Climate Change Effects on Lake Erie Yellow Perch Reproduction and Recruitment

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Troy Mason Farmer, B.S., M.S.

Graduate Program in Evolution, Ecology, and Organismal Biology

The Ohio State University

2013

Dissertation Committee:

Dr. Stuart A. Ludsin, Co-advisor

Dr. Elizabeth A. Marschall, Co-advisor

Dr. Konrad Dabrowski

Dr. Maria N. Miriti

Troy M. Farmer

2013

ABSTRACT

Climate warming is expected to positively affect cool-water, temperate fish populations by lengthening the growing season and expanding thermal habitats suitable for positive growth. Yet, little is known about how a corresponding shortened winter might affect temperate fish populations, especially for species that require a prolonged period of cold temperature during the winter prior to spawning for proper ovary development. Additionally, events such as hypolimnetic hypoxia (O2 < 2 mg/L), are expected to increase with continue warming. We hypothesized that climate change would negatively affect temperate fish populations by 1) increasing bottom hypoxia during summer, which can reduce energy reserves (fish condition) prior to winter, when ovaries develop for many species, and 2) increasing winter water temperature, which could increase basal metabolic rates during winter (i.e., reduce energy available for ovary development) and disrupt thermal requirements necessary for proper ovary development.

To test these hypotheses, we investigated the effects of winter temperature and female condition on Lake Erie yellow perch Perca flavescens reproductive development, egg and larval quality, and ultimately, fall juvenile abundance (a strong predictor of future recruitment to the yellow perch fishery in Lake Erie). Towards this end, we conducted laboratory experiments, a multi-year field study, and historical analyses. In our laboratory experiments, female yellow perch exposed to a long winter produced

ii higher quality eggs (i.e., in terms of size, energetic, and lipid content) that both hatched at higher rates and produced larger larvae than lower quality eggs from females exposed to a short winter (Chapters 2 and 3). Counter to our hypotheses, reduced female condition entering winter did not adversely affect reproductive success (Chapter 3). Additionally, field and laboratory studies found that when spring warming happened extremely early, yellow perch spawning did not fully adjust, increasing the possibility of a mis-match between first-feeding larvae and their zooplankton prey (Chapter 2). Finally, we show through historical analyses that the negative effect of warm winters on Lake Erie yellow perch juvenile abundance appears to be consistent over 42 years (i.e., 1969-2010), and has persisted throughout a large-scale, nutrient-driven regime shift and restructuring of the food-web due to numerous introductions of invasive species (Chapter 4).

Our research offers a previously unrecognized mechanism by which climate change can threaten temperate fish populations, through reductions in reproductive success. Our results also may have relevance to fisheries managers seeking to better anticipate the responses of fish populations to climate change. Specifically, given that our study has identified mechanisms that appear to be responsible for long-term population dynamics, our findings may allow for managers to monitor the appropriate variables (i.e., winter thermal regime) necessary to predict annual recruitment to the fishery for Lake Erie yellow perch. Additionally, because our study species has similar life-history and physiological requirements not unlike many other cool-water, temperate fishes, our findings may have relevance to fish populations in many ecosystems.

iii

Dedicated to Karen, Landon, Mom, and Dad

AKNOWLEDGEMENTS

I sincerely thank my co-advisors, Drs. Stuart Ludsin and Elizabeth Marschall for the exceptional guidance and support that they provided during the past several years. I am grateful for all of their insightful advice (and helpful edits) that largely shaped the direction and focus of my dissertation research. In addition to being top-notch scientists, they are even better people, who maintain their humanity while operating at incredibly high levels. In this manner, they have served as excellent role models. Along these lines,

I also thank Dr. Roy Stein for serving as a mentor during the past several years. I also am grateful to Dr. Konrad Dabrowski for serving as a committee member and for sharing with me his expert knowledge in many areas, particularly in spawning yellow perch in the laboratory. I also am greatly appreciative of Dr. Maria Miriti for serving as a committee member and for offering numerous helpful comments regarding this work.

I owe a great deal of thanks to the students and staff of the OSU Aquatic Ecology

Laboratory and the Aquaculture Laboratory for their assistance in conducting this research and for making life in the lab enjoyable. Specifically, former head-technicians

Alex Johnson, Jeramy Pinkerton, Theo Gover, and Chelsea Schmit deserve special thanks for their dedication to all aspects of this project, and for putting up with my frequent and varied requests. Simply put, the success of this research project would not have been possible without their hard work and dedication. Special thanks are due to Melissa

Marburger and Margarita Talavera both of whom assisted me with countless purchasing and HR requests. Additionally, numerous OSU graduate students, visiting scientists, post-doctoral researchers, and undergraduate student volunteers deserve thanks for conducting field collections, caring for live yellow perch in the lab, processing samples, and providing helpful suggestions. While a comprehensive list of those deserving thanks would fill multiple pages, I owe a great deal of gratitude to Mike Bahler, Ben Bolam,

Ruth Briland, Reed Brodnick, Emily Burbacher, Alex Chen, Chelsea Coble, Paris

Collingsworth, Kristen DeVanna, Chris French, Seyoum Gebremariam, Dave Glover,

Brittany Gunther, Christina Haska, Paul Hurtado, Breanna Humbarger, Jahn Kallis, Jeff

Kemper, Bryan Kinter, Karolina Kwasek, Cassie May, Kevin Pangle, Tim Parker,

Thomas Peterson, Nevine Abou Shabana, Tyler Stuebe, Jason Van Tassell, and Erich

Williams for their contributions and support.

I greatly appreciate assistance from biologists and research staff at the Ohio

Division of Wildlife (DOW) Fairport and Sandusky Fisheries Stations. Special thanks are due to Carey Knight and Ann Marie Gorman for sharing their expert knowledge of

Lake Erie yellow perch with me, and for involving me in a variety of related research activities being conducted out of the Fairport office. For their help with coordinating logistics and conducting field collections John Deller, Travis Hartmann, Jeff Tyson, Eric

Wiemer, and Chris Vandergoot also deserve special thanks. I also appreciate technical assistance from supervisors and staff from the Ohio DOW Hebron State Fish Hatchery, who modified the AEL pool facility so that this research could be conducted. Special thanks go out to Elmer Heyob and Brian Kitchen for their assistance. Also, thanks to

Mort Pugh and his staff at the Ohio DOW St. Mary’s State Fish Hatchery for providing feeder minnows during our experiments. Additionally, I am grateful to the biologists and staff from the USGS Lake Erie Biological Station. Thanks to Patrick Kocovsky, Richard

Kraus, Dale Hale and Tim Cherry, who made spring sampling trips in western Lake Erie possible.

Finally, I would like to thank my family for their enduring love and support during this time. Karen and Landon, thank you for tolerating the many hours that I spent in the lab and field and for sustaining me with your unconditional love. Mom, Dad, and

Kyle, thank you for your support through all these years of graduate school, and the many years before. Thanks also to Elise, Kyle, Kelley, Lauren, David, Angie, Max, and Sarah, for your support. It has been trying to be so far from everyone that we love, and I know the sacrifices have not been ours alone. Your constant love and support helped Karen,

Landon, and I make it though the past few years.

Support for this research was provided by the Federal Aid in Sport Fish

Restoration Program (F-69-P, Fish Management in Ohio), administered jointly by the

U.S. Fish and Wildlife Service and the Ohio DOW as State Project FADR62. Additional support was provided by the Great Lakes Fishery Commission, the International

Associate for Great Lakes Research, and by a Presidential Fellowship awarded by The

Ohio State University.

vii

VITA

April 10, 1982 …………………………… Born – Birmingham, Alabama

2004……………………………………… B.S. Graduation with university honors, cum laud, Auburn University

2008………………………………………. M.S. Auburn University

2008-2012………………………………… Graduate Teaching and Research Associate The Ohio State University

2013…………………………………...... University Presidential Fellow The Ohio State University

PUBLICATIONS

Farmer, T.M., D.R. DeVries, R.A. Wright, and J.E. Gagnon. 2013. Using seasonal variation in otolith microchemical composition to indicate largemouth bass and southern flounder residency patterns across an estuarine salinity gradient. Transactions of the American Fisheries Society 142: 1415-1429. Lowe, M.L., S.A. Ludsin, B.J. Fryer, R.A. Wright, D.R. DeVries, T.M. Farmer. 2012. Response to “Comment on "Otolith Microchemistry Reveal Substantial Use of Freshwater by Southern Flounder in the Northern Gulf of Mexico” by Pedro Morais. Estuaries and Coasts 35: 107-110. Farmer, T.M., Wright, R.A., DeVries, D.R. 2010. Mercury concentration in two estuarine fish populations across a seasonal salinity gradient. Transactions of the American Fisheries Society 139: 1896-1912.

FIELDS OF STUDY

Major Field: Evolution, Ecology, and Organismal Biology

viii

TABLE OF CONTENTS

Abstract ...... ii Dedication ...... iv Aknowledgements...... v Vita ...... viii List of Tables ...... xii List of Figures ...... xiv

Chapter 1: Introduction ...... 1 References ...... 8

Chapter 2: Winter warming effects on reproductive success threaten cool-water fish populations ...... 16 Supplementary materials and methods ...... 22 Historical analysis of recruitment and ice cover ...... 24 Laboratory experiment ...... 25 Timing of spawning in Lake Erie ...... 34 References ...... 36

Chapter 3: Winter warming and reduced energy reserves affect reproduction in a cool- water, temperate fish species ...... 50 Introduction ...... 50 Methods ...... 53 Overview ...... 53 Fish collection...... 54 Experimental design and procedures ...... 55

Sample processing and analysis ...... 59 Data analysis ...... 62 Results ...... 64 Pre-experiment reproductive development ...... 64 Spawning, fertilization, and embryo hatching success ...... 64 Fecundity and egg quality...... 66 Male reproductive quality ...... 68 Discussion ...... 69

Chapter 4: The role of climate change, invasive species, and nutrient-driven regime shift in a Great Lakes fish population ...... 97 Introduction ...... 97 Methods ...... 101 Study site ...... 101 Index of year perch fall juvenile abundance ...... 104 Explanatory variables ...... 105 Data analysis ...... 109 Results ...... 112 Trends in yellow perch juvenile abundance and spawning-stock size ...... 112 Comparing multiple regression and stock-recruit models ...... 113 West basin juvenile abundance ...... 114 Central basin juvenile abundance ...... 115 Assessing correlation and collinearity ...... 117 Discussion ...... 119 References ...... 125

Appendix A: Predicting yellow perch juvenile abudance in Lake Erie given future, projected changes in climate ...... 143 Introduction ...... 143 Methods ...... 144 Model overview ...... 144 Calculating historical larval production ...... 145 Relating historical larval abundance to year-class strength ...... 146 Creating current probability distributions of recruitment ...... 147 Creating future probability distributions of recruitment ...... 148 Results ...... 149

Discussion ...... 152 References ...... 154

Bibliography ...... 162

LIST OF TABLES Table Page

2.1. Associated p-values for Pearson correlation coefficients testing the response of fecundity (number of eggs per g of ribbon), egg quality, hatching success (%), and larval size to indices of female yellow perch size (nearest 1 mm total length), age, and condition (both spring and fall separately), and incubation temperature (ºC). An index of female condition was created from residuals of length-mass relationships developed for all individuals measured in the spring (time of spawning) and fall (initiation of winter experiment). Bold values indicate p < 0.05. These results show that all of our metrics of egg and larval quality were unrelated to individual female attributes or incubation temperatures...... 42 4.1. Explanatory and response variables used in our investigation of factors affecting yellow perch juvenile (age-0) abundance in the west (WB) and central basins (CB) of Lake Erie, 1969-2010 (CPUE: catch per unit effort; # individuals per trawling minute)...... 133

4.2. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the west basin of Lake Erie, 1969-1985. For each model, the overall 2 p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semipartial correlation coefficient (sr2: the amount of unique variance it explains) also is presented...... 134

4.3. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the west basin of Lake Erie, 1986-2010. For each model, the overall 2 p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semi partial correlation coefficient (sr2: the amount of unique variance it explains) also is presented...... 135

xii

4.4. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the central basin of Lake Erie, 1969-1985. For each model, the 2 overall p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semipartial correlation coefficient (sr2: the amount of unique variance it explains) also is presented...... 136

4.5. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the central basin of Lake Erie, 1986-2010. For each model, the 2 overall p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semipartial correlation coefficient (sr2: the amount of unique variance it explains) also is presented...... 137 A1. Probability of achieving designated year-class strengths (Failed: cpue age-0 yellow perch CPUE < 5; Weak: 5-14 CPUE; Moderate: 15-29 CPUE; Strong: 30-60 CPUE; Very Strong: CPUE > 60) under current (1975-2010) and future (mid- century: 2046- 2065) emissions scenarios (A1b: high, A2: moderate, B1: low; IPCC 2007) for management unit two in Lake Erie...... 156

xiii

LIST OF FIGURES Figure Page

2.1. Yellow perch juvenile (age-0) abundance, plotted against mean ice cover during the winter (February-March; % areal coverage) prior to hatching in western and central Lake Erie (1973-2010). Vertical dashed lines (and associated p-values) indicate threshold values of mean winter ice cover as indicated by two-dimensional Kolmogorov-Smirnov tests (Garvey et al. 1998). Above the threshold, juvenile abundance (a strong predictor of recruitment to the fishery at age-2 and future harvest [Figures 2.2, 2.3]) may be high or low, but below the threshold only low abundances occur. Thus, winter duration may ‘set the stage’ for future high recruitment to the fishery. Juvenile abundance was determined from annual, fisheries-independent Ohio Division of Wildlife bottom trawling surveys, is presented as catch-per-unit-effort (CPUE)...... 43 2.2. Yellow perch cohort size at age-2 in year t + 2 plotted against juvenile abundance in year t (year in which each cohort was hatched) for western and central Lake Erie from 1987-2010. Estimates of cohort size at age-2 for western and central Lake Erie are from the Great Lakes Fishery Commission Lake Erie Yellow Perch Task Group (YPTG 2013). Results indicate juvenile (age-0) abundance is a good proxy for future recruitment to age-2, when yellow perch typically become reproductively mature and enter the fishery...... 44 2.3. Lifetime harvest by cohort (cumulative harvest ages 2– 5) plotted against juvenile abundance in year t (year in which each cohort was hatched) for western and central Lake Erie from 1987-2010 (YPTG 2013). These results show that a strong annual cohort, which is only possible after long, cold winters, can support the fishery for many years afterwards...... 45 2.4. Relationships between: A) winter duration and individual yellow perch egg mass (mg); B) hatching success (%) and individual egg mass (mg); and C) larval total length (mm) and individual egg mass (mg). All data were collected during a controlled laboratory experiment and are presented as tank means (± 1 SE). Females exposed to a long winter produced large eggs that hatched at higher rates and produced larger larvae than small eggs...... 46

xiv

2.5. Mean daily water temperature (°C) from short (50 d < 5°C) and long (110 d < 5°C) winter-duration treatments used in laboratory experiments with Lake Erie yellow perch. The duration of spawning is shown for each treatment. Labels indicate the dates and water temperatures during which females spawned. Despite earlier arrival of seeming suitable spawning temperatures (8-11°C) following a short winter, females did not spawn during these temperatures, but instead delayed spawning until the typical spawning period (i.e., April-May [Hokanson 1977]). ... 47 2.6. Mean (± 1 SE) bottom water temperature measured in western Lake Erie near Sandusky, Ohio during the yellow perch spawning season, 2010-2012...... 48 2.7. Probability that a yellow perch female was spent (i.e., had completed spawning) given bottom water temperature (°C) during spring 2010-2012 spawning seasons in western Lake Erie. Colored lines determined by logistic regression; dashed lines indicate 95% confidence intervals. Solid black lines indicate probabilities determined for 8, 10, and 12°C each year. Lines with different letters differed among years (p < 0.05). Despite earlier arrival of seemingly suitable spawning temperatures (8-11°C) following a short winter in 2012, females did not spawn earlier, but instead delayed spawning until the typical spawning period (April-May [Hokanson 1977])...... 49 3.1. Mean (± 1 SE) residuals (derived from sex-specific total length – mass relationships) for male and female Lake Erie yellow perch from high (ad libitum; female N = 49; male N = 27) and low (maintenance ration; female N = 49; male N = 27) body-condition treatments. Means with different lowercase letters within a panel differed in Tukey’s honestly significant difference post-hoc comparison (p < 0.05)...... 86 3.2. Mean (± 1 SE) growth from late July through early October for male and female Lake Erie yellow perch in high (female N = 49; male N = 27) and low (female N = 49; male N = 27) body-condition treatments. Means with different lowercase letters differed after correction for multiple comparisons (Tukey’s honestly significant difference post-hoc comparison; p < 0.05)...... 87 3.3. Female and male yellow perch gonad mass (g) versus total length for individuals sacrificed from high and low body-condition treatments prior to the start of the experiment (during the first week of October)...... 88 3.4. Hatching success relative to fertilization success for Lake Erie yellow perch eggs across all experimental treatments. The dashed line indicates the 1:1 line...... 89 3.5. Mean (± 1 SE) hatching success (to the nearest %) of yellow perch eggs plotted against mean (± 1 SE) residuals (derived from female total length – mass relationships) of female Lake Erie yellow perch condition. Negative residual values indicate females in relatively poorer condition, whereas positive residuals indicate females in relatively better condition...... 90

3.6. Lake Erie yellow perch fecundity (eggs per female) shown as a function of total length. A single relationship describes the number of eggs produced from females in both the short and long winter-duration treatments in our experiment...... 91 3.7. Mean (± 1 SE) fecundity residuals (from Figure 3.6) plotted against mean (± 1 SE) residuals (derived from sex-specific total length – mass relationships) for Lake Erie yellow perch. Negative residual values indicate females in relatively poorer condition while positive residuals indicate females in relatively better condition.. 92 3.8. Mean (± 1 SE) gonadosomatic index (GSI: [gonad mass/body mass]*100) for female yellow perch hand-stripped of eggs at the conclusion of both the short and long winter-duration treatments. Means with different lowercase letters within a panel differed (Tukey’s honestly significant difference post-hoc comparison; p < 0.05)...... 93 3.9. Mean (± 1 SE) egg diameter for Lake Erie yellow perch hand-stripped of eggs in the short and long winter-duration treatments. Means with different lowercase letters within a panel differed (Tukey’s honestly significant difference post-hoc comparison; p < 0.05)...... 94 3.10. Scatterplots of mean (± 1 SE) hatching success relative to mean (± 1 SE) egg mass and mean (± 1 SE) energy density for fertilized eggs from Lake Erie yellow perch in our laboratory experiments...... 95 3.11. Individual sperm motility in male yellow perch exposed to short versus long winter-durations in a controlled laboratory experiment. Individual measurements of sperm motility (short winter N = 23; long winter N = 34) were collected from males (short winter N = 12 males; long winter N = 18 males) selected haphazardly during the spawning season for fertilization of egg samples...... 96 4.1. Winter (February-March) ice cover (% areal coverage lakewide) and Lake Erie age-1+ white perch abundance (catch per unit effort [CPUE] for the west and central basins, 1969-2010). The vertical dashed line indicates the location where data were split for analysis (1969-1985 vs. 1986-2010)...... 138 4.2. Yellow perch juvenile (age-0) abundance (bars = catch-per-unit-effort [CPUE]) and spawning stock size (lines = age-3+ population size in millions of fish) for the west and central basins of Lake Erie. Juvenile abundance was estimated from Ohio Division of Wildlife (DOW) bottom trawl surveys conducted in fall 1969-2010. Age-3+ population size was obtained from Lake Erie Yellow Perch Task Group reports (YPTG 2013). The vertical dashed line indicates the location where data were split for analysis (1969-1985 vs. 1986-2010)...... 139

4.3. Akaike variable weights (ωi) for explanatory variables used in multiple regression models to explain yellow perch juvenile abundance in the west and central basins of Lake Erie, 1969-1985 and 1986-2010. Akaike variable weights range from 0-1, and are the relative importance of each explanatory variable in each sub set of models...... 140

xvi

4.4. Yellow perch juvenile (age-0) abundance catch per unit effort (CPUE) in western Lake Erie for the early (1969-1985) and late (1986-2010) periods. For each period, ice cover classification is indicated and juvenile abundance is plotted against both May air temperature (°C) and Maumee River discharge (m3/s)...... 141 4.5. Yellow perch juvenile (age-0) abundance catch per unit effort (CPUE) in central Lake Erie for the early (1969-1985) and late (1986-2010) periods, plotted against Maumee River discharge (m3/s). During both periods, ice cover classification also is indicated...... 142 A1. Mean (± SE) number of days Lake Erie water temperature < 5°C under current (1975–2010) and predicted future environmental conditions at mid-century (2046– 2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007). Lowercase letters indicate means that differed significantly (Tukey’s honestly significant difference post-hoc comparison; p < 0.05)...... 157 A2. Probability distribution of hatching success (% Hatched) for yellow perch eggs given current (1975-2010) and predicted future environmental conditions at mid- century (2046-2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007)...... 158 A3. Probability distribution of total larval production (in millions; estimated each spring from age-specific adult size distributions and relationships developed from our laboratory experiment) given current (1975-2010) and predicted future environmental conditions at mid-century (2046-2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007)...... 159 A4. Age-0 yellow perch catch-per-unit-effort (CPUE) during October 1975-2010 for central Lake Erie (management unit 2) versus total larval yellow perch production (in millions; estimated each spring from age-specific adult size distributions and relationships developed from our laboratory experiment). Lines plotted are for 99th, 90th, 75th, 50th, 25th, and 1st regression quantile estimates...... 160 A5. Probability distribution of predicted yellow perch catch-per-unit-effort (CPUE) in October given current (1975-2010) and predicted future environmental conditions at mid-century (2046-2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007)...... 161

xvii

CHAPTER 1

INTRODUCTION

Climate change is predicted to affect marine (Grantham et al. 2004; Mackenzie et al. 2007; Kerr et al. 2009), freshwater (Meisner et al. 1987; Magnuson et al. 1997; Kling et al. 2003; Schindler et al. 2005) and estuarine ecosystems (Genner et al. 2004; Willard and Bernhardt 2011) across the globe. The world’s oceans, which began warming around the turn of the 20th century due to climate change (Rhein et al. 2013), experienced a rapid increase in the rate of warming during the latter half of the 20th and start of the 21st centuries (Balmaseda et al. 2013). This increased rate of warming was largely driven by the fact that marine systems have absorbed about 90% of the total heat that has been added to the climate system since the mid-1970s (Bindoff and Willebrand 2007; Rhein et al.

2013; IPCC 2013). Large, north-temperate freshwater systems such as the Laurentian

Great Lakes (Assel et al. 1995; McCormick and Fahnenstiel 1999; Jones et al. 2006) also have experienced dramatic warming trends during the latter half of the 20th century.

Further, in the case of the Great Lakes, water temperature has been increasing twice as fast as air temperature, due largely to a reduction in winter ice cover (Austin and Colman 2007;

Wang et al. 2012). While previous modeling studies have explored the effects of a changing climate on fish communities (Shuter and Meisner 1992; Stefan et al. 1996;

Magnuson et al. 1997;

Mackenzie-Grieve and Post 2006; Sharma et al. 2007; Cline et al. 2013), they have focused primarily on how warmer temperatures during the growing season would affect thermal habitat availability, expected growth rate, and species range. Because climate change also is anticipated to alter precipitation patterns (Kunkel et al. 2002; Kling et al.

2003), nutrient dynamics (Kerfoot et al. 2008), water levels (Hartmann 1990; Mortsch and

Quinn 1996; Kling et al. 2003), nursery and spawning habitat (Mortsch 1998; Jones et al.

2006), as well as the timing, magnitude, and duration of thermal stratification (McCormick and Fahnenstiel 1999; Austin and Colman 2007), hypoxia (Fang et al. 2004; Hawley et al.

2006), and ice cover (Assel et al. 1995, 2003; Magnuson 2010; Wang et al. 2012), one should expect climate change to have multiple interactive or synergistic effects on fish populations and communities (Brook et al. 2008). In turn, climate change would be expected to lead to “ecological surprises” (Paine et al. 1998; Doak et al. 2008) especially when we consider that the effects of climate change may interact with other anthropogenic stressors (e.g., effects of invasive species, eutrophication, overfishing, habitat loss) to influence species composition, distribution, and production (Vitousek

1994; Jackson et al. 2001; Ficke et al. 2007; Hellmann et al. 2008). As global fisheries provide over 3 billion people with greater than 15% of their annual protein intake (FAO

2010), sustainable management of fisheries stocks in the face of a changing climate is critical to securing food for the world’s growing population (Pauly et al. 2002) and maintaining biodiverse (Folke et al. 2004) and biocomplex ecosystems (Hilborn et al.

2003) that are both resistant and resilent to environmental change.

Climate change, whether considered solo or in tandem with other anthropogenic perturbations, has the potential to directly and indirectly influence the vital rates of fish populations (e.g., growth, survival, and reproduction; Kerr et al. 2009; Ottersen et al.

2010). One vital rate that appears highly susceptible to climate impacts is reproduction, as temperature is a strong regulator of reproductive processes (e.g., vitellogenesis and ovulation) in temperate fish species (Dabrowski et al. 1996; Wang et al. 2010). The impact of climate change on fish populations via reproductive pathways, however, remains largely unexplored (e.g., Donelson et al. 2010; Pankhurst and Munday 2011), despite knowledge from laboratory experiments that pre-spawning temperature can influence reproductive development and spawning (Burt et al. 2011). Notably, other recent studies have investigated how climate change might affect fish populations by impacting early life processes, such as survival of eggs and larvae (Svendsen et al. 1995;

Casselman 2002; Kӧster et al. 2005; Vikebø et al. 2005). While we cannot discount the importance of investigations that explore climate effects on early life growth and survival, the lack of attention paid to understanding how climate can influence reproductive output of adult female spawners is a major gap in both the freshwater and marine climate change literature.

Our research addresses this major information gap by proposing and investigating mechanisms through which climate may affect fish reproductive biology and physiology.

The central thesis of our research is that climate change will negatively affect

reproductive development and spawning by 1) increasing hypoxia (O2 < 2 mg/L), which could reduce energy reserves prior to winter, when ovaries develop for many fishes, and

2) increasing winter temperature, which could increase basal metabolic rates during winter (i.e., reduce energy available for ovary development) and disrupt thermal requirements necessary for proper ovary development. Recent evidence from the

Laurentian Great Lakes indicates a positive effect of cold winter temperatures on juvenile abundance, with the highest abundances following long, cold winters and lowest abundances following short, warm winters (Crane 2007; Ludsin 2000). Additionally, experimental and field-observational research in North American and European ecosystems has demonstrated that fecundity and spawning success of cool-water fish populations can decline when adult females do not experience a prolonged period of cold temperatures during winter (Jones et al. 1972; Hokanson 1977; Ciereszko et al. 1997;

Sandstrom et al. 1997; Luksiene et al. 2000; Migaud et al. 2002; Wang et al. 2006).

Furthermore, events such as hypolimnetic hypoxia, which reduce female condition prior to winter, may exacerbate the negative effects of warm winters on reproductive development. Hypoxia is predicted to increase with continued climate warming (Fang et al. 2004) and has been shown to reduce benthic foraging, energetic condition, and growth potential of several fish species in the Laurentian Great Lakes (Roberts et al. 2009; Arend et al. 2010; Scavia et al. in revision). Reduced energetic condition has been shown to reduce egg production (Bunnell et al. 2005, 2009; Kennedy et al. 2008; Gregersen et al.

2011) and increase the incidence of skipped spawning (Scott and Crossman 1974;

Rideout et al. 2005; Jørgensen et al. 2006; Skjæraasen et al. 2009, 2012). Despite this evidence, however, the biological mechanisms underlying the relationships among winter temperature, female energetic condition, and female reproductive development are

4 unknown and no study to date has explored the cohort-level response of such relationships.

Herein, we investigated the effects of winter temperature and female condition on

Lake Erie yellow perch Perca flavescens reproductive development, egg quality, spawning success, larval quality, and ultimately fall juvenile abundance (a strong predictor of future recruitment to the yellow perch fishery in Lake Erie [YPTG 2013]).

While the mechanisms being explored herein are relevant to a wide array of species, yellow perch is an ideal study species for three key reasons. First, it is widely distributed throughout North America and is of recreational, commercial, and economic importance across the Great Lakes basin (Craig 2000). Second, an abundance of spatially explicit biological data are available for this species (e.g., age-specific estimates of population size, size distributions, and biomass; age-, size-, and sex-specific maturation schedules).

Finally, yellow perch in Lake Erie and across the Great Lakes, like many other Great

Lakes fish species, exhibit extreme annual variability in first-year survival (YPTG 2013), which appears to be driven by meteorological forces that are largely dictated by climate

(Ludsin 2000; Crane 2007).

Lake Erie is one of the five North American Laurentian Great Lakes that together compose, by surface area, the largest freshwater system on earth (Bolsenga and

Herdendorf 1993). A growing body of evidence exists to indicate that north temperate systems such as the Laurentian Great Lakes have been experiencing a warming trend during recent decades (Assel et al. 1995; McCormick and Fahnenstiel 1999; Jones et al.

2006; Wang et al. 2012). For example, over Lake Erie, average winter air temperature

5 increased by 2.4ºC (r = 0.39, P = 0.003) and the annual average number of days with air temperatures below freezing decreased by 16 d during 1956-2012 (r = -0.34, P = 0.01).

Water temperatures and the duration of summer stratification also have increased across the Great Lakes during the 20th century (McCormick and Fahnenstiel 1999), and general circulation models (GCMs) indicate that this warming trend is expected to continue throughout this century (Magnuson et al. 1997; Lofgren et al. 2002; Kling et al. 2003;

Hayhoe et al. 2010). Also, seasonal hypoxia in Lake Erie’s central basin has been increasing in recent years (Hawley et al. 2006; Rucinski et al. 2010), which is strongly related to the strength and duration of thermal stratification (Burns et al. 2005; Zhou et al.

2013). Hypoxia has been shown to have negative effects on yellow perch foraging

(Roberts et al. 2009), growth (Arend et al. 2010), body condition (Scavia et al. in revision) and, possibly, energy reserves entering the winter.

To investigate the mechanisms by which climate change may affect yellow perch reproduction and, subsequently, population dynamics, we conducted analyses of historical data, laboratory experiments, and a multi-year field study. In Chapter 2 we present long-term (1973-2010) field patterns, which show that short, warm winters were associated with failed recruitment events and then reveal through experimentation and field studies how winter warming effects on egg and larval quality and on spawning time underlie these patterns. In Chapter 3, we present additional results from laboratory experiments that link egg quality and hatching success to winter temperature and female condition entering winter. Finally in Chapter 4, we show that the negative effect of warm winters on yellow perch juvenile abundance in Lake Erie appears to be consistent over 42

6 years, and has persisted throughout a large-scale nutrient-driven regime shift (Ludsin et al. 2001) and restructuring of the food web due to numerous introductions of invasive species (Mills et al. 1993; Ricciardi and MacIsaac 2000).

While our primary objective was to understand the mechanisms by which climate change can affect cool-water fishes, such as yellow perch, our results also have relevance to their management. Knowledge of the various ways in which climate change can influence the reproductive success and demographics of Lake Erie yellow perch would enhance the ability of agencies seeking to explain inter-annual variation in past fishery dynamics, and also better anticipate responses of the fishery to future climate conditions.

It also could indicate the appropriate variables to monitor (e.g., winter thermal regime) to predict annual recruitment to the fishery in the future. Additionally, the ability to explain inter-annual variation in population (and fishery) dynamics lends credibility to management agencies (which increases support for management actions), and allows agencies to guide user-group expectations given future, predicted environmental conditions. Finally, given that yellow perch has life-history and physiological requirements not unlike many other cool-water, temperate fishes, insights provided herein will have relevance to management of fish populations in many ecosystems.

References

Food and Agricultural Organization of the United Nations (FAO). 2010. The state of the world's fisheries and aquaculture. Rome.

Arend, K. K., D. Beletsky, J. V. DePinto, S. A. Ludsin, J. J. Roberts, D. K. Rucinski, D. Scavia, D. J. Schwab, and T. O. Hӧӧk. 2010. Seasonal and interannual effects of hypoxia on fish habitat quality in central Lake Erie. Freshwater Biology 56:366- 383.

Assel, R. A., D. M. Robertson, M. H. Hoff, and J. H. Selgeby. 1995. Climatic-change implications from long-term (1823-1994) ice records for the Laurentian Great Lakes. Pages 383-386 in D. A. Rothrock, editor Annals of Glaciology, Vol 21: Proceedings of the International Symposium on the Role of the Cryosphere in Global Change.

Assel, R. A., D. M. Robertson, M. H. Hoff, and J. H. Selgeby. 1995. Climatic-change implications from long-term (1823-1994) ice records for the Laurentian Great Lakes.

Austin, J. A., and S. M. Colman. 2007. Lake Superior summer water temperatures are increasing more rapidly than regional air temperatures: A positive ice-albedo feedback. Geophysical Research Letters 34(6).

Balmaseda, M. A., K. E. Trenberth, and E. Källén. 2013. Distinctive climate signals in reanalysis of global ocean heat content. Geophysical Research Letters 40:1754- 1759.

Bindoff, N. L., and J. Willebrand. 2007. Observations: Oceanic Climate Change and Sea Level. In: Climate Change 2007: The Physical Science Basis. Cambridge University Press, New York.

Bolsenga, S. J., and C. E. Herdendorf. 1993. Lake Erie and Lake St. Clair Handbook. Wayne State University Press, Detriot.

Brook, B. W., N. S. Sodhi, and C. J. A. Bradshaw. 2008. Synergies among extinction drivers under global change. Trends in Ecology & Evolution 23:453-460.

Bunnell, D. B., S. R. David, and C. P. Madenjian. 2009. Decline in bloater fecundity in Southern Lake Michigan after decline of Diporeia. Journal of Great Lakes Research 35:45-49.

Bunnell, D. B., M. A. Scantland, and R. A. Stein. 2005. Testing for evidence of maternal effects among individuals and populations of white crappie. Transactions of the American Fisheries Society 134:607-619.

Burns, N. M., D. C. Rockwell, P. E. Bertram, D. M. Dolan, and J. J. H. Ciborowski. 2005. Trends in temperature, Secchi depth, and dissolved oxygen depletion rates in the central basin of Lake Erie, 1983-2002. Journal of Great Lakes Research 31:35-49.

Burt, J. M., S. G. Hinch, and D. A. Patterson. 2011. The importance of parentage in assessing temperature effects on fish early life history: a review of the experimental literature. Reviews in Fish Biology and Fisheries 21:377-406.

Casselman, J. M. 2002. Effects of temperature, global extremes, and climate change on year-class production of warmwater, coolwater, and coldwater fishes in the Great Lakes Basin. Pages 39-59 in N. A. McGinn, editor. Fisheries in a Changing Climate, volume 32.

Ciereszko, R. E., K. Dabrowski, A. Ciereszko, J. Ebeling, and J. S. Ottobre. 1997. Effects of temperature and photoperiod on reproduction of female yellow perch Perca flavescens: plasma concentrations of steroid hormones, spontaneous and induced ovulation, and quality of eggs. Journal of the World Aquaculture Society 28:344- 356.

Cline, T. J., V. Bennington, and J. F. Kitchell. 2013. Climate change expands the spatial extent and duration of preferred thermal habitat for Lake Superior fishes. Plos One 8(4).

Craig, J. F. 2000. Percid fishes: systematics, ecology and exploitation. Blackwell Sciences, Oxford.

Crane, V. 2007. Lower trophic level and climate influences on western Lake Erie fish recruitment 1988 through 2005. Master's thesis. The Ohio State University, Columbus.

Dabrowski, K., R. E. Ciereszko, A. Ciereszko, G. P. Toth, S. A. Christ, D. ElSaidy, and J. S. Ottobre. 1996. Reproductive physiology of yellow perch (Perca flavescens): Environmental and endocrinological cues. Journal of Applied Ichthyology- Zeitschrift Fur Angewandte Ichthyologie 12:139-148.

Doak, D. F., J. A. Estes, B. S. Halpern, U. Jacob, D. R. Lindberg, J. Lovvorn, D. H. Monson, M. T. Tinker, T. M. Williams, J. T. Wootton, I. Carroll, M. Emmerson, F. Micheli, and M. Novak. 2008. Understanding and predicting ecological dynamics: Are major surprises inevitable? Ecology 89:952-961.

Donelson, J. M., P. L. Munday, M. I. McCormick, N. W. Pankhurst, and P. M. Pankhurst. 2010. Effects of elevated water temperature and food availability on the reproductive performance of a coral reef fish. Marine Ecology Progress Series 401:233-243.

Fang, X., H. G. Stefan, J. G. Eaton, J. H. McCormick, and S. R. Alam. 2004. Simulation of thermal/dissolved oxygen habitat for fishes in lakes under different climate scenarios - Part 1. Cool-water fish in the contiguous US. Ecological Modelling 172:13-37.

Ficke, A. D., C. A. Myrick, and L. J. Hansen. 2007. Potential impacts of global climate change on freshwater fisheries. Reviews in Fish Biology and Fisheries 17:581- 613.

Folke, C., S. Carpenter, B. Walker, M. Scheffer, T. Elmqvist, L. Gunderson, and C. S. Holling. 2004. Regime shifts, resilience, and biodiversity in ecosystem management. Annual Review of Ecology Evolution and Systematics 35:557-581.

Genner, M. J., D. W. Sims, V. J. Wearmouth, E. J. Southall, A. J. Southward, P. A. Henderson, and S. J. Hawkins. 2004. Regional climatic warming drives long-term community changes of British marine fish. Proceedings of the Royal Society of London Series B-Biological Sciences 271:655-661.

Grantham, B. A., F. Chan, K. J. Nielsen, D. S. Fox, J. A. Barth, A. Huyer, J. Lubchenco, and B. A. Menge. 2004. Upwelling-driven nearshore hypoxia signals ecosystem and oceanographic changes in the northeast Pacific. Nature 429:749-754.

Gregersen, F., L. A. Vollestad, K. Ostbye, P. Aass, and O. Hegge. 2011. Temperature and food-level effects on reproductive investment and egg mass in vendace, Coregonus albula. Fisheries Management and Ecology 18:263-269.

Hartmann, H. C. 1990. Climate change impacts on Laurentian Great Lakes levels. Climatic Change 17:49-67.

Hawley, N., T. H. Johengen, Y. R. Rao, D. Rueberg, D. Beletsky, S. A. Ludsin, B. J. Eadie, D. S. Schwab, and S. B. Croley. 2006. Lake Erie hypoxia prompts Canda- U.S. Study. EOS Transactions 87:313-324.

Hayhoe, K., J. VanDorn, T. Croley, II, N. Schlegal, and D. Wuebbles. 2010. Regional climate change projections for Chicago and the US Great Lakes. Journal of Great Lakes Research 36:7-21.

Hellmann, J. J., J. E. Byers, B. G. Bierwagen, and J. S. Dukes. 2008. Five potential consequences of climate change for invasive species. Conservation Biology 22:534-543.

Hilborn, R., T. P. Quinn, D. E. Schindler, and D. E. Rogers. 2003. Biocomplexity and fisheries sustainability. Proceedings of the National Academy of Sciences of the United States of America 100:6564-6568.

Hokanson, K. E. F. 1977. Temperature requirements of some percids and adaptations to seasonal temperature cycle. Journal of the Fisheries Research Board of Canada 34:1524-1550.

Intergovernmental Panel on Climate Change (IPCC). 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T. F., D. Qin, G. -K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P. M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, in press.

Jackson, J. B. C., M. X. Kirby, W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes, S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson, R. S. Steneck, M. J. Tegner, and R. R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629-638.

Jones, B. R., K. E. F. Hokanson, and J. H. McCormick. 1972. Winter temperature requirements for maturation and spawning of yellow perch Perca flavescens (Mitchill). M. Marios, editor Towards a plan of action for mankind. Proc. World Conf. Vol. 3 Biological balance and thermal modification. Pergamon Press, NY.

Jones, M. L., B. J. Shuter, Y. M. Zhao, and J. D. Stockwell. 2006. Forecasting effects of climate change on Great Lakes fisheries: models that link habitat supply to population dynamics can help. Canadian Journal of Fisheries and Aquatic Sciences 63:457-468.

Jørgensen, C., B. Ernande, O. Fiksen, and U. Dieckmann. 2006. The logic of skipped spawning in fish. Canadian Journal of Fisheries and Aquatic Sciences 63:200-211.

Kennedy, B. P., K. H. Nislow, and C. L. Folt. 2008. Habitat-mediated foraging limitations drive survival bottlenecks for juvenile salmon. Ecology 89:2529-2541.

Kerfoot, W. C., J. W. Budd, S. A. Green, J. B. Cotner, B. A. Biddanda, D. J. Schwab, and H. A. Vanderploeg. 2008. Doughnut in the desert: Late-winter production pulse in southern Lake Michigan. Limnology and Oceanography 53:589-604.

Kerr, L. A., W. J. Connelly, E. J. Martino, A. C. Peer, R. J. Woodland, and D. H. Secor. 2009. Climate Change in the U. S. Atlantic Affecting Recreational Fisheries. Reviews in Fisheries Science 17:267-289.

Kling, G. W., K. Hayhoe, L. B. Johnson, J. J. Magnuson, S. Polasky, S. K. Robinson, B. J. Shuter, M. M. Wander, D. J. Wuebbles, D. R. Zak, R. L. Lindroth, S. C. Moser, and M. L. Wilson. 2003. Confronting climate change in the Great Lakes region: impacts on our communities and ecosystems. Union of Concerned Scientists and Ecological Society of America, Cambridge, MA. 11

Kӧster, F. W., C. Mollmann, H. H. Hinrichsen, K. Wieland, J. Tomkiewicz, G. Kraus, R. Voss, A. Makarchouk, B. R. MacKenzie, M. A. St John, D. Schnack, N. Rohlf, T. Linkowski, and J. E. Beyer. 2005. Baltic cod recruitment - the impact of climate variability on key processes. ICES Journal of Marine Science 62:1408-1425.

Kunkel, K. E., N. E. Westcott, and D. A. R. Kristovich. 2002. Assessment of potential effects of climate change on heavy lake-effect snowstorms near Lake Erie. Journal of Great Lakes Research 28:521-536.

Lofgren, B., F. H. Quinn, A. H. Clites, R. A. Assel, A. J. Eberhardt, and C. L. Luukkonen. 2002. Evaluation of potential impacts on Great Lakes water resources based on climate scenarios of two GCM’s. Journal of Great Lakes Research 28:537-554.

Ludsin, S. A. 2000. Exploration of spatiotemporal patterns in recruitment and community organization of Lake Erie fishes: a multiscale, mechanistic approach. PhD dissertation. The Ohio State University, Columbus.

Ludsin, S. A., M. W. Kershner, K. A. Blocksom, R. L. Knight, and R. A. Stein. 2001. Life after death in Lake Erie: Nutrient controls drive fish species richness, rehabilitation. Ecological Applications 11:731-746.

Luksiene, D., O. Sandstrom, L. Lounashelmo, and J. Andersson. 2000. The effects of thermal effluent exposure on the gametogenesis of female fish. Journal of Fish Biology 56:37-50.

Mackenzie-Grieve, J. L., and J. R. Post. 2006. Projected impacts of climate warming on production of lake trout (Salvelinus namaycush) in southern Yukon lakes. Canadian Journal of Fisheries and Aquatic Sciences 63:788-797.

Mackenzie, B. R., H. Gislason, C. Mollmann, and F. W. Kӧster. 2007. Impact of 21st century climate change on the Baltic Sea fish community and fisheries. Global Change Biology 13:1348-1367.

Magnuson, J. J. 2010. History and heroes: the thermal niche of fishes and long-term lake ice dynamics. Journal of Fish Biology 77:1731-1744.

Magnuson, J. J., K. E. Webster, R. A. Assel, C. J. Bowser, P. J. Dillon, J. G. Eaton, H. E. Evans, E. J. Fee, R. I. Hall, L. R. Mortsch, D. W. Schindler, and F. H. Quinn. 1997. Potential effects of climate changes on aquatic systems: Laurentian Great Lakes and Precambrian Shield Region. Hydrological Processes 11:825-871.

McCormick, M. J., and G. L. Fahnenstiel. 1999. Recent climatic trends in nearshore water temperatures in the St. Lawrence Great Lakes. Limnology and Oceanography 44:530-540.

Meisner, J. D., J. L. Goodier, H. A. Regier, B. J. Shuter, and W. J. Christie. 1987. An assessment of the effects of climate warming on Great Lakes basin fishes. Journal of Great Lakes Research 13:340-352.

Migaud, H., P. Fontaine, I. Sulistyo, P. Kestemont, and J. N. Gardeur. 2002. Induction of out-of-season spawning in Eurasian perch Perca fluviatilis: effects of rates of cooling and cooling durations on female gametogenesis and spawning. Aquaculture 205:253-267.

Mills, E. L., J. H. Leach, J. T. Carlton, and C. L. Secor. 1993. Exotic species in the Great Lakes - a history of biotic crises and anthropogenic introductions. Journal of Great Lakes Research 19:1-54.

Mortsch, L. D. 1998. Assessing the impact of climate change on the Great Lakes shoreline wetlands. Climatic Change 40:391-416.

Mortsch, L. D., and F. H. Quinn. 1996. Climate change scenarios for Great Lakes Basin ecosystem studies. Limnology and Oceanography 41:903-911.

Ottersen, G., S. Kim, G. Huse, J. J. Polovina, and N. C. Stenseth. 2010. Major pathways by which climate may force marine fish populations. Journal of Marine Systems 79:343-360.

Paine, R. T., M. J. Tegner, and E. A. Johnson. 1998. Compounded perturbations yield ecological surprises. Ecosystems 1:535-545.

Pankhurst, N. W., and P. L. Munday. 2011. Effects of climate change on fish reproduction and early life history stages. Marine and Freshwater Research 62:1015-1026.

Pauly, D., V. Christensen, S. Guenette, T. J. Pitcher, U. R. Sumaila, C. J. Walters, R. Watson, and D. Zeller. 2002. Towards sustainability in world fisheries. Nature 418:689-695.

Rhein, M., S. R. Rintoul, S. Aoki, E. Campos, D. Chambers, R. A. Feely, S. Gulev, G. C. Johnson, S. A. Josey, A. Kostianoy, C. Mauritzen, D. Roemmich, L. D. Talley, and F. Wang. 2013. Observations: Ocean. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, United Kingdom and New York, NY, USA.

Ricciardi, A., and H. J. MacIsaac. 2000. Recent mass invasion of the North American Great Lakes by Ponto-Caspian species. Trends in Ecology & Evolution 15:62-65.

Rideout, R. M., G. A. Rose, and M. P. M. Burton. 2005. Skipped spawning in female iteroparous fishes. Fish and Fisheries 6:50-72. 13

Roberts, J. J., T. O. Hӧӧk, S. A. Ludsin, S. A. Pothoven, H. A. Vanderploeg, and S. B. Brandt. 2009. Effects of hypolimnetic hypoxia on foraging and distributions of Lake Erie yellow perch. Journal of Experimental Marine Biology and Ecology 381:S132-S142.

Rucinski, D. K., D. Beletsky, J. V. DePinto, D. J. Schwab, and D. Scavia. 2010. A simple 1-dimensional, climate based dissolved oxygen model for the central basin of Lake Erie. Journal of Great Lakes Research 36:465-476.

Sandstrom, O., I. Ambrahamsson, J. Andersson, and M. Vetemaa. 1997. Temperature effects on spawning and egg development in Eurasian perch. Journal of Fish Biology 51:1015-1024.

Scavia, D., J.D. Allan, K.A Arend, S. Bartell, D. Beletsky, N.S. Bosch, S.B. Brandt, R.D. Briland, I. Daloğlu, J.V. DePinto, D.M. Dolan, M.A. Evans, T.M. Farmer, D. Goto, H. Han, T.O. Höök, R. Knight, S.A. Ludsin, D. Mason, A.M. Michalak, R.P. Richards, J.J. Roberts, D.K. Rucinski, E. Rutherford, D.J. Schwab, T. Sesterhenn, H. Zhang, and Y. Zhou. in revision. Assessing and addressing the re- eutrophication of Lake Erie: central basin hypoxia. Journal of Great Lakes Research.

Schindler, D. E., D. E. Rogers, M. D. Scheuerell, and C. A. Abrey. 2005. Effects of changing climate on zooplankton and juvenile sockeye salmon growth in southwestern Alaska. Ecology 86:198-209.

Scott, W. B., and E. J. Crossman. 1974. Fisheries research board of Canada bulletin no. 184: fresh water fishes of Canada. Bulletin of the Fisheries Research Board of Canada: 966 pp.

Sharma, S., D. A. Jackson, C. K. Minns, and B. J. Shuter. 2007. Will northern fish populations be in hot water because of climate change? Global Change Biology 13:2052-2064.

Shuter, B. J., and J. D. Meisner. 1992. Tools for assessing the impact of climate change on frehswater fish populations. GeoJournal 28:7-20.

Skjæraasen, J. E., J. Kennedy, A. Thorsen, M. Fonn, B. N. Strand, I. Mayer, and O. S. Kjesbu. 2009. Mechanisms regulating oocyte recruitment and skipped spawning in Northeast Arctic cod (Gadus morhua). Canadian Journal of Fisheries and Aquatic Sciences 66:1582-1596.

Skjæraasen, J. E., R. D. M. Nash, K. Korsbrekke, M. Fonn, T. Nilsen, J. Kennedy, K. H. Nedreaas, A. Thorsen, P. R. Witthames, A. J. Geffen, H. Hoie, and O. S. Kjesbu. 2012. Frequent skipped spawning in the world's largest cod population. Proceedings of the National Academy of Sciences of the United States of America 109:8995-8999. 14

Stefan, H. G., M. Hondzo, X. Fang, J. G. Eaton, and J. H. McCormick. 1996. Simulated long-term temperature and dissolved oxygen characteristics of lakes in the north- central United States and associated fish habitat limits. Limnology and Oceanography 41:1124-1135.

Svendsen, E., A. Aglen, S. A. Iversen, D. W. Skagen, and O. Smestad. 1995. Influence of climate on recruitment and migration of fish stocks in the North Sea. Canadian Special Publication of Fisheries and Aquatic Sciences 121:641-653.

Vikebø, F., S. Sundby, B. Adlandsvik, and O. Fiksen. 2005. The combined effect of transport and temperature on distribution and growth of larvae and pelagic juveniles of Arcto-Norwegian cod. ICES Journal of Marine Science 62:1375- 1386.

Vitousek, P. M. 1994. Beyond global warming - ecology and global change. Ecology 75:1861-1876.

Wang, J., X. Bai, H. Hu, A. Clites, M. Colton, and B. Lofgren. 2012. Temporal and spatial variability of Great Lakes ice cover, 1973-2010. Journal of Climate 25:1318-1329.

Wang, N., J.-N. Gardeur, E. Henrotte, M. Marie, P. Kestemont, and P. Fontaine. 2006. Determinism of the induction of the reproductive cycle in female Eurasian perch, Perca fluviatilis: Identification of environmental cues and permissive factors. Aquaculture 261:706-714.

Wang, N., F. Teletchea, P. Kestemont, S. Milla, and P. Fontaine. 2010. Photothermal control of the reproductive cycle in temperate fishes. Reviews in Aquaculture 2:209-222.

Willard, D. A., and C. E. Bernhardt. 2011. Impacts of past climate and sea level change on Everglades wetlands: placing a century of anthropogenic change into a late- Holocene context. Climatic Change 107:59-80.

Yellow Perch Task Group (YPTG). 2013. Report of the Lake Erie yellow perch task group. Great Lakes Fishery Commission, Ann Arbor, MI.

Zhou, Y., D. R. Obenour, D. Scavia, T. H. Johengen, and A. M. Michalak. 2013. Spatial and Temporal Trends in Lake Erie Hypoxia, 1987-2007. Environmental Science & Technology 47:899-905.

CHAPTER 2

WINTER WARMING EFFECTS ON REPRODUCTIVE SUCCESS THREATEN

COOL-WATER FISH POPULATIONS

To properly assess the effects of climate change on fish populations and communities, studies of physiological mechanisms that relate growth, survival, and reproductive performance to thermal and climate-driven stressors are required (Portner and Farrell 2008; Eliason et al. 2011). As examples, the effects of thermal oxygen limitation on survival (Portner and Knust 2007) and thermal habitat availability on expected growth rates (Magnuson and Destasio 1997; Neuheimer et al. 2011) have been evaluated, in turn helping to understand distributional shifts in response to increased temperatures (Perry et al. 2005; Sharma et al. 2007). Based on these findings, warming is expected to benefit temperate (cool-water) fishes by increasing thermal growth habitat and lengthening the growing season (Magnuson and Destasio 1997; Magnuson et al.

1997). The impact of climate change on fish populations via reproductive pathways, however, remains largely unexplored (e.g., Donelson et al. 2010; Pankhurst and Munday

2011) despite knowledge from laboratory experiments that pre-spawning temperature can influence reproductive development and spawning (Burt et al. 2011). Further, because

16 such studies have used domesticated or lab-reared broodstock (not wild fish) in combination with environmentally irrelevant photothermal conditions (Burt et al. 2011), the potential for climate warming to drive the dynamics of fish populations through effects on reproductive success remains unknown.

This gap in knowledge is important for three reasons. First, because the number of eggs produced annually sets the absolute maximum potential cohort size for that year

(Houde 1987), those aspects of climate change that affect reproductive output may strongly impact population dynamics and viability (Kerr et al. 2009; Ottersen et al. 2010).

Second, temperature can be a critical determinant of reproductive processes (e.g., vitellogenesis, ovulation, hatching success) in temperate fishes (Wang et al. 2010), with experimental and field-observational research in North American and European ecosystems documenting reductions in fecundity, oocyte quality, and spawning success when adult females do not experience a prolonged period of cold temperatures during the winter prior to spawning (Hokanson 1977; Ciereszko et al. 1997; Sandstrom et al. 1997;

Webb et al. 2001; Migaud et al. 2002; Clark et al. 2005; Brown et al. 2006). Further, we have documented a positive threshold relationship between an index of winter longevity

(February-March ice cover) and the summertime abundance of juvenile yellow perch

(Perca flavescens) (Figure 2.1). Specifically, short, warm winters with little ice cover were always followed by weak recruitment to the juvenile stage, which is a strong predictor of recruitment to the fishery at age-2 (Figure 2.2) and future lifetime harvest of that cohort (Figure 2.3). By contrast, high juvenile abundance only occurred after long, cold winters (Figure 2.1). Third, a shortened winter (and early spring onset) may alter the

17 phenology (i.e., timing) of spawning, eventual larval fish emergence, and prey production, thereby increasing the potential for a mismatch between first-feeding larvae and their planktonic prey during spring (Cushing 1990; Edwards and Richardson 2004;

Winder and Schindler 2004).

Despite this support for our hypothesis that climate change will negatively affect reproductive success of temperate fishes by increasing the frequency of short, warm winters, the biological mechanisms underlying the relationship between winter temperature and reproductive success remain largely hypothetical. Further, no study has yet related the laboratory response of individual fish to population patterns in the wild.

Herein, we report findings from a laboratory experiment and field study conducted with

Lake Erie yellow perch, which quantified the effects of winter longevity on the timing of reproduction, spawning success, fecundity, egg quality, hatching success, and larval quality. We used Lake Erie yellow perch as our study organism because i) it develops ovaries during winter and spawns during spring, a common reproductive strategy in temperate regions (Wang et al. 2010) and ii) it, along with its congener, European perch

Perca fluviatilis, is ecologically, economically, and culturally important in North

America and Europe.

While laboratory experimentation revealed that fecundity was entirely a function of female size (table S1) and did not differ between long and short winter-duration treatments (p = 0.44), it also showed that females exposed to a long winter produced larger eggs than those exposed to a short winter (Figure 2.4). Further, large eggs had higher energetic content (R2 = 0.60, p = 0.02), total lipids (R2 = 0.63, p = 0.03), and

18 neutral lipids (R2 = 0.70, p = 0.02) than small eggs and hatched at higher rates and produced larger larvae than small eggs (Figure 2.4). Because large size-at-hatching typically confers larval growth and survival advantages (Miller et al. 1988; Einum and

Fleming 2000), long, cold winters should result in higher numbers of successful larvae than short, warm winters via high hatching rates of large eggs and high survival rates of large larvae. We also found egg quality, hatching success, and larval size to be unrelated to incubation temperatures and female size, age, and condition (Table 2.1), pointing to winter-duration effects on gametogenesis as the causal mechanism for observed differences between experimental treatments.

One potential explanation for large eggs being produced after long winters is that cold temperatures allow for more energy to be allocated to reproduction than under warmer conditions, when maintenance metabolic demands are elevated (Wootton 1998).

Alternatively, warm temperatures may disrupt hormone and sex steroid production, reducing maternal investment in oocytes (Webb et al. 2001; Clark et al. 2005) and inducing apoptotic processes in the ovary (Ito et al. 2008). Additional laboratory and modeling studies are required to understand whether warm winters reduce egg size by affecting endocrine processes, energy, lipid, or essential fatty acid allocation patterns, or all of the above.

Beyond finding that larval quantity and size were reduced following short winters, we documented differences in the timing of spawning in response to spring warming in both the laboratory and wild. In the short-winter treatment, females did not spawn at the temperatures commonly observed for this species in the wild (8-11°C) (Hokanson 1977),

19 which occurred in mid-March in our short-winter treatment. Instead, these females delayed spawning until the historically more typical spawning-period (April-May)

(Hokanson 1977), even though temperatures (~15°C) greatly exceeded those typically observed during yellow perch spawning in the wild (Figure 2.5). A similar phenomenon also was observed in Lake Erie. Despite the much earlier arrival of warm spring temperatures during 2012 relative to 2010 and 2011 (Figure 2.6), yellow perch did not spawn earlier during 2012 (Figure 2.7).

These findings suggest that yellow perch cannot spawn before a certain date. But, after that date has passed, spawning becomes dependent on temperature. Previous laboratory experiments with yellow perch support this conclusion (Hokanson 1977;

Dabrowski et al. 1996). Photoperiod, therefore, may play a role in controlling the final stages of ovulation and spawning, as has been suggested for this species (Dabrowski et al.

1996; Abdulfatah et al. 2013). Or, perhaps the timing of spawning is imprinted genetically, as recently documented for some other temperate fishes (Ottera et al. 2012).

Regardless of the mechanism, inability to adjust spawning in response to early spring warming may result in low larval survival, if hatching does not coincide with peak zooplankton densities (Cushing 1990; Edwards and Richardson 2004; Winder and

Schindler 2004; Wright and Trippel 2009). Although other studies have documented earlier spawning by fish in response to earlier arrival of spring (Ahas 1999; Wedekind and Kueng 2010), our results highlight that the ability to shift spawning times in response to a large phase-shift in the spring thermal regime may be species-specific and dependent on a complex set of conditions (not just temperature). Thus, when the timing of peak

20 zooplankton production in temperate regions primarily tracks temperature (e.g.,

Richardson 2008), winter warming should be expected to negatively affect larval survival and subsequent population growth potential for temperate fishes (Cushing 1990; Wright and Trippel 2009) in the absence of any rapid life-history change or evolution.

Although the photothermal requirements for proper hormonal stimulation of reproductive development and spawning are complex and species-specific (Wang et al.

2010), our conclusions seem applicable to other ecologically and economically important spring-spawning temperate fishes with potentially similar photothermal requirements.

For example, a recent classification of temperate fishes into functional groups based on their reproductive strategies grouped yellow perch with a wide diversity of early-spring spawning species (e.g., European perch, walleye Sander vitreus, pike-perch Sander lucioperca, striped bass Morone saxatilis; Wang et al. 2010). These species also require a long winter chilling duration to complete vitellogenesis, followed by a rise in spring temperature, with photoperiod also seemingly important to the completion of ovulation and spawning (Dabrowski et al. 1996; Wang et al. 2010).

While many climate-driven stressors have been proposed to affect older fish larvae, juveniles, and adults (Ficke et al. 2007; Ottersen et al. 2010; Pankhurst and

Munday 2011), their importance as a driver of recruitment and subsequent population dynamics is likely to be reduced, if the total number of potential offspring has already been severely limited by altered thermal regimes during reproductive development and spawning periods. In this way, winter warming, through its effects on reproductive success, may ‘set the stage’ (i.e., potential number of pre-recruits) upon which these other

21 factors eventually can act. For temperate species in particular, our findings are significant in that they show how warming may negatively affect a group of fishes previously expected to benefit from warming (Magnuson and Destasio 1997; Magnuson et al. 1997). Continued investigation into how climate warming influences fish reproductive success offers a means to not only begin to understand past fishery dynamics (e.g., Lake Erie yellow perch), but also could help identify a potential impediment to recovery for other populations currently at historically low levels.

Supplementary materials and methods

Below we provide details on our study system and study species. Additionally, we provide details concerning our historical analysis of yellow perch recruitment, the experiment, and the determination of spawning time. All field collections and laboratory experiments were conducted according to animal use guidelines outlined in IACUC protocol # 2009A0073 at The Ohio State University.

Study system and species

As with many freshwater and marine systems globally, the Laurentian Great

Lakes already have begun to warm (McCormick and Fahnenstiel 1999; Jones et al. 2006).

Over Lake Erie, average winter air temperature increased by 2.4ºC (r = 0.39, p = 0.003) and the annual average number of days with air temperatures below freezing decreased by 16 d during 1956-2012 (r = -0.34, p = 0.01). Additionally, average air temperatures are expected to continue to rise in the Great Lakes basin, with predicted increases of 2-

5°C during winter by 2050, depending on emission scenarios (Hayhoe et al. 2010).

Further, in the case of the Great Lakes, water temperature has been increasing twice as fast as air temperature (Austin and Colman 2007), due largely to a reduction in winter ice cover, which has decreased 72% over the past 4 decades(Wang et al. 2012).

Yellow perch is a common, cool-water iteroparous fish that is widespread across the Atlantic, Great Lakes, and Mississippi River basins of North America. This species is particularly important in the Laurentian Great Lakes basin, where it serves as an important consumer in the middle of the food web (Craig 2000) and supports valuable commercial and recreational fisheries. Yellow perch support Lake Erie’s largest commercial fishery and second most valuable recreational fishery (YPTG 2013).

Recruitment to the fishery has been quite variable in Lake Erie through time (YPTG

2013), with no strong year-classes being produced since 2003 (YPTG 2013).

In Lake Erie, as across much of its range, yellow perch develop ovaries during winter months (Henderson et al. 2000) and spawn during spring (mid-April-May

(Collingsworth and Marschall 2011) across the lake, with different local spawning stocks existing that mix in the open lake as adults (Sepulveda-Villet and Stepien 2011;

Kocovsky et al. 2013). Eggs hatch and develop into pelagic larvae during May through

June, depending on lake basin, with larvae becoming demersal after about 25-35 d

(Gopalan et al. 1998; Ludsin 2000). Juveniles recruit to fishery-independent assessment gear by August of their first year of life and eventually become reproductively mature and enter the fishery at age-2. Juvenile abundance is a strong predictor of recruitment to the fishery at age-2 (Figure 2.2). In turn, because strong recruitment events to the fishery

23 at age-2 support the fishery for many years afterwards (Figure 2.3), factors (e.g., winter duration) that influence juvenile abundance leave a long legacy that is evident in fishery harvest.

Historical analysis of recruitment and ice cover

Analysis of historical Lake Erie yellow perch population dynamics (1973-2010) used juvenile (age-0) yellow perch catch rates generated annually during October bottom trawl surveys conducted by the Ohio Department of Natural Resources-Ohio Division of

Wildlife, as coordinated by the Great Lakes Fishery Commission Lake Erie Yellow Perch

Task Group. Yellow perch juvenile abundance from 1973-2010 was estimated as the number of juveniles caught per minute of bottom trawling. As sampling was conducted by multiple vessels during these years, we applied vessel-specific fishing power corrections to juvenile abundances from 1982-2010 to standardize catches (Tyson et al.

2006). Juvenile abundances from 1973-1981 had no fishing power corrections applied, due to a lack of studies comparing older and modern research vessels. However, using only data from 1982-2010, with fishing power corrections applied, we obtained similar significant two-dimensional thresholds for yellow perch juvenile fall abundance and ice cover data from the previous winter (sensu Figure 2.1). Thus, we feel confident that sampling biases do not underlie our observed relationships.

As juvenile survey designs have changed during the past 35+ years, we used only yellow perch data from twelve fixed, historical sites, sampled consistently during this time. Historical sites were spread across the Ohio waters of western (N=4) and central

(N=8) Lake Erie (see Ludsin et al. 2001). Annual juvenile abundances from historical sites correlated strongly with overall annual abundance calculated using all sites in both the western (r = 0.92; ~ 80 sites per year since 1987) and central Lake Erie (r = 0.94; ~

40 sites per year since 1990), indicating historical sites closely track population-level variation in juvenile abundance.

Ice cover trends from 1973-2010 for Lake Erie were compiled from ice charts collected by the Canadian Ice Service and NOAA National Ice Center and summarized by the NOAA Great Lakes Environmental Research Laboratory, Ann Arbor, Michigan,

USA (Wang et al. 2012). Daily values of lake area ice cover were obtained by interpolating between ice chart values, originally obtained at approximately weekly intervals. Mean February-March lake area ice cover was calculated by averaging daily ice cover values during these two months. As ice cover reaches its peak on Lake Erie during February through March (Wang et al. 2012), we selected this timeframe to distinguish between years of low and high ice cover. Late winter (February – March) ice cover was highly correlated (r = -0.86, p < 0.0001) with an independent index of mean winter air temperature (January-March; 1973-2010) developed for Lake Erie (T. Farmer unpublished data), indicating ice cover is a valid indicator of winter severity.

Laboratory experiment

We conducted a controlled laboratory experiment with these yellow perch during

October 2011 through June 2012 to quantify the effects of winter duration on the timing of spawning, ovarian development, fecundity, egg hatching success, and the quality of

25 eggs and larvae. Below, we provide details regarding the field collections of fish used in the experiment, experimental design, sample processing, and data analysis.

Field-collections. The central basin yellow perch population is the largest in Lake

Erie, representing 66% and 70% of the population in terms of lake-wide abundance and biomass, respectively, during 1987-2010 (YPTG 2013). For this reason, we collected live male and female yellow perch for this study from central Lake Erie near Fairport

Harbor, Ohio, USA (41° 46’ N, 81° 21’ W). All individuals were collected via bottom trawling conducted aboard the Ohio Department of Natural Resources – Division of

Wildlife’s R/V Grandon during April-May 2010-2011 (when sex could be determined by external examination). Immediately upon collection, all individuals were placed into live wells and ‘fizzed’ with a hypodermic needle to prevent over-inflation of the gas bladder

(Keniry et al. 1996), a common occurrence that can cause mortality in fish caught in deep water and brought to the surface. Fizzing reduced post-capture mortality and did not adversely affect long-term survival, as previously reported (Keniry et al. 1996). After collection, fish were held in 2,500-L circular tanks from May – October 2011, which were supplied with constant aeration and flow-through water. All individuals were fed live fathead minnows (Pimephales promelas) and were tagged with passive integrated transponder (PIT) tags to monitor individual growth and maturation.

Experimental design. Our experiment quantified the effects of winter duration

(number of days < 5°C, levels = 50 and 110 d) on the timing of reproduction, spawning success, fecundity, egg quality, embryo hatching success, larval size. Winter-duration treatments and the rates of fall cooling and spring warming (both 0.25°C/d) used in our

26 experiment were based on historical (1994-2010) field measurements of water temperature collected from central Lake Erie (41° 32’ 53’’ N, 81° 44’ 60’’ W) at a

Cleveland, Ohio water intake (TMF, unpublished data). Our two winter-duration treatment levels were intended to simulate historical (110 d) and future (50 d) conditions for Lake Erie. Our short winter duration (50 d) equaled the number of days < 5°C recorded during winter 2002, an ice-free winter, the frequency of which has been increasing during the past several decades (Wang et al. 2012).

The experiment was conducted in walk-in environmental chambers (N=2) in which all physicochemical conditions were controlled. Each chamber represented a single winter-duration treatment and contained a recirculating system with six 189-L tanks. Recirculating systems were supplied with continuous aeration along with physical and biological filtration to maintain water quality. Ammonia, nitrite, nitrate, and pH were measured daily and water changes were conducted as needed to maintain high levels of water quality (i.e., unionized ammonia < 0.1 ppm; nitrates < 40 ppm). Hobo® loggers (Onset®) recorded water temperature every 2 h in each recirculating system, and temperature and dissolved oxygen were measured daily with a YSI 550a handheld meter.

Lighting in all rooms was provided by incandescent lights and controlled by a digital system that simulated daily reductions (during the fall) and increases (during winter and spring) in photoperiod so as to mimic the photoperiod at Cleveland, Ohio. Additionally, at dawn and dusk, 1 h of increasing and decreasing light intensity preceded day and night periods, respectively.

The experiment began the first week of October. At this time, all fish were briefly anesthetized in MS-222, measured (nearest 1 mm total length, TL), weighed

(nearest 1 g wet mass), and scanned for PIT tags. Subsequently, individuals were randomly assigned to a winter-duration treatment and tank. Each individual tank contained 12-15 yellow perch (8-9 females and 3-7 males, with varying numbers used to try and standardize initial biomass), for a total of 49 females and 28 males and 49 females and 26 males in the long and short winter-duration treatments, respectively. During the experiment, all yellow perch were fed daily maintenance rations of live fathead minnows.

Maintenance rations were determined from an existing bioenergetics model (Kitchell et al. 1977), based on daily temperatures and the mass of individuals in each tank.

To simulate the end of winter and onset of spring, water temperatures were increased based on historical Lake Erie water temperature data (as described above).

During spring warming, females were monitored much of the day and were hand-stripped of eggs once signs of ovulation were present. Eggs were fertilized using the dry method with composite milt samples from males in the same treatment (see Spawning and fertilization). Unfertilized egg samples also were collected to quantify fecundity, egg size, and energetic and lipid content (see Fecundity and egg quality). Fertilized eggs were incubated and hatched under controlled conditions, with hatching success calculated as the number of hatched larvae per number of fertilized eggs per sample (see Hatching success). All hatched larvae were preserved for counts and measurements of larval size

(see Larval size and quality). The date and water temperature at which spawning occurred also recorded and used in the determination of spawning time.

During the experiment, some females suffered mortality (short winter N=19; long winter N=15). The majority (56%) of mortalities occurred within three weeks of the start of the experiment, likely due to transfer stress and failure to acclimate to laboratory conditions. No mortalities occurred once spawning began. Once spawning began, another group of randomly selected female yellow perch in the short (N=14) and long

(N=14) winter-duration treatments were euthanized to assess reproductive development

(results not presented). The remaining females in both the short and long winter-duration treatments all spawned, with some being hand-stripped (short winter N=7; long winter

N=10) and some spawning in tanks (short winter N=9; long winter N=10).

Spawning and fertilization. Once females showed signs of ovulation (i.e., swollen and slightly reddish genital papilla, bulging of the ovary towards the exterior), they were removed from tanks, briefly anesthetized in MS-222, scanned for PIT tags, dried with a cloth, and gentle pressure was applied to the abdomen to strip ovulated egg ribbons.

Eggs were expressed into a dry pan and their mass recorded. From each female stripped of eggs, two subsamples of the egg mass were fertilized with another subsample used to determine fecundity and quantify egg quality. For each fertilization event, two 2-g egg masses were fertilized with a fresh composite of milt sample from three males within the same winter-duration treatment as the stripped female. Milt from the three males was composited into Moore’s extender (Rinchard et al. 2005), where it was diluted 20-fold.

Each 2-g egg mass was subsequently fertilized with a concentration of 100,000 spermatozoa per egg. Additionally, we analyzed both individual and composite milt samples from each fertilization event for percent sperm motility, duration of sperm

29 motility, and sperm density (Rinchard et al. 2005). Results showed that none of these milt quality metrics were related to hatching success (all p > 0.05). Some females released egg ribbons spontaneously in tanks, in between our monitoring activities. We did not attempt to assess fertilization or hatching success of spontaneously released eggs, as in-tank egg ribbon release can result in highly variable male fertilization success rates

(i.e., 40 – 85% Kayes 1977). However, the timing of these spontaneous spawning events was recorded and used in our determination of spawning time.

Hatching success. Fertilized eggs were placed into mesh-covered jars in upwelling California-style tray incubators. The incubators were supplied with water from a partially recirculating system equipped with a chiller to maintain water temperatures at optimal levels for yellow perch egg incubation (13.5-18°C) during the course of the experiment. Temperature and dissolved oxygen were recorded daily and eggs were monitored carefully for the first presence of eyed embryos. When fertilized eggs reached the eyed-embryo stage (8-10 d, depending on water temperature), the samples were moved into clear, plastic 500-mL jars that were filled with water and sealed to quantify hatching success. An air stone in each jar provided vigorous aeration to assist with hatching. Hatching jars also were held in a bath of flow-through water to maintain stable temperatures.

Eggs were checked every 12-24 h for hatching. Once hatched larvae were visible, an 1800-μm sieve was used to separate hatched larvae from un-hatched eggs. All hatched larvae were immediately euthanized and preserved in 3% buffered glutaraldehyde

(Dabrowski and Bardega 1982). After collection of hatched larvae, all un-hatched eggs

30 were returned to the hatching jar, with fresh water. The hatching success of each fertilized egg sample was determined by dividing the total number of hatched larvae collected by the total number of eggs in each sample (determined following methods for fecundity estimation described below).

Fecundity. Fecundity was determined from egg ribbons of females that were hand-stripped and measured as the number of eggs per g of ribbon. Briefly, three subsamples of each egg ribbon (~0.5 g each) were collected and weighed, and the number of eggs in each subsample was counted under a dissecting microscope. The total number of eggs produced by each female was estimated by multiplying the number of eggs per g of ribbon by the overall ribbon mass (Collingsworth and Marschall 2011). While fecundity does not necessarily indicate fertility, which was tested with hatching tests, it can provide an objective measure of reproductive output (Moyle and Cech 2000).

Egg quality. While hatching success can be considered a measure of egg quality, we also quantified other measures, including individual egg mass, energetic density, total lipids, and neutral lipid composition. These additional metrics of egg quality were intended to assist in our investigation of possible mechanisms underlying variation in hatching success and also to determine if other, more easily collected, measures of egg quality could be used in future studies to accurately predict hatching success. The number of eggs per g of ribbon was originally quantified for each female to determine fecundity (see Fecundity above). The inverse of this metric provides the mass of an individual egg for each female, which we used as our proxy of individual egg size. Egg size has been found to be an important predictor of egg quality, as larger eggs generally

31 produce larger offspring that have higher rates of survival (Chambers and Leggett 1996;

Chambers and Waiwood 1996; Einum and Fleming 2000). Total energetic content of egg ribbon subsamples was quantified with bomb calorimetry (Collingsworth and Marschall

2011). Briefly, ovaries were dehydrated in a drying oven (65–70°C; 48-72 h), homogenized into a fine powder, compressed into a small pellet, and 2-3 replicate samples for each ovary were combusted in a Parr® oxygen bomb calorimeter. Total energy density was expressed as calories per g of wet mass. We also quantified the percentages of total lipids, neutral lipids, and phospholipids in each hand-stripped egg mass. Total lipids were extracted from ovaries after homogenization in chloroform- methanol according to (Folch et al. 1957; Czesny and Dabrowski 1998). The organic solvent was evaporated under a stream of nitrogen and the lipid content determined gravimetrically.

Larval size and quality. To determine if larval size metrics and quality differed between winter-duration treatments, we measured four metrics of size and quality in preserved larvae: TL; eye diameter; yolk-sac volume; and body depth at the insertion of the anal fin. Because previous research has found larval traits such as these to be correlated with female spawner size and/or age (Heyer et al. 2001; Berkeley et al. 2004;

Venturelli et al. 2010), as well as incubation temperature (Teletchea et al. 2009; Burt et al. 2011), we included female size, age, condition (residuals of female length-mass relationships), and incubation temperature as covariates in our analyses.

Data analysis. We used generalized linear mixed models (GLMMs; PROC GLM,

SAS v. 9.3) to test whether fecundity, egg quality, egg hatching success, and larval size

32 differed between winter-duration treatments (considered a fixed categorical effect in our models). Prior to using GLMMs, we assessed if response variables were related to female size (i.e., TL), age, and condition (residual of length-mass relationship, calculated separately for both spring [at time of spawning] and fall [when winter-duration treatments began]). Additionally, we assessed if response variables for embryo hatching success and larval size were related to incubation temperature. If significant relationships were found, we used residuals from size-specific relationships as our response variable in our

GLMMs, to remove the effect of female size, age, condition, or incubation temperature.

Using data from our long winter-duration treatment (as our replication was highest for this treatment), we tested for tank effects using tanks as random categorical effects (i.e., replicates nested within winter duration) in a GLMM (PROC MIXED, SAS v.9.3).

Finding no tank effects (all p > 0.05) for any of our response variables, we calculated means of each response variable by tank (our experimental unit) and proceeded with testing for treatment effects. Finally, we used regression techniques (linear and non- linear) to evaluate the relationship between fertilization rates and hatching success with our metrics of egg quality (i.e., individual egg mass, energy density, total and neutral lipids) to determine if a relationship existed between variables. Prior to conducting statistical tests, we verified that each response variable was normally distributed.

Additionally, we analyzed residuals from each model to verify that assumptions of normality, constant variance, independence, and (when appropriate) linearity were met.

Timing of spawning in Lake Erie

We sampled yellow perch weekly in central Lake Erie during spring 2010-2012 to determine the timing of spawning. Individuals were collected near Sandusky, OH (41°

30’ N, 82° 37’ W) by bottom trawling two nearshore-to-offshore transects. Transects were divided into five 1.5-m depth contours (5, 7.5, 9, 10.5, and 12+ m depth contours) with two trawls conducted per depth contour. Bottom water temperature (Figure 2.6) was measured prior to each trawl during 2011-2012 and for each depth contour in 2010. All female yellow perch collected were euthanized, dissected, and classified as either immature, mature (gravid but not spawning), spawning, or spent, using a previously developed criteria based on macroscopic inspection of gonads (Treasurer and Holliday

1981): 1) immature females had a small thread-like transparent ovary; 2) gravid females had clearly visible eggs with the ovary filling about two-thirds of the body cavity; 3) spawning females expressed egg ribbons with gentle pressure; and 4) spent females had empty, flaccid ovaries that were reddish gray. Following these criteria we classified all field-collected female yellow perch in 2010 (N=551), 2011 (N=429), and 2012 (N=279) as immature, gravid, spawning, or spent.

While our study only spanned three years, these years varied greatly in spring temperature. Weekly spring water temperatures measured during 2010-2012 (Figure 2.7) indicate that warming rates were similar among the three years (one-way ANOVA: p =

0.17). However, spring 2010 and 2012 were both significantly warmer than spring 2011

(one-way ANOVA, Tukey’s hsd: 2010 vs. 2011 p = 0.01; 2012 vs. 2011: p = 0.003).

Based on air temperatures, March-May 2012 was the warmest on record for Ohio (as it

34 was for the contiguous U.S.), whereas 2011 and 2010 were ranked as the 92nd and 115th warmest, respectively, out of 119 annual observations (1895-2013; http://www.ncdc.noaa.gov/cag/time-series/us).

Logistic regression was used to relate the number of spawning or spent yellow perch to water temperature recorded for each trawl. Confidence intervals generated were used to determine if the timing of spawning in each year differed in response to annual variation in temperature.

References

Abdulfatah, A., P. Fontaine, P. Kestemont, S. Milla, and M. Marie. 2013. Effects of the thermal threshold and the timing of temperature reduction on the initiation and course of oocyte development in cultured female of Eurasian perch Perca fluviatilis. Aquaculture 376:90-96.

Ahas, R. 1999. Long-term phyto-, ornitho- and ichthyophenological time-series analyses in Estonia. International Journal of Biometeorology 42:119-123.

Berkeley, S. A., C. Chapman, and S. M. Sogard. 2004. Maternal age as a determinant of larval growth and survival in a marine fish, Sebastes melanops. Ecology 85:1258- 1264.

Brown, N. P., R. J. Shields, and N. R. Bromage. 2006. The influence of water temperature on spawning patterns and egg quality in the Atlantic halibut (Hippoglossus hippoglossus L.). Aquaculture 261:993-1002.

Chambers, R. C., and W. C. Leggett. 1996. Maternal influences on variation in egg sizes in temperate marine fishes. American Zoologist 36:180-196.

Chambers, R. C., and K. G. Waiwood. 1996. Maternal and seasonal differences in egg sizes and spawning characteristics of captive Atlantic cod, Gadus morhua. Canadian Journal of Fisheries and Aquatic Sciences 53:1986-2003.

Ciereszko, R. E., K. Dabrowski, A. Ciereszko, J. Ebeling, and J. S. Ottobre. 1997. Effects of temperature and photoperiod on reproduction of female yellow perch Perca flavescens: Plasma concentrations of steroid hormones, spontaneous and induced ovulation, and quality of eggs. Journal of the World Aquaculture Society 28:344- 356.

Clark, R. W., A. Henderson-Arzapalo, and C. V. Sullivan. 2005. Disparate effects of constant and annually-cycling daylength and water temperature on reproductive maturation of striped bass (Morone saxatilis). Aquaculture 249:497-513.

Collingsworth, P. D., and E. A. Marschall. 2011. Identifying relationships between catches of spawning condition yellow perch and environmental variables in the western basin of Lake Erie. Transactions of the American Fisheries Society 140:31-36.

Craig, J. F. 2000. Food and Feeding. Pages 78-105 in J. F. Craig, editor. Percid fishes: systematics, ecology and exploitation. Blackwell Sciences, Oxford.

Cushing, D. H. 1990. Planktonic production and year-class strength in fish populations - an update of the match mismatch hypothesis. Advances in Marine Biology 26:249-293.

Czesny, S., and K. Dabrowski. 1998. The effect of egg fatty acid concentrations on embryo viability in wild and domesticated walleye (Stizostedion vitreum). Aquatic Living Resources 11:371-378.

Dabrowski, K., and R. Bardega. 1982. The changes of fish larvae dimensions due to fixation in different preservatives. Zool. Jb. Anat.:509-516.

Edwards, M., and A. J. Richardson. 2004. Impact of climate change on marine pelagic phenology and trophic mismatch. Nature 430:881-884.

Einum, S., and I. A. Fleming. 2000. Highly fecund mothers sacrifice offspring survival to maximize fitness. Nature 405:565-567.

Eliason, E. J., T. D. Clark, M. J. Hague, L. M. Hanson, Z. S. Gallagher, K. M. Jeffries, M. K. Gale, D. A. Patterson, S. G. Hinch, and A. P. Farrell. 2011. Differences in Thermal Tolerance Among Sockeye Salmon Populations. Science 332:109-112.

Ficke, A. D., C. A. Myrick, and L. J. Hansen. 2007. Potential impacts of global climate change on freshwater fisheries. Reviews in Fish Biology and Fisheries 17:581- 613.

Folch, J., M. Lees, and G. H. Sloane-Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissue. Journal of Biological Chemistry 226:497-509. 37

Garvey, J. E., E. A. Marschall, and R. A. Wright. 1998. From star charts to stoneflies: Detecting relationships in continuous bivariate data. Ecology 79:442-447.

Gopalan, G., D. A. Culver, L. Wu, and B. K. Trauben. 1998. Effects of recent ecosystem changes on the recruitment of young-of-the-year fish in western Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 55:2572-2579.

Hayhoe, K., J. VanDorn, T. Croley, II, N. Schlegal, and D. Wuebbles. 2010. Regional climate change projections for Chicago and the US Great Lakes. Journal of Great Lakes Research 36:7-21.

Henderson, B. A., T. Trivedi, and N. Collins. 2000. Annual cycle of energy allocation to growth and reproduction of yellow perch. Journal of Fish Biology 57:122-133.

Heyer, C. J., T. J. Miller, F. P. Binkowski, E. M. Caldarone, and J. A. Rice. 2001. Maternal effects as a recruitment mechanism in Lake Michigan yellow perch (Perca flavescens). Canadian Journal of Fisheries and Aquatic Sciences 58:1477- 1487.

Hokanson, K. E. F. 1977. Temperature requirements of some percids and adaptations to seasonal temperature cycle. Journal of the Fisheries Research Board of Canada 34:1524-1550.

Houde, E. D. 1987. Fish early life dynamics and recruitment variability. Pages 17-29 in, volume American Fisheries Soceity Symposium.

Ito, L. S., C. Takahashi, M. Yamashita, and C. A. Strussmann. 2008. Warm Water Induces Apoptosis, Gonadal Degeneration, and Germ Cell Loss in Subadult Pejerrey Odontesthes bonariensis (Pisces, Atheriniformes). Physiological and Biochemical Zoology 81:762-774.

Kayes, T. 1977. Reproductive biology and artificial propogation methods for adult perch. University of Wisconsin Sea Grant Program, Madison, WI.

Keniry, M. J., W. A. Brofka, W. H. Horns, and J. E. Marsden. 1996. Effects of decompression and puncturing the gas bladder on survival of tagged yellow perch. North American Journal of Fisheries Management 16:201-206.

Kitchell, J. F., D. J. Stewart, and D. Weininger. 1977. Applications of a bioenergetics model to yellow perch (Perca-flavescens) and walleye (Stizostedion-vitreum- vitreum). Journal of the Fisheries Research Board of Canada 34:1922-1935.

Kocovsky, P. M., T. J. Sullivan, C. T. Knight, and C. A. Stepien. 2013. Genetic and morphometric differences demonstrate fine-scale population substructure of the yellow perch Perca flavescens: need for redefined management units. Journal of Fish Biology 82:2015-2030.

Magnuson, J. J., and B. T. Destasio. 1997. Thermal niche of fishes and global warming. Pages 377-408 in C. M. Wood, and D. G. McDonald, editors. Society for Experimental Biology Seminar Series; Global warming: Implications for freshwater and marine fish, volume 61.

McCormick, M. J., and G. L. Fahnenstiel. 1999. Recent climatic trends in nearshore water temperatures in the St. Lawrence Great Lakes. Limnology and Oceanography 44:530-540.

Miller, T. J., L. B. Crowder, J. A. Rice, and E. A. Marschall. 1988. Larval size and recruitment mechanisms in fishes: Toward a conceptual framework. Canadian Journal of Fisheries and Aquatic Sciences 45:1657-1670.

Moyle, P. B., and J. J. Cech. 2000. Fishes: an introduction to ichtyology, 4th edition. Prentice Hall, Englewood Cliffs, NJ, USA.

Neuheimer, A. B., R. E. Thresher, J. M. Lyle, and J. M. Semmens. 2011. Tolerance limit for fish growth exceeded by warming waters. Nature Climate Change 1:110-113.

Ottera, H., A. L. Agnalt, A. Thorsen, O. S. Kjesbu, G. Dahle, and K. Jorstad. 2012. Is spawning time of marine fish imprinted in the genes? A two-generation experiment on local Atlantic cod (Gadus morhua L.) populations from different geographical regions. ICES Journal of Marine Science 69:1722-1728.

Ottersen, G., S. Kim, G. Huse, J. J. Polovina, and N. C. Stenseth. 2010. Major pathways by which climate may force marine fish populations. Journal of Marine Systems 79:343-360.

Pankhurst, N. W., and P. L. Munday. 2011. Effects of climate change on fish reproduction and early life history stages. Marine and Freshwater Research 62:1015-1026.

Perry, A. L., P. J. Low, J. R. Ellis, and J. D. Reynolds. 2005. Climate change and distribution shifts in marine fishes. Science 308:1912-1915.

Portner, H. O., and A. P. Farrell. 2008. Physiology and Climate Change. Science 322:690-692.

Portner, H. O., and R. Knust. 2007. Climate change affects marine fishes through the oxygen limitation of thermal tolerance. Science 315:95-97.

Richardson, A. J. 2008. In hot water: zooplankton and climate change. ICES Journal of Marine Science 65:279-295.

Rinchard, J., K. Dabrowski, J. J. Van Tassell, and R. A. Stein. 2005. Optimization of fertilization success in Sander vitreus is influenced by the sperm : egg ratio and ova storage. Journal of Fish Biology 67:1157-1161.

Sandstrom, O., I. Ambrahamsson, J. Andersson, and M. Vetemaa. 1997. Temperature effects on spawning and egg development in Eurasian perch. Journal of Fish Biology 51:1015-1024.

Sepulveda-Villet, O. J., and C. A. Stepien. 2011. Fine-scale population genetic structure of the yellow perch Perca flavescens in Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 68:1435-1453.

Sharma, S., D. A. Jackson, C. K. Minns, and B. J. Shuter. 2007. Will northern fish populations be in hot water because of climate change? Global Change Biology 13:2052-2064.

Teletchea, F., J. N. Gardeur, E. Kamler, and P. Fontaine. 2009. The relationship of oocyte diameter and incubation temperature to incubation time in temperate freshwater fish species. Journal of Fish Biology 74:652-668.

Treasurer, J. W., and F. G. T. Holliday. 1981. Some aspects of the reproductive biology of perch Perca fluviatilis L. - A histological description of the reproductive cycle. Journal of Fish Biology 18:359-376.

Tyson, J. T., T. B. Johnson, C. T. Knight, and M. T. Bur. 2006. Intercalibration of research survey vessels on Lake Erie. North American Journal of Fisheries Management 26:559-570.

Venturelli, P. A., C. A. Murphy, B. J. Shuter, T. A. Johnston, P. J. V. de Groot, P. T. Boag, J. M. Casselman, R. Montgomerie, M. D. Wiegand, and W. C. Leggett. 2010. Maternal influences on population dynamics: evidence from an exploited freshwater fish. Ecology 91:2003-2012.

Wang, J., X. Bai, H. Hu, A. Clites, M. Colton, and B. Lofgren. 2012. Temporal and spatial variability of Great Lakes ice cover, 1973-2010. Journal of Climate 25:1318-1329.

Wang, N., F. Teletchea, P. Kestemont, S. Milla, and P. Fontaine. 2010. Photothermal control of the reproductive cycle in temperate fishes. Reviews in Aquaculture 2:209-222.

Webb, M. A. H., J. P. Van Eenennaam, G. W. Feist, J. Linares-Casenave, M. S. Fitzpatrick, C. B. Schreck, and S. I. Doroshov. 2001. Effects of thermal regime on ovarian maturation and plasma sex steroids in farmed white sturgeon, Acipenser transmontanus. Aquaculture 201:137-151.

Wedekind, C., and C. Kueng. 2010. Shift of spawning season and effects of climate warming on developmental stages of a grayling (Salmonidae). Conservation Biology 24:1418-1423.

Winder, M., and D. E. Schindler. 2004. Climatic effects on the phenology of lake processes. Global Change Biology 10:1844-1856.

Wootton, R. J. 1998. Reproduction. Ecology of Teleost Fishes, 2nd edition. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Wright, P. J., and E. A. Trippel. 2009. Fishery-induced demographic changes in the timing of spawning: consequences for reproductive success. Fish and Fisheries 10:283-304.

Yellow Perch Task Group (YPTG). 2013. Report of the Lake Erie yellow perch task group. Great Lakes Fishery Commission, Ann Arbor, MI.

Response Variable Spring Female Fall Female Incubation Female Size Female Age Condition Condition Temperature Fecundity <.0001 0.99 0.29 0.28 -

Egg quality Individual egg mass (mg) 0.41 0.09 0.28 0.76 - Calories per egg 0.19 0.23 0.38 0.71 - Total lipids (µg) per egg 0.47 0.09 0.27 0.99 - Total neutral lipids (µg) per egg 0.50 0.08 0.21 0.92 -

Hatching success 0.69 0.13 0.22 0.65 0.51

Larval size metrics Larval total length (mm) 0.26 0.30 0.45 0.38 0.92 Larval body depth (mm) 0.66 0.48 0.76 0.54 0.89 Larval yolk sac volume (mm3) 0.83 0.72 0.40 0.91 0.71

Table 2.1. Associated p-values for Pearson correlation coefficients testing the response of fecundity (number of eggs per g of ribbon), egg quality, hatching success (%), and larval size to indices of female yellow perch size (nearest 1 mm total length), age, and condition (both spring and fall separately), and incubation temperature (ºC). An index of female condition was created from residuals of length-mass relationships developed for all individuals measured in the spring (time of spawning) and fall (initiation of winter experiment). Bold values indicate p < 0.05. These results show that all of our metrics of egg and larval quality were unrelated to individual female attributes or incubation temperatures.

Figure 2.1. Yellow perch juvenile (age-0) abundance, plotted against mean ice cover during the winter (February-March; % areal coverage) prior to hatching in western and central Lake Erie (1973-2010). Vertical dashed lines (and associated p-values) indicate threshold values of mean winter ice cover as indicated by two-dimensional Kolmogorov-Smirnov tests (Garvey et al. 1998). Above the threshold, juvenile abundance (a strong predictor of recruitment to the fishery at age-2 and future harvest [Figures 2.2, 2.3]) may be high or low, but below the threshold only low abundances occur. Thus, winter duration may ‘set the stage’ for future high recruitment to the fishery. Juvenile abundance was determined from annual, fisheries-independent Ohio Division of Wildlife bottom trawling surveys, is presented as catch-per-unit-effort (CPUE).

Figure 2.2. Yellow perch cohort size at age-2 in year t + 2 plotted against juvenile abundance in year t (year in which each cohort was hatched) for western and central Lake Erie from 1987-2010. Estimates of cohort size at age-2 for western and central Lake Erie are from the Great Lakes Fishery Commission Lake Erie Yellow Perch Task Group (YPTG 2013). Results indicate juvenile (age-0) abundance is a good proxy for future recruitment to age-2, when yellow perch typically become reproductively mature and enter the fishery.

Figure 2.3. Lifetime harvest by cohort (cumulative harvest ages 2– 5) plotted against juvenile abundance in year t (year in which each cohort was hatched) for western and central Lake Erie from 1987-2010 (YPTG 2013). These results show that a strong annual cohort, which is only possible after long, cold winters, can support the fishery for many years afterwards.

Figure 2.4. Relationships between: A) winter duration and individual yellow perch egg mass (mg); B) hatching success (%) and individual egg mass (mg); and C) larval total length (mm) and individual egg mass (mg). All data were collected during a controlled laboratory experiment and are presented as tank means (± 1 SE). Females exposed to a long winter produced large eggs that hatched at higher rates and produced larger larvae than small eggs.

Figure 2.5. Mean daily water temperature (°C) from short (50 d < 5°C) and long (110 d < 5°C) winter-duration treatments used in laboratory experiments with Lake Erie yellow perch. The duration of spawning is shown for each treatment. Labels indicate the dates and water temperatures during which females spawned. Despite earlier arrival of seeming suitable spawning temperatures (8-11°C) following a short winter, females did not spawn during these temperatures, but instead delayed spawning until the typical spawning period (i.e., April-May [Hokanson 1977]).

Figure 2.6. Mean (± 1 SE) bottom water temperature measured in western Lake Erie near Sandusky, Ohio during the yellow perch spawning season, 2010-2012.

Figure 2.7. Probability that a yellow perch female was spent (i.e., had completed spawning) given bottom water temperature (°C) during spring 2010-2012 spawning seasons in western Lake Erie. Colored lines determined by logistic regression; dashed lines indicate 95% confidence intervals. Solid black lines indicate probabilities determined for 8, 10, and 12°C each year. Lines with different letters differed among years (p < 0.05). Despite earlier arrival of seemingly suitable spawning temperatures (8- 11°C) following a short winter in 2012, females did not spawn earlier, but instead delayed spawning until the typical spawning period (April-May [Hokanson 1977]).

CHAPTER 3

WINTER WARMING AND REDUCED ENERGY RESERVES

AFFECT REPRODUCTION IN A COOL-WATER, TEMPERATE FISH SPECIES

Introduction

Modeling studies have suggested that climate change will positively affect cool- water, temperate fishes by lengthening the growing season and enhancing the availability of favorable thermal habitat for growth (Shuter and Meisner 1992; Stefan et al. 1996;

Magnuson et al. 1997; Mackenzie-Grieve and Post 2006; Sharma et al. 2007; Cline et al.

2013). However, because climate change also has the potential to alter precipitation patterns (Kunkel et al. 2002; Kling et al. 2003), nutrient dynamics (Kerfoot et al. 2008), water levels (Hartmann 1990; Mortsch and Quinn 1996; Kling et al. 2003; Angel and

Kunkel 2010; MacKay and Seglenieks 2013), nursery and spawning habitat (Mortsch

1998; Jones et al. 2006), as well as the timing, magnitude, and duration of thermal stratification (McCormick and Fahnenstiel 1999; Austin and Colman 2007), hypoxia (Fang et al. 2004; Hawley et al. 2006), and ice cover (Assel et al. 2003; Magnuson 2010; Wang et al. 2012), making predictions based on thermal habitat conditions alone is insufficient. In fact, one should expect climate change to have interactive effects on fish populations and communities (Brook et al. 2008; Doney et al. 2012), which will increase the likelihood of

50 climate-driven ecological “surprises” (Paine et al. 1998; Doak et al. 2008). This likelihood only increases when we consider that effects of climate change may interact with other anthropogenic stressors (e.g., effects of invasive species, eutrophication, overfishing, habitat loss) to influence species composition, distribution, and production

(Vitousek 1994; Jackson et al. 2001; Ficke et al. 2007; Hellmann et al. 2008).

Alteration of demographic rates (e.g., birth, death) is one way that climate change may influence fish populations (Kerr et al. 2009; Ottersen et al. 2010). Given that the maximal number of potential future survivors to a population in a year is set by the number of eggs produced that year (Houde 1987), those aspects of climate change that affect reproductive output (i.e., potential recruit production) may have a disproportionately large impact on population dynamics and viability. Indeed, previous studies have investigated how climate change might influence fish populations via effects on pre-recruit (e.g., eggs, larvae) survival (Svendsen et al. 1995; Casselman 2002; Kӧster et al. 2005; Vikebø et al. 2005). Few studies, however, have explored the link between climate change and fish reproductive biology and physiology in wild fish populations

(note the lack of studies reviewed in Drinkwater et al. 2010), which we view as a major gap in both the freshwater and marine climate change literature for two key reasons.

First, temperature has been shown to be a key determinant of reproductive processes

(e.g., vitellogenesis, ovulation) in temperate fishes (Dabrowski et al. 1996; Wang et al.

2010). Experimental and field-observational research has demonstrated that fecundity and spawning success of temperate fish species can decline when adult females do not experience a prolonged period of cold temperatures during the winter preceding the

51 spring spawning season (Jones et al. 1972; Hokanson 1977; Ciereszko et al. 1997;

Sandstrom et al. 1997; Luksiene et al. 2000; Webb et al. 2001; Migaud et al. 2002; Clark et al. 2005; Brown et al. 2006; Wang et al. 2006). We also have documented a positive threshold relationship between a proxy for winter severity (ice cover) and an annual index of juvenile (age-0) yellow perch abundance in Lake Erie’s western and central basin populations, which reveals that short, warm winters were always followed by weak recruitment, whereas strong recruitment events were possible only after long, cold winters (Chapter 2). Second, seasonal (mid-summer through early fall) hypolimnetic hypoxia (dissolved oxygen < 2 mg/L), which has re-emerged as a water quality impairment in Lake Erie during recent decades (Hawley et al. 2006; Rucinski et al. 2010) and is predicted to increase with continued climate warming (Kling et al. 2003; Fang et al. 2004; Ficke et al. 2007), may exacerbate the negative effects of warm winters on reproductive development by reducing female energetic condition prior to winter.

Indeed, poor energetic condition has been shown to reduce egg production (Bunnell et al.

2005; Kennedy et al. 2008; Bunnell et al. 2009; Gregersen et al. 2011) and increase the incidence of skipped spawning (Scott and Crossman 1974; Rideout et al. 2005;

Skjæraasen et al. 2012) in several temperate fishes. Further, field and modeling data from central Lake Erie have shown that hypolimnetic hypoxia can reduce food consumption, energetic condition, and growth potential of several fishes in Lake Erie, including yellow perch (Roberts et al. 2009; Arend et al. 2011; Scavia et al. in revision).

Despite much evidence to support our hypothesis that climate change will negatively affect temperate fishes that develop ovaries during winter by reducing 1)

52 exposure of potential spawners to the thermal requirements necessary for proper ovary development during winter and 2) the energetic condition of potential spawners prior to winter, the biological mechanisms underlying relationships between winter temperature, female condition, and reproductive success remain largely unexplored. Certainly, such information would be critical to efforts to forecast the impacts of continued climate change on temperate fish populations (Chapter 2). Towards this end, we conducted a controlled laboratory experiment to quantify the independent and interactive effects of reduced body condition entering winter and winter duration on Lake Erie yellow perch reproductive success. We used Lake Erie yellow perch as our study organism because 1) it is a cool-water, temperate species that develops ovaries during winter and spawns during spring, a common reproductive strategy among temperate fishes (Wang et al.

2010), and 2) it, and its congener (Eurasian perch Perca fluviatilis), are ecologically, economically, and culturally important across much of North America and Europe.

Methods

Overview

We conducted a controlled laboratory experiment with yellow perch during summer 2011 through spring 2012 to quantify the independent and interactive effects of winter duration and adult body condition on ovarian development, male and female gamete production and quality, female spawning success, egg hatching success, and the quality of eggs and larvae. For our experiment, we used wild-caught Lake Erie yellow perch. Most previous studies that have explored the impacts of temperature on

53 reproductive success used domesticated fish (e.g., hatchery broodstock), which would be expected to elicit an unnatural response to environmental stimuli, owing to selection over multiple generations for traits favorable to commercial production (Heath et al. 2003).

Details regarding each aspect of our experiment are given below. The experiment and the field collections were conducted according to animal use guidelines outlined in

IACUC protocol # 2009A0073 at The Ohio State University (OSU).

Fish collection

Live male and female yellow perch were collected via bottom trawling off of the

R/V Grandon in central Lake Erie (near Fairport Harbor, OH) during April-May 2011, a period when sex could be easily determined by visual examination. Because over- inflation of the gas bladder is a common occurrence in fish caught in deep water and brought to the surface, which, if not relieved, can result in mortality due to gas embolisms or internal organ damage (Keniry et al. 1996), all captured individuals were placed into a live well and ‘fizzed’ with a hypodermic needle immediately upon collection to release air pressure from the gas bladder. This ‘fizzing’ process has been shown to reduce post- capture mortality in yellow perch, with no adverse effects on long-term survival (Keniry et al. 1996).

Live yellow perch were transported from Lake Erie to an outdoor pool facility at the OSU’s Aquatic Ecology Laboratory (Columbus, OH), where they could experience natural, seasonal variation in temperature and photoperiod. All individuals were held in

2,500-L circular tanks that were supplied with constant aeration and flow-through water

54 and fed live (or occasionally frozen) fathead minnows (Pimephales promelas). All individuals were PIT-tagged to allow us to identify individuals for estimation of growth and maturation.

Experimental design and procedures

Body-condition treatments. To create fish of varying energetic condition prior to winter, so as to test the importance of adult condition on reproductive success, we exposed fish to 1 of 2 feeding regimes during late summer through early fall. During the third week of July 2011, all individuals were briefly anesthetized in MS-222 (to reduce handling stress), measured (nearest 1 mm), weighed (nearest 0.1 g), scanned for PIT tags, and randomly assigned to 1 of 2 feeding regimes (maintenance ration or ad libitum). The low feeding-level (maintenance ration) regime was intended to simulate the effect of hypolimnetic hypoxia on yellow perch condition entering the winter (i.e., create individuals of low energetic condition), whereas the unlimited feeding-level (ad libitum ration) regime was intended to simulate to effect of a hypoxia-free environment (i.e., create individuals of high energetic condition), in which yellow perch foraging was not limited by the presences of hypoxia (per Roberts et al. 2009, 2011, 2012). Daily maintenance ration (the amount of food that met all basal metabolic demands without allowing for growth or increased fat reserves) for our simulated hypoxia (low feeding- level ) treatment was calculated using a previously published bioenergetics model for yellow perch (Kitchell et al. 1977). Model inputs included water temperature (recorded daily), individual body mass (recorded during the third week of July), and the energy

55 density of yellow perch and their feed (fathead minnows; TMF, unpublished data).

During the first week of October 2011, at which time we ended differential feeding regimes, we randomly selected and euthanized both male (N= 47) and female (N = 9) yellow perch from each feeding regime to quantify individual condition (i.e., energy density, mesenteric fat mass; see protocol details below in Sample processing and analysis) and reproductive development (i.e., gonadosomatic index) entering winter.

Our feeding regimes successfully created two distinct levels of body condition prior to winter. To verify this, we created total length (TL) - mass relationships for females (log10[TL] = 3.20 x log10 [mass] - 5.40; P < 0.0001; R2 = 0.94) and males

(log10[TL] = 3.11 x log10[mass] - 5.17; P < 0.0001; R2 = 0.94). Using the residual value for each individual as a relative index of body condition (i.e., positive residual indicates a higher than average mass for a given length), we determined that residual condition was higher in the ad libitum (high ration) treatment than in the maintenance (low) ration treatment for both females (ANOVA F1,96 = 26.0 , P < 0.0001) and males (ANOVA F1,53

= 4.24 , P = 0.045; Figure 3.1). Growth in mass from the start (late July) to the end of summer (early October) also was greater in the high (ad libitum) versus low

(maintenance) ration treatments for both females (ANOVA F1,97 = 85.4; P < 0.0001) and males (ANOVA F1,53 = 31.5; P < 0.0001; Figure 3.2), with fish TL having no effect on female (P = 0.54) or male (P = 0.16) growth in mass.

Winter-duration treatments. During October 2011 through June 2012, we simulated two winters of varying duration: 1) a long, cold winter, with 110 d less than <

5°C; and 2) a short, warm winter, with only 50 d less than 5°C. The duration of winter

56 and rates of fall cooling and spring warming (0.2°C/d) that were used in our experiment were based on historical (1990-2010) field measurements of water temperature collected from Lake Erie at a Cleveland, OH water intake (TMF, unpublished data). Our two winter-duration treatment levels were intended to simulate historical (110 d) and future

(50 d) conditions for Lake Erie. Both levels of modeled winter duration were observed in our historical range of winter durations (1990-2010), the shortest (50 d) being recorded during the warm El Niño winter of 2002.

During the first week of October 2011, prior to placement in winter-duration treatments, all individuals were briefly anesthetized in MS-222, measured, weighed, and scanned for PIT tags. Subsequently, individuals were randomly assigned to a winter- duration treatment and randomly placed into tanks inside walk-in, temperature-control rooms. Each room represented a single winter-duration treatment and contained a recirculating system with six 189 L tanks (3 tanks/body-condition treatment). Each individual tank contained a similar fish biomass that consisted of 12-15 yellow perch (8-9 females and 3-7 males). In total, 49 females and 28 males were initially used in the long winter-duration treatment, whereas 49 females and 26 males were initially used in the short winter-duration treatment. Unfortunately, some individuals died (females: short winter N=19, long winter N=15; males: short winter N=10, long winter N=5) during the experiment, leaving 34 females and 23 males to complete the long-winter treatment and

30 females and 16 males to complete the short-winter treatment. The majority (61%) of mortalities occurred within three weeks of the start of the experiment, likely due to failure to acclimate to laboratory conditions. No mortalities occurred once spawning

57 began. Once spawning began, another group of randomly selected female yellow perch in the short (N=14) and long (N=14) winter treatments were euthanized to assess reproductive development (results not presented).

Experimental conditions. Lighting in all rooms was provided by incandescent lights and controlled by a digital system that simulated daily reductions (during the fall) and increases (during winter and spring) in photoperiod so as to mimic the photoperiod at

Cleveland, OH. Additionally, at dawn and dusk, 1 hour of increasing and decreasing light intensity preceded day and night periods. Recirculating systems were supplied with continuous aeration along with physical and biological filtration to maintain water quality. Ammonia, nitrite, nitrate, and pH were measured daily and water changes were conducted as needed to maintain high levels of water quality (i.e., unionized ammonia <

0.1 ppm; nitrates < 40 ppm). Hobo® loggers (Onset®) recorded water temperature every

2 hours in each recirculating system, and temperature and dissolved oxygen were measured daily with a YSI 550a handheld meter. During the entire experiment, all individuals were fed daily maintenance rations of live fathead minnows, determined from an existing bioenergetics model (Kitchell et al. 1977).

Spawning, fertilization, and euthanization. At the end of each simulated winter, water temperature was slowly increased to simulate spring warming. Females were monitored closely (checked every 30 min from first light until 3 hours after last light) for external signs of ovulation. Once females showed signs of ovulation, they were removed from tanks, briefly anesthetized in MS-222, scanned for PIT tags, dried with a cloth, and gentle pressure was applied to the abdomen to strip ovulated egg ribbons. Eggs were

58 expressed into a dry pan and their mass recorded. Subsamples of these stripped eggs were used to determine reproductive development, to quantify fecundity, and to assess egg quality (see Sample processing and analysis for details). Additionally, four 2-g egg masses were fertilized with fresh milt that was collected for each fertilization event from the three males within the same winter-duration/body-condition treatment as the stripped female. As with females, males were briefly anesthetized in MS-222, scanned for PIT tags, dried with a cloth, and gentle pressure was applied to the abdomen to express milt.

Milt from three males was composited into Moore’s extender, where it was diluted 20- fold (Rinchard et al. 2005). Each 2-g egg mass was subsequently fertilized with a concentration of 100,000 spermatozoa per egg (Kwasek 2012). We also analyzed both individual and composite milt samples from each fertilization event for percent sperm motility, duration of sperm motility, and sperm density (following Rinchard et al. 2005).

Sample processing and analysis

Fertilization and hatching success. Fertilized eggs were transported in fresh water to incubating and hatching facilities at OSU’s Aquaculture Laboratory (Columbus, OH), where they were immediately placed into mesh-covered jars in upwelling California-style tray incubators. California trays were supplied with water from a recirculating system that was equipped with a chiller to maintain water temperatures at levels typically observed in the wild for yellow perch egg incubation (i.e., 12-18°C; Hokanson 1977).

Temperature and dissolved oxygen were recorded daily and eggs were watched carefully for presence of eyed-embryos. When fertilized eggs reached the eyed-embryo stage (8-

10 d, depending on water temperature), 2 of the 4 fertilized egg samples were removed from incubators and viewed under a dissecting microscope to quantify fertilization success. To do so, all eyed-eggs and dead eggs were counted in a single field of vision at

10x magnification (Dabrowski et al. 1996). This process was repeated up to 10 times for each egg mass.

The remaining two fertilized egg samples were moved from California trays into hatching jars for hatching success determination. Clear, plastic 500-mL hatching jars were filled with water and sealed with an air stone inside to provide vigorous aeration that assisted with hatching (as suggested in Hart et al. [2006]). Hatching jars were held in a bath of flow-through water to maintain stable temperatures (13-20°C, depending on when hatching occurred during the spring). Eggs were checked every 12-24 hours for hatching. Once hatched larvae were visible, an 1800 µm sieve was used to separate hatched larvae from un-hatched eggs. All hatched larvae were immediately euthanized and preserved in 3% buffered glutaraldehyde (Dabrowski and Bardega 1982). After collection of hatched larvae, all un-hatched eggs were returned to the hatching jar, with fresh water. The hatching success of each fertilized egg sample was determined by dividing the total number of hatched larvae collected by the total number of eggs in each sample (see Fecundity section below).

Reproductive development. Ovarian development was measured by quantifying the percent of females in each treatment that ovulated during the experiment, with the gonadosomatic index (gonad mass x 100 / body mass; GSI) determined by weighing the mass (nearest 0.1 g) of each egg ribbon spawned by each individual. A female was

60 considered to be ovulating if it was able to be hand-stripped of egg or it spontaneously released eggs in its tank.

Fecundity. While fecundity does not necessarily indicate fertility, which was tested with fertilization and hatching tests, it can provide an objective measure of reproductive output (Moyle and Cech 2000). Thus, we determined fecundity from egg ribbons of females that were hand-stripped. To do so, a subsample (0.2-0.25 g) of each egg ribbon was collected and weighed, and the number of eggs in each subsample was counted under a dissecting microscope. With this information, we could estimate the total number of eggs produced by each female.

Egg quality. Egg size has been found to be a useful proxy of egg quality, as larger eggs generally produce larger offspring that have higher rates of survival (see reviews by Chambers and Leggett 1996 and Chambers and Waiwood 1996). Further, while the effects of maternal traits, such as size and age, on egg size have been previously investigated (Heyer et al. 2001; Johnston et al. 2007; Venturelli et al. 2010), the effects of environmental variables such as winter duration, in combination with female condition, have not. Thus, in addition to using fertilization and hatching success to assess egg quality, we measured egg size (both egg diameter and individual egg mass) and egg energetic density. Quantifying these metrics also provided a way to help explain variation in fertilization and hatching success and if these more easily collected measures of egg quality could be used as a proxy of fertilization and hatching success. To estimate egg mass for each female, we used the inverse of our fecundity estimate (number of eggs per g of ovaries; see above). To quantify egg diameter, a subsample (~0.2 g) of

61 unfertilized eggs was collected, fixed for 12-24 hours in Davidson’s solution, and then preserved in 70% ethyl alcohol. On a later date, these eggs were digitally photographed using a dissecting microscope equipped with a Nikon® DS Fi2 digital camera. The diameter of 30 randomly selected eggs per sample was measured using Nikon® NIS

Elements v. 4.0 to quantify individual egg size for each female.

Total energetic content of ovary subsamples was quantified with bomb calorimetry (following Collingsworth and Marschall 2011). Briefly, ovaries were dehydrated in a drying oven (65–70°C; 48-72 hours), homogenized into a fine powder, compressed into a small pellet, and 2-3 replicate samples from each ovary were combusted in a Parr® oxygen bomb calorimeter. Total energy density was expressed as calories per g of wet mass.

Data analysis

We used generalized linear mixed models (GLMMs; PROC GLM, SAS v. 9.3) to test for effects of winter duration and female condition on each response described above.

Winter-duration was considered a fixed categorical effect in all GLMMs. By contrast, body-condition was considered fixed categorical effect in an initial set of GLMMs, whereas it was treated as a continuous covariate in a second set, allowing us to assess if individual measures improved our ability to explain variation in our response variables.

Prior to building any GLMM, we assessed if the response variable was related to female size (i.e., TL and mass). If a significant relationship was found, we used residuals from this size-specific relationship as our response variable in our GLMMs, to remove the

62 confounding effect of female size. Additionally, using data from our long winter- duration treatment (as our replication was highest for this treatment), we tested for tank effects using tank as a random categorical variable (i.e., replicates nested within winter- duration and body-condition treatment combinations) in a GLMM (PROC MIXED, SAS v. 9.3). Finding no tank effects (P > 0.05) for any of our variables, we calculated means of each response variable for each tank (our experimental unit) and proceeded with testing for treatment effects. We also tested for an interaction between fixed categorical effects (i.e., winter duration*body condition). If the interaction term was not significant, it was removed from the model. In this manner, we tested for the effect of winter temperature by accounting for the effect of body condition on each of our response variables. Finally, we used regression techniques to evaluate the relationship between fertilization rates and hatching success with our metrics of egg quality (i.e., egg mass, diameter, energetic density) to determine if a relationship existed between variables.

Prior to conducting statistical tests, we verified that each response variable was normally distributed. Additionally, we analyzed residuals from each model to verify that assumptions of normality, constant variance, independence, and (when appropriate) linearity were met. As needed, response variables were transformed to meet assumptions.

For example, proportion data were arcsine-square root transformed prior to analysis. In some cases, when transformations failed to meet assumptions, non-parametric statistics were used instead of generalized linear models. The α-level for all statistical tests was set at 0.05.

Results

Pre-experiment reproductive development

The initiation of reproductive development did not differ between body-condition treatments. In males and females that were sacrificed at the end of summer (first week of

October), gonad mass increased with increasing fish TL (males: P < 0.0001; R2 = 0.75;

Females: P=0.03. R2 = 0.49). The relationship between fish TL and gonad mass did not differ between feeding regimes (females: ANCOVA F2,6 = 2.11; P = 0.2; males:

ANCOVA F2,44 = 3.03; P = 0.09; Figure 3.3). Although the largest females in the ad libitum (high) feeding regime had greater gonad mass than the largest females in the maintenance ration (low) feeding treatment (Figure 3.3), we were not able to detect significant differences in gonad mass between these groups, possibly due to low sample size.

Spawning, fertilization, and embryo hatching success

We successfully simulated winters of different duration and created conditions that allowed for apparently normal spawning. Observed winter duration in both treatments matched target durations for number of days with temperatures < 5°C (i.e., short = 50 d, long = 110 d < 5°C; see Supplemental Materials and Methods in Chapter 2).

Likewise, regardless of treatment, all females that were allowed to progress towards spawning eventually spawned, with viable eggs found in all treatments. However, important differences in the timing of spawning, quality of eggs produced, and hatching success of eggs existed among treatments.

Spawning phenology. Both the timing of and temperatures at which spawning occurred differed between winter-duration treatments, with timing of spawning being defined as when the female showed signs of ovulation and was hand-stripped of eggs or when eggs were spontaneously released in a tank. Females from the short-winter treatment spawned at earlier dates but at higher temperatures than females in the long- winter treatments (Chapter 2). No obvious differences in the timing of spawning between females of high versus low body condition were found within each winter duration treatment (TFM, unpublished data).

Fertilization and embryo hatching success. Of the females that entered the final stages of ovulation, ~50% were able to be hand-stripped of eggs, which were fertilized under controlled conditions (short N=7, long N=10). Fertilization and embryo hatching success of these eggs was highly variable (short mean=46%, range=6-80%; long mean=82%, range=43-98%). Furthermore, by regressing fertilization success against hatching success, we found that our index of fertilization success largely over-predicted embryo hatching success (Figure 3.4). Therefore, in subsequent analyses that relate egg quality to egg survival, we used embryo hatching success, as we believe this metric is a truer measure of egg viability.

As categorical variables, neither winter duration (P = 0.25) nor female condition

(P = 0.50) treatments affected hatching success. However, a model that had female condition as a continuous variable (i.e., the residual value from pre-winter length-mass relationship) along with winter duration in it explained a great deal of variance in embryo

2 hatching success (ANOVA F2,7 = 4.59; P = 0.07; R = 0.65), although the overall model

65 was not statistically significant. In this model, winter duration (P = 0.03) had a positive effect on embryo hatching success, with higher success in the long winter compared to the short winter (Figure 3.5). Female condition had a weak negative effect on embryo hatching success (P = 0.055; Figure 3.5). Within both winter-duration treatments, embryos from females in poor fall condition hatched at higher rates than those from females in relatively better fall condition.

Fecundity and egg quality

Fecundity and GSI. Fecundity increased with TL (fecundity = 451 x TL [mm] –

82128; P < 0.0001; R2 = 0.65; see Figure 3.6). As categorical variables, neither winter duration (P = 0.55) nor female condition (P = 0.29) affected fecundity. However, a model with winter duration and a continuous measure of female condition (i.e., the residual value from the pre-winter TL-mass relationship) explained a high degree of

2 variance in fecundity (ANOVA F2,5 = 3.12; P = 0.13; R = 0.55), although the overall model was not statistically significant. In this model, winter duration (P = 0.08) had a weak (albeit non-significant) negative effect on fecundity, with somewhat higher egg production in the short winter compared to the long winter (Figure 3.7). Within both winter-duration treatments, females in poorer fall condition produced slightly fewer eggs than those in relatively better fall condition (P = 0.065; Figure 3.7). While fecundity was weakly affected by winter duration and female condition, winter duration had a strong

2 positive effect on GSI (ANOVA F1,6 = 14.52; P = 0.009; R = 0.71; Figure 3.8), with females in the long winter-duration treatment producing larger egg ribbons than those in

66 the shorter winter-duration treatment. We did not detect an effect of female condition, treated as either a classification (P = 0.08) or continuous (P = 0.17) variable, on GSI.

Egg size and energy density. The increase in egg ribbon size was accompanied by an increase in the size of individual eggs. Females in the long winter-duration treatment produced larger eggs, in terms of both mass (Chapter 2) and diameter (ANOVA

2 F1,6 = 7.9 , P = 0.03; R = 0.57, Figure 3.9), than those in the short winter-duration treatment.

Energy density of egg ribbons declined with the longevity of winter and was lower in the low body-condition treatment than the high body-condition treatment in both

2 the short and long winter-duration treatments (ANOVA F2,5 = 27.6, P = 0.002; R =

0.92). While the energy density of egg ribbons was lower in the long versus short winter- duration treatment, the mass of individual eggs showed an opposite effect (see Chapter

2). Because the increase in egg mass (~40%) was about double the decline in caloric density (~20%), the energy density per egg did not differ between winter-duration treatments. Subsequently, we scaled egg energy data to the individual egg level (μg/egg wet mass), to account for changes in egg mass between winter-duration treatments.

Relating metrics of egg quality against hatching success, we found that egg mass was positively correlated with hatching success for (Figure 3.10). We also found that egg energy was positively correlated with hatching success (Figure 3.10).

Male reproductive quality

While some measures of male reproductive development and quality differed between winter-duration treatments, others appeared unaffected. Specifically, sperm motility (the percentage of milt that was activated by water contact) declined at the end of the long winter treatment (Figure 3.12). By contrast, other measures of sperm quality

(i.e., duration of motility) and production (i.e., sperm density) did not differ between treatments throughout the experiment. To assess if repeated stripping of milt from males lowered sperm quality or production, we looked at how mean motility, the mean duration of motility, and mean sperm density changed through time in individuals for which we had at least three measures of sperm quality. Finding no evidence that repeated stripping of males influenced any of these metrics (TMF, unpublished data), we averaged measures of sperm quality and production for each individual male to investigate treatment effects.

The duration of motility did not differ between winter-duration or body-condition treatments and was similar for all males (35 ± 6 sec [mean ± 1 SD]). Similarly, sperm density (after correcting for an overall negative relationship with male size: sperm density

= -0.11 x mass [g] + 68.7, P = 0.016, R2 = 0.19) appeared unaffected by winter-duration

(P > 0.1) or body-condition treatments (P > 0.1), and was similar between all males (57.8

± 7.9 % spermatocrit [mean ± 1 SD]). By contrast, sperm motility was lower in the long winter-duration treatment than in the short-winter treatment (Kruskal Wallis χ-squared =

6.33; d.f. = 1; P = 0.012), owing to a decline during the tail end of spawning in our long winter-duration treatment (Figure 3.11). This decline in sperm motility, however, was unrelated to hatching success, and we found no correlation between hatching success and

68 any other measure of sperm quality or production. Additionally, including sperm motility as a covariate in the previously defined relationship between hatching success and egg mass did not explain any additional variance (P = 0.31).

Discussion

Our experimental results, combined with previously published findings from this experiment (Chapter 2), indicate a novel mechanism by which climate change may influence yellow perch reproductive success and, subsequently, recruitment in the

Laurentian Great Lakes. Lake Erie yellow perch exposed to a long winter produced higher quality (larger eggs with more calories per egg) than those exposed to a short winter. These higher quality eggs produced from females exposed to a long winter duration hatched at higher rates and produced larger larvae (Chapter 2) than Lake Erie females exposed to a short winter.

These results indicate a surprising way in which winter warming may negatively affect yellow perch recruitment across the Great Lakes (also see Chapter 2). Previous modeling has indicated that warm-, cool- (including yellow perch), and even some coldwater fishes could benefit from climate change in the Great Lakes basin, owing primarily to an increase in thermal habitat available for growth (Stefan et al. 1996; Minns et al.

1997; Cline et al. 2013). Winter warming also would be expected to increase over-winter survival of juveniles of warm- and cool-water species such as smallmouth bass

(Micropterus dolomieu), white perch (Morone americana), and yellow perch (Johnson and Evans 1990; Shuter and Post 1990). Further, as the inability of smallmouth bass and

69 yellow perch juveniles to grow large enough to survive winter temperatures sets the northern range limits of these species (Shuter and Post 1990), enhanced growth and reduced winter temperatures, owing to expected climate warming, could result in range expansion (Magnuson et al. 1997). Expected increases in storm (precipitation) events and nutrient runoff during spring under future CO2 emission scenarios (Magnuson et al.

1997; Kling et al. 2003) also would be expected to benefit yellow perch recruitment via river-induced nutrient and turbidity plumes that increase foraging opportunities and reduce predation risk for yellow perch larvae (Ludsin 2000; Reichert et al. 2010; Ludsin et al. 2011; Pangle et al. 2012).

Our study, as well as others, suggests that the anticipated effects of climate on

Great Lakes fishes will not be all positive, however. For example, in addition to affecting the total number of larvae produced, climate warming scenarios also predict increased nutrient loading (via enhanced spring precipitation events), increased summer temperature, and reduced summer water levels (due to decreased summer precipitation and increased evapotranspiration) (Magnuson et al. 1997; Lofgren et al. 2002; Kling et al.

2003) that will lead to an increase in the intensity and duration of hypolimnetic hypoxia

(Kling et al. 2003; Fang et al. 2004). This is problematic because hypoxia, which has been increasing in the central basin of Lake Erie during recent years (Burns et al. 2005;

Hawley et al. 2006; Rucinski et al. 2010), has been shown to reduce consumption, the potential for growth, and energetic condition of both cool-water and cold-water (e.g., rainbow smelt Osmerus mordax) fishes by both forcing these species into sub-optimal

(warmer) surface and nearshore waters and reducing access to benthic food sources (e.g.,

70 chironomids) (Vanderploeg et al. 2009; Roberts et al. 2009, 2011). As percids such as yellow perch and walleye are capital breeders that allocate energy reserves accumulated during the summer and early fall to reproduction during late fall, when vitellogenesis begins (Dabrowski et al. 1996), female condition entering the winter is likely to determine the amount of energy available for reproduction (Henderson et al. 1996;

Henderson et al. 2000).

We did not, however, find that females entering the winter in poor condition had reduced reproductive output. We did not find that reduced condition entering winter reduced the overall amount of energy allocated to reproduction. In support of this notion, females in poor body condition had similar spring GSI values compared to females in the high body-condition treatment, despite delayed reproductive development females entering the winter. These findings indicate that reduced body condition entering winter, which can result from hypolimnetic hypoxia formation during late summer and early fall

(Roberts 2009, 2010, 2011; Scavia et al. in revision), may not reduce the overall amount of energy allocated to reproduction in a major way. Importantly, however, even if we conclude that hypoxia does not affect reproductive output by limiting energy reserves, hypoxia may negatively affect total egg production the following year by reducing growth and, consequently, female size in the fall. Because total egg production was shown to be a function of female size, smaller female yellow perch would produce fewer eggs than larger females, thus limiting the potential number of pre-recruits to the population (Houde 1987).

While our results suggest a possible mechanism through which climate change may affect yellow perch reproductive output and, subsequently, recruitment, we did not document complete reproductive failure in response to short, warm winters. This result was particularly surprising given previous research that indicated yellow and Eurasian perch females would not spawn following abnormally short winters (Jones et al. 1972;

Hokanson 1977; Ciereszko et al. 1997; Sandstrom et al. 1997; Lukšienê et al. 2000;

Migaud et al. 2002). If female yellow perch have a minimum duration of cold winter temperatures that is required for proper reproductive development (as suggested by previous investigators; Jones et al. 1972; Hokanson 1977; Ciereszko et al. 1997), our experimental design did not breach this limit for yellow perch from Lake Erie.

Importantly, many of these previous studies exposed female yellow or Eurasian perch to minimum winter temperatures that were warmer (Hokanson 1977: 6, 8, and 10°C;

Ciereszko et al. 1997: 7 and 11°C; Migaud et al. 2002: 6°C) than the minimum winter temperatures (4 – 5°C) used in our study, which mirrored those observed in Lake Erie over the past two decades. While our study did not replicate the dramatic effects observed in other studies (i.e., complete lack of spawning; Jones et al. 1972; Hokanson

1977; Ciereszko et al. 1997; Lukšienê et al. 2000), our finding that long winters produced larger eggs (also see Chapter 2) than shorter winters is supported by at least one previous study. Migaud et al. (2002) found that females exposed to 6°C water for 5 mos produced larger oocytes and had larger GSIs than those exposed to only 3 mos of 6°C water.

While the exact mechanism causing larger, higher quality eggs to be produced after long winters is unknown, we speculate that longer winters allow for a prolonged period of low

72 metabolic rates, which allow for a larger proportion of energy to be devoted to egg production. Controlled laboratory experiments and/or bioenergetics modeling would be needed to appropriately test this hypothesis.

Other studies, which were not investigating winter effects on reproduction, also found that larger Eurasian perch eggs typically hatched at higher rates (Olin et al. 2012) and, for walleye, led to higher juvenile survival rates (Venturelli et al. 2010) relative to their smaller counterparts. In a recent review, Kamler (2005) indicated that three primary periods of mortality exist for larvae prior to exogenous feeding: 1) early mortality associated with fertilization success; 2) mortality associated with hatching; and 3) starvation mortality that occurs when yolk reserves are depleted under absence of external food. Our hatching success metrics, as well as those reported in Chapter 2, quantified survival through the first two of these periods, and indicated that larvae from larger eggs (produced after long winters) survived these two periods at disproportionately higher rates than small larvae. Finally, large eggs also produced large larvae (also see

Chapter 2), which are more likely to survive the third period of elevated mortality than small larvae (Miller et al. 1988),and to survive in the presence of larval yellow perch predators such as invasive white perch (Ludsin et al. 2011). Therefore, our experimental results suggest that eggs and larvae produced from long winters should have greater chances of surviving all three of these critical mortality periods, and be better positioned to survive to the juvenile stage, than smaller eggs and larvae.

In contrast to the negative effects of short winters on female reproductive success, all measures of male reproductive success (sperm density, percent of motile sperm, and

73 duration of motility) indicated that males were capable of fertilizing eggs, even those exposed to our short winter-duration treatment. However, at the end of our long winter- duration treatment, Lake Erie males experienced a decline in sperm motility. This result was not surprising given that male yellow perch typically show a readiness to spermiate in January (Ciereszko et al. 1998). An extended period of low temperature (long winter) resulted apparently in sperm “aging” and deterioration of sperm quality. This finding is the first description of this phenomenon in yellow perch, although it has been documented in rainbow trout (Oncorhynchus mykiss) following an extended reproductive season (Ciereszko and Dabrowski 1995).

The lack of an effect of sperm quality on fertilization rate, and in turn, embryo hatching success may be an artifact of our fertilization methods. The spermatozoa to egg ratio used in our study was 100,000 to 1, which was double the optimum level found for other fishes (e.g., walleye = 50,000 spermatozoa per egg; Rinchard et al. 2005a). We chose this high ratio so as to ensure maximal fertilization rates. Thus, our use of more spermatozoa per egg may have prevented the observed decline in male milt motility from negatively affecting fertilization and egg hatching success. What remains unknown, however, is if this physiological response of males to excessively long winters serves as a mechanism to limit the positive effect of long winters on yellow perch reproduction and recruitment in the wild.

Understanding the effects of winter temperature on reproductive output and quality may allow Lake Erie fisheries managers to 1) better explain historically variable yellow perch recruitment in Lake Erie (YPTG 2013) and 2) better anticipate the effects of

74 climate change so that they can manage user-group expectations, keeping them from exceeding the ability of Lake Erie to produce yellow perch. Our mechanistic study also should serve as an impetus for future research into the effect of climate change on reproductive output and recruitment both within and outside the Great Lakes basin, given that 1) yellow perch and its congener, Eurasian perch, are economically and ecologically important species across North America and Europe, and 2) numerous other species across the world have similar reproductive characteristics as yellow perch (i.e., spring spawners with winter ovary development; e.g., European perch, walleye Sander vitreus, pike-perch Sander lucioperca, striped bass Morone saxatilis; Wang et al. 2010).

While our study has identified one particular mechanism by which climate change can affect reproductive output and recruitment, we suspect that many others exist, which have yet been explored. If so, the likelihood that climate change will lead to ecological surprises (Paine et al. 1998; Doak et al. 2008) will only increase, especially when considering that the effects of climate change are likely interact with other anthropogenic stressors (e.g., altered system productivity, invasive species, exploitation; Perry et al.

2010). For this reason, we encourage integrative research approaches that combine controlled laboratory experiments, field observations, and modeling. Such approaches offer the best means not only to avoid major ecological surprises that can have devastating ecological and socioeconomic impacts, but also to provide critical ecological understanding that can increase the ability of mangers to overcome the complex challenge in sustaining fisheries in the face of a changing climate.

References

Angel, J. R., and K. E. Kunkel. 2010. The response of Great Lakes water levels to future climate scenarios with an emphasis on Lake Michigan-Huron. Journal of Great Lakes Research 36:51-58.

Arend, K. K., D. Beletsky, J. V. DePinto, S. A. Ludsin, J. J. Roberts, D. K. Rucinski, D. Scavia, D. J. Schwab, and T. O. Hoeoek. 2011. Seasonal and interannual effects of hypoxia on fish habitat quality in central Lake Erie. Freshwater Biology 56:366-383.

Assel, R., K. Cronk, and D. Norton. 2003. Recent trends in Laurentian Great Lakes ice cover. Climatic Change 57:185-204.

Berkeley, S. A., C. Chapman, and S. M. Sogard. 2004. Maternal age as a determinant of larval growth and survival in a marine fish, Sebastes melanops. Ecology 85:1258- 1264.

Brook, B. W., N. S. Sodhi, and C. J. A. Bradshaw. 2008. Synergies among extinction drivers under global change. Trends in Ecology & Evolution 23:453-460.

Bunnell, D. B., C. P. Madenjian, J. D. Holuszko, J. V. Adams, and J. R. P. French. 2009. Expansion of Dreissena into offshore waters of Lake Michigan and potential impacts on fish populations. Journal of Great Lakes Research 35:74-80.

Chambers, R. C., and W. C. Leggett. 1996. Maternal influences on variation in egg sizes in temperate marine fishes. American Zoologist 36:180-196.

Ciereszko, A., and K. Dabrowski. 1995. Sperm quality and ascorbic acid concentration in rainbow trout semen are affected by dietary vitamin C: an across season study. Biol. Reprod. 52:982-988.

Ciereszko, R. E., K. Dabrowski, A. Ciereszko, J. Ebeling, and J. S. Ottobre. 1997. Effects of temperature and photoperiod on reproduction of female yellow perch Perca flavescens: Plasma concentrations of steroid hormones, spontaneous and induced ovulation, and quality of eggs. Journal of the World Aquaculture Society 28:344- 356.

Ciereszko, R. E., K. Dabrowski, A. Ciereszko, and J. S. Ottobre. 1998. Plasma concentrations of steroid hormones in male yellow perch, Perca flavescens: The effect of age and photothermal manipulation. Environmental Biology of Fishes 51(1):97-105.

Cline, T. J., V. Bennington, and J. F. Kitchell. 2013. Climate Change Expands the Spatial Extent and Duration of Preferred Thermal Habitat for Lake Superior Fishes. Plos One 8.

Doney, S. C., M. Ruckelshaus, J. E. Duffy, J. P. Barry, F. Chan, C. A. English, H. M. Galindo, J. M. Grebmeier, A. B. Hollowed, N. Knowlton, J. Polovina, N. N. Rabalais, W. J. Sydeman, and L. D. Talley. 2012. Climate Change Impacts on Marine Ecosystems. Pages 11-37 in C. A. Carlson, and S. J. Giovannoni, editors. Annual Review of Marine Science, volume 4.

Drinkwater, K. F., G. Beaugrand, M. Kaeriyama, S. Kim, G. Ottersen, R. I. Perry, H. O. Portner, J. J. Polovina, and A. Takasuka. 2010. On the processes linking climate to ecosystem changes. Journal of Marine Systems 79(3-4):374-388.

Ficke, A. D., C. A. Myrick, and L. J. Hansen. 2007. Potential impacts of global climate change on freshwater fisheries. Reviews in Fish Biology and Fisheries 17:581- 613.

Hart, S. D., D. L. Garling, and J. A. Malison. 2006. Yellow perch (Perca flavescens) culture guide. North Central Regional Aquaculture Center, Iowa State University, Ames, Iowa.

Hartmann, D. L. 1990. Modeling climate change. Pages 97-140 in G. J. Macdonald, and L. Sertorio, editors. Global Climate and Ecosystem Change, Maratea, Italy.

Heath, D. D., J. W. Heath, C. A. Bryden, R. M. Johnson, and C. W. Fox. 2003. Rapid evolution of egg size in captive salmon. Science 299:1738-1740.

Hellmann, J. J., J. E. Byers, B. G. Bierwagen, and J. S. Dukes. 2008. Five potential consequences of climate change for invasive species. Conservation Biology 22:534-543.

Henderson, B. A., T. Trivedi, and N. Collins. 2000. Annual cycle of energy allocation to growth and reproduction of yellow perch. Journal of Fish Biology 57:122-133.

Henderson, B. A., J. L. Wong, and S. J. Nepszy. 1996. Reproduction of walleye in Lake Erie: Allocation of energy. Canadian Journal of Fisheries and Aquatic Sciences 53:127-133.

Hokanson, K. E. F. 1977. Temperature requirements of some percids and adaptations to seasonal temperature cycle. Journal of the Fisheries Research Board of Canada 34:1524-1550.

Houde, E. D. 1987. Fish early life dynamics and recruitment variability. Pages 17-29 in, volume American Fisheries Soceity Symposium.

Hutchings, J. A. 1991. Fitness consequences of variation in egg size and food abundance in brook trout Salvelinus fontinalis. Evolution 45:1162-1168.

Johnson, T. B., and D. O. Evans. 1990. Size-dependent winter mortality of young-of-the- year white perch: climate warming and invasion of the Laurentian Great Lakes. Transactions of the American Fisheries Society 119:301-313.

Johnston, T. A., M. D. Wiegand, W. C. Leggett, R. J. Pronyk, S. D. Dyal, K. E. Watchorn, S. Kollar, and J. M. Casselman. 2007. Hatching success of walleye embryos in relation to maternal and ova characteristics. Ecology of Freshwater Fish 16:295-306.

Kamler, E. 2005. Parent-egg-progeny relationships in teleost fishes: an energetics perspective. Reviews in Fish Biology and Fisheries 15:399-421. 79

Kennedy, B. P., K. H. Nislow, and C. L. Folt. 2008. Habitat-mediated foraging limitations drive survival bottlenecks for juvenile salmon. Ecology 89:2529-2541.

Kwasek, K.A. 2012. The nutritional and genetic effects on body growth, reproduction and molecular mechanisms responsible for muscle growth in yellow perch Perca flavescens. Ph.D. dissertation. The Ohio State University, Columbus.

Ludsin, S., K. Pangle, L. Carreon-Martinez, N. Legler, J. Reichert, D. D. Heath, B. Fryer, T. Johnson, J. Tyson, T. Höök, G. Leshkevich, D. Mason, D. Bunnell, C. Mayer, T. Johengen, H. Vanderploeg, and A. Drelich. 2011. River discharge as a predictor of Lake Erie yellow perch recruitment. Final completion report. Great Lakes Fishery Commission, Fisheries Research Program, Ann Arbor, MI.

Ludsin, S. A. 2000. Exploration of spatiotemporal patterns in recruitment and community organization of Lake Erie fishes: a multiscale, mechanistic approach. Ph.D. dissertation. The Ohio State University, Columbus.

Luksiene, D., O. Sandstrom, L. Lounashelmo, and J. Andersson. 2000. The effects of thermal effluent exposure on the gametogenesis of female fish. Journal of Fish Biology 56:37-50.

MacKay, M., and F. Seglenieks. 2013. On the simulation of Laurentian Great Lakes water levels under projections of global climate change. Climatic Change 117:55- 67.

Magnuson, J. J. 2010. History and heroes: the thermal niche of fishes and long-term lake ice dynamics. Journal of Fish Biology 77:1731-1744.

McCormick, M. J., and G. L. Fahnenstiel. 1999. Recent climatic trends in nearshore water temperatures in the St. Lawrence Great Lakes. Limnology and Oceanography 44:530-540.

Minns, C. K., J. A. Lane, and C. B. Portt. 1997. Analyses of the habitat requirements of Great Lakes fishes. 40th Conference of the International Association for Great Lakes Research:98-98. 81

Mortsch, L. D. 1998. Assessing the impact of climate change on the Great Lakes shoreline wetlands. Climatic Change 40:391-416.

Mortsch, L. D., and F. H. Quinn. 1996. Climate change scenarios for Great Lakes Basin ecosystem studies. Limnology and Oceanography 41:903-911.

Moyle, P. B., and J. J. Cech. 2000. Fishes: an introduction to ichtyology, 4th edition. Prentice Hall, Englewood Cliffs, NJ, USA.

Olin, M., J. Jutila, H. Lehtonen, M. Vinni, J. Ruuhijarvi, S. Estlander, M. Rask, A. Kuparinen, and J. Lappalainen. 2012. Importance of maternal size on the reproductive success of perch, Perca fluviatilis, in small forest lakes: implications for fisheries management. Fisheries Management and Ecology 19:363-374.

Ottersen, G., S. Kim, G. Huse, J. J. Polovina, and N. C. Stenseth. 2010. Major pathways by which climate may force marine fish populations. Journal of Marine Systems 79:343-360.

Paine, R. T., M. J. Tegner, and E. A. Johnson. 1998. Compounded perturbations yield ecological surprises. Ecosystems 1:535-545.

Pangle, K. L., T. D. Malinich, D. B. Bunnell, D. R. DeVries, and S. A. Ludsin. 2012. Context-dependent planktivory: interacting effects of turbidity and predation risk on adaptive foraging. Ecosphere 3:art114.

Perry, R. I., P. Cury, K. Brander, S. Jennings, C. Möllmann, and B. Planque. 2010. Sensitivity of marine systems to climate and fishing: concepts, issues and management responses. Journal of Marine Systems 79:427-435.

Reichert, J. M., B. J. Fryer, K. L. Pangle, T. B. Johnson, J. T. Tyson, A. B. Drelich, and S. A. Ludsin. 2010. River-plume use during the pelagic larval stage benefits recruitment of a lentic fish. Canadian Journal of Fisheries and Aquatic Sciences 67:987-1004.

Rideout, R. M., G. A. Rose, and M. P. M. Burton. 2005. Skipped spawning in female iteroparous fishes. Fish and Fisheries 6:50-72.

Roberts, J. J., S. B. Brandt, D. Fanslow, S. A. Ludsin, S. A. Pothoven, D. Scavia, and T. O. Hoeoek. 2011. Effects of hypoxia on consumption, growth, and RNA:DNA ratios of young yellow perch. Transactions of the American Fisheries Society 140:1574-1586.

Roberts, J. J., P. A. Grecay, S. A. Ludsin, S. A. Pothoven, H. A. Vanderploeg, and T. O. Hoeoek. 2012. Evidence of hypoxic foraging forays by yellow perch (Perca flavescens) and potential consequences for prey consumption. Freshwater Biology 57:922-937.

Sandstrom, O., I. Ambrahamsson, J. Andersson, and M. Vetemaa. 1997. Temperature effects on spawning and egg development in Eurasian perch. Journal of Fish Biology 51:1015-1024.

Scott, W.B., and E.J. Crossman. 1973. Freshwater fishes of Canada. Bulletin 184. Fisheries Research Board of Canada. 966 pp

Sharma, S., D. A. Jackson, C. K. Minns, and B. J. Shuter. 2007. Will northern fish populations be in hot water because of climate change? Global Change Biology 13:2052-2064.

Shuter, B. J., and J. D. Meisner. 1992. Tools for assessing the impact of climate change on frehswater fish populations. GeoJournal 28:7-20.

Shuter, B. J., and J. R. Post. 1990. Climate, population viability, and the zoogeography of temperate fishes.Transactions of the American Fisheries Society 119:314-336.

Vanderploeg, H. A., S. A. Ludsin, S. A. Ruberg, T. O. Hoeoek, S. A. Pothoven, S. B. Brandt, G. A. Lang, J. R. Liebig, and J. F. Cavaletto. 2009. Hypoxia affects spatial distributions and overlap of pelagic fish, zooplankton, and phytoplankton in Lake Erie. Journal of Experimental Marine Biology and Ecology 381:S92- S107.

Vitousek, P. M. 1994. Beyond global warming: ecology and global change. Ecology 75:1861-1876.

Wang, J., X. Bai, H. Hu, A. Clites, M. Colton, and B. Lofgren. 2012. Temporal and spatial variability of Great Lakes ice cover, 1973-2010. Journal of Climate 25:1318-1329.

Wang, N., F. Teletchea, P. Kestemont, S. Milla, and P. Fontaine. 2010. Photothermal control of the reproductive cycle in temperate fishes. Reviews in Aquaculture 2:209-222.

Yellow Perch Task Group (YPTG). 2013. Report of the Lake Erie yellow perch task group. Great Lakes Fishery Commission, Ann Arbor, MI.

Figure 3.1. Mean (± 1 SE) residuals (derived from sex-specific total length – mass relationships) for male and female Lake Erie yellow perch from high (ad libitum; female N = 49; male N = 27) and low (maintenance ration; female N = 49; male N = 27) body- condition treatments. Means with different lowercase letters within a panel differed in Tukey’s honestly significant difference post- hoc comparison (p < 0.05).

Figure 3.2. Mean (± 1 SE) growth from late July through early October for male and female Lake Erie yellow perch in high (female N = 49; male N = 27) and low (female N = 49; male N = 27) body-condition treatments. Means with different lowercase letters differed after correction for multiple comparisons (Tukey’s honestly significant difference post-hoc comparison; p < 0.05).

Figure 3.3. Female and male yellow perch gonad mass (g) versus total length for individuals sacrificed from high and low body- condition treatments prior to the start of the experiment (during the first week of October).

Figure 3.4. Hatching success relative to fertilization success for Lake Erie yellow perch eggs across all experimental treatments. The dashed line indicates the 1:1 line.

Figure 3.5. Mean (± 1 SE) hatching success (to the nearest %) of yellow perch eggs plotted against mean (± 1 SE) residuals (derived from female total length – mass relationships) of female Lake Erie yellow perch condition. Negative residual values indicate females in relatively poorer condition, whereas positive residuals indicate females in relatively better condition.

Figure 3.6. Lake Erie yellow perch fecundity (eggs per female) shown as a function of total length. A single relationship describes the number of eggs produced from females in both the short and long winter-duration treatments in our experiment.

Figure 3.7. Mean (± 1 SE) fecundity residuals (from Figure 3.6) plotted against mean (± 1 SE) residuals (derived from sex-specific total length – mass relationships) for Lake Erie yellow perch. Negative residual values indicate females in relatively poorer condition while positive residuals indicate females in relatively better condition.

Figure 3.8. Mean (± 1 SE) gonadosomatic index (GSI: [gonad mass/body mass]*100) for female yellow perch hand-stripped of eggs at the conclusion of both the short and long winter-duration treatments. Means with different lowercase letters within a panel differed (Tukey’s honestly significant difference post-hoc comparison; p < 0.05).

Figure 3.9. Mean (± 1 SE) egg diameter for Lake Erie yellow perch hand-stripped of eggs in the short and long winter-duration treatments. Means with different lowercase letters within a panel differed (Tukey’s honestly significant difference post-hoc comparison; p < 0.05).

Figure 3.10. Scatterplots of mean (± 1 SE) hatching success relative to mean (± 1 SE) egg mass and mean (± 1 SE) energy density for fertilized eggs from Lake Erie yellow perch in our laboratory experiments.

Figure 3.11. Individual sperm motility in male yellow perch exposed to short versus long winter-durations in a controlled laboratory experiment. Individual measurements of sperm motility (short winter N = 23; long winter N = 34) were collected from males (short winter N = 12 males; long winter N = 18 males) selected haphazardly during the spawning season for fertilization of egg samples.

CHAPTER 4

THE ROLE OF CLIMATE CHANGE, INVASIVE SPECIES, AND NUTRIENT-

DRIVEN REGIME SHIFT IN A GREAT LAKES FISH POPULATION

Introduction

Climate change and invasive species spread are two key human-driven processes affecting ecosystem functioning, community composition, and population dynamics

(Vitousek 1994; Hellmann et al. 2008). While climate warming can negatively affect physiological processes of resident species, including growth and reproduction (Portner and Knust 2007; Neuheimer et al. 2011; Baumann et al. 2012;Chapter 2), invasive species can have negative effects on resident species through trophic interactions (e.g. competition, predation; Ruiz et al. 1999; Grosholz 2002). In addition to these independent effects, climate change is expected to facilitate the introduction, establishment, and spread of certain invasive species (Stachowicz et al. 2002; Rahel and

Olden 2008; Walther et al. 2009). In this manner, the combination of climate change and invasive species spread may have synergistic effects on resident species (Brook et al.

2008; Doney et al. 2012), increasing the likelihood for “ecological surprises” (Paine et al.

1998; Doak et al. 2008).

As a large, north-temperate freshwater ecosystem, The Laurentian Great Lakes have warmed rapidly during the past several decades. Water temperature has increased twice as fast as air temperatures in some lakes (Austin and Colman 2007) with ice cover correspondingly decreasing by more than 70% (Wang et al. 2012). Concurrently, the

Great Lakes also have experienced a large influx of non-native, invasive species due largely to ballast water transport of Atlantic and Eurasian aquatic species through the St.

Lawrence Seaway (Mills et al. 1993; Ricciardi and MacIsaac 2000). While both warming and species invasions are likely to have had significant negative effects on native Great Lakes species, drawing direct causal relationships among warming, invasions, and native species abundance has been difficult, owing to a background of constantly changing conditions, driven largely by anthropogenic stressors (i.e., nutrient- driven regime shifts [Ludsin 2000, Ludsin et al. 2001], overfishing [Francis et al. 1996;

Wilberg et al. 2005]; habitat loss [Casselman and Lewis 1996]).

Despite this challenge, several recent studies of Lake Erie yellow perch (Perca flavescens) have documented mechanistic linkages between climate change and invasive species and indices of juvenile abundance (Reichert et al. 2010; Carreon-Martinez 2011;

Ludsin et al. 2011; Farmer et al. 2013; Chapter 2). In Lake Erie, fall yellow perch juvenile abundance is a strong predictor of future yellow perch recruitment to the commercial and recreational fisheries (which occurs at age-2) and a given cohort’s future contribution to harvest (Ludsin 2000; YPTG 2013; see Chapter 1). Therefore, mechanisms affecting juvenile abundance would be expected to have lasting implications for yellow perch population dynamics in Lake Erie. Investigating a previously

98 documented (Ludsin 2000) positive correlation between river discharge and fall yellow perch juvenile abundance in western Lake Erie, Ludsin et al. (2011) found that large spring sediment plumes from the Maumee River into western Lake Erie provided larval yellow perch refuge from predation, enhancing survival and year-class strength (Reichert et al. 2010; Ludsin et al. 2011). Further, genetic analysis of predator diets revealed invasive age-1+ white perch (Morone americana) were particularly voracious predators of larval yellow perch in western Lake Erie (Carreon-Martinez 2011; Ludsin et al. 2011).

Together these studies suggest that, in the absence of a large spring-time sediment plume, invasive white perch predation has a strong negative effect on larval yellow perch survival and, subsequently, fall juvenile abundance (Ludsin et al. 2011).

Investigating the role of climate change in regulating yellow perch populations, another set of studies found that short, warm winters decreased yellow perch egg quality, embryo hatching success, and larval size-at-hatching (Farmer et al. 2013; Chapter 2).

Additionally, field and laboratory studies found that when spring warming happened extremely early, yellow perch spawning did not fully adjust, increasing the possibility of a mis-match between first-feeding larvae and their zooplankton prey (Farmer et al. 2013;

Chapter 2). Together, i) declines in hatching success that directly reduce the number of potential recruits, ii) reductions in larval size-at-hatching that lead to reduced growth and survival during the larval stage (Miller et al. 1988; Einum and Fleming 2000; Ludsin et al. 2001), and iii) mis-matches in timing between larvae and their prey, reducing larval survival (Cushing 1990; Edwards and Richardson 2004; Wright and Trippel 2009), provide biological mechanisms by which short, warm winters may negatively affect

99 yellow perch fall juvenile abundance. Additionally, these findings provide mechanistic explanations for previously detected positive relationships between yellow perch fall juvenile abundance and indices of winter duration (Ludsin 2000; Crane 2007). Together, these previous studies (Reichert et al. 2010; Ludsin et al. 2011; Farmer et al. 2013;

Chapter 2) provide strong evidence for mechanistic linkages between climate, invasive species, and yellow perch year-class strength.

We assessed the independent and combined effects of climate change and an invasive species (white perch) on Lake Erie yellow perch fall juvenile abundance using historical datasets from western and central Lake Erie (1969-2010). Although our primary interest was to evaluate the effects of winter duration and white perch abundance on yellow perch recruitment, we also included other environmental variables that were expected to mediate relationships between yellow perch, invasive species, and climate change (i.e., river discharge [Ludsin 2000; Reichert et al. 2010; Ludsin et al. 2011]) or that had previously been shown to influence yellow perch recruitment independently (i.e., spring temperature [Eshenroder 1977; Henderson and Nepszy 1988; Crane 2007], spring warming rate [Clady 1976; Eshenroder 1977; Henderson and Nepszy 1988], and adult spawning stock size [Henderson and Nepszy 1988; Redman et al. 2011; Forsythe et al.

2012]). By including these additional variables, I sought to understand the effects of warm winters and predation by an invasive species on yellow perch recruitment while accounting for annual environmental stochasticity and variation in spawning stock size.

Ultimately, my goal in this analysis was not to explain all of the variability in yellow perch recruitment, but rather to test the hypotheses that i) increased frequency of short,

100 warm winters and ii) increased larval predation pressure from an invasive species have both negatively affected yellow perch fall juvenile abundance in Lake Erie in recent years.

Methods

Study site

Lake Erie is 1 of the 5 Laurentian Great Lakes that together compose, by surface area, the largest freshwater system on earth (Fuller et al. 1995). Among the Great Lakes,

Lake Erie is the southern-most, shallowest, and warmest and has the shortest water retention time (Fuller et al. 1995). These characteristics have made Lake Erie the most productive of the Great Lakes across all trophic levels. Due to its geomorphology, Lake

Erie has three distinct basins (west, central, and east) that decrease in depth, temperature, and productivity moving from west to east (see Bolsenga and Herdendorf 1993 for details). This decrease in productivity is driven both by physical factors (temperature and depth), as well as by the fact that > 90% of water and nutrients entering Lake Erie do so from western tributaries (i.e., Detroit, Maumee, and Sandusky Rivers; Ludsin et al. 2001;

OEPA 2013). For the purpose of this study, I will focus on yellow perch populations inhabiting the western and central basins of Lake Erie, for which long-term data exist.

Evidence is mounting to indicate that north temperate systems such as the

Laurentian Great Lakes have been warming during recent decades (Assel et al. 1995;

McCormick and Fahnenstiel 1999; Jones et al. 2006). Water temperatures and the duration of summer stratification also have increased across the Great Lakes during the

101

20th century (McCormick and Fahnenstiel 1999), and general circulation models (GCMs) indicate that this warming trend is expected to continue throughout this century

(Magnuson et al. 1997; Lofgren et al. 2002; Kling et al. 2003), with predicted increases of 2-5 °C during winter by 2050, depending on emission scenarios (Hayhoe et al. 2010).

Concurrently with warming, the Great Lakes also have experienced a dramatic increase in the number of introduced species (Mills et al. 1993; Ricciardi and MacIsaac

2000). Lake Erie, along with the other Great Lakes have been invaded by an estimated

145 aquatic species since the early 1800s (Mills et al. 1993; Riccardi and MacIssac

2000). Specifically, the invasions of Dreissenid mussels (Dreissena polymorpha,

Dreissena bugensis; Griffiths et al. 1991) and round gobies (Neogobius melanostomus;

Charlebois et al. 2001) have garnered much attention and appear to have had large effects on ecosystem processes and food webs (Vanderploeg et al. 2001; Hecky et al. 2004;

Johnson et al. 2005). I do not dispute the likelihood that these large-scale ecological changes brought on by invasive species affected yellow perch fall juvenile abundance through a variety of mechanisms. Herein, however, I focused specifically on white perch, as this invasive species has the potential to have strong, direct negative effects on yellow perch larval survival (Carreon-Martinez 2011; Ludsin et al. 2011) and, subsequently, year-class strength.

White perch are an invasive predator of (Roseman et al. 1996; Madenjian et al.

2000; Carreon-Martinez 2011) and competitor with (Parrish and Margraf 1990; Gopalan et al. 1998) larval and juvenile yellow perch. White perch established in Lake Erie in the

1970s (Stapanian et al. 2007) and maintained low population levels until they

102 dramatically increased in abundance following a period of short, warm winters in the mid-1980s (Gopalan et al. 1998; Stapanian et al. 2007; see Figure 4.1). This period of increased white perch abundance correlated with a decline in yellow perch juvenile abundance (Stapanian et al. 2007; see Figure 4.2). This climate-mediated expansion of an invasive predator of yellow perch may have served to limit the number of potential yellow perch recruits surviving through the first summer of life.

In addition to climate-driven warming and re-structuring of the food web by invasive species, Lake Erie experienced large-scale oligotrophication in the mid-1980s.

This was due, in large part, to planned phosphorus (P) abatement programs implemented in the 1970s as part of the US-Canada Great Lakes Water Quality Agreement, along with the zebra mussel invasion (ca. 1987; Griffiths et al. 1991). Together, these changes reduced P levels (Dolan 1993; Neilson et al. 1995) and led to oligotrophication, evidenced by decreased phytoplankton biomass (Nicholls and Hopkins 1993; Nicholls

1999; Boegman et al. 2008), decreased zooplankton biomass (Makarewicz et al. 1998;

Ludsin et al. 2001), enhanced water clarity (Johansson et al. 1999; Weisgerber 1999), increased benthic macroinvertebrate abundance (Krieger et al. 1996; Tyson and Knight

2001), and shifts in fish community composition (Ludsin et al. 2001; Ludsin and Stein

2001). While our study seeks to understand the role of climate change and an invasive species on Lake Erie yellow perch recruitment, our analysis also must account for this large-scale regime shift in productivity, which occurred during the years of interest for this study.

103

Index of yellow perch fall juvenile abundance

Our analysis used yellow perch juvenile (age-0) abundance data, expressed as catch per unit effort (# individuals per minute of trawling; CPUE), generated annually during a long-term fall bottom trawl survey conducted by the Ohio Department of

Natural Resources-Ohio Division of Wildlife (DOW) during 1969-2010. As sampling was conducted by multiple vessels during these years, we applied vessel-specific fishing power corrections to juvenile abundances from 1982-2010 to standardize catches (Tyson et al. 2006). Juvenile abundances from 1969-1981 had no fishing power corrections applied, due to a lack of studies comparing older and modern research vessels. These data were used under the assumption that the magnitude of variation between strong and weak year-classes far exceeded the rather small variation due to differences in capture efficiency among vessels during this time, an assumption supported by more recent

(1982-2010) comparisons between fishing power corrections (Tyson et al. 2006) and annual variation in year-class strength.

As evidence now exists that separate yellow perch stocks exist within Lake Erie

(Sepulveda-Villet et al. 2009; Kocovsky and Knight 2012) and abiotic and biotic conditions vary spatially across the lake, models were parameterized separately for both the western and central basins of Lake Erie as these large sections of the lake are currently thought to represent distinct populations (YPTG 2013). Because survey designs have changed during the past 35+ years, we used only yellow perch data from twelve fixed, historical sites, sampled consistently during this time. Historical sites were spread across the Ohio waters of western (N=4 sites) and central (N=8 sites) Lake Erie (see

104

Ludsin et al. 2001). Annual juvenile abundances from historical sites correlated strongly with overall annual abundance calculated using all sites in both the western (r = 0.92; ~

80 sites per year since 1987) and central Lake Erie (r = 0.94; ~ 40 sites per year since

1990), indicating historical sites closely track population-level variation in juvenile abundance.

Explanatory variables

Ice cover. Ice cover trends during 1973-2010 for Lake Erie were compiled from ice charts collected by the Canadian Ice Service and NOAA National Ice Center and summarized by the NOAA Great Lakes Environmental Research Laboratory, Ann Arbor,

MI, USA (Wang et al. 2012). Daily values of lake area ice cover were obtained by interpolating between ice chart values, originally obtained at approximately weekly intervals. Mean February-March lake area ice cover was calculated by averaging daily ice cover values during these two months. As ice cover reaches its peak on Lake Erie during February through March (Wang et al. 2012), we selected this timeframe to distinguish between years of low and high ice cover. Because winter ice cover data were unavailable during 1969-1972, we used a relationship between mean winter air temperature and ice cover (estimated from 1973-2010 data; r = -0.86, p < 0.0001) to estimate 1969-1972 ice cover. As these estimates of ice cover were annual lake-wide averages, identical annual ice cover values were used for both western and central Lake

Erie models for 1969-1972.

105

A previous analysis found a positive threshold relationship between Lake Erie winter (February-March) ice cover and yellow perch juvenile abundance in both western and central Lake Erie (1973-2010; Chapter 2). In both the west and central basins, short, warm winters with little ice cover are always followed by low juvenile abundance; high juvenile abundance occurs only after long, cold winters with high ice cover. Using these previously detected thresholds (west basin 66% ice cover and central basin 71% ice cover; see Chapter 2 Figure 1.1), we created a binary variable for each annual ice cover value to indicate whether ice cover was above or below the threshold value for each basin. In this manner, a continuous index of winter ice cover was transformed into a class variable for use in subsequent model selection.

White perch. Annual abundances of white perch (age-1+) were obtained from the previously described October bottom trawl surveys conducted by Ohio DOW.

Abundance was expressed as catch per minute of bottom trawling (CPUE), and was calculated separately for western and central Lake Erie. This index of white perch abundance was intended to serve as a proxy for predation pressure on larval yellow perch during the spring. Given previous field observations (Gopalan et al. 1998) and the fact the white perch are predators of larval yellow perch (Roseman et al. 1996; Madenjian et al.2000; Carreon-Martinez 2011), we expected yellow perch fall juvenile abundance to be negatively related to white perch abundance.

Yellow perch. Spawning-stock size of adult (age-3+) yellow perch (in millions of fish) was obtained from a statistical catch-at-age analysis using the Auto Differentiation

Model Builder computer program (ADMB; see Myers and Bence 2001 and YPTG 2013

106 for details), which is informed by both fishery-independent and fishery-dependent data.

Separate estimates of adult population sizes were generated for both western and central

Lake Erie for 1969-2012. While some yellow perch reach maturity by mature at age-2, percent maturity at age-2 is highly variable (range: 6 – 97% from 1992-2010; ODOW

2011). As age-specific maturity information was unavailable for all years of this study, we included only age-3 and older fish in any calculation of spawning stock size, as a conservative measure. Virtually all female yellow perch are mature by age-3 (ODOW

2011). For western Lake Erie, spawning stock size was estimated for management unit one (MU1), whereas adult population size for the central basin was the sum from MU2 and MU3 (see YPTG 2013 for detailed description of MUs).

River discharge. We accounted for the known positive effect of spring river discharge on fall yellow perch juvenile abundance using Maumee River discharge

(recorded daily to the nearest m3/s) provided by the U.S. Geological Survey (USGS) from station #04193500 at Waterville, OH. As previous research has shown that April-May

Maumee River discharge is a strong predictor of yellow perch year-class strength (Ludsin

2000; Ludsin et al. 2011), we averaged mean daily flow over this period to obtain annual mean spring (April-May) flow for each year. Since the mid-1980s, following the oligotrophication of Lake Erie, Maumee River discharge has been inversely related to water clarity in Lake Erie during spring (Ludsin et al. 2011). Given that the Maumee

River contributes the vast majority of sediments to Lake Erie and its spring flows are highly correlated with other smaller tributaries of Lake Erie (see Ludsin et al. 2011), we used annual spring Maumee River discharge as an indicator of sediment turbidity levels

107 in both our west and central basin models. Given findings from previous studies, we hypothesized that spring (April-May) Maumee River discharge would be positively related to yellow perch fall juvenile abundance in both basins.

Spring temperature and warming rate. As no continuous dataset of Lake Erie water temperatures was available from 1969-2010, we created a continuous index of regional air temperature (nearest 1°C) using data provided by the National

Oceanographic and Atmospheric Administration’s National Climatic Data Center

(NOAA-NCDC). To accomplish this, we first averaged daily minimum and maximum temperatures recorded at two Ohio weather stations (Toledo [NOAA-NCDC station

#94830] and Cleveland [NOAA-NCDC station #14820]) and then averaged daily mean temperatures across both sites. Similar to the approach in Ludsin (2000), we smoothed annual temperature profiles using a 3-d moving average. Using this approach, we developed a single temperature dataset for Lake Erie derived from both a west basin

(Toledo) and central basin (Cleveland) site. Our measure of spring temperature was average May temperature, as this is the typical period of larval yellow perch growth and development in Lake Erie (Gopalan et al. 1998; Ludsin 2000; Reichert et al. 2011; Ludsin et al. 2011). Previous studies found positive relationships between spring water temperature and yellow perch year-class strength (Clady 1976; Eshenroder et al. 1977;

Henderson and Nepszy 1988). Our measure of spring warming rate was the slope of mean daily spring (April-May) temperatures plotted against day-of-the-year, as previous studies have found that increased warming rates during spawning and larval growth periods (i.e., April-May) also positively correlate with strong yellow perch year-classes

108

(Clady 1976; Eshenroder 1977; Henderson and Nepszy 1988). Based on these previous studies, we hypothesized that both spring temperature and warming rate would positively correlate with yellow perch fall juvenile abundance.

Data analysis

Given that our analysis spanned 42 years (1969-2010) and that well documented and dramatic changes in thermal regimes (Assel et al. 1995; McCormick and Fahnenstiel

1999; Wang et al. 2012), community composition (Ludsin et al. 2001; Ludsin and Stein

2001), and productivity (Dolan 1993; Neilson et al. 1995) occurred in Lake Erie during this time, we considered the possibility that mechanisms driving yellow perch recruitment may also have changed during this period. To account for this possibility, we split the datasets for west and central Lake Erie into two time periods. The earlier time period

(1969-1985) represents an environment with high frequency of cold, long winters, increased productivity, and low white perch abundance (Figure 4.1). The later period

(1986-2010) represents a period of warmer winters (specifically with increased frequency of short, warm ice-free winters; Figure 4.1) lower productivity, and increased white perch abundance (Figure 4.1). This approach represents an alternative to incorporating interaction terms into linear models (Ludsin 2000).

To quantify the effect of yellow perch juvenile abundance, we used both multiple regression (i.e., linear models) and modified Ricker stock-recruit models that were fit with environmental terms (Hilborn and Walters 1992). For Ricker stock-recruit models, we log transformed both sides of the equation (loge[R] = loge[S] + a – bS + c1X1 + c2X2 +

109

ε) where R was an index of recruitment (i.e., juvenile abundance), S was a measure of spawning-stock size, X1 and X2 were environmental variables (winter ice cover, adult white perch abundance), and ε was the error term. We used non-linear regression (PROC

NLIN; SAS Institute 2007) to estimate a and b parameters.

To determine the most parsimonious model for each time period and basin, we used second-order Akaike’s information criteria (AICc: Burnham and Anderson 2002), which is suggested when the ratio of sample size to the number of parameters is low

(ratio < 40; Burnham and Anderson 2002). AICc accounts for goodness of fit, while assessing a penalty for model complexity, thereby favoring models with reduced complexity that explain a high degree of variability. Here,

( ) ( )

where ni = the number of observations used to estimate model i, RSSi = the residual sum of squares from least squares regression analysis for model i, and ki = the number of independent variables in model i. Individual AICci values are not informative, as they are affected by arbitrary constants and sample size. Therefore, to assessed model confidence we used two approaches. First, I re-scaled each AICci value to

Δi = AICci – AICc(min)

110 where AICc(min) is the minimum of all AICci values (Burnham and Anderson 2002). The

Δi are easily interpretable and allow for simple comparison and ranking of candidate models. Typically, models with Δi ≤ 2 have substantial support (Burnham and Anderson

2002), compared to other models. Secondly, we calculated the Akaike weight

0.5( AIC ) e i c(min)   i 0.5( AIC ) re j c(min) j

where r = the number of models examined. Akaike weights are the normalized relative likelihood that a given model is the best among the subset of models (Burnham and

Anderson 2002). For example, a model with a ωi = 2.0 is twice as likely to be the ‘best’ model as one with ωi = 1.0. To determine the relative importance of each explanatory variable, we calculated the Akaike weight by summing ωi over all models in which a given explanatory variable appeared. To meet assumptions of normality and linearity, explanatory and response variables were loge transformed prior to analysis, as needed.

To guard against the possibility that our ‘best’ models were not simply the ‘best’ bad models selected from a suite of bad models, we also included null models (i.e., intercept only models) in our model selection analysis for each basin and time period. If a null model had Δi ≤ 2, we disregarded our ‘best’ models and concluded that none of our candidate models adequately explained the variance in our response variable. In this manner, our null models can be thought of as a p-value for our ‘best’ models, indicating if these models explained more variation than expected by random chance alone.

111

Finally, as collinearity among explanatory variables may lead to regression coefficients that area unstable (i.e., have large SEs) and have counterintuitive signs

(Belsley et al. 1980), we calculated Pearson’s correlation coefficients among all response variables used in each sub-set of models. Additionally, to aide in detection of collinearity and multicollinearity (i.e., correlations among three or more variables), we calculated condition indices (CIs). Small CIs (<10) indicate that collinearity has not caused adverse effects on estimated parameters (Belsley et al. 1980), whereas moderate to strong collinearity among variables exists with CIs of 30-100 (Belsley et al. 1980). When CIs indicated collinearity was present, we also calculated variance decomposition proportions

(VPs) to assist in identifying which variables were responsible for collinearity. Collinear variables were identified as those with VP > 0.5. Subsequently, those variables responsible for high CIs were assessed for degraded coefficients (Belsley et al. 1980).

Results

Trends in yellow perch juvenile abundance and spawning-stock size

During both early (1969-1985) and late (1986-2010) periods, juvenile abundance was highly variable in both western and central Lake Erie (Figure 4.2). No consistent trends were evident during the early (both basins p ≥ 0.38) or late periods (both basins p

≥ 0.12) in either basin. However, during both early (r = 0.63, p = 0.0007) and late periods (r = 0.57, p = 0.02), annual abundance estimates were correlated between west and central basins, indicating that large, regional processes may have affected larval and juvenile survival during both periods. Despite this correlation, a high degree of variance

112 remained unexplained in each basin, indicating more-local factors also played a role in regulating juvenile abundance.

Spawning-stock size was higher in the central basin than the west basin during both the early (p < 0.0001) and late (p = 0.001) period. Within each basin, spawning stock-size differed between early and late periods. Spawning-stock size decreased from the early to late period in the west basin (p = 0.002) but increased from the early to late period in the central basin (p = 0.07). Additionally, spawning-stock size increased with time in the central basin during the late-period (p = 0.003), with the highest observed sizes occurring near the end of our record (Figure 4.2).

Comparing multiple regression, stock-recruit, and null models

Although we used both multiple regression (i.e., linear models) and Ricker stock- recruit models for each basin in both early and late periods, simple multiple regression models performed substantially better than Ricker stock-recruit models across all basins and time periods. Ricker stock-recruit models were not included in the sub set of acceptable models (Δi ≤ 2) for either basin in early or late time periods. In the west basin, acceptable multiple regression models were about 3-7 times as likely as the best Ricker models in both the early and late periods (based on Akaike weights, ωi). In the central basin, acceptable multiple regression models were about 5-23 times as likely as the best

Ricker models in both the early and late periods. Additionally, null models were not included in the sub set of acceptable models (Δi ≤ 2) for either basin in early or late time periods. Subsequently, we focus on acceptable multiple regression models for assessing

113 the effects of explanatory variables on yellow perch juvenile abundance in the west and central basins.

West basin juvenile abundance

Overall, seven and nine acceptable models (Δi ≤ 2) were found for the early and late periods, respectively, in the west basin (Tables 3.2 and 3.3). Acceptable west basin models during the early period (1969-1985) explained 55-70% of the variation in yellow perch juvenile abundance, with all seven of these models having May temperature and winter ice cover in them (Table 4.2). Both May temperature and winter ice cover had positive coefficients indicating that long, cold winters with a high degree of ice cover, followed by warm May temperatures tended to correlate with high juvenile abundance the following fall (Table 4.2; Figure 4.3). In each acceptable model, these two variables explained the vast majority of the total variance, with other terms explaining only a small amount of additional variance in the west basin (Table 4.2). Accordingly, Akaike weights for May temperature (0.91) and winter ice cover (0.85) far exceeded those of all other variables during the early period (Figure 4.3). While other terms explained little additional variance, each of the four other explanatory variables (spring warming rate,

Maumee River discharge, white perch abundance, and spawning-stock size) were all included in at least one of the acceptable models. All coefficients for these variables were negative (Table 4.2).

Acceptable models during the late period (1986-2010) in the west basin explained

56-61% of the variance in yellow perch juvenile abundance. Maumee River discharge

114 and winter ice cover were included in all nine of the acceptable models in this basin with coefficients being positive for both variables in every model (Table 4.3). Both Maumee

River discharge and winter ice cover had positive coefficients indicating that long, cold winters with a high degree of ice cover, followed by wet springs with high Maumee River discharge tended to correlate with high fall juvenile abundance (Table 4.3; Figure 4.4).

In each of these nine models, these two variables explained the vast majority of the total variance, with other terms explaining only a small additional amount of variance (Table

4.3). Accordingly, Akaike weights for Maumee River discharge (0.99) and winter ice cover (0.97) far exceeded those of all other variables during the early period (Figure 4.3).

While other terms explained little additional variation, the other four explanatory variables (spring warming rate, May temperature, adult white perch abundance, and yellow perch spawning-stock size) were all included in at least one of the acceptable models. All signs for coefficients of these variables were in a direction consistent with expectations (MayTemp [+], WP [-], YPSpawn [-]), save for spring warming rate, which had a negative coefficient, counter to expectations (Table 4.3).

Central basin juvenile abundance

Overall, five acceptable models (Δi ≤ 2) were found for both the early and late periods in the central basin, respectively (Tables 3.4 and 3.5). Acceptable models during the early central basin period (1969-1985) explained from 49-61% of the variation in juvenile abundance. During this period in the central basin, winter ice cover and

Maumee River discharge were included in each of the five acceptable models and had the

115 highest overall Akaike weights, 0.84 and 0.74, respectively (Figure 4.3). In each acceptable model, winter ice cover had a positive coefficient, whereas Maumee River discharge had a negative coefficient (Table 4.4). While these two variables were always the most important predictor variables, yellow perch spawning-stock size and spring warming rate (both negative coefficients in all models) were included in 3 and 2 of the 5 acceptable models, respectively (Table 4.4). Akaike weights for these variables

(YPSpawn = 0.66; SpringWarm = 0.54; Figure 4.3) also indicated they were important in explaining additional variation in yellow perch juvenile abundance. The remaining two variables, adult white perch (negative coefficient) and May temperature were only included in one and zero of the acceptable models, respectively. Additionally, Akaike weights for these variables (WP = 0.33; MayTemp = 0.31; Figure 4.3) indicated that they were relatively less important than other variables in the central basin during the early period.

Acceptable models during the late central basin period (1986-2010) explained 51-

57% of the variation in juvenile abundance. During this period in the central basin, winter ice cover and Maumee River discharge were included in each of the five acceptable models (Table 4.5) and had the highest overall Akaike weights, 0.90 and 0.88, respectively (Figure 4.3). Counter to findings from the early period in the central basin,

Maumee River discharge had a positive coefficient in all acceptable models during the late period (Figure 4.4), coinciding with results from the west basin (Table 4.3), again similar to the west basin (Table 4.3). Winter ice cover maintained a positive coefficient in all acceptable models during the late period (Table 4.5; Figure 4.4). While these two

116 variables always were important predictors of juvenile abundance, spring warming rate

(negative coefficient in all models), May temperature (positive coefficients in all models), and adult white perch abundance (negative coefficient in all models) were included in 4, 3, and 3 of the 5 acceptable models, respectively (Table 4.5). Akaike weights for these variables (SpringWarm = 0.74; MayTemp = 0.50; AdtWPer = 0.45;

Figure 4.3) also indicated that these variables were important in explaining additional variation in yellow perch juvenile abundance. The remaining variable, yellow perch spawning-stock size was not included in any of the acceptable models. Additionally,

Akaike weights for yellow perch spawning-stock size (0.28) indicated it was relatively less important than other variables in the central basin during the late period.

Assessing correlation and collinearity

The only correlation detected among explanatory variables was a weak, negative correlation between spring (April-May) warming rate and winter ice cover in both the west (r = -0.35, p = 0.09) and central (r = -0.38, p = 0.06) basins during the more recent time period (1986-2010). This indicated winters with high ice cover tended to be followed by more slowly warming springs during this time. In assessing collinearity for the west basin early period (1969-1986), moderate collinearity (CI ≥ 30) was detected in the first (MayTemp, Ice, WP, SpringWarm; CI = 30.5) and fourth (MayTemp, Ice,

SpringWarm, WP, YPSpawn; CI = 31.6) ranked acceptable models (Table 4.2). Variance decomposition proportions (all VP ≥ 0.92) indicated that May temperature and intercept values were the collinear variables in each of these models. However, as coefficient

117 estimates for May temperature were relatively stable among all models (range = 0.34-

0.46; Table 4.2) and coefficient sign (+) was consistent, it does not appear that collinearity significantly degraded coefficient estimates.

Moderate collinearity also was detected in one of the acceptable models during both the early and late periods in the central basin. In the early central basin period moderate collinearity (CI = 31.1) was found for the second ranked acceptable model (i.e.,

Ice, RiverDisch, YPSpawn, SpringWarm; Table 4.4). Variance decomposition proportions (VP ≥ 0.96) indicated that yellow perch spawning-stock size and intercept values were responsible for collinearity. However, as coefficient estimates for spawning- stock size were stable among all models (range = -0.95 to -1.1; Table 4.4) and coefficient sign (-) was consistent, collinearity did not appear to dramatically degrad coefficient estimates. In the late central basin period, moderate collinearity (CI = 31.9) was found for the third ranked acceptable model (i.e., Ice, RiverDisch, SpringWarm, MayTemp, WP;

Table 4.5). Variance decomposition proportions (VP ≥ 0.86) indicated that May temperature was responsible for collinearity. However, as coefficient estimates for May temperature were stable among all models (range = 0.16-0.17; Table 4.5) and coefficient sign (+) was consistent, collinearity did not appear to significantly degrade coefficient estimates. No other instances of collinearity were detected. Therefore, although moderate collinearity was detected in a sub-set of my acceptable models, further investigation revealed that this weak collinearity did not likely degrade model results.

118

Discussion

Across basins (i.e., west and central) and time periods (i.e.., 1969-1985 and 1986-

2010), Lake Erie yellow perch juvenile abundance appeared to benefit from long, cold winters, signified by high ice cover. This finding supports other findings, generated from previous lab and field results (Hokanson 1977; Ciereszko et al. 1997; Farmer et al.

2013, Chapter 2 and 3), which showed that short, warm winters can negatively affect yellow perch reproductive success. Previous investigations (Ludsin 2000; Crane 2007) also found that long, cold winters prior to spawning positively correlated with high fall yellow perch juvenile abundance. Given the increasing frequency of short, warm winters on Lake Erie and that continued winter warming is expected (Kling et al. 2003; Hayhoe et al. 2010), yellow perch recruitment to the juvenile stage is likely to be severely limited by warm winters. Because yellow perch juvenile abundance is a strong predictor of both future recruitment to the fishery and a given cohort’s lifetime contribution to harvest in

Lake Erie (Chapter 2), the negative effects of short, warm winters may have serious economic and ecological implications. As yellow perch populations are currently near historically low levels in western and central Lake Erie (YPTG 2013), warm winters may act as an impediment to recovery for this system’s populations.

I also found evidence to suggest that the establishment of large invasive white perch populations in Lake Erie is negatively affecting yellow perch juvenile abundance in both basins. This finding is consistent with the hypothesis that increased white perch abundance has led to higher levels of predation on larval and juvenile yellow perch

(Madenjian et al.2000; Carreon-Martinez 2011; Ludsin et al. 2011), reducing survival and

119 negatively affecting yellow perch juvenile abundance. The negative effect of white perch was detected during both early and late time periods, in both basins, despite the fact that white perch were present in relatively low numbers during the early time period. While the direction of the white perch effect (-) was consistent with this proposed mechanism, the magnitude was lower than that of several other environmental variables (i.e., winter ice cover, May temperature, spring warming rate, river discharge) in some basins and periods. This finding may indicate that, while white perch negatively affect yellow perch larval and juvenile survival, other environmentally driven variables can be of greater importance in regulating juvenile abundance some of the time.

Dividing the yellow perch juvenile abundance series into separate time periods to account for changing nutrient levels in Lake Erie (1969-1985 = eutrophic conditions;

1986-2010 = oligotrophic conditions) offered insights into how changing productivity levels altered the importance of environmental variables affecting yellow perch recruitment. During the early period, Maumee River discharge was negatively related to yellow perch abundance in both basins. During this early period, which was characterized by high productivity (Dolan 1993; Neilson et al. 1995), high phytoplankton production (Nicholls and Hopkins 1993; Nicholls et al. 1999; Boegman et al. 2008), and, consequently, low water clarity, water clarity was not driven by Maumee River discharge

(Ludsin et al. 2011). Thus, during the early time period, Maumee River discharge should not be expected to affect predation rate on larval yellow perch (sensu Pangle et al. 2012).

Additionally, during this period of high productivity (due largely to high levels of non- point source loading of P), high Maumee River flows during the spring may have

120 provided excess nutrients to Lake Erie, contributing to reduced water quality (Peters et al.

2009) and may have increased the severity and duration of hypolimnetic hypoxia

(dissolved O2 < 2 mg/L; Rosa and Burns 1987). If so, hypoxia and low water quality may have negatively affected growth and survival of larval and juvenile yellow perch.

Indeed, it is noteworthy that the relative negative effect of Maumee River discharge during 1969-1985 was greater in the central basin, where the severity and duration of hypoxia would be expected to be greater, than the in the west basin, where hypoxic conditions occur less frequently and are of shorter duration (Zhou et al. 2013).

Counter to the early eutrophic period (1969-1985), results from the later oligotrophic period (1986-2010) indicate Maumee River discharge had a strong positive effect on yellow perch juvenile abundance in both the west and central basins of Lake

Erie. During this time, reduced phytoplankton production (Nicholls and Hopkins 1993;

Nicholls et al. 1999; Boegman et al. 2008) may have heightened the importance of sediment turbidity (largely driven by spring flows from the Maumee River) to provide a refuge from predation for yellow perch larvae (as proposed by Ludsin et al. 2011).

However, high Maumee River discharge events during this later oligotrophic period

(1986-2010) also contributed significantly to total lake-wide P loading, indicating the potential for bottom-up productivity increases that could also positively affect larval yellow perch (Ludsin et al. 2011). However, recent findings from field collections, laboratory experiments, and modeling suggest that top down (i.e., refuge from predation) effects of Maumee River discharge on yellow perch larval survival rates were more important than bottom up effects (productivity driven increases in zooplankton

121 availability; Reichert et al. 2010; Ludsin et al. 2011). The importance of river discharge following the oligotrophication of Lake Erie also highlights the potential importance of larval predators such as white perch, in regulating yellow perch juvenile abundance. In fact, the dramatic increase in white perch abundance during the late period (1986-2010) may have functionally elevated the importance of turbidity as a refuge from predation

(Ludsin et al. 2011). Along this same line of reasoning, the negative relationship between yellow perch stock size and juvenile abundance across both basins and time periods may highlight the importance of predation on larvae as a key process affecting juvenile abundance. Indeed, in other systems, high spawning-stock biomass has been linked to density-dependent reductions in recruitment due to cannibalism by adult spawners (Smith and Reay 1991; Link et al. 2012).

Spring temperature has been previously found to be positively related to yellow perch fall juvenile abundance (Clady 1976; Eshenroder et al. 1977; Henderson and

Nepszy 1988; Ludsin 2000). Herein, I also found that overall spring (May) temperature was positively correlated with yellow perch juvenile abundance. Warm spring temperatures have been found to increase larval yellow perch growth rates (Mills et al.

1989; Ludsin 2000), which likely leads to higher larval survival rates (Miller et al. 1988).

While I found May temperature to be a strong predictor in the west basin during the earlier, more productive time period, it was less important in the later oligotrophic time period when compared to other variables (i.e., winter ice cover, river discharge). As spring temperatures were relatively constant across both early and late time periods, this

122 reduction in importance may be due to a fundamental shift in the mechanisms regulating yellow perch as Lake Erie experienced oligotrophication.

While spring warming rate also was important, I found that slow warming correlated with high fall juvenile abundance. This runs counter to findings of previous studies, which had documented that rapid spring warming resulted in both greater egg survival to larval stages (Clady 1976) and in increased abundance of juvenile (age-0) and age-1 yellow perch in the Canadian waters of western Lake Erie (Henderson and Nepszy

1988). The negative relationship between spring warming and fall yellow perch juvenile abundance documented in this study could be a spurious correlation or it may be due to slight differences in methods for defining the relevant ‘spring’ period between studies.

Importantly, it should be noted that a weak negative correlation existed between spring warming rate and winter ice cover during the late period, where slow warming springs tended to follow winters with high ice cover. Given this, albeit weak, correlation, caution is warranted in interpretation of spring warming rate and yellow perch juvenile abundance relationships.

In conclusion, my findings support the hypothesis that increased frequency of short warm winters and the introduction of an invasive predator on yellow perch larvae

(i.e., white perch), have negatively affected yellow perch juvenile abundance in Lake

Erie. Yellow perch juvenile abundance was negatively related to white perch abundance across both time periods in both basins, consistent with the hypothesis of high white perch predation levels on larval yellow perch (Carreon-Martinez 2011; Ludsin et al.

2011). Additionally, findings indicate that the negative effect of white perch may be

123 mediated by river discharge, indicating that climate-driven precipitation patterns may dictate the relative magnitude of this negative effect.

The negative effect of warm winters on yellow perch juvenile abundance appears to be consistent over 42 years of data, and has persisted throughout a large-scale nutrient- driven regime shift (Dolan 1993; Nicholls et al. 1999; Ludsin et al. 2001; Tyson and

Knight 2001) and restructuring of the food web due to numerous introductions of invasive species (Mills et al. 1993; Ricciardi and MacIsaac 2000). When these results are considered along with the evidence from the lab and field that warm winters negatively affect yellow perch reproductive success (Hokanson 1977; Farmer et al. 2013; Chapters 2 and 3), it seems highly likely that warm winters have affected yellow perch populations by reducing reproductive success. Given the increased frequency of warm winters in recent years and the likelihood of continued winter warming across the Great Lakes region (Kling et al. 2003; Hayhoe et al. 2010), it appears that, in the absence of any life- history change or evolution, climate change may have begun to limit the potential for

Lake Erie yellow perch population growth, and may continue to do so into the foreseeable future.

124

References

Baumann, H., S. C. Talmage, and C. J. Gobler. 2012. Reduced early life growth and survival in a fish in direct response to increased carbon dioxide. Nature Climate Change 2:38-41.

Belsley, D. A., E. Kuh, and R. E. Welsch. 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. Wiley, Hoboken, New Jersey.

Boegman, L., M. R. Loewen, D. A. Culver, P. F. Hamblin, and M. N. Charlton. 2008. Spatial-dynamic modeling of algal biomass in Lake Erie: relative impacts of dreissenid mussels and nutrient loads. Journal of Environmental Engineering- Asce 134:456-468.

Bolsenga, S. J., and C. E. Herdendorf. 1993. Lake Erie and Lake St. Clair Handbook. Wayne State University Press, Detriot.

Brook, B. W., N. S. Sodhi, and C. J. A. Bradshaw. 2008. Synergies among extinction drivers under global change. Trends in Ecology & Evolution 23:453-460.

Burnham, K. P., and D. R. Anderson. 2002. Model Selection and Multimodel Inference. Springer-Verlag, New York.

Carreon-Martinez, L. 2011. Larval Yellow Perch (Perca flavescens) and predation in the Maumee and Detroit River Plumes in the Western Basin of Lake Erie. PhD dissertation. University of Windsor, Windsor, Onatrio

Carreon-Martinez, L., T. B. Johnson, S. A. Ludsin, and D. D. Heath. 2011. Utilization of stomach content DNA to determine diet diversity in piscivorous fishes. Journal of Fish Biology 78:1170-1182.

Casselman, J. M., and C. A. Lewis. 1996. Habitat requirements of northern pike (Esox lucius). Canadian Journal of Fisheries and Aquatic Sciences 53:161-174.

125

Charlebois, P. M., L. D. Corkum, D. J. Jude, and C. Knight. 2001. The round goby (Neogobius melanostomus) invasion: Current research and future needs. Journal of Great Lakes Research 27:263-266.

Clady, M. D. 1976. Influence of temperature and wind on survival of early stages of yellow perch, Perca flavescens. Journal of the Fisheries Research Board of Canada 33:1887-1893.

Crane, V. 2007. Lower trophic level and climate influences on western Lake Erie fish recruitment 1988 through 2005. Master's thesis. The Ohio State University, Columbus.

Cushing, D. H. 1990. Planktonic production and year-class strength in fish populations - an update of the match mismatch hypothesis. Advances in Marine Biology 26:249-293.

Dolan, D. M. 1993. Point-source loadings of phosphorus to Lake Erie: 1986-1990. Journal of Great Lakes Research 19:212-223.

Doney, S. C., M. Ruckelshaus, J. E. Duffy, J. P. Barry, F. Chan, C. A. English, H. M. Galindo, J. M. Grebmeier, A. B. Hollowed, N. Knowlton, J. Polovina, N. N. Rabalais, W. J. Sydeman, and L. D. Talley. 2012. Climate change impacts on marine ecosystems. Pages 11-37 in C. A. Carlson, and S. J. Giovannoni, editors. Annual Review of Marine Science, volume 4.

Edwards, M., and A. J. Richardson. 2004. Impact of climate change on marine pelagic phenology and trophic mismatch. Nature 430:881-884.

Einum, S., and I. A. Fleming. 2000. Highly fecund mothers sacrifice offspring survival to maximize fitness. Nature 405:565-567.

Eshenroder, R. L. 1977. Effect on intensified fishing, species changes, and spring water temperature on yellow perch, Perca flavescens, in Saginaw Bay. Journal of the Fisheries Research Board of Canada 34:1830-1838.

Farmer, T. M., K. Dabrowski, E. A. Marschall, and S. A. Ludsin. 2013. Winter warming effects on yellow perch reproduction and recruitment. Great Lakes Fishery Commission, Ann Arbor, MI.

126

Forsythe, P. S., J. C. Doll, and T. E. Lauer. 2012. Abiotic and biotic correlates of yellow perch recruitment to age-2 in southern Lake Michigan, 1984-2007. Fisheries Management and Ecology 19:389-399.

Francis, J. T., S. R. Robillard, and J. E. Marsden. 1996. Yellow perch management in Lake Michigan: A multi-jurisdictional challenge. Fisheries 21:18-20.

Fuller, K., H. Shear, and J. Wittig. 1995. The Great Lakes: An Environmental Atlas and Resource Book. Government of Canada and U. S. Environmental Protection Agency, Chicago.

Griffiths, R. W., D. W. Schloesser, J. H. Leach, and W. P. Kovalak. 1991. Distribution and dispersal of the zebra mussel (Dreissena polymorpha) in the Great Lakes region. Canadian Journal of Fisheries and Aquatic Sciences 48:1381-1388.

Grosholz, E. 2002. Ecological and evolutionary consequences of coastal invasions. Trends in Ecology & Evolution 17:22-27.

Hayhoe, K., J. VanDorn, T. Croley, II, N. Schlegal, and D. Wuebbles. 2010. Regional climate change projections for Chicago and the US Great Lakes. Journal of Great Lakes Research 36:7-21.

Hecky, R. E., R. E. H. Smith, D. R. Barton, S. J. Guildford, W. D. Taylor, M. N. Charlton, and T. Howell. 2004. The nearshore phosphorus shunt: a consequence of ecosystem engineering by dreissenids in the Laurentian Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 61:1285-1293.

Hellmann, J. J., J. E. Byers, B. G. Bierwagen, and J. S. Dukes. 2008. Five potential consequences of climate change for invasive species. Conservation Biology 22:534-543.

Henderson, B. A., and S. J. Nepszy. 1988. Recruitment of yellow perch (Perca flavescens) affected by stock size and water temperature in Lakes Erie and St Clair 1965-85. Journal of Great Lakes Research 14:205-215.

Hilborn, R., and C. J. Walters. 1992. Quantitative Fisheries Stock Assessment Choice Dynamics and Uncertainty. Routledge, Chapman and Hall, Inc, New York, New York; London, England.

Johansson, O. E., C. Dumitru, and D. M. Graham. 1999. Estimation of zooplankton mean length for use in an index of fish community structure and its application to Lake Erie. Journal of Great Lakes Research 25:179-186. 127

Johnson, T. B., D. B. Bunnell, and C. T. Knight. 2005. A potential new energy pathway in central Lake Erie: the round goby connection. Journal of Great Lakes Research 31:238-251.

Kocovsky, P. M., and C. T. Knight. 2012. Morphological evidence of discrete stocks of yellow perch in Lake Erie. Journal of Great Lakes Research 38:534-539.

Krieger, K. A., D. W. Schloesser, B. A. Manny, C. E. Trisler, S. E. Heady, J. J. H. Ciborowski, and K. M. Muth. 1996. Recovery of burrowing mayflies (Ephemeroptera: Ephemeridae: Hexagenia) in western Lake Erie. Journal of Great Lakes Research 22:254-263.

Link, J. S., S. M. Lucey, and J. H. Melgey. 2012. Examining cannibalism in relation to recruitment of silver hake Merluccius bilinearis in the U.S. northwest Atlantic. Fisheries Research 114:31-41.

128

Ludsin, S. A., and R. A. Stein. 2001. Species interactions among young-of-year fishes of Lake Erie. Ohio Department of Natural Resources, Division of Wildlife, Project F-69-P, Final Report, Columbus, OH.

Madenjian, C. P., R. L. Knight, M. T. Bur, and J. L. Forney. 2000. Reduction in recruitment of white bass in Lake Erie after invasion of white perch. Transactions of the American Fisheries Society 129:1340-1353.

Makarewicz, J. C., P. Bertram, and T. W. Lewis. 1998. Changes in phytoplankton size- class abundance and species composition coinciding with changes in water chemistry and zooplankton community structure of Lake Michigan, 1983 to 1992. Journal of Great Lakes Research 24:637-657.

McCormick, M. J., and G. L. Fahnenstiel. 1999. Recent climatic trends in nearshore water temperatures in the St. Lawrence Great Lakes. Limnology and Oceanography 44:530-540.

Myers, R. A., and J. R. Bence. 2001. The 2001 assessment of perch in Lake Erie; a review. Great Lakes Fishery Commission, Lake Erie Yellow Perch Task Group, Ann Arbor, MI.

Neilson, M., S. L’Italien, V. Glumac, D. Williams, and P. Bertram. 1995. Nutrients: trends and system response. State of the Lake Erie Conference (SOLEC) working group paper, EPA 905-R-95-015. U.S.E.P.A., Chicago.

Neuheimer, A. B., R. E. Thresher, J. M. Lyle, and J. M. Semmens. 2011. Tolerance limit for fish growth exceeded by warming waters. Nature Climate Change 1:110-113.

Nicholls, K. H. 1999. Evidence for a trophic cascade effect on north-shore western Lake Erie phytoplankton prior to the zebra mussel invasion. Journal of Great Lakes Research 25:942-949.

129

Nicholls, K. H., and G. J. Hopkins. 1993. Recent changes in Lake Erie (north shore) phytoplankton - cumulative impacts of phosphorus loading reductions and the zebra mussel introduction. Journal of Great Lakes Research 19:637-647.

Ohio Division of Wildlife (ODOW). 2011. Ohio Lake Erie's Fisheries, 2010. Annual status report. Ohio Department of Natural Resources, Division of Wildlife, Lake Erie Fisheries Units, Fairport and Sandusky.

Ohio Environmental Protection Agency (OEPA). 2013. Ohio Lake Erie Phosphorus Task Force II Final Report. Columbus, OH. 109p.

Paine, R. T., M. J. Tegner, and E. A. Johnson. 1998. Compounded perturbations yield ecological surprises. Ecosystems 1:535-545.

Parrish, D. L., and F. J. Margraf. 1990. Interactions between white perch (Morone americana) and yellow perch (Perca flavescens) in Lake Erie as determined from feeding and growth. Canadian Journal of Fisheries and Aquatic Sciences 47:1779- 1787.

Portner, H. O., and R. Knust. 2007. Climate change affects marine fishes through the oxygen limitation of thermal tolerance. Science 315:95-97.

Rahel, F. J., and J. D. Olden. 2008. Assessing the effects of climate change on aquatic invasive species. Conservation Biology 22:521-533.

Redman, R. A., S. J. Czesny, J. M. Dettmers, M. J. Weber, and D. Makauskas. 2011. Old Tales in Recent Context: Current Perspective on Yellow Perch Recruitment in Lake Michigan. Transactions of the American Fisheries Society 140:1277-1289.

Ricciardi, A., and H. J. MacIsaac. 2000. Recent mass invasion of the North American Great Lakes by Ponto-Caspian species. Trends in Ecology & Evolution 15(2):62- 65.

Roseman, E. F., E. L. Mills, J. L. Forney, and L. G. Rudstam. 1996. Evaluation of competition between age-0 yellow perch (Perca flavescens) and gizzard shad (Dorosoma cepedianum) in Oneida Lake, New York. Canadian Journal of Fisheries and Aquatic Sciences 53:865-874.

Ruiz, G. M., P. Fofonoff, A. H. Hines, and E. D. Grosholz. 1999. Non-indigenous species as stressors in estuarine and marine communities: Assessing invasion impacts and interactions. Limnology and Oceanography 44:950-972. 130

Sepulveda-Villet, O. J., A. M. Ford, J. D. Williams, and C. A. Stepien. 2009. Population genetic diversity and phylogeographic divergence patterns of the yellow perch (Perca flavescens). Journal of Great Lakes Research 35:107-119.

Smith, C., and P. Reay. 1991. Cannibalism in teleost fish. Reviews in Fish Biology and Fisheries 1:41-64.

Stachowicz, J. J., J. R. Terwin, R. B. Whitlatch, and R. W. Osman. 2002. Linking climate change and biological invasions: Ocean warming facilitates nonindigenous species invasions. Proceedings of the National Academy of Sciences of the United States of America 99:15497-15500.

Stapanian, M. A., M. T. Bur, and J. V. Adams. 2007. Temporal trends of young-of-year fishes in Lake Erie and comparison of diel sampling periods. Environmental Monitoring and Assessment 129:169-178.

Tyson, J. T., T. B. Johnson, C. T. Knight, and M. T. Bur. 2006. Intercalibration of research survey vessels on Lake Erie. North American Journal of Fisheries Management 26:559-570.

Tyson, J. T., and R. L. Knight. 2001. Response of yellow perch to changes in the benthic invertebrate community of western Lake Erie. Transactions of the American Fisheries Society 130:766-782.

Vanderploeg, H. A., J. R. Liebig, W. W. Carmichael, M. A. Agy, T. H. Johengen, G. L. Fahnenstiel, and T. F. Nalepa. 2001. Zebra mussel (Dreissena polymorpha) selective filtration promoted toxic Microcystis blooms in Saginaw Bay (Lake Huron) and Lake Erie. Canadian Journal of Fisheries and Aquatic Sciences 58:1208-1221.

Vitousek, P. M. 1994. Beyond global warming: ecology and global change. Ecology 75:1861-1876.

Walther, G.-R., A. Roques, P. E. Hulme, M. T. Sykes, P. Pysek, I. Kuehn, M. Zobel, S. Bacher, Z. Botta-Dukat, H. Bugmann, B. Czucz, J. Dauber, T. Hickler, V. Jarosik, M. Kenis, S. Klotz, D. Minchin, M. Moora, W. Nentwig, J. Ott, V. E. Panov, B. Reineking, C. Robinet, V. Semenchenko, W. Solarz, W. Thuiller, M. Vila, K. Vohland, and J. Settele. 2009. Alien species in a warmer world: risks and opportunities. Trends in Ecology & Evolution 24:686-693.

Wang, J., X. Bai, H. Hu, A. Clites, M. Colton, and B. Lofgren. 2012. Temporal and spatial variability of Great Lakes ice cover, 1973-2010. Journal of Climate 25:1318-1329.

131

Weisgerber, K. M. 1999. Lower trophic level dynamics in western Lake Erie: changes in biomass, clearance, and nutrient excretion rates in crustacean zooplankton versus zebra mussels. Master’s thesis. The Ohio State University, Columbus.

Wilberg, M. J., J. R. Bence, B. T. Eggold, D. Makauskas, and D. F. Clapp. 2005. Yellow Perch Dynamics in Southwestern Lake Michigan during 1986–2002. North American Journal of Fisheries Management 25:1130-1152.

Wright, P. J., and E. A. Trippel. 2009. Fishery-induced demographic changes in the timing of spawning: consequences for reproductive success. Fish and Fisheries 10:283-304.

Yellow Perch Task Group (YPTG). 2013. Report of the Lake Erie yellow perch task group. Great Lakes Fishery Commission, Ann Arbor, MI.

132

1969 - 1985 1986 - 2010 Variable Abbreviation Mean ± 1SE Range Mean ± 1SE Range Juvenile (Age-0) Abundance (CPUE) - WB 74 ± 21 2 - 287 110 ± 39 3 - 974 - CB 15 ± 4 0 - 54 22 ± 5 0 - 91

Ap ril-May Air Warming Rate (°C/d) Spring Warm 0.19 ± 0.02 0.05 - 0.35 0.20 ± 0.01 0.09 - 0.38

Mean April-May Maumee River Discharge (m3/s) RiverDisch 490 ± 40 165 - 770 460 ± 29 208 - 738

Mean February-March Ice Cover (% lake area) Ice 65 ± 12 0 - 100 28 ± 9 0 - 100

133

Mean May Air Temperature (°C) MayT emp 14.8 ± 0.4 12.5 - 18.2 15.4 ± 0.4 11.6 - 19.5

Age -3+ Spawning Stock (millions of fish) YPSp awn - WB 44.8 ± 4.5 15.8 - 92.8 23.9 ± 3.0 4.3 - 60.5 - CB 38.5 ± 3.5 18.6 - 67.8 60.5 ± 9.4 11.4 - 218.7

Age -1+ White Perch Abudance (CPUE) WP - WB 1 ± 1 0 - 13 13 ± 3 0 - 70 - CB 5 ± 4 0 - 73 36 ± 6 2 - 109

Table 4.1. Explanatory and response variables used in our investigation of factors affecting yellow perch juvenile (age-0) abundance in the west (WB) and central basins (CB) of Lake Erie, 1969-2010 (CPUE: catch per unit effort; # individuals per trawling minute).

133

YP Spring May River Intercept Ice WP 2 Spawn Warm Temp Disch p R Δi ωi coefficient -2.49 1.18 -4.74 0.43 -0.44 0.005 0.68 0.00 0.12 (sr2) (0.21) (0.07) (0.31) (0.09)

coefficient -2.27 1.35 0.35 0.053 0.004 0.63 0.13 0.11 (sr2) (0.28) (0.23) (0.12)

coefficient -2.82 1.08 -5.78 0.46 0.004 0.63 0.45 0.10 (sr2) (0.17) (0.12) (0.34)

coefficient -1.81 -0.01 1.23 -5.75 0.43 -0.46 0.012 0.70 1.42 0.06 (sr2) (0.03) (0.21) (0.09) (0.28) (0.09)

coefficient -2.61 1.27 0.36 0.004 0.55 1.51 0.06

134 (sr2) (0.28) (0.27)

coefficient -2.21 -0.01 1.12 -6.73 0.46 0.01 0.64 1.93 0.05 (sr2) (0.02) (0.17) (0.13) (0.32)

coefficient -1.78 1.42 0.34 -0.0008 -0.51 0.01 0.64 1.96 0.05 (sr2) (0.3) (0.21) (0.01) (0.12)

Table 4.2. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the west basin of Lake Erie, 1969- 2 1985. For each model, the overall p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semipartial correlation coefficient (sr2: the amount of unique variance it explains) also is presented.

134

YP Spring May River Intercept Ice WP 2 Spawn Warm Temp Disch p R Δi ωi coefficient 2.03 -0.02 1.15 0.004 0.0003 0.59 0.00 0.13

(sr2) (0.05) (0.24) (0.30)

coefficient 1.38 1.3 0.005 0.0001 0.56 0.14 0.12

(sr2) (0.26) (0.30)

coefficient 1.76 1.45 0.005 -0.44 0.0003 0.59 0.20 0.12

(sr2) (0.28) (0.28) (0.03)

coefficient 2.29 -0.01 1.3 0.004 -0.38 0.0006 0.61 0.56 0.10

(sr2) (0.03) (0.26) (0.29) (0.03)

coefficient 2.48 -0.02 1.07 -1.75 0.004 0.0008 0.60 1.35 0.07

2 135 (sr ) (0.05) (0.22) (0.01) (0.32)

coefficient 1.18 1.32 0.01 0.005 0.0005 0.56 1.77 0.05

(sr2) (0.25) (0.0002) (0.31)

coefficient 1.83 -0.02 1.16 0.01 0.04 0.0009 0.59 1.78 0.05

(sr2) (0.05) (0.23) (0.0004) (0.31)

coefficient 2.03 1.39 -1.16 0.04 -0.42 0.001 0.59 1.81 0.05

(sr2) (0.25) (0.003) (0.31) (0.03)

coefficient 1.65 1.46 0.01 0.05 -0.44 0.001 0.59 2.00 0.05

(sr2) (0.27) (0.001) (0.29) (0.03)

Table 4.3. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the west basin of Lake Erie, 1986- 2 2010. For each model, the overall p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semi partial correlation coefficient (sr2: the amount of unique variance it explains) also is presented.

135

YP Spring May River Intercept Ice WP 2 Spawn Warm Temp Disch p R Δi ωi coefficient 6.87 -1.01 1.25 -0.003 0.009 0.57 0.0 0.11 (sr2) (0.1) (0.27) (0.2)

coefficient 7.14 -0.95 1.12 -3.86 -0.003 0.02 0.61 0.5 0.09 (sr2) (0.12) (0.28) (0.06) (0.15)

coefficient 3.27 1.45 -0.003 0.009 0.49 1.1 0.06 (sr2) (0.31) (0.18)

coefficient 7.26 -1.1 1.14 -0.003 -0.17 0.02 0.60 1.2 0.06 (sr2) (0.13) (0.24) (0.20) (0.03)

136

coefficient 3.82 1.29 -4.32 -0.003 0.02 0.54 1.4 0.06

(sr2) (0.34) (0.06) (0.14)

Table 4.4. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the central basin of Lake Erie, 1969- 2 1985. For each model, the overall p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semipartial correlation coefficient (sr2: the amount of unique variance it explains) also is presented.

136

YP Spring May River Intercept Ice WP 2 Spawn Warm Temp Disch p R Δi ωi coefficient -0.64 1.29 -6.41 0.16 0.003 0.0024 0.55 0.00 0.13 sr2 (0.20) (0.12) (0.05) (0.18)

coefficient 1.93 1.1 -6.39 0.003 0.0017 0.51 0.28 0.11 sr2 (0.19) (0.14) (0.18)

coefficient -0.17 1.24 -5.5 0.17 0.003 -0.57 0.0042 0.57 0.66 0.09 sr2 (0.20) (0.08) (0.07) (0.19) (0.03)

coefficient 2.5 1.04 -5.6 0.003 -0.5 0.0036 0.53 1.16 0.07 sr2 (0.20) (0.11) (0.19) (0.03)

137

coefficient -1.25 1.5 0.17 0.003 -0.8 0.0047 0.51 1.87 0.05

sr2 (0.23) (0.05) (0.18) (0.05)

Table 4.5. Acceptable (Δi ≤ 2) multiple regression models for predicting yellow perch juvenile abundance in the central basin of Lake Erie, 1986- 2 2010. For each model, the overall p-value, total variation explained (R ), the AICc difference (Δi: difference in AICc between the given model and the model with the lowest AICc score), and the Akaike weight (ωi: the relative likelihood or probability that the given model is the 'best' model among all models investigated) are given. Variable abbreviations are presented in Table 4.1. Each independent variable's semipartial correlation coefficient (sr2: the amount of unique variance it explains) also is presented.

137

Figure 4.1. Winter (February-March) ice cover (% areal coverage lakewide) and Lake Erie age-1+ white perch abundance (catch per unit effort [CPUE] for the west and central basins, 1969-2010). The vertical dashed line indicates the location where data were split for analysis (1969-1985 vs. 1986-2010).

138

Figure 4.2. Yellow perch juvenile (age-0) abundance (bars = catch-per-unit-effort [CPUE]) and spawning stock size (lines = age-3+ population size in millions of fish) for the west and central basins of Lake Erie. Juvenile abundance was estimated from Ohio Division of Wildlife (DOW) bottom trawl surveys conducted in fall 1969-2010. Age-3+ population size was obtained from Lake Erie Yellow Perch Task Group reports (YPTG 2013). The vertical dashed line indicates the location where data were split for analysis (1969-1985 vs. 1986-2010).

139

140

Figure 4.3. Akaike variable weights (ωi) for explanatory variables used in multiple regression models to explain yellow perch juvenile abundance in the west and central basins of Lake Erie, 1969-1985 and 1986-2010. Akaike variable weights range from 0-1, and are the relative importance of each explanatory variable in each sub set of models.

140

141

Figure 4.4. Yellow perch juvenile (age-0) abundance catch per unit effort (CPUE) in western Lake Erie for the early (1969-1985) and late (1986-2010) periods. For each period, ice cover classification is indicated and juvenile abundance is plotted against both May air temperature (°C) and Maumee River discharge (m3/s).

141

142

Figure 4.5. Yellow perch juvenile (age-0) abundance catch per unit effort (CPUE) in central Lake Erie for the early (1969-1985) and late (1986-2010) periods, plotted against Maumee River discharge (m3/s). During both periods, ice cover classification also is indicated.

142

APPENDIX A

PREDICTING YELLOW PERCH JUVENILE ABUDANCE IN LAKE ERIE GIVEN

FUTURE, PROJECTED CHANGES IN CLIMATE

Introduction

Field observations from Lake Erie suggest that winter temperature, in part, drives yellow perch year-class strength (as determined from annual indices of juvenile abundance in October; Chapter 2 and 4), which is a strong predictor of future recruitment to the fishery at age-2 (Chapter 1; Ludsin 2000; YPTG 2011). For example, during the period 1973-2010, a negative threshold relationship was shown to exist between winter temperature and yellow perch year-class strength (Chapter 1), wherein yellow perch year- classes were consistently low after warm winters in both western and central Lake Erie.

Additionally, Crane (2007) explored numerous abiotic (n = 13) and biotic (n = 7) correlates of yellow perch and walleye recruitment across four different habitat types in western Lake Erie, finding that the previous winter’s temperature occurred in all of the best models for yellow perch recruitment and the majority of models for walleye recruitment. Similar to our own data, winter temperature was negatively related to year- class strength for both yellow perch and walleye with the strongest year-classes following long, cold winters (Crane 2007).

143

Using findings from previously conducted laboratory experiments (Chapters 2 and 3), which quantified the role of winter duration and female condition entering winter on yellow perch reproductive success, we sought to use a population-modeling approach to understand what effect, if any, the relationship between winter duration/female condition and egg quality might have on yellow perch recruitment in both western and central Lake Erie. More specifically, we used our model to both explain the previously documented historical (1975–2010) relationships between winter temperature and yellow perch year-class strength (Ludsin 2000; Crane 2007; Chapter 2), as well as how predicted climate change would be expected to influence Lake Erie yellow perch recruitment through effects on reproduction.

Methods

Model overview

Our goal was to derive a model that used annually varying environmental conditions to explain annually varying historical (1975-2010) yellow perch recruitment

(as measured by year-class strength) and then to use the model to ask how projected changes in environmental conditions (e.g., winter temperature) might affect future (2046-

2065) yellow perch recruitment. Initially, we planned to include the effect of summer hypoxia on fall female condition in this model. However, finding in our experiment that fall female condition did not have a strong negative effect on reproductive output or quality (see Chapter 3), our model only considered the effect of winter duration on juvenile abundance.

144

As a first step, we used recent Lake Erie adult yellow perch population size structure data, an index of juvenile yellow perch abundance (year-class strength), and winter temperature data to assess the ability of our model to explain annual variation in past recruitment. Afterwards, we used the model to forecast Lake Erie yellow perch recruitment under several predicted climate scenarios. This entire modeling process consisted of four steps: 1) calculating historical larval production; 2) relating historical larval abundance to year-class strength; 3) creating probability distributions of recruitment given current conditions; and 4) creating future probability distributions of recruitment given future, projected climate conditions.

Calculating historical larval production

We related our experimental data to Lake Erie agency-derived estimates of age- specific, adult yellow perch size distributions (in October; YPTG 2011) to predict spring egg production and quality (i.e., egg size, which was found to be important in our experiment). Our experimental data allowed us to use an index of annual winter duration

(number of days with water temperature ≤ 5°C at the Cleveland, OH water intake; 1994 –

2010) to predict egg size the following spring (egg mass = 1.55 + 0.014[# d water temp ≤

5°C]; R2 = 0.50). To obtain estimates of egg size for early years in our dataset (1975–

1993), for which we did not have water temperature data, we developed a relationship between the number of days with mean daily air temperature ≤ 0°C each winter (derived from a regional index of mean daily air temperature averaged from Cleveland-Hopkins

International Airport and Toledo Express Airport, OH) and the number of days water

145 temperature was less than 5°C (# d water temp ≤ 5°C = 41.2 + 1.1[# d air temp ≤ 0°C];

R2 = 0.58). With an annual measure of egg mass from 1975–2010 in hand, our experimental data allowed us to calculate annual hatching success as a function of egg mass (see Chapter 1; % hatching = -0.50 + 0.37[egg mass]; R2 = 0.68). Finally, by combining our estimate of total egg production with annual hatching success, we calculated an annual index of total larval production (TLP) for the period 1975-2010.

Relating historical larval abundance to year-class strength

Using historical (1975-2010) annual estimates of TLP and estimates of year-class strength (i.e., age-0 yellow perch catch-per-unit-effort [CPUE] during October; YPTG

2011), we tested the hypothesis that high population TLP results in strong year-classes.

Originally, we planned to test this hypothesis for both Management Unit 2 and 3 in Lake

Erie, which are located in the central basin. However, due to a low frequency (and, in some cases, lack) of sampling in Management Unit 3 across years, this analysis was conducted only for Management Unit 2.

Previous research has demonstrated that abundance during August and October of the first year of life is a strong predictor of future recruitment to the fishery at age-2

(Chapter 2; Ludsin 2000; YPTG 2011). Indeed, we found a positive relationship between

TLP and recruitment in Management Unit 2 during 1975–2010 (Figure A4; CPUE age-0 yellow perch = 8.74 + 0.54(TLP in millions); P = 0.02; R2 = 0.16). With this relationship added to our model, we gained the ability to predict recruitment the following year as a function of adult total egg production (based on adult size distributions in October) and

146 winter duration. While we found a significant positive relationship between TLP and recruitment, this relationship had a great deal of variance associated with it (Figure A4); thus, we described it with a probabilistic model. In simulations with the resulting model, this part of the life cycle was assumed to have a stochastic relationship between the input of eggs and the output of recruitment (year-class strength).

Creating current probability distributions of recruitment

The goal of this step was to produce a probability distribution of annual recruitment expected under current environmental conditions. Using our winter-duration index based on Lake Erie water temperatures (described above) and annual estimates of total egg production, we drew randomly from the observed distribution of winter conditions and from the observed distribution of total egg production. We applied the model that we derived in Steps 1-2 above to these input data to get recruitment as output.

To model a stochastic representation of the transition from larvae to recruits, we used quantile regression (Cade et al. 1999) to define the 1st, 25th, 50th, 75th, 90th and 99th quantiles of the TLP to recruits relationship (Figure A4). Using these quantiles to give the probability of achieving specific levels of recruitment each year given TLP, we generated a recruitment probability distribution from 10,000 repetitions of this simulation process.

147

Creating future probability distributions of recruitment

Ultimately, we used this model to ask how changes in environmental conditions predicted by climate models should change our expectations about yellow perch recruitment. We used downscaled and bias-corrected climate predictions of mean daily air temperature at mid-century (2046–2065) from three General Circulation Models used in IPCC (2007) (CCCMA: Canadian Centre for Climate Modeling and Analysis; GFDL:

Geophysical Fluid Dynamics Laboratory; IPSL: Institute Simon Pierre Simon Leplace).

We used downscaled air temperature predictions from each of the 1/8° grids that were nearest Cleveland-Hopkins International Airport and Toledo Express Airport. Exactly as we created a historical index of air temperatures in Step 1, we derived a regional index of future mean daily air temperature by averaging daily predictions from both of these grids for each model. We ran each model under three future greenhouse gas emissions scenarios (A1B: high; A2: moderate; and B1: low; IPCC 2007). Within each emissions scenario, daily regional predictions were averaged across all three models to generate a single index of the number of days with mean daily air temperatures < 0°C during 2046–

2065. We altered our original distributions of Lake Erie winter water temperatures (using the previously defined relationship between air and water temperatures from Step 1) to reflect changes expected by 2046–2065 under each emission scenario. From these new distributions, we again drew randomly from winter environmental conditions and applied the model that we derived in steps 1-3 to these input data, exactly as done with the current environmental conditions in modeling Step 3 (using current levels of total egg

148 production). We repeated this process for 10,000 repetitions to create probability distributions of expected annual recruitment under future climate conditions.

Results

Our index of winter duration (number of days with water temperature ≤ 5°C) significantly differed from the current period (1975–2010) to all three mid-century scenarios (Figure A1). Winters were longer in the current time period than in any of the mid-century scenarios. Among mid-century scenarios, the highest emission scenario

(A1B) had the shortest mean winter duration, whereas the moderate and low emission scenarios (A2 and B1, respectively) had similar mean winter durations (Figure A1).

Predicted hatching success (determined as a function of winter duration and egg mass) also differed between current and future scenarios (ANOVA P < 0.0001, R2 = 0.42;

Figure A2); being highest in the current period (64 ± 9% [mean ± std. dev.]) and lowest in the A1B scenario (47 ± 7%). Predicted hatching success was similar among the A2

(53 ± 7%) and B1 (54 ± 9%) emission scenarios, which was not surprising given the similar winter durations predicted for these emissions scenarios.

Distributions of TLP (generated from 10,000 repetitions using randomly selected values from distributions of winter duration and total egg production) were positively skewed for current and all predicted future mid-century scenarios (Figure A3). However, careful examination of these distributions reveals subtle differences. The current distribution of TLP (1975–2010) ranged from 0.5 to 158 million larvae and had a mode of 17.4 million. However, predicted distributions of TLP from all emissions scenarios

149 had an abbreviated range, compared to the current distribution (A1B: 0.4 – 116; A2: 0.2 –

123; B1: 0.4 – 137 million larvae), with the smallest range occurring in the highest emissions scenarios (A1B; Figure A3). Also, the mode of the distribution shifted to lower values in all predicted future distributions of TLP (A1B: 12.5; A2: 10.6; B1: 11.8 million larvae), compared to the current distribution (Figure A3).

As evidence that high TLP is positively related to recruitment success, we documented a positive relationship between current annual estimates of TLP and recruitment in management unit 2 from 1975–2010 (Figure A4; CPUE age-0 yellow perch = 8.74 + 0.54(TLP in millions); P = 0.02; R2 = 0.16), although this relationship was highly variable. Generally, the potential for strong year-classes increased as TLP increased. Specifically, the years with the highest recruitment followed springs with the highest levels of TLP. However, we found the potential for low recruitment across all levels of TLP (Figure A4).

Using estimates of TLP generated under each of our 10,000 random repetitions for each climate scenario, we generated predicted recruitment using the stochastic transition from larvae to recruits (depicted by quantiles in Figure A4). Probability distributions of predicted recruitment are bimodal, with a high mode at extremely low

CPUE and a lower mode at intermediate levels of CPUE (Figure A5). While distributions of predicted recruitment are qualitatively quite similar among current and future winter conditions, there is a slight truncation of recruitment probability distributions with increasing emissions scenarios. For example, under current climate conditions, mean recruitment is 27.6 CPUE with a range of 0 – 212 CPUE (Figure A5).

150

In the high emissions scenario (A1B) the mean (20.2 CPUE) and range (0 – 185 CPUE) of predicted recruitment are lower, while in the moderate (A2: mean = 23.1; range = 0 –

190 CPUE) and low (B1: mean = 23.1; range = 0 – 205 CPUE) emissions scenarios, the mean and range of predicted recruitment are intermediate to those in the current and high emissions scenario. To further clarify the importance of these slight differences among probability distributions, we classified each predicted year-class as either failed (CPUE <

5), weak (CPUE = 5 – 14), moderate (CPUE = 15 – 29), strong (CPUE = 30 – 60), or very strong (CPUE > 60). These classifications were derived based on the observed relationship between age-0 CPUE and the resultant biomass of age-2 yellow perch produced two years later (1975 – 2010; YPTG 2011). In Management Unit 2, year- classes with a CPUE < 5 typically produced a mean age-2 biomass of 0.7 million kg

(range = 0.15 – 1.13 million kg), and supported little commercial and recreational harvest in future years (YPTG 2011). Weak, moderate, and strong year-classes produced a mean age-2 biomass of 2.1, 2.9, and 3.1 million kg, respectively (ranges = 0.8 – 4.3, 0.9 – 5.4, and 1.6 – 4.9 mil kg). Very strong year-classes produced a mean age-2 biomass of 6.7 million kg (range 1.7 – 11.1 million kg). Further, these very strong year-classes typically dominated commercial and recreational harvest even after years of natural and fishing mortality (e.g., the 2003 year-class [age-0 CPUE = 91] dominated harvest in all four Lake

Erie management units as recently as 2010; YPTG 2011). Results of these classifications for each scenario are presented in Table A1. Moving from current conditions to the high emission mid-century scenario (A1B), the probability of achieving failed and moderate year-class increased, slightly, whereas the probability of achieving very strong year-

151 classes declined moving from current conditions to the high emission mid-century scenario (A1B).

Discussion

By incorporating our experiment results into a hind-casting model, we found that total larval production (a metric that included the effects of winter duration on egg quality and hatching success) was positively related to historical (1975–2010) variation in yellow perch recruitment in Lake Erie (although the relationship was highly variable). Next, by incorporating our results into a forecasting model that included future, predicted downscaled regional winter temperature data for the mid-20th century (2046-2065), we found evidence to suggest that hatching success of yellow perch eggs will decline under high (A1B), moderate (A2), and low (B2) greenhouse gas emission scenarios (IPCC

2007), with the largest decline predicted under the highest emission scenario. However, when future hatching success was combined with estimates of total egg production to obtain a probability distribution of total larval production, we found that the predicted effects of climate change on total larval production were less than those on hatching success alone. The results of those simulations were conservative, however, as we used current (1975-2010) estimates of age and size distribution of spawners to generate total egg production for all simulations, including those for mid-century total larval production estimates. We also did not allow the slight declines in year-class strength with climate change to translate into future declines in spawners. Inclusion of this feedback would have caused stronger negative effects of climate on simulated year-class strength.

152

Our efforts herein, represent a first step in forecasting how a previously detected mechanistic effect of climate change will impact an ecologically and economically important fish population. Recent efforts to predict recruitment have indicated that studies using traditional relationships between spawning stock size (or biomass) and larval and/or juvenile abundance can be greatly enhanced when environmental variation is included (Stige et al. 2013). Furthermore, when future, predicted environmental indices are included in forecasting models, preserving the probability distribution of extreme events, as conducted here, is likely critical for generating informative predictions

(Katz and Brown 1992). Given that the Great Lakes have warmed considerably over the past several decades (Assel et al. 1995; McCormick and Fahnenstiel 1999; Jones et al.

2006) and are predicted to continue warming into the foreseeable future (Magnuson et al.

1997; Lofgren et al. 2002; Kling et al. 2003; Hayhoe et al. 2010), reliable forecasts based on mechanistic linkages between spawner abundance, the environment, and annual recruitment are essential for sustainably managing fish populations in response to continued climate change.

153

References

Crane, V. 2007. Lower trophic level and climate influences on western Lake Erie fish recruitment 1988 through 2005. Master's thesis. The Ohio State University, Columbus.

Hayhoe, K., J. VanDorn, T. Croley, II, N. Schlegal, and D. Wuebbles. 2010. Regional climate change projections for Chicago and the US Great Lakes. Journal of Great Lakes Research 36:7-21.

Intergovernmental Panel on Climate Change (IPCC). 2007. Climate Change 2007: The Scientific Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge Univ. Press, New York.

Katz, R. W., and B. G. Brown. 1992. Extreme events in a changing climate - variability is more important than averages. Climatic Change 21:289-302.

154

McCormick, M. J., and G. L. Fahnenstiel. 1999. Recent climatic trends in nearshore water temperatures in the St. Lawrence Great Lakes. Limnology and Oceanography 44:530-540.

Stige, L. C., M. E. Hunsicker, K. M. Bailey, N. A. Yaragina, and G. L. Hunt, Jr. 2013. Predicting fish recruitment from juvenile abundance and environmental indices. Marine Ecology Progress Series 480:245-261.

Yellow Perch Task Group (YPTG). 2011. Report of the Lake Erie yellow perch task group. Great Lakes Fishery Commission, Ann Arbor, MI.

155

Scenario Failed Weak Moderate Strong Very Strong

Current 32% 19% 13% 24% 11%

Low 34% 19% 14% 25% 8%

Moderate 34% 19% 15% 25% 8%

High 35% 18% 16% 25% 6%

Table A1. Probability of achieving designated year-class strengths (Failed: age-0 yellow perch CPUE [catch per unit effort] < 5; Weak: 5-14 CPUE; Moderate: 15-29 CPUE; Strong: 30-60; Very Strong: CPUE > 60) under current (1975-2010) and future (mid- century: 2046-2065) emissions scenarios (A1B: high, A2: moderate, B1: low; IPCC 2007) for management unit two in Lake Erie.

156

120

C P < 0.0001

o R2 = 0.42 110 a

100

90 b b

80 b

Days Water Temperature < 5 Water< Temperature Days 70 Current Low Moderate High

Emission Scenario

Figure A1. Mean (± SE) number of days Lake Erie water temperature < 5°C under current (1975–2010) and predicted future environmental conditions at mid-century (2046–2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007). Lowercase letters indicate means that differed significantly (Tukey’s honestly significant difference post-hoc comparison; p < 0.05).

157

Current 0.3

0.2

0.1

0.0 0.3 Moderate Emissions Scenario

0.2

0.1

0.0

0.3 Low Emissions Scenario

Probability

0.2

0.1

0.350.0 High Emissions Scenario 0.300.3

0.25

0.200.2

0.15

0.100.1

0.05

0.000.0 20 40 60 80 100 % Hatched Figure A2. Probability distribution of hatching success (% Hatched) for yellow perch eggs given current (1975-2010) and predicted future environmental conditions at mid- century (2046-2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007).

158

0.20 Current

0.15

0.10

0.05

0.000.20 Low Emissions Scenario

0.15

0.10

0.05

0.000.20 Moderate Emissions Scenario

Probability 0.15

0.10

0.05

0.000.20 High Emissions Scenario 0.15

0.10

0.05

0.00 0 20 40 60 80 100 120 140 160 Total Larval Production (in millions) Figure A3. Probability distribution of total larval production (in millions; estimated each spring from age-specific adult size distributions and relationships developed from our laboratory experiment) given current (1975-2010) and predicted future environmental conditions at mid-century (2046-2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007).

159

99th 90th 100 75th

50th 40

Age-0 Yellow Perch (CPUE) Perch Yellow Age-0 25th 0 1st

0 20 40 60 80

Total Larval Production (millions)

Figure A4. Age-0 yellow perch catch-per-unit-effort (CPUE) during October 1975-2010 for central Lake Erie (management unit 2) versus total larval yellow perch production (in millions; estimated each spring from age-specific adult size distributions and relationships developed from our laboratory experiment). Lines plotted are for 99th, 90th, 75th, 50th, 25th, and 1st regression quantile estimates.

160

0.30 0.25 Current 0.15

0.10

0.05

0.00 0.25 Low Emissions Scenario 0.15

0.10

0.05

0.00 0.25 Moderate Emissions Scenario 0.15

Probability

0.10

0.05

0.00 0.25 High Emissions Scenario 0.15

0.10

0.05

0.00 0 50 100 150 200 Age-0 Yellow Perch (CPUE) Figure A5. Probability distribution of predicted yellow perch catch-per-unit-effort (CPUE) in October given current (1975-2010) and predicted future environmental conditions at mid-century (2046-2065) under three emissions scenarios (A1B: high, A2: moderate; B1: low; IPCC 2007).

161

BIBLIOGRAPHY

Ahas, R. 1999. Long-term phyto-, ornitho- and ichthyophenological time-series analyses in Estonia. International Journal of Biometeorology 42:119-123.

Angel, J. R., and K. E. Kunkel. 2010. The response of Great Lakes water levels to future climate scenarios with an emphasis on Lake Michigan-Huron. Journal of Great Lakes Research 36:51-58.

Assel, R., K. Cronk, and D. Norton. 2003. Recent trends in Laurentian Great Lakes ice cover. Climatic Change 57:185-204.

Balmaseda, M. A., K. E. Trenberth, and E. Källén. 2013. Distinctive climate signals in reanalysis of global ocean heat content. Geophysical Research Letters 40:1754- 1759.

Baumann, H., S. C. Talmage, and C. J. Gobler. 2012. Reduced early life growth and survival in a fish in direct response to increased carbon dioxide. Nature Climate Change 2:38-41.

162

Belsley, D. A., E. Kuh, and R. E. Welsch. 1980. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. Wiley, Hoboken, New Jersey.

Berkeley, S. A., C. Chapman, and S. M. Sogard. 2004. Maternal age as a determinant of larval growth and survival in a marine fish, Sebastes melanops. Ecology 85:1258- 1264.

Bindoff, N. L., and J. Willebrand. 2007. Observations: Oceanic Climate Change and Sea Level. In: Climate Change 2007: The Physical Science Basis. Cambridge University Press, New York.

Bolsenga, S. J., and C. E. Herdendorf. 1993. Lake Erie and Lake St. Clair Handbook. Wayne State University Press, Detriot.

Brook, B. W., N. S. Sodhi, and C. J. A. Bradshaw. 2008. Synergies among extinction drivers under global change. Trends in Ecology & Evolution 23:453-460.

Bunnell, D. B., S. R. David, and C. P. Madenjian. 2009. Decline in bloater fecundity in Southern Lake Michigan after decline of Diporeia. Journal of Great Lakes Research 35:45-49.

Burnham, K. P., and D. R. Anderson. 2002. Model Selection and Multimodel Inference. Springer-Verlag, New York.

163

Carreon-Martinez, L., T. B. Johnson, S. A. Ludsin, and D. D. Heath. 2011. Utilization of stomach content DNA to determine diet diversity in piscivorous fishes. Journal of Fish Biology 78:1170-1182.

Casselman, J. M., and C. A. Lewis. 1996. Habitat requirements of northern pike (Esox lucius). Canadian Journal of Fisheries and Aquatic Sciences 53:161-174.

Chambers, R. C., and W. C. Leggett. 1996. Maternal influences on variation in egg sizes in temperate marine fishes. American Zoologist 36:180-196.

Charlebois, P. M., L. D. Corkum, D. J. Jude, and C. Knight. 2001. The round goby (Neogobius melanostomus) invasion: Current research and future needs. Journal of Great Lakes Research 27:263-266.

Ciereszko, A., and K. Dabrowski. 1995. Sperm quality and ascorbic acid concentration in rainbow trout semen are affected by dietary vitamin C: an across season study. Biol. Reprod. 52:982-988.

164

Clady, M. D. 1976. Influence of temperature and wind on survival of early stages of yellow perch, Perca flavescens. Journal of the Fisheries Research Board of Canada 33:1887-1893.

Cline, T. J., V. Bennington, and J. F. Kitchell. 2013. Climate change expands the spatial extent and duration of preferred thermal habitat for Lake Superior fishes. Plos One 8(4).

Craig, J. F. 2000. Food and Feeding. Pages 78-105 in J. F. Craig, editor. Percid fishes: systematics, ecology and exploitation. Blackwell Sciences, Oxford.

Crane, V. 2007. Lower trophic level and climate influences on western Lake Erie fish recruitment 1988 through 2005. Master’s thesis. The Ohio State University, Columbus.

Cushing, D. H. 1990. Planktonic production and year-class strength in fish populations - an update of the match mismatch hypothesis. Advances in Marine Biology 26:249-293.

Czesny, S., and K. Dabrowski. 1998. The effect of egg fatty acid concentrations on embryo viability in wild and domesticated walleye (Stizostedion vitreum). Aquatic Living Resources 11:371-378.

Dabrowski, K., and R. Bardega. 1982. The changes of fish larvae dimensions due to fixation in different preservatives. Zool. Jb. Anat.:509-516.

165

Dolan, D. M. 1993. Point-source loadings of phosphorus to Lake Erie: 1986-1990. Journal of Great Lakes Research 19:212-223.

Edwards, M., and A. J. Richardson. 2004. Impact of climate change on marine pelagic phenology and trophic mismatch. Nature 430:881-884.

Einum, S., and I. A. Fleming. 2000. Highly fecund mothers sacrifice offspring survival to maximize fitness. Nature 405(6786):565-567.

Farmer, T. M., K. Dabrowski, E. A. Marschall, and S. A. Ludsin. 2013. Winter warming effects on yellow perch reproduction and recruitment. Great Lakes Fishery Commission, Ann Arbor, MI.

Ficke, A. D., C. A. Myrick, and L. J. Hansen. 2007. Potential impacts of global climate change on freshwater fisheries. Reviews in Fish Biology and Fisheries 17:581- 613.

166

Folch, J., M. Lees, and G. H. Sloane-Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissue. Journal of Biological Chemistry 226:497-509.

Food and Agricultural Organization of the United Nations (FAO). 2010. The state of the world's fisheries and aquaculture. Rome.

Francis, J. T., S. R. Robillard, and J. E. Marsden. 1996. Yellow perch management in Lake Michigan: A multi-jurisdictional challenge. Fisheries 21:18-20.

Fuller, K., H. Shear, and J. Wittig. 1995. The Great Lakes: an environmental atlas and resource book. Government of Canada and U. S. Environmental Protection Agency, Chicago.

Garvey, J. E., E. A. Marschall, and R. A. Wright. 1998. From star charts to stoneflies: Detecting relationships in continuous bivariate data. Ecology 79:442-447.

167