances on salmon populations.

84

5.2. Introduction

Billions of dollars are spent annually to mitigate impacts of human activities on biotic

communities and abiotic processes. The stakes of such ecological and economic trade-offs are

high, as total investments in energy, water, and infrastructure development projects are

expected to exceed $53 trillion (US) worldwide between 2010 and 2030 (OECD 2012 in Tallis et

al. 2015). Regulatory agencies often require that developers apply the mitigation hierarchy,

which consists in first trying to avoid, then minimize, and finally compensate for the impact of

projects on ecosystems and biodiversity (Quétier and Lavorel 2011; Tallis et al. 2015). Around

the world, compensation activities in particular are often guided by a requirement for “No Net

Loss” of biodiversity, where losses from development are required to be fully compensated for

so that no declines in biodiversity persist after development (Bull et al. 2013). To ensure the

success of biodiversity compensation, researchers and managers develop measures, called

equivalency metrics, that compute the quantity of compensation needed to offset the losses due

to development. In practice, the difficulty of developing metrics that capture biodiversity impacts

inclusively (Bull et al. 2013) mean that compensation often targets a single, economically

important, species, or is measured in terms of lost habitat, individuals, or productivity of

populations. Equivalency analysis thus balances factors like severity and duration of impacts,

value of affected habitats or populations, and uncertainty in outcomes of both development and

offsets to determine how much compensation is needed (Quétier and Lavorel 2011).

Equivalency metrics are particularly challenging to develop given that losses from development

actions are immediate but difficult to predict in advance, while gains from compensation are

often delayed, leading to a lot of uncertainty in the outcomes (Bradford 2017). Compensation

also suffers from challenges in ensuring compliance by developers and dealing with large

uncertainties in the outcomes of offsetting projects (Maron et al. 2016; Bull et al. 2013; Gardner

et al. 2013). Finally, questions have been raised as to whether compensation activities may do

more harm than good by creating the perception that negative consequences from development

can be effectively mitigated despite few successful examples (Moore and Moore 2013), and by

focusing attention on symptoms rather than causes of development (Beechie and Bolton 1999).

Despite their shortcomings, compensation activities and offsetting are mandated through

legislation in 45 countries, and are under development in another 27 (Bull et al. 2013). For

example, the US Clean Water Act Section 404 (1972) regulates the discharge of dredged

material into aquatic habitats (Hough and Robertson 2009). Historically this act focused

85

primarily on wetland management, but Section 404 includes a permit program that requires

mitigation for all activities that impact fish or wildlife habitat and the use of the mitigation

hierarchy (Doyle and Shields 2012). In Canada, the Fisheries Act is one of the most stringent

pieces of legislation protecting wildlife, and it was amended in 2019 to include the mitigation

hierarchy for protecting the productivity of fisheries. Specifically, activities other than fishing that

result in the death of fish can now be authorized by the Minister, with a provision to consider

offsetting (s34(1)f). Generally, most compensation activities involve creating or restoring

habitats, which is often done in the USA and Canada in “like-for-like” schemes, where

compensation activities replace lost areas of a given habitat by an equal or higher amount of the

same habitat to account for uncertainty (McKenney and Kiesecker 2010; Quétier and Lavorel

2011). However, such an approach does not consider the productive capacity of habitats,

defined as the maximum biomass that can be sustained over time (DFO 1998), that will be

damaged or created. Further, the ratio of lost habitat to created habitat (termed “compensation

ratios”) are generally lower than what is required to achieve “No Net Loss” (e.g. 2:1). As a

consequence, “No Net Loss” of productive capacity remains difficult to achieve in practice (e.g.

25% of cases studied in Canada, Quigley and Harper 2006).

When “like-for-like” offsetting is not feasible, a different approach called “out-of-kind”

offsets is employed, where impacts on one population or location are compensated for by

improving a different population or location, or easing pressures from a different threat on the

targeted population (Bull et al. 2013; Tallis et al. 2015). For example, fish mortality due to

entrainment by water intake structures of a nuclear plant were proposed to be compensated by

removing a dam 50 km inland from the facility (Barnthouse et al. 2019). The added challenge of

“out-of-kind” offsetting is that losses and gains are often in different units or locations, as was

the case when mortality of Golden eagles (Aquila chrysaetos) from wind energy development

was offset by reducing lead poisoning from ammunition throughout Wyoming state (Cochrane et

al. 2015). Thus, instead of focusing on replacing lost productive capacity of habitats by creating

adjacent new ones as in “like-for-like” offsetting, “out-of-kind” offsetting aims at maintaining the

productive capacity of a population overall, defined as the sum of production from all stocks in

an area (Minns 1997). In that regard, recent amendments to the Fisheries Act in Canada (2019)

raise the possibility of using habitat creation as an “out-of-kind” offset for chronic anthropogenic

mortality. However, to our knowledge, there has never been an attempt to equate, in an “out-of-

kind” offsetting scheme, chronic anthropogenic mortality to the habitat compensation that would

be required to appropriately compensate for the expected population-level impacts.

86

Coho salmon and other species of Pacific salmon (Oncorhynchus sp.) are ecologically,

culturally, and economically important species in the PNW (Gende et al. 2002; Naiman et al.

2002). They face many threats and have declined dramatically in abundance over the last

decades (Slaney et al. 1996; Gustafson et al. 2007; Crozier et al. 2019). The main causes of

decline are poor marine survival, commercial and recreational harvest, climate change,

competition from hatchery fish, and the degradation of freshwater habitats that salmon rely on

for spawning and rearing (Kareiva et al. 2000; Chittenden et al. 2010). Degradation of

freshwater habitats include the alteration of in-stream and adjacent terrestrial habitats from

forestry practices (Scrivener and Brownlee 1989; Hartman et al. 1996; Tschaplinski and Pike

2017), the creation of barriers to migration, and the entrainment of fish in water intake structures

of small and large hydropower dams (e.g. in the Columbia River system, Levin and Tolimieri

2001). Freshwater habitat degradation is especially threatening for small populations that are

often additionally challenged by low ocean survival (Nickelson and Lawson 1998; Oosterhout et

al. 2005). As a consequence, an estimated $1 billion (US) a year has been spent on restoration

of streams and rivers within the continental US since 1990 (Bernhardt et al. 2005). Though

researchers and managers increasingly recognize that restoration actions need to approximate

historical disturbance regimes and conserve ecosystem functions to be effective over time

(Waples et al. 2009), most restoration actions still focus on small-scale actions because of

logistic, political, financial, or technical constraints (Roni et al. 2008). For example, restoration

activities often alter the structure of in-channel habitats through the addition of woody debris or

spawning gravels (Roni and Quinn 2001; Roni et al. 2008). In addition, off-channel habitats, like

sloughs, side-channels, beaver ponds, or temporary or permanent flooded areas of the

floodplains of rivers, are also often the target for restoration or compensation actions (Morley et

al. 2005; Roni et al. 2006). For example, a major program in the watershed of the Chilliwack

River in British Columbia, Canada, restored a total of 157,000 km2 of rearing and overwintering

habitat in the floodplain of the large river (69 linear km of habitat accessible to salmon), mostly

by constructing or restoring side-channels and pond complexes (Ogston et al. 2015).

Constructed groundwater side-channels are particularly important habitats for rearing and

overwintering juvenile coho salmon (Morley et al. 2005; Rosenfeld et al. 2008). Recent reviews

reported production values ranging from 0.0143 to 3.25 coho smolts per m2 of constructed side-

channels (Rosenfeld et al. 2008; Roni et al. 2010), while densities ranged from 0.01 to 1.3

smolts per m2 in constructed ponds (Roni et al. 2006; Rosenfeld et al. 2008). The benefits of

restoration projects are based on the assumption that the productivity of constructed off-channel

habitats remains constant over time (Roni et al. 2008). Overall, in- and off-channel

87

compensation habitats are considered to be important and necessary tools to limit or reverse

the decline of coho salmon in the PNW (Nickelson and Lawson 1998; Ogston et al. 2015).

Here we use a matrix population model developed for coho salmon (Oncorhynchus

kisutch) in the Pacific North-West (PNW) to evaluate, in an “out-of-kind” offsetting scheme, the

degree of compensation that would result in No Net Loss of productivity for coho populations

impacted by anthropogenic development activities. Specifically, we evaluated how much

compensation habitat (in m2), under a range of assumptions, is required to offset chronic

mortality (in % of lost individuals) induced by anthropogenic development projects such that

coho population productivity is maintained (i.e. stable population size). Further, we assessed

how chronic mortality of different coho life history stages influenced the effectiveness of

freshwater compensation habitat for emergent population dynamics. Developing a quantitative

framework for balancing development losses and mitigation gains for threatened populations of

coho salmon serves the important call for research into addressing the uncertainty in offset

analysis (Bull et al. 2013). Ultimately, protecting and restoring threatened salmon populations in

the PNW rest upon our ability to evaluate the efficacy of policies and practices of industry and

regulators so that we can achieve both our economic and ecological targets.

5.3. Methods

5.3.1. Scenarios of compensation

To evaluate the amount of habitat compensation required to maintain productivity for

coho populations impacted by anthropogenic mortality, we modelled the impacts of chronic

mortality and the offsetting potential of constructed compensation habitat separately for three

life stages of coho salmon: egg, parr, and smolt/adult. We expected that the value of

constructed compensation habitats to the overall population dynamics of coho salmon would

depend on which life history stage experienced additional mortality. We built scenarios that

varied the chronic mortality imposed on each life stage and the number of smolts produced by

the compensation habitat, and used a 3-year matrix population model with density-dependent

survival in the fry to parr transition (Chapter 4) to assess impacts to population size. In the first

set of scenarios, termed Egg scenarios, we imposed a range of chronic anthropogenic mortality

on eggs, which occurs prior to the density-dependent freshwater survival bottleneck. This set of

scenarios represents a wide range of activities known to cause additional mortality to the egg

development stage, including increased sediment deposition or scour of spawning beds

88

following deforestation in the watershed (Scrivener and Brownlee 1989), thermal extremes

(Tang et al. 1987), and dewatering of nests below dams (Becker and Neitzel 1985, Casas-Mulet

et al. 2005). The second set of scenarios, termed Parr scenarios, corresponded to a range of

chronic anthropogenic mortality from development or operational activities to parr, including

mortality from being entrained in a dam or powerplant (Muir et al. 2001; Barnthouse 2013),

lower flows and warmer temperatures in streams due to climate change (Crozier et al. 2008b),

and degradation in habitat quality needed during freshwater residency due to forestry practices

or urbanization (Hartman et al. 1996; Chittenden et al. 2010). In the final set of scenarios,

termed Smolt/Adult scenarios, we imposed a range of chronic mortality on smolts or adults to

represent examples such as smolt entrainment in, or passage over, dams (Levin and Tolimieri

2001; Schreck et al. 2006), adult mortality during upstream migration due to warmer

temperatures downstream of dams or powerplants (Baisez et al. 2011), adult thermal mortality

from climate change (Farrell et al. 2008), and adult mortality due to recreational freshwater

fisheries (Nguyen et al. 2014). In both the Parr and the Smolt/Adult scenarios, individuals

suffered chronic mortality after the density-dependent survival bottleneck. However, in the Parr

scenarios, we assume that the compensation habitats were constructed downstream of the

source of mortality so that smolts produced in compensation habitats were not exposed to the

simulated anthropogenic mortality. In comparison, in the Smolt/Adult scenarios, chronic

anthropogenic mortality impacted smolts or adults from both the compensation habitat and main

population.

Precise estimates of mortality by classes of anthropogenic disturbance are very rare,

and relating stage-specific vital rates (e.g. survival) to degraded habitat conditions is difficult

(Hayes et al. 2009). Consequently, we chose to run simulations across a broad range of chronic

annual mortality, from 2-20% per year, to represent disturbances resulting in relatively small to

very large annual mortality. We chose the upper limit (20% of mortality per year) to coincide with

data on juvenile entrainment in spillways or turbines of dams (Muir et al. 2001). We applied

these values of annual mortality to egg, parr, or smolt/adult to explore the potential population-

level consequences of chronic anthropogenic impacts, and the scope for compensation habitats

to ameliorate those effects. We calculated how many smolts would need to be produced from

constructed compensation habitats in order to maintain population productivity (i.e. achieve

offsetting equivalency), defined as population abundance returning to baseline (pre-impact)

levels within 45 years. To do so, we ran simulations for each chronic mortality rate and impacted

life stage over a range of compensation habitat sizes constructed for coho salmon in the PNW

89

(Figure 5.1, n= 27 sites, Morley et al. 2005; Roni et al. 2006; Rosenfeld et al. 2008), and noted if

offsetting equivalency was achieved. We focused on constructed groundwater side-channels as

compensation habitats because of their importance for rearing and overwintering coho juveniles

(Morley et al. 2005; Rosenfeld et al. 2008), and to simplify the range of sizes and types of off-

channel habitats included in our simulations.

5.3.2. Deterministic matrix model

To evaluate the potential of constructed side-channels to offset mortality from

anthropogenic sources at the level of population abundance and dynamics, we created a

deterministic matrix model with a one-year time step (sensu Chapter 4, Gibeau and Palen in

review). Our model represents a typical 3-year coho salmon life-cycle (Nickelson and Lawson

1998; Bradford et al. 2000), where spawning occurs in late fall, eggs incubate over winter, and

fry emerge in the spring (Table 5-1). Juveniles typically remain in freshwater until their second

spring, before migrating to the ocean to spend another 18 months before returning to

freshwaters to spawn (Groot et al. 1995). The first year (F13) includes adult fecundity, egg, and

fry survival, while the second year (a21) includes parr and smolt survival, as well as the early

months of adult survival in the ocean. The third year (a32) is made up of adult ocean survival as

well as migration upstream into spawning habitats. We modeled freshwater density dependence

survival, after fry emergence by assuming a moderate (i.e. a = 0.5) survival bottleneck (f(φspr)) to

represent average conditions observed from 10 coho populations in the PNW (mean stream

length = 9 km, mean stream drainage area = 32 km2, Bradford 1995; Korman and Tompkins

2014). The shape of the density-dependent relationship was parameterized according to the

Beverton-Holt relationship:

Dsmolt(t+1) = α × Dfry(𝑡)

1+[(α

k )×Dfry(𝑡)]

(eq. 1)

Where α is the number of smolts per fry at the origin, and k is the carrying capacity for

smolts in the stream (# smolts per km). The carrying capacity for smolts (k) was set to 15,318,

which corresponded to the mean k over the 10 creeks (1702 smolts/km), multiplied by 9 km to

simulate populations inhabiting a hypothetical stream, which was the average length of the 10

creeks (average stream area of 32 km2). We did not repeat the elasticity analysis performed in

Chapter 4 since it was done on a deterministic form of the model and its results are directly

applicable here

90

We initially ran the deterministic model in the absence of added anthropogenic mortality

to establish final population sizes under baseline conditions, using a starting population size

corresponding to the stable age distribution for the deterministic model without density

dependence. This baseline stable age distribution served as the starting population vector (808

adults, 13,379 parr, and 5249 smolts) for all subsequent simulations. We assumed that the

constructed side-channels were always fully seeded (i.e. producing smolts at the maximum

capacity specified in each scenario). To do so, the model allocated spawning adults first to the

constructed side-channel (Na_compH) to equal the number needed to produce the required

number of smolts (varying based on the size of compensation simulated in each scenario), with

the reminder of adults spawning in the main channel. We computed Na_compH by using the

average vital rates as:

𝑁𝑎𝑐𝑜𝑚𝑝𝐻 =𝑁𝑠_𝑐𝑜𝑚𝑝𝐻

(𝑃𝑓𝑒𝑚 ∗ 𝐹𝑒𝑔𝑔𝑠 ∗ 𝜙𝑒𝑚 ∗ 𝜙𝑓_𝑠)

where Ns_compH is the number of smolts contributed by the compensation habitat (which

varies based on the size of side-channels constructed in each scenario), Pfem is the proportion of

females returning to spawn (mean = 0.452), Feggs is the number of eggs produced per female

(mean = 2597), ϕem is the survival rate from eggs to fry emergence (mean = 0.223), and ϕf_s is

the average survival rate from fry to smolts in the absence of density dependence (0.075)

(Table 5-1, Chapter 4 estimation and sources of vital rates). Based on a widespread assumption

used as justification for the size of compensation habitats, we modeled compensation habitat

such that productive capacity did not change through time (i.e. same number of smolts were

produced annually). Finally, the Egg, Parr and Smolt/Adult scenarios differed based on when

the chronic anthropogenic mortality (ϕdist) was applied in the coho life cycle in relation to when

the smolts produced by the compensation habitat (Ns_compH) entered the main population (Ns)

(Table 5-1).

Our simulations further explored how the productive capacity (i.e. quality, expressed in #

of smolts produced per m2) of the compensation habitat influenced the size needed to achieve

offsetting equivalency. We first assumed a mean production by side-channels of 0.47 smolts per

m2, corresponding to the mean smolt production from 33 constructed side-channels sampled in

the PNW (Rosenfeld et al. 2008; Roni et al. 2010; Figure 4.5b). Consequently, for each

scenario, the given size of compensation habitat modelled was converted to Ns_compH, the

number of smolts contributed by the compensation habitat to the main population, by multiplying

91

by 0.47 (e.g. a constructed side-channel measuring 1000 m2 produced 470 smolts). We then

varied the productivity of compensation habitats (smolts produced per m2) to evaluate how our

conclusions changed regarding the size of constructed side-channels required for equivalency.

We computed the 25th (0.1 smolt / m2), 50th (0.18 smolts / m2), and 75th (0.54 smolts /m2)

percentiles of the smolt production from side-channels in the PNW (Rosenfeld et al. 2008; Roni

et al. 2010) and re-ran all scenarios. We ran each simulation for 45 years and compared the

final population size under baseline conditions to those for all ranges of chronic mortality,

compensation habitat sizes and productivity, for the Egg, Parr, and Smolt/Adult scenarios. All

analyses were performed in program R (version 3.5.1, R Core Team 2013).

5.4. Results

5.4.1. Effects of chronic mortality without compensation

The chronic anthropogenic mortality we imposed in our models had different

consequences for coho populations in the absence of constructed side-channels and varied

depending on which life stage experienced the chronic mortality. Overall, in the absence of

compensation, final population sizes were closer to the baseline abundances when mortality

affected eggs than if chronic mortality targeted parr or smolts (Figure 5.2). When we modeled

egg mortality without compensation, the consequences for final population sizes were minimal

as much of the extra mortality was buffered by the natural density-dependent bottleneck

occurring after the mortality. For example, the final population sizes in scenarios with chronic

egg mortality, ranging from 2 to 20%, were always at least 97% of the final baseline population

(Figure 5.2). However, when chronic mortality was imposed on parr or smolts/adults, the final

population size declined as chronic mortality increased (Figure 5.2).

5.4.2. Effects of habitat compensation

Populations impacted by chronic mortality affecting all life stages benefitted from

contributions of smolts by constructed side-channels, as final population sizes increased linearly

with sizes of constructed side-channels until, and beyond, achieving offsetting equivalency (i.e.

reaching adult abundances corresponding to baseline levels). The importance of side-channel

compensation habitat varied for each life stage. For example, relatively small amounts of

constructed side-channels were needed to achieve offsetting equivalency when mortality

affected eggs, but the size of side-channels needed to reach offsetting equivalency increased

92

quickly when chronic mortality affected parr and smolts/adults (Figure 5.2). For example, a side-

channel as small as 1058 m2 (i.e. smaller than the 10th percentile of side-channels built in the

PNW) compensated for chronic annual mortality <20% acting on the egg stage (Figure 5.2a). In

fact, a side-channel of 1000 m2 was sufficient to equal or even increase the final adult

population sizes compared to the baseline scenario across all levels of egg chronic mortality

(Figure 5.3). Comparatively, more compensation habitat was needed to achieve offsetting

equivalency when smolts/adults were affected rather than parr, especially as the intensity of

chronic mortality increased >5% (Figure 5.2b).

Our results suggest that the average size of constructed side-channels in the PNW (i.e.

3400 m2) is likely to be effective at offsetting chronic annual coho mortality of up to 10%

affecting parr, but would not be effective for similar amount of extra mortality impacting smolts or

adults (Figure 5.2, Figure 5.3). Our models suggest that when chronic mortality of parr or

smolt/adults exceeds 15%, the size of compensation habitat required to achieve offsetting

equivalency would need to be larger than the 90th percentile of sizes typically built in the PNW

(i.e. 5680 m2, Figure 5.2). In the worst-case scenario, when we imposed 20% of extra mortality

from anthropogenic development activities on smolts or adults, achieving offsetting equivalency

would require at least 8259 m2 of compensation habitat (Figure 5.2), corresponding to more

than twice the average size of the side-channels currently built in the PNW.

5.4.3. Productivity of compensation habitats

Our results suggest that our assumptions regarding the productivity of compensation

habitats (i.e. the number of smolts produced by the constructed side-channels) have a large

effect on the size of compensation needed to achieve offsetting equivalency. Indeed, running

simulations with median productivity (0.18 smolts per m2) or the 25th percentile of productivity

(0.1 smolts per m2) (Rosenfeld et al. 2008; Roni et al. 2010) increased the size of side-channels

required to achieve offsetting equivalency as compared to sizes needed with mean productivity

(0.47 smolts per m2), especially for parr and smolts/adults (Figure 5.4). For all levels of mortality

affecting parr or smolts/adults, the size of constructed side-channels would need to increase by

2.5-fold to achieve equivalency if we assumed median productivity, or by 4.5-fold if we assumed

25th percentile productivity, as compared to mean quality (Figure 5.4). In the worst-case

scenario we simulated, 20% annual chronic mortality affecting smolts/adults, reaching offsetting

equivalency would require a 6- to 11-fold increase in size of side-channels if we assumed either

a median or 25th-percentile level of productivity.

93

5.5. Discussion

We used a matrix population model to perform an equivalency analysis in an ”out-of-

kind” scheme by simulating how chronic mortality applied to three different life stages of coho

salmon translated into the minimum amount of compensation habitat needed to mitigate the

effects of chronic disturbance and meet the requirement for No Net Loss of population

productivity. Our models suggest that the mean size of constructed side-channels typically built

in the PNW would be sufficient to compensate for chronic mortality of up to 20% annually if it

affected the egg stage, but only up to 7% if it affected parr, smolts, or adults. Moreover, the

same mean size of constructed side-channels would be insufficient if chronic mortality climbs

above 10% for parr or smolts/adults, or if the productivity of side channels is lower than the

mean values used in our baseline scenario. When we lowered the smolt productivity in modelled

side-channels, we found that the size of compensation habitats would need to be much larger to

maintain population sizes in the face of chronic anthropogenic mortality. A more precautionary

approach to building side-channels would be to assume less than ideal productivity in

compensation habitats, which we simulated using median productivity (0.18 smolts/m2) or the

25th percentile of productivity (0.1 smolts/m2) observed in constructed side-channels surveyed

in the PNW (Rosenfeld et al. 2008; Roni et al. 2010). Our results indicate that lower productivity

would require side-channels 2.5 to 4.5 times larger than those required when we assumed

mean productivity (sensu Rosenfeld et al. 2008; Roni et al. 2010). Additionally, our results

suggest that compensation habitats are more effective at mitigating chronic mortality if they are

built downstream of where disturbances occur, such that smolts they produce are not affected

by chronic mortality experienced by the population in the main channel. For example, our results

show that smaller constructed side-channels (or of lower productivity) could compensate for

chronic mortality to parr but not for smolts or adults. This finding highlights the additional

challenge posed if compensation (construction of habitats in freshwater) were to address

impacts such as high ocean mortality. Overall, if chronic mortality also affects the smolts

contributed by the compensation habitats (e.g. in cases of smolts entrainment into downstream

dams or increased ocean mortality), achieving offsetting equivalency will require more

compensation habitat to be built in freshwater.

As our modelling of an “out-of-kind” compensation scheme highlights, designing effective

side-channels for compensation is complex. Our results emphasize the importance of

considering variation in productivity when deciding on the optimal size of compensation habitat

94

needed. However, the size of constructed side-channels may also impact quality of the

compensation habitat. For example, studies have noted a decline in smolt density with

increasing side-channel sizes, suggesting that smaller side-channels may be more ecologically

efficient than large side-channels (Keeley et al. 1996; Rosenfeld et al. 2008). Building

numerous, but smaller, side-channels is also often more technically and economically feasible

than large ones. As such, our recommendation to increase the size of compensation habitat to

acknowledge the uncertainty around the quality of smolt production may be best achieved by

building more side-channels rather than one larger side-channel. In our models, we assumed

that larger compensation habitats could be equally as productive as smaller habitats, but if

productivity declines with increasing size or through time due to degradation, our results may

underestimate the sizes needed to maintain population equivalency.

Adding to the complexity of designing constructed side-channels of adequate size and

quality, the size and the number of side-channels are frequently chosen based on availability of

land and costs of restoration rather than based on ecological bottlenecks or the potential for

success (Rosenfeld et al. 2008). Such an opportunistic selection of compensation sites is stated

as one reason why “No Net Loss” of productivity is often not achieved in practice (Quigley and

Harper 2006). Our results suggest that side-channels constructed downstream of where the

anthropogenic impacts occur would be more effective at compensating for chronic mortality.

However, downstream reaches of streams and rivers are often less stable geomorphologically

(Naiman et al. 2000; Naiman et al. 2008), and resources and efforts in building compensation

habitats in low-elevation floodplains could be lost if, or when, natural large flow events re-shuffle

the landscape. Other challenges in building side-channels as compensation habitats include the

need for connectivity between the compensation habitats and main channel, which is crucial to

ensure the success of mitigation (Anderson et al. 2014; Crozier et al. 2019). For example,

potential spatial clustering of juveniles and limited migration may lower the carrying capacity of

the new habitat by limiting the number of fry that move in from spawning or rearing grounds

(Walters et al. 2013b). Finally, salmon need dynamic and diverse freshwater habitats to thrive,

and compensation projects must sustain natural evolutionary processes in order to promote the

health and resilience of the species (Waples et al. 2009; Beechie et al. 2010; Booth et al. 2016),

especially considering the added challenges posed by climate change. For example, improving

temperature and flow regimes through riparian restoration and increased food availability may

improve the tolerance of salmon populations to warmer water temperatures induced by climate

change (Crozier et al. 2019). Most Pacific salmon populations face multiple threats

95

simultaneously, and cumulative impacts from multiple sources of anthropogenic mortality on

more than one life-stage at a time will increase the need for mitigation and compensation. As an

indication, simple extensions of our model show that the size of compensation habitat needed

for equivalency would be 2X higher or more (range: +193-230%) if chronic mortality affected all

three life stages simultaneously instead of just to parr, or smolts/adults individually, and would

increase by 14-16X (1450-1675%) compared to when only eggs were affected by chronic

mortality. As such, our simulations of only a single life stage experiencing chronic mortality

suggest that our results are likely to be underestimates of the minimum sizes of constructed

side-channels needed to maintain population sizes in the face of multiple, cumulative sources of

anthropogenic mortality (e.g. Hartman et al. 1996). Overall, the importance of side-channels for

coho juveniles and the fact that quality overwintering habitat, especially, is limiting coho smolt

production in the PNW (Nickelson and Lawson 1998; Morley et al. 2005; Oosterhout et al. 2005)

should not discourage attempts to build efficient compensation habitat for coho salmon despite

the complexity and challenges involved.

At the population level, our results highlight that understanding how mitigation may affect

salmon depends on knowledge of the density-dependent mechanisms in the population(s)

affected. For example, natural density-dependent survival bottlenecks may compensate for

some of the mortality imposed by anthropogenic disturbances if extra mortality imposed by

anthropogenic disturbances occur before density dependence (see Chapter 4, Hodgson et al.

2017). The modest population-level impacts of additional egg mortality in our scenarios show

further the potential for density-dependent survival to compensate for some anthropogenic

mortality, and how natural density-dependent bottlenecks may influence the need for

compensation habitat. However, it is risky to rely on natural density dependence to compensate

for chronic anthropogenic mortality as it remains very challenging to detect if, when and how

strong density-dependent survival bottlenecks occur in freshwater for specific populations (Rose

et al. 2001; Milner et al. 2003; Walters et al. 2013b). Moreover, density-dependent survival may

only compensate for additional mortality if population sizes approach carrying capacity

(Goodwin et al. 2006), which may be uncommon for many anadromous salmon populations

affected by anthropogenic disturbances (Achord et al. 2003), unless anthropogenic disturbances

also lower carrying capacities of freshwater habitats (Greene and Beechie 2004). Other forms of

density dependence may enhance the efficacy of compensation habitat in mitigating effects of

disturbances. For example, density-dependent migration, whereby individuals in areas with high

densities move to areas with lower densities, allows the compensation habitat to be more

96

effectively populated (Greene and Beechie 2004). Overall, long-term studies detailing

population dynamics of local populations appear crucial to adequately design effective

compensation habitats.

Our model assumed that quality of the compensation habitat remained constant over

time, which is an unrealistic expectation especially if adult abundances were low. We do not

expect that assumption to largely affect model outcome however, as our model was

deterministic and always reached equilibrium (given the density-dependent function included)

after a few generations. However, our results highlight the need for continuous maintenance, or

at least that monitoring compensation effectively span several years, which is seldom done in

practice. For example, offsetting projects were monitored for an average of only 3.7 years in

Canada (Harper and Quigley 2005). Furthermore, compliance success does not always equate

to ecological success (Quigley and Harper 2006), and the high inter-annual variability in salmon

population dynamics (e.g. Beamish et al. 1999) means responses to mitigation have to be large

to be detected (Roni et al. 2010). The efficiency of compensation habitat may particularly need

to be assessed over time given the predicted changes in stream flow and temperature regimes

due to climate change (Kerkhoven and Yew Gan 2011; Crozier et al. 2019). For example,

increases in peak flows may cause failure of in-channel structures or reduce efficiency of off-

channel constructed features (Beechie et al. 2013). Furthermore, a substantial amount of

uncertainty remains in our modelling results due to the uncertainty around mortality rates for

various sources of anthropogenic mortality. Quantifying how mortality from anthropogenic

sources of mortality (e.g. forestry, entrainment in dam structures, harvesting, climate change

etc.) translates into population-level mortality rates is a crucial piece to designing enough and

effective measures of habitat compensation. Overall, the need for long-term maintenance and

monitoring, as well as empirical data to quantify expected mortality rates, cannot be stressed

enough so we can collectively better manage our threatened rivers and salmon populations.

5.6. Conclusion

Our results suggest that compensation habitats for coho salmon in freshwater have the

potential to mitigate chronic mortality. However, we suggest that achieving offsetting

equivalency may require larger side-channels than those that are typically built in the PNW,

especially when we incorporate more risk-adverse assumptions about smolt productivity in

constructed side-channels. Overall, our results illustrate the high potential of compensation

habitat for mitigating chronic mortality of coho due to anthropogenic disturbances. The new

97

habitats of the constructed side-channels may also opportunistically help other components and

species of the ecosystems (Ogston et al. 2015). However, focus on building compensation

habitat should not be made to the detriment of avoidance and reduction of anthropogenic

effects, which come first in the mitigation hierarchy to offset disturbances. Our results further

show that matrix population models can be effectively used to quantitatively estimate the

uncertainty in “out-of-kind” offset analysis. We conclude that accounting for uncertainty in

productivity of compensation over time requires increasing the size of compensation habitat

generally built. Ultimately, we hope that our results will help guide industry and regulators to

better evaluate both the potential and limitations of compensation habitat in mitigating the

population-level impacts of chronic mortality in freshwater, and help protect and restore

threatened salmon populations.

Acknowledgements

We thank K. Wilson, D.A Greenberg, A. Cantin, R. Murray, J.W. Moore, and M. J.

Bradford for help with conceptualizing the analysis and comments that greatly improved the

manuscript. This work was funded by grants from NSERC Discovery Grant, the Gordon and

Betty Moore, and Wilburforce Foundations to WJP.

98

Table 5-1. Deterministic matrix model and scenarios per life stage.

[

0 0 𝐹13

𝑎21 0 00 𝑎32 0

] x (

𝑁𝑝

𝑁𝑠 + 𝑁𝑠_𝑐𝑜𝑚𝑝𝐻

𝑁𝑎 − 𝑁𝑎_𝑐𝑜𝑚𝑝𝐻

) = (

𝑁𝑝

𝑁𝑠

𝑁𝑎

)

Transition rate aij Parameter equations

Egg scenarios

Eggs to parr F13 Pfem * (Feggs * φdist) * φem * ƒ(φspr)

Parr to smolt a21 φoceY2

Smolt to adult a32 φoceY3

Parr scenarios (chronic mortality before compensation)

Eggs to parr F13 Pfem * Feggs * φem * ƒ(φspr) * φdist

Parr to smolt a21 φoceY2

Smolt to adult a32 φoceY3

Smolt/Adult scenarios (chronic mortality after compensation)

Eggs to parr F13 Pfem * Feggs * φem * ƒ(φspr)

Parr to smolt a21 φoceY2

Smolt to adult a32 φdist * φoceY3

Pfem= proportion of females spawning (0.452); Feggs= # of eggs per female (2597); φem= survival from hatching to emergence (0.223); f(φf_spr)= density-dependent survival of fry through spring bottleneck; φoceY2= survival of smolt through 1st summer and fall in ocean (0.392); φoceY3= survival of adults in ocean (Year 3, 0.154); φdist= survival after chronic anthropogenic disturbance; Na= Number of adults; Na_compH= Minimum number of adults needed to seed the compensation habitat; Np= Number of parr; Ns= Number of smolts; Ns_compH= Number of smolts contributed by the compensation habitat

99

Figure 5.1. Distribution of (a) the size (m2) of 27 side-channel habitats built in the PNW (from Morley et al. 2005; Roni et al. 2006; Rosenfeld et al. 2008), and (b) the productivity of 33 side-channel habitats (# of smolts per m2) built in the PNW (from Rosenfeld et al. 2008; Roni et al. 2010).

100

Figure 5.2. Proportion of the final size of the baseline population represented by populations subjected to chronic anthropogenic mortality (2 to 20%) affecting a) Eggs, and b) Parr and Smolts/Adults as a function of the size of side-channel compensation habitat (m2). A relative population size of 1 implies that the population modelled had the same final population size as the baseline population (n= 808 adults). The grey box represents the 10th to the 90th percentile, and the vertical line indicates the average size of compensation habitat built in the PNW (based on 27 sites, from Morley et al. 2005; Roni et al. 2006; Rosenfeld et al. 2008). Note that the Y-axis is different for the two panels.

101

Figure 5.3. Change in final number of adults for impacted populations compared to baseline abundances (%, colors) over a range of sizes of constructed side-channels (y-axis), annual chronic anthropogenic mortality (x-axis), and life stage (eggs, parr, smolts/adults) affected by the chronic mortality (panels), assuming mean productivity in constructed side-channels. 0 (white cells) indicates no change in final abundances of impacted populations compared to baseline, while positive values (cool shades) mean compensation increased the final number of adults in impacted populations and negative values (warm shades) mean final size of impacted populations decreased despite the amount of compensation habitat added. The black lines indicate offsetting equivalency for the combination of compensation habitat sizes and chronic mortality.

102

Figure 5.4. Range in sizes of compensation habitat (m2) required to effectively offset annual chronic mortality ranging from 2 to 20% affecting, a) Eggs, b) Parr, and c) Smolts/Adults, assuming varying productivity (# of smolts produced per m2) for the constructed side-channels. Horizontal black lines indicate the 25th to 75th percentiles of productivity of compensation habitat, black triangles represent mean number of smolts contributed, and stars represent median number of smolts contributed by compensation habitat (based on data from 33 sites, from Rosenfeld et al. 2008; Roni et al. 2010). The grey box represents the mean (vertical line) and 10th to 90th percentiles (shaded) of sizes of compensation habitats built in the PNW (n= 27 sites, data from Morley et al. 2005; from Roni et al. 2006; Rosenfeld et al. 2008).

103

General conclusions

In the last three decades, the global emergence of RoR hydropower in BC and

worldwide has increased pressure on salmonid fishes by altering their freshwater habitats

through the diversion of waters for electricity production. Understanding how RoR hydropower

alters natural flow regimes and subsequently the survival, growth, bioenergetics, and population

dynamics of salmonid fishes can help to effectively design future developments that balance

conserving biodiversity with economic needs. My thesis filled important knowledge gaps around

potential impacts from RoR hydropower on stream ecosystems in general, and salmonid

populations in particular. I used a combination of published research, empirical data and models

to summarize and explore different pathways of impacts by which RoR hydropower may affect

salmonids, and make recommendations for future development and conservation.

In Chapter 2, I reviewed the scientific literature and general salmonid ecology to

summarize three main pathways of impacts through which RoR hydropower may impact

salmonids: the reduction of flow in the bypassed reach, the presence of low-head dams

impounding rivers, and the production of anthropogenic flow fluctuations. This chapter highlights

both the paucity of peer-reviewed literature dedicated specifically to RoR hydropower (n=31

peer-reviewed papers, of which only 10 targeted specifically salmonids), and the need to

conduct more research in specific areas where most impacts are expected to occur. I highlight

that RoR dams may divert a considerable proportion (up to 97%) of incoming river flow away

from the main river channel, and that the magnitude of flow diversion by RoR hydropower varies

by season and is proportionally higher when flows in rivers are naturally low. I emphasized that

flow regulation for RoR hydropower is unique in how it changes the seasonality, magnitude,

frequency, and duration of flow events, both in the bypassed (particularly increasing low flow

periods) and downstream (producing anthropogenic flow fluctuations) reaches of regulated

rivers. However, the fact that low-head dams can be overtopped at high flows may reduce the

potential for discontinuities in habitat, barriers to migration, and impacts on stream

geomorphology, though many uncertainties remain. All these pathways of effect from RoR

hydropower have the potential to alter habitat quantity and quality in regulated streams, and to

negatively affect salmonid survival, growth, and fitness. The remaining major uncertainties I

104

identified as avenues for future research are the impacts of extended low flow periods in the

bypassed reaches, the consequences of anthropogenic flow fluctuations, and the cumulative

effects of multiple RoR projects on the same river systems. I used the knowledge gaps identified

through this synthesis to guide the development of my subsequent thesis chapters.

In Chapter 3, I evaluated the impacts of extended low flow periods created from flow

diversion in bypassed reaches. To do so, I quantified the increases in water temperature

observed in bypassed reaches and simulated potential consequences to the bioenergetics of

resident salmonids. I found that the annual increases in water temperatures in bypassed

reaches due to flow diversion by RoR hydropower averaged 0.5 and 0.8℃ in the two creeks

surveyed and reached a monthly maximum of 1.4℃. Using observed (increased) and predicted

temperatures (natural, without flow diversion), I found that increasing water temperatures from

flow diversion by RoR hydropower could have negative effects on rainbow trout (Oncorhynchus

mykiss) by reducing their growth if food availability does not compensate for increased

temperatures. Alternatively, warmer waters due to flow diversion could increase the growth

potential for fish if food availability and fish consumption increased with temperature and

temperature remains below the species’ optima. This chapter highlights the need for empirical

data on fish consumption and food availability to better understand how aquatic and terrestrial

invertebrates respond to flow diversion by RoR hydropower, given their importance in mediating

effects of warmer waters on fish metabolism and growth.

Flows downstream of RoR infrastructures can fluctuate unexpectedly and differ in timing,

frequency, magnitude, and duration as compared to classic reservoir-storage hydropower (see

Chapter 2). In Chapter 4, I explored the consequences of RoR-induced flow fluctuations to coho

salmon population dynamics using stochastic stage-structured matrix models that integrate

density dependence. In general, I found that RoR-induced flow fluctuations may impact

population dynamics of coho salmon by increasing the probability of quasi-extinction and

reducing adult population sizes. Importantly, the impacts of anthropogenic flow fluctuations to

coho salmon varied with both the strength and timing of density-dependent survival in

freshwater habitats. When density dependence was weak, even rare and mild flow fluctuations

substantially affected population dynamics. When density dependence was strong, coho

populations had a greater capacity to compensate for the loss of fry from RoR-induced flow

fluctuations. Demographic compensation was strongest when the timing of density dependence

was best aligned to the timing of flow fluctuations (i.e. occurred in the winter after most flow

fluctuation events). However, frequent and deadly anthropogenic fluctuations overwhelmed

105

even the strongest winter density-dependent survival bottleneck. Overall, Chapter 4

demonstrates that consequences from RoR-induced flow fluctuations should not be taken

lightly, and that while natural density dependent survival bottlenecks may help compensate for

some of the extra anthropogenic mortality induced, their timing and strength are critical.

Finally, in Chapter 5, I explored how the building of compensation habitats in freshwater

could mitigate chronic anthropogenic mortality and maintain abundance and resilience of

salmon populations. I ran simulations to cover a variety of sources of human activities known to

affect anadromous salmons, like modifications to riverine and terrestrial landscapes such as

large dams or forestry practices, harvesting, or increased ocean or freshwater mortality due to

climate change. I used a deterministic version of the stage-structured matrix model developed in

Chapter 4 and ran simulations for a range of chronic mortality affecting three different life

stages, where I varied the size and the productivity (in terms of production of smolts) of side-

channels built as habitat compensation in impacted streams. My results show that very limited

amount of compensation habitat would be needed to compensate for the loss of eggs, given the

presence of a natural density-dependent survival bottleneck occurring after the chronic

mortality. However, the mean size of constructed side-channels typically built in the PNW would

only be enough to compensate for chronic mortality below 10% for parr or smolts/adults. Also, I

conclude in Chapter 5 that compensation habitats in freshwater have the potential to mitigate

chronic mortality of parr, smolt or adult coho, but often require larger areas than are typically

built in the PNW, especially if we want to use precaution and incorporate more realistic

assumptions about smolt productivity in constructed side-channels. Overall, the simulations I

performed in Chapter 5 illustrate both the large potential and the large uncertainties about the

use of habitat compensation in freshwater to offset the loss of coho of various life stages.

My thesis highlights both the potential for RoR hydropower to impact salmonids in

regulated rivers, and the need for more research and data to confirm the trends identified by my

modelling and simulations. For example, results from Chapter 3 show the potential of flow

diversion to increase water temperature in bypassed reaches with possible consequences for

fish bioenergetics, especially if food becomes limited. However, sampling invertebrate

abundance and consumption by fish in RoR-regulated rivers, as well as measuring growth of

fish in the same locations as where temperature is monitored, would be great next steps to

further our understanding of the bioenergetics consequences of RoR hydropower on resident

salmonids. Chapter 4 illustrates the potential for RoR-induced flow fluctuations to influence

emergent population dynamics, and the importance of density-dependent survival bottlenecks to

106

mitigate the loss of fry. Next steps of research should include a thorough quantitative

assessment of mortality rates of salmon fry in relation to specific flow fluctuations events of

varying magnitude. That work would imply detailed population sampling both before and

immediately after anthropogenic flow fluctuations downstream of RoR facilities. The importance

of density-dependent survival shown in this chapter also emphasizes the importance of

adequately assessing the presence, strength and timing of density-dependent bottlenecks in

specific populations prior to the development of RoR hydropower facilities if one expects such

bottlenecks to mitigate loss of fry from RoR-induced flow fluctuations. Chapter 5 quantifies the

high uncertainty in the size and productivity of compensation habitat needed to offset chronic

mortality and maintain population sizes. Again, precise and empirical estimates of mortality

rates of eggs, parr, smolts, and adults in relation to various anthropogenic disturbances are

needed to ground-truth the results of the simulations. Overall, results from all three quantitative

chapters strongly support the need for long-term, detailed, empirical datasets in regulated rivers

to confirm, infirm and/or quantify the magnitude of the impacts identified by the modelling and

simulations I performed.

My thesis fills in some of the knowledge gaps identified around effects of small-scale,

RoR hydropower on stream ecosystems, and represents a unique and important contribution to

increase our state of knowledge in that domain. I also hope that my thesis can serve as a

cautionary tale illustrating that impacts from RoR hydropower may not be as negligible as their

small physical footprint suggests (in support of Couto and Olden 2018). However, among the

many threats facing anadromous and resident salmonids around the world, the impacts from

RoR hydropower may not be a primary concern. That would especially be the case if, at least a

minimum of, effort is made to mitigate unavoidable consequences of RoR hydropower on

regulated rivers (e.g. by requiring minimum environmental flows, maintenance of compensation

habitat, technical upgrades to limit flow fluctuations, etc.). Yet, most fish populations face

multiple sources of disturbances that are likely to have additive or synergistic effects on their

abundance and survival. As such, assessing impacts from RoR hydropower (or any other

anthropogenic disturbance) in isolation from other sources threatening populations locally or

globally could carry grave consequences for salmonids and other aquatic species (e.g. Kibler

and Tullos 2013). For example, studies showed interactions between climate change and water

diversion increased threats for growth, movement and survival of Chinook salmon (Walters et al.

2013a), or how increased variability in extreme flow events due to climate change may further

challenge the survival of coho salmon in freshwater (Ohlberger et al. 2018). Increases in water

107

temperature expected following climate change (e.g. Isaak et al. 2016) could also act

synergistically with those I modelled in bypassed reaches of streams regulated by RoR

hydropower in Chapter 3 and add to bioenergetics challenges for stream salmonids. As such,

the need for mitigation of negative consequences from development and for restoration and

compensation actions appear unavoidable. We collectively will likely need to embrace the

building of compensation habitats as a mean to increase carrying capacity in freshwater

ecosystems and enhance resilience of salmonids populations, despite the uncertainties and

limitations that I noted in Chapter 5.

In conclusion, my thesis highlights that the rapid rise of RoR hydropower around the

world presents both promising opportunities for increasing renewable energy production, and

important challenges in predicting impacts to stream ecosystems and their inhabitants. As the

global need for RoR hydropower increases, concerted efforts to study the impacts of RoR

hydropower dams on salmonids is crucial. In parallel, changes in regulatory frameworks are

needed to ensure more resources are devoted to long-term monitoring of regulated rivers, as

well as to cumulative impacts of all sources of anthropogenic disturbances acting on stream

ecosystems. Attention should also be given to the fact that many small RoR facilities are

needed to produce the same amount of energy as fewer large facilities, thus potentially yielding

higher landscape-scale impacts per unit of energy than large hydropower dams (Bakken et al.

2012; Kibler and Tullos 2013). As unpopular as these recommendations may be for project

developers, I believe we have a collective responsibility to look after our threatened freshwater

ecosystems and their stream-dwelling organisms. This responsibility is greater under the

terrifying threats of climate change that are likely to exacerbate consequences of other, milder

sources of anthropogenic disturbances on streams and rivers. Simultaneously, we cannot afford

to snob the great potential of RoR hydropower to respond in a reasonably clean and local way

to the important need for renewable energies as we transition our economies away from fossil

fuel. Done properly, RoR hydropower holds promise for providing both some of the energy

needed to sustain our economies, and the opportunity to preserve our natural resources and

ecosystems that are also critical to healthy communities and economies.

Overall, I hope that my work through this thesis will help further advance the

conversation around RoR hydropower and its impact on freshwater ecosystems by clarifying

what are the potential pathways of impacts, and their possible magnitude. I also hope that my

thesis can bring glimmers of hope that human interventions like the building of compensation

habitats can help in mitigating unavoidable consequences of human activities. However, I also

108

want to caution against the mirage of compensation (sensu Moore and Moore 2013). We need

to keep investing enough to clarify the many remaining uncertainties, and compensation and

restoration work needs to respect its place in the mitigation hierarchy, i.e. come after avoidance

and minimization of effects. Finally, I hope that my work can empower us all to act and devote

the time and energy in pursuing fundamental research. We also need to implement concurrently

the many social and technical solutions that already exist since the urgency to protect our

threatened salmonid populations and river ecosystems is clear. Hopefully my work can add

strength to the power of science and ensure that evidence-based decisions are made for

sustainable management of salmonid populations and stream ecosystems.

109

References

Abbasi, T., and Abbasi, S.A. 2011. Small hydro and the environmental implications of its

extensive utilization. Renewable and Sustainable Energy Reviews 15: 2134–2143.

doi:10.1016/j.rser.2010.11.050.

Achord, S., Levin, P.S., and Zabel, R.W. 2003. Density-dependent mortality in Pacific salmon:

the ghost of impacts past? Ecology Letters 6: 335–342.

Allen, D., Rosenfeld, J., and Richards, J. 2016. Physiological basis of metabolic trade-offs

between growth and performance among different strains of rainbow trout. Canadian

Journal of Fisheries and Aquatic Science 73: 1493–1506. doi:10.1139/cjfas-2015-0429.

Almodόvar, A., and Nicola, G.G. 1999. Effects of a small hydropower station upon brown trout

Salmo trutta L. in the river Hoz Seca (Tagus Basin, Spain) one year after regulation.

Regulated Rivers: Research & Management 15: 477–484.

Altukhow, Y.P., Salmenkova, E.A., and Omelchenko, V.T. 2000. The theoretical principles of

population genetics. In Salmonid Fishes: Population biology, genetics and management,

English translation edited by J. Thorpe and G. Carvalho. Blackwell, United Kingdom. pp.

1–32.

Anderson, D., Moggridge, H., Shucksmith, J.D., and Warren, P.H. 2015. Quantifying the impact

of water abstraction for low head “run of the river” hydropower on localized river

channel hydraulics and benthic macroinvertebrates. River Research and Applications: 1–

12. doi:10.1002/rra.2992.

Anderson, D., Moggridge, H., Warren, P., and Shucksmith, J. 2014. The impacts of “run-of-

river” hydropower on the physical and ecological condition of rivers. Water and

Environment Journal. doi:10.1111/wej.12101.

Anderson, E.P., Freeman, M.C., and Pringle, C.M. 2006. Ecological consequences of

hydropower development in Central America: impacts of small dams and water diversion

on neotropical stream fish assemblages. River Research and Applications 22: 397–411.

doi:10.1002/rra.899.

Aroonrat, K., and Wongwises, S. 2015. Current status and potential of hydro energy in Thailand:

a review. Renewable and Sustainable Energy Reviews 46: 70–78.

doi:10.1016/j.rser.2015.02.010.

Ayllόn, D., Railsback, S.F., Harvey, B.C., García Quirós, I., Nicola, G.G., Elvira, B., and

Almodόvar, A. 2019. Mechanistic simulations predict that thermal and hydrological

effects of climate change on Mediterranean trout cannot be offset by adaptive behaviour,

evolution, and increased food production. Science of the Total Environment 693.

doi:10.1016/j.scitotenv.2019.133648.

Bacon, P.J., Gurney, W.S.C., Jones, W., Mclaren, I.S., and Youngson, A.F. 2005. Seasonal

growth patterns of wild juvenile fish: partitioning variation among explanatory variables,

based on individual growth trajectories of Atlantic salmon (Salmo salar) parr. Journal of

Animal Ecology 74: 1–11.

Baisez, A., Bach, J.-M., Leon, C., Parouty, T., Terrade, R., Hoffman, M., and Laffaille, P. 2011.

Migration delays and mortality of adult Atlantic salmon (Salmo salar) en route to

spawning grounds on the River Allier, France. Endangered Species Research 15: 265–

270. doi:10.3354/esr00384.

110

Baker, D.W., Bledsoe, B.P., Albano, C.M., and Poff, N.L. 2011. Downstream effects of

diversion dams on sediment and hydraulic conditions of Rocky Mountain streams. River

Research and Applications 27: 388–401. doi:10.1002/rra.1376.

Bakken, T.H., Sundt, H., Ruud, A., and Harby, A. 2012. Development of small versus large

hydropower in Norway -- comparison of environmental impacts. Energy Procedia 20:

185–199. doi:10.1016/j.egypro.2012.03.019.

Barnthouse, L.W. 2013. Impacts of entrainment and impingement on fish populations: a review

of the scientific evidence. Environmental Science & Policy 31: 149–156.

doi:10.1016/j.envsci.2013.03.001.

Barnthouse, L.W., Boreman, J., Christensen, S.W., Goodyear, C.P., van Winkle, W., and

Vaughan, D.S. 1984. Population biology in the courtroom: the Hudson River controversy.

BioScience 34(1): 14–19.

Barnthouse, L.W., Fietsch, C.-L., and Snider, D. 2019. Quantifying restoration offsets at a

nuclear power plant in Canada. Environmental Management 64: 593–607.

doi:10.1007/s00267-019-01214-2.

Battin, J., Wiley, M.W., Ruckelshaus, M.H., Palmer, R.N., Korb, E., Bartz, K.K., and Imaki, H.

2007. Projected impacts of climate change on salmon habitat restoration. PNAS 104(16):

6720–6725.

Bauersfeld, K. 1978. Stranding of juvenile salmon by flow reductions at Mayfield dam on the

Cowlitz River 1976. State of Washington, Department of Fisheries, Olympia, Technical

report 36.

Beakes, M.P., Moore, J.W., Hayes, S.H., and Sogard, S.M. 2014. Wildfire and the effects of

shifting stream temperature on salmonids. Ecosphere 5(5): 1–14. doi:10.1890/ES13-

00325.1.

Beamish, R.J., Noakes, D.J., McFarlane, G.A., Klyashtorin, L., Ivanov, V.V., and Kurashov, V.

1999. The regime concept and natural trends in the production of Pacific salmon.

Canadian Journal of Fisheries and Aquatic Science 56: 516–526.

Beechie, T., and Bolton, S. 1999. An approach to restoring salmonid habitat-forming processes

in Pacific Northwest watersheds. Fisheries 24(4): 6–15.

Beechie, T., Imaki, H., Greene, J., Wade, A., Wu, H., Pess, G.R., Roni, P., Kimball, J., Stanford,

J., Kiffney, P., and Mantua, N. 2013. Restoring salmon habitat for a changing climate.

River Research and Applications 29: 939–960. doi:10.1002/rra.2590.

Beechie, T.J., Sear, D.A., Olden, J.D., Pess, G.R., Buffington, J.M., Moir, H., Roni, P., and

Pollock, M.M. 2010. Process-based principles for restoring river ecosystems. BioScience

60(3): 209–222.

Becker, C.D., and Neitzel, D.A. 1985. Assessment of intergravel conditions influencing egg and

alevin survival during salmonid redd dewatering. Environmental Biology of Fishes 12(1):

33–46.

Bell, E., Kramer, S., Zajanc, D., and Aspittle, J. 2008. Salmonid fry stranding mortality

associated with daily water level fluctuations in Trail Bridge Reservoir, Oregon. North

American Journal of Fisheries Management 28: 1515–1528. doi:10.1577/M07-026.1.

Bendall, B., Moore, A., Maxwell, D., Davidson, P., Edmonds, N., Archer, D., Solomon, D.,

Greest, V., Wyatt, R., and Broad, K. 2012. Modelling the migratory behaviour of

salmonids in relation to environmental and physiological parameters using telemetry

data. Fisheries Management and Ecology 19: 475–483. doi:10.1111/j.1365-

2400.2011.00811.x.

111

Bennett, T.R., Roni, P., Denton, K., McHenry, M., and Moses, R. 2015. Nomads no more: early

juvenile coho salmon migrants contribute to the adult return. Ecology of Freshwater Fish

24: 264–275.

Benstead, J.P., March, J.G., Pringle, C.M., and Scatena, F.N. 1999. Effects of a low-head dam

and water abstraction on migratory tropical stream biota. Ecological Applications 9(2):

656–668.

Bernhardt, E.S., Palmer, M.A., Allan, J.D., Alexander, G., Barnas, K., Brooks, S., Carr, J.,

Clayton, S., Dahm, C., Follstad-Shah, J., Galat, D., Gloss, S., Goodwin, P., Hart, D.,

Hassett, B., Jenkinson, R., Katz, S., Kondolf, G.M., Lake, P.S., Lave, R., Meyer, J.L.,

O’Donnell, T.K., Pagano, L., Powell, B., and Sudduth, E. 2005. Synthesizing U.S. river

restoration efforts. Science 308(5722): 636–637.

Beverton, R.J.H., and Holt, S.J. 1957. On the dynamics of exploited fish populations.

Bilotta, G.S., Burnside, N.G., Gray, J.C., and Orr, H.G. 2016. The effects of run-of-river

hydroelectric power schemes on fish community composition in temperate streams and

rivers. Plos One: 1–15. doi:10.1371/journal.pone.0154271.

Bisson, P.A., Gregory, S.V., Nickelson, T.E., and Hall, J.D. 2008. Chapter 15: the Alsea

watershed study: a comparison with other multi-year investigations in the Pacific

Northwest. In Hydrological and biological responses to forest practices, Stednick, J.D.

pp. 259–289.

Bonneau, J.L., and Scarnecchia, D.L. 1998. Seasonal and diel changes in habitat use by juvenile

bull trout (Salvelinus confluentus) and cutthroat trout (Oncorhynchus clarki) in a

mountain stream. Canadian Journal of Zoology 76: 783–790.

Booth, D.B., Scholz, J.G., Beechie, T.J., and Ralph, S.C. 2016. Integrating limiting-factors

analysis with process-based restoration to improve recovery of endangered salmonids in

the Pacific Northwest, USA. Water 8(174). doi:doi:10.3390/w8050174.

Boubée, J.A.T., and Williams, E.K. 2006. Downstream passage of silver eels at a small

hydroelectric facility. Fisheries Management and Ecology 13: 165–176.

Bradford, M.J. 1994. Trends in the abundance of Chinook salmon (Onchorrhunchus

tshawytscha) of the Nechako River, British Columbia. Canadian Journal of Fisheries and

Aquatic Science 51: 965–973.

Bradford, M.J. 1995. Comparative review of Pacific salmon survival rates. Canadian Journal of

Fisheries and Aquatic Science 52: 1327–1338.

Bradford, M.J. 2017. Accounting for uncertainty and time lags in equivalency calculations for

offsetting in aquatic resources management programs. Environmental Management

60(4): 588–597. doi:10.1007/s00267-017-0892-6.

Bradford, M.J., and Heinonen, J.S. 2008. Low flows, instream flow needs and fish ecology in

small streams. Canadian Water Resources Journal 33(2): 165–180.

doi:10.4296/cwrj3302165.

Bradford, M.J., Korman, J., and Higgins, P.S. 2005. Using confidence intervals to estimate the

response of salmon populations (Oncorhynchus spp.) to experimental habitat alterations.

Canadian Journal of Fisheries and Aquatic Science 62: 2716–2726.

Bradford, M.J., Myers, R.A., and Irvine, J.R. 2000. Reference points for coho salmon

(Oncorhynchus kisutch) harvest rates and escapement goals based on freshwater

production. Canadian Journal of Fisheries and Aquatic Science 57: 677–686.

112

Bradford, M.J., Taylor, G.C., Allan, J.A., and Higgins, P.S. 1995. An experimental study of the

stranding of juvenile Coho salmon and rainbow trout during rapid flow decreases under

winter conditions. North American Journal of Fisheries Management 15: 473–479.

Brett, J.R. 1995. Energetics. In Physiological Ecology of Pacific Salmon, Groot, C., L. Margolis,

and W.C. Clarke. in cooperation with the Government of Canada, Department of

Fisheries and Oceans, UBC Press, Vancouver. pp. 3–68.

Bryant, M.D. 2009. Global climate change and potential effects on Pacific salmonids in

freshwater ecosystems of southeast Alaska. Climatic Change 95: 169–193.

doi:10.1007/s10584-008-9530-x.

Budy, P., and Luecke, C. 2014. Understanding how lake populations of arctic char are structured

and function with special consideration of the potential effets of climate change: a multi-

faceted approach. Oecologia 176: 81–94. doi:10.1007/s00442-014-2993-8.

Bull, J.W., Suttle, K.B., Gordon, A., Singh, N.J., and Milner-Gulland, E.J. 2013. Biodiversity

offsets in theory and practice. Fauna & Flore International 47(3): 369–380.

Bunn, S.E., and Arthington, A.H. 2002. Basic principles and ecological consequences of altered

flow regimes for aquatic biodiversity. Environmental Management 30(4): 492–507.

doi:10.1007/s00267-002-2737-0.

Butler, S.E., and Wahl, D.H. 2011. Distribution, movements and habitat use of channel catfish in

a river with multiple low-head dams. River Research and Applications 27: 1182–1191.

doi:10.1002/rra.1416.

Čada, G.F. 2001. The development of advanced hydroelectric turbines to improve fish passage

survival. Fisheries 26(9): 14–23. doi:10.1577/1548-

8446(2001)026<0014:TDOAHT>2.0.CO;2.

Capra, H., Sabaton, C., Gouraud, V., Souchon, Y., and Lim, P. 2003. A population dynamics

model and habitat simulation as a tool to predict brown trout demography in natural and

bypassed stream reaches. River Research and Applications 19: 551–568.

Casas-Mulet, R., Salveit, S.J., and Alfredsen, K. 2014. The survival of Atlantic salmon (Salmo

salar) eggs during dewatering in a river subjected to hydropeaking. River Research and

Applications. doi:10.1002/rra.2827.

Cattanéo, F., Lamouroux, N., Breil, P., and Capra, H. 2002. The influence of hydrological and

biotic processess on brown trout (Salmo trutta) population dynamics. Canadian Journal of

Fisheries and Aquatic Science 59(1): 12–22. doi:10.1139/F01-186.

Chapman, D.W. 1962. Aggressive behavior in juvenile Coho salmon as a cause of emigration.

Journal Fisheries Research Borad of Canada 19(6): 1047–1080.

Chipps, S.R., and Wahl, D.H. 2008. Bioenergetics modeling in the 21st century: reviewing new

insights and revisiting old contraints. Transactions of the American Fisheries Society

137(1): 298–313. doi:10.1577/T05-236.1.

Chittaro, P.M., Zabel, R.W., Haught, K., Sanderson, B.L., and Kennedy, B.P. 2014. Spatial and

temporal patterns of growth and consumption by juvenile spring/summer Chinook

salmon Oncorhynchus tshawytscha. Environmental Biology of Fishes 97: 1397–1409.

doi:10.1007/s10641-014-0230-2.

Chittenden, C.M., Melnychuk, M.C., Welch, D.W., and McKinley, R.S. 2010. An investigation

into the poor survival of an endangered coho salmon population. PLoS ONE 5(5):

e10869. doi:10.1371/journal.pone.0010869.

113

Chiyembekezo, S.K. 2013. Energy situation, potential and application status of small-scale

hydropower systems in Malawi. Renewable and Sustainable Energy Reviews 26: 1–19.

doi:10.1016/j.rser.2013.05.034.

Cochrane, J.F., Londorf, E., Allison, T.D., and Sanders-Reed, C.A. 2015. Modeling with

uncertain science: estimating mitigation credits from abating lead poisoning in Golden

Eagles. Ecological Applications 25(6): 1518–1533.

Connors, B.M., Marmorek, D.R., Olson, E., Hall, A.W., de la Cueva Bueno, P., Bensen, A.,

Bryan, K., Perrin, C., Parkinson, E., Abraham, D., Alexander, C., Murray, C., Smith, R.,

Grieg, L., and Farrell, G. 2014. Independent review of run-of-river hydroelectric projects

and their impacts on salmonid species in British Columbia. Pacific Salmon Foundation.

Available from http://www.psf.ca/programs/independent-review-run-of-river-hydro-

projects.

Coulter, D.P., Sepúlveda, M.S., Troy, C.D., and Höök, T.O. 2016. Species-specific effects of

subdaily temperature fluctuations on consumption, growth and stress responses in two

physiologically similar fish species. Ecology of Freshwater Fish 25: 465–475.

doi:10.1111/eff.12227.

Couto, T.B., and Olden, J.D. 2018. Global proliferation of small hydropower plants -- science

and policy. Frontiers in Ecology and Environment. doi:10.1002/fee.1746.

Crone, R.A., and Bond, C.E. 1976. Life history of coho salmon, Oncorhynchus kisutch, in Sashin

Creek, Southeastern Alaska. Fisheries Bulletin 74: 897–923.

Crozier, L.G., Hendry, A.P., Lawson, P.W., Quinn, T.P., Mantua, N.J., Battin, J., Shaw, R.G.,

and Huey, R.B. 2008a. Potential responses to climate change in organisms with complex

life histories: evolution and plasticity in Pacific salmon. Evolutionary Applications 1:

252–270. doi:10.1111/j.1752-4571.2008.00033.x.

Crozier, L.G., McClure, M.M., Beechie, T., Bograd, S.J., Boughton, D.A., Carr, M., Cooney,

T.D., Dunham, J.B., Greene, C.M., Haltuch, M.A., Hazen, E.L., Holzer, D.M., Huff,

D.D., Johnson, R.C., Jordan, C.E., Kaplan, I.C., Lindley, S.T., Mantua, N.J., Moyle, P.B.,

Myers, J.M., Nelson, M.W., Spence, B.C., Weitkamp, L.A., Williams, T.H., and Willis-

Norton, E. 2019. Climate vulnerability assessment for Pacific salmon and steelhead in the

California Current Large Marine Ecosystem. PLoS ONE: 1–49.

doi:10.1371/journal.pone.0217711.

Crozier, L.G., Zabel, R.W., and Hamlet, A.F. 2008b. Predicting differential effects of climate

change at the population level with life-cycle models of spring Chinook salmon. Global

Change Biology 14: 236–249.

Crozier, L.G., Zabel, R.W., Hockersmith, E.E., and Achord, S. 2010. Interacting effects of

density and temperature on body size in multiple populations of Chinook salmon. Journal

of Animal Ecology 79: 342–349. doi:10.1111/j.1365-2656.2009.01641.x.

Csiki, S.J.C., and Rhoads, B.L. 2010. Hydraulic and geomorphological effects of run-of-river

dams. Progress in Physical Geography 34(6): 755–780. doi:10.1177/0309133310369435.

Csiki, S.J.C., and Rhoads, B.L. 2013. Influence of four run-of-river dams on channel

morphology and sediment characteristics in Illinois, U.S.A. Geomorphology.

doi:10.1016/j.geomorph.2013.10.009.

Cunningham, C.J., Westley, P.A.H., and Adkison, M.D. 2018. Signals of large scale climate

drivers, hatchery enhancement, and marine factors in Yukon River Chinook salmon

survival revealed with a Bayesian life history model. Global Change Biology 24: 4399–

4416. doi:10.1111/gcb.14315.

114

Cyr, J.-F., Landry, M., and Gagnon, Y. 2011. Methodology for the large-scale assessment of

small hydroelectric potential: application to the province of New Brunswick (Canada).

Renewable Eergy 36: 2940–2950. doi:10.1016/j.renene.2011.04.003.

Dibble, K.L., Yackulic, C.B., Kennedy, T.A., and Budy, P. 2015. Flow management and fish

density regulate salmonid recruitment and adult size in tailwaters across western North

America. Ecological Applications 25(8): 2168–2179.

Doyle, M.W., and Shields, F.D. 2012. Compensation mitigation for streams under the Clean

Water Act: reassessing science and redirecting policy. Journal of the American Water

Resources Association 48(3): 494–509. doi:10.1111/j.1752-1688.2011.00631.x.

Dudgeon, D. 2010. Prospects for sustaining freshwater biodiversity in the 21st century: linking

ecosystem structure and function. Environmental Sustainability 2: 422–430.

Einum, S., Sundt-Hansen, L., and Nislow, K.H. 2006. The partitioning of density-dependent

dispersal, growth and survival throughout ontogeny in a highly fecund organism. Oikos

113: 489–496.

Elder, T., Woodley, C.M., Weiland, M.A., and Strecker, A.L. 2016. Factors influencing the

survival of outmigration juvenile salmonids through multiple dam passages: an

individual-based approach. Ecology and Evolution 6(16): 5881–5892.

doi:10.1002/ece3.2326.

Englund, G., and Malmqvist, B. 1996. Effects of flow regulation, habitat area and isolation on

the macroinvertebrates fauna of rapids in north Swedish rivers. Regulated Rivers:

Research & Management 12: 433–445.

Evans, A., Strezov, V., and Evans, T.J. 2009. Assessment of sustainability indicators for

renewable energy technologies. Renewable and Sustainable Energy Reviews 13: 1082–

1088. doi:https://doi.org/10.1016/j.rser.2008.03.008.

Farrell, A.P., Hinch, S.G., Cooke, S.J., Patterson, D.A., Crossin, G.T., Lapointe, M., and Mathes,

M.T. 2008. Pacific Salmon in hot water: applying aerobic scope models and biotelemetry

to predict the success of spawning migrations. Physiological and Biochemical Zoology

81(6): 697–709. doi:10.1086/592057.

Finstad, A.G., Forseth, T., Jonsson, B., Bellier, E., Hesthagen, T., Jensen, A.J., Hessen, D.O.,

and Foldvik, A. 2011. Competitive exclusion along climate gradients: energy efficiency

influences the distribution of two salmonid fishes. Global Change Biology 17: 1703–

1711. doi:10.1111/j.1365-2486.2010.02335.x.

Fodrie, F.J., Able, K.W., Galvez, F., Heck, K.L.Jr., Jensen, O.P., Lόpez-Duarte, P.C., Martin,

C.W., Turner, R.E., and Whitehead, A. 2014. Integrating organismal and population

responses of estuarine fishes in Macondo Spill Research. BioScience 64(9): 778–788.

Freckleton, R.P., Silva Matos, D.M., Bovi, M.L.A., and Watkinson, A.R. 2003. Predicting the

impacts of harvesting using structured population models: the importance of density-

dependence and timing of harvest for a tropical palm tree. Journal of Applied Ecology

40: 846–858.

Freeman, M.C., Bowen, Z.H., Bovee, K.D., and Irwin, E.R. 2001. Flow and habitat effects on

juvenile fish abundance in natural and altered flow regimes. Ecological Applications

13(1): 179–190.

Fu, X.C., Tao, T., Wanxiang, J., Fengqing, L., Naicheng, W., Shuchan, Z., and Qinghua, C.

2008. Impacts of small hydropower plants on macroinvertebrate communities. Acta

Ecologica Sinica 28(1): 45–52.

115

Fuller, T.K., Venditti, J.G., Nelson, P.A., and Palen, W.J. 2016. Modeling grain size adjustments

in the downstream reach following run-of-river development. Water Resources Research

52. doi:10.1002/2015WR017992.

Gardner, T.A., Von Hase, A., Brownlie, S., Ekstrom, J.M.M., Pilgrim, J.D., Savy, C.E.,

Stephens, R.T.T., Treweek, J., Ussher, G.T., Ward, G., and Ten Kate, K. 2013.

Biodiversity offsets and the challenge of achieving no net loss. Conservation Biology

27(6): 1254–1264. doi:10.1111/cobi.12118.

Geist, D.R., Arntzen, E.V., Murray, C.J., McGrath, K.E., Bott, Y.-J., and Hanrahan, T.P. 2008.

Influence of river level on temperature and hydraulic gradients in chum and fall Chinook

salmon spawning areas downstream of Bonneville Dam, Columbia River. North

American Journal of Fisheries Management 27: 30–41. doi:10.1577/M07-009.1.

Gende, S.M., Edwards, R.T., Willson, M.F., and Wipfli, M.S. 2002. Pacific salmon in aquatic

and terrestrial ecosystems. BioScience 52(10): 917–928.

Gibeau, P., Connors, B.M., and Palen, W.J. 2017. Run-of-River hydropower and salmonids:

potential effects and perspective on future research. Canadian Journal of Fisheries and

Aquatic Science 74: 1135–1149. doi:10.1139/cjfas-2016-0253.

Goodyear, C.P. 1977. Assessing the impact of power plant mortality on the compensatory

reserve of fish populations. Proceedings of the Conference on Assessing the Effects of

Power-Plant-Induced Mortality on Fish, pages 186-195. Populations. Available from

https://www.researchgate.net/profile/C_Goodyear/publication/245508321_Assessing_the

_Impact_of_Power_Plant_Mortality_on_the_Compensatory_Reserve_of_Fish_Populatio

ns/links/5a7c7d67aca272669a2ce6b5/Assessing-the-Impact-of-Power-Plant-Mortality-

on-the-Compensatory-Reserve-of-Fish-Populations.pdf [accessed 21 August 2019].

Goodwin, N.B., Grant, A., Perry, A.L., Dulvy, N.K., and Reynolds, J.D. 2006. LIfe history

correlates of density-dependent recruitment in marine fishes. Canadian Journal of

Fisheries and Aquatic Science 63: 494–509.

Gouraud, V., Capra, H., Sabaton, C., Tissot, L., Lim, P., Vandewalle, F., Fahrner, G., and

Souchon, Y. 2008. Long-term simulations of the dynamics of trout populations on river

reaches bypassed by hydroelectric installations - analysis of the impact of different

hydrological scenarios. River Research and Applications 24: 1185–1205.

doi:10.1002/rra.1129.

Greene, C.M., and Beechie, T.J. 2004. Consequences of potential density-dependent mechanisms

on recovery of ocean-type chinook salmon (Oncorhynchus tshawytscha). Canadian

Journal of Fisheries and Aquatic Science 61: 590–602.

Groot, C., Margolis, L., and Clarke, W.C. 1995. Physiological ecology of Pacific salmon. In

UBC Press, Vancouver.

Grumbine, R.E., and Pandit, M.K. 2013. Threats from India’s Himalaya dams. Science 339: 36–

37.

Gustafson, R.G., Waples, R.S., Myers, J.M., Weitkamp, L.A., Bryant, G.J., Johnson, O.W., and

Hard, J.J. 2007. Pacific Salmon extinctions: quantifying lost and remaining diversity.

Conservation Biology 21(4): 1009–1020.

Habit, E., Belk, M.C., and Parra, O. 2007. Response of the riverine fish community to the

construction and operation of a diversion hydropower plant in central Chile. Aquatic

Conservation: Marine and Freshwater Ecosystems 17: 37–49. doi:10.1002/aqc.774.

116

Hall, J.D. 2008. Chapter 5: Salmonid populations and habitat. In Hydrological and biological

responses to forest practices; the Alsea watershed study, John D. Stednick. Springer. pp.

78–105.

Halleraker, J.H., Salveit, S.J., Harby, A., Arnekleiv, J.V., Fjeldstad, H.-P., and Kohler, B. 2003.

Factors influencing stranding of wild juvenile brown trout (Salmo trutta) during rapid and

frequent flow decreases in an artificial stream. River Research and Applications 19: 589–

603.

Hansen, M.J., Boisclair, D., Brandt, S.B., Hewett, S.W., Kitchell, J.F., Lucas, M.C., and Ney, J.J.

1993. Applications of bioenergetics models to fish ecology and management: where do

we go from here. Transactions of the American Fisheries Society 122: 1019–1030.

Hanson, P.C., Johnson, T.B., Schindler, D.E., and Kitchell, J.F. 1997. Fish bioenergetics 3.0.

Technical report WISCU-T-97-001, University of Wisconsin Sea Grant Institute,

Madison, WI.

Harby, A., and Halleraker, J.H. 2001. Ecological impacts of hydropeaking in rivers. Hydropower

and Dams (4): 132–135.

Harper, D.J., and Quigley, J.T. 2005. No net loss of fish habitat: a review and analysis of habitat

compensation in Canada. Environmental Management 36: 343–355.

Hartman, G.F., Scrivener, J.C., and Miles, M.J. 1996. Impacts of logging in Carnation Creek, a

high-energy coastal stream in British Columbia, and their implication for restoring fish

habitat. Canadian Journal of Fisheries and Aquatic Science 53 (suppl. 1): 237:251.

Hartman, K.J., and Kitchell, J.F. 2008. Bioenergetics modeling: progress since the 1992

Symposium. Transactions of the American Fisheries Society 137(1): 216–223.

doi:10.1577/T07-040.1.

Harvey, B.C., Nakamoto, R.J., and White, J.L. 2006. Reduced streamflow lowers dry-season

growth of rainbow trout in a small stream. Transactions of the American Fisheries

Society 135(4): 998–1005. doi:10.1577/T05-233.1.

Harvey, B.C., White, J.L., and Nakamoto, R.J. 2009. The effect of deposited fine sediment on

summer survival and growth of rainbow trout in riffles of a small stream. North

American Journal of Fisheries Management 29(2): 434–440. doi:10.1577/M08-074.1.

Hayes, D., Jones, M., Lester, N., Chu, C., Doka, S., Netto, J., Stockwell, J., Thompson, B.,

Minns, C.K., Shuter, B., and Collins, N. 2009. Linking fish population dynamics to

habitat conditions: insights from the application of a process-oriented approach to several

Great Lakes species. Reviews in Fish Biology and Fisheries 19: 295–312.

Heggenes, J., and Røed, K.H. 2006. Do dams increase genetic diversity in brown trout (Salmo

trutta)? Microgeographic differentiation in a fragmented river. Ecology of Freshwater

Fish 15: 366–375. doi:10.1111/j.1600-0633.2006.00146.x.

Hilborn, R., and Walters, C.J. 1992. Quantitative fisheries stock assessment: choice, dynamics

and uncertainty.

Hodgson, E.E., Essington, T.E., and Halpern, B.S. 2017. Density dependence governs when

population responses to multiple stressors are magnified or mitigated. Ecology 98(10):

2673–2683.

Holtby, L.B. 1988. Effects of logging on stream temperatures in Carnation Creek, British

Columbia, and associated impacts on the Coho salmon (Oncorhynchus kisutch).

Canadian Journal of Fisheries and Aquatic Science 45: 502–515.

Holtby, L.B., and Healey, M.C. 1990. Sex-specific life history tactics and risk-taking in coho

salmon. Ecology 71(2): 678–690.

117

Honea, J.M., McClure, M.M., Jorgensen, J.C., and Scheuerell, M.D. 2016. Assessing freshwater

life-stage vulnerability of an endangered Chinook salmon pop to climate change

influences on stream habitat. Climate Research 71: 127–137.

Hough, P., and Robertson, M. 2009. Mitigation under Section 404 of the Clean Water Act: where

it comes from, what it means. Wetlands Ecological Management 17: 15–33.

Hunter, M.A. 1992. Hydropower flow fluctuations and salmonids: a review of the biological

effects, mechanical causes, and options for mitigation. Technical report prepared for the

State of Washington Department of Fisheries, Number 119. Olympia, Washington

State,USA. 58 pages.

Hvidsten, N.A. 1985. Mortality of pre-smolt Atlantic salmon, Salmo salar L., and brown trout,

Salmo trutta L., caused by fluctuating water levels in the regulated River Nidelva, central

Norway. Journal of Fish Biology 27: 711–718.

Imholt, C., Malcolm, I.A., Bacon, P.J., Gibbins, C.N., Soulsby, C., Miles, M., and Fryer, R.J.

2011. Does diurnal temperature variability affect growth in juvenile Atlantic salmon

Salmo salar? Journal of Fish Biology 78: 436–448. doi:10.1111/j.1095-

8649.2010.02838.x.

Isaak, D.J., Young, M.K., Luce, C.H., Hostetler, S.W., Wenger, S.J., Peterson, E.E., ver Hoef,

J.M., Groce, M.C., Horan, D.L., and Nagel, D.E. 2016. Slow climate velocities of

moutain streams portend their role as refugia for cold-water biodiversity. PNAS 113(16):

4374–4379. doi:10.1073/pnas.1522429113.

Jaccard, M., Melton, N., and Nyboer, J. 2011. Institutions and processes for scaling up

renewables: Run-of-river hydropower in British Columbia. Energy Policy 39: 4042–

4050.

Jackson, R.B., Carpenter, S.R., Dahm, C.N., McKnight, D.M., Naiman, R.J., Postel, S.L., and

Running, S.W. 2001. Water in a changing world. Ecological Applications 11(4): 1027–

1045.

Jelks, H.L., Walsh, S.J., Burkhead, N.M., Contreras-Balderas, S., Dίaz-Pardo, E., Hendrickson,

D.A., Mandrak, N.E., McCormick, F., Nelson, J.S., Platania, S.P., Porter, B.A., Renaud,

C.B., Schmitter-Soto, J.J., Taylor, E.B., and Warren, M.L.Jr. 2008. Conservation status of

imperiled North American freshwater and diadromous fishes. Fisheries 33(8): 372–407.

Jenkins, A.R., and Keeley, E.R. 2010. Bioenergetic assessment of habitat quality for stream-

dwelling cutthroat trout (Oncorhynchus clarkii bouvieri) with implications for climate

change and nutrient supplementation. Canadian Journal of Fisheries and Aquatic Science

67: 371–385. doi:10.1139/F09-193.

Jenkins, T.M., Diehl, S., Kratz, K.W., and Cooper, S.D. 1999. Effects of population density on

individual growth of brown trout in streams. Ecology 80(3): 941–956.

Jesus, T., Formigo, N., Santos, P., and Tavares, G.R. 2004. Impact evaluation of the Vila Viçosa

small hydroelectric power plant (Portugal) on the water quality and on the dynamics of

the benthic macroinvertebrate communities of the Ardena river. Limnetica 23(3–4): 241–

256.

Jowett, I.G., and Biggs, B.J.F. 2006. Flow regime requirements and the biological effectiveness

of habitat-based minimum flow assessments for six rivers. International Journal of River

Basin Management 4(3): 179–189.

Jowett, I.G., Richardson, J., and Bonnett, M.L. 2005. Relationship between flow regime and fish

abundances in a gravel-bed river, New Zealand. Journal of Fish Biology 66: 1419–1436.

doi:10.1111/j.1095-8649.2005.00693.x.

118

Kareiva, P., Marvier, M., and McClure, M. 2000. Recovery and management options for

Spring/Summer Chinook salmon in the Columbia River Basin. Science 290: 977–981.

Kawaguchi, Y., and Nakano, S. 2001. Contribution of terrestrial invertebrates to the annual

resource budget for salmonids in forest and grassland reaches of a headwater stream.

Freshwater Biology 46: 303–316.

Keeley, E.R. 2001. Demographic responses to food and space competition by juvenile steelhead

trout. Ecology 82(5): 1247–1259.

Keeley, E.R., Slaney, P.A., and Zaldokas, D. 1996. Estimates of production benefits for salmonid

fishes from stream restoration initiatives. Province of British Columbia, Ministry of

Environment, Lands and Parks, and Ministry of Forests. Watershed Restoration

Management Report4: 22p.

Kerkhoven, E., and Yew Gan, T. 2011. Differences and sensitivities in potential hydrologic

impact of climate change to regional-scale Athabasca and Fraser basins of the leeward

and winward sides of the Canadian Rocky Mountains respectively. Climatic Change 106:

583–607. doi:10.1007/s10584-010-9958-7.

Kibler, K.M. 2011. Development and decommissioning of small dams: analysis of impact and

context. PhD dissertation, Water Resources Engineering, Oregon State University,

Corvalis, Oregon, United States of America.

Kibler, K.M., and Tullos, D.D. 2013. Cumulative biophysical impact of small and large

hydropower development in Nu River, China. Water Resources Research 49: 3104–3118.

doi:10.1002/wrcr.20243.

Kondolf, G.M. 1997. Hungry water: effects of dams and gravel mining on river channels.

Environmental Management 21(4): 533–551.

Korman, J., and Campana, S.E. 2009. Effects of hydropeaking on nearshore habitat use and

growth of age-0 rainbow trout in a large regulated river. Transactions of the American

Fisheries Society 138(1): 76–87.

Korman, J., Sawada, J., and Bradford, M.J. 2019. Evaluation framework for assessing potential

Pacific Salmon Commission reference points for population status and associated

allowable exploitation rates for Strait of Georgia and Fraser River Coho Salmon

Management Units. Canadian Science Advisory Secretariat, Research Document

2019/001.

Korman, J., and Tompkins, A. 2014. Estimating regional distributions of freshwater stock

productivity, carrying capacity, and sustainable harvest rates for Coho salmon using a

hierarchical Bayesian modelling approach. Canadian Science Advisory Secretariat,

Research Document 2014/089. Available from http://www.dfo-mpo.gc.ca/csas-sccs/ csas-

sccs@dfo-mpo.gc.ca.

Korman, J., Walters, C.J., Martell, S.J.D., Pine III, W.E., and Dutterer, A. 2011. Effects of flow

fluctuations on habitat use and survival of age-0 rainbow trout (Oncorhunchus mykiss) in

a large, regulated river. Canadian Journal of Fisheries and Aquatic Science 68: 1097–

1109. doi:10.1139/F2011-045.

Kubečka, J., Matĕna, J., and Hartvich, P. 1997. Adverse ecological effects of small hydropower

stations in the Czech Republic: 1. bypass plants. Regulated Rivers: Research &

Management 13: 101–113.

Lake, P.S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North

American Benthological Society 19(4): 573–592.

119

Lande, R., Engen, S., and Sӕther, B.-E. 2003. Stochastic population dynamics in ecology and

conservation. In Oxford University Press. Oxford.

Larinier, M. 2008. Fish passage experience at small-scale hydro-electric power plants in France.

Hydrobiologia 609: 97–108. doi:10.1007/s10750-008-9398-9.

Leach, J.A., Moore, R.D., Hinch, S.G., and Gomi, T. 2012. Estimation of forest harvesting-

induced stream temperature changes and bioenergetic consequences for cutthroat trout in

a coastal stream in British Columbia, Canada. Aquatic Sciences 74: 427–441.

doi:10.1007/s00027-011-0238-z.

Levin, P.S., and Tolimieri, N. 2001. Differences in the impacts of dams on the dynamics of

salmon populations. Animal Conservation 4: 291–299.

Lobon-Cervia, J. 2009. Why, when and how do fish populations decline, collapse and recover?

The example of brown trout (Salmo trutta) in Rio Chaballos (northwestern Spain).

Freshwater Biology 54: 1149–1162.

Lucas, M.C., Bubb, D.H., Jang, M., Ha, K., and Masters, J.E.G. 2009. Availability of and access

to critical habitats in regulated rivers: effects of low-head barriers on threatened

lampreys. Freshwater Biology 54: 621–634. doi:10.1111/j.1365-2427.2008.02136.x.

Lytle, D.A., and Poff, N.L. 2004. Adaptation to natural flow regimes. Trends in Ecology &

Evolution 19(2): 94–100. doi:10.1016/j.tree.2003.10.002.

Malcolm, I.A., Gibbins, C.N., Soulsby, C., Tetzlaff, D., and Moir, H.J. 2012. The influence of

hydrology and hydraulics on salmonids between spawning and emergence: implications

for the management of flows in regulated rivers. Fisheries Management and Ecology 19:

464–474. doi:10.1111/j.1365-2400.2011.00836.x.

Mantua, N., Tohver, I., and Hamlet, A. 2010. Climate change impacts on streamflow extremes

and summertime stream temperature and their possible consequences for freshwater

salmon habitat in Washington State. Climatic Change 10(102): 187–223.

doi:10.1007/s10584-010-9845-2.

Maron, M., Ives, C.D., Kujala, H., Bull, J.W., Maseyk, F.J., Bekessy, S., Gordon, A., Watson,

J.E.M., Lentini, P.E., Gibbons, P., Possingham, H.P., Hobbs, R.J., Keith, D.A., Wintle,

B.A., and Evans, M.C. 2016. Taming a wicked problem: resolving controversies in

biodiversity offsetting. BioScience 66(6): 489–498.

Matthews, W.J., and Marsh-Matthews, E. 2003. Effects of droughts on fish across axes of space,

time and ecological complexity. Freshwater Biology 48: 1232–1253.

McCarthy, M.A. 1996. Red kangaroo (acropus rufus) dynamics: effects of rainfall, density

dependence, harvesting and environmental stochasticity. Journal of Applied Ecology

33(1): 45–53.

McKenney, B., and Kiesecker, J.M. 2010. Policy development for biodiversity offsets: a review

of offset frameworks. Environmental Management 45: 165–176.

McManamay, R.A., Peoples, B.K., Orth, D.J., Dolloff, C.A., and Matthews, D.C. 2015. Isolating

causal pathways between flow and fish in the regulated river hierarchy. Canadian Journal

of Fisheries and Aquatic Science 72: 1731–1748. doi:10.1139/cjfas-2015-0227.

McPhail, J.D. 2007. The freshwater fishes of British Columbia. The University of Alberta Press,

Edmonton, Alberta.

Milner, N.J., Elliott, J.M., Armstrong, J.D., Gardiner, R., Welton, J.S., and Ladle, M. 2003. The

natural control of salmon and trout populations in streams. Fisheries Research 62: 111–

125.

120

Minns, C.K. 1997. Quantifying “no net loss” of productivity of fish habitats. Canadian Journal of

Fisheries and Aquatic Science 54(10): 2463–2473. doi:10.1139/cjfas-54-10-2463.

Minns, C.K., Randall, R.G., Smokorowski, K.E., Clarke, K.D., Vélez-Espino, A., Gregory, R.S.,

Courtenay, S., and LeBlanc, P. 2011. Direct and indirect estimates of the productive

capacity of fish habitat under Canada’s Policy for the Management of Fish Habitat:

where have we been, where are we now, and where are we going? Canadian Journal of

Fisheries and Aquatic Science 68: 2204–2227. doi:10.1139/F2011-130.

Mitro, M.G., and Zale, A.V. 2002. Seasonal survival, movement, and habitat use of age-0

rainbow trout in the Henrys Fork of the Snake River, Idaho. Transactions of the

American Fisheries Society 131: 271–286.

Moore, J.W., Beakes, M.P., Nesbitt, H.K., Yeakel, J.D., Patterson, D.A., Thompson, L.A.,

Phillis, C.C., Braun, D.C., Favaro, C., Scott, D., Carr-Harris, C., and Atlas, W.I. 2015.

Emergent stability in a large, free-flowing watershed. Ecology 96(2): 340–347.

Moore, K.D., and Moore, J.W. 2013. Ecological restoration and enabling behavior: a new

metaphorical lens? Conservation Letters 6: 1–5. doi:10.1111/j.1755-263X.2012.00300.x.

Morley, S.A., Garcia, P.S., Bennett, T.R., and Roni, P. 2005. Juvenile salmonid (Oncorhynchus

spp.) use of constructed and natural side channels in Pacific Northwest rivers. Canadian

Journal of Fisheries and Aquatic Science 62: 2811–2821. doi:10.1139/F05-185.

Morris, S.E. 1992. Geomorphic assessment of the effects of flow diversion on anadromous fish

spawning habitat: Newhalem Creek, Washington. Professional Geographer 44(4): 444–

452.

Mortensen, E. 1977. Density-dependent mortality of trout fry (Salmo trutta L.) and its

relationship to the management of small streams. Journal of Fish Biology 11: 613–617.

Mueller, M., Pander, J., and Geist, J. 2011. The effects of weirs on structural stream habitat and

biological communities. Journal of Applied Ecology 48: 1450–1461. doi:10.1111/j.1365-

2664.2011.02035.x.

Muir, W.D., Smith, S.G., Williams, J.G., and Sandford, B.P. 2001. Survival of juvenile

salmonids passing through bypass systems, turbines and spillways with and without flow

deflectors at Snake River Dams. North American Journal of Fisheries Management 21:

135–146.

Murchie, K.J., Hair, K.P.E., Pullen, C.E., Redpath, T.D., Stephens, H.R., and Cooke, S.J. 2008.

Fish response to modified flow regimes in regulated rivers: research methods, effects and

opportunities. River Research and Applications 24: 197–217. doi:10.1002/rra.1058.

Myers, R.A. 2002. Recruitment: understanding density-dependence in fish population. In

Handbook of fish biology and fisheries, Vol.1: Fish Biology, Hart, P., Reynolds J. eds.

Oxford: Blackwell Publishing. pp. 123–148.

Myrvold, K.M., and Kennedy, B.P. 2015a. Interactions between body mass and water

temperature cause energetic bottlenecks in juvenile steelhead. Ecology of Freshwater

Fish 24: 373–383. doi:10.1111/eff.12151.

Myrvold, K.M., and Kennedy, B.P. 2015b. Density dependence and its impact on individual

growth rates in an age-structured stream salmoni population. Ecosphere 6(12): 281.

doi:http://dx.doi.org/10.1890/ES15-00390.1.

Myrvold, K.M., and Kennedy, B.P. 2018. Increasing water temperatures exacerbate the potential

for density dependence in juvenile steelhead. Canadian Journal of Fisheries and Aquatic

Science 75: 897–907. doi:10.1139/cjfas-2016-0497.

121

Nagrodski, A., Raby, G.D., Hasler, C.T., Taylor, M.K., and Cooke, S.J. 2012. Fish stranding in

freshwater systems: sources, consequences, and mitigation. Journal of Environmental

Management 103: 133–141. doi:10.1016/j.jenvman.2012.03.007.

Naiman, R.J., Bilby, R.E., and Bisson, P.A. 2000. Riparian ecology and management in the

Pacific Coastal Rain Forest. BioScience 50(11): 996–1011.

Naiman, R.J., Bilby, R.E., Schindler, D.E., and Helfield, J.M. 2002. Pacific salmon, nutrients,

and the dynamics of freshwater and riparian ecosystems. Ecosystems 5: 399–417.

doi:10.1007/s10021-001-0083-3.

Naiman, R.J., Latterell, J.J., Pettit, N.E., and Olden, J.D. 2008. Flow variability and the

biophysical vitality of river systems. Geoscience 340: 629–643.

Nguyen, V.M., Martins, E.G., Robichaud, D., Raby, G.D., Donaldson, M.R., Lotto, A.G.,

Willmore, W.G., Patterson, D.A., Farrell, A.P., Hinch, S.G., and Cooke, S.J. 2014.

Disentangling the roles of air exposure, gill net injury, and facilitated recovery of the

postcapture and release mortality and behavior of adult migratory sockeye salmon

(Oncorhynchus nerka) in freshwater. Physiological and Biochemical Zoology 87(1): 125–

135.

Nickelson, T.E., and Lawson, P.W. 1998. Population viability of coho salmon, Oncorhynchus

kisutch, in Oregon coastal basins: application of a habitat-based life cycle model.

Canadian Journal of Fisheries and Aquatic Science 55: 2383–2392.

Niemi, G.J., DeVore, P., Detenbeck, N., Taylor, D., Lima, A., Pastor, J., Yount, J.D., and

Naiman, R.J. 1990. Overview of case studies on recovery of aquatic systems from

disturbance. Environmental Management 14(5): 571–587.

Nilsson, C., Reidy, C.A., Dynesius, M., and Revenga, C. 2005. Fragmentation and flow

regulation of the world’s large river systems. Science 308: 405–408.

doi:10.1126/science.1107887.

Nislow, K.H., Einum, S., and Folt, C.L. 2004. Testing predictions of the critical period for

survival concept using experiments with stocked Atlantic salmon. Journal of Fish

Biology 65 (Supplement A): 188–200. doi:10.1111/j.1095-8649.2004.00561.x.

Nislow, K.H., and Armstrong, J.D. 2012. Towards a life-history-based management framework

for the effects of flow on juvenile salmonids in streams and rivers. Fisheries Management

and Ecology 19: 451–463. doi:10.1111/j.1365-2400.2011.00810.x.

OECD. 2013. Global energy trends to 2035. World Energy Outlook, pp.55-98.

Ogston, L., Gidora, S., Foy, M., and Rosenfeld, J.S. 2015. Watershed-scale effectiveness of

floodplain habitat restoration for juvenile coho salmon in the Chilliwack River, British

Columbia. Canadian Journal of Fisheries and Aquatic Science 72: 479–490.

doi:dx.doi.org/10.1139/cjfas-2014-0189.

Ohlberger, J., Buehrens, T.W., Brenkman, S.J., Crain, P., Quinn, T.P., and Hilborn, R. 2018.

Effects of past and projected river discharge variability on freshwater production in an

anadromous fish. Freshwater Biology 63: 331–340. doi:10.1111/fwb.13070.

Oosterhout, G.R., Huntington, C.W., Nickelson, T.E., and Lawson, P.W. 2005. Potential benefits

of a conservation hatchery program for supplementing Oregon coast coho salmon

(Oncorhynchus kisutch) populations: a stochastic model investigation. Canadian Journal

of Fisheries and Aquatic Science 62: 1920–1935.

Ovidio, M., Capra, H., and Philippart, J.-C. 2008. Regulated discharge produces substantial

demographic changes on four typical fish species of a small salmonid stream.

Hydrobiologia 609: 59–70. doi:10.1007/s10750-008-9399-8.

122

Petersen, J.H., and Kitchell, J.F. 2001. Climate regimes and water temperature changes in the

Columbia River: bioenergetic implications for predators of juvenile salmon. Canadian

Journal of Fisheries and Aquatic Science 58(9): 1831–1841.

Petersen, J.H., and Paukert, C.P. 2005. Development of a bioenergetics model for humpback

chub and evaluation of water temperature changes in the Grand Canyon, Colorado River.

Transactions of the American Fisheries Society 134(4): 960–974. doi:10.1577/T04-090.1.

Pittock, J., and Hartmann, J. 2011. Taking a second look: climate change, periodic relicensing

and improved management of dams. Marine and Freshwater Research 62: 312–320.

doi:10.1071/MF09302.

Poff, N.L., Allan, D.J., Bain, M.B., Karr, J.R., Prestegaard, K.L., Richter, B.D., Sparks, R.E., and

Stromberg, J.C. 1997. The natural flow regime. BioScience 47(11): 769–784.

Poff, N.L., and Hart, D.D. 2002. How dams vary and why it matters for the emerging science of

dam removal. BioScience 52(8): 659–668. doi:10.1641/0006-

3568(2002)052[0659:HDVAWI]2.0.CO;2.

Poff, N.L., and Matthews, J.H. 2013. Environmental flows in the Anthropocence: past progress

and future prospects. Current Opinion in Environmental Sustainability 5: 667–675.

doi:10.1016/j.cosust.2013.11.006.

Poff, N.L., and Ward, J.V. 1990. Physical habitat template of lotic systems: recovery in the

context of historical pattern of spatiotemporal heterogeneity. Environmental Management

14(5): 629–645.

Poff, N.L., and Zimmerman, J.K.H. 2010. Ecological responses to altered flow regimes: a

literature review to inform the science and management of environmental flows.

Freshwater Biology 55: 194–205. doi:10.1111/j.1365-2427.2009.02272.x.

Poole, G.C., and Berman, C.H. 2001. An ecological perspective on in-stream temperature:

natural heat dynamics and mechanisms of human-caused thermal degradation.

Environmental Management 27(6): 787–802. doi:10.1007/s002670010188.

van Poorten, B.T., and McAdam, S.O. 2010. Estimating differences in growth and metabolism in

two spatially segregated groups of Columbia River white sturgeon using a field-based

bioenergetics model. The Open Fish Science Journal 3: 132–141.

van Poorten, B.T., and Walters, C.J. 2010. Estimation of bioenergetics parameters for rainbow

trout (Oncorhynchus mykiss) using capture-recapture data with comparison to estimates

from a laboratory-based model. The Open Fish Science Journal 3: 69–79.

Postel, S., Daily, G.C., and Ehrlich, P.R. 1996. Human appropriation of renewable fresh water.

Science 271(5250): 785–788.

Power, M. 1997. Assessing the effects of environmental stressors on fish populations. Aquatic

Toxicology 39: 151–169.

Quétier, F., and Lavorel, S. 2011. Assessing ecological equivalence in biodiversity offset

schemeL key issues and solutions. Biological Conservation 144: 2991–2999.

doi:10.1016/j.biocon.2011.09.002.

Quigley, J.T., and Harper, D.J. 2006. Effectiveness of fish habitat compensation in Canada in

achieving No Net Loss. Environmental Management 37(3): 351–366.

R Core Team. 2013. R: a language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, Austria.

Railsback, S.F., and Rose, K.A. 1999. Bioenergetics modeling of stream trout growth:

temperature and food consumption effects. Transactions of the American Fisheries

Society 128(2): 241–256.

123

Rand, P.S., Steward, D.J., Seelbach, P.W., Jones, M.L., and Wedge, L.R. 1993. Modeling

steelhead population energetics in Lakes Michigan and Ontario. Transactions of the

American Fisheries Society 122(5): 977–1001.

Ratner, S., Lande, R., and Roper, B.B. 1997. Population viability analysis of spring Chinook

salmon in the South Umpqua River, Oregon. Conservation Biology 11(4): 879–889.

Reice, S.R., Wissmar, R.C., and Naiman, R.J. 1990. Disturbance regimes, resilience, and

recovery of animal communities and habitats in lotic ecosystems. Environmental

Management 14(5): 647–659.

Renöfält, B.M., Jansson, R., and Nilsson, C. 2010. Effects of hydropower generation and

opportunities for environmental flow management in Swedish riverine ecosystems.

Freshwater Biology 55: 49–67. doi:10.1111/j.1365-2427.2009.02241.x.

Resh, V.H., Brown, A.V., Covich, A.P., Gurtz, M.E., Hiram, W.L., Minshall, G.W., Reice, S.I.,

L., A.L., Bruce, J.B., and Wissmar, R.C. 1988. The role of disturbance in stream ecology.

Journal of the North American Benthological Society 7(4): 433–455.

Richter, B.D., Baumgartner, J.V., Wigington, R., and Braun, D.P. 1997. How much water does a

river need? Freshwater Biology 37: 231–249.

Ricker, W.E. 1992. Back-calculation of fish lengths based on proportionality between scale and

length increments. Canadian Journal of Fisheries and Aquatic Science 49: 1018–1026.

Robson, A., Cowx, I.G., and Harvey, J.P. 2011. Impact of run-of-river hydro-schemes upon fish

populations. Scotland & Northern Ireland Forum for Environmental Research

(SNIFFER).

Roni, P., Hanson, K., and Beechie, T. 2008. Global review of the physical and biological

effectiveness of stream habitat rehabilitation techniques. North American Journal of

Fisheries Management 28: 856–890. doi:10.1577/M06-169.1.

Roni, P., Morley, S.A., Garcia, P., Detrick, C., King, D., and Beamer, E. 2006. Coho salmon

smolt production from constructed and natural floodplain habitats. Transactions of the

American Fisheries Society 135: 1398–1408. doi:10.1577/T05-296.1.

Roni, P., Pess, G., Beechie, T., and Morley, S. 2010. Estimating changes in coho salmon and

steelhead abundance from watershed restoration: how much restoration is needed to

measurably increase smolt production? North American Journal of Fisheries

Management 30: 1469–1484. doi:10.1577/M09-162.1.

Roni, P., and Quinn, T.P. 2001. Density and size of juvenile salmonids in response to placement

of large woody debris in western Oregon and Washington streams. Canadian Journal of

Fisheries and Aquatic Science 58: 282–292. doi:10.1139/cjfas-58-2-282.

Rose, K.A., Cowan, J.H.J., Winemiller, K.O., Myers, R.A., and Hilborn, R. 2001. Compensatory

density dependence in fish populations: importance, controversy, understanding and

prognosis. Fish and Fisheries 2: 293–327.

Rosenberg, D.M., Berkes, F., Bodaly, R.A., Hecky, R.E., Kelly, C.A., and Rudd, J.W.N. 1997.

Large-scale impacts of hydroelectric development. Environ. Rev. 5: 27–54.

Rosenfeld, J.S., Bouwes, N., Wall, C.E., and Naman, S.M. 2013. Successes, failures, and

opportunities in the practical application of drift-foraging models. Environmental Biology

of Fishes. doi:10.1007/s10641-013-0195-6.

Rosenfeld, J.S., Raeburn, E., Carrier, P.C., and Johnson, R. 2008. Effects of side channel

structure on productivity of floodplain habitats for juvenile coho salmon. North American

Journal of Fisheries Management 28: 1108–1119.

124

Sabaton, C., Souchon, Y., Capra, H., Gouraud, V., Lascaux, J.-M., and Tissot, L. 2008. Long-

term brown trout populations responses to flow manipulation. River Research and

Applications 24: 476–505. doi:DOI: 10.1002/rra.1130.

Santos, J.M., Ferreira, M.T., Pinheiro, A.N., and Bochechas, J.H. 2006. Effects of small

hydropower plants on fish assemblages in medium-sized streams in central and northern

Portugal. Aquatic Conservation: Marine and Freshwater Ecosystems 16: 373–388.

doi:10.1002/aqc.735.

Santos, J.M., Silva, A., Katopodis, C., Pinheiro, P., Pinheiro, A., Bochechas, J., and Ferreira,

M.T. 2012. Ecohydraulics of pool-type fishways: Getting past the barriers. Ecological

Engineering 48: 38–50. doi:10.1016/j.ecoleng.2011.03.006.

Santucci, V.J.Jr., Gephard, S.R., and Pescitelli, S.M. 2005. Effects of multiple low-head dams on

fish, macroinvertebrates, habitat, and water quality in the Fox River, Illinois. North

American Journal of Fisheries Management 25(3): 975–992. doi:10.1577/M03-216.1.

Sauterleute, J.F., Hedger, R.D., Hauer, C., Pulg, U., Skoglund, H., Sundt-Hansen, L.E., Bakken,

T.H., and Ugedal, O. 2016. Modelling the effects of stranding on the Atlantic salmon

population in the Dale River, Norway. Science of the Total Environment 573: 574–584.

Schaefer, M.S. 1954. Some aspects of the dynamics of populations important to the management

of the commercial marine fisheries. Inter-American tropical tuna commission, Bulletin

Vol.1 No.2.

Schrek, C.B., Stahl, T.P., Davis, L.E., Roby, D.D., and Clemens, B.J. 2006. Mortality estimates

of juvenile spring-summer chinook salmon in the Lower Columbia River and Estuary,

1992-1998: Evidence for Delayed Mortality. Transactions of the American Fisheries

Society 135(2): 457–475.

Scott, W.B., and Crossman, E.J. 1973. Freshwater fishes of Canada. In Bulletin 184. Fisheries

Research Board of Canada.

Scrivener, J.C., and Brownlee, M.J. 1989. Effects of forest harvesting on spwning gravel and

incubation survival of chum (Oncorhynchus keta) and coho salmon (O. kisutch) in

Carnation Creek, British Columbia. Canadian Journal of Fisheries and Aquatic Science

46: 681–696.

Sedell, J.R., Reeves, G.H., Hauer, F.R., Stanford, J.A., and Hawkins, C.P. 1990. Role of refugia

in recovery from disturbances: modern fragmented and disconnected river systems.

Environmental Management 14(5): 711–724.

Senay, C., Taranu, Z.E., Bourque, G., MacNaughton, C.J., Lanthier, G., Harvey-Lavoie, S., and

Boisclair, D. 2016. Effects of river scale flow regimes and local scale habitat properties

on fish community attributes. Aquatic Sciences. doi:10.1007/s00027-016-0476-1.

Shapovalov, L., and Taft, A.T. 1954. The life histories of the steelhead rainbow trout (Salmo

gairdneri gairdneri) and silver salmon (Oncorhynchus kisutch) with special reference to

Waddell Creek, California, and recommendations regarding their management. State of

California Department of Fish and Game, Fish Bulletin No. 98.

Sharma, R., Cooper, A.B., and Hilborn, R. 2005. A quantitative framework for the analysis of

habitat and hatchery practices on Pacific salmon. Ecological Modelling 183: 231–250.

Shaw, T.L. 2004. The implications of modern hydro-electric power schemes for fisheries. Water

and Environment Journal 18(4): 207–209.

Shuter, B.J. 1990. Population-level indicators of stress. American Fisheries Society 8: 145–166.

125

Skalski, J.R., Mathur, D., and Heisey, P.G. 2002. Effects of turbine operating efficiency on smolt

passage survival. North American Journal of Fisheries Management 22(4): 1193–1200.

doi:10.1577/1548-8675(2002)022<1193:EOTOEO>2.0.CO;2.

Slaney, T.L., Hyatt, K.D., Northcote, T.G., and Fielden, R.J. 1996. Status of anadromous salmon

and trout in British Columbia and Yukon. Fisheries 21(10): 20–35.

Smith, B.D. 2000. Trends in wild adult steelhead (Oncorhynchus mykiss) abundance for

snowmelt-driven watersheds of British Columbia in relation to freshwater discharge.

Canadian Journal of Fisheries and Aquatic Science 57(2): 285–297.

Smith, S.D., Wellington, A.B., Nachlinger, J.L., and Fox, C.A. 1991. Functional responses of

riparian vegetation to streamflow diversion in the Eastern Sierra Nevada. Ecological

Applications 1(1): 89–97.

Solazzi, M.F., Nickelson, T.E., Johnson, S.L., and Rodgers, J.D. 2000. Effects of increasing

winter rearing habitat on abundance of salmonids in two coastal Oregon streams.

Canadian Journal of Fisheries and Aquatic Science 57: 906–914.

Sopinka, A., van Kooten, G.C., and Wong, L. 2013. Reconciling self-sufficiency and renewable

energy targets in a hydro dominated system: the view from British Columbia. Energy

Policy 61: 223–229. doi:10.1016/j.enpol.2013.05.068.

Spänhoff, B. 2014. Current status and future prospects of hydropower in Saxony (Germany)

compare to trends in Germany, the European Union, and the world. Renewable and

Sustainable Energy Reviews 30: 518–525. doi:10.1016/j.rser.2013.10.035.

Spidle, A.P., Quinn, T.P., and Bentzen, P. 1998. Sex-biased marine survival and growth in a

population of coho salmon. Journal of Fish Biology 52: 907–915.

Stradmeyer, L., Hojesjo, J., Griffiths, S.W., Gilvear, D.J., and Armstrong, J.D. 2008.

Competition between brown trout and Atlantic salmon parr over pool refuges during

rapid dewatering. Journal of Fish Biology 72: 848–860.

Strayer, D.L., and Dudgeon, D. 2010. Freshwater biodiversity conservation: recent progress and

future challenges. Journal of the North American Benthological Society 29(1): 344–358.

doi:10.1899/08-171.1.

Sutherland, W.J., Dias, M.P., Dicks, L.V., Doran, H., Entwistle, A.C., Fleishman, E., Gibbons,

D.W., Hails, R., Hughes, A.C., Hughes, J., Kelman, R., Le Roux, X., LeAnstey, B.,

Lickorish, F.A., Maggs, L., Pearce-Higgins, J.W., Peck, L.S., Pettorelli, N., Pretty, J.,

Spalding, M.D., Tonneijck, F.H., Wentworth, J., and Thornton, A. 2020. A horizon scan

of emerging global biological conservation issues for 2020. Trends in Ecology &

Evolution 35(1): 81–90.

Suttle, K.B., Power, M.E., Levine, J.M., and McNeely, C. 2004. How fine sediment in riverbeds

impairs growth and survival of juvenile salmonids. Ecological Applications 14(4): 969–

974.

Sweka, J.A., and Hartman, K.J. 2008. Contribution of terrestrial invertebrates to yearly brook

trout prey consumption and growth. Transactions of the American Fisheries Society

137(1): 224–235.

Tallis, H., Kennedy, C.M., Ruckelshaus, M., Goldstein, J., and Kiesecker, J.M. 2015. Mitigation

for one & all: an integrated framework for mitigation of development impacts on

biodiversity and ecosystem services. Environmental Impact Assessment Review 55: 21–

34. doi:10.1016/j.eiar.2015.06.005.

126

Tang, J., Bryant, M.D., and Brannon, E.L. 1987. Effect of temperature extremes on the mortality

and development rates of coho salmon embryos and alevins. North American Journal of

Aquaculture 49(3): 167–174. doi:10.1577/1548-8640(1987)49<167:EOTEOT>2.0.CO;2.

Taylor, L.A. 2010. A comparison of the aquatic impacts of large hydro and small hydro projects.

Master of Resource Management, School of Resource and Environmental Management,

Simon Fraser University, Burnaby, British Columbia, Canada.

Taylor, N., and Walters, C.J. 2010. Estimation of bioenergetics parameters for a stunted northern

pikeminnow population of south central British Columbia. The Open Fish Science

Journal 3: 110–121.

Thompson, B.E., and Hayes, D.B. 2010. Estimating bioenergetics model parameters for fish with

incomplete recapture histories. Canadian Journal of Fisheries and Aquatic Science 67:

1075–1085. doi:10.1139/F10-048.

Thorson, J.T., Scheuerell, M.D., Buhle, E.R., and Copeland, T. 2014. Spatial variation buffers

temporal fluctuations in early juvenile survival for an endangered Pacific salmon. Journal

of Animal Ecology 83: 157–167.

Tickner, D., Opperman, J.J., Abell, R., Acreman, M., Arthington, A.H., Bunn, S.E., Cooke, S.J.,

Dalton, J., Darwall, W., Edwards, G., Harrison, I., Hughes, K., Jones, T., Leclère, D.,

Lynch, A.J., Leonard, P., McClain, M.E., Muruven, D., Olden, J.D., Ormerod, S.J.,

Robinson, J., Tharme, R.E., Thieme, M., Tockner, K., Wright, M., and Young, L. 2020.

Bending the curve of global freshwater biodiversity loss: an emergency recovery plan.

BioScience: 1–13. doi:10.1093/biosci/biaa002.

Tilman, D., Lehman, C.L., and Bristow, C.E. 1998. Diversity-stability relationships: statistical

inevitability or ecological consequence? The American Naturalist 151(3): 277–282.

Tschaplinski, P.J., and Pike, R.G. 2017. Carnation Creek watershed experiment - long-term

responses of coho salmon populations to historic forest practices. Ecohydrology: 1–12.

doi:10.1002/eco.1812.

Ugedal, O., Næsje, T.F., Thorstad, E.B., Forseth, T., Saksga ̊rd, L.M., and Heggberget, T.G.

2008. Twenty years of hydropower regulation in the River Alta: Long-term changes in

abundance of juvenile and adult Atlantic salmon. Hydrobiologia 609: 9–23.

Valero, E. 2012. Characterization of the water quality status on a stretch of River Lérez around a

small hydroelectric power station. Water 4: 815–834. doi:10.3390/w4040815.

Vannote, R.L., and Sweeney, B.W. 1980. Geographic analysis of thermal equilibria: a conceptual

model for evaluating the effect of natural and modified thermal regimes on aquatic insect

communities. The American Naturalist 115(5): 667–695.

Vörösmarty, C.J., and Sahagian, D. 2000. Anthropogenic disturbance of the terrestrial water

cycle. BioScience 50(9): 753–765.

Wainwright, T.C., and Waples, R.S. 1998. Prioritizing Pacific Salmon stocks for conservation:

response to Allendorf et al. Conservation Biology 12(5): 1144–1147.

Walters, A.W., Bartz, K.K., and McClure, M.M. 2013a. Interactive effects of water diversion

and climate change for juvenile Chinook salmon in the Lemhi River Basin (U.S.A.).

Conservation Biology 27(6): 1179–1189.

Walters, A.W., Copeland, T., and Venditti, D.A. 2013b. The density dilemma: limitations on

juvenile production in threatened salmon populations. Ecology of Freshwater Fish 22:

508–519.

Walters, A.W., and Post, D.M. 2008. An experimental disturbance alters fish size structure but

not food chain length in streams. Ecology 89(12): 3261–3267.

127

Walters, A.W., and Post, D.M. 2011. How low can you go? Impacts of a low-flow disturbance

on aquatic insect communities. Ecological Applications 21(1): 163–174.

Wang, G., Fang, Q., Zhang, L., Chen, W., Chen, Z., and Hong, H. 2010. Valuing the effects of

hydropower development on watershed ecosystem services: Case studies in the Jiulong

River Watershed, Fujian Province, China. Estuarine, Coastal and Shelf Science 86: 363–

368. doi:10.1016/j.ecss.2009.03.022.

Waples, R.S., Beechie, T., and Pess, G.R. 2009. Evolutionary history, habitat disturbance

regimes, and anthropogenic changes: what do these mean for resilience of Pacific salmon

populations? Ecology and Society 14(1): 3.

Waples, R.S., Pess, G.R., and Beechie, T. 2008. Evolutionary history of Pacific salmon in

dynamic environments. Evolutionary Applications 1: 189–206. doi:10.1111/j.1752-

4571.2008.00023.x.

Ward, B.R., and Slaney, P.A. 1993. Egg-to-smolt survival and fry-to-smolt density dependence

of Keogh River steelhead trout. In Production of juvenile Atlantic salmon Salmo salar, in

natural waters, R.J. Gibson and R. E. Cutting. pp. 209–2017.

Ward, J.V., and Stanford, J.A. 1983. The serial discontinuity concept in lotic ecosystems. In

Dynamics of lotic ecosystems, T.D. Fontain and S.M. Bartell. Ann Arbor Science, Ann

Arbor, Michigan. pp. 29–42.

Weitkamp, D.E., and Katz, M. 1980. A review of dissolved gas supersaturation literature.

Transactions of the American Fisheries Society 109(6): 659–702.

Weltman-Fahs, M., and Taylor, J.M. 2013. Hydraulic fracturing and brook trout habitat in the

Marcellus Shale region: potential impacts and research needs. Fisheries 38(1): 4–15.

doi:10.1080/03632415.2013.750112.

Williams, J.E., and Miller, R.R. 1990. Conservation status of the North American fish fauna in

fresh water. Journal of Fish Biology 37 (Supplement A): 79–85.

Wisdom, M.J., Mills, L.S., and Doak, D.F. 2000. Life stage simulation analysis: estimating vital-

rate effects on population growth for conservation. Ecology 81(3): 628–641.

Wohl, E. 2006. Human impacts to mountain streams. Geomorphology 79: 217–248.

doi:10.1016/j.geomorph.2006.06.020.

Wu, N., Cai, Q., and Fohrer, N. 2012. Development and evaluation of a diatom-based index of

biotic integrity (D-IBI) for rivers impacted by run-of-river dams. Ecological Indicators

18: 108–117. doi:10.1016/j.ecolind.2011.10.013.

Wu, N., Tang, T., Fu, X., Jiang, W., Li, F., Zhou, S., Cai, Q., and Fohrer, N. 2010a. Impacts of

cascade run-of-river dams on benthic diatoms in the Xiangxi River, China. Aquatic

Sciences 72: 117–125. doi:10.1007/s00027-009-0121-3.

Wu, N., Tang, T., Zhou, S., Jia, X., Li, F., Liu, R., and Cai, Q. 2009. Changes in benthic algal

communities following construction of a run-of-river dam. Journal of the North American

Benthological Society 28(1): 69–79. doi:10.1899/08-047.1.

Wu, N.C., Jiang, W.X., Fu, X.C., Zhou, S.C., Li, F.Q., Cai, Q.H., and Fohrer, N. 2010b.

Temporal impacts of a small hydropower plant on benthic algal community.

Fundam.Appl. Limnol 177(4): 257–266. doi:10.1127/1863-9135/2010/0177-0257.

Yeakel, J.D., Moore, J.W., Guimaraes Jr., P.R., and de Aguiar, M.A.M. 2014. Synchronisation

and stability in river metapopulation networks. Ecology Letters 17: 273–283.

doi:10.1111/ele.12228.

128

Young, P.S., Cech, J.J. jr, and Thompson, L.C. 2011. Hydropower-related pulsed-flow impacts

on stream fishes: a brief review, conceptual model, knowledge gaps, and research needs.

Revies in Fish Biology and Fisheries 21: 713–731. doi:10.1007/s11160-011-9211-0.

Zabel, R.W., Scheuerell, M.D., McClure, M.M., and Williams, J.G. 2006. The interplay between

climate variability and density dependence in the population viability of Chinook salmon.

Conservation Biology 20(1): 190–200.

Zhou, S., Tang, T., Wu, N., Fu, X., and Cai, Q. 2008. Impacts of a small dam on riverine

zooplankton. Internat. Rev. Hydrobiol 93(3): 297–311. doi:10.1002/iroh.200711038.

Zhou, S., Tang, T., Wu, N., Fu, X., Jiang, W., Li, F., and Cai, Q. 2009. Impacts of cascaded

small hydropower plants on microzooplankton in Xiangxi River, China. Acta Ecologica

Sinica 29: 62–68. doi:10.1016/j.chnaes.2009.04.008.

129

Appendix A

Non-empirical peer-reviewed papers on RoR hydropower.

Papers Category Country

Abbasi and Abbasi 2001 Global and historical perspective on emergence of ROR hydropower, general review of potential

effects on ecosystem, cumulative impacts India

Anderson et al. 2014 Review of ROR hydropower impacts on river

ecosystems UK, Europe

Aroonrat and Wongwises 2015 Geography, economy and politics of RoR

hydropower Thailand

Bakken et al. 2012 Comparison of small to large hydropower,

cumulative impacts Norway

Chiyembekezo 2013 Geography, economy and politics of RoR

hydropower Malawi

Cyr et al. 2001 Geography, economy and politics of RoR

hydropower Canada

Grumbine and Pumdit 2013 Geography, economy and politics of RoR

hydropower India

Jaccard et al. 2001 Geography, economy and politics of RoR

hydropower Canada

Kibler 2011 Comparison of small to large hydropower,

cumulative impacts China

Kotchen et al. 2006 Cost benefit analysis of hydropower dam

relicensing USA

Poff and Hart 2002 Classification of small dams and review of

science of dam removal USA

Renofalt et al. 2010 Effects of hydropower and environmental flow management on river ecosystems, cumulative

impacts Sweden

Shaw 2004 Implications of modern hydro-electric plants for

fisheries and natural hydrographs

Sopinka et al. 2013 Geography, economy and politics of RoR

hydropower Canada

Spänhoff 2014 Geography, economy and politics of RoR

hydropower Germany and Europe

Wang et al. 2010 Geography, economy and politics of RoR

hydropower China

130

Appendix B

Figure 3B-1. Mean daily flow (m3s-1) diverted for hydroelectricity production (plant flow) and left in the bypassed reach (darker lines, left Y-axis) in Fire Creek from 2010 to 2013, and corresponding proportion of diverted flow (pale gray line, right axis). Flow in years pre-flow diversion or in 2014 were not available.

0

5

10

15

20

25

30

35a) 2010

0

5

10

15

20

25

30

35

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov DecMonth

c) 2012

Ave.

daily

flo

w (

cm

s)

b) 2011 Plant f low

By passed f low

% f low div erted

0

20

40

60

80

100

Pro

port

ion o

f div

ert

ed f

low

(%

)

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov DecMonth

c) 2013

0

20

40

60

80

100

131

Figure 3B-2. Mean daily flow (m3s-1) diverted for hydroelectricity production (plant flow) and left in the bypassed reach (darker lines, left Y-axis) in Douglas Creek from 2010 to 2013, and corresponding proportion of diverted flow (pale gray line, right axis). Flow in years pre-flow diversion or in 2014 were not available.

0

10

20

30

40

a) 2010

0

10

20

30

40

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov DecMonth

c) 2012

Ave.

daily

flo

w (

cm

s)

b) 2011Plant f low

By passed f low

% f low div erted

0

20

40

60

80

100

Pro

port

ion o

f div

ert

ed f

low

(%

)

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov DecMonth

c) 2013

0

20

40

60

80

100

132

Appendix C

Table 3C-1. Four food availability scenarios used in the bioenergetics simulations.

Scenario Ecological assumptions in

context of RoR hydropower Bioenergetics consequences

for fisha

Bioenergetics settings

parameters fixed parameters solved

Unlimited Consumption (UC)

↑↑ food availability, no limits on capacity to catch and process

food C and G are maximized pCmax=1 Total C and final weight

Constant Consumption (CC)

No change in food availability or ability to catch food

↑ in T= ↑ in metabolism but no change in C; ↓ in final weight

Total C in bypassed reach= total C under

natural baseline T Final weight

Increased Consumption (IC)

↑ food availability and processing with ↑ in T; catching

food facilitated by ↓ in Q

C: not limiting, and ↑ capacity to process food= ↑ in final

weight

pCmax in bypassed reach= pCmax in natural

baseline T Total C and final weight

Reduced Consumption (RC)

↓ in Q= ↓ rate of invertebrate drift ; ↓ in capacity to catch food

↓ in total C + ↑ in water T= ↓ in final weight

Total C in bypassed reach = 75% of C in natural baseline T

Final weight

a Q= discharge, C= consumption, G= growth, T= temperature

133

Appendix D

Figure 3D-1. Temperatures in the upstream and bypassed reaches from 2007 to 2014 in Fire Creek.

0

5

10

15

a) 2007

0

5

10

15

c) 2010

Tem

pera

ture

(C

)

0

5

10

15

e) 2012

0

5

10

15

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

g) 2014

Upstream reach

Bypassed reach

b) 2008

d) 2011

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

f) 2013

134

Figure 3D-2. Temperatures in the upstream and bypassed reaches from 2007 to 2014 in Douglas Creek.

0

5

10

15

a) 2007

0

5

10

15

c) 2010

Tem

pera

ture

(C

)

0

5

10

15

e) 2012

0

5

10

15

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

g) 2014

Upstream reach

Bypassed reach

b) 2008

d) 2011

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month

f) 2013

135

Appendix E

Table 3E-1. Mean annual growth (g) observed in the bypassed reaches of the two streams and simulated under the four food availability scenarios for the three years modelled.

Stream Age Scenarios Annual growth (mean, g) in:

2010-2011 2011-2012 2013-2014

Douglas Creek

from 0+ to 1+

Observed 4.7 4.4 8.4

Reduced C 3.8 3.9 3.8

Constant C 6.3 6.4 6.2

Increased C 8.3 8.1 7.9

Unlimited C 228.0 153.7 262.4

from 1+ to 2+

Observed 7.9 9.4 12.0

Reduced C 5.5 5.7 5.4

Constant C 11.9 12.1 11.8

Increased C 16.0 15.7 15.3

Unlimited C 322.8 227.4 365.7

Fire Creek

from 0+ to 1+

Observed 5.6 5.7 7.5

Reduced C 3.4 3.5 3.4

Constant C 5.7 5.7 5.6

Increased C 8.5 9.2 8.1

Unlimited C 225.4 169.9 269.6

from 1+ to 2+

Observed 9.8 8.1 10.0

Reduced C 4.9 4.9 4.7

Constant C 10.6 10.6 10.4

Increased C 16.2 17.5 15.4

Unlimited C 314.6 244.6 369.4

136

Figure 3E-1. Frequency distributions of Age 2+ O.mykiss sampled in the bypassed reach of Douglas Creek from 2010 to 2014, a) weight (n= 48 fish, mean: 20.5g), and b) estimated specific growth rate (SGR, vertical bars) with predicted SGR of Age 2+ O.mykiss under reduced C, constant C, and increased C scenarios (dashed lines, right axis).

137

Figure 3E-2. Frequency distributions of Age 1+ O.mykiss sampled in the bypassed reach of Fire Creek from 2010 to 2014, a) weight (n= 193 fish, mean: 7.3g), and b) estimated specific growth rate (SGR, vertical bars) with predicted SGR of Age 1+ O.mykiss under reduced C, constant C, and increased C scenarios (dashed lines, right axis).

138

Figure 3E-3. Frequency distributions of Age 2+ O.mykiss sampled in the bypassed reach of Fire Creek from 2010 to 2014, a) weight (n= 112 fish, mean: 19.15g), and b) estimated specific growth rate (SGR, vertical bars) with predicted SGR of Age 2+ O.mykiss under reduced C, constant C, and increased C scenarios (dashed lines, right axis).

139

Appendix F

Figure 3F-1. Variation in mean annual fish weight (g) for Age 1+ O. mykiss in a) Douglas Creek and b) Fire Creek, and for Age 2+ O. mykiss in c) Douglas Creek and d) Fire Creek under each of the four food availability scenarios. The three lines per scenario represent the averages for all fish simulated per year. Final weights under Scenario 1 Unlimited C reached up to 270g for Age 1+ fish and 315g for Age 2+ fish.

140

Appendix G

Characteristics of streams regulated by RoR hydropower in the coastal mountains of British Columbia, Canada. Frequency and timing of flow fluctuations in these streams from 2011 and 2016 were computed from monitoring reports made available by the Ministry of Forests, Lands, Natural Resource Operations and Rural Development of British Columbia, Canada.

Creek Length of

bypassed reach (km)

Mean annual discharge (cms)

Max discharge diverted (cms)

Capacity (MW)

Ashlu 5 -- 29 49.9

Fitzsimmons 4.5 -- 4 7.5

Lamont 3.2 -- 8.7 27

McNair -- -- 3.3 9.8

Fire 4.3 5.2 10.5 23

Douglas 3 6.5 11.3 27

Stokke 3.1 -- 8.4 22

Tippela 2.2 -- 7.2 18