Variable movements after catch-and-release tournament angling and isotopic niches

of walleye (Sander vitreus) and sauger (S. canadensis)

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

For the Degree of

Master of Science

in

Biology

University of Regina

by

Jessica Carroll Butt

Regina, Saskatchewan

November, 2016

© 2016: J. C. Butt

UNIVERSITY OF REGINA

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Jessica Carroll Butt, candidate for the degree of Master of Science in Biology, has presented a thesis titled, Variable movements after catch-and-release angling and isotope niches of walleye (Sander vitreus) and sauger (S. canadensis), in an oral examination held on November 14, 2016. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner: Dr. Douglas Chivers, University of Saskatchewan

Supervisor: Dr. Christopher Somers, Department of Biology

Committee Member: Dr. Mark Brigham, Department of Biology

Committee Member: Dr. Richard Manzon, Department of Biology

Chair of Defense: Dr. Maria Velez, Department of Geology

! II

ABSTRACT

Walleye (Sander vitreus) and sauger (S. canadensis) are closely related freshwater species that are ecologically and economically important throughout North America.

These two species are sympatric in many areas, and are often regulated as a single entity.

However, the important similarities and differences that exist between these species in the context of various management issues remain uncertain.

The first chapter of my thesis addresses the effects of catch-and-release fishing that is part of angling tournaments on walleye and sauger. Catch-and-release tournaments are popular in Saskatchewan and other areas, but frequently prompt management concerns regarding fish health and mortality. Potential sub-lethal effects on movement and habitat use are particularly poorly studied. I tracked both walleye and sauger after catch-and-release using telemetry following four 2015 tournaments. My objectives were to: (1) describe post-release movements; and (2) evaluate factors that may affect post- release behaviour. The initial movements I observed were highly variable; for example, some fish moved less than 1 km, whereas others moved long distances, the maximum recorded being 31 km. All large movements I observed were made by walleye, while the farthest a sauger moved was 2 km. Fish exited the release area (500 m) faster (2.6±1.1 days versus 3.9±1.6 – 4.4±1.7 days) and moved farther in the first 48 hours

(3,712.8±4,427.9 m versus 608.1±392.1 – 1065.3±687.8 m) at one tournament compared to the other three. Stress scores were significantly lower at this tournament (3.50±0.73 versus 4.60±1.17 – 4.75±1.48), and an important negative predictor of fish movement.

However, there were few common predictors of movement among tournaments. I confirmed delayed mortality of tagged fish at only one tournament, which was high ! III

(45%) and mostly due to avian predation. I conclude that walleye and sauger behaviour after tournaments varies widely, likely due to varying tournament and environmental conditions. The understanding of factors that modify fish behaviour remains limited and warrants further study.

The second chapter of my thesis assesses the ecological relationship between walleye and sauger when they co-exist. Limited information is available regarding potential niche partitioning across different time scales by these sympatric congeners, though I expected niche overlap to be high during early ages and lower later in life. I used multi-tissue (liver, muscle, and bone layers) isotopic niche analysis to quantify overlap by these species in two large reservoirs in Saskatchewan. Population niche size was larger for sauger (1.73 – 3.60 ‰2) than walleye (0.51 – 2.61 ‰2) in Lake Diefenbaker, but this trend was reversed in Tobin Lake (sauger = 0.54 – 2.57 ‰2; walleye = 1.30 – 2.94 ‰2).

Inter-species niche overlap at the population level also varied by site (Diefenbaker = 3-

19%, Tobin = 9-31%), and niche partitioning was evident in both systems at all time scales. The niche size of individual fish and levels of intra- and inter-specific overlap were highly variable by site. Walleye and sauger appear to use different resources but the factors influencing variation in niche size and overlap are currently unclear.

My findings indicate there is considerable variation in the response of walleye and sauger to catch-and-release angling, as well as in their resource use and niche overlap.

The important differences I describe between these species suggest that a re-evaluation of regulating them as one entity may be warranted.

! IV

ACKNOWLEDGEMENTS

Funding and support was provided by the Canada Foundation for Innovation, the

Fish and Wildlife Development Fund through the Saskatchewan Ministry of

Environment, the Saskatchewan Wildlife Federation, the Regina Fish & Game League, the Saskatchewan Walleye Trail, and angler donations. I received additional support from the University of Regina, as well as the Institute of Environmental Change and Society

(IECS).

I thank my field crew, Una Goncin, Tom Morgan, and Tony Fernando, for their invaluable assistance in chasing fish. I also appreciate the help of various other volunteers who gave their time on short notice. They include Rebecca Eberts, Tera Edkins, Gabe

Foley, Shayna Hamilton, Andrea Murillo, Kelsey Marchand, Adam Matichuk, Natalie

Taylor, Tyler Vanstone, and Lindy Whitehouse. I thank Rebecca Eberts a second time for her extensive help with my stable isotope analysis and her unending patience as I learned the many necessary procedures. Additionally, I thank Kayla Balderson Burak for GIS help, Leanne Heisler and Gavin Simpson for help solving various statistical conundrums, and Bjoern Wissel for stable isotopes guidance.

Finally, I thank my supervisor Chris Somers, who saw enough potential in this

Yank to bring her North. I couldn’t have asked for a better mentor for the many adventures and learning curves in this project. Thank you for letting me “tour the lakes” of Saskatchewan, putting up with my sometimes constant visits to your office, and giving me the support and advice I needed to become the biologist I am today.

! V

POST DEFENSE ACKNOWLEDGEMENTS

I would like to thank my External Examiner, Dr. Doug Chivers, for thought- provoking questions, insightful discussion, and a thoroughly enjoyable defense.

! VI

DEDICATION

This thesis is dedicated to my family – Mum, Dad, and little Sis – thanks for always believing in me, even when I ran away to Canada.

! VII

TABLE OF CONTENTS

ABSTRACT………………………………………………………………………………II

ACKNOWLEDGEMENTS……………………………………………………………...IV

POST DEFENSE ACKNOWLEDGEMENTS…………………………………………...V

DEDICATION…………………………………………………………………………...VI

TABLE OF CONTENTS……………………………………………………………….VII

LIST OF TABLES………………………………………………………………………..X

LIST OF FIGURES……………………………………………………………………..XII

LIST OF ABBREVIATIONS & SYMBOLS………………………………………….XIV

1. CHAPTER 1: GENERAL INTRODUCTION……………………………………...1

1.1. Life history and ecology of walleye and sauger……..…………………………...1

1.2. Walleye and sauger in Saskatchewan…………………………………………….5

1.3. Research interest #1: Catch-and-release tournament angling…………………….9

1.4. Research interest #2: Niche ecology of sympatric species……………………...10

1.5. Research objectives……………………………………………………………..11

2. CHAPTER 2: Variable short-term movements by walleye (Sander vitreus) and

sauger (S. canadensis) after catch-and-release at angling tournaments………….12

2.1. INTRODUCTION………………………………………………………………12

2.2. MATERIALS & METHODS…………………………………………………...16

2.2.1. Study sites……………………………………………………………….16

2.2.2. Fish capture & handling………………………………………………...20

2.2.3. Angler survey…………………………………………………………...21

2.2.4. Fish stress assessment…………………………………………………..21 ! VIII

2.2.5. Application of radio or acoustic transmitters.…………………………..22

2.2.6. Fish tracking…………………………………………………………….23

2.2.7. Post-tournament behaviour……………..……………………………….24

2.2.7.1. General fish movement……………………………………………..25

2.2.7.2. Departure from release area……...... ……………………………….25

2.2.7.3. Fish mortality……………………………………………………….26

2.2.7.4. Statistical analyses…………………………………………………..26

2.3. RESULTS……………………………………………………………………….27

2.3.1. Angler survey…………………………………………………………...28

2.3.2. Fish stress assessment…………………………………………………..29

2.3.3. Fish tracking…………………………………………………………….29

2.3.4. Departure from release area……..……………………………………...33

2.3.5. Movement in 48 hours…………………………………………………..37

2.3.6. Overall movement………………………………………………………44

2.4. DISCUSSION…………………………………………………………………...49

3. CHAPTER 3: Sympatric walleye (Sander vitreus) and sauger (S. canadensis) in

large reservoirs: variable niche size and overlap across multiple time scales…...59

3.1. INTRODUCTION………………………………………………………………59

3.2. MATERIALS & METHODS…………………………………………………...63

3.2.1. Study sites & fish collection………………………………………………63

3.2.2. Multi-tissue processing & analyses……………………………………….66

3.2.3. Statistical analyses………………………………………………………...69

3.3. RESULTS……………………………………………………………………….70 ! IX

3.3.1. Niche overlap……………………………………………………………..70

3.3.2. Niche size…………………………………………………………………72

3.3.3. Individual niches………………………………………………………….73

3.4. DISCUSSION…………………………………………………………………...85

3.4.1. Niche overlap…………………………………………………………….85

3.4.2. Niche size………………………………………………………………...88

3.4.3. Individual niches…………………………………………………………90

3.4.4. Differences among study sites…………………………………………...91

3.4.5. Other points………………………………………………………………92

4. CHAPTER 4: CONCLUSIONS & FUTURE DIRECTIONS………………….....95

5. LITERATURE CITED………………………………………………………………100

6. APPENDICES……………………………………………………………………….130

6.1. Appendix I: Angler survey…………………………………………………….130

! X

LIST OF TABLES

Table 2.1. Summary information for angling tournaments held on three lakes in

Saskatchewan in 2015……………………………………………………………………18

Table 2.2. Summary of responses from anglers to a survey that was conducted at each tournament……………………………………………………………………………….30

Table 2.3. Mean scores for all components of the stress test performed on tagged fish, including RAMP assessment and swim score……….…………………………………..31

Table 2.4. Fish movement, behaviour, and mortality after each angling tournament based on telemetry data…………………...…………………………………………………….35

Table 2.5. Model selection based on AIC assessing relationship between predictor variables, and time to exit the release area and movement in 48 hours at the Riverhurst

Walleye Classic…………………………………………………………………………..38

Table 2.6. Model selection based on AIC assessing relationship between predictor variables, and time fish took to exit the release area, movement in 48 hours, and mortality at the Saskatchewan Landing Walleye Tournament……………………………………..39

! XI

Table 2.7. Model selection based on AIC assessing relationship between predictor variables, and time fish took to exit the release area, movement in 48 hours, and overall movement rate at the at the Last Mountain Fall Classic…………………………………41

Table 2.8. Model selection based on AIC assessing relationship between predictor variables, and time fish took to exit the release area, movement in 48 hours, and overall movement rate at the at the Nipawin Vanity Cup………………………………………..42

Table 3.1. Summary information for the four tournaments at which walleye and sauger were sampled for stable isotope analysis………………………………………………...64

Table 3.2. Details regarding walleye and sauger sampled for liver, muscle, and opercula bones from three lakes in Saskatchewan during 2014 and 2015………………………...71

Table 3.3. Isotopic means (δ13C and δ15N) and niche size for multiple tissues sampled from walleye and sauger in three lakes…………………………………………………..74

Table 3.4. Niche metrics for individual walleye and sauger quantified using five temporal periods and metrics for total population niche for both species at three locations………75

! XII

LIST OF FIGURES

Figure 1.1. Walleye (Sander vitreus) and sauger (S. canadensis)………………………..2

Figure 2.1. Map showing locations of four catch-and-release tournaments in

Saskatchewan………………………………………………………………………….…17

Figure 2.2. Box-and-whisker plot of total stress scores for fish at tournaments…...……32

Figure 2.3. Proportion of total tagged fish found on each of seven days of tracking after tournaments, based on radio and acoustic telemetry…………….………………………34

Figure 2.4. Proportion of tagged fish that exited the release area (swam > 500 m) during three defined stages of dispersal…………………………………………………………36

Figure 2.5. Maps for the Riverhurst Walleye Classic and Saskatchewan Landing Walleye

Tournament summarizing minimum paths fish traveled during the week of tracking…..45

Figure 2.6. Maps for the Last Mountain Fall Classic and Nipawin Vanity Cup summarizing minimum paths fish traveled during the week of tracking………………...47

Figure 3.1. Maps indicating the locations of the three lakes in Saskatchewan that were sampled for walleye and sauger for stable isotope analysis……………………………..65

! XIII

Figure 3.2. Example of a walleye operculum, which were sampled from fish for stable isotope analysis, illustrating the three segments used……………………………………68

Figure 3.3. Bayesian standard ellipses representing the isotopic niches (δ15N vs. δ 13C) of walleye and sauger from Lake Diefenbaker at five temporal periods…………………...76

Figure 3.4. Bayesian standard ellipses representing the isotopic niches (δ 15N vs. δ 13C) of walleye and sauger from Tobin Lake at five temporal periods………………………….78

Figure 3.5. Bayesian standard ellipses representing the isotopic niches (δ 15N vs. δ 13C) of walleye from Last Mountain Lake at five temporal periods……………………………..80

Figure 3.6. Standard ellipses representing the isotopic niches (δ 15N vs. δ 13C) for individual walleye and sauger from Lake Diefenbaker, as well as the total isotopic niche for each population………………………………………………………………...82

Figure 3.7. Standard ellipses representing the isotopic niches (δ 15N vs. δ 13C) for individual walleye and sauger from Tobin Lake, as well as the total isotopic niche for each population…………………………………………………………………………..83

Figure 3.8. Standard ellipses representing the isotopic niches (δ 15N vs. δ 13C) for individual walleye from Last Mountain Lake, as well as the total isotopic niche for the population………………………………………………………………………………..84 ! XIV

LIST OF ABBREVIATIONS & SYMBOLS

δ Delta notation used when reporting the ratio of heavy (e.g., 15N) to light (e.g., 14N) isotopes. Expressed in units of per mill (‰). Values are derived from the equation: δ Heavy Isotope = (Rsample – Rstandard / Rstandard) x 1000, where R is the ratio of heavy to light isotopes.

‰ Per mill or parts per thousand, the unit used for describing the ratio of heavy to light isotopes

AIC Akaike’s Information Criterion, a model selection criteria that is used to rank competing models and weigh the relative support for each

AICc Second-order AIC, used to correct for small sample sizes

ΔAIC Measure of each model within AIC relative to the best model, where Δi < 2 suggests substantial evidence for the model

ANOVA Analysis of Variance, statistical test used to analyze the differences among group means

GLM Generalized Linear Model, a flexible generalization of the ordinary linear regression, which allows for response variables that have non-normal error distributions k Number of fitted parameters, including the intercept, within a model lme4 Package in R used to fit and analyze linear mixed models

MuMIn Package in R used for model selection and model averaging based on information criteria, including AICc

RAMP test Reflex Action Mortality Predictor test, quantifies the level of impairment of a series of reflexes to measure fish stress

SD Standard Deviation

SEA Standard Ellipse Area, represents the core 40% data points and is a measure of isotopic niche size. This size is directly related to the diversity of resources used by the consumer.

SEAB Standard Ellipse Area calculated using Bayesian statistics, which involves 10,000 replications and includes niche size uncertainty

! XV

SEAC Standard Ellipse Area, corrected for small sample size

SIAR Stable Isotope Analysis in R, a package in R used for analyzing stable isotope data

SIBER Stable Isotope Bayesian Ellipses in R, a set of functions within the SIAR R-package used for constructing Bayesian ellipses and quantifying niche parameters

TL Total length of fish, measured from nose to end of tail wi Akaike weight, a measure of strength of evidence for a model, based on the ratio of the ΔAIC value of the model relative to the whole set of candidate models

! 1

1. CHAPTER 1: GENERAL INTRODUCTION

1.1. Life history and ecology of walleye and sauger

Walleye (Sander vitreus) and sauger (S. canadensis) are related fish species from within the percid family that are endemic to North America (Billington and Sloss 2011).

These species have been identified as closest relatives with a divergence time estimated at 3.12 ± 1.33 million years before present (Billington et al. 1990). This relatively recent divergence leads to the possibility of hybridization; male sauger and female walleye have been known to naturally hybridize and produce “saugeye” (Billington and Sloss 2011).

Rates of natural hybridization can be high in some areas (10-22%; Flammang and Willis

1993; Leary and Allendorf 1997; Billington et al. 2006) and low in others (>5%;

Bingham et al. 2012). Additionally, walleye and sauger populations can have high potential for competition in areas where they overlap (Bellgraph et al. 2008). Competition potential between the two species largely stems from significant similarities in their life history and ecology (Haxton 2015).

In general, both species can be characterized as having an elongate body, two large dorsal fins (the first spiny), and large teeth (Scott and Crossman 1973). They do exhibit several unique species-specific characteristics (described in Fig. 1.1), and there are also differences in their growth patterns and maximum size. Walleye, as a popular sport fish and sought-after trophy species, can be as large as 790 mm (11 kg), while sauger tend to be much smaller and reach a maximum size of 490 mm (3.8 kg; Mittelbach and Persson 1998).

! 2

Figure 1.1. Photographs of (a) sauger and (b) walleye from Saskatchewan (photo credit:

C. Somers). Color in both species varies with habitat and can range from sandy to dull brown in sauger and olive-brown to golden-brown in walleye. In addition, walleye often exhibit golden flecks on their scales. There are several major physical appearance differences between the two species. The first dorsal fin of sauger has alternating distinct black and yellow spots on the membrane between the spines while the walleye first dorsal fin is dusky, clear, or vaguely speckled but without the distinct rows of spots. The body of sauger also exhibits three or four dark brown saddles and the lower lobe of the caudal fin sometimes has a faint white tip. Walleye have no saddling on the body but do have barring on the back, as well as distinct white coloring on the lower lobe of the caudal fin.

Differentiation of the sexes based on physical characteristics is difficult in both species. ! 3

Walleye and sauger are top-level predators native to the freshwater systems of the central United States and Canada (Scott and Crossman 1973). Their native distributions in North America are similar, though walleye inhabit a higher percentage of available freshwater habitats compared to sauger (walleye = 32%, sauger = 10%; Bozek et al.

2011a). Both species range from Quebec and Vermont in the east, Alberta and Montana in the west, and Louisiana in the south. Their only difference in range is that sauger are only found as far north as the middle of Saskatchewan and Manitoba, while walleye extend farther north into parts of the Northwest Territories and Yukon (Billington et al.

2011). Both species have been introduced into freshwater systems within and outside of their native ranges (Fuller 2010a; Fuller 2010b). Walleye are tolerant of a wide range of environmental conditions and recruit successfully in an array of aquatic systems. They are abundant in both large, turbid lakes, which represent the most common habitat in which they occur, and large rivers (Scott and Crossman 1973; Bozek et al. 2011a).

Sauger tolerance to environmental conditions is less flexible. They are generally considered to be a riverine species but can also occur in large lakes, and they are associated with high turbidity from suspended clay particles and deep waters (Scott and

Crossman 1973; Bozek et al. 2011a). As the two species both prefer to inhabit large, turbid rivers and lakes, they are often sympatric (Bellgraph et al. 2008; Haxton 2015).

They also exhibit syntopy in areas, using the same habitats at the same time, thus facilitating hybridization.

Both species undertake similar large migrations to overwintering and spawning grounds (Bellgraph 2008). Spawning by both species is determined by temperature (6-

11°C for both; Mittelbach and Persson 1998), and takes place for several weeks in the ! 4 spring or early summer, depending on location. Northern populations of walleye do not spawn in years when temperatures are too low. Walleye and sauger are external fertilizing broadcast spawners, and while walleye tend to spawn earlier than sauger, spawning periods overlap in many systems (Nelson and Walburg 1977). Walleye spawning begins shortly after the ice breaks in spring, with sauger spawning immediately after, usually at the end of May or beginning of June (Scott and Crossman 1973; Nelson and Walburg 1977). Both species prefer gravel or rubble substrate in the shallows of lakes and rivers to spawn, and may use the same gravel shoals (Scott and Crossman 1973;

Barton and Barry 2011). Males arrive at spawning grounds first and stay after females have departed. Spawning generally occurs at night in water 6-36 m deep, where several males attend a single female (Rawson 1957). Females drop their eggs (total can range from 9,000 to 96,000 for sauger and be as high as 600,000 for walleye), where they fall between boulders or gravel and are fertilized by males; fertilized eggs harden in the water and become non-adhesive and semi-buoyant (Scott and Crossman 1973). They hatch in

12-18 days for walleye and 25-29 days for sauger; differences in this timing are not well understood but likely depend on water temperature (Koenst and Smith 1976; Barton and

Barry 2011).

Sauger and walleye are visual predators that have evolved to exploit turbid, low- light conditions, taking advantage of the light-gathering effect of tapetum lucida layers in their eyes (Kelso 1978). This eye layer effectively magnifies the amount of naturally occurring light and is therefore extremely sensitive to bright daylight (Ali and Anctil

1977). As such, feeding patterns reflect the environment, where fish in clearer waters and small systems will feed exclusively at night, while those in more turbid waters and large, ! 5 deeper systems that provide shelter during daylight will feed intermittently throughout the day (Scott and Crossman 1973). Both species swim in schools while they feed, often overlapping. They exhibit ontogenetic dietary shifts as they grow from the fry and juvenile life stages to adulthood. Their diet depends on location and availability, but generally begins with a reliance on zooplankton and insect larvae. A rapid shift to piscivory within the first year of life alters the diet to comprise a wide variety of small fishes, including cisco (Coregonus spp.), emerald shiners (Notropis atherinoides), perch

(Perca spp.), burbot (Lota lota), flathead chub (Platygobio gracilis), gizzard shad

(Dorosoma cepedianum), rainbow smelt (Osmerus mordax), sticklebacks (Gasterosteidae spp.), and even small walleye and sauger, as well as various invertebrates (leeches, crayfish, and insects; Rawson 1957; Mittelback and Persson 1998; Paradis et al. 2006;

Chipps and Graeb 2011).

1.2. Walleye and sauger in Saskatchewan

The province of Saskatchewan in central Canada supports an abundance of aquatic systems – an estimated 50,000 lakes and rivers contain 67 species of fish, 58 of which are native (Somers and Morris 2015). These systems are economically important as they support significant recreational and commercial fisheries (Ashcroft et al. 2006).

The popularity of recreational fishing in Saskatchewan has grown considerably since

1950 and as recently as 2010, anglers invested almost $500 million in the province (Sport

Fishing 2010). Saskatchewan has a population of 1.13 million people (Bureau of

Statistics 2015) and approximately 180,000 residents of all ages (16%) participate in sport fishing (Sport Fishing 2010). Many lakes and rivers in Saskatchewan, including ! 6 several recently constructed reservoirs in the southern portion, also have high commercial fishery potential. The commercial gill net fishery has been a productive operation for over 120 years. Today, most commercial fishing occurs on more northern lakes in the province and generates between $5-7 million annually (Ashcroft et al. 2006). Though a wide variety of species are targeted, five species are most important in terms of fishing effort and harvest: Northern pike (Esox lucius), yellow perch (Perca flavescens), lake trout (Salvelinus namaycush), lake whitefish (Coregonus clupeaformis), and walleye

(Ashcroft et al. 2006).

Both walleye and sauger are valued in certain recreational, commercial, and subsistence fisheries in Canada. For example, both species are historically important for the Lake Winnipeg fishery in Manitoba (Johnston et al. 2012). However, in

Saskatchewan, walleye are highly valuable but sauger are significantly less so. Walleye have increased in popularity among anglers since the 1980’s and are currently the most popular sport fish, comprising 41% of the Saskatchewan recreational harvest in 2010, compared to 30% for pike (the next most popular; Sport Fishing 2010). Commercially, walleye are the most valuable freshwater fish both in Saskatchewan and Canada

(Fisheries and Oceans Canada 2014). Of the 3 million dollars earned by the

Saskatchewan commercial fishery in 2014, walleye brought in almost half (Fisheries and

Oceans Canada 2014). Additionally, walleye represent an important part of the diet, culture, and traditions of indigenous people in Saskatchewan (Ashcroft et al. 2006).

Recreational, commercial, and Aboriginal harvest of sauger in Saskatchewan is minimal, and anglers generally catch sauger when targeting walleye (Ashcroft et al. 2006).

Furthermore, walleye, sauger, and hybrids (saugeye) are regulated as the same entity for ! 7 the purposes of recreational angling, where the catch limit is based on any combination of the three (Saskatchewan Anglers’ Guide 2015). While the economic importance of sauger is significantly less than that of walleye in Saskatchewan, this species is ecologically important, as their presence in aquatic systems adds to larger biodiversity and supports the health of all components of the system (Dunne et al. 2002; Worm et al. 2006).

Managers tasked with maintaining populations of walleye and sauger in

Saskatchewan face many challenges. Overfishing is a significant concern for managers wherever economically valuable populations occur. Walleye are in high demand in

Saskatchewan, both by anglers and commercial net fisheries, and therefore the possibility of over-exploitation is high (Ashcroft et al. 2006). In addition to the threat of over- harvesting, walleye and sauger populations can also be negatively impacted by factors like changing prey communities (Roseman et al. 2014) and loss of spawning grounds

(Thompson 2009; Wallace et al. 2010), both of which are possible as habitat alterations continue to occur across the province. To compound the issue, the more environmentally constrained sauger are also declining in many parts of their range due to development projects, habitat loss, and fragmentation (Bingham et al. 2012).

In areas where walleye and sauger populations overlap, there is a potential for interspecific competition caused by resource limitation, a situation where walleye are phenotypically favored because of their larger size (Bellgraph et al. 2008). Competition is also prevalent when non-native walleye invade areas where native sauger are present

(Bellgraph et al. 2008; Graeb et al. 2010), as well as situations where native populations of both species are brought together artificially through human disturbance, such as dam and reservoir construction (Wallace et al. 2010; Bozek et al. 2011a). For example, the ! 8 construction of two dams produced Lake Diefenbaker, an important reservoir in southern

Saskatchewan. This development brought together historically separate riverine populations of walleye and sauger. While recent reports have indicated these populations are healthy and stable, there is potential for competition (Wallace et al. 2004; Wallace et al. 2010).

Management efforts also need to focus on the potential impacts of hybridization on native walleye and sauger. While this phenomenon occurs naturally in many areas, hybridization rates as high as 22% were reported after the creation of Lake Diefenbaker, and the phenomenon has been observed to occur in other reservoirs in Saskatchewan as well (Wallace et al. 2010). The presence of hybrid saugeye in these systems has the potential to compromise the genetic integrity of both parent species by decreasing genetic diversity, as has been illustrated in other systems (Fiss et al. 1997; Graeb et al. 2010;

Quist et al. 2010). Saskatchewan has also developed an extensive walleye stocking program, where eggs are collected during spawning from multiple lakes in the southern part of the province and the resulting fry are used to stock many lakes and maintain walleye populations (Wallace et al. 2010). Up to 100 million eggs are taken every year from fish from Lake Diefenbaker alone (Wallace et al. 2004). While these stocking activities are used as a method of walleye management, stocking has been shown to affect genetic structure and disrupt local adaptation of native fish (Waterhouse et al.

2014). In the event walleye populations collapse in the province, using hatchery fish may not be useful in aiding recovery (Garner et al. 2013). Overall, walleye and sauger are both ecologically important species in Saskatchewan as top-level predators, as well as economically important. The many possible management concerns for their maintenance ! 9 in the province provides incentive to better understand the ecology of both species, especially in a province where research on them has been minimal.

1.3. Research interest #1: Catch-and-release tournament angling

Recent global estimates predict that 60% (estimated 30 billion fish released of 47 billion total caught) of fish caught worldwide are released back into their environments

(Arlinghaus et al. 2007). Catch-and-release as a management technique presumes fish survive unharmed. In Saskatchewan, the number of fish released back into water bodies has more than doubled since 1980, increasing from 30% to 70% released in just 20 years

(Ashcroft et al. 2006). However, the success of catch-and-release angling in terms of fish survival can be highly variable. The process physiologically impacts fish in the form of injuries, air exposure, and exhaustion (Cooke et al. 2002; Richard et al. 2014; Havn et al.

2015). While many fish survive largely unharmed, many others are significantly impacted or die (Cooke et al. 2002; Cooke and Philipp 2004).

The most extreme forms of catch-and-release occur at angling tournaments. At these competitive events, anglers attempt to catch as many large fish as possible (within the defined limit). Retained fish are kept in a livewell and smaller fish are released and replaced by larger ones as necessary throughout the 8 to 10 hours of each tournament day. The final fish are transported to land, where they are weighed and released back into the water; dead fish brought to the weigh-in are assessed a penalty. In Saskatchewan, the popularity of catch-and-release tournaments has increased since their implementation in the 1970’s. In 2003, approximately 11,000 anglers participated in 133 licensed events, 10 of which were large professional tournaments that targeted walleye (Ashcroft et al. 2006). ! 10

Tournaments involve extra elements that exacerbate the standard effects of catch-and- release previously listed. Factors such as extended livewell confinement under poor conditions (e.g., high water temperature, crowded tank), increased air exposure as a result of weigh-in procedures, and high lake temperatures also impact fish (Gingerich et al.

2007; Sullivan et al. 2015). Research on walleye caught at angling tournaments has shown that these fish experience significant physiological disturbance, correlated with the stressors of angling and the weigh-in procedure (Killen et al. 2003; Killen et al. 2006).

However, how walleye move and behave once they are released has not been investigated. Furthermore, the relationship between post-release movement and fish treatment at tournaments is poorly understood. Therefore, my goal was to assess how this extreme version of catch-and-release affects the post-release movements of walleye and sauger.

1.4. Research interest #2: Niche ecology of sympatric species

Walleye and sauger are thought to use similar resources and habitats, though not always in the same location. In areas where populations of the two species exist sympatrically, there is potential for resource use overlap (Bellgraph et al. 2008). This also introduces the possibility of competition between the species, which is dependent on a variety of factors, including resource and space availability (MacLean and Magnuson

1977; Crowder 1990). Therefore, understanding resource use by populations of walleye and sauger is useful for understanding their competition potential, as well as their relationship in general. A common and useful technique for quantifying resource use and niche overlap is stable isotope analysis (SIA), which provides a long-term perspective on ! 11 diet. Using SIA to describe the niche of walleye and sauger populations, particularly in river systems, has recently been acknowledged as an important research need for these species (Baccante et al. 2011). Furthermore, rudimentary isotopic analysis has shown that diet overlap of walleye and sauger varies with environmental conditions, but is higher in riverine food webs than in large lakes (Fincel et al. 2015). However, there has not been any work done on the isotopic niche relationships of these species. As such, my second research objective was to use SIA to evaluate isotopic niche size and potential overlap of sympatric walleye and sauger in several Saskatchewan reservoir systems.

1.5. Research objectives

Given that walleye and sauger are currently managed as a single entity in

Saskatchewan, part of my aim in the current work was to examine similarities and differences between the species in different contexts that could be useful for improving their management. Therefore, the overall purpose of my thesis research was to address questions about how walleye and sauger respond to catch-and-release angling, as well as their resource use and niche in areas of co-occurrence. My specific research objectives were as follows:

1) Examine the short-term behaviour and movement of walleye and sauger

after catch-and-release at angling tournaments, and assess the

importance of a variety of tournament-related fish handling variables

for predicting behaviour after release

2) Quantify and compare the isotopic niche size and potential overlap of

walleye and sauger in two reservoirs ! 12

2. CHAPTER 2

Variable short-term movements by walleye (Sander vitreus) and sauger (S. canadensis) after catch-and-release at angling tournaments

2.1. INTRODUCTION

In North America, 25,000-30,000 catch-and-release angling tournaments occur in freshwater environments each year, creating significant concern for fisheries management

(Kerr and Kamke 2003; Schramm and Hunt 2007). By far the majority of these events are in warm water systems and target black bass (Micropterus spp) (Schramm et al. 1991).

However, walleye (Sander vitreus) tournaments have grown in popularity since the

1970’s, especially in regions that do not support bass populations (Schramm et al. 1991).

Catch-and-release tournaments for any freshwater species are associated with a range of management concerns. These include physical injury, immediate or delayed mortality, and disruption of normal movement, such as fish displacement and stockpiling near release areas (Kerr and Kamke 2003). The potential importance of these effects on bass caught at tournaments has been well studied, but remains comparatively understudied for walleye.

The major assumption underlying catch-and-release tournaments is that returning live fish to their source waterbody effectively conserves fish populations. Accordingly,

Canadian fisheries managers and tournament organizers implemented the concept of catch-and-release tournaments in the late 1980’s in an effort to reduce mortality and minimize the depletion of native walleye populations (Patalas 1996). However, despite good intentions, catch-and-release tournaments can dramatically affect fish mortality. ! 13

Immediate (pre-release) mortality rates at walleye tournaments in Wisconsin were estimated to be as high as 80%, likely due to high lake temperatures (Hoffman et al.

1996). Short-term (several hours after weigh-in) and long-term (5 days) delayed mortality at 14 walleye tournaments located in Michigan, Minnesota, the Dakotas, and Wisconsin were assessed by retaining a subset of fish in pens. Short-term mortality was as high as

54%, while long-term mortality was up to 100%, depending on conditions (Schramm et al. 2010). Mortality rates are highly variable and likely a reflection of the tournament conditions (both environmental and fish treatment). However, the specific tournament conditions that affect mortality and the efficacy of management for walleye survival remain uncertain.

Catch-and-release tournaments operate under the assumption that the behaviour of released fish is relatively unaffected. Brown et al. (2015) assessed the dispersal of largemouth bass (Micropterus salmoides) in response to simulated tournament displacement and found little evidence of long-term aggregation around the release area

(hereafter known as stockpiling). Most fish (87%) departed from the site of release (<500 m) within 21 days and 70% of the fish moved to an adjacent bay within 40 days. Though unable to compare the behaviour of disrupted fish to unaffected largemouth bass, Brown et al. (2015) concluded that the species copes with tournament practices and only made recommendations to limit displacement distance. In contrast, data for both largemouth and spotted bass (M. punctulatus) showed highly altered behaviour in displaced fish that experienced simulated tournament treatment, compared with control fish that were caught and released at their site of capture (Hunter and Maceina 2008). These authors found the average total distance moved after 10 weeks was much higher for displaced fish (7.4 km) ! 14 compared with control fish (1.3 km); the energy necessary for this extra movement could negatively impact growth, survival, and reproduction (Hunter and Meceina 2008). Wilde

(2003) indicated that research conclusions about bass dispersal after tournaments were contradictory, where rates of return to capture sites, fish dispersal, and recapture differed widely. It is therefore important to understand how differing catch-and-release conditions impact fish behaviour. Research has been limited on post-release behaviours, which include capacity to disperse from the release area, short-term movement rates and distances, and the ability to return to suitable habitats and environmental conditions.

There have been no studies on the post-release behaviours of tournament-caught walleye.

Increases in mortality and changes to fish behaviour are presumably linked to the levels of physiological stress experienced during catch-and-release. Contributors to stress are many, and at a minimum include several aspects directly related to angling, tournament handling, and environmental conditions. Stress levels can be influenced by angling and hooking injuries (Cooke et al. 2002; Arlinghaus et al. 2009), capture depth

(Talmage and Staples 2011), livewell conditions (Sullivan et al. 2015), and weigh station treatment (Suski et al. 2004). Stress is also related to holding time, air exposure, fish size, and water temperature (Maynard et al. 2013). A common method of quantifying stress is to assess a series of reflexes and their level of impairment. The reflex action mortality predictor (RAMP) test assesses reflex impairment and is a useful quantification of fish stress that can be rapidly measured in the field (Davis 2010). Different combinations of reflexes have been used for different species and collectively have been shown to effectively capture stress levels (Brownscombe et al. 2013; McArley and Herbert 2014). ! 15

However, the RAMP test was designed to predict mortality, so its utility for predicting fish behaviour after catch-and-release is not known.

In Saskatchewan, Canada, walleye comprise 41% of the annual recreational harvest, making them the most popular game species in the province (Sport Fishing,

2010). As a result, most tournaments in Saskatchewan target this species. However, walleye, the closely related sauger (Sander canadensis), and their hybrid (saugeye) are regulated as the same entity for recreational angling, where the catch limit is based on any combination of the three (Saskatchewan Anglers’ Guide, 2015). Thus, walleye tournaments in Saskatchewan also include sauger in systems where both species are present (Bozek et al. 2011a). Angling tournaments take place throughout the open water fishing season across a diverse range of lakes, which deviate in location, size, depth, salinity, and average temperature. The tournaments also vary in size, from small, individual events to larger ones that are part of a connected annual series (e.g., the

Saskatchewan Walleye Trail). Consequently, tournaments in Saskatchewan differ substantially in levels of funding, infrastructure, fish handling procedures, weigh station design, and rules and regulations. Currently, the understanding of how these varying conditions affect post-release fish behaviour is limited.

My goal was to use radio and acoustic telemetry to examine the short-term post- release behaviour of walleye and sauger after angling tournaments in Saskatchewan. My over-arching hypothesis was that a range of variables associated with angling and tournament procedures affect fish stress levels and thereby post-release behaviours such as dispersal and movement. My specific objectives were to: (1) document and describe short-term post-release movements for walleye and sauger at a number of tournaments ! 16 that varied in environmental conditions; and (2) assess the importance of tournament- related variables as predictors for fish behaviour. Based on studies in other fish species that have shown altered behaviour due to stress (Gravel and Cooke 2008; Arlinghaus et al. 2009), I expected fish to make shorter movements and take longer to leave the release area as stress levels increased.

2.2. MATERIALS & METHODS

2.2.1. Study sites.− I collected data at four catch-and-release walleye tournaments on three lakes in Saskatchewan during 2015 (Fig. 2.1). Two tournaments were held on Lake

Diefenbaker: the Riverhurst Walleye Classic (RH) on June 20 and 21, and the

Saskatchewan Landing Walleye Tournament (SL) on July 18 and 19. I also studied caught and released fish at the Last Mountain Fall Walleye Classic (LM) on Last

Mountain Lake (September 11 and 12), and the Nipawin Vanity Cup (NV) on Tobin

Lake (October 3 and 4). Lake Diefenbaker and Last Mountain Lake are located in the southern fishing zone of the province, while Tobin Lake is in the central zone

(Saskatchewan Anglers’ Guide, 2015). The tournaments differed substantially in size, entry fees, prize value, area fished, and time of year (Table 2.1). The livewell limit for all tournaments was 5 fish.

Lake Diefenbaker (51°01’N 106°50’W) is a large reservoir (max. length = 225 km; max. width = 6 km; surface area = 430 km2) created by two dams on the South

Saskatchewan and Qu’Appelle Rivers; it supports 26 fish species (State of Lake

Diefenbaker, 2012). The lake has a mean depth of 21.6 m (maximum 66 m). Last

Mountain Lake (51°06’N 105°15’W) is southern Saskatchewan’s largest natural ! 17

NVTB

Saskatoon SLDF LM

Regina RH

150 km

!

FigureDF 2.1 . Locations of the four tournaments at which data were collected in

Saskatchewan, Canada in 2015. Abbreviations are as follows: RH = Riverhurst Walleye

Classic; SL = Saskatchewan Landing Walleye Tournament; LM = Last Mountain Fall

Classic; NV = Nipawin Vanity Cup.

TB

! 18

Table 2.1. Summary information for the four tournaments examined in this study. The dates of the tournament in 2015 and the dates fish were tracked using telemetry are listed. The entry fee for each team and the value of the first prize is listed, as well as the maximum number of teams that could register for the event (Max. teams). The area of the fishing zones within tournament boundaries

(km2) and the maximum distance anglers could travel to the weigh station (Max. dist.; km) are also included. Fish information, including number of fish that were tagged at each tournament by species (W=Walleye, S=Sauger), and mean and range for total lengths (cm) of tagged fish are provided.

Tournament Lake Tourn. / Mean Entry Max. Area Max. Tracking Species Number Mean TL±SD (range) Track. surface Fee / 1st teams (km2) dist. technology of fish (cm) dates temp Prize (km) (°C) Riverhurst Lake June 20, 18 $200 / 106 65 18 Radio W, S 15 W W: 56.9±6.7 (47 – 72) Walleye Diefenbaker 21 / June $5,000 5 S S: 43.2±4.6 (40 – 47) Classic (RH) 21-27

Saskatchewan Lake July 18, 19.1 $310 / 125 68 35 Radio W, S 15 W W: 60.6±10.2 (43 – 76) Landing Diefenbaker 19 / July $10,000 5 S S: 44.5±1.8 (42 – 47) Walleye 19-25 Tournament (SL)

Last Mountain Last Sept. 11, 15.3 $610 / 120 40 10 Acoustic W 10 W W: 55.8±9.2 (43 – 70) Fall Classic Mountain 12 / $15,000 (LM) Lake Sept. 12- 19

! 19

Nipawin Tobin Lake Oct. 3, 4 / 11 $2,000 / 160 35 20 Acoustic W1 16 W W: 60.4±11.9 (48 – 76) Vanity Cup Oct. 4-10 $100,000 (NV)

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! 1 Both walleye and sauger are present in Tobin Lake but only walleye were weighed-in that fit within the size specifications for the acoustic tags during the 2015 event. ! 20 waterbody. Like Lake Diefenbaker, it is long and narrow (max. length = 93 km; max. width = 4.5 km; surface area = 233 km2) supports a variety of fish species and has a mean depth of 7.6 m (maximum 31.5 m). Tobin Lake (53°35’N 103°30’W) is a large reservoir formed by a dam on the North Saskatchewan River with a surface area of 228 km2 and a mean depth of 8 m (maximum 26.4 m) (Durbin 1981). This reservoir is composed of two unique basins: the first is riverine and parallels the original Saskatchewan River valley, while the second is much larger and lake-like (Warwick and Tisdale 1988). Tobin Lake supports around 25 species of fish and is one of the best walleye fishing locations in the province; the Canadian ice fishing record for walleye (18.3 lbs) was caught there in 2005

(“Walleye”, 2013).

2.2.2. Fish capture & handling.− Walleye and sauger were angled by rod and reel by competitors within set tournament boundaries. Each tournament occurred over the course of two days, during which anglers attempted to catch both the largest fish and largest total weight. Total weight was based on a maximum of 5 fish per day, of which 2 fish were allowed to be over the slot limit (dependent on lake; commonly 55 cm) or all 5 could be under. Anglers generally caught 5 fish as soon as possible each day of competition, and then substituted larger fish if they were caught (referred to as “culling” in bass tournaments; Schramm et al. 1987). As a result, fish brought to the weigh-in at the end of the day spent varying amounts of time in the livewell. At weigh-in, officials checked all fish to determine whether they were alive or dead; fish that were dead were assessed a weight penalty. Once checked, anglers transported their fish from the livewell to the weigh scale in dry plastic boxes for weighing. ! 21

I selected individual walleye and sauger for radio or acoustic tag application immediately after they were weighed and registered. Selected fish were directly transferred to a 120-L tank receiving flow-through lake surface water from a portable pump. I tagged approximately the same number of fish from three common size classes

(40-52 cm, 52-64 cm, 64 cm+) at each tournament. These categories provided a range of sizes for assessment; the minimum length of fish to which I applied a tag was 40 cm.

Importantly, my selection of fish was not entirely random within each size class; I avoided individuals that were visibly moribund, showed signs of extensive discoloration or bleeding, or had obvious external injuries.

2.2.3. Angler survey.− I interviewed the angler who caught each selected fish to collect data on a number of aspects of fish treatment prior to release (see 6.1 Appendix I). The survey included having anglers mark capture location on a map, and questions regarding depth of capture and length of time the fish was retained (rounded to the nearest 1 hour).

For some fish I was unable to obtain these data and did not include them in the multivariate modeling analyses described below.

2.2.4. Fish stress assessment.− Prior to tagging, I performed a stress assessment on each fish using a modified version of the RAMP assessment (Davis 2010). I rapidly checked for the presence or absence of five reflexes (reflex present = 1, absent = 2); if there was doubt that the reflex was present, the fish was assigned a status of impaired. The reflexes assessed were: (1) body flex when restrained; (2) dorsal fin flare when restrained; (3) operculum closed when manually opened; (4) V-O test where the eye maintains a level ! 22 pitch when the fish is rolled on its side; and (5) orientation, where the fish is able to right itself within 3 seconds after being turned on its side. Impairment of these reflex responses is correlated with stress and increased likelihood of mortality (Davis 2010).

Additionally, I assigned a quantitative swimming behaviour score to each tagged fish after release based on the following scale: (0) normal swimming behaviour with rapid movement away from release site; (1) general normal swimming behaviour but with some visible signs of fatigue, such as slow swimming; (2) random and erratic swimming behaviour; (3) fish appeared to passively descend; (4) fish displayed partial lateral equilibrium loss; (5) fish displayed complete equilibrium loss (modified from Gingerich et al. 2007).

The RAMP score for all three reflexes and swimming behaviour scores were summed for each fish to generate a cumulative stress score (i.e., 1 + 2 + 1 (reflexes) + 1

(swimming) = 5 (total score)). Of the five RAMP reflex responses I tested, I found that three varied in their possible response at each tournament (i.e. presence or absence of response). These reflexes were the body flex, the operculum close, and orientation. The other two tests (dorsal fin flare and V-O test) were either always absent or always present, and for this reason I chose not to include them in the final stress score. Higher values for the total score indicate more stress; the possible range of stress scores values was 3-11.

2.2.5. Application of radio or acoustic transmitters.− Following RAMP assessment, each selected fish was fitted with an externally attached radio or acoustic transmitter. I opted for external attachment over surgically implanting the tags to avoid the additional ! 23 stressors of anesthesia and surgery. To begin, I inserted a FT-4 cinch-up tag (Floy

Manufacturing, Seattle, Washington) with a unique numeric identifier into the tissue behind the first dorsal fin using a sterile stainless steel 14-gauge applicator. This is essentially a modified Carlin tag (Carlin 1955), which provides a secure anchor for externally attaching a tracking device in a way where it will release and fall off after a short-term study.

Radio (Lotek MCFT2-3A and MCFT2-3BM) and acoustic tags (Vemco V9-2x,

V9T-2x, and V13-1x) with external mounting caps were used for my study. The external mounting cap was attached to the FT-4 cinch-up tag via a loop of Monocryl 2-0 dissolvable suture. This suture material dissolves in approximately 21-30 days, allowing the radio or acoustic tag to release (modified tagging methods from Kaintz and Bettoli

2010). The tagging process took approximately 2 minutes, during which fish were kept submerged in the holding tank to limit air exposure. Following tag application, the fish were immediately released following the same protocol as other fish in the tournament.

The nature of release varied with tournament and included manually releasing fish from a dock or via a PVC chute filled with running water. After release, tagged fish were monitored for initial mortalities. Fish that died immediately after release were collected from the water and their tag was recovered.

2.2.6. Fish tracking.− I used radio telemetry at the RH and SL tournaments. I tracked fish using a handheld receiver (Lotek SRX 600 Telemetry Receiver, Lotek, Ontario, Canada) and a combination of a portable three-element Yagi antenna and a larger five-element antenna affixed to a 3-m mast on the bow of the research boat (Legend Extreme, 20.5 ! 24 feet). Acoustic telemetry was performed manually to track fish after the LM and NV tournaments, using a handheld receiver (Vemco VR-100 Acoustic Tracking Receiver,

Vemco, Nova Scotia, Canada) and a VH110 directional hydrophone.

Each tagged fish was manually tracked for 7 days following the tournament; mortality studies have indicated this is the period in which acute effects from stress, generally reflected by delayed mortality, are most significant (Patalas 1996; Schramm et al. 2010). Tracking began the day after the fish was tagged; attempts were made to find each fish once per day, although this was not always possible. For both telemetry technologies, I tracked fish by driving along transects throughout the search area. Once a signal was detected, I travelled until I reached a location in which the signal from the tag was equally strong in the four primary directions. I recorded this as the location of the fish using a handheld global positioning system (Garmin, etrex 20, Kansas, USA) with a precision of ±5 m. If the tag had a temperature or motion sensor, the data from this sensor were manually recorded. The depth at each location was measured using a sonar depth sounder.

I searched as much area as possible during daylight hours to find all tagged fish.

The range of detection for the radio tags was estimated to be approximately 100 to 300 m, while the range for the acoustic tags was estimated at 700 to 1000 m (varied by location). These approximations were based on field observations from the known locations of tags on non-mobile fish; the search pattern I used at each location (location and density of transects) was adjusted to accommodate these ranges.

2.2.7. Post-tournament behaviour. ! 25

2.2.7.1. General fish movement.− The total minimum distance a fish moved was measured based on the cumulative distance between daily relocation points. When a fish was located more than once in a 24-hour period, the last location recorded in the day was used to calculate total net distance moved. Average daily rate of movement was calculated for fish at LM and NV by dividing the total distance moved by the time in days that passed until the fish was found for the final time. This metric was only calculated for these tournaments because fish were tracked with acoustic tags, which enabled more consistent and reliable relocations. All fish were treated the same way for this calculation, regardless of the number of days they were relocated. Using the responses from the angler surveys, I also calculated the shortest distance each fish would have been transported from their capture location to the weigh station, defined as

“distance transported in livewell”. All movement data were analyzed using ArcMap 10.2

(Environmental Systems Research Institute 2015).

2.2.7.2. Departure from release area.− I defined the release area as a 500-m radius around the tournament weigh station (as in Brown et al. 2015). Fish that left this area were considered to have dispersed from the release site. I recorded the number of hours

(in 24-hr intervals) required for each fish to leave the 500-m buffer zone; a value that ranged from ≤24 to 168. Fish that never left the release area during the 7 days of tracking were given a designation of 168+. Dispersal was compared with established criteria, defined as immediate (1-2 days), delayed (3-4 days), and extended (5+ days) (adapted from Kaintz and Bettoli 2010).

! 26

2.2.7.3. Fish mortality.− I confirmed mortality events based on observation of carcasses or tracking of radio-signals to predators (American white pelicans; Pelecanus erythrorhynchos). The location of fish inside of pelicans was confirmed by both the location of the signal and the temperature sensing function of radio tags, which indicated temperatures over 35°C (much higher than possible for fish in water). I did not classify mortality based on lack of fish movement.

2.2.7.4. Statistical analyses.− One-way Analysis of Variance (ANOVA) was used to test for differences among values for predictor (time in livewell, distance transported in livewell, depth of capture, stress score) and response (days to exit release area, total distance moved in first 48 hours, rate of movement) variables across tournaments (p ≤

0.05). If a significant main effect was obtained, a post-hoc Tukey test was used to identify specific pair-wise differences.

I built generalized linear models (GLM) and used model selection based on

Akaike’s Information Criterion (AIC) to test hypotheses about relationships between capture and handling predictor variables and fish movement. I generated models independently for each tournament, which was necessary due to large differences in conditions, as well as the switch in telemetry technologies (radio vs. acoustic). To determine whether the predictor variables (total length, depth of capture, distance transported in livewell, time in livewell, stress score, and species) were correlated to each other, I ran a series of Pearson Correlation tests. Time in livewell was correlated with another variable with r > 0.7 at more than one tournament, so I eliminated it from my models. Species was only in the models for the first two tournaments (RH and SL) as I ! 27 only tagged walleye at the second two (LM and NV). I ran GLMs for several dependent variables that were measures of fish behaviour: time to leave release area (days), total distance moved in first 48 hours, and daily rate of movement (for only LM and NV). I also ran a separate model for SL to assess factors affecting confirmed mortalities. All models were run with a Gamma error distribution and log link function except the mortality model, which was run with a binomial error distribution.

Maximum likelihoods from all possible subsets of each global model containing all predictors were used to obtain AIC scores and build the candidate set of explanatory models. AICc was used to correct for small sample sizes (Grueber et al. 2011). I calculated the number of parameters (k) as 2 (1 for the intercept and 1 for the random factor) + 1 for each model parameter (Barton 2013). The top or most parsimonious model was that with the lowest AIC score (i.e. greatest support) out of the candidate set. Model averaging using the natural average method was performed using all models within 2

ΔAIC units of the top model (Burnham et al. 2011). Inference was made from parameters whose confidence intervals did not pass through 0. All statistical analyses were conducted in R Project for Statistical Computing 3.3.0 (R Development Core Team

2016), using packages lme4 (version 1.1-12; Bates et al. 2012) and MuMIn (version

1.15.6; Barton 2013).

2.3. RESULTS

I tracked 66 fish (56 walleye over four tournaments, 10 sauger over two tournaments) using telemetry at four angling tournaments in Saskatchewan (Table 2.1).

All tagged fish except for one at RH were found at least once. Sauger large enough to be ! 28 tagged (TL > 40 cm) were only caught at two tournaments. There were also 5 fish (1 walleye, 4 sauger; all at SL) that died immediately after release and were removed from the study before any tracking data were collected. I observed confirmed mortalities at only one of the four tournaments studied. SL had nine confirmed mortalities (2 sauger, 7 walleye): three fish were found dead floating in the water and six fish were confirmed to have been depredated by American white pelicans. Multivariate modeling for mortality at

SL indicated a weak relationship with total length; smaller fish were more likely to die

(Table 2.6). Depth of capture was also included in the model averaging, showing fish caught at shallower depths were more likely to die; however, the 95% confidence interval of this variable passed through zero, indicating that the influence of depth was minimal.

The accelerometers in the radio tags were initially used to designate fish as suspected mortalities, based on several consecutive Inactive codes. However, I found the motion sensors would sometimes switch back to the Active code; additionally, there were fish at each tournament that would remain in one place for several days and then began moving again. I concluded that accelerometers and lack of fish movement were unreliable indicators of mortality and did not include suspected mortalities in this analysis.

2.3.1. Angler survey.− I was able to collect survey data from 75 to 100% of anglers who caught fish I chose to tag, depending on the tournament (Table 2.2). The mean depth of capture varied significantly by tournament (Table 2.2; ANOVA, F(3,52)=42.01, p<0.001). Post-hoc analysis revealed that fish at LM were caught at deeper depths than at the other three tournaments. Fish at NV were caught at intermediate depths while fish at

RH and SL were caught at similar shallow depths. The mean reported distance fish were ! 29 transported from their site of capture to the weigh station ranged widely both within and between events, but was significantly different by tournament (Table 2.2; ANOVA,

F(3,52)=14.10, p<0.001). Post-hoc analysis showed that fish were transported significantly farther at SL than any of the other tournaments. Values for the other three tournaments were not significantly different. The time that fish were kept in the livewell was generally much less than the total period available (8 hours), and did not vary significantly by tournament (Table 2.2; ANOVA, F(3,52)=2.25, p=0.09).

2.3.2. Fish stress assessment.− The mean scores for each reflex response varied with tournament (Table 2.3). NV consistently had the lowest mean score for all three reflexes.

For swimming scores, SL had the highest mean while NV had the lowest (Table 2.3).

When comparing species overall stress scores, walleye at SL had a similar mean score

(4.20±1.14) to RH (4.42±1.74). In contrast, the score for sauger at SL (6.40±1.14) was higher than at RH (5.40±2.07). Overall, total stress score varied among tournaments (Fig.

2.2; ANOVA, F(3,61)=2.88, p=0.04). SL had the highest overall (4.75±1.48) score, which was significantly different from the lowest mean score at NV (3.50±0.73). Mean stress scores for RH (4.68±1.83) and LM (4.60±1.17) were intermediate and not significantly different from those for SL or NV.

2.3.3. Fish tracking.− Using radiotelemetry, an average of 14 out of 20 (70%) fish were relocated each day for RH and 9 out of 20 (45%) for SL. In contrast, using acoustic telemetry I relocated an average of 9 out of 10 (90%) fish daily for LM, and 16 out of 16 ! 30

Table 2.2. Summary of angler survey responses. Includes number of surveys successfully received, and mean values for the answers to several of the questions: time fish were held in livewell (hours), minimum distance fish were transported from the site of capture to the weigh station (km), and depth at which fish were caught (m). All mean values are expressed ±SD; the total range of values is provided.

Tournament Number of Time in Distance Depth of angler surveys livewell transported in capture (m) (total possible) (hours) livewell (km) Riverhurst 15 (20) 5.9±1.8 5.1±4.4 4.5±1.1 Walleye (1-7) (1.2-18.7) (2.4-6.4) Classic (RH)

Saskatchewan 16 (20) 4.0±2.0 18.9±9.5 4.2±0.9 Landing (1-8) (1.4-31.4) (2.7-5.5) Walleye Tournament (SL)

Last Mountain 10 (10) 5.6±1.8 10.7±3.6 8.6±1.6 Fall Classic (1-7) (3.8-14.7) (6.7-12.2) (LM)

Nipawin 16 (16) 4.9±2.5 8.9±3.8 6.0±0.7 Vanity Cup (1-8) (2.1-14.0) (4.9-7.6) (NV)

! 31

Table 2.3. Summary of components of stress test performed on walleye and sauger.

Includes mean (±SD) values for body flex, operculum close, orientation reflexes (from

RAMP test), and swim score. RAMP test values are based on presence / absence of a reflex and are either 1 or 2. The swim score is based on different categories of behaviour and ranges from 0 to 5.

Tournament Body flex Operculum Orientation score Swim score score close score Riverhurst 1.16±0.37 1.32±0.48 1.37±0.50 0.84±1.21 Walleye (0-4) Classic (RH)

Saskatchewan 1.35±0.49 1.15±0.37 1.20±0.41 1.05±1.19 Landing (0-3) Walleye Tournament (SL)

Last Mountain 1.30±0.48 1.20±0.42 1.20±0.42 0.90±0.88 Fall Classic (0-2) (LM)

Nipawin 1.13±0.34 1.13±0.34 1.00±0.00 0.25±0.45 Vanity Cup (0-1) (NV)

! 32

!

Figure 2.2. Box-and-whisker plot of overall stress scores for all fish at four tournaments.

Each box encompasses the middle 50% of the score distribution, with the median indicated by the horizontal line in the interior of the box. Whiskers extend to maximum and minimum values that are within 3/2 times of interquartile range. Outliers are represented by open circles. Stress scores were calculated by summing the scores for the three reflex tests and the swimming score. Abbreviations are as follows: RH = Riverhurst

Walleye Classic; SL = Saskatchewan Landing Walleye Tournament; LM = Last

Mountain Fall Classic; NV = Nipawin Vanity Cup.

! 33

(100%) for NV. The proportion of fish relocated daily varied substantially for RH and

SL; I was more successful with daily relocations at the beginning of the tracking period, and lost contact with more fish as time progressed (Fig. 2.3). There were no differences in the number of relocations between walleye and sauger. In contrast, I consistently relocated almost all tagged fish daily for both LM and NV, and at both events 100% of tagged fish were relocated for several consecutive days (Fig. 2.3).

2.3.4. Departure from release area.− I determined a minimum value for the number of days it took each fish to leave the release area except for 7 fish at SL; these fish were confirmed mortalities before they were able to exit. The mean number of days to exit the release area varied significantly by tournament (Table 2.4; ANOVA, F(3, 54)=3.92, p=0.01). Post-hoc analysis showed that fish at NV exited more rapidly than those from

RH. Values for fish from SL and LM were not significantly different from NV but were more similar to those of RH. Fish at NV left the release area in a manner similar to patterns established by Kaintz and Bettoli (2010), where 80% of fish dispersal was immediate. In contrast, dispersal of a large (30-40%) portion of fish at RH and LM was either delayed or extended. SL followed a similar trend, though dispersal proportions at this tournament were affected by the 40% of fish that were mortalities before exiting

(Fig. 2.4).

For RH and LM, the top model for days to exit showed the intercept to be the most important, indicating none of the predictor variables had more support than the null model (Tables 2.5, 2.7). In contrast, stress score, distance transported, and species were ! 34

1 0.9 0.8 0.7 0.6 RH 0.5 SL 0.4 0.3 LM 0.2 NV Proportion of Fish Found Proportion 0.1 0 1 2 3 4 5 6 7 Day of Tracking !

Figure 2.3. Proportion of total tagged walleye and sauger found for seven days of tracking after four angling tournaments. Radio telemetry was used for RH and SL while acoustic telemetry was used for LM and NV. Abbreviations are as follows: RH =

Riverhurst Walleye Classic; SL = Saskatchewan Landing Walleye Tournament; LM =

Last Mountain Fall Classic; NV = Nipawin Vanity Cup.

! 35

Table 2.4. Summary of telemetry data from the four angling tournaments in

Saskatchewan that were the basis for this study. Includes mean days to exit the release area, mean total distance moved in the first 48 hours after release (m), and mean rate of movement (m/ day) with ranges of values. Also provided is the number of confirmed mortalities.

Tournament Days to exit Total distance in Rate of Confirmed first 48 hrs (m) movement (m / mortalities day) Riverhurst 4.4±1.7 723.7±1,087.9 N/A 0 / 20 Walleye (2-8) (25-4,717) Classic (RH) n=19 n=18

Saskatchewan 3.9±1.6 1065.3±687.8 N/A 9 / 20 Landing (2-7) (308-1,893) Walleye n=13 n=8 Tournament (SL)

Last Mountain 4.3±2.2 608.1±392.1 935.6±1,018.3 0 / 10 Fall Classic (1-9) (147-1,120) (108-3,435) (LM) n=10 n=7

Nipawin 2.6±1.1 3,712.8±4,427.9 1,235.1±1,051.3 0 / 16 Vanity Cup (2-6) (388-15,262) (129-3,048) (NV) n=16 n=16

! 36

1 0.9 0.8 0.7 0.6 0.5 RH 0.4 SL 0.3 LM Proportion of Fish Proportion 0.2 0.1 NV 0 1-2 days 3-4 days 5+ days Dead before leaving Periods to Exit Release Area !

Figure 2.4. Proportion of tagged walleye and sauger that exited release area (swam

>500m) during three time periods. These time periods reflect defined stages of dispersal: immediate (1-2 days), delayed (3-4 days), and extended (5+ days) (adapted from Kaintz and Bettoli 2010). Also included is the “dead before leaving” category, which indicates fish that were confirmed mortalities before they left the release area. Abbreviations are as follows: RH = Riverhurst Walleye Classic; SL = Saskatchewan Landing Walleye

Tournament; LM = Last Mountain Fall Classic; NV = Nipawin Vanity Cup.

! 37 all part of the top model for SL, though a competing model used in the model averaging was the null (Table 2.6). Both relationships for stress score and distance transported in this model were positive, though weak, and indicated that more stressed fish or those transported farther in a livewell took longer to leave the release area. Stress score and total length were both important variables in the top model for NV (Table 2.8); these relationships, though weak, showed less stressed and larger fish took longer to exit the release area. Depth of capture was included in the averaged model for NV but its confidence interval passed through zero.

2.3.5. Movement in 48 hours.− I determined the total initial distance fish moved by calculating the distance for the first 48 hours after release. I was unable to calculate the distance for some fish, as they were not located on day 2 of tracking (2 fish at RH, 12 fish at SL, 3 fish at LM). Total distance traveled during the first 48 hours varied significantly by tournament (Table 2.4; ANOVA, F(3,45)=4.40, p=0.008). Post-hoc analysis indicated fish at NV moved farther in the first 48 hours after release than fish at RH, SL, and LM.

The top models for RH and LM showed the intercept to be the most important for predicting movement in the first 48 hours after release, indicating that the predictor variables were at best, marginally important (Tables 2.5, 2.7). For SL, species and stress score were part of the top model for predicting movement. Both relationships were relatively strong; the model showed that fish moved farther in the first 48 hours with every three-fold increase in stress level. This model actually found all predictor variables to be important but only these two variables had an estimate different from zero (Table

2.6). Stress score and total length were important predictor variables for NV (Table 2.8). ! 38

Table 2.5. Top model, intercept-only model, and all models within 2 delta Akaike information criterion (ΔAIC) units of the top model (candidate set) explaining fish behaviour at the Riverhurst Walleye Classic. The two behaviours modeled here are time it took to leave the release area (exit) and total distance moved in the first 48 hours after release (48 hours). Model averaging is also included with parameter estimates, standard errors, and 95% confidence intervals (CIs). Abbreviations used for variables are

DistTrans = distance transported in livewell and DepthCapt = depth of capture.

Exit Model k AIC ΔAIC wi Top and Intercept only 2 59.6 0 0.354 competing Stress 3 61.5 1.87 0.139 models Ad Hoc - DistTrans 3 62.4 2.76 0.089

Parameter Estimate Standard 95% CI 95% CI Relative Error (lower) (upper) importance Model Intercept 1.118 0.532 0.415 1.774 averaging Stress 0.076 0.018 -0.053 0.227 0.282

48 hours Top and Model k AIC ΔAIC wi competing Intercept only 2 193.7 0 0.349 models Species 3 194.9 1.28 0.184 Ad Hoc – DistTrans 3 196.7 3.04 0.076 Parameter Estimate Standard 95% CI 95% CI Relative Model Error (lower) (upper) importance averaging Intercept 5.812 0.411 5.103 6.741 Species 0.689 0.468 -0.323 1.547 0.35

!

!

!

!

! ! 39

Table 2.6. Top model, intercept-only model, and all models within 2 delta Akaike information criterion (ΔAIC) units of the top model (candidate set) explaining fish behaviour at the Saskatchewan Landing Walleye Tournament. The two behaviours modeled here are time it took to leave the release area (exit) and total distance moved in the first 48 hours after release (48 hours). Confirmed mortality at this tournament was also modeled. Model averaging is also included where necessary with parameter estimates, standard errors, and 95% confidence intervals (CIs). Abbreviations used for variables are DistTrans = distance transported in livewell and DepthCapt = depth of capture.

Exit Top and Model k AIC ΔAIC wi competing DistTrans + Species + Stress 5 26.9 0 0.430 models Intercept only 2 28.3 1.40 0.213 Ad Hoc - DistTrans 3 29.8 2.88 0.102 Parameter Estimate Standard 95% CI 95% CI Relative Error (lower) (upper) importance Model Intercept 0.123 0.027 0.070 0.176 averaging DistTrans 0.003 0.001 0.002 0.005 0.67 Species -0.139 0.025 -0.188 -0.088 0.67 Stress 0.055 0.008 0.039 0.069 0.67

48 hours Model k AIC ΔAIC wi DepthCapt + DistTrans + Species + Stress + TL 7 -5.3 0 1.00 Ad Hoc – Intercept only 2 117.1 122.42 0.00 Parameter Estimate Standard Error 95% CI 95% CI (lower) (upper) Intercept 8.683 3.449 1.923 15.442 DeptCapt -0.854 0.463 -1.762 0.054 DistTrans 0.512 0.270 -0.018 1.042 Species -7.840 3.534 -14.766 -0.913 Stress 3.448 1.609 0.295 6.601 TL -0.139 0.030 -0.307 0.028

Mortality Top and Model k AIC ΔAIC wi competing TL 2 21.8 0 0.251 ! 40 models DepthCapt + TL 3 23.8 2.00 0.090 Intercept only 1 24.2 2.45 0.074 Parameter Estimate Standard 95% CI 95% CI Relative Model Error (lower) (upper) importance averaging Intercept 11.412 5.882 1.403 25.800 DepthCapt -0.201 0.204 -0.645 0.188 0.26 TL -0.142 0.071 -0.320 -0.025 1.00 !

!

! 41

Table 2.7. Top model, intercept-only model, and all models within 2 delta Akaike information criterion (ΔAIC) units of the top model (candidate set) explaining fish behaviour at the Last Mountain Fall Classic. The three behaviours modeled here are time it took to leave the release area (exit), total distance moved in the first 48 hours after release (48 hours), and movement rate during the week of tracking. Model averaging is also included where necessary with parameter estimates, standard errors, and 95% confidence intervals (CIs). Abbreviations used for variables are DistTrans = distance transported in livewell and DepthCapt = depth of capture.

Exit Model k AIC ΔAIC wi Top and Intercept Only 2 47.9 0 0.406 competing TL 3 49.0 1.07 0.238 models Ad hoc - Stress 3 50.0 2.11 0.142 Parameter Estimate Standard 95% CI 95% CI Relative Model Error (lower) (upper) importance averaging Intercept -0.171 0.251 -0.660 0.329 TL 0.008 0.005 -0.002 0.017 0.37

48 hours Model k AIC ΔAIC wi Intercept only 2 108.6 0 0.806 Ad Hoc – TL 3 112.9 4.32 0.093 Parameter Estimate Standard Error 95% CI 95% CI (lower) (upper) Intercept 6.009 1.439 3.225 8.916

Movement rate Model k AIC ΔAIC wi Stress 3 160.1 0 0.550 Ad Hoc – Intercept Only 2 162.4 2.39 0.166 Parameter Estimate Standard Error 95% CI 95% CI (lower) (upper) Intercept 4.024 2.741 2.309 5.918 Stress 0.566 0.305 0.187 0.948 !

! ! 42

Table 2.8. Top model, intercept-only model, and all models within 2 delta Akaike information criterion (ΔAIC) units of the top model (candidate set) explaining fish behaviour at the Nipawin Vanity Cup. The three behaviours modeled here are time it took to leave the release area (exit), total distance moved in the first 48 hours after release (48 hours), and movement rate in the week of tracking. Model averaging is also included where necessary with parameter estimates, standard errors, and 95% confidence intervals

(CIs). Abbreviations used for variables are DistTrans = distance transported in livewell and DepthCapt = depth of capture.

Exit Model k AIC ΔAIC wi Top and Stress + TL 4 34.7 0 0.403 competing Stress 3 35.9 1.19 0.223 models DepthCapt + Stress + TL 5 36.5 1.81 0.163 Ad hoc – DepthCapt + Stress 4 37.6 2.86 0.096 Intercept only 2 47.0 12.35 0.001 Parameter Estimate Standard 95% CI 95% CI Relative Error (lower) (upper) importance Model Intercept 0.346 0.197 -0.036 0.738 averaging Stress -0.133 0.026 -0.183 -0.080 1.00 TL 0.004 0.002 0.001 0.008 0.72 DepthCapt 0.014 0.009 -0.004 0.032 0.21

48 hours Model k AIC ΔAIC wi Top and Stress 3 289.3 0 0.336 competing Stress + TL 4 289.7 0.41 0.273 models DistTrans + Stress + TL 5 290.9 1.65 0.147 Ad Hoc – DistTrans + Stress 4 291.5 2.28 0.107 Intercept only 2 300.2 10.90 0.001 Parameter Estimate Standard 95% CI 95% CI Relative Error (lower) (upper) importance Model Intercept 14.346 1.419 12.056 16.767 averaging Stress -1.118 0.252 -1.591 -0.608 1.00 TL -0.031 0.015 -0.059 -0.003 0.56 DistTrans -0.082 0.051 -0.176 0.015 0.19

Movement rate Model k AIC ΔAIC wi Top and DistTrans + Stress + TL 5 258.2 0 0.339 competing Stress 3 259.6 1.41 0.167 ! 43 models Ad hoc – DistTrans + Stress 4 260.3 2.11 0.118 Intercept only 2 264.3 6.10 0.016 Parameter Estimate Standard 95% CI 95% CI Relative Error (lower) (upper) importance Model Intercept 12.866 1.535 9.765 16.096 averaging Stress -0.717 0.273 -1.285 -0.118 1.00 DistTrans -0.137 0.055 -0.243 -0.025 0.67 TL -0.038 0.017 -0.071 -0.004 0.67

! 44

The relationship with size was relatively weak; as fish got larger, the initial distance they traveled decreased. The relationship to stress was stronger, where an increase in stress also caused initial fish movement to decrease. In addition, distance transported was part of the competing model but the 95% confidence interval passed through zero, indicating its effect was minimal.

2.3.6. Overall movement.− Total fish movement during the week of tracking ranged widely. Movement at RH ranged from three walleye moving under 1 km, 10 walleye moving between 1 and 9 km, and a single walleye moving 18 km (Fig. 2.5a, b). All five sauger at this tournament moved less than 2 km. At SL, discounting fish that were mortalities, three fish moved under 1 km (1 walleye, 2 sauger), six fish moved between 1 and 9 km (5 walleye, 1 sauger) and two walleye moved large distances (10 and 16 km).

Of the fish that were mortalities, seven moved less than 1 km before dying (6 walleye, 1 sauger). Additionally, 1 walleye moved 1.2 km and 1 sauger moved 2.2 km, both over four days before dying (Fig. 2.5c, d). The movement at LM ranged from one walleye moving under 1 km, six walleye moving between 1 and 9 km, to three walleye moving large distances (10, 16, and 31 km; Fig. 2.6a, b). Finally, walleye movement at NV showed zero fish moving less than 1 km, nine fish moving between 1 and 9 km, and seven fish moving large distances (10, 14, 18, 18.5, 19, 22, and 24 km; Fig. 2.6c, d).

The mean rate of movement was calculated for walleye at LM and NV (Table

2.4), and it did not significantly vary between these two tournaments (ANOVA, F(1,

24)=0.51, p=0.481). The top model for LM indicated stress was the only variable that predicted movement rate (Table 2.7). This model had a weak, positive relationship where ! 45

a b

d c

!

! 46

Figure 2.5. Maps of the minimum path fish tracked using radio telemetry traveled after the Riverhurst Walleye Classic (a, b) and the Saskatchewan Landing Walleye

Tournament (c, d). Panels (a) and (c) show all fish at both tournaments that moved greater than 2 km while panels (b) and (d) show fish paths less than 2 km. Note differences in scale between maps of limited and significant movement. Paths made by sauger are indicated by red lines. The location of the weigh station site is indicated by a yellow circle. ! 47

a b

c d

!

! 48

Figure 2.6. Maps of the minimum path fish tracked using acoustic telemetry traveled after the Last Mountain Fall Classic (a, b) and the Nipawin Vanity Cup (c, d). Panels (a) and (c) show all fish at both tournaments that moved greater than 5 km while panels (b) and (d) show fish paths less than 5 km. Note differences in scale between maps of limited and significant movement. The location of the weigh station site is indicated by a yellow circle.

! 49 movement rate of fish increased with stress level. The top model for NV showed total length, stress score, and distance transported were all important variables (Table 2.8). All relationships for these variables were weak and negative; movement rate increased with lower stress levels, smaller fish, and smaller livewell transport distances.

2.4. DISCUSSION

Walleye and sauger made highly variable movements after catch-and-release both within and among tournaments. My data showed a 15 to 30-fold difference between the shortest and longest minimum distance moved, depending on the tournament. Walleye exhibited more movement at LM and NV, where many fish moved greater distances and fewer fish made limited movements. The farthest a fish moved during a week of tracking was 31 km at LM, but my data also showed four fish at NV moving around 20 km, a considerable distance. All fish that made long-distance movements (>20 km) were walleye. At the two tournaments where sauger were tagged, movements by this species were limited and the longest observed movement was by a sauger that eventually died (2 km). The large movements I observed for walleye contrast those documented for the frequently studied black bass. Limited initial dispersal is common after tournaments, where both largemouth and smallmouth bass remain within a 1-2 km radius of their release site for months (Gilliland 1999; Bunt et al. 2002). Bass normally maintain relatively small home ranges (Hunter and Maceina 2008; Ahrenstorff et al. 2009), so the influence of catch-and-release on their movements is unclear. Bass movement can be larger seasonally, with rates highest in the spring during spawning (7.3 km/day), and lowest during colder months (2.2 km/day) (Hanson et al. 2007). In contrast, walleye ! 50 migrate extensively (up to 100s of km) to access spawning grounds or foraging areas

(Hayden et al. 2014b; Haxton et al. 2015), which partly explains the observed difference in patterns after catch-and-release tournaments. However, walleye and sauger movement during the non-spawning season (May to November; similar to the timing of the tournaments studied here) is generally much more limited (Bellgraph et al. 2008). I conclude that walleye are capable of making long-distance movements after tournaments, but the factors that prompt this behaviour are unclear.

The time fish took to leave the release area varied between tournaments, with fish leaving most quickly at NV (80% within the first 2 days), while dispersal at the other three tournaments was a mix of immediate, delayed (3-4 days), and extended (5+ days)

(Kaintz and Bettoli 2010). Similar initial dispersal behaviours have been described for various species of bass, where half of tagged largemouth bass in North Carolina dispersed more than 500 m within 7 days (Brown et al. 2015). Smallmouth bass in Tennessee showed little evidence of stockpiling and most fish left the study area even more quickly, within 5 days of release (Kaintz and Bettoli 2010). In contrast, largemouth bass in

Arizona and Nevada had limited dispersal in the first 7 days after release, with 63% of fish last located less than 500 m from the release site (Wilde and Paulson 2003).

However, as I have already described, walleye exhibited longer movements than bass, so it is unclear whether the 500-m buffer is an appropriate benchmark for assessing walleye dispersal. Correspondingly, I found that departure from the release area was not consistently predicted by any of the variables I measured. Stress was an important variable at two tournaments (SL and NV), but the direction of the relationship was opposite. At SL stressed fish left more slowly, whereas at NV stressed fish left more ! 51 quickly. Of the studies summarized above, only Brown et al. (2015) attempted to determine the effects of stress on movement. They measured blood cortisol concentrations but found no relationship with post-release dispersal. Overall, factors that affect walleye and sauger dispersal from the release area are uncertain and need to be further investigated.

The variables I measured did not consistently predict fish movement in the first

48 hours after release. None of the variables were important for RH or LM, while several variables were important for SK and NV, though stress was the only one in common. The relationship between stress and initial movement at these two tournaments was opposite, where stressed fish at SL moved farther, while stressed fish at NV moved less. Initial movement is an important behaviour to understand in relation to stress as the first few days after release are when most acute effects from catch-and-release manifest, as well as when a large portion of delayed morality occurs (Patalas 1996; Schramm et al. 2010). A study on the behaviour of caught-and-released bonefish (Albula spp.), which fight to exhaustion during angling, found that in the hours after release, fish stopped moving often and required extended periods of rest. In contrast, bonefish given a period of recovery swam more consistently (Cooke and Philipp 2004). Species was also an important predictor of movement in 48 hours at SL, where the relationship was extremely strong. There was almost an eight-fold change in this relationship, where walleye were more likely to move farther than sauger. The strength of this relationship was likely related to a small sample size, where only one sauger, compared to 7 walleye were used in the model. I also saw less movement and higher mean stress scores for sauger at SL, which lends support to this relationship. Comparative short-term movements by walleye ! 52 and sauger have not been previously reported. However, sauger are capable of moving large distances in other contexts; for example, sauger in a river in Kentucky moved more than 200 km in less than 10 days (Pegg et al. 1997). Overall, the factors that affect initial dispersal for both species remain uncertain.

The measured predictor variables only partially predicted the rate of movement of fish in the full seven day tracking period. The rates of movement at LM and NV were not significantly different and stress was an important predictor variable at both tournaments, in addition to several other variables at NV. However, the relationship between stress and movement rate at both tournaments was weak, as well as in opposite directions (positive at LM, negative at NV) and therefore not consistent. The range of rates of movement were similar at both tournaments, with some fish moving slowly over the course of the week and some fish moving quickly (3,000+ m / day). Of the studies that have examined the dispersal of fish after tournaments, none have assessed rate of movement as a response variable. Most studies use the total distance moved as a metric, though rate of movement is directly related (Hunter and Maceina 2008; Kaintz and Bettoli 2010). Wilde

(2003) assessed the data from 12 studies on the dispersal of tournament-caught largemouth and smallmouth bass, and examined short-term (14 to 40 days) dispersal as an important behaviour after release. The studies they summarized found this time period to be important; however, I only collected data for 7 days. Increasing the time period of tracking would say more about the overall response of walleye and sauger to these tournaments. Additionally, a different metric of movement, such as displacement distance, may be more useful in describing post-release behaviour. ! 53

Of the various post-release behaviours described, none were consistently predicted by the variables I expected to be important. There are likely additional variables that are important and need to be assessed, such as water surface temperature (Patalas

1996) and air exposure (Gingerich et al. 2007). Importantly, I expected stress to be the most useful predictor of post-release behaviour, and while this variable did appear in multiple relationships, it was not consistent. My data suggest that indices I recorded may not be reliable indicators of fish stress, and vice-versa. Previous work employed cortisol concentrations, but these were also unrelated to dispersal (Brown et al. 2015). The lack of a clear relationship between fish behaviour and stress suggests the various initial behaviours I observed may be adaptive to recovering from stress, ranging from a “fight or flight” response to limited movement to facilitate recovery (O’Connor et al. 2010). The link between movement and stress appears to be more complex than initially assumed, where stress does not simply impair movement, and is therefore not a straightforward predictor.

It is also possible that my measure of stress was unable to resolve differences in the levels of stress actually present. I expected stress to affect fish movement and dispersal time, where time to leave the release area would increase, and movement in the first 48 hours would be reduced in response to stress and exhaustion. However, the only tournament where stress score was an important variable was NV, where the mean stress score was significantly lower and the range of stress scores was much narrower than at other tournaments. My modeling showed stress at NV had a negative relationship with all the behaviours I defined. The strongest relationship was with movement in the first 48 hours, where increasing stress resulted in a decrease in the distance a fish would move, as ! 54 predicted if stress impairs movement. I also observed the greatest overall movement at this tournament, both in terms of how quickly fish left the release area and how far they moved in 48 hours. In contrast, data from tournaments where my assessment indicated fish were more stressed but where fish moved less did not result in a significant relationship between stress and behaviour. These findings suggest that the RAMP test and swim score measures I used were not able to accurately resolve differences among fish that were past a certain threshold of stress. The understanding of the relationship between fish stress and post-release behaviour is limited, but stress has been linked to some of the fish treatments I observed in this study, such as extensive handling and air exposure (Cooke et al. 2002; Gingerich et al. 2007; Arlinghaus et al. 2009; McArley and

Herbert 2014) and high rates of delayed mortality (Davis 2010). Work on signs of barotrauma, which can be physical signs of stress, has shown fish with reduced indicators of barotrauma exit the release area much more quickly than those with severe barotrauma

(Gravel and Cooke 2008). In terms of how I quantified stress and post-release behaviour, it seems that the relationship between high stress and fish movement may not be as straightforward as I predicted.

The current version of the RAMP test may not be suitable for studying fish movement after catch-and-release and other sub-lethal effects. The variation I recorded in

RAMP scores was limited, likely because the reflexes were assessed as present or absent rather than as a spectrum. Different levels of fish stress may be indicated by more subtle differences and this assessment was too coarse to detect differences among similarly stressed fish. Additionally, it may be difficult to identify differences between highly stressed fish. While the RAMP test was originally designed to predict mortality rather ! 55 than behaviour (Davis 2010), it has been used to successfully assess post-stress condition of fish, where reflex impairment was linked to duration of chasing and air exposure, as well as measures of physiological stress (muscle pH, plasma lactate; McArley and

Herbert 2014). The RAMP test has also been found to predict higher rates of migration failure in coho salmon (Oncorhychus kisutch; Raby et al. 2012). Therefore, the RAMP test has been successfully linked to stress and impaired fish movement. However, as the test was unable to separate fish with apparently similar levels of high stress in the current study, it may be useful to build on the original system of stress scoring I used or introduce an alternative. For example, Maynard et al. (2013) used a variety of physical indicators for stress, most of which were related to injuries or indications of barotrauma, to evaluate stress for largemouth and smallmouth bass at tournaments. They found elevated stress levels were correlated to lake temperature, fish size, and livewell transport distance. A test that combines measurements of barotrauma, behaviour, and reflexes may better capture the wide possible spectrum of stress than the test I used which did not appear to be sensitive enough.

Delayed mortality of tagged fish was rare except at SL, where it was high (45%).

I avoided tagging fish that appeared obviously moribund, and stress scores for the tagged fish were not excessively high within the available range. Of the fish that were confirmed to have died, total RAMP scores were moderate (3-4, maximum possible was 6).

Previous work has documented a similar lack of correlation between delayed mortality and increased reflex impairment (McArley and Herbet 2014). Furthermore, my results conflict with the original purpose of the RAMP test to predict morality (Davis 2010). I also recorded a range of swimming scores for the delayed mortalities. Five of the nine ! 56 mortalities had a swimming score of 0, which was the best possible score and meant the fish exhibited normal swimming behaviour. Additionally, the highest score for a fish that died was 3, a moderate value indicating the fish passively sank. There was a fish at this tournament that generated a swimming score of 4, displaying partial equilibrium loss, and was still alive at the end of tracking. Of the five fish that died at SL during tagging and were removed from the study, two had swimming scores of 5 (the worst score, where fish displayed complete equilibrium loss) but the other three had low scores (0-1). Previous studies that have used a swim assessment on rainbow trout (Oncorhynchus mykiss) have indicated stress has little to no effect on swimming performance (Gregory and Wood

1999). Furthermore, swimming away to escape is considered the predominant behavioural reaction to stress in fish and is therefore likely only impaired by extreme physiological stress (Peake et al. 1997). Thus, neither the RAMP nor swimming scores reliably predicted walleye immediate or delayed mortality in my application; only a few of the scores suggested that the tagged fish were at risk for dying.

The relationship between mortality rates of tagged fish at SL and those for the tournament as a whole is currently unclear. I selected fish non-randomly for tagging, so I cannot assume that the delayed mortality rate I observed is directly representative.

However, the high mortality rate I observed in apparently healthy fish suggests that the overall delayed mortality rate for the tournament in general was even higher. SL is the only tournament I studied that occurred in summer, and it continues to take place because of a grandfather clause excluding pre-existing tournaments from regulations preventing competitive fishing events in the months of July and August in Saskatchewan (Section

11.1, Fisheries Regulations 2015). Given the dates of SL, it is not surprising that it had ! 57 the highest surface water temperatures (mean 19.1°C). High water temperatures have been shown to significantly affect mortality rates of fish at tournaments (Patalas 1996;

Gingerich et al. 2007). Furthermore, modeling for mortality at this tournament showed fish size to be the only important predictor variable, where smaller fish were more likely to die. However, the mean size of fish that died was not significantly different from the overall mean size for either species. Hoffman et al. (1996) recommended increasing size limits as a method for reducing mortality at tournaments, while other research has indicated larger fish are more likely to become exhausted to the point of death (Maynard et al. 2013). Overall, while I cannot describe the relationship between the delayed mortality I observed and overall tournament mortality, my data does tentatively support previous work on the effects of water temperature on mortality.

American white pelicans depredated six of the fish that were confirmed mortalities, a phenomenon which has not been documented before. These birds do not dive for fish and are generally restricted to foraging for prey in the upper 1.25 m of the water column (Anderson 1991). It is therefore likely that rather than true predation, pelicans were scavenging for moribund fish that had floated to the surface, or had died previously. While uncommon in freshwater systems like the ones I visited, mortality resulting from predation is significant following catch-and-release angling in saltwater environments. Opportunistic predation by large fish, such as sharks, has been documented for Atlantic sailfish (Istiophorus platypterus, Jolley and Irby 1979), tarpon

(Megalops atlanticus, Edwards 1998), and bonefish (Albula spp., Cooke and Philipp

2004). My study is the first to observe similar opportunistic depredation of sport fish in freshwater environments after catch-and-release and tournaments. Furthermore, my ! 58 observations of moribund, floating fish, both during the SL tournament and afterwards while tracking, suggest that the largest factor affecting their condition was barotrauma.

This is a condition that is not specific to tournament angling, and rather occurs after bringing a fish up from deep waters (Gravel and Cooke 2008). Barotrauma is a common problem for catch-and-release angling, though is it is generally worse during summer months when fish seek out deeper cooler water. Relief methods for this common condition may offer a potential intervention in future years.

Identifying fish mortality based on lack of movement proved to be unreliable. The radio tags I used contained accelerometers, but some fish would remain in one place long enough to trigger the Inactive code and then would begin moving again. I had similar experiences with fish using acoustic tags, which did not have motion sensors. For example, one fish at the NV event moved approximately 2 km and then remained in the same place for the rest of the tracking period. I often would note fish that stopped moving like this as suspected mortalities. However, an angler caught this fish 14 km away from its last known location in healthy condition 12 days after tracking had concluded. Thus, it is clear that inactivity is not necessarily a sign of mortality, and there appears to be period of recovery some fish require after release. This is in contrast to previous methods where fish were considered dead after remaining in the same location for several consecutive fixes, though in most cases the last fix was recorded more than a month after release and was therefore more certain (Kaintz and Bettoli 2010). Identifying mortality via telemetry appears to be difficult because fish may be behaving in an unexpected way and mortality cannot be confirmed.

! 59

3. CHAPTER 3

Sympatric walleye (Sander vitreus) and sauger (S. canadensis) in large reservoirs: variable resource use and overlap across multiple time scales

3.1. INTRODUCTION

Walleye (Sander vitreus) and sauger (S. canadensis) are closely related species that are sympatric over much of their range in North America, but tend to occupy different habitats. In general, walleye are more lacustrine compared to the more riverine sauger (Scott and Crossman 1973; Bozek et al. 2011a). The two species also occupy different parts of the water column; walleye feed in shallower, pelagic areas, compared to the deeper, more demersal sauger (Swenson and Smith 1976; Haxton 2015; Sheppard et al. 2015). However, walleye and sauger are both top predators that require similar dietary resources and will eat the same prey, especially when the species coexist and resources are limited (Bellgraph et al. 2008). Therefore, niche overlap is possible, whereby niche reflects the full range of habitat and dietary resources used (Swenson and Smith 1976;

Hutchinson 1991; Sheppard et al. 2015). Dietary overlap between these fish varies by aquatic system and is more likely to be similar in riverine habitats than in large lakes

(Fincel et al. 2015). There is also evidence that interactions between walleye and sauger increase in areas where anthropogenic changes to habitat have occurred, such as the construction of dams and reservoirs (Rawson and Scholl 1978; Gangl et al. 2000;

Bellgraph et al. 2008). However, the circumstances leading to variation in resource use and niche overlap are poorly understood. ! 60

Levels of resource use overlap by walleye and sauger potentially vary at different ages. Both species spawn at similar times (April-June) and use gravel substrates in the shallows of rivers and lakes for egg deposition (Nelson and Walburg 1977; Bozek et al.

2011a; Bozek et al. 2011b). Sympatric populations often migrate to and use the same shoals of gravel to spawn (Bellgraph et al. 2008). In such situations, substantial resource use overlap could theoretically begin early in life, and persist across several stages of development. Larval walleye and sauger, for example, feed on similar resources, such as zooplankton, as they become free swimming and move away from spawning grounds

(Chipps and Graeb 2011). However, despite apparent similarity in early ages, there has been no measurement of resource use overlap during these periods. Furthermore, as walleye and sauger grow, they undergo similar ontogenetic shifts to diets consisting of benthic invertebrates and then fish (Nelson 1968; Fox and Flowers 1990; Kolar et al.

2003; Hoxmeier et al. 2004). It is likely during the change to a piscivorous diet that major habitat use differences occur, although this has also not been investigated. To date, the resource use overlap by adult, sympatric walleye and sauger has only been described using stomach content analysis (Swenson 1977; Bellgraph et al. 2008; Sheppard et al.

2015), which limits temporal perspectives on resource use. Walleye tend to be larger than sauger and appear to be more flexible in their habitat use (Mittelbach and Persson 1998;

Johnston et al. 2012). These traits may allow walleye to exploit a wider range of prey, and thereby exhibit a larger niche than the more constrained sauger. However, we currently lack any perspective on how the dietary niche of these species compare across ages. ! 61

Stable isotope analysis (SIA) offers quantitative insight into the resource use of populations of coexisting species. SIA provides the opportunity to describe and quantify trophic interactions between sympatric species, as well as food web structure (Leibold

1995; Jackson and Britton 2013; Staudinger et al. 2014; Eberts et al. 2016). Carbon and nitrogen isotopes are most commonly used to describe ecological niches, as they are passed from food source to consumer with characteristic trophic enrichment (France

1995; Perga and Gerdeaux 2003). The isotopic niche can then be quantified in two- dimensional space using δ13C and δ15N values (e.g., Abbey-Lee et al. 2013; Berkström et al. 2013; Hayden et al. 2014a). Isotopic niches effectively represent both the identity and diversity of resource use (Newsome et al. 2007). Furthermore, Bayesian metrics using standard ellipses are especially well suited for statistically comparing resource use among individuals and populations (Jackson et al. 2011). To date, SIA of walleye is limited

(Overman and Parrish 2001; Fincel et al. 2012), and only basic comparisons of walleye and sauger trophic similarity have been undertaken (description of similarity in δ13C and

δ15N values, no niche metrics; Fincel et al. 2015). Consequently, there is a need to more fully examine and describe the niche dynamics of these species when they co-occur.

Analysis of tissues with different turnover rates provides a multi-temporal scale perspective on resource use (Dalerum and Angerbjörn 2005). Soft tissues like liver (days to weeks) and muscle (months) can be used to compare recent resource use of individuals or populations (Tieszen et al. 1983; Perga and Gerdeaux 2003). Tissues with longer turnover times can represent resource use over a time frame of years (Krueger and

Sullivan 1984; Hobson and Sease 1998; Clementz 2012). Many hard tissues grow incrementally such that layers are deposited during periods of fast and slow growth; these ! 62 layers may be used to assess resource use during particular time periods (Hobson and

Sease 1998; Campana and Thorrold 2001; Overman and Parrish 2001; Morbey et al.

2007). In fish, scales and otoliths are most commonly used to quantify long-term resource use (Wainwright et al. 1993; Perga and Gerdeaux 2003; Hutchinson and Trueman 2006;

Morbey et al. 2007), though opercular bones have also recently been analyzed (Syväranta et al. 2008; Eberts et al. 2016). The use of multiple tissues to compare resource use of walleye and sauger at several temporal scales has not been undertaken.

SIA provides the opportunity to quantify niches of individuals within a population. The overall niche of a population comprises biotic and abiotic interactions, while individuals within the population may be heterogeneous in their use of varying resources and conditions (Bolnick et al. 2010). It is therefore important to not only quantify the overall niche of a population, but also how individuals contribute to that niche (Araújo and Gonzaga 2007). Multi-tissue analysis is useful for such an approach, where several measurements from the same individual over different time scales can give a comprehensive perspective on niche (e.g., Knoff et al. 2008; Fincel et al. 2012; Heady and Moore 2013; Vander Zanden et al. 2013; Eberts et al. 2016). To fully understand how the niches of walleye and sauger compare when the species co-occur, quantifying individual niches within a population is necessary (Baccante et al. 2011).

My purpose was to compare resource use of sympatric populations of walleye and sauger in central Canada using carbon and nitrogen stable isotopes. I focused on two large and economically important reservoirs in Saskatchewan, where walleye and sauger co-occur in environments that were extensively altered by the construction of dams

(Royer 1972; Wallace et al. 2010). My overall objectives were to compare resource use ! 63 between walleye and sauger at multiple ages, as well as to determine how individuals contribute to population niches. I hypothesized that niche size would vary with species, and predicted that walleye would have different and larger niches than the more constrained sauger as adults. I also predicted that the niches of both species would be similar in size and composition at earlier ages.

3.2. MATERIALS & METHODS

3.2.1. Study sites & fish collection.− I worked with the organizers of catch-and-release tournaments to capitalize on extensive angler efforts to sample walleye and sauger at my study sites. I collected tissues from fish for SIA exclusively from mortalities that resulted from tournament activities. As such, my sampling was inherently biased towards fish susceptible to angling and occupying habitats targeted by tournament contestants.

However, tournaments process large numbers of fish (1,200 – 1,600 in a weekend) of all sizes from within extensive boundaries; thus, they are likely representative of the lake area sampled. I targeted three major lakes in Saskatchewan during 2014 and 2015 (Table

3.1; Fig. 3.1). Fish were sampled at the Saskatchewan Landing Walleye Tournament held on Lake Diefenbaker in July (2014, 2015), the Nipawin Vanity Cup held on Tobin Lake in October (2014, 2015), and the Saskatchewan Premier’s Walleye Cup on Tobin Lake in

August (2014). Both tournaments on Tobin Lake were held in the same location with identical boundaries. Additionally, walleye were sampled from the Last Mountain Fall

Classic on Last Mountain Lake in September (2015) to provide niche data for this species in the absence of sauger and in a natural lake. Study sites are hereafter referred to by lake, not tournament (Lake Diefenbaker = DF; Tobin Lake = TB; Last Mountain Lake = LM). ! 64

Table 3.1. Summary information for the four tournaments that provided fish tissue samples for this study. The maximum number of teams that could register for the event

(Max. teams) is listed, as well as the area of the fishing zones within tournament boundaries (km2). This area constitutes the section of the lake sampled.

Tournament Name of lake Month / Year(s) Max. teams Area (km2) Saskatchewan Lake Diefenbaker July / 2014, 2015 125 68 Landing (DF) Walleye Tournament

Nipawin Vanity Tobin Lake (TB) October / 2014, 160 25 Cup 2015

Saskatchewan Tobin Lake August / 2014 160 25 Premiers Walleye Cup

Last Mountain Last Mountain September / 2015 120 40 Fall Classic Lake (LM)

! 65

TBTB

Saskatoon DFDF LM

Regina

150 km

DF DF

TB LM TB

Figure 3.1. Locations of the three lakes in Saskatchewan, Canada where walleye and sauger were sampled from angling tournaments. Abbreviations of lakes are: DF = Lake

Diefenbaker; LM = Last Mountain Lake; TB = Tobin Lake. Arrows indicate the area of the lake where the tournaments were held. Inset maps provide more detailed images of the lakes and boxed sections depict the area sampled by anglers. ! 66

Lake Diefenbaker (51°01’N 106°50’W) is a large reservoir (max. length = 225 km; max. width = 6 km; surface area = 430 km2) created by two dams on the South

Saskatchewan and Qu’Appelle Rivers; it supports 26 fish species (State of Lake

Diefenbaker, 2012). The lake has a mean depth of 21.6 m (maximum 66 m); the area where the Saskatchewan Landing tournament occurs is the most riverine portion of the lake. Walleye and sauger occur throughout the lake, which is known to produce hybrids between the two species (saugeye). Tobin Lake (53°35’N 103°30’W) is a large reservoir with a surface area of 228 km2 and a mean depth of 8 m (maximum 26.4 m; Durbin

1981). The reservoir is composed of two unique basins: the more southern is riverine while the northern is larger and more lacustrine. Both Tobin Lake tournaments were held on the riverine portion. Tobin Lake supports 25 species of fish and is home to one of the most valuable walleye sport fisheries in the province. Similar to Lake Diefenbaker, both walleye and sauger occur in Tobin Lake and hybrids have been reported. Last Mountain

Lake (51°06’N 105°15’W) is southern Saskatchewan’s largest natural waterbody. This long, narrow prairie lake (max. length = 93 km; max. width = 4.5 km; surface area = 233 km2) supports a variety of fish species, including walleye but not sauger, and has a mean depth of 7.6 m (maximum 31.5 m). I chose to include walleye from this location to provide some perspective on the niche of walleye in the absence of sauger in a lake that is part of the same river system as both reservoirs.

3.2.2. Multi-tissue processing & analyses.− Total length and mass were measured for all fish, and dorsal muscle and liver samples were collected for SIA. Additionally, both opercular bones were taken for SIA and to estimate minimum age (Le Cren 1947). The ! 67 muscle and liver samples were rinsed with distilled water, dried in a drying oven at 50°C for 72 hours, ground to a fine powder using a dental amalgamator, and weighed (0.5-1.0 mg) into tin capsules for mass spectrometry analysis.

I chose opercula from both species of fish for SIA that were at least 5 years old and showed clear and distinct annuli. To compare resource use during several distinct time periods, I segregated desired age sections for analysis following the methods of

Eberts et al. (2016). I etched three age segments into each operculum along the annuli using a scalpel as follows: (1) year 1, all bone from the origin to the first annulus; (2) years 2-4, all bone from the first to the fourth annulus; and (3) years 5+, all bone from the fourth annulus onward (Fig. 3.2). Once etched, the opercula were soaked in distilled water overnight (12-14 hours) until soft. Each age segment was cut out using dissecting scissors and then cut into small pieces. Bone samples were decalcified with 1.2N HCl

(Perga and Gerdeaux 2003) to remove inorganic carbon, rinsed twice with distilled water, and dried in a drying oven at 50°C for 72 hours. Finally, the years 2-4 and 5+ bone segments were ground to a fine powder using a dental amalgamator and samples were weighed (0.5-1.0 mg) into tin capsules. Year 1 bone segments were too small to be ground, so I analyzed a small piece of the sample from closest to the origin, which represented the earliest possible section of growth. This section was weighed whole (0.3-

1.0 mg) into tin capsules; all bone samples were analyzed using mass spectrometry.

All samples were analyzed for carbon (C) and nitrogen (N) stable isotopes using the facilities of the Institute of Environmental Change and Society (IECS) at the

University of Regina with a Thermo Finnigan Delta V isotope ratio mass spectrometer. ! 68

Figure 3.2. Example of walleye operculum showing an illustration of the three age segments used for stable isotopes analysis. These sections are: (A) years 0-1; (B) years 2-

4; and (C) years 5+. In this example, the fish was determined to be 14 years old, so section (C) includes years 5-14. All material outside the dotted line, which is lateral bone growth and the origin, was not included in the analysis.

! 69

Laboratory standards of wheat and bovine liver, unknowns, and replicates of standards and unknowns were all run in each batch of samples. Replicates of unknowns and standards differed by less than 0.2‰. Isotope values are reported as delta values

13 15 13 (δ CVPDB and δ NAIR) in units of per mil (‰) (Bond and Hobson 2012). δ C values for muscle and liver samples were arithmetically corrected for lipids when C/N ratios were

>3.5 following the methods of McConnaughey and McRoy (1979).

3.2.3. Statistical analyses.− Isotopic niches were quantified using δ13C and δ15N values with standard ellipses, which contain the core 40% of data points (Jackson et al. 2011).

Standard ellipse area (SEA) was compared between species at each sampling location for each of the five tissue samples (liver, muscle, and three bone sections). The five tissue values were then used to produce individual standard ellipses for each fish. All SEA were corrected for small sample size (SEAC) (Jackson et al. 2011). I used SEAC to calculate the degree of isotopic niche overlap for all ellipses (population and individual level), which represented a quantitative measure of dietary similarity. One-way Analysis of

Variance (ANOVA) was used to test for differences among niche metrics (p ≤ 0.05). If a significant main effect was obtained, a post-hoc Tukey test was used to identify specific pair-wise differences.

I compared resource use diversity among populations by statistically comparing niche sizes using Bayesian ellipses in R (SIBER; Jackson et al. 2011). All population level standard ellipses for walleye and sauger had a SEA calculated with 10,000 replications. This produced a Bayesian SEA (SEAB) paired with niche size uncertainty.

This measurement allowed us to test the probability that one standard ellipse (and thus ! 70 isotopic niche) was larger than another (significant probability: P ≥ 0.95; Jackson et al.

2011). All statistical analyses were conducted in R Project for Statistical Computing 3.3.0

(R Development Core Team 2016), using the Stable Isotope Analysis in R (SIAR) package (version 4.2; Parnell and Jackson 2015).

3.3. RESULTS

I collected muscle and liver tissues from a total of 156 walleye from three lakes and 75 sauger from two lakes in Saskatchewan (Table 3.2). Total length of walleye varied among lakes (Table 3.2; ANOVA, F(2,153)=12.11, p<0.001); post-hoc testing revealed that fish were significantly larger at LM than at DF and TB. Sauger were significantly larger at DF than TB (Table 3.2; ANOVA, F(1,73)=26.46, p<0.001). A subset of 99 fish was used for operculum analysis and the creation of individual isotopic niches (19 walleye, 20 sauger from DF; 20 walleye, 20 sauger from TB; 20 walleye from LM).

Mean age of walleye used for bone SIA from all three locations were not significantly different (TB: 9.5±3.6 years, max=16 years; DF: 8.7±1.3 years, max=12 years; LM:

9.5±2.9 years, max=17 years respectively; ANOVA, F(2,56)=0.47, p=0.63). Sauger from

DF were significantly older than those from TB (DF: 9.1±1.4 years, max=12 years; TB:

6.9±1.7 years, max=12 years; ANOVA, F(1,38)=19.24, p<0.001). This corresponds to their larger mean size.

3.3.1. Niche overlap.− For all tissues, walleye from DF had lower mean δ 13C values and higher δ 15N values compared with sauger (Table 3.3; Fig. 3.3a-e). In contrast, walleye from TB had lower δ 15N values than sauger (Table 3.3; Fig. 3.4a-e). δ 13C values for the ! 71

Table 3.2. Walleye and sauger sampled for liver, muscle, and opercular bones from three lakes in Saskatchewan (DF = Lake Diefenbaker; TB = Tobin Lake; LM = Last Mountain

Lake) during 2014 and 2015. Mean values for total length, mass, and age are given ±SD.

Ages were determined using both opercula; if the ages were different, the lower age was used.

Total Length (mm) Mass (g) Age (years) DF Walleye 2014 20 471.0±86.6 1027.5±673.7 7.2±2.1 2015 31 466.5±47.9 875.8±301.2 6.9±1.7 Sauger 2014 18 447.5±42.0 780.6±258.9 8.0±1.8 2015 28 464.5±39.0 771.4±240.6 8.7±1.7 TB Walleye 2014 35 421.0±90.8 840.0±910.0 5.6±2.4 2015 31 453.9±127.6 1209.7±1300.3 6.6±3.7 Sauger 2014 5 413.0±51.9 744.0±310.8 5.8±0.8 2015 24 403.1±47.8 655.8±289.0 6.1±2.1 LM Walleye 2015 39 536.8±122.3 1980.8±1328.2 7.2±3.2

! 72 two species at TB were similar (Table 3.3; Fig. 3.4a-e). Comparisons between species for all three locations indicate δ 13C values for liver and muscle were always lower compared to bone samples (Table 3.3).

Walleye and sauger niche overlap at DF was highest for liver and muscle, each at

19%. In comparison, niche overlap was much lower for the three sections of operculum, with 3% overlap for years 0-1, 10% for years 2-4, and 8% for years 5+. At TB, niche overlap between walleye and sauger was more variable but generally higher. The lowest overlap was for muscle (9%), while the highest was for liver (31%). Niche overlap for the three bone segments was intermediate (years 0-1: 18%, years 2-4: 18%, years 5+: 13%). I also quantified overlap between the total niches created by all tissue values for both species. The overall overlap of walleye and sauger niches at DF was 15% (Fig. 3.6), contrasting with the much higher overlap of 32% at TB (Fig. 3.7).

3.3.2. Niche size.− Isotopic niche breadth analysis (SEAC) of fish from DF indicated that niche sizes based on analysis of liver and muscle were similar for both species (liver:

P=0.93, muscle: P=0.78; SEAC, Table 3.3; Fig. 3.3a, b). In contrast, isotopic niches for sauger were larger than walleye for all three bone segments with a significant probability

(P=1.00; 0-1: 4.2x larger, 2-4: 5.3x, 5+: 3.5x; SEAC, Table 3.3; Fig. 3.3c-e). Relative niche size between walleye and sauger varied with tissue in TB. SEAC estimates for walleye were significantly larger than sauger for muscle but not significantly different for liver (muscle: P=0.99; liver: P=0.69; SEAC, Table 3.3; Fig. 3.4a, b). Conversely, differences in SEAC between walleye and sauger were not significant for bone segments

0-1 and 2-4, while walleye had a significantly larger isotopic niche compared to sauger ! 73

for the 5+ bone segment (0-1: P=0.86, 2-4: P=0.58, 5+: P=0.99; SEAC, Table 3.3; Fig.

3.4c-e). Only walleye were sampled from LM and the niche size of all sample types was similar, except for the year 0-1 sample, which was 2.8x larger than the next largest SEAC

(SEAC, Table 3.3; Fig. 3.5a-e).

I created one total standard ellipse for both species using all tissue values. The total niche size of sauger was significantly larger than walleye at DF (P=0.98; Pop SEAC,

Table 3.4; Fig. 3.6). In contrast, total ellipses at TB revealed walleye had a larger niche than sauger (P=1.00; Pop SEAC, Table 3.4; Fig. 3.7). The total niche size of walleye at

LM was 2x larger than walleye at DF and 1.5x larger than walleye at TB (Pop SEAc,

Table 3.4; Fig. 3.8).

3.3.3. Individual niches.− I quantified individual lifetime niches using all five tissue samples for 59 walleye from three lakes and 40 sauger from two lakes. At DF, individual niche size was similar for both species (Table 3.4; ANOVA, F(1,37)=0.51, p=0.47).

Intra-specific niche overlap among individual walleye was significantly higher (14.5% ±

10.9%) than that for sauger (11.3% ± 10.5%), as well as inter-specific overlap among individuals of the two species (9.7% ± 10.4%; Fig. 3.6; ANOVA, F(2,738)=12.39, p<0.001). In contrast, niche size for individual walleye and sauger at TB differed; walleye had significantly larger niches (Table 3.4; ANOVA, F(1,38)=6.47, p=0.015). In general, individual niche overlap was also higher at this lake, regardless of comparison.

Intra-specific overlap for walleye (15.3% ± 10.7%) was significantly lower than that for sauger (18.7% ± 10.9%). Inter-specific overlap among individuals of both species (15.9%

± 11.6%) was similar to that of walleye, and significantly lower than for sauger (Fig. 3.7; ! 74

Table 3.3. Stable isotopes data summary for walleye and sauger sampled from three lakes in Saskatchewan (DF = Lake Diefenbaker; TB = Tobin Lake; LM = Last Mountain

Lake) during 2014 and 2015. Isotopic values are presented as mean ± SD in per mil (‰) with associated C/N ratios. Standard Ellipse Area (SEAC) describes the isotopic niche size (‰2); δ13C values are lipid corrected.

13 15 2 δ C (‰) δ N (‰) C/N SEAC (‰ ) DF Walleye Liver −28.6±0.8 +18.3±1.1 7.1±1.6 2.61 Muscle −27.6±0.9 +18.8±0.8 3.8±0.2 2.08 Years 0-1 −25.7±0.5 +18.3±0.5 3.4±0.2 0.72 Years 2-4 −25.4±0.2 +18.6±0.6 3.4±0.1 0.51 Years 5+ −25.5±0.3 +18.6±0.6 3.3±0.1 0.56 Sauger Liver −27.6±1.3 +17.7±0.9 6.6±1.6 3.60 Muscle −27.0±0.7 +18.1±0.8 3.8±0.1 1.73 Years 0-1 −25.1±1.0 +16.9±1.1 3.5±0.2 3.00 Years 2-4 −24.6±1.1 +17.3±1.5 3.3±0.1 2.70 Years 5+ −24.9±0.7 +17.6±1.0 3.3±0.1 1.96 TB Walleye Liver −27.3±0.8 +16.7±1.1 7.9±2.5 2.94 Muscle −26.4±0.5 +17.3±0.8 3.8±0.1 1.30 Years 0-1 −24.4±0.8 +16.2±0.9 3.4±0.1 2.30 Years 2-4 −24.1±0.7 +16.4±0.8 3.3±0.1 1.82 Years 5+ −24.1±0.7 +17.1±0.9 3.4±0.1 1.95 Sauger Liver −27.5±1.0 +17.2±0.9 9.3±4.0 2.57 Muscle −26.4±0.3 +18.2±0.6 3.9±0.1 0.54 Years 0-1 −24.9±0.7 +16.7±0.6 3.5±0.3 1.49 Years 2-4 −24.3±0.8 +17.2±0.7 3.4±0.1 1.68 Years 5+ −24.6±0.4 +17.6±0.6 3.3±0.1 0.68 LM Walleye Liver −24.8±0.7 +14.7±0.6 10.0±4.4 1.49 Muscle −23.8±0.5 +16.1±0.6 3.8±0.1 0.84 Years 0-1 −21.0±1.3 +14.1±1.3 3.4±0.1 4.73 Years 2-4 −20.6±0.9 +14.2±0.7 3.4±0.1 1.66 Years 5+ −21.0±1.2 +15.5±0.5 3.4±0.1 1.65

! 75

Table 3.4. Niche metrics for individual walleye and sauger in three Saskatchewan lakes

(DF = Lake Diefenbaker; TB = Tobin Lake; LM = Last Mountain Lake). Mean individual niche size (Ind. mean SEAC) was calculated using SEAC values from individual standard ellipses for 59 walleye and 40 sauger. The population value (Pop. SEAC) represents the niche size of the population using all five sampled tissues. The mean percentage of the total population niche used by individual walleye and sauger is indicated.

Ind. mean SEAC Pop. SEAC Mean % of total (± SD) DF Walleye 2.3±1.0 3.11 23.1±7.5 Sauger 2.6±1.4 4.16 19.8±11.7 TB Walleye 5.1±2.6 5.50 24.7±8.2 Sauger 3.4±1.5 3.76 26.9±9.4 LM Walleye 8.0±3.9 6.95 30.9±11.1

! 76

a 20 N 18 15 δ 16 14 −32 −30 −28 −26 −24 −22

20 b δ13C N 18 15 δ 16 14 −32c −30 −28 −26 −24 −22 20 13C

) δ ‰ N 18 15 δ N ( 16 15 δ 14 −32 −30 −28 −26 −24 −22

20 d δ13C N 18 15 δ 16 14 −32 −30 −28 −26 −24 −22

13 20 e δ C N 18 15 δ 16 14 −32 −30 −28 −26 −24 −22 δ13Cδ13 C(‰)

! 77

Figure 3.3. Walleye (triangles; black ellipses) and sauger (circles; grey ellipses) δ13C and

δ15N values in units per mil (‰) for: (a) liver, (b) muscle, (c) years 0-1 bone segment, (d) years 2-4 bone segment, and (e) years 5+ bone segment for fish sampled from Lake

Diefenbaker (Walleye n=51, Sauger n=46). Values for bone segments represent the organic fraction only. Standard ellipses are plotted using the data points for walleye and sauger to describe isotopic niche. Convex hulls (dashed) outline all data points for each species. Ranges on both axes are the same. ! 78

a 19 N 17 15 δ 15 13 −30 −28 −26 −24 −22

19 b 13 δ C N 17 15 δ 15 13 c −30 −28 −26 −24 −22 ) 19 δ13C ‰ N 17 15 δ N ( 15 15 δ 13 −30 −28 −26 −24 −22

19 d 13 δ C N 17 15 δ 15 13 e −30 −28 −26 −24 −22 19 δ13C N 17 15 δ 15 13 −30 −28 −26 −24 −22 δ13δC13C (‰)

! 79

Figure 3.4. Walleye (triangles; black ellipses) and sauger (circles; grey ellipses) δ13C and

δ15N values in units per mil (‰) for: (a) liver, (b) muscle, (c) years 0-1 bone segment, (d) years 2-4 bone segment, and (e) years 5+ bone segment for fish sampled from Tobin

Lake (Walleye n=66, Sauger n=29). Values for bone segments represent the organic fraction only. Standard ellipses are plotted using the data points for walleye and sauger to describe isotopic niche. Convex hulls (dashed) outline all data points for each species.

Ranges on both axes are the same.

! 80

a 18 N 16 15 δ 14 12 −26 −24 −22 −20 −18

18 b δ13C N 16 15 δ 14 12 c −26 −24 −22 −20 −18 ) 18 δ13C ‰ N 16 15 δ N ( 14 15 δ 12 −26 −24 −22 −20 −18

13 18 d δ C N 16 15 δ 14 12 e −26 −24 −22 −20 −18 18 δ13C N 16 15 δ 14 12 −26 −24 −22 −20 −18 δ13δC13C (‰)

! 81

Figure 3.5. Walleye δ13C and δ15N values in units per mil (‰) for: (a) liver, (b) muscle,

(c) years 0-1 bone segment, (d) years 2-4 bone segment, and (e) years 5+ bone segment for fish sampled from Last Mountain Lake (n=39). Values for bone segments represent the organic fraction only. Standard ellipses are plotted using the data points for walleye to describe isotopic niche. Convex hulls (dashed) outline all data points. Ranges on both axes are the same.

! 82 20 ) 18 ‰ N 15 δ N ( 16 15 δ 14

−30 −28 −26 −24 −22

13 13 δ Cδ (C‰)

Figure 3.6. All tissue values (muscle, liver, and three bone segments) combined to quantify individual isotopic niches for walleye (thin black ellipses; n=19) and sauger

(thin red ellipses; n=20) at Lake Diefenbaker. Total population isotopic niches are indicated for walleye (thick black ellipse) and sauger (thick red ellipse). Convex hulls

(dashed) outline all data points for total population ellipses only.

! 83 20 ) 18 ‰ N 15 δ N ( 15 16 δ 14 −30 −28 −26 −24 −22

13 13 δ Cδ (C‰)

Figure 3.7. All tissue values (muscle, liver, and three bone segments) combined to quantify individual isotopic niches for walleye (thin black ellipses; n=20) and sauger

(thin red ellipses; n=20) at Tobin Lake. Total population isotopic niches are indicated for walleye (thick black ellipse) and sauger (thick red ellipse). Convex hulls (dashed) outline all data points for total population ellipses only.

! 84 18 ) ‰ N 16 15 δ N ( 15 14 δ 12 −26 −24 −22 −20 −18

13 13 δ Cδ (‰C )

Figure 3.8. All tissue values (muscle, liver, and three bone segments) combined to quantify individual isotopic niches for walleye (thin black ellipses; n=20) at Last

Mountain Lake. Total population isotopic niche is indicated for walleye (thick black ellipse). Convex hull (dashed) outlines all data points for total population ellipse only.

! 85

ANOVA, F(2,777)=6.39, p=0.0018). Mean niche size of individual walleye at LM was

3.5x larger than walleye at DF and 1.5x larger than walleye at TB (Ind Mean SEAC,

Table 3.4). Intra-specific overlap among individual walleye was also higher at 24.2% ±

12.5% (Fig. 3.8).

Finally, I calculated the average proportion of the total population niche that individual fish used at each lake (Table 3.4). Individual walleye at DF used significantly less of the total population niche than walleye at LM, while walleye at TB were not significantly different from either location (ANOVA, F(2,56)=4.10, p=0.0217).

Individual sauger at DF used significantly less of the total population niche than sauger at

TB (ANOVA, F(1,38)=4.53, p=0.0398). Both species at DF and TB used similar proportions of their population niche (DF: ANOVA, F(1,37)=1.05, p=0.3117; TB:

ANOVA, F(1,38)=0.68, p=0.4150).

3.4. DISCUSSION

3.4.1. Niche overlap.− Walleye and sauger exhibited consistent niche partitioning at all time scales based on our analysis from both reservoirs. Significant niche overlap between species based on SIBER metrics (SEAC) has been defined as greater than 60%, which indicates potential for diet competition (Guzzo et al. 2013). All niche overlap values quantified in the current study, regardless of tissue type or lake, were much lower than

60%. Thus, walleye and sauger were spatially and ecologically segregated at all ages

(Haxton 2015). I predicted that niche overlap would be minimal among adult fish of both species (based on muscle, liver, and years 5+ bone), as they likely co-exist via niche partitioning during this time period (Ali and Anctil 1977; Keough et al. 1996; Bellgraph ! 86 et al. 2008; Bozek et al. 2011a). Though the fish I sampled were caught in similar areas of the lake, they likely partition water depths and target prey species found at those depths (Haxton 2015; Sheppard et al. 2015). In contrast, I expected niche overlap to be much more extensive at the earlier ages analyzed (years 0-1 bone layers), when fish are similar in size and more likely to occupy similar habitats and consume similar prey

(Bozek et al. 2011b). The understanding of walleye and sauger feeding ecology during their early life history is limited (Chipps and Graeb 2011), and currently I can only offer several potential explanations for the unexpected lack of overlap based on analysis of the

0-1 bone layer. Walleye and sauger may actually begin consistent ecological segregation at a very young age, or their major ontogenetic dietary shift occurs early in life and the 0-

1 layer of the operculum represents different patterns in the two species. It is also possible that the operculum is not biologically inactive and the entire structure continues to incorporate more recent isotopic information. Unfortunately, I am unable to distinguish among these possible explanations.

Levels of niche overlap between walleye and sauger differed substantially in the two study systems. I expected to observe similar levels of niche overlap at both locations, as high dietary overlap has been described in riverine reservoir environments previously

(Fitz and Holbrook 1978; Mero 1992; Bellgraph et al. 2008). However, my data showed that overlap of the total niches of walleye and sauger was more than double at TB (32%), compared to DF (15%). Diet overlap between these species has been shown to vary with environmental conditions and habitat structure, as well as baseline and prey fish composition, which may explain the differences I observed (Fincel et al. 2015; Sheppard et al. 2015). The general habitat and size of the reservoirs I sampled also differed. TB is a ! 87 smaller lake and offers less habitat heterogeneity, especially in the area where anglers sampled; this would suggest a lower possibility for niche partitioning. The species composition in the two reservoirs is very similar, with mostly suckers and shiners as small prey fish and the same species of larger, predatory fish (Durbin 1981; State of Lake

Diefenbaker 2012). However, the relative abundance of these species in the reservoirs is unknown. Ultimately, overlap between walleye and sauger differed with reservoir, but the precise differences in the reservoirs that caused varying niche overlap need to be investigated to facilitate broader generalization.

Based on isotope values, walleye were feeding at higher trophic positions compared to sauger at DF and lower trophic positions at TB. Previous research has shown that trophic position, both within and between populations in different lakes, can be variable (Trippel and Beamish 1993; Vander Zanden and Rasmussen 1996).

Variability across lakes can largely be explained by differences in the lakes, while variability within lakes reflects the extent of differences in individual foraging (Vander

Zanden et al. 2000). Differences in mean δ 15N values within locations, such as I describe, may also imply differences in diet composition (Fincel et al. 2015). I also observed an interesting pattern where the δ13C values of liver and muscle were always lower (more depleted) than those of the three bone segments, regardless of location. Previous work has shown similar patterns, where soft tissues like muscle were more δ13C-depleted than hard tissues, like scales (Fincel et al. 2012; Heady and Moore 2013). These differences are likely reflective of differences in tissue specific discrimination (Schmidt et al. 2004).

The values also reflect specific periods of time; in this study, soft tissues represent a short time window during summer feeding, while hard tissues represent the average resource ! 88 use of fish as they make movements to other habitats throughout the year (Guzzo et al.

2013).

3.4.2. Niche size.− In contrast to my prediction that sauger would have consistently smaller niches than walleye, relative niche size for the two species varied by location. My data showed that in DF, the niche size of sauger was significantly larger than that of walleye for bone samples, as well as for the combined analysis of all tissue types. In contrast, niche size in TB was similar for the two species for three of the tissue types, but walleye had significantly larger niches based on muscle, the year 5+ bone segment, and the combined analysis of all tissues. My prediction about relative niche size was based on sauger being more constrained in their use of specific habitats and prey (e.g., Bozek et al.

2011a; Johnston et al. 2012). Based on my SIA, sauger may be more flexible than previously considered, and in this case consistently use more diverse resources than walleye over significant periods of time in DF. Other niche size comparisons for these two species have not previously been undertaken, but general diet studies have indicated that walleye do not always exhibit the high flexibility in resource use that is expected

(Hobson et al. 2012; Herbst et al. 2016). Sauger have been shown to exhibit high levels of variation in their diet in certain environments, including reservoirs (Fincel et al. 2015).

Furthermore, sauger have differences in the structure of their retina, compared with walleye, which may allow them to forage more efficiently under reduced light conditions and enable them to exploit a wider variety of prey (Ali and Anctil 1977; Swenson 1977).

Overall, I found that niche size can vary with location and sauger can exhibit larger ! 89 niches than walleye in some situations. However, it is unclear what factors contributed to this variable niche size.

I also predicted the niche size of both species would be most similar during earlier ages, when the fish are comparable in size (Bozek et al. 2011b). As the fish aged, I expected niche size to differ more, reflecting the tendency for ecological segregation as the two species undergo ontogenetic shifts. My data showed that in DF, sauger had a significantly larger niche size than walleye for the bone section representing the earliest stage of life, as well as for the two bone sections representing the later ages of the fish.

Additionally, niche sizes in this reservoir for muscle and liver, which characterized different levels of short-term resource use, were similar. In contrast, niche sizes in TB were similar for the first two time periods (years 0-1 and 2-4), but walleye exhibited larger niches for the years 5+ and muscle samples. Thus, my findings were more consistent with my predictions at one reservoir (TB) than the other (DF). While niche size of these species has not been compared previously at the early ages I examined, there are potential differences in spawning that may explain the unexpected findings at DF.

The populations I sampled may not spawn in the same location, so the resource use of fry and juvenile fish would be naturally different based on available prey and baseline at each location (Ross 1986; De la Morinière et al. 2003; Kolar et al. 2003). Alternatively, if the populations are spawning in the same area, sauger tend to spawn later in the spring than walleye, when there may be a higher abundance of available resources (Scott and

Crossman 1973; Nelson and Walburg 1977; Bethea et al. 2004). Finally, there is the possibility that the sections of bone sampled here are not biologically inactive, and may actually reflect resource use during later time periods rather than those expected. ! 90

3.4.3. Individual niches.− Niche breadth and overlap of individual walleye and sauger varied among locations. The mean niche breadth of both species at DF were similar, while walleye at TB had significantly larger mean niche sizes than sauger. It is interesting that while individual fish at TB followed patterns in niche breadth exhibited by the total population, individual walleye and sauger at DF used similar sized niches, even though the overall niche of sauger was larger than that of walleye at this lake. This suggests that the sauger population in DF was made up of more individuals behaving uniquely in terms of resource use (Beaudoin et al. 1999; Bolnick et al. 2002; Anderson et al. 2009; Bolnick et al. 2010). Niche overlap was also highly variable among locations. My data showed that at DF, individual walleye overlapped most with other walleye, as expected. In comparison, sauger overlapped as much with walleye as they did with other sauger. In contrast, sauger at TB overlapped most with each other, while walleye were more likely to overlap with sauger than they were with other walleye. At LM, inter-specific walleye overlap was the highest of all locations, suggesting a possible difference in resource use by walleye that do not co-occur with sauger. The variation I saw in niche size and overlap among individuals is likely a reflection of differences in diet composition between study lakes (Sheppard et al. 2015). For example, walleye are largely piscivorous but will supplement their diet with invertebrates based on availability (Chipps and Graeb 2011). It is possible some individuals within these populations were supplementing their diet with other prey sources. Ultimately, I observed individuals within populations acting differently among the study lakes, but the precise reasons for why this occurred require further study. ! 91

The extent individual walleye and sauger contributed to the population niche also varied by reservoir. Walleye at DF contributed significantly less to the population niche than walleye at LM, while walleye contributions at TB were not different from the other lakes. Sauger at DF contributed significantly less to the population niche than sauger at

TB. Therefore at DF, I saw the highest degree of fish using different resources compared to the total population. Both walleye and sauger are considered opportunistic feeders, meaning they eat what is available rather than searching out a specific prey source

(Chipps and Graeb 2011). My data reflect this notion; individual fish may only partially overlap with the population niche because they feed opportunistically and therefore are not always using similar resources to the population as a whole (Grey 2001). Fish may also be moving to, and feeding in, locations with different baseline isotopic values (Barks et al. 2010; Bugajski et al. 2013). The differences I saw by location suggest that at DF, where individual contributions by both species were lowest, fish have a wider diversity of prey to feed on. In contrast, I found higher contributions to the population niche at TB, where perhaps there is a lower diversity of prey or habitats.

3.4.4. Differences among study sites.− The reservoirs studied here are connected through the greater Saskatchewan River system, and superficially are similar in terms of their structure and general characteristics. However, location-specific features may account for some of the differences observed in walleye and sauger niche. Lake Diefenbaker is approximately two times larger than Tobin Lake, and is also three times deeper on average. This difference in size and depth may provide more opportunity for both species to select different (and possibly more preferred) habitats in Lake Diefenbaker (e.g., ! 92

Seehausen and Bouton 1997), and may also explain why sauger had larger niches and greater individual variability compared to walleye in this reservoir. Conversely, the smaller and shallower Tobin Lake may provide a lower variety of habitats, causing overall niche overlap to be higher than in Lake Diefenbaker. Differences in overlap may also be explained by differences in the density of prey species, which can change with lake size and depth (Holmgren and Appelberg 2000; Radke and Eckmann 2001; Haertel et al. 2002; Olin et al. 2002; Mehner et al. 2005). Finally, isotopic baseline can be influenced by source waters (Hobson et al. 2012), as well as lake size and terrestrial input

(Post 2002). If the isotopic baselines of these lakes are significantly different and spatially variable, differences in niche size and overlap of walleye and sauger would reflect this.

3.4.5. Other points.− I took advantage of tournament angling to collect fish, and the anglers at these tournaments do not catch fish randomly. Rather, they focus on known areas of walleye habitat where they are most likely to catch large fish. They also do not target sauger when fishing, so any fish of this species that they keep would most likely be caught in areas where walleye and sauger co-occur. This is helpful for the questions I was attempting to answer in the current study, but also means the sauger I sampled were potentially more similar to walleye in resource use than sauger from other areas of the study lakes. The anglers in these tournaments also only fished a specific portion of the lake, rather than the whole thing. However, alternative methods of sampling fish in previous SIA research include seine netting and fish traps (De la Morinière et al. 2003), gill netting (Fincel et al. 2012; Eberts et al. 2016; Fincel et al. 2015), electrofishing ! 93

(Jackson and Britton 2013), and angling (Grey 2001). These methods also target specific lake areas and often do not cover the entire study location. In addition, the fish I sampled at tournaments were mortalities and there is evidence that fish do not die randomly as a result of angling activities. Rather, many factors can influence whether a fish will die; numerous aspects of tournaments can cause varying levels of stress for fish that may lead to death, including air exposure, livewell conditions, and water temperature (Gingerich et al. 2007; Havn et al. 2015; Sullivan et al. 2015). Fish caught at certain depths may also experience barotrauma, which can be difficult to recover from (Gravel and Cooke 2008).

Finally, fish size and condition can influence the likelihood of mortality (Maynard et al.

2013). However, I took subsets of the fish that died at these tournaments and attempted to sample a wide range of fish sizes, making my sampling as random as possible. Overall, sampling fish for SIA in this manner is likely comparable to other methods in terms of how representative the fish I sampled are of general relationships.

An important assumption I made when choosing to use different segments of bone in this research was that each section became metabolically inert after its growth was complete. Consequently, each bone layer should permanently reflect the isotopic values of prey consumed during the deposition period. Based on this assumption, I expected to see a distinct isotopic shift as the fish went through ontogenetic shifts from zooplankton to piscivory (Keough et al. 1996). However, SIA of my three bone segments did not show this resource shift; δ13C and δ15N values for both species at all locations did not change across bone layers. Thus, I cannot confirm that the bone layers I analyzed in walleye and sauger actually reflect resource use during the specified ages. This potential problem has been previously highlighted in the use of scales for SIA (e.g., Wainright et al. 1993; ! 94

Hutchinson and Trueman 2006). However, Morbey et al. (2007) separated scales into sections to represent the early and late life periods in lake trout (Salvelinus namaycush) and successfully demonstrated a significant shift in δ13C values between scale centers and edges, consistent with an ontogenetic diet shift. The only work done to date on several sections of opercula did not show any shift in resource use between the juvenile and adult bone sections in two species of whitefish (Coregonus spp.), but it is unclear whether such a shift could be detected in the mixed populations assessed (Eberts et al. 2016). In the absence of other data, my results suggest that walleye and sauger opercula may not be suitable for the study of specific age layers. However, these species also experience a shift to piscivory at a very early stage of life (Chipps and Graeb 2011). There is a possibility that my sampling of the year 0-1 bone segment was simply not specific enough to capture this shift.

! 95

4. CHAPTER 4: CONCLUSIONS AND FUTURE DIRECTIONS

In many areas, including Saskatchewan, recreational and commercial angling for walleye is economically important, creating a significant need to better understand many aspects of the ecology of this species. Additionally, the closely related sauger often co- exists with walleye, including in Saskatchewan lakes that are popular angling destinations, where the two species are caught and regulated as the same entity. Sauger are much less studied than walleye in general, but how the two species interact and relate, and how they are similar or differ, is poorly understood. Therefore, the major objective of this work was to address several important questions regarding walleye and sauger, with a focus on populations in Saskatchewan.

The results from Chapter 1, which assessed the movement of both species after catch-and-release tournaments, revealed high variation in fish movement and also showed that walleye have the capability of moving great distances in the short time period I evaluated. In contrast, my results indicate that sauger move much less in that time. As such, I generally observed two classes of fish behaviour after release (as described in O’Conner et al. 2010). The first was a “fight or flight” response, where fish immediately left the release area and continued to move very quickly and over significant distances for at least the first few days, if not the entire tracking period. The second response was little or no movement for varying periods of time after release followed by general movement again, suggesting the need for a period of recovery. While the understanding of normal walleye movement in the systems I studied is limited, neither of the two types of movement I saw are generally reflective of normal movements by these fish. Both walleye and sauger make mostly small movements during the non-spawning ! 96 season, when their priority is to find food (Bellgraph et al. 2008). Therefore, it appears that the large movements by walleye I saw are possibly abnormal. Furthermore, previous work has shown that swimming only becomes impaired in fish that have experienced extreme physiological stress (Peake et al. 1997). I saw high numbers of fish with limited initial movement, which suggests they experienced high enough levels of stress to necessitate staying in one place for an extended period of time. I especially observed this limited movement in sauger, where all tagged fish of this species only moved 2 km or less in the week of tracking. This suggests that sauger are more susceptible to stress and require more recovery time than walleye.

These movement patterns are of particular interest in terms of management because they detail how fish are affected by tournament activities – they either exhibit abnormally large movements or they require a recovery period. Both altered behaviours likely require energy, either to swim far or recover, that would otherwise be used for feeding or reproduction (Cooke et al. 2002; Wilde 2003; Hunter and Maceina 2008). As both tournament organizers and managers aspire to maintain healthy populations of walleye and sauger, the post-release behaviours I observed are important as they likely impact these populations. Interventions need to be put in place to help “normalize” movement, which requires further work to understand what normal behaviour is. There is also a need for a better understanding of how walleye and sauger are different, to explain why they exhibit different responses to tournaments, as observed here. Walleye are considered the hardier of the two species, as they are much larger and exploit more habitats, which may mean they respond to stress better (Johnston et al. 2012). However, this indicates the two species need to be treated differently at tournaments to ! 97 acknowledge their unique tolerances. For example, if sauger are more sensitive to air exposure or handling time, methods could be put in place to limit these treatments.

Alternatively, if sauger respond poorly to high water temperatures, tournaments during summer months might sanction that sauger cannot be weighed-in.

My results also showed that the variables I measured related to fish treatment were not necessarily suitable predictors of fish movement. However, there are many other treatment variables that I did not include in this study that are likely important and need to be investigated in the context of post-release movement. Those variables include degree of air exposure at several time windows, including during capture and weigh-in, and a more comprehensive assessment of livewell conditions, including temperature and the number of fish in the livewell (Gingerich et al. 2007; Sullivan et al. 2015). I was most surprised that the stress variable, which I expected to be important, was not a consistent predictor of behaviour. This variable was only important at one tournament (NV), where

I saw the lowest stress scores and also the most movement. This suggests that this factor is likely still important for predicting fish movement, but the test I used did not resolve high levels of stress adequately. A test including signs of barotrauma, as well as behaviour and reflexes, would likely be more indicative and should be developed. Stress can also be quantified using certain measures of blood chemistry, like plasma cortisol, which could be incorporated into the test (Pottinger 1998). My current understanding of how fish stress affects post-release behaviour appears to be quite limited but improving this understanding would be extremely helpful in implementing management strategies at tournaments to prevent stress. ! 98

In Chapter 2, I used stable isotope analysis to quantify the niches of walleye and sauger in two reservoirs. My findings indicate that the species co-exist via niche partitioning at all the time scales I measured. I also saw variation among locations in regards to resource use and niche overlap. Thus, my comparison of sympatric populations in two reservoirs showed the niche relationship between these species can vary with location. I also noted that walleye living in a system without sauger appear to use resources differently than those co-existing with sauger. Further work should attempt to explain this variation, which is likely due to differences in the reservoirs I sampled.

Describing the prey fish communities in each lake, as well as the overall food web and baseline, could be further used to explain why these species appear to interact differently in different places (i.e., Keough et al. 1996; Hobson et al. 2012; Abrantes et al. 2014).

For example, knowing the baseline of each lake would help confirm whether the varying niche overlap and resource use I saw actually indicated different uses of resources or if they were simply a reflection of differing baselines. Assessing the composition of prey fish communities in these lakes could also inform which species are present, where they exist in the lake, and what their relative abundance is, all of which indicates what and where the walleye and sauger in these lakes are eating. This information is generally helpful when managing the species, as it shows what prey species are important and whether there is any competition potential.

While my work is the first to use stable isotopes to compare the resource use of walleye and sauger when they co-exist, populations like this occur throughout North

America. It is therefore likely that differences in resource use and niche overlap also occur in other areas based on location. As the differences I saw were most likely a ! 99 reflection of baseline, prey source, and environmental structure of the reservoirs I sampled, understanding how these differences affected the isotope values of my sampled fish would aid in establishing a broader general expectation of walleye-sauger relationships in similar lakes (Baccante et al. 2011). Understanding the relationship between lake characteristics and the population dynamics of walleye and sauger informs management techniques, indicating whether certain lakes need to be uniquely managed based on their potential for high resource overlap and competition (Fincel et al. 2015).

Overall, I sought to investigate several ecological questions regarding walleye and sauger and describe the similarities and differences between these closely related species.

I found, in both movement response to catch-and-release and resource use, that the species are different from each other. As has been summarized, walleye and sauger differ in their life history and this advocates that harvest regulations related to size and limit need to be species specific if they are to be effective (Haxton 2015). The differences I noted further support this statement. Differences in the response to catch-and-release tournaments and the stress they cause between walleye and sauger suggests tournament organizers and government authorities need to treat these species separately when they implement strategies to minimize sub-lethal effects and mortality. Furthermore, in aquatic systems where these species exhibit differences in resource use and consistent niche partitioning, the species should be managed separately as they are using the resources and habitat of the lake or river uniquely. If the species continue to be treated as the same, sauger will likely continue to be more impacted by tournaments and standard catch-and- release than walleye, while the resources and habitat of one species will be better managed, providing a competitive edge for that species. ! 100

5. LITERATURE CITED

Abbey‐Lee, R.N., Gaiser, E.E., and Trexler, J.C. 2013. Relative roles of dispersal dynamics and competition in determining the isotopic niche breadth of a wetland fish.

Freshwater Biol. 58(4): 780-792. doi:10.1111/fwb.12084.

Abrantes, K.G., Barnett, A., and Bouillon, S. 2014. Stable isotope-based community metrics as a tool to identify patterns in food web structure in east African estuaries.

Funct. Ecol. 28(1): 270-282. doi:10.1111/1365-2435.12155.

Ahrenstorff, T.D., Sass, G.G., and Helmus, M.R. 2009. The influence of littoral zone coarse woody habitat on home range size, spatial distribution, and feeding ecology of largemouth bass (Micropterus salmoides). Hydrobiologia 623(1): 223-233. doi:10.1007/s10750-008-9660-1.

Ali, M.A. and Anctil, M. 1977. Retinal structure in the walleye (Stizostedion vitreum vitreum) and sauger (Stizostedion canadense). J. Fish. Res. Board Can. 34(10): 1467-

1474. doi:10.1139/f77-211.

Anderson, J.G. 1991. Foraging behaviour of the American white pelican (Pelecanus erythrorhyncos) in western Nevada. Colon. Waterbird. 14(2): 166-172. doi:10.2307/1521506.

! 101

Anderson, O.R.J., Phillips, R.A., Shore, R.F., McGill, R.A.R., McDonald, R.A., and

Bearhop, S. 2009. Diet, individual specialization and breeding of brown skuas

(Catharacta antarctica lonnbergi): An investigation using stable isotopes. Polar Biol. 32:

27-33. doi:10.1007/s00300-008-0498-9.

Araújo, M.S. and Gonzaga, M.O. 2007. Individual specialization in the hunting-wasp

Trypoxylon (Trypargilum) albonigrum (Hymenoptera: Cabronidae). Behav. Ecol.

Sociobiol. 61(12): 1855-1863. doi:10.1007/s00265-007-0425-z.

Arlinghaus, R., Cooke, S.J., Lyman, J., Policansky, D., Schwab, A., Suski, C., Sutton,

S.G., and Thorstad, E.B. 2007. Understanding the complexity of catch-and-release in recreational fishing: an integrative synthesis of global knowledge from historical, ethical, social, and biological perspectives. Rev. Fish. Sci. 15(1-2): 75-167. doi:10.1080/10641260601149432.

Arlinghaus, R., Klefoth, T., Cooke, S.J., Gingerich, A., and Suski, C. 2009. Physiological and behavioural consequences of catch-and-release angling on northern pike (Esox lucius

L.). Fish. Res. 97(3): 223-233. doi:10.1016/j.fishres.2009.02.005.

Ashcroft, P., Duffy, M., Dunn, C., Johnston, T., Koob, M., Merkowsky, J., Murphy, K.,

Scott, K., and Senik, B. 2006. The Saskatchewan fishery: History and current status.

Department of Environment, Saskatchewan, Canada.

! 102

Baccante, D.A., Barton, B.A., Bozek, M.A., and Bruner, J.C. 2011. Future research needs of walleye and sauger. Fisheries 36(12): 618-619. doi:10.1080/03632415.2011.640231.

Barks, P.M., Doucette, J.L., and Somers, C.M. 2010. Lack of angling-sized yellow perch in a Canadian boreal lake: potential influences of growth rate, diet, and predation by double-crested cormorants. Trans. Am. Fish. Soc. 139(4): 1029-1040. doi:10.1577/T09-

174.1.

Barton, K. 2013. MuMIn: multi-model inference. R Package version 1.9.0.

Barton, B.A. and Barry, T.P. 2011. Reproduction and environmental biology. In Biology, management, and culture of walleye and sauger. Edited by B. A. Barton. American

Fisheries Society, Bethesda, Maryland. pp. 199-231.

Bates, D., Maechler, M., and Bolker, B. 2012. lme4: linear mixed effects models using

S4 classes. R Package version 0.999999-0.

Beaudoin, C.P., Tonn, W.M., Prepas, E.E., and Wassenaar, L.I. 1999. Individual specialization and trophic adaptability of northern pike (Esox lucius): An isotope and dietary analysis. Oecologia 120(3): 386-396. doi:10.1007/s004420050871.

! 103

Bellgraph, B.J., Guy, C.S., Gardner, W.M., and Leathe, S.A. 2008. Competition potential between saugers and walleyes in nonnative sympatry. T. Am. Fish. Soc. 137(3): 790-800. doi:10.1577/T07-102.1.

Berkström, C., Jörgensen, T.L., and Hellström, M. 2013. Ecological connectivity and niche differentiation between two closely related fish species in the mangrove-seagrass- coral reef continuum. Mar. Ecol. Prog. Ser. 477: 201-215. doi:10.3354/meps10171.

Bethea, D.M., Buckel, J.A., and Carlson, J.K. 2004. Foraging ecology of the early life stages of four sympatric shark species. Mar. Ecol. Prog. Ser. 268: 245-264. doi:10.3354/meps268245.

Billington, N., Hebert, P.D.N., and Ward, R.D. 1990. Allozyme and mitochondrial DNA variation among three species of Stizostedion (Percidae): phylogenetic and zoogeographical implications. Can. J. Fish. Aquat. Sci. 47(6): 1093-1102. doi:10.1139/f90-126.

Billington, N., Koigi, R.N., Sloss, B.L., Franckowiak, R.P., and Xiong, J. 2006. Genetic variation and hybridization with walleye in Montana sauger populations determined by protein electrophoresis and microsatellite analysis. Report of the Department of

Biological and Environmental Sciences, Troy University, and Molecular Conservation

Genetics Laboratory, University of Wisconsin–Stevens Point, to Montana Department of

Fish, Wildlife and Parks, Helena.

! 104

Billington, N. and Sloss, B.L. 2011. Molecular systematics of Sander, and hybridization between walleye and sauger. In Biology, management, and culture of walleye and sauger.

Edited by B. A. Barton. American Fisheries Society, Bethesda, Maryland. pp. 85-104.

Bingham, D.M., Leary, R.F., Painter, S., and Allendorf, F.W. 2012. Near absence of hybridization between sauger and introduced walleye despite massive releases. Conserv.

Genet. 13(2): 509-523. doi:10.1007/s10592-011-0303-0.

Bolnick, D.I., Yang, L.H., Fordyce, J.A., Davis, J.M., and Svanbäck, R. 2002. Measuring individual-level resource specialization. Ecology 83(10): 2936-2941. doi:10.1890/0012-

9658(2002)083[2936:MILRS]2.0.CO;2.

Bolnick, D.I., Ingram, T., Stutz, W.E., Snowberg, L.K., Lau, O.L., and Paull, J.S. 2010.

Ecological release from interspecific competition leads to decoupled changes in population and individual niche width. P. Roy. Soc. Lond. B Bio. 277(1689): 1789-1797. doi:10.1098/rspb.2010.0018.

Bond, A.L. and Hobson, K.A. 2012. Reporting stable-isotope ratios in ecology: recommended terminology, guidelines and best practices. Waterbirds 35(2): 324-331. doi:10.1675/063.035.0213.

Bozek, M.A., Haxton, T.J., and Raabe, J.K. 2011a. Walleye and sauger habitat. In

Biology, management, and culture of walleye and sauger. Edited by B.A. Barton. ! 105

American Fisheries Society, Bethesda, Maryland. pp. 133-178.

Bozek, M.A., Baccante, D.A., and Lester, N.P. 2011b. Walleye and sauger life history. In

Biology, management, and culture of walleye and sauger. Edited by B.A. Barton.

American Fisheries Society, Bethesda, Maryland. pp. 233-301.

Brown, D.T., Rice, J.A., Suski, C.D., and Derek Aday, D. 2015. Dispersal patterns of coastal largemouth bass in response to tournament displacement. N. Am. J. Fish. Manage.

35(3): 431-439. doi:10.1080/02755947.2015.1009660.

Brownscombe, J.W., Thiem, J.D., Hatry, C., Cull, F., Haak, C.R., Danylchuk, A.J., and

Cooke, S.J. 2013. Recovery bags reduce post-release impairments in locomotory activity and behaviour of bonefish (Albula spp.) following exposure to angling-related stressors.

J. Exp. Mar. Biol. Ecol. 440: 207-215. doi:10.1016/j.jembe.2012.12.004.

Bugajski, A., Reudink, M.W., Doucette, J.L., Franks, S.E., Wissel, B., and Somers, C.M.

2013. The complexity of cormorants: stable isotopes reveal multiple prey sources and feeding site switching. Can. J. Fish. Aquat. Sci. 70(2): 271-279. doi:10.1139/cjfas-2012-

0237.

Bunt, C.M., Cooke, S.J., and Philipp, D.P. 2002. Mobility of riverine smallmouth bass related to tournament displacement and seasonal movements. Am. Fish. S. S. 31.

! 106

Bureau of statistics. 2015. Government of Saskatchewan, Saskatchewan, Canada.

Accessed July 2016. .

Campana, S. E. and Thorrold, S. R. 2001. Otoliths, increments, and elements: keys to a comprehensive understanding of fish populations?. Can. J. Fish. Aquat. Sci. 58(1): 30-38. doi:10.1139/f00-177.

Carlin, B. 1955. Tagging of salmon smolts in the River Lagan. Rep. Inst. Freshwat. Res.

Drott. 36: 57-74.

Chipps, S.R and Graeb, B.D.S. 2011. Feeding ecology and energetics. In Biology, management, and culture of walleye and sauger. Edited by B.A. Barton. American

Fisheries Society, Bethesda, Maryland. pp. 303-319.

Clementz, M.T. 2012. New insight from old bones: stable isotope analysis of fossil mammals. J. Mammal. 93(2): 368-380. doi:10.1644/11-MAMM-S-179.1c.

Cooke, S.J. and Philipp, D.P. 2004. Behaviour and mortality of caught-and-released bonefish (Albula spp.) in Bahamian waters with implications for a sustainable recreational fishery. Biol. Conserv. 118(5): 599-607. doi:10.1016/j.biocon.2003.10.009.

Cooke, S.J., Schreer, J.F., Wahl, D.H., and Philipp, D.P. 2002. Physiological impacts of catch-and-release angling practices on largemouth bass and smallmouth bass. In ! 107

American Fisheries Society Symposium 30. Edited by J. A. Lucy and A. L. Studholme.

American Fisheries Society. Bethesda, Maryland. pp. 489-512.

Crowder, L.B. 1990. Community ecology. In Methods for fish biology. Edited by C. B.

Schreck and P. B. Moyle. American Fisheries Society, Bethesda, Maryland. pp 609-632.

Dalerum, F. and Angerbjörn, A. 2005. Resolving temporal variation in vertebrate diets using naturally occurring stable isotopes. Oecologia 144(4): 647-658. doi:10.1007/s00442-005-0118-0.

Davis, M.W. 2010. Fish stress and mortality can be predicted using reflex impairment.

Fish. Fish. 11(1): 1-11. doi:10.1111/j.1467-2979.2009.00331.x.

De la Morinière, E.C., Pollux, B.J.A., Nagelkerken, I., Hemminga, M.A., Huiskes,

A.H.L., and Van der Velde, G. 2003. Ontogenetic dietary changes of coral reef fishes in the mangrove-seagrass-reef continuum: stable isotopes and gut-content analysis. Mar.

Ecol. Prog. Ser. 246: 279-289. doi:10.3354/meps246279.

Dunne, J.A., Williams, R.J., and Martinez, N.D. 2002. Networks topology and biodiversity loss in food webs: Robustness increases with connectance. Santa Fe Institute

Working Paper 02-03-013.

! 108

Durbin, J.K. 1981. Tobin Lake fishery 1979 – 17 years after impoundment. Fisheries

Technical Report 81-5. Fisheries Research Program, Saskatchewan Tourism and

Renewable Resources, Saskatoon, Saskatchewan, Canada.

Eberts, R.L., Wissel, B., Manzon, R.G., Wilson, J.Y., Boreham, D.R., and Somers, C.M.

2016. Consistent differential resource use by sympatric lake (Coregonus clupeaformis) and round (Prosopium cylindraceum) whitefish in Lake Huron: a multi-time scale isotopic niche analysis. Can. J. Fish. Aquat. Sci. 73(7): 1072-1080. doi:10.1139/cjfas-

2015-0324.

Edwards, R.E. 1998. Survival and movement patterns of released tarpon (Megalops atlanticus). Gulf Mexico Sci. 16(1): 1.

The Fisheries Regulations. 2015. Available from the Government of Saskatchewan,

Saskatchewan, Canada.

Flammang, M.K. and Willis, D.W. 1993. Comparison of electropherograms and external characteristics for distinguishing juvenile walleyes and saugeyes. Prairie Naturalist 25(3):

255-255. doi:008357452.

Fincel, M.J., Vandehey, J.A., and Chipps, S.R. 2012. Non‐lethal sampling of walleye for stable isotope analysis: a comparison of three tissues. Fisheries Manage. Ecol. 19(4):

283-292. doi:10.1111/j.1365-2400.2011.00830.x. ! 109

Fincel, M.J., Chipps, S.R., Graeb, B.D.S., and Brown, M.L. 2015. Diet breadth and variability in Sander spp. inferred from stable isotopes. River Res. Appl. 32(5): 984-991. doi:10.1002/rra.2916.

Fisheries and Oceans Canada, 2014. Canadian Fisheries Statistics 2004. Department of

Fisheries and Oceans, Ottawa, Canada.

Fiss, F.C., Sammons, S.M., Bettoli, P.W., and Billington, N. 1997. Reproduction among saugeyes (Fx hybrids) and walleyes in Normandy Reservoir, Tennessee. N. Am. J. Fish.

Manage. 17(1): 215-219. doi:10.1577/1548-8675(1997)017<0215:RASFHA>2.3.CO;2.

Fitz, R.B. and Holbrook, J.A. 1978. Sauger and walleye in Norris Reservoir, Tennessee.

In Selected Coolwater Fishes of North America. Edited by R.L. Kendall. American

Fisheries Society, Special Publication 11, Bethesda, Maryland. pp. 82-88.

Fox, M.G. and Flowers, D.D. 1990. Effect of fish density on growth, survival, and food consumption by juvenile walleyes in rearing ponds. T. Am. Fish. Soc. 119(1): 112-121. doi:10.1577/1548-8659(1990)119<0112:EOFDOG>2.3.CO;2.

France, R.L. 1995. Carbon-13 enrichment in benthic compared to planktonic algae: foodweb implications. Mar. Ecol. Prog. Ser. 124: 307-312. doi:10.3354/meps124307.

! 110

Fuller, P. 2010a. Sander vitreus. USGS nonindigenous aquatic species database.

Gainesville, Florida. Accessed August 2016.

.

Fuller, P. 2010b. Sander canadensis. USGS nonindigenous aquatic species database.

Gainesville, Florida. Accessed August 2016.

.

Gangl, R.S., Pereira, D.L., and Walsh, R.J. 2000. Seasonal movements, habitat use, and spawning areas of walleye Stizostedion vitreum and sauger S. canadense in Pool 2 of the upper Mississippi River. Investigational Report. Minnesota Department of Natural

Resources, Minnesota, U.S.A.

Garner, S.R., Bobrowicz, S.M., and Wilson, C.C. 2013. Genetic and ecological assessment of population rehabilitation: walleye in Lake Superior. Ecol. Appl. 23(3):

594-605. doi:10.1890/12-1099.1.

Gilliland, E.R. 1999. Dispersal of black bass following tournament release in an

Oklahoma reservoir. In Proceedings of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 53. pp. 144-149.

Gingerich, A.J., Cooke, S.J., Hanson, K.C., Donaldson, M.R., Hasler, C.T., Suski, C.D. and Arlinghaus, R. 2007. Evaluation of the interactive effects of air exposure duration ! 111 and water temperature on the condition and survival of angled and released fish. Fish.

Res. 86(2): 169-178. doi:10.1016/j.fishres.2007.06.002.

Graeb, B.D., Willis, D.W., Billington, N., Koigi, R.N., and VanDeHey, J.A. 2010. Age- structured assessment of walleyes, saugers, and naturally produced hybrids in three

Missouri River reservoirs. N. Am. J. Fish. Manage. 30(4): 887-897. doi:10.1577/M09-

091.1.

Gravel, M.A. and Cooke, S.J. 2008. Severity of barotrauma influences the physiological status, postrelease behaviour, and fate of tournament-caught smallmouth bass. N. Am. J.

Fish. Manage. 28(2): 607-617. doi:10.1577/M07-013.1.

Gregory, T.R. and Wood, C.M. 1999. The effects of chronic plasma cortisol elevation on the feeding behaviour, growth, competitive ability, and swimming performance of juvenile rainbow trout. Physiol. Biochem. Zool. 72(3): 286-295. doi:10.1086/316673.

Grey, J. 2001. Ontogeny and dietary specialization in brown trout (Salmo trutta L.) from

Loch Ness, Scotland, examined using stable isotopes of carbon and nitrogen. Ecol.

Freshw. Fish 10(3): 168-176. doi:10.1034/j.1600-0633.2001.100306.x.

Grueber, C.E., Nakagawa, S., Laws, R.J., and Jamieson, I.G. 2011. Multimodel inference in ecology and evolution: challenges and solutions. J. Evol. Biol. 24(4): 699-711. doi:10.1111/j.1420-9101.2010.02210.x. ! 112

Guzzo, M.M., Haffner, G.D., Legler, N.D., Rush, S.A., and Fisk, A.T. 2013. Fifty years later: trophic ecology and niche overlap of a native and non-indigenous fish species in the western basin of Lake Erie. Biol. Invasions 15(8): 1695-1711. doi:10.1007/s10530-012-

0401-z.

Haertel, S.S., Baade, U., and Eckmann, R. 2002. No general percid dominance at mesotrophic lake conditions; insights from the quantification of predator-prey interactions. Limnologica 32(1): 1-13. doi:10.1016/S0075-9511(02)80011-1.

Hanson, K.C., Cooke, S.J., Suski, C.D., Niezgoda, G., Phelan, F.J.S., Tinline, R., and

Philipp, D.P. 2007. Assessment of largemouth bass (Micropterus salmoides) behaviour and activity at multiple spatial and temporal scales utilizing a whole-lake telemetry array.

Hydrobiologia 582(1): 243-256. doi:10.1007/s10750-006-0549-6.

Havn, T.B., Uglem, I., Solem, Ø., Cooke, S.J., Whoriskey, F.G., and Thorstad, E.B.

2015. The effect of catch‐and‐release angling at high water temperatures on behaviour and survival of Atlantic salmon Salmo salar during spawning migration. J. Fish Biol.

87(2): 342-359. doi:10.1111/jfb.12722.

Haxton, T. 2015. Characteristics and spatial segregation of sympatric saugers and walleyes in the Ottawa River, Canada. N. Am. J. Fish. Manage. 35(3): 418-430. doi:10.1080/02755947.2015.1017117. ! 113

Haxton, T., Nienhuis, S., Punt, K., and Baker, T. 2015. Assessing walleye movement among reaches of a large, fragmented river. N. Am. J. Fish. Manage. 35(3): 537-550. doi:10.1080/02755947.2015.1012278.

Hayden, B., Harrod, C., and Kahilainen, K.K. 2014. Lake morphometry and resource polymorphism determine niche segregation between cool‐and cold‐water‐adapted fish.

Ecology 95(2): 538-552. doi:10.1890/13-0264.1.

Hayden, T.A., Holbrook, C.M., Fielder, D.G., Vandergoot, C.S., Bergstedt, R.A.,

Dettmers, J.M., Krueger, C.C., and Cooke, S.J. 2014. Acoustic telemetry reveals large- scale migration patterns of walleye in Lake Huron. PloS ONE 9(12): e114833. doi:10.1371/journal.pone.0114833.

Herbst, S.J., Roth, B.M., Hayes, D.B., and Stockwell, J.D. 2016. Walleye foraging ecology in an interconnected chain of lakes influenced by nonnative species. T. Am. Fish.

Soc. 145(2): 319-333. doi:10.1080/00028487.2015.1121924.

Heady, W.N. and Moore, J.W. 2013. Tissue turnover and stable isotope clocks to quantify resource shifts in anadromous rainbow trout. Oecologia 172(1): 21-34. doi:10.1007/s00442-012-2483-9.

Hobson, K.A. and Sease, J.L. 1998. Stable isotope analyses of tooth annuli reveal temporal dietary records: an example using stellar sea lions. Mar. Mam. Sci. 14(1): 116- ! 114

129. doi:10.1111/j.1748-7692.1998.tb00694.x.

Hobson, K.A., Ofukany, A., Soto, D.X., and Wassenaar, L.I. 2012. An isotopic baseline

(δ13C, δ15N) for fishes of Lake Winnipeg: implications for investigating impacts of eutrophication and invasive species. J. Great Lakes Res. 38: 58-65. doi:10.1016/j.jglr.2010.11.008.

Hoffman, G.C., Coble, D.W., Frie, R.V., Copes, F.A., Bruch, R.M., and Kamke, K.K.

1996. Walleye and sauger mortality associated with live-release tournaments on the Lake

Winnebago system, Wisconsin. N. Am. J. Fish. Manage. 16(2): 364-370. doi:10.1577/1548-8675(1996)016<0364:WASMAW>2.3.CO;2.

Holmgren, K. and Appelberg, M. 2000. Size structure of benthic freshwater fish communities in relation to environmental gradients. J. Fish Biol. 57(5):

1312-1330. doi:10.1111/j.1095-8649.2000.tb00489.x.

Hoxmeier, R.J.H., Wahl, D.H., Hooe, M.L., and Pierce, C.L. 2004. Growth and survival of larval walleyes in response to prey availability. T. Am. Fish. Soc. 133(1): 45-54. doi:10.1577/T01-082.

Hunter, R.W. and Maceina, M.J. 2008. Dispersal of tournament-displaced largemouth bass and spotted bass in Lake Martin, Alabama. N. Am. J. Fish. Manage. 28(3): 678-687. doi:10.1577/M07-082.1. ! 115

Hutchinson, G.E. 1991. Population studies: Animal ecology and demography. B. Math.

Bio. 53(1-2): 193-213. doi:10.1016/S0092-8240(05)80046-1.

Hutchinson, J.J. and Trueman, C.N. 2006. Stable isotope analyses of collagen in fish scales: limitations set by scale architecture. J. Fish Biol. 69(6): 1874-1880. doi:10.1111/j.1095-8649.2006.01234.x.

Jackson, M.C. and Britton, J.R. 2013. Variation in the trophic overlap of invasive

Pseudorasbora parva and sympatric cyprinid fishes. Ecol. Freshw. Fish 22(4): 654-657. doi:10.1111/eff.12063.

Jackson, A.L., Inger, R., Parnell, A.C., and Bearhop, S. 2011. Comparing isotopic niches widths among and within communities: SIBER – Stable Isotope Bayesian Ellipses in R.

J. Anim. Ecol. 80: 595-602. doi:10.1111/j.1365-2656.2011.01806.x.

Johnston, T.A., Lysak, W., and Leggett, W.C. 2012. Abundance, growth, and life history characteristics of sympatric walleye (Sander vitreus) and sauger (Sander canadensis) in

Lake Winnipeg, Manitoba. J. Great Lakes Res. 38(3): 35-46. doi:10.1016/j.jglr.2010.06.009.

! 116

Jolley Jr, J.W. and Irby Jr, E.W. 1979. Survival of tagged and released Atlantic sailfish

(Istiophorus platypterus: Istiophoridae) determined with acoustical telemetry. B. Mar.

Sci. 29(2): 155-169.

Kaintz, M.A. and Bettoli, P.W. 2010. Dispersal of smallmouth bass from a simulated tournament weigh-in site. N. Am. J. Fish. Manage. 30(4): 976-982. doi:10.1577/M10-

033.1.

Kelso, J.R.M. 1978. Diel rhythm in activity of walleye, Stizostedion vitreum vitreum. J.

Fish Biol. 12(6): 593-599. doi:10.1111/j.1095-8649.1978.tb04206.x.

Keough, J.R., Sierszen, M.E., and Hagley, C.A. 1996. Analysis of a Lake Superior coastal food web with stable isotope techniques. Limnol. Oceanogr. 41(1): 136-146. doi:10.4319/lo.1996.41.1.0136.

Kerr, S.J. and Kamke, K.K. 2003. Competitive fishing in freshwaters of North America: a survey of Canadian and US jurisdictions. Fisheries 28(3): 26-31. doi:10.1577/1548-

8446(2003)28[26:CFIFON]2.0.CO;2.

Killen, S.S., Suski, C.D., Morrissey, M.B., Dyment, P., Furimsky, M., and Tufts, B.L.

2003. Physiological responses of walleyes to live-release angling tournaments. N. Am. J.

Fish. Manage. 23(4): 1238-1246. doi:10.1577/M02-164.

! 117

Killen, S.S., Suski, C.D., Cooke, S.J., Philipp, D.P., and Tufts, B.L. 2006. Factors contributing to the physiological disturbance in walleyes during simulated live-release angling tournaments. T. Am. Fish. Soc. 135(3): 557-569. doi:10.1577/T05-168.1.

Knoff, A., Hohn, A., and Macko, S. 2008. Ontogenetic diet changes in bottlenose dolphins (Tursiops truncatus) reflected through stable isotopes. Mar. Mammal Sci. 24(1):

128-137. doi:10.1111/j.1748-7692.2007.00174.x.

Koenst, W.M. and Smith Jr, L.L. 1976. Thermal requirements of the early life history stages of walleye, Stizostedion vitreum vitreum, and sauger, Stizostedion canadense. J.

Fish. Board Can. 33(5): 1130-1138. doi:10.1139/f76-141.

Kolar, C.S., Wahl, D.H., and Hooe, M.L. 2003. Piscivory in juvenile walleyes: relative importance of prey species, timing of spawning of prey fish, and density on growth and survival. T. Am. Fish. Soc. 132(4): 679-690. doi:10.1577/T99-068.

Krueger, H.W. and Sullivan, C.H. 1984. Models for carbon isotope fractionation between diet and bone. Stable Iso. Nutri. 258: 205-220. doi:10.1021/bk-1984-0258.ch014.

Leary, R.W. and Allendorf, F.W. 1997. Introgression between introduced walleye and native sauger in Fort Peck Reservoir and the Yellowstone River, Montana. Report to

Montana Department of Fish, Wildlife & Parks, Helena.

! 118

Le Cren, E. D. 1947. The determination of the age and growth of the Perch (Perca fluviatilis) from the opercular bone. J. Anim. Ecol. 16(2): 188-204.

Leibold, M.A. 1995. The niche concept revisited: mechanistic models and community context. Ecology 76: 1371-1382. doi:10.2307/1938141.

MacLean, J. and Magnuson, J.J. 1977. Species interactions in percid communities. J.

Fish. Board Can. 34(10): 1941-1951. doi:10.1139/f77-260.

Maynard, G.A., Mihuc, T.B., Schultz, R.E., Sotola, V.A., Reyes, A.J., Malchoff, M.H., and Garneau, D.E. 2013. Use of external indicators to evaluate stress of largemouth

(Micropterus salmoides) and smallmouth (M. dolomieu) bass at tournaments. Open Fish

Sci. J. 6: 78-86. doi:10.2174/1874401X01306010078.

McArley, T.J. and Herbert, N.A. 2014. Mortality, physiological stress and reflex impairment in sub-legal Pagrus auratus exposed to simulated angling. J. Exp. Mar. Biol.

Ecol. 461: 61-72. doi:10.1016/j.jembe.2014.07.016.

McConnaughey, T. and McRoy, C.P. 1979. Food-web structure and the fractionation of carbon isotopes in the Bering Sea. Mar. Biol. 53: 257-262. doi:10.1007/BF00952434.

Mehner, T., Diekmann, M., Brämick, U., and Lemcke, R. 2005. Composition of fish communities in German lakes as related to lake morphology, trophic state, shore structure ! 119 and human!use intensity. Freshwater Biol. 50(1): 70-85. doi:10.1111/j.1365-

2427.2004.01294.x.

Mero, S.W. 1992. Food habits of walleye and sauger in Lake Sakakawea, North Dakota.

Master’s Thesis. South Dakota State University, Brookings.

Mittelbach, G.G. and Persson, L. 1998. The ontogeny of piscivory and its ecological consequences. Can. J. Fish. Aquat. Sci. 55(6): 1454-1465. doi:10.1139/f98-041.

Morbey, Y.E., Vascotto, K., and Shuter, B.J. 2007. Dynamics of piscivory by lake trout following a smallmouth bass invasion: A historical reconstruction. T. Am. Fish. Soc.

136(2): 477-483. doi:10.1577/T06-070.1.

Nelson, W.R. 1968. Reproduction and early life history of sauger, Stizostedion canadense, in Lewis and Clark Lake. T. Am. Fish. Soc. 97(2): 159-166. doi:10.1577/1548-8659(1968)97[159:RAELHO]2.0.CO;2.

Nelson, W.R. and Walburg, C.H. 1977. Population dynamics of yellow perch (Perca flavescens), sauger (Stizostedion canadense), and walleye (S. vitreum vitreum) in four main stem Missouri River reservoirs. J. Fish. Board Can. 34(10): 1748-1763. doi:10.1139/f77-240.

! 120

Newsome, S.D., Martinez del Rio, C., Bearhop, S., and Phillips, D.L. 2007. A niche for isotopic ecology. Front. Ecol. Environ. 5(8): 429-436. doi:10.1890/060150.01.

O’Connor, C.M., Gilmour, K.M., Arlinghaus, R., Hasler, C.T., Philipp, D.P., and Cooke,

S.J. 2010. Seasonal carryover effects following the administration of cortisol to a wild teleost fish. Physiol. Biochem. Zool. 83(6): 950-957. doi:10.1086/656286.

Olin, M., Rask, M., Ruuhijarvi, J., Kurkilahti, M., Ala-Opas, P., and Ylonen, O. 2002.

Fish community structure in mesotrophic and eutrophic lakes of southern Finland: the relative abundances of percids and cyprinids along a trophic gradient. J. Fish Biol. 60(3):

593-612. doi:10.1111/j.1095-8649.2002.tb01687.x.

Overman, N. C. and Parrish, D. L. 2001. Stable isotope composition of walleye: 15N accumulation with age and area-specific differences in δ13C. Can. J. Fish. Aquat. Sci.

58(6): 1253-1260. doi:10.1139/f01-072.

Paradis, Y., Bertolo, A., Perot, A., Debas, S., and Magnan, P. 2006. Do benthivory and piscivory result in similar growth in walleye?. J. Fish Biol. 69(5): 1317-1329. doi:10.1111/j.1095-8649.2006.01187.x.

Parnell, A. and Jackson, A. 2015. SIAR: Stable Isotope Analysis in R. R Package version

4.2.

! 121

Patalas, J. 1996. Summary of walleye mortality studies at live-release tournaments in

Alberta 1991-1994. R. L. & L. Environmental Services Ltd., Edmonton, Alberta, Canada.

Peake, S., McKinley, R.S., and Barth, C. 1997. Effect of recovery parameters on critical swimming speed of juvenile rainbow trout (Oncorhynchus mykiss). Can. J. Zoolog.

75(10): 1724-1727. doi: 10.1139/z97-800.

Pegg, M.A., Bettoli, P.W., and Layzer, J.B. 1997. Movement of saugers in the lower

Tennessee River determined by radio telemetry, and implications for management. N.

Am. J. Fish. Manage. 17(3): 763-768. doi:10.1577/1548-

8675(1997)017<0763:MOSITL>2.3.CO;2.

Perga, M.E. and Gerdeaux, D. 2003. Using the δ13C and δ15N of whitefish scales for retrospective ecological studies: changes in isotope signatures during the restoration of

Lake Geneva, 1980-2001. J. Fish Biol. 63(5): 1197-1207. doi:10.1046/j.1095-

8649.2003.00239.x.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83: 703-718. doi:10.1890/0012-

9658(2002)083[0703:USITET]2.0.CO;2.

Pottinger, T.G. 1998. Changes in blood cortisol, glucose and lactate in carp retained in anglers' keepnets. J. Fish Biol. 53(4): 728-742. ! 122 doi:10.1111/j.1095-8649.1998.tb01828.x.

Quist, M.C., Stephen, J.L., Lynott, S.T., Goeckler, J.M., and Schultz, R.D. 2010. An evaluation of angler harvest of walleye and saugeye in a Kansas reservoir. J. Freshwater

Ecol. 25(1): 1-7. doi:10.1080/02705060.2010.9664351.

Raby, G.D., Donaldson, M.R., Hinch, S.G., Patterson, D.A., Lotto, A.G., Robichaud, D.,

English, K.K., Willmore, W.G., Farrell, A.P., Davis, M.W., and Cooke, S.J. 2012.

Validation of reflex indicators for measuring vitality and predicting the delayed mortality of wild coho salmon bycatch released from fishing gears. J. Appl. Ecol. 49(1): 90-98. doi:10.1111/j.1365-2664.2011.02073.x.

Radke, R.J. and Eckmann, R. 2001. No general percid dominance under mesotrophic lake conditions: a test of several hypotheses. Limnologica 31(1): 37-44. doi:10.1016/S0075-

9511(01)80046-3.

Rawson, D.S. 1957. The life history and ecology of the yellow walleye, Stizostedion vitreum, in Lac La Ronge, Saskatchewan. T. Am. Fish. Soc. 86(1): 15-37. doi:10.1577/1548-8659(1956)86[15:TLHAEO]2.0.CO;2.

Rawson, M.R. and Scholl, R.L. 1978. Reestablishment of sauger in western Lake Erie. In

Selected coolwater fishes of North America. Edited by R.L. Kendall. American Fisheries

Society, Special Publication 11, Bethesda, Maryland. pp. 261-265. ! 123

Richard, A., Bernatchez, L., Valiquette, E., and Dionne, M. 2014. Telemetry reveals how catch and release affects prespawning migration in Atlantic salmon (Salmo salar). Can. J.

Fish. Aquat. Sci. 71(11): 1730-1739. doi:10.1139/cjfas-2014-0072.

Roseman, E.F., Schaeffer, J.S., Bright, E., and Fielder, D.G. 2014. Angler-caught piscivore diets reflect fish community changes in Lake Huron. T. Am. Fish. Soc. 143(6):

1419-1433. doi:10.1080/00028487.2014.945659.

Ross, S.T. 1986. Resource partitioning in fish assemblages: a review of field studies.

Copeia 1986(2): 352-388. doi:10.2307/1444996.

Royer, L.M. 1972. Limnology and fisheries of Lake Diefenbaker, an impoundment on the

South Saskatchewan River, 1967-69. Technical report. Saskatchewan Fisheries

Laboratory, Fisheries and Wildlife Branch, Department of Natural Resources,

Saskatchewan, Canada.

Saskatchewan Anglers’ Guide. 2015. Available from the Government of Saskatchewan,

Saskatchewan, Canada.

Schmidt, K., McClelland, J.W., Mente, E., Montoya, J.P., Atkinson, A., and Voss, M.

2004. Trophic-level interpretation based on δ15N values: implications of tissue-specific ! 124 fractionation and amino acid composition. Mar. Ecol. Prog. Ser. 266: 43-58. doi:10.3354/meps266043.

Schramm Jr, H.L., Haydt, P.J., and Portier, K.M. 1987. Evaluation of prerelease, postrelease, and total mortality, of Largemouth Bass caught during tournaments, in two

Florida lakes. N. Am. J. Fish. Manage. 7(3): 394-402. doi:10.1577/1548-

8659(1987)7<394:EOPPAT>2.0.CO;2.

Schramm Jr, H.L., Armstrong, M.L., Fedler, A.J., Funicelli, N.A., Green, D.M., Hahn,

J.L., Lee, D.P., Manns Jr, R.E., Quinn, S.P., and Waters, S.J. 1991. Sociological, economic, and biological aspects of competitive fishing. Fisheries 16(3): 13-21. doi:10.1577/1548-8446(1991)016<0013:SEABAO>2.0.CO;2.

Schramm Jr, H.L. and Hunt, K.M. 2007. Issues, benefits, and problems associated with fishing tournaments in inland waters of the United States: a survey of fishery agency administrators. Fisheries 32(5): 234-243. doi:10.1577/1548-

8446(2007)32[234:IBAPAW]2.0.CO;2.

Schramm Jr, H.L., Vondracek, B., French, W.E., and Gerard, P.D. 2010. Factors associated with mortality of walleyes and saugers caught in live-release tournaments. N.

Am. J. Fish. Manage. 30(1): 238-253. doi:10.1577/M09-003.1.

Scott, W.B. and Crossman, E.J. 1973. The freshwater fishes of Canada. Fish. Res. Board ! 125

Can. Ottawa, Ontario.

Seehausen O. and Bouton N. 1997. Microdistribution and fluctuations in niche overlap in a rocky shore cichlid community in Lake Victoria. Ecol. Freshw. Fish 6: 161-173. doi:10.1111/j.1600-0633.1997.tb00159.x.

Sheppard, K.T., Davoren, G.K., and Hann, B.J. 2015. Diet of walleye and sauger and morphological characteristics of their prey in Lake Winnipeg. J. Great Lakes Res. 3(41):

907-915. doi:10.1016/j.jglr.2015.05.006.

Somers, C.M. and Morris, M. 2015. Saskatchewan Fishes, A Pocket Naturalist Guide.

Waterford Press, Dunedin, FL, USA.

Sport Fishing in Saskatchewan. 2010. Summary report. Government of Saskatchewan,

Saskatchewan, Canada.

State of Lake Diefenbaker. 2012. Water Security Agency, Saskatchewan, Canada.

Staudinger, M.D., McAlarney, R.J., McLellan, W.A., and Ann Pabst, D. 2014. Foraging ecology and niche overlap in pygmy (Kogia breviceps) and dwarf (Kogia sima) sperm whales from waters of the US mid‐Atlantic coast. Mar. Mammal Sci. 30(2): 626-655. doi:10.1111/mms.12064.

! 126

Sullivan, C., Hasler, C., and Suski, C.D. 2015. Do live-well temperatures differ from ambient water during black bass tournaments?. N. Am. J. Fish. Manage. 35(5): 1064-

1069. doi:10.1080/02755947.2015.1079575.

Suski, C.D., Killen, S.S., Cooke, S.J., Kieffer, J.D., Philipp, D.P., and Tufts, B.L. 2004.

Physiological significance of the weigh-in during live-release angling tournaments for largemouth bass. T. Am. Fish. Soc. 133(6): 1291-1303. doi:10.1577/T03-145.1.

Swenson, W.A. 1977. Food consumption of walleye (Stizostedion vitreum vitreum) and sauger (S. canadense) in relation to food availability and physical conditions in Lake of the Woods, Minnesota, Shagawa Lake, and western Lake Superior. J. Fish. Board Can.

34(10): 1643-1654. doi:10.1139/f77-229.

Swenson, W.A. and Smith Jr, L.L. 1976. Influence of food competition, predation, and cannibalism on walleye (Stizostedion vitreum vitreum) and sauger (S. canadense) populations in Lake of the Woods, Minnesota. J. Fish. Board Can. 33(9): 1946-1954. doi:10.1139/f76-248.

Syväranta, J., Vesala, S., Rask, M., Ruuhijärvi, J., and Jones, R.I. 2008. Evaluating the utility of stable isotope analyses of archived freshwater sample materials. Hydrobiologia

600(1): 121-130. doi:10.1007/s10750-007-9181-3.

! 127

Talmage, P.J. and Staples, D.F. 2011. Mortality of walleyes angled from the deep waters of Rainy Lake, Minnesota. N. Am. J. Fish. Manage. 31(5): 826-831. doi:10.1080/02755947.2011.623759.

Thompson, A.L. 2009. Walleye habitat use, spawning behaviour, and egg deposition in

Sandusky Bay, Lake Erie. Doctoral dissertation. Ohio State University.

Tieszen, L.L., Boutton, T.W., Tesdahl, K.G., and Slade, N.A. 1983. Fractionation and turnover of stable carbon isotopes in animal tissues: implications for δ13C analysis of diet.

Oecologia 57(1-2): 32-37. doi:10.1007/BF00379558.

Trippel, E.A. and Beamish, F.W.H. 1993. Multiple trophic level structuring in

Salvelinus–Coregonus assemblages in boreal forest lakes. Can. J. Fish. Aquat. Sci. 50(7):

1442-1455. doi:10.1139/f93-165.

Vander Zanden, M.J. and Rasmussen, J.B. 1996. A trophic position model of pelagic food webs: impact on contaminant bioaccumulation in lake trout. Ecol. Monogr. 66(4):

451-477. doi:10.2307/2963490.

Vander Zanden, M.J., Shuter, B.J., Lester, N.P., and Rasmussen, J.B. 2000. Within-and among-population variation in the trophic position of a pelagic predator, lake trout

(Salvelinus namaycush). Can. J. Fish. Aquat. Sci. 57(4): 725-731. doi:10.1139/f00-011.

! 128

Vander Zanden, H.B., Bjorndal, K.A., and Bolten, A.B. 2013. Temporal consistency and individual specialization in resource use by green turtles in successive life stages.

Oecologia 173(3): 767-777. doi:10.1007/s00442-013-2655-2.

Wainright, S.C., Fogarty, M.J., Greenfield, R.C., and Fry, B. 1993. Long-term changes in the Georges Bank food web: trends in stable isotopic compositions of fish scales. Mar.

Biol. 115(3): 481-493. doi:10.1007/BF00349847.

Wallace, R., Jensen, R., and Murphy, K.M. 2004. Monitoring hatchery use of spawning walleye in Lake Diefenbaker, Saskatchewan. In Proceedings of PERCIS III, the Third

International Percid Fish Symposium. Edited by T.P. Barry and J. A. Malison. University of Wisconsin, Madison, Wisconsin. pp. 127-128.

Wallace, R., Jensen, R., and Merkowsky, J. 2010. Assessment of fish stocks in Lake

Diefenbaker, 2004. Ministry of Environment, Fish and Wildlife Branch, Saskatchewan,

Canada.

Walleye – Saskatchewan Record. 2013. Record Fish Canada. Accessed May 2016.

.

Warwick, W.F. and Tisdale, N.A. 1988. Morphological deformities in Chironomus,

Cryptochironomus, and Procladius larvae (Diptera: Chironomidae) from two differentially stressed sites in Tobin Lake, Saskatchewan. Can. J. Fish. Aquat. Sci. 45(7):

1123-1144. doi:10.1139/f88-136.

! 129

Waterhouse, M.D., Sloss, B.L., and Isermann, D.A. 2014. Relationships among walleye population characteristics and genetic diversity in northern Wisconsin lakes. T. Am. Fish.

Soc. 143(3): 744-756. doi:10.1080/00028487.2014.880742.

Wilde, G.R. 2003. Dispersal of tournament-caught black bass. Fisheries 28(7): 10-17. doi:10.1577/1548-8446(2003)28[10:DOTBB]2.0.CO;2.

Wilde, G.R. and Paulson, L.J. 2003. Movement and dispersal of tournament-caught largemouth bass in Lake Mead, Arizona-Nevada. J. Freshwater Ecol. 18(2): 339-342. doi:10.1080/02705060.2003.9664502.

Worm, B., Barbier, E.B., Beaumont, N., Duffy, J.E., Folke, C., Halpern, B.S., Jackson,

J.B., Lotze, H.K., Micheli, F., Palumbi, S.R., and Sala, E. 2006. Impacts of biodiversity loss on ocean ecosystem services. Science 314(5800): 787-790. doi:10.1126/science.1132294.

! 130

6. APPENDICES

6.1. Appendix I: Angler survey

Name:

Where are you from?

How long have you been an angler?

1) Where did you catch the fish? Please mark on the map.

2) What depth did you catch the fish at?

3) How long was the fish kept in your live well?

4) What temperature do you keep your live well at?

5) What type of bait did you use to catch the fish?

6) Did you use any kind of live well additive? What kind?