Richard R. Budnik. A Dissertation

Transcription

1 ASSESSMENT OF SITE-FIDELITY AND STRAYING IN LAKE ERIE STEELHEAD TROUT Richard R. Budnik A Dissertation Submitted to the Graduate College of Bowling Green State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY May 2017 Committee: Jeffrey Miner, Advisor Enrique Gomezdelcampo Graduate Faculty Representative Paul Moore Daniel Wiegmann Kevin Pangle

2 2017 Richard Budnik All Rights Reserved

3 iii ABSTRACT Jeffrey Miner, Advisor This dissertation examines straying in Lake Erie steelhead (Oncorhynchus mykiss) and investigates how stocking practices can influence the propensity of steelhead to stray. The Lake Erie steelhead fishery generates millions of dollars in revenue each year for local economies in New York, Ohio, Pennsylvania, Michigan, and the province of Ontario. This fishery is overwhelmingly dominated by stocking from each of the four contiguous US states, and approximately 1.8 million juvenile steelhead are stocked each year to sustain population numbers. River-spawned salmonids generally exhibit high rates of philopatry, while for stockedfish straying rates of up to 15% are common. In Chapter I, we quantify the proportion of straying adult steelhead in five Lake Erie tributaries using state-hatchery specific otolith chemical signatures to identify sources. We also investigate the prevalence of naturally produced fish and identify spatial differences in the proportion of strays at different stream locations within two Lake Erie tributaries. Because straying proportions were found to be high in New York, in Chapter II an otolith back-calculation method was used to investigate the influence of size at stocking on the survival of juvenile steelhead released by the New York hatchery program. In Chapter III, we investigate additional drivers of straying by using dual-frequency identification sonar (DIDSON) to estimate survival and tributary residence time of juvenile steelhead stocked into a small Lake Erie tributary. Patterns in emigration, and the role of environmental factors and individual size on emigration timing were also investigated. In Chapter IV, we identify the prevalence of aragonite versus vaterite sagittal otoliths in steelhead raised in Lake Erie hatcheries. We then present a technique to use sagittal otoliths that have transitioned from

4 aragonite to vaterite to help develop otolith chemistry signatures of steelhead from different iv hatchery sources in Lake Erie. This research was an attempt to identify potential drivers of straying in Lake Erie steelhead. From an applied management perspective, we attempted to identify straying proportions in different Lake Erie tributaries and examine how differential stocking practices may influence overall survival and return rates. Straying is a complex process driven by a number of different biotic and abiotic factors. The identification of mechanisms that increase straying, especially those which can be controlled by hatchery managers, would be invaluable and greatly advance our ability to manage steelhead populations.

5 v This dissertation is dedicated to Amanda Browne for her unwavering love and support and to my parents Raymond and Bonnie Budnik for convincing me that life is too short to work in sales.

6 vi ACKNOWLEDGMENTS I would like to express my appreciation to my adviser, Dr. Jeff Miner, for the influence and guidance he has given me throughout my academic career. He has been instrumental for my success within my doctoral degree and through his hard work and dedication has inspired me to become a better scientist and a better person. I would also like to thank Dr. John Farver and my committee members, Dr. Enrique Gomezdelcampo, Dr. Kevin Pangle, Dr. Paul Moore, and Dr. Daniel Wiegmann, for assisting me with the construction of my project. The feedback given has been extremely insightful and I have gained a greater understanding of the scientific field through our interactions. This work would not have been possible without the generous support of Trout Unlimited (Western New York and Red House Brook Chapters) and the Larry and Linda Oman Graduate Scholarship. Further monetary support was provided by the Department of Biological Sciences at Bowling Green State University who provided tuition waivers, assistantships, and travel support. Additional travel funds were provided by the North Central Division and Ohio Chapters of the American Fisheries Society. This study would not have been possible without the cooperation of all the state hatchery staff and managers: Martha Wolgamood and Matt Hughes (Wolf Lake, MI), Andy Greulich (Salmon River, NY), Dave Insley and Andy Jarrett (Castalia, OH), Scott Fedei, Mark Haffley, Lew Kerner, Craig Lucas, and Scott Morgan, and Ray Youngs (Fairview, PA), Robert Brown (Linesville, PA), Craig Vargason (Tionesta, PA), and Bob Hetz (3-C-U Private, PA). Adult steelhead trout field collections were conducted with Jim Francis at the Michigan Department of Natural Resources (MDNR), Chuck Murray, Tim Wilson, Mike Freeman and others from the Pennsylvania Fish and Boat Commission (PFBC), and Kevin Kayle and Phil Hillman of the Ohio Department of Natural Resources (ODNR). Special thanks to Jim Markham at the New York

7 Department of Environmental Conservation (NYDEC) for assisting with adult steelhead trout vii and naturally reproduced yearling collections and Will Miller from the Seneca Nation of Indians for help with sampling on Cattaraugus Creek. Tom MacDougall at the Ontario Ministry of Natural Resources (OMNR) provided the yearlings from the Grand River, Ontario. I had the opportunity to work with so many extraordinary people at Bowling Green and would like to acknowledge my lab mates at the Aquatic Ecology and Fisheries Laboratory and the graduate and undergraduate students who have helped me along the way. Thank you, Ty, Basilius, Chris Boehler, Kevin Bland, Karen Burris, Camille Caryer, Andrew Clark, Emily Davenport, Jeremiah Davis, Sayali Gore, Nathanyal Hartkop, Will Hudacek, Jordan Jackson, Jamie Johnson, Jamie Justice, Chris Kemp, Audrey Maran, Dani McNeil, Jacob Miller, Jamie Russell, Megan Shortridge, Sadie Slamka, Kami Stamey, Sara Stienecker, Lauren Stewart, Nate Stott, and Jacob Sublett. Additional thanks to Dr. Shannon Pelini and Dr. Kevin McCluney for their advice with this project and to Dr. Mike McKay and Dr. Robert Huber for getting me on the ice every Friday morning. Mohamed Shaheen and JC Barrette at the Great Lakes Institute for Environmental Research (GLIER) at the University of Windsor provided great conversation and help with the LA-ICP-MS. Finally, I would like to thank Amanda Browne and my family for their love and support along the way. I never would have made it this far without their constant encouragement and positivity.

8 vii TABLE OF CONTENTS Page INTRODUCTION... 1 CHAPTER I: APPLICATION OF OTOLITH MICROCHEMISTRY TO INVESTIGATE THE ORIGIN AND STRAYING OF STEELHEAD IN LAKE ERIE TRIBUTARIES... 4 Introduction... 4 Methods... 7 Otolith Chemical Signatures... 7 Five-Tributary Analysis ( )... 8 New York-Tributary Analysis (2015) Results Otolith Chemical Signatures Five-Tributary Analysis ( ) New York-Tributary Analysis (2015) Discussion CHAPTER II: USE OF OTOLITH BACK-CALCULATION METHOD TO ESTIMATE THE SIZE AT STOCKING OF ADULT STEELHEAD IN NEW YORK TRIBUTARIES Introduction Methods Results Discussion... 27

9 viii CHAPTER III: USE OF DUAL-FREQUENCY IDENTIFICATION SONAR (DIDSON) TO QUANTIFY EMIGRATION PATTERNS OF JUVENILE STEELHEAD STOCKED IN A LAKE ERIE TRIBUTARY Introduction Methods Study Area Field Methods and Equipment Data Analysis Results Discussion CHAPTER IV: TRASH OR TREASURE? USE OF SAGITTAL OTOLITHS PARTIALLY COMPOSED OF VATERITE FOR HATCHERY STOCK DISCRIMINATION IN STEELHEAD Introduction Methods Results Discussion CHAPTER V: SUMMARY AND CONCLUSIONS REFERENCES... 57

10 ix LIST OF FIGURES Figure Page 1 Adult Steelhead Collection Sites Regression of Steelhead Total Length Vs. Otolith Radius Map of Trout Run, PA DIDSON Configuration at Trout Run, PA Steelhead Emigration Out of Trout Run in 2014 and Lengths of Juvenile Steelhead in Trout Run Lorentzian Peak Model for Steelhead Emigration Plot of Elemental Concentrations in Vaterite Otolith Quadratic Discriminant Function Analysis for Juvenile Steelhead with Vaterite Otoliths... 85

11 x LIST OF TABLES Table Page 1 Juvenile Steelhead Descriptions Adult Steelhead Descriptions Canonical Scoring Coefficients for Discriminant Analyses Jackknifed Classification Accuracies of Discriminant Analyses Site-fidelity and Straying of Adult Steelhead in Lake Erie Tributaries in Fall 2009 and Spring Site-fidelity and Straying of Adult Steelhead Captured in Lake Erie Tributaries of New York in Spring and Fall of Non-linear Regression Models for Environmental Predictor Variables of Emigration Rate Calcium Carbonate Polymorphs Present in Each Otolith Region of Hatchery Steelhead Average Concentrations of Elements in Each Region of Vaterite Juvenile Steelhead Otoliths from Different Sources that were Included in Quadratic Discriminant Function Analysis Cross-validation Summary from the Quadratic Discriminant Function Analysis Run on Hatchery Steelhead Elemental Signatures Total Canonical Structure Coefficients for Quadratic Discriminant Function Analysis... 96

12 1 INTRODUCTION Stocking is a common practice used to enhance fisheries worldwide (Cowx, 1998). Although stocking is often successful, hatchery and stocking practices can elicit behavioral changes, leading to differences in reproductive success and migratory behavior of fish (Fleming et al., 1997; Yokota et al., 2006). Thus, it is important for fishery managers to make informed decisions to support recreational fisheries while attempting limit the impacts of stocking. In the United States, freshwater fishing produces billions of dollars in economic revenue each year and trout fishing is among the most popular freshwater fishing pursuits (USDOI et al., 2011). The Lake Erie steelhead (Oncorhynchus mykiss) fishery alone generates millions of dollars annually for local economies in Michigan, New York, Ohio, Pennsylvania, and the province of Ontario (Murray and Shields, 2004; Kelch, 2006). Although there is some open lake harvest of steelhead in Lake Erie, the clear majority of fishing pressure occurs through tributary angling (CWTG, 2016). Because natural reproduction is not sufficient to sustain this fishery (Thompson and Ferreri, 2002), close to 1.8 million steelhead are stocked into Lake Erie bays and tributaries each year. Stocking duties are shared by several government agencies and the number of individuals released differs by state (2015 stocking totals: Michigan-64,735 yearlings; Ontario-70,250 adults; Pennsylvania-1,079,019 yearlings; Ohio-421,740 yearlings; New York- 1,378 fingerlings and 152,545 yearlings, CWTG. 2016). Stocking location (mouth vs upstream) is also variable by state, and the timing of releases varies by state, year, and tributary. Naturalized Lake Erie strain steelhead make up most releases (58%), followed by Lake Michigan (28%), and Lake Ontario strain (14%). Lake Erie strains are collected from Trout Run in Pennsylvania, Lake Michigan strains are collected from the Manistee River in Michigan, and

13 2 Lake Ontario strains are collected from the Salmon River in New York and the Ganaraska River in Ontario (CWTG, 2016). Steelhead are the migratory potamodromous life-history variant of rainbow trout (Daugherty et al., 2003) and were first introduced to the Great Lakes region in the late 1800s for sport fishing purposes (MacCrimmon and Gots, 1972). The natural life cycle of steelhead begins in streams when fertilized eggs hatch after 4-7 weeks of incubation. Newly hatched fry feed on plankton and other microscopic organisms as they grow and develop. After 1-3 years, young steelhead smolt and emigrate out into the ocean or a large lake. Following emigration, juvenile steelhead spend 1-3 years foraging and growing to sexual maturity in the open water before returning to stream tributaries to spawn (Biette et al., 1981). Typically, most steelhead return to the tributary from which they were spawned, or stocked, but straying to other tributaries for spawning is not uncommon (Labelle, 1992; Quinn, 1993; Keefer et al., 2008). In the Lake Erie region, steelhead spend approximately one year in the hatchery before release (CWTG, 2016). Because different agencies are used to raise steelhead, individuals often experience different early-life conditions in their hatchery of origin, resulting in different outcomes later in life. Environmental conditions, crowding, space availability, food availability, and food quality can all vary depending on the hatchery, leading to differences in sizes at stocking and potentially behavior. Additionally, agencies often use differential stocking practices and release steelhead: 1) at different times of year 2) at different distances from tributary mouths 3) at different distances from hatcheries 4) under different prevailing environmental conditions. It is integral to recognize optimal conditions and strategies, as incorrect management decisions can reduce survivorship and increase straying rates.

14 3 Because a tremendous amount of time and effort are put into stocking steelhead each year, identifying the origin of adult steelhead spawning in Lake Erie tributaries is crucial to help inform managers about the site-fidelity of the fish they stock, and the overall effectiveness of stocking efforts. In Chapter I we use discriminant analysis to identify source-specific chemical signatures in steelhead otoliths, and use these signatures to quantify stocking site-fidelity and straying by adult steelhead. In Chapters II and III we investigate potential mechanisms of straying in Lake Erie steelhead, with a focus on ways managers can improve stocking practices by understanding these drivers. Lastly, in Chapter IV we identify the prevalence of vateritic sagittal otoliths in Lake Erie steelhead and develop an analytical procedure that can be used to gain valuable insights into the causes of vaterite deposition and its potential negative consequences.

15 4 CHAPTER I: APPLICATION OF OTOLITH MICROCHEMISTRY TO INVESTIGATE THE ORIGIN AND STRAYING OF STEELHEAD IN LAKE ERIE TRIBUTARIES 1 Introduction Homing to natal sites is a basic life-history trait of anadromous salmonids which increases the likelihood of successful reproduction for adult fish, and of survival in juveniles (Hendry et al., 2004; Quinn, 2005). Mature fish that spawn in locations other than the sites in which they originated are considered strays (Quinn, 1993). Straying is not uncommon and successful reproduction of strays likely occurs at low levels in salmonid populations (Reisenbichler, et al. 1992). Straying allows salmonids to colonize new habitat (Milner and Bailey, 1989), avoid locally unfavorable conditions (Leider, 1989), maintain genetic diversity (Horrall, 1981), and support metapopulations (Hanski and Gilpin, 1997). Conversely, straying can negatively impact salmonid populations when hatchery fish interbreed with wild individuals, reducing genetic diversity (Emlen, 1991; Waples, 1991; Adkison, 1995; Felsenstein, 1997). Fishery managers may also be concerned about straying if a sizeable portion of the fish they stock do not return to the waters in which they were released, reducing angler catch rates within their management area. In the Lake Erie, the steelhead (Oncorhynchus mykiss) fishery generates millions of dollars in revenue each year for local economies in Michigan, New York, Ohio, Pennsylvania, and the province of Ontario (Murray and Shields, 2004; Kelch et al., 2006). Although there is some open lake harvest of steelhead in Lake Erie, most fishing pressure occurs through tributary angling, when adult fish return to spawn in tributaries (CWTG, 2016). Because natural 1 In review as: Budnik, R. R., C. T. Boehler, J. E. Gagnon, J. R. Farver, and J. G Miner. Application of Otolith Microchemistry to Investigate the Origin and Straying of Steelhead in Lake Erie Tributaries. Transactions of the American Fisheries Society.

16 5 reproduction is not sufficient to sustain this fishery (Thompson and Ferreri, 2002), close to 1.8 million steelhead are stocked annually into Lake Erie bays and tributaries. State fisheries management agencies are responsible for the majority (92%) of all steelhead stocking effort in Lake Erie (2015: Michigan-64,735 yearlings; Pennsylvania-1,079,019 yearlings; Ohio-421,740 yearlings; New York-1,378 fingerlings and 152,545 yearlings), and approximately 8% of the steelhead stocking is through sportsmen s organizations in Pennsylvania (61,232 yearlings) and Ontario (56,700 yearlings) (CWTG, 2016). Each state hatchery system obtains broodstock from a different source and the average size at stocking is variable. Stocking dates and locations (near mouth vs upstream) also vary by state, year, and tributary. The differential stocking practices used throughout Lake Erie may influence survivorship, spawning site-fidelity, and straying of steelhead. Agencies that stock smaller individuals likely have increased mortality as smaller individuals are much less likely to survive after stocking (Ward, 1989; Slaney et al., 1993; Seelbach et al., 1994). Additionally, early emigration due to high spring discharge events, or near-mouth stocking, could cause steelhead to leave a stream before imprinting occurs (Daugherty et al., 2003), leading to increased straying. Along with size and location, the strain of steelhead stocked by each state, and the procedures for collecting broodstock, may influence site-fidelity and the timing of spawning runs (Keefer and Caudill, 2014). Different strains of steelhead are stocked by each state in Lake Erie and broodstock collection procedures vary (multiple-season collections vs. spring only collections). Wild juvenile steelhead that are present in some Lake Erie tributaries may also influence sitefidelity as they can produce pheromones that attract homing adults (Groot et al., 1986; Quinn and Tolson, 1986; Courtenay et al., 1997), and could potentially attract straying individuals.

17 6 With the amount of effort and resources put into stocking, identifying the origin of spawning steelhead is crucial to help inform managers about straying among tributaries and the overall effectiveness of stocking efforts. In this study, we used discriminant analysis to identify source-specific chemical signatures in steelhead otoliths, and used these signatures to examine the origin and straying of steelhead among Lake Erie tributaries. Additionally, to confirm consistent differences in otolith chemistry among Lake Erie hatcheries, fish from different year classes were used to develop otolith chemical signatures. Otolith chemistry is a popular and effective method for discriminating stocks of fish (Campana, 1999; Elsdon et al., 2008), and has been used widely in the Great Lakes (Pangle et al., 2010; Reichert et al., 2010; Hayden et al., 2011). Most notably, it has been demonstrated that the otolith chemistry from salmonids reared at different hatcheries in the Great Lakes are significantly different, allowing stocks of chinook salmon (Oncorhynchus tshawytscha; Marklevitz et al., 2011) and steelhead (Boehler et al., 2012) to be accurately discriminated. Although salmonid homing to the river of origin has been extensively documented (Quinn et al., 1989; Quinn, 1993; Dittman et al., 2010; Keefer and Caudill, 2014), the fidelity of fish to their natal (or stocked) site is not as well understood (Dittman et al., 2010). The selection of spawning sites likely involves a complex trade-off among selective pressures to home, spawning habitat selection, competition, and mate choice (Dittman and Quinn, 1996; Hendry et al., 2004). The weighting of these tradeoffs may be different in fish that do not home to their stream of origin (strays). For example, assuming selective pressures to home are weighted less in straying fish, you might expect strays to be more exploratory, and more likely to travel further upstream to find suitable spawning sites. Besides Boehler et al. (2012) who reported a slightly higher proportion of strays at upstream locations in a Lake Erie tributary in Pennsylvania, little

18 7 work has focused on the prevalence of strays at different locations within the same tributary. To address this question, we also implemented a second, finer-scale, approach to examine the origin and straying of steelhead within Lake Erie tributaries at different upstream locations. Methods In this study, we considered adult steelhead strays if they returned to a tributary located in a state or province different than the one in which they were released or spawned (e.g., a Pennsylvania hatchery fish collected in a New York tributary as an adult). As such, there may have been straying among tributaries within each state, in which case the reported proportions of strays are conservative. In addition, neither steelhead collected in New York identified as Cattaraugus Creek origin (i.e., wild reproduction), nor steelhead identified as unknown origin were considered strays. Otolith Chemical Signatures For stock identification, hatchery yearling steelhead (mean total length and standard error = 163 ± 4 mm; range = mm) were provided by managers and personnel from the state hatchery systems in Lake Erie: Wolf Lake State Fish Hatchery (Michigan), Castalia State Fish Hatchery (Ohio), Salmon River Fish Hatchery (New York), Linesville State Fish Hatchery (Pennsylvania), Tionesta State Fish Hatchery (Pennsylvania), Fairview State Fish Hatchery (Pennsylvania), and from the private 3-C-U, Erie County Cooperative Trout Nursery (Pennsylvania). Yearlings were collected from multiple cohorts ( ; Table 1), and analyses of their otoliths demonstrated that otolith chemical signatures within hatcheries were consistent between years. Individuals from the Pennsylvania hatcheries were eventually pooled

19 8 because similarities in the otolith chemistry signature of each hatchery made discrimination difficult. Additional details for pooling of individuals from the Pennsylvania hatcheries can be found in Boehler et al. (2012). Wild juvenile steelhead were also collected from two Lake Erie tributaries where natural recruitment occurs. In Cattaraugus Creek, NY, wild yearlings were collected with the assistance of the New York State Department of Environmental Conservation (NYDEC) in July Fish were also collected in August 2010 from Grand River by the Ontario Ministry of Natural Resources. These fish were interpreted to be resident rainbow trout (non-migratory) because they had typical stream rainbow trout coloration and were larger than yearling steelhead collected from other systems in Lake Erie (mean total length and standard error = 194 ± 6 mm; range = mm). Regardless of whether they were resident or migratory, they provided an accurate otolith chemical signature for Grand River, ON. Five-Tributary Analysis ( ) Adult steelhead were collected in conjunction with the Michigan Department of Natural Resources, the Ohio Department of Natural Resources, the Pennsylvania Fish and Boat Commission, and the NYDEC during the fall of 2009 and spring of Individuals were obtained from five Lake Erie tributaries: Huron River (MI), Vermilion River (OH), 16-mile Creek (PA), Chautauqua Creek (NY), and Cattaraugus Creek (NY) (Figure 1). Collections were obtained near the mouth of each tributary ( 4 km upstream), except for Huron River where collections occurred 15 km upstream (Table 2). Once steelhead were collected, sagittal otoliths were removed and prepared for microchemical analysis. Procedures for preparation of otoliths and subsequent chemical analysis

20 9 have been presented previously (e.g. see Secor et al., 1991; Hayden et al., 2011; Boehler et al., 2012). Briefly, otoliths were cleaned of organic material by sonication in hydrogen peroxide (3% V:V), and air dried. Otoliths were embedded in a two-part epoxy (West System 105 Epoxy Resin and 206 Slow Hardener ) and sectioned in the transverse plane using a low-speed wafer saw with diamond-tipped blade. Both sides of the cut otolith cross-sections were wet polished using ultrapure (Milli-Q ) water with 3M -brand silicon carbide sandpaper and lapping film (particle size: 20 μm, 10 μm, 6 μm, 2 μm) to a thickness of approximately 200 μm. Polished otoliths were mounted on standard petrographic microscope slides using epoxy. Mounted slides were triple-rinsed and sonicated for five minutes with ultrapure water, then covered and allowed to dry overnight, and stored in clean Petri dishes until analyses were performed. Otoliths were analyzed using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) at the Great Lakes Institute for Environmental Research (GLIER; University of Windsor, ON). A description and operating conditions for the LA-ICP-MS can be found in Boehler et al. (2012). The theoretical concentration of calcium in stoichiometric calcium carbonate (400,432 μg Ca g -1 CaCO3) was used as an internal standard to correct for ablation yield differences between external calibration standard and the otoliths, and the occurrence of mass 120 (measured as 120 Sn, tin isotope), an indicator of contamination from the epoxy bonding medium, was also quantified (Ludsin et al., 2006; Reichert et al., 2010). Data processing and calculations of detection limits were performed using a Microsoft Excel spreadsheet macro (Yang, 2003) based on algorithms developed by Longerich et al. (1996). A total of 20 isotopes of 15 elements ( 7 Li, 25 Mg, 33 S, 39 K, 43 Ca, 44 Ca, 55 Mn, 57 Fe, 63 Cu, 65 Cu, 66 Zn, 67 Zn, 86 Sr, 87 Rb, 88 Sr, 120 Sn, 133 Cs, 137 Ba, 138 Ba, 208 Pb) were collected, of which only 86 Sr and 138 Ba (reported as total Sr and Ba) were used in the discrimination analysis. Sr and Ba have

21 10 previously shown to be important discriminators of natal origins in the Great Lakes (Brazner et al., 2004; Ludsin et al., 2006; Hand et al., 2008; Pangle et al., 2010; Boehler et al., 2012), and this approach was used because Sr and Ba had concentrations with lower variance within hatcheries compared to other elements analyzed. Sr and Ba concentrations were always greater than the limits of detection (LOD; mean LOD concentration in ppm and standard error: Sr 1.24±0.03 and Ba 0.044±0.002). Elemental concentrations of otolith regions were standardized using procedures outlined by Boehler et al. (2012). Briefly, otoliths from steelhead hatchery yearlings were standardized from core (0%) to edge (100%), and otoliths from wild stocks (Grand River and Cattaraugus Creek) were standardized from core (0%) to first annulus (100%). Average concentrations for Sr and Ba were calculated for each 5% otolith increment, but the inner 10% (50-60 µm) of all otoliths was excluded from subsequent analyses to minimize potential maternal influences (Chittaro et al., 2006; Macdonald et al., 2008; Wolff et al., 2012). The remaining 18 Sr and 18 Ba variables were entered into a linear discriminant analysis (LDA) and systematically removed with backwards stepwise variable selection, until the combination of variables yielding the lowest number of misclassified fish was obtained. To remove impacts of multicollinearity, any otolith regions that were continuous and highly correlated (pairwise correlation >0.70) were pooled (McGarigal et al., 2000). In total, eight otolith region variables (5 Sr and 3 Ba) were included in the final discriminant analyses (Table 1). Because differences in the crystalline structure (aragonite or vaterite) of otoliths causes variation in otolith elemental signatures, only otoliths composed entirely of aragonite were included in all subsequent analyses (Zimmerman and Nielsen, 2003; Gibson-Reinemer, et al. 2009). Vateritic otolith sections were identified

22 11 based on their characteristically low levels of Sr and high levels of Mg (Gauldie, 1996; Melancon et al., 2005; Gibson-Reinemer, et al. 2009). Initially, an attempt was made to separate all source populations of steelhead in a single LDA; however, due to similar mean Ba concentrations in the outer otolith regions between the Pennsylvania hatcheries and Cattaraugus Creek, and heavy weighting of the outer regions when all source populations were included in the analysis (especially the Castalia, OH yearlings), discrimination was relatively poor. Therefore, two discriminant analyses were used to maximize source population identification (Curtis et al., 2014). A primary analysis was utilized to discriminate yearlings from Wolf Lake State Hatchery (MI), Castalia State Hatchery (OH), Salmon River State Hatchery (NY), Pennsylvania (hatcheries pooled), and Grand River, ON (see Boehler et al., 2012). Then, to maximize discriminating power between the Pennsylvania hatcheries and wild Cattaraugus Creek, NY otolith chemical signatures, a secondary LDA was employed. This secondary analysis excluded all other source populations (Castalia, Wolf Lake, Salmon River, and Grand River), changing the between group covariance matrix and subsequent variable weightings. Thus, the overall classification accuracy between Pennsylvania and Cattaraugus Creek signatures was improved while including the same Sr and Ba otolith regions used in the primary LDA. Only adult steelhead identified as Pennsylvania hatchery origin in the primary LDA were included in the secondary LDA (Pennsylvania and Cattaraugus Creek only). A multivariate analysis of variance was used to test whether the elemental signatures of steelhead source populations in Lake Erie were significantly different. Likelihood probabilities that all yearlings belonged to their correct origin were calculated using a jackknife procedure (Thorrold et al., 1998; Wells et al., 2003; Walther and Thorrold, 2008) for both the primary and secondary discriminant analyses. The high jackknifed classification accuracies and low numbers

23 12 of misclassified fish (see results) in the linear discriminant analyses without removal of statistical outliers suggest these analyses were robust and not sample size limited (McGarigal et al., 2000). For a fish to be correctly classified it needed a probability >0.50 of belonging to its actual (pactual) hatchery of origin. We report the mean p-actual for source populations, allowing individuals that classify correctly, but with lower likelihood (i.e., fish with p-actual of 0.60), to have more influence in the overall mean classification (Boehler et al., 2012). All statistical analyses were conducted using JMP (copyright 2015 SAS Institute Inc). Unlike the hatchery yearling otoliths that were partitioned into 5% increments from core to the true otolith edge, adult steelhead otoliths were analyzed from core (0%) to the demarcation at stocking (100%). All otolith material that occurred after the shift in otolith chemistry between the known hatchery, or first year chemical signature, and the consistent Lake Erie signature (or potential stream signature in some instances) was ignored. This allowed the source population of unknown origin adults to be identified with the linear discriminant analyses generated from the same otolith regions as juveniles. After the primary analysis of adult steelhead, only adults identified as Pennsylvania hatchery origin in the first LDA were included in the secondary LDA to differentiate adults of Pennsylvania hatchery origin from Cattaraugus Creek origin. Like the yearling steelhead, adults had to have a p-actual >0.50 of belonging to one of the known source populations or they were considered of unknown origin. We estimated straying within each tributary as the percentage of the total sample of steelhead (hatchery and wild) that was identified as a stray. Differences in straying between tributaries and between seasons were statistically analyzed using Pearson s χ 2. We could not

24 13 calculate a stray rate (i.e., the percentage of a source group that strayed) because we could not account for all adult returns from a given source. New York-Tributary Analysis (2015) To confirm consistently high straying proportions in the New York tributaries (see results), and to identify proportional differences in straying within these tributaries, adult steelhead were collected from Cattaraugus and Chautauqua Creeks in the fall and spring of Adults were collected from three different locations in 2015: Cattaraugus Creek - 3, 27, and 59 km upstream from the mouth; Chautauqua Creek <1, 3, and 7 km upstream from the mouth. These distances represent the near maximum range of each creek system that can be traversed by adult steelhead. Adults were collected from all locations in both the spring and fall of 2015 from Chautauqua Creek and from the downstream section of Cattaraugus Creek. Fish were only collected from the two upstream locations of Cattaraugus Creek during the fall of 2015 due to logistical constraints. All otoliths were removed and prepared using the same methods described in the fivetributary analysis. Otoliths were also analyzed using the same methods and settings described by Boehler et. al (2012), but, because of an upgrade of the LA-ICP-MS at the GLIER, a PhotonMachines 193 nm excimer laser coupled to an Agilent 7900 fast-scanning ICP-QMS was used. The proportion of adults from each stocking source and the proportion of strays were then determined for each tributary section and each season in both New York tributaries. Differences in straying between seasons and between stream sections were statistically analyzed using Pearson s χ 2.

25 14 Results Otolith Chemical Signatures The otolith microchemistry of state hatchery steelhead yearlings (Wolf Lake, MI; Castalia, OH; Salmon River, NY; and Pennsylvania) along with the Grand River, ON yearlings was found to be significantly different with the primary LDA (overall MANOVA: Wilk s λ F32,629 = 64.25; P < ). The first two canonical axes accounted for 99.6% of the total variance (CA1 = 94.1%, eigen value = 30.5; CA2 = 5.5%, eigen value = 1.8, see canonical scoring coefficients, Table 3). The mean jackknifed classification accuracies ranged from 72.1% (Michigan hatchery) to 100.0% (Ohio hatchery). Overall, the primary LDA accurately classified 94.9% (224 of 236) of the juvenile steelhead (Table 4). Because discrimination between the Pennsylvania hatchery fish and the naturally reproduced steelhead from Cattaraugus Creek was critical to our conclusions, the secondary LDA was employed to maximize the discrimination of these two source populations (overall MANOVA: Wilk s λ F8,165 = 36.65; P < ). This secondary LDA utilized the same otolith regions as the primary LDA, and had high mean jackknifed classification accuracies (Pennsylvania = 88.6%, Cattaraugus Creek = 90.1%), with 91.4% (159 of 174) of the juvenile steelhead being correctly classified (Table 4). Canonical axis-1 explained 100% of the total variance (eigen value = 1.8) in the secondary LDA (canonical scoring coefficients, Table 3). Five-Tributary Analysis ( ) The incidences of straying in Chautauqua and Cattaraugus Creeks in New York were 72% and 88%, respectively (fall and spring combined; Table 5) and the proportions of strays in these creeks were similar in both fall and spring sampling periods (Pearson s χ 2, all p > 0.05).

26 15 Pennsylvania hatchery origin fish made up the greatest proportion of fish identified (Chautauqua: fall 67%, spring 29%; Cattaraugus: fall 60%, spring 43%) except during spring Chautauqua Creek collections when a higher proportion of Ohio hatchery fish were captured (39%). The proportions of Pennsylvania fish were significantly lower during spring collection periods (Pearson s χ 2, both p < ), and during both spring collection periods the decrease in Pennsylvania hatchery fish coincided with a proportional increase in Ohio hatchery fish (Chautauqua: 34% increase; Cattaraugus: 9% increase). The other Lake Erie tributaries outside of New York had straying instances of 23% (16- mile Creek, PA), 18% (Vermilion River, OH), and 57% (Huron River, MI) (fall and spring combined; Table 5). The proportion of strays for each tributary was similar in both fall and spring sampling periods (Pearson s χ 2, all p > 0.05), except in Vermilion River where significantly more strays were identified in fall (Pearson s χ 2, p = ). In Vermilion River individuals from every state stocking program were identified, while in 16-mile Creek only strays from the New York and Michigan hatcheries were found. Wild fish made up 9% of the catch of steelhead in Lake Erie tributaries during the study period (Table 5). Wild Cattaraugus Creek origin steelhead were identified in 16-mile Creek along with Chautauqua and Cattaraugus Creeks (1.6%, 4.5%, and 2.7% of total catch, respectively), and Grand River origin fish were identified in each tributary sampled except Chautauqua Creek (Huron River, MI =43.2%, Vermilion, OH =2.1%, 16-mile Creek, PA =3.2%, Chautauqua Creek, NY =0.0%, Cattaraugus Creek, NY =2.7%, all percentages of total catch).

27 16 New York-Tributary Analysis (2015) The proportions of strays identified in Cattaraugus Creek (80%) and Chautauqua Creek (84%) at the locations closest to the mouth were like those from (Pearson s χ 2, spring and fall combined; both comparisons p > ). Overall, 74% of all individuals captured in Cattaraugus Creek and 68% of all individuals captured in Chautauqua Creek were strays (spring and fall combined; Table 6). In both seasons, the proportions of strays were similar in all stream sections of Cattaraugus Creek (Pearson s χ 2, all comparisons p > 0.05), and in Chautauqua Creek only the proportion of strays found in the furthest upstream section (7 km) was significantly greater than the tributary section closest to the mouth (<1 km) (Pearson s χ 2 comparison of <1 km site and 7 km site: fall p = ; spring p = , all other comparisons p > 0.05). When comparing fall and spring sampling periods, the proportions of strays were similar in all stream sections (Pearson s χ 2, all comparison p > 0.05), but seasonal comparisons could not be made for the two upstream sections of Cattaraugus Creek as no fish were collected during spring sampling periods (Table 6). In 2015 Pennsylvania hatchery fish made up half of the total individuals collected in Chautauqua Creek, followed by New York hatchery (21%), Ohio hatchery (12%), wild New York (11%), Michigan hatchery (5%), and wild Ontario (1%) fish. Like Chautauqua Creek, collections in Cattaraugus Creek consisted mostly of Pennsylvania hatchery steelhead (57%). Wild Cattaraugus Creek individuals were the next most common (17%), followed by New York hatchery (10%), and Michigan and Ohio hatchery (8% each). The proportion of Pennsylvania hatchery fish identified was significantly lower in spring than fall (Pearson s χ 2, p = 0.006). Conversely, the proportion of Ohio hatchery fish identified was higher during the spring season (Pearson s χ 2, p < 0.001). The percentage of New York hatchery fish was higher in spring than

28 17 fall (Pearson s χ 2, 0.034) and no proportional difference was identified for wild Cattaraugus Creek individuals among seasons (Table 6). The proportion of Michigan hatchery steelhead found in each section of the New York tributaries varied (range: 0-20%). Michigan individuals were identified in each tributary section of both tributaries and during most sampling periods. Throughout 2015 sampling only one wild Ontario individual was identified, and was found in the downstream section of Cattaraugus Creek (Table 6). Discussion Hatchery stocks and wild fish from New York and Ontario were clearly discriminated using elemental signatures from otolith transects. The high classification accuracies obtained, despite fish from multiple year-cohorts being used, implies that otolith chemical signatures of steelhead in Lake Erie are relatively stable. The discriminant analyses generated here could be used for future studies if the ambient water chemistry in hatcheries (or in natal tributaries) remains similar among years. The percentages of stray steelhead in New York tributaries were higher (72%-88% of total catch) than in the other Lake Erie tributaries sampled (18-57% of total catch), and when compared to those found in coastal rivers in Oregon (4-43%; Schroeder et al., 2001). These straying values should be considered conservative because we identified hatchery stocking source by state, and not specific tributary of stocking. In Vermilion River strays were mostly of Pennsylvania hatchery origin, while in 16-mile Creek most strays were of New York hatchery origin. This may be representative of a preference to stray to geographically proximate locations rather than to travel longer distances, a tendency that has been previously identified in straying

29 18 steelhead (Lirette and Hooton, 1988; Schroeder et al., 2001). In Huron River, 43% of the adult steelhead collected were identified as Grand River, ON origin. Although the jackknifed classification accuracy of Grand River juveniles was high (91%), we cannot be certain that some of the adult steelhead classified as Grand River origin did not originate from other sources for which we did not have otolith chemistry data. For example, both the Grand River and the southeastern shoreline of Lake Huron share the same geological terrane (Marklevitz et al., 2011), potentially making the otolith chemical signatures of steelhead originating from this region difficult to discriminate from the Grand River signature. The possibility of misclassification of Lake Huron origin steelhead is also feasible considering the connectivity that exists between Lake Huron and Lake Erie waters and the high proportion of adult steelhead captured in Huron River predicted to be of wild Grand River, ON origin. The adult steelhead in Cattaraugus and Chautauqua Creeks were overwhelmingly Pennsylvania (52% of total New York catch) and Ohio hatchery origin fish (17% of total New York catch). Stocking practices and their relationship with stream imprinting may play a substantive role in the high proportion of strays found in New York streams, as both Pennsylvania and Ohio stocking programs implement near-mouth releases. For example, Ohio stocks approximately 60,000 steelhead yearlings into Vermillion River <2 km from Lake Erie annually. Similarly, Pennsylvania stocks approximately 20,000 steelhead yearlings into 16-mile Creek <1 km from the lake. Short emigration distance may result in poor stream imprinting for these fish, especially in years with high tributary flow rates, resulting in rapid emigration (Melnychuk et al., 2010). In addition to near-mouth stocking, the large number of juvenile steelhead stocked by both Pennsylvania and Ohio may contribute to the large percentage of individuals from these

30 19 stocking programs captured in New York tributaries. Pennsylvania stocks approximately 1.1 million steelhead yearlings annually into 13 streams, and Ohio stocks approximately 450,000 yearlings into five streams. High densities of smolts can promote rapid emigration from these systems (Keeley, 2001), decreasing stream imprinting time. Rapid emigration and the close proximity of some Pennsylvania and Ohio tributaries to New York streams may combine to contribute to straying potential. For example, the mouth of 20-mile Creek, which is stocked with approximately 160,000 Pennsylvania hatchery yearlings each year (> 50% of the entire NY annual stocking program in Lake Erie), is <17 km from the mouth of Chautauqua Creek, which is stocked with approximately 40,000 New York hatchery yearlings. In fact, NYDEC biologists (J. Markham, Lake Erie Fisheries Unit, NYDEC, Dunkirk, NY, personal communication, 2013) find yearling steelhead of stocking size in lower Chautauqua Creek in spring before they stock fish (i.e., PA stocks up to one month earlier than NY). Thus, rapid emigration and exploration of nearby tributaries may be causative factors that lead some of the steelhead stocked into Pennsylvania and Ohio streams to select New York waters to spawn. Even if rapid emigration is not leading to increased rates of straying in Pennsylvania and Ohio hatchery fish, many individuals would still be expected to stray to New York tributaries simply due to the large number of individuals stocked, and the propensity of steelhead to stray to geographically proximate locations (Lirette and Hooton, 1988; Schroeder et al., 2013). Straying percentages could also be exacerbated by poor returns of New York hatchery fish. The Salmon River State Hatchery yearlings released into New York tributaries are stocked at <150 mm on average (mean total length ranges from mm; J. Markham, Lake Erie Fisheries Unit, NYDEC, Dunkirk, NY, personal communication, 2013), which is below the approximate size threshold for juvenile steelhead emigration (Wagner et al., 1963; Seelbach, 1987; Peven et al.,

31 ; Quinn, 2005). If Salmon River State Hatchery yearlings are too small to physiologically smolt and are not ready to emigrate, then their longer residency may lead to substantive instream mortality because of high summer temperatures, predation, and competition with wild yearlings (Keeley, 2001; Hill et al., 2006). Slaney et al. (1993) reported a three-fold increase in survival for steelhead smolts stocked at 180 mm compared to those stocked at 145 mm. Thus, poor contribution of the Salmon River State Hatchery stock to New York streams may contribute to the higher than expected proportion of Pennsylvania and Ohio hatchery steelhead in New York tributaries. The identification of wild origin adult steelhead in Lake Erie tributaries (15% of total catch) suggests that naturally reproduced individuals are contributing to the fishery. This moderate natural reproduction is another potential driver of the large proportion of Pennsylvania and Ohio strays identified in New York tributaries. New York streams stay sufficiently cool during most summers to allow young steelhead to remain in the system throughout the year and it is possible that odors from naturally reproduced fish in the New York tributaries attract adult steelhead trout from other systems. Unlike most Lake Erie tributaries that support very little successful reproduction due to high summer temperatures (Thompson and Ferreri, 2002), Chautauqua and Cattaraugus Creeks support steelhead year around (Goehle, 1999; this study), potentially allowing continuous odor marking of stream substrate to occur. Atlantic salmon (Salmo salar) parr can recognize conspecific pheromones deposited in stream substrates (Stabell, 1987), but the concentration of the odor also influences whether a preference is observed (Courtenay et al., 2001). While adult salmonids can also discriminate between populations (Stabell, 1992; Brown and Brown, 1996), the role chemical odors (i.e., pheromones) play in homing is still uncertain (Quinn, 2005; Ueda, 2012).

32 21 We found no evidence that straying individuals were more likely to travel further upstream to find suitable spawning sites. Little difference in straying proportions was identified in each section of Cattaraugus and Chautauqua Creeks indicating that straying individuals utilized both upstream and downstream sections of these tributaries. Oncorhynchus species commonly non-directly home to two or more alternate sites before reaching their final spawning destination, but this non-direct homing is confined to the lower reaches of nearby tributaries (Burger et al., 1995; Keefer et al., 2008). The similar proportions of strays identified in the lower, middle, and upper reaches of the New York tributaries, suggests that these individuals are in fact strays, and are not just individuals utilizing additional tributaries on their way to their final spawning site. Compared to fall sampling periods, the proportions of Pennsylvania fish identified in Chautauqua Creek were lower during spring sampling periods in all tributary locations. It is unclear whether the reduction in the proportion of Pennsylvania fish during the spring season is the result of less Pennsylvania fish straying into New York tributaries, an increase in Ohio fish straying in the spring, or some combination of the two. Each year Pennsylvania performs six adult steelhead collections from November-February to obtain broodstock for their hatchery program (Vargason, 2013). Because spring sampling occurred in April, it is possible that many the Lake Erie strain steelhead stocked by Pennsylvania had already spawned and left the New York tributaries before the sampling occurred. Conversely, peak spawning of Manistee strain steelhead stocked by the state of Ohio occurs in March through April (Seelbach, 1993; Seelbach et al., 1994) and may be responsible for the increased number of Ohio strays present in Cattaraugus and Chautauqua Creeks during spring sampling periods.

33 22 In summary, our results provide state fishery managers with new and valuable information regarding the presence of straying adult steelhead in Lake Erie tributaries. Future otolith chemistry studies, including source populations from multiple Great Lakes, could greatly increase the knowledge gained about movement of steelhead within and between systems (Landsman et al., 2011). For example, Barnett-Johnson et al. (2010) found that coupling otolith chemistry with genetic analysis further increased stock resolution. There is also a need to quantify run numbers into Lake Erie tributaries so that straying rates and total numbers of strays can be determined.

34 23 CHAPTER II: USE OF OTOLITH BACK-CALCULATION METHOD TO ESTIMATE THE SIZE AT STOCKING OF ADULT STEELHEAD IN NEW YORK TRIBUTARIES Introduction Based on otolith microchemical analyses a large proportion of steelhead stocked into Lake Erie tributaries by state fish hatcheries in Pennsylvania and Ohio have been identified in New York tributaries (Cattaraugus and Chautauqua Creeks) during fall and spring spawning runs (see Chapter I). This suggests that either a large number of individuals from Pennsylvania and Ohio hatcheries are straying into New York tributaries and/or steelhead stocked by the state of New York are not returning in large numbers. Here we investigate size at stocking as a potential mechanism for poor returns of New York hatchery fish. Hatchery steelhead may fail to become migratory when stocked too small (Wagner et al., 1963; Slaney and Harrower, 1981, Rempel et al., 1984, Ward and Slaney, 1990) making it important for managers to release fish that are the correct size. The threshold size for steelhead emigration is mm (Wagner et al., 1963; Chrisp and Bjornn, 1978; Bjornn et al., 1979; Seelbach, 1987; Peven et al., 1994; Quinn, 2005) and the average size at stocking for most agencies in the Lake Erie region is above this threshold (In 2014: Michigan-193 mm, Pennsylvania-177 mm, and Ohio-171 mm), except for New York where the average size at stocking is <150 mm (118 mm in 2014; CWTG, 2015). Juvenile steelhead can reach, and exceed, 150 mm within one year in the hatchery; however, not all individuals are able to reach 150 mm after a year of hatchery life. Residual hatchery fish that do no emigrate in the first year of release are prone to high mortality (Seelbach, 1987), usually occurring during summer months (Chrisp and Bjornn, 1978).

35 24 Size-biased survival has been identified in stocked juvenile steelhead in Lake Michigan (Seelbach et al., 1994) and in other steelhead populations along the Pacific coast. Slaney et al. (1993) determined that return rates of stocked steelhead were highly correlated with size, and found that survival tripled as the length of individuals at time of stocking increased from 145 to 180 mm. Ward et al. (1989) also reported that survival of wild steelhead in the Pacific Ocean was positively correlated with juvenile size, and that individuals had a 10% increase in survival for every 10 mm increase in length (when individuals were between 140 and 260 mm). To investigate the impact of stocking size on the survival of steelhead released in Lake Erie, we used an otolith back-calculation procedure to estimate the size at stocking of adults captured in Chautauqua and Cattaraugus Creeks in New York. Like the annular growth rings found in trees, layers of calcium are deposited in otoliths at different rates throughout the year forming visible annuli (growth rings) that can reveal the approximate age of fish (Pannella, 1971, 1974). Back-calculation using these rings has been commonly used by fisheries scientists and ecologists to infer the size of fish at ages prior to capture (Francis, 1995). Typically, three steps are performed during back-calculation: verification of the periodicity of annulus formation, establishment of an otolith radius (OR)-total length (TL) relation, and the estimation of size at the time of annulus formation (Schirripa, 2002). Here we used a modified back-calculation method that utilizes both traditional identification of otolith annuli and otolith microchemistry. Because Lake Erie steelhead are stocked at approximately one year of age, the 1 st annulus of adult steelhead otoliths, which demarcate one year of life, can be used to estimate the size of a fish at time of stocking. However, due to different prevailing hatchery conditions each year and variable stocking dates (CWTG, 2014; 2015; 2016), the formation of the first annulus may not accurately represent the time of stocking. Inaccuracies would be expected if, before the time of

36 25 stocking, the 1 st annulus is not formed or additional otolith growth past the 1 st annulus occurs. Because otoliths do not undergo chemical resorption, their chemical composition can provide a permanent chronological record of the environment(s) in which fish have resided during their lifetime (Secor, 1991; Campana and Thorrold, 2001). Here I identify the section of adult steelhead otoliths where a shift in otolith microchemistry occurred (from hatchery chemical signature to Lake Erie chemical signatures), and use this site to perform back-calculation estimates of the size at stocking of adult steelhead captured in Chautauqua and Cattaraugus Creeks, NY. I then use these back-calculated sizes to compare the average size at stocking of surviving adults from the New York hatchery compared to those surviving adults that strayed to New York tributaries from other hatcheries around Lake Erie. Methods Yearlings from each of the five state hatcheries that raise steelhead for Lake Erie stockings (spring 2015 collection: Salmon River, NY n=9; Tionesta, PA n =12; Fairview, PA n = 8; Castalia, OH n =13; and Wolf Lake, MI n=10) and adult steelhead from Chautauqua and Cattaraugus Creeks, NY (2015 total for both creeks combined: fall n=32; spring n=20) were used to develop a regression of TL and OR. All individuals used in the regression had been previously measured for TL and had had their otoliths removed, prepared, and analyzed to determine otolith microchemistry. Additionally, using LDA the hatchery origin of adult steelhead had been determined (see Chapter 1). For otolith radius measurements, otoliths that had been polished to ~200 μm and affixed to petrographic microscope slides were placed on a dark board under reflected light and otoliths were examined with a dissecting microscope to identify length from core to edge (OR). An additional measurement was taken from the core to the demarcation of

37 26 stocking on adult otoliths and all measurements were taken in the same plane. Otolith material is burned away during analysis with LA-ICP-MS making laser traverses visible on otoliths after analysis. Therefore, the demarcation of stocking could be identified by locating the section of the laser traverse where a shift in otolith chemistry, from the hatchery signature, to the Lake Erie signature occurred. When laser traverses were not in the same plane as otolith length measurements, the demarcation site was identified on the laser traverse and a corresponding otolith annuli was followed until it aligned with the otolith measurement path, and a length measurement was made. When the demarcation of stocking did not directly lineup with an annulus, the demarcation site in relation to the two annuli, in which the demarcation of stocking site was between, was used for measurement (ie: if the demarcation of stocking was half way between two annuli, then both these annuli would be followed to the otolith measurement plane and a site half way between the two annuli in the otolith measurement plane was used). Because Lake Erie steelhead are stocked as yearlings, the demarcation of stocking was verified by its proximity to the first otolith annulus. The continuous translucent zones of otoliths were assumed to be annuli and the white, opaque zones were considered winter growth (McKern et al., 1974: Bagenal and Tesch, 1978: Chilton and Beamish, 1982: Barber and McFarlane, 1987; Peven et al., 1994). Caution was used not to read the "metamorphic" check (the translucent zone that occurs around the nucleus of the otolith at hatching) as the 1 st annulus (McKern et al., 1974; Rybock et al., 1975; Martin, 1978; Peven et al., 1994). The relationship between TL and OR was not linear therefore a polynomial regression was developed. This was regression was forced through the origin because otoliths are formed early in the development of steelhead embryos (McKern, 1971). The length of adult otoliths, from core to the demarcation of stocking, were each plugged into the regression equation to

38 27 back-calculate size at stocking. To identify any size difference between the size at stocking for surviving adult steelhead from the New York hatchery and the other Lake Erie hatcheries, twosample t-tests were used to compare back-calculated sizes for each group. A second t-test was performed to compare the difference in TL between juvenile steelhead collected from the New York hatchery prior to stocking (average size of all fish stocked), and the estimated TL at time of stocking for adult steelhead (average size at stocking of fish that survived to adulthood). Results Regression analysis showed a strong and significant relationship between fish TL and OR when a quartic function was used (r 2 = 0.95, n=104, p < 0.001; Figure 2). Back-calculation of size at stocking for adult fish produced a mean TL of mm. Average estimated TL of individuals predicted to be from the New York hatchery ( ) was significantly smaller than the estimated average TL of fish predicted to be from other hatcheries ( ; p = 0.004). Conversely, the mean estimated size at stocking for adult steelhead from New York were significantly larger than the mean TL of juvenile steelhead from the New York hatchery that were measured prior to stocking ( ; p = 0.001) Discussion Typically, the periodicity of annulus formation in otoliths would result in a linear relationship between TL and OR (Schirripa, 2002). This relationship has been identified in the steelhead life-history variant, rainbow trout, however, it is not present in steelhead because the otoliths of adults show a different growth pattern compared to juveniles, possibly due to different phylogenetic factors (Rybock, 1973). Because this non-linear growth pattern was identified in

39 28 the steelhead otoliths that we measured, a quartic function was used for regression analysis. Usefulness of the regression equation in estimating fish TL from OR was supported by a high correlation coefficient and the significant relationship between fish TL and OR. It is still unclear why adult steelhead otoliths show a different growth pattern compared to juveniles, but an examination of otoliths from steelhead of intermediate size ( mm) would help provide additional information on the relationship between TL and OR in steelhead. Our results confirm that size at stocking impacts the survival of juvenile steelhead. The average size at stocking for adult steelhead from the New York hatchery was significantly greater than the average size of New York hatchery steelhead measured prior to stocking. This suggests that larger individuals from the New York hatchery are more likely to survive and make spawning runs, especially when you consider that <9% of the juvenile steelhead stocked by New York are 142 mm. Although larger New York hatchery fish were more likely to survive and spawn, the average back-calculated size at stocking for adult steelhead from the other Lake Erie hatcheries was still significantly larger than the average back-calculated size of adult steelhead from the New York hatchery. Survival of stocked juvenile steelhead has been shown to be positively correlated with size (Ward et al., 1989) so it is not surprising that larger individuals stocked by New York would have a greater chance of survival. To increase survivorship of hatchery steelhead, the New York stocking program should increase the average size at stocking by at least 20 mm, from ~120 mm to ~140 mm. Survival of stocked fish would likely be increased further by stocking steelhead at 168 mm, the average size at stocking for adult steelhead from other Lake Erie hatcheries. Although larger juvenile steelhead were more likely to survive and enter spawning tributaries than smaller steelhead, individuals as small as 107 mm were estimated to have

40 29 survived after stocking. Because steelhead stocked at this small a size would be unlikely to emigrate to Lake Erie in their first year (Seelbach, 1987), some second-year tributary survival is likely occurring despite the high mortality associated with high summer temperatures (Chrisp and Bjornn, 1978), and these smaller individuals, even if they do not emigrate to Lake Erie after their first year, are contributing to the fishery.

41 30 CHAPTER III: USE OF DUAL-FREQUENCY IDENTIFICATION SONAR (DIDSON) TO QUANTIFY EMIGRATION PATTERNS OF JUVENILE STEELHEAD STOCKED IN A LAKE ERIE TRIBUTARY 2 Introduction Each year approximately 2 million juvenile steelhead (Oncorhynchus mykiss) are stocked into Lake Erie tributaries to sustain a valuable sport fishery (CWTG, 2015). These fish are released with the expectation that they will emigrate to Lake Erie, grow to a size desired by anglers, and then return to the tributary in which they were stocked during spawning runs, where they become available for anglers. Ideally, stocked individuals emigrate shortly after release because juvenile salmonid mortality increases with increased emigration time (Seelbach et al., 1994; Welch et al., 2004; Melnychuk et al., 2007; Welch et al., 2008). Although increased tributary residence time increases mortality, imprinting of juvenile salmonid species to an olfactory cue can take days (Nevitt et al., 1994; Dittman et al., 1996; Yamamoto et al., 2010). Thus, to maximize return rates optimal stocking strategies would limit residence time, while allowing sufficient time for tributary imprinting. In addition to timing, stocking steelhead at the optimal size is important to maximize adult returns. Size-biased survival has been identified in stocked juvenile steelhead in Lake Michigan (Seelbach et al., 1994), and on the Pacific Coast juvenile steelhead stocked at a size of 180 mm have a three times higher survival rate compared to individuals stocked at 145 mm (Slaney et al., 1993). Survival of wild steelhead in the Pacific Ocean is also positively correlated with juvenile size, and a 10% increase in survival for every 10 mm increase in length has been identified in fish between the sizes of 140 and 260 mm (Ward et al., 1989). Individuals stocked 2 In reveiw as: Budnik, R. R., and J. G. Miner. Use of Dual-Frequency Identification Sonar (DIDSON) to Quantify Emigration Patterns of Juvenile Steelhead Stocked in a Lake Erie Tributary. Journal of Great Lakes Research.

42 31 at smaller sizes (< mm) are especially susceptible to mortality as they are less likely to become migratory in the first year after stocking (Wagner et al., 1963; Chrisp and Bjornn, 1978; Bjornn et al., 1979; Seelbach, 1987; Peven et al., 1994; Quinn, 2005). This failure to emigrate often exposes fish to low water levels and high water temperatures during summer months, leading to increased mortality (Chrisp and Bjornn, 1978; Seelbach, 1987). The average size at stocking for most agencies in the Lake Erie region is above the mm size threshold (In 2014: Michigan-193 mm, Pennsylvania-177 mm, and Ohio-171 mm); however, not all individuals are able to reach this size after a year of hatchery life and the Salmon River State Hatchery, which raises steelhead for release into New York s Lake Erie tributaries, consistently produces individuals with an average size at stocking of <150 mm TL (118 mm in 2014; CWTG, 2015). Besides influencing survival, size at stocking may also affect the timing at which juvenile steelhead emigrate. Earlier downstream movements have been identified in larger young-of-theyear brown trout (Salmo trutta; Holmes et al, 2014) and some work done in a Lake Erie tributary suggests larger juvenile steelhead emigrate earlier than smaller individuals (CWTG, 2014). Juvenile salmonid movement usually peaks in autumn or spring (Bjornn, 1971; Hayes, 1988; Jonsson, 1991; Hvidsten et al., 1995), and environmental variables such as photoperiod, temperature, and flow can initiate juvenile salmonid migration (Jonsson and Jonsson, 2011). Salmonid smolts also time emigration movements based on diel patterns, a strategy believed to be related to predator avoidance (Rieman et al., 1991; Poe et al., 1991; Leduc et al., 2010), turbidity (Gregory and Levings, 1998), and flow (Greenstreet, 1992). Here, we quantify emigration patterns of stocked yearling steelhead using Dual Frequency Identification Sonar (DIDSON), in a Lake Erie tributary that experiences elevated

43 32 water temperatures in mid-summer typical of most tributaries in Lake Erie. Our main objectives were to: 1) determine the viability of using DIDSON to monitor the movements of juvenile fish in a small stream system 2) estimate survival of juvenile steelhead stocked into a small Lake Erie tributary 3) quantify patterns of emigration timing 4) assess the role that environmental factors and individual size have on emigration timing. DIDSON allows fish emigration to be measured in real time without tagging and can be effective under most environmental conditions (i.e., high flow, high turbidity, nocturnal conditions). DIDSON is split beam sonar that records near video quality images in underwater environments using sound. In high frequency mode ( MHz) mode, fish as small as mm can be observed and measured (Holmes et al., 2006; Boswell et al., 2007; Doehring et al., 2011). DIDSON has been used to successfully monitor salmon in highabundance situations, such as in Alaska (Maxwell and Gove, 2004; Brazil and Buck, 2011), Idaho (Kucera and Faurot, 2005; Kucera and Orme, 2006), Washington (Galbreath and Barber, 2005), and British Columbia (Cronkite et al., 2006; Holmes et al., 2006). DIDSON has also been used to monitor steelhead in central California, providing accurate counts of steelhead compared with dam and weir counts (Pipal et al., 2010; Pipal et al., 2012). Although salmonid migration has been extensively studied using DIDSON, most work has focused on the movement of adults. Methods Study Area Trout Run is a small Lake Erie tributary located in Erie County, Pennsylvania (Figure 3). The stream is 10.4 km in length with an average width of 4 m, average channel depth of 0.5 m, and pools as deep as 1.2 m are present. Substrate is predominantly sand, silt, and small to medium sized gravel; however, from the lake, the first kilometer of stream section is

44 33 characterized by slate-shale bedrock bottom. Trout Run is protected from angling year-around because it serves as the primary source of broodstock for Pennsylvania s steelhead stocking program. Approximately 150 m from the mouth of the tributary is a manmade structure that serves to aggregate spawning adult steelhead. The structure has four concrete blocks (two on each side of the structure) that streamline flow and create a 4.2 X 6.1 m pool with a concrete bottom. The lower end of the structure (closer to the mouth of tributary) creates a small waterfall that is impassable to any fish that are unable to jump over the barrier which is approximately 2 m high (Figure 4). Annually, juvenile steelhead stocking into Trout Run takes place in early April and occurred on April 4, 7, and 10 in 2014, and April 8 and 9 in Stocking occurred approximately 2 km upstream of the tributary mouth (Figure 1) and a total of 46,250 individuals were released each year. The total length (TL) of a sub-sample of fish was measured prior to stocking in both 2014 (n=50) and 2015 (n = 50) and the mean TL S.E. was determined for all fish measured. Field Methods and Equipment A DIDSON (Sound Metrics, Lake Forest Park, Washington) operating in high frequency mode (1.8 MHz) was placed 150 m from the mouth of Trout Run within the manmade structure described above (Figures 3 and 4). In this mode, an image is created with a 29 horizontal and a 12 vertical field of view. A window length (length of the ensonified view) of 5 m was used with a window start of 0.42 m (i.e., any targets that passed within 0.42 m of the DIDSON were undetectable). Data was captured at 7 frames s -1 and focus and receiver gains were automatically set. The DIDSON was connected to a laptop running DIDSON software (Sound Metrics, Lake

45 34 Forest Park, Washington. Version ) and was cradled in a small chassis. The chassis was attached to a large metal frame to prevent movement and aimed slightly downward into the substrate to ensonify the stream bottom. Inside the manmade structure, stream width did not change, however depth was variable. Depth varied from 0.3 m 1.3 m depending on flow, and the water surface was ensonified during the entire study. Footage was recorded from April 3 May 12, 2014 and April 7- June 23, During the first six days of the study in 2014, two non-optimal DIDSON configurations were used. Because these configurations caused milling fish to repeatedly move in and out of the viewing cone, they were abandoned for a third configuration that reduced the number of milling fish and maximized our ability to count individuals as they emigrated over the barrier (Figures 4). Along with DIDSON footage, water temperature ( C), discharge (m 3 /s; Marsh-McBirney, Model 2000 Flo-Mate, Frederick, Maryland), precipitation (cm), cloud cover (%), moon illumination (%), and photoperiod (mins of daylight) were recorded for each date. Precipitation and cloud cover data was obtained from the National Climatic Data Center station at the Erie International Airport (Erie, PA) which is <8 km from Trout Run. After the DIDSON was removed, an electrofishing survey was performed on June 11, 2014 and June 22, Trout Run was divided into three sections; a lower section that was characterized by slate-shale bedrock bottom, along with a middle and upper section where substrate was predominantly sand, silt, and small to medium sized gravel. The lower section began at the DIDSON location and extended 1.75 km upstream (Figure 2). The middle section extended from the end of the lower section to the stocking site, and the upper section extended from the stocking site to a barrier where additional upstream migration was unlikely (Figure 2). To estimate the residual population of steelhead in Trout Run, two 50-meter portions of each

46 35 tributary section were sampled with equal effort using backpack electrofishing (Aquashock Model AP1, Sevier County, Tennessee). A population estimate was then determined for each tributary section using the depletion method (Lockwood and Schneider, 2000) and a total population estimate was determined by extrapolation for each reach. All juvenile steelhead captured during electrofishing surveys were removed, placed in live wells, counted, and measured for total length. After sampling of each tributary section was completed, individuals were released. Data Analysis DIDSON footage was saved in one-hour files and each file was visually reviewed to estimate the number of juvenile steelhead emigrating per hour. All footage was evaluated by the same reviewer (reviewer 1) in 2014 and one of three reviewers (reviewers 1, 2, or 3) evaluated each file from Reviews were conducted at an initial playback speed of 100 frames per second, which is 14 times the speed of recording (7 frames/second). Because aggregations of fish are more difficult to enumerate during DIDSON file processing (Able et al., 2014), files that contained many targets in the view frame at once were slowed to ensure each emigrating fish could be counted. Targets were counted as emigrants if they exited the field of view on the downstream side of the screen (over the barrier). Milling fish were not counted until they exited the field of view. In both years recording began approximately 24 hours before stocking and footage was reviewed to ensure no other species that could be misidentified as steelhead, due to similar size, were present in the study area. Additionally, visual observations of fish in the pool were possible

47 36 during daylight hours. Besides steelhead, only small minnows (Cyprinidae spp.) and a catfish (Ictaluridae spp.) were visually identified in the pool, and all individuals were <90 mm. To verify DIDSON counts, a seine was placed directly downstream of the study area, below the waterfall, during hour-long periods (n=6) when DIDSON footage was recorded. The seine covered the entire width and depth of Trout Run and was used to collect fish emigrating over the barrier. The most experienced reviewer (reviewer 1) analyzed all files associated with these seining periods and a linear statistical model (Neter et al., 1985) was fitted to assess potential differences between DIDSON and seine counts. If DIDSON and seine counts agreed, then the regression slope would be coincident with a line with a slope of 1.0. To identify potential bias among reviewers, a set of the same hour-long DIDSON files (n=15) were evaluated by each reviewer. These files contained either a low (~1-19; n = 5), intermediate (~20-90; n = 5), or high (~90+; n = 5) number of targets and counts estimated for each file were used to develop linear regressions of less experienced reviewer counts (reviewer 2 or 3) versus the counts estimated by reviewer 1. Once developed, the regression slopes were compared to 1.0. If counts resulted in a regression slope not coincident with 1.0, all files analyzed by the less-experienced reviewer were corrected using the regression equation. These corrected hourly counts were then used to calculate diel totals, along with daily and yearly emigration numbers. The percent emigration (%E) and percent survival (%S) were determined for each year using the following equations: % = T % = ( + ) T where, s = estimated # of surviving individuals remaining in Trout Run e = estimated # of emigrants

48 37 T = total # of individuals stocked Using the manual fish-measuring feature included with the DIDSON software, DIDSONbased-total-lengths (DTLs) were estimated for 100 randomly selected individuals from each day of recording. The size range of juvenile steelhead stocked into Trout Run by the Pennsylvania Fish and Boat Commission (PFBC) can range from mm (CWTG, 2013; 2014; 2015; prestocking length measurements), thus, only fish between mm were included in our analysis. This excluded adult steelhead that could jump the barrier and any fish that were below the smallest size steelhead were stocked. Interference (noise) from stationary structures was removed from the images by using DIDSON s algorithm for dynamic background removal, set to default parameters (factor A ¼ 0.95, factor B ¼ 0.05). To achieve the best possible estimates of fish length, efforts were made to measure during frames when fish were actively moving and when fish appeared to display their full length (Burwen et al., 2010). To validate the accuracy of the DIDSON fish-measuring feature, and the above procedures for selecting images for length estimation, the study area was cleared of any fish and a juvenile steelhead of known TL was added to the pool. DIDSON footage was recorded until the fish swam through the field of view and the DTL of the fish was measured. This process was repeated with different known length juvenile steelhead (n=28) and a linear statistical model (Neter et al., 1985) was fitted to assess the relationship between DTL and TL measurements. If DTL and TL values agreed then the slope of the regression would be coincident with a line with a slope of 1, therefore, to ensure DTL estimates were accurate we determined if the slope of the regression was significantly different than 1. Water temperature ( C), discharge (m 3 /s), precipitation (cm), moon illumination (% moon lit), cloud cover (%), and photoperiod (mins of daylight) were selected as potential

49 38 predictor variables for emigration rate (estimated daily number of emigrants/maximum possible number of individuals remaining in tributary). Because of the non-parametric and non-linear nature of our data, non-linear regression models (cubic, exponential, logistic, quadratic, Gaussian, and Lorentz) were used to determine the relationship between each variable and emigration rate. The best fitting regression model for each variable was selected using AIC, where AIC values were adjusted for small sample size (AICc; Burnham and Anderson, 2002). AICc weights (AICc Wt.) were calculated for each variable and the model with the lowest AICc and the highest AICc Wt was considered the best model. Once the best non-linear fit was identified for each variable, the best overall predictor variable was determined using r 2. Correlations between variables were identified using Spearman s correlation test (p < 0.05), and if a correlation was identified between two variables, only the variable with the highest r 2 was included in the final analysis. An attempt was also made to develop multiple regression models, but because these models only slightly improved predicting power (all models <4 % increase in r 2 compared to best fitting single independent variable model) they were not considered when ranking predictor models. Results In 2014, 33,088 emigrants were identified during DIDSON video processing (Figure 5) and 40,438 were identified in A total of 2,208 hours of footage were recorded during the study with 739 hours recorded in 2014 and 1,469 hours recorded in DIDSON counts estimated by reviewer 1 had a strong linear relationship with seine counts (y = x , r 2 = 0.99, RMSE = 1.10), and the slope was not significantly different than 1.0 (one-way ANOVA, p = 0.07). Strong linear relationships were also identified among reviewers (r2 vs r1: y

50 39 = x , r 2 = 0.99, RMSE = 8.15; r3 vs r1: y = 1.085x , r 2 = 0.99, RMSE = 4.20), but counts of the less-experienced reviewers were corrected because the resulting regression slopes were significantly different than 1.0 (both comparisons one-way ANOVA, p < 0.001). Both regression intercepts of the reviewer comparisons were negative, and the slopes were slightly greater than one. Thus, there was a slight positive bias in the DIDSON files estimated to have 18 targets analyzed by reviewer 2 and in those estimated to have 9 targets reviewed by reviewer 3. For example, the corrected count, based on DIDSON files analyzed by reviewer 2, would be 53 if 50 fish were counted; and the expected count, based on files analyzed by reviewer 3, would be 54 if 50 fish were counted. When corrected for reviewer bias, the total number of emigrants estimated for 2015 was 41,712 (Figure 5), an increase of 3.1%. The first emigrating fish were detected within one hour of stocking in both years, and an increase in the daily number of emigrants began one week after stocking with peak emigration occurring after 3 weeks. A sharp decline in emigration occurred in the first week of May in 2015 and a similar decline appeared to take place in 2014, although the DIDSON was removed before this could be confirmed. A greater number of hatchery steelhead were detected during nocturnal periods in both years (2014: 68%; 2015: 51%; Figure 5) even though this period was after the vernal equinox. Overall, the average number of hatchery steelhead detected per hour was significantly greater during night than day periods (Kruskal-Wallis, Z = 9.76, p < 0.001). The maximum number of hourly and daily emigrants were 291 and 2,777, respectively, in 2014 and 276 and 3,253 in 2015 (Figure 5). The %E and %S were lowered in 2014 compared to 2015 (2014: %E = 71.5, %S = 77.8 ± 0.5; 2015: %E = 90.2, %S = 91.5 ± 0.2), thus, stocked juvenile steelhead remaining in Trout Run at the time of June population estimation were more abundant in 2014 (June 11: 2,888 ± 187) than in 2015 (June 22: 592 ± 98).

51 40 Lengths measured manually from DIDSON images had a strong linear relationship with actual TL (y = 0.91x , r 2 = 0.94, RMSE = 6.99), and the slope of the regression line was not significantly different than 1.0 (one-way ANOVA, p = 0.06). The DTL of emigrant fish was significantly greater than fish measured prior to stocking (one-tailed t-test: t = 5.44, P < 0.001). Additionally, the lengths of fish captured during electrofishing surveys (remaining fish) were significantly smaller than the DTL of emigrated fish (one-tailed t-test: t = , P < 0.001) and the length of fish that were measured prior to stocking (one-tailed t-test: t = , P < 0.001). Although the average size of emigrant fish was larger than those measured prior to stocking, the proportion of individuals that were 150 mm was not significantly different between groups (Prior to stocking 89.1%; Emigrants 90.1%; Pearson s χ 2, P = 0.609); however, the proportion of individuals 150 mm that remained in Trout Run after the study period was significantly lower compared to pre-stocking and emigrant fish (36.9%; Pearson s χ 2, both comparisons P < 0.001; Figure 6a). Individuals that emigrated during the first two weeks of tributary residence were significantly larger than fish that emigrated after this period (pairwise Tukey-Kramer HSD, 10+ weeks P = 0.082; all other comparisons P < 0.001), and fish that emigrated 2-4 weeks after stocking were significantly larger than those with a tributary residence time of 4-10 weeks (pairwise Tukey-Kramer HSD, 10+ weeks P = 0.202; all other comparisons P < 0.001). The size range of emigrating individuals was wide, but excluding outliers (10 th to 90 th percentile), there was a narrowing in the size range of emigrating fish as the study progressed (0-2 weeks: mm; 2-4 weeks: mm; 4-6 weeks: ; 6-8 weeks: ; 8-10 weeks: ; 10+ weeks: ). This narrowing occurred in the upper portion of the size ranges as the lower portions remained relatively stable (total change -3 mm; Figure 4b).

52 41 Timing of emigration was also analyzed as a function of several abiotic factors. First, significant correlations between photoperiod and temperature (Spearman s rho = ; p < 0.001) and cloud cover and precipitation (Spearman s rho = 0.640; p < 0.001) were identified. Therefore, only photoperiod and precipitation were considered when ranking the importance of variables as the best fitting model because each of these variables performed better than temperature and cloud cover, respectively. We found that a Lorentzian peak regression (Johnson et al., 1994) using photoperiod as the independent variable explained the most variation in emigration rate (r 2 = 0.72; Table 7; Figure 7). The equation for the Lorentzian peak regression model was: (a ) [( h ) + ] Where: a = Peak Value b = Growth Rate c = Critical point The practical utility of the remaining non-linear models was limited for each variable, as none explained more than 18% of the variation in emigration rate. Discussion In both years of our study, at least 71% of the steelhead stocked into Trout Run were detected emigrating out of the tributary. These estimates of emigrating individuals are conservative as additional steelhead likely emigrated out of Trout Run during periods when the DIDSON was not operational. The high number of emigrating individuals detected suggests that the percentage of stocked juvenile steelhead that emigrate out of small tributaries is very high

53 42 compared to those stocked into larger tributary systems. Bjornn et al. (1978) estimated percent emigration of stocked juvenile steelhead to be 32-54% in the Columbia and Snake Rivers, while Seelbach (1987) working in a large Lake Michigan tributary estimated percent emigration to be 47%. Percent emigration is likely high in Trout Run because of the small size of the tributary and the short emigration distance that fish need to travel to reach Lake Erie. DIDSON and electrofishing both failed to identify any large piscivores in Trout Run, except for a small number of adult steelhead that were identified in the proximity of the DIDSON on several occasions. No predatory behavior was observed by these adult steelhead and juvenile steelhead retreated upstream of the DIDSON pool when adults were present. These observations and the lack of other large piscivores in Trout Run suggests that aquatic predation is minimal and any additional predation on stocked juvenile steelhead likely comes from terrestrial sources. Although juvenile steelhead mortality does increase with increased emigration distance (Seelbach et al., 1994; Welch et al., 2004; Melnychuk et al., 2007; Welch et al., 2008), the high survival-to-emigration of steelhead in Trout Run supports the practice by Great Lakes fishery managers of stocking a considerable distance upstream in these typically short tributaries to maximize the potential for imprinting while experiencing relatively low mortality. Photoperiod was the best predictor of steelhead emigration timing. Because peak emigration occurred when photoperiod was ~840 minutes (Apr 30), it may be beneficial for managers to stock early enough for steelhead to have sufficient time to imprint before this photoperiod is reached. Wagner (1974) identified photoperiod as the main environmental factor controlling the onset of the parr-smolt transformation in steelhead, with increasing day length as the main photoperiod constituent stimulating smolting. Wedemeyer (1980) reviewing the early literature on smolting in hatchery salmonids also identified photoperiod as the primary factor

54 43 influencing smolt migration. Water temperature can also influence the timing and amount of smolting that occur (Wagner, 1974; Wedemeyer, 1980), but any influence temperature had on the current study could not be determined due to a high correlation with photoperiod. Within the correct season and temperature range, the rate of migration for juvenile salmonids can also be directly related to water velocity (Raymond, 1968; Smith et al., 2002; Connor et al., 2003), and Carlsen et al. (2004) reported that juveniles anticipate increases in water depth and have increased movement as water levels rise. In Trout Run, no correlation between increased discharge (r 2 = 0.18), or periodicity of discharge, and emigration rate were identified. Although it is unclear why discharge did not influence emigration, the small size of Trout Run may have been a factor, as other instances of discharge influencing emigration have occurred in larger river systems. The period and timing of emigration observed here were like those found in both Great Lakes and Pacific Coast tributaries (Figure 5; Wagner, 1968; Bjornn et al., 1979; Seelbach, 1987). Peak emigration timing of approximately 3 weeks suggests that most individuals spend the appropriate amount of time needed to imprint in Trout Run; however, many individuals (11.7% of total emigrants) emigrated <10 days after stocking, which is less than the time needed for juvenile salmonids to imprint on olfactory cues (Nevitt et al., 1994; Dittman et al., 1996; Yamamoto et al., 2010). This early emigration may lead to poor imprinting and could potentially lead these early-emigrating individuals to stray more frequently. Westley et al. (2013) identified an increased propensity of adult steelhead to stray when displacement (distance from hatchery to stocking site) was increased, thus, the importance of tributary residence time is likely dependent on the displacement of fish being stocked. Managers that transport hatchery fish considerable

55 44 distances may increase imprinting by stocking their fish earlier in the year and by stocking as far upstream as is practical. It is unclear how early emigration influences overall survival of steelhead stocked in Trout Run but Romer et. al (2012) found that juvenile steelhead who emigrated during peak emigration periods in Oregon had higher estuary survival, possibly due to survival benefits associated with schooling behavior. Conversely, early emigration of hatchery steelhead could provide a fitness advantage, as early emigrating fish would gain access to open water foraging opportunities sooner than fish that emigrated later. If this is true in Lake Erie steelhead, it may be advantageous for larger individuals, who would be less susceptible to predation due to their large size, to emigrate earlier than smaller individuals. We provide some evidence for size dependent emigration timing, as larger steelhead were more likely to emigrate earlier than their smaller counterparts. Although this size effect was significant, the size range of fish that emigrated throughout the study period was very wide and implies that other factors are also influencing emigration timing. Evidence for an emigration size threshold of mm was provided by the fact that most juvenile steelhead remaining in Trout Run at the end of each study year were small, with 63% of individuals having a total length of < 150 mm. This suggests that juvenile steelhead stocked below 150 mm are more likely to remain in Lake Erie tributaries for an additional year where they are exposed to a high mortality risk due to summer tributary conditions. The New York State stocking program may be especially susceptible to low rates of immediate emigration because the average size of stocked individuals is only ~120 mm annually (CWTG, 2015). Although the non-emigrant juvenile steelhead remaining in Trout Run were significantly smaller than fish measured in the other groups, size is certainly not the only factor that influences the

56 45 decision for fish to emigrate as individuals as large as 200 mm were present in Trout Run in June. Along with size and photoperiod, diel patterns also influenced steelhead emigration and the number of juvenile steelhead detections was significantly higher during nocturnal periods. In the Sacramento/San Joaquin watershed, migrating juvenile steelhead exhibited variable diel preferences depending on river location with the amount of nighttime versus daytime detections of emigrating juvenile steelhead varying from 41-63% (Chapman et al., 2013). The amount of nighttime versus daytime detections in our study fell in the higher part of this range with 59% of detections occurring at night. This propensity to emigrate during nighttime periods is likely a predator avoidance strategy, which has been implicated for juvenile salmonids in several systems (Rieman et al., 1991; Poe et al., 1991; Leduc et al., 2010). Our findings provide useful information for managers stocking juvenile steelhead and indicate that multiple factors, including photoperiod and length at stocking, are primary factors in emigration timing. The methods presented here were effective in quantifying the emigration of juvenile steelhead in real time, providing constant information which has been lacking in other studies on salmonid emigration.

57 46 CHAPTER IV: TRASH OR TREASURE? USE OF SAGITTAL OTOLITHS PARTIALLY COMPOSED OF VATERITE FOR HATCHERY STOCK DISCRIMINATION IN STEELHEAD Introduction Otoliths are bony structures used for hearing, balance, gravity sensation, and linear acceleration in teleost fishes (Fay, 1980; Popper and Lu, 2000). Three different otolith types exist, the sagittae, asteriscus, and lapillus; each contained in an interconnected chamber within the inner ear (Dale, 1976; Moyle and Cech, 1996). All otoliths are composed of calcium carbonate, and different crystal polymorphs of calcium carbonate are linked with each otolith type. Calcium carbonite is most commonly deposited as vaterite in the asteriscus, and in the lapillus and sagittae, aragonite is usually deposited (Campana, 1999; Falini et al., 2005). Although these polymorph associations exist, the composition of sagittal otoliths can switch from aragonite to vaterite, and vaterite replacement has been identified in several salmonid species including chinook salmon (Oncorhynchus tshawyscha; Gauldie, 1986; 1996), coho salmon (Oncorhynchus kisutch; Gauldie et al., 1997), lake trout (Salvelinus namaycush; Melancon et al., 2005) and steelhead (Oncorhynchus mykiss; Campana, 1983). Vaterite replacement impairs sagittal otolith function leading to significant hearing loss (Oxman et al., 2007; Reimer et al., 2016). The consequences of this hearing loss for predator aversion and mortality are unclear (Oxman et al., 2007; Reimer et al., 2016), but asymmetrical otoliths, like those formed during vaterite replacement, can lead to greater difficulties in detecting suitable settlement habitats in larval fish (Gagliano et al., 2008). If otolith asymmetry limits navigation and habitat selection in later life stages, the asymmetry in otoliths due to

58 47 vaterite replacement could contribute to the low ocean survival typical of hatchery salmonids (Beamish et al., 2012; Moore et al., 2012). Generally, fewer than 10% of wild fish have sagittal otoliths that contain vaterite, but much higher occurrences have been identified (Morat et al., 2008). The prevalence of vaterite otoliths in hatchery fish is also higher than in wild fish, and percentages as high as 50-60% have been reported (Watson, 1964; Peck, 1970; Bowen et al., 1999; Sweeting et al., 2003). Reimer et al. (2016) reviewing the existing literature for vaterite prevalence in sagittal otoliths found that hatchery populations had a 10.4 times higher incidence of vaterite compared to wild populations, regardless of species. They also identified a 3.7 times higher vaterite prevalence in hatchery raised Atlantic salmon, compared to wild fish, and verified that vateritic otoliths are common in hatchery salmon worldwide [Atlantic salmon (Salmo salar) - Australia 57%; Scotland 58%; Canada 30%; and Chile 64%: Rainbow trout (Oncorhynchus mykiss) Chile 48%]. Because otoliths exhibit continual growth, do not reabsorb, and reflect ambient water chemistry within their microchemical composition (Campana,1999; Melancon et al., 2008), they can provide chronological information on the environment in which fish have inhabited. Otolith chemistry has proven to be an effective way to discriminate stocks of fish originating in distinct geographic locations (Campana, 1999; Brazner et al., 2004; Ludsin, 2006; Elsdon et al., 2008; Hand et al., 2008; Pangle et al., 2010; Reichert et al., 2010; Hayden et al., 2011) and laserablation-inductively-coupled-plasma-mass-spectrometry (LA-ICP-MS) has been used to analyze the chemistry of different otolith sections, allowing the detection of small-scale differences in otolith signatures, such as temporal changes in hatchery water sources (Marklevitz et al., 2011; Boehler et al., 2012).

59 48 Sagittae are most commonly used for otolith chemistry and although proven to be effective, it is recommended that sagittae containing vaterite be removed from analyses due to variation in otolith chemical signatures between aragonite and vaterite otoliths (Melancon et al., 2005; Zimmerman and Nielson, 2003). Because sagittal otoliths containing vaterite are not uncommon in fish, these omissions cause large proportions of individuals to be excluded from analyses, and life history information about fish with vaterite otoliths to be lost. Developing a technique to use vateritic otoliths for similar stock discriminations as those performed with aragonite otoliths would be beneficial to fishery managers as it would provide additional information on the contribution of individuals from different geographic locations to fish populations. Here we investigate the prevalence of aragonite versus vaterite sagittae in steelhead raised in Lake Erie hatcheries. We then present a technique to use saggital otoliths that have transitioned from aragonite to vaterite to discriminate the otolith chemistry signatures of steelhead from different hatchery sources in Lake Erie. Methods Hatchery yearling steelhead (mean total length and standard error = 159 ± 6.3 mm; range = mm) were provided by managers and personnel from the state hatchery systems in Lake Erie: Castalia State Fish Hatchery (Ohio; n = 15), Fairview State Fish Hatchery (Pennsylvania; n = 15), Salmon River State Fish Hatchery (New York; n = 17), Tionesta State Fish Hatchery (Pennsylvania; n = 15), and Wolf Lake State Fish Hatchery (Michigan; n = 15) before stocking in Once steelhead were collected, sagittal otoliths were removed and prepared for microchemical analysis. Procedures for preparation of otoliths and subsequent chemical analysis have been presented previously (e.g. see Secor et al., 1991; Hayden et al.,

60 ; Boehler et al., 2012). Briefly, otoliths were cleaned of organic material by sonication in hydrogen peroxide (3% V:V), and air dried. Otoliths were embedded in a two-part epoxy (West System 105 Epoxy Resin and 206 Slow Hardener ) and sectioned in the transverse plane using a low-speed wafer saw with diamond-tipped blade. Both sides of the cut otolith crosssections were wet polished using ultrapure (Milli-Q ) water with 3M -brand silicon carbide sandpaper and lapping film (particle size: 20 μm, 10 μm, 6 μm, 2 μm) to a thickness of approximately 200 μm. Polished otoliths were mounted on standard petrographic microscope slides using epoxy. Mounted slides were triple-rinsed and sonicated for five minutes with ultrapure water, then covered and allowed to dry overnight, and stored in clean Petri dishes until the analysis was performed. LA-ICP-MS analysis was performed on otoliths using a PhotonMachines 193 nm laser coupled to the Agilent 7900 fast-scanning ICP-QMS at the Great Lakes Institute for Environmental Research (GLIER, University of Windsor, ON). A description of operating conditions can be found in Boehler et al. (2012). The theoretical concentration of calcium in stoichiometric calcium carbonate (400,432 μg Ca g -1 CaCO3) was used as an internal standard to correct for ablation yield differences between external calibration standard and the otoliths, and the occurrence of mass 120 (measured as 120 Sn, tin isotope), an indicator of contamination from the epoxy bonding medium, was also quantified (Ludsin et al., 2006; Reichert et al., 2010). Data processing and calculations of detection limits were performed using a Microsoft Excel spreadsheet macro (Yang, 2003) based on algorithms developed by Longerich et al. (1996). Chemical information on four elements was collected: barium (Ba), magnesium (Mg), manganese (Mn), and strontium (Sr). These elements were selected as they have previously been used in stock discrimination in Lake Erie fishes (Brazner et al., 2004; Ludsin et al., 2006; Hand

61 50 et al., 2008; Pangle et al., 2010; Boehler et al., 2012). The concentrations of all elements were above background and always greater than the limits of detection. To characterize variation in the otolith elemental signatures of Lake Erie steelhead, otolith sections along the LA-ICP-MS traverse were standardized into 5% regions from core (0%) to edge (100%) and average elemental concentrations of all elements were calculated for each region (Boehler et al., 2012). In each otolith, 5% regions were classified as aragonite, vaterite, or transitional (transition from aragonite to vaterite; Table 8). Vateritic otolith sections have characteristically low levels of Sr and Ba and high levels of Mg (Gauldie, 1996; Melancon et al., 2005; Gibson-Reinemer et al., 2009), therefore a switch from aragonite to vaterite was identified when Sr and Ba levels dropped rapidly and Mg levels increased sharply (Figure 8). Once otolith regions were classified for each fish, only the regions composed of entirely of vaterite in all individuals were used for further analysis. The three outer sections of all otoliths were vaterite (85-90, 90-95, and %) resulting in 3 variables for each element (12 total). Wilcoxon tests with Bonferroni adjustments were used to test for differences in the elemental signatures from each hatchery sources. To test our ability to discriminate among hatchery sources all variables were entered into a quadratic discriminant function analysis (QDFA) and systematically removed with backwards stepwise variable selection until the combination of variables yielding the lowest number of misclassified fish was obtained. A QDFA was used because this procedure does not require data to meet the assumptions of homogeneity of variance and multivariate normality (McGarigal et al., 2000). In total, 5 otolith region variables were included in the final QDFA. The QDFA used a jackknife cross-validation procedure to determine classification accuracy.

62 51 Results The overall prevalence of vateritc otoliths in Lake Erie hatchery steelhead was 29.9% (23/77) and proportions of vateritic individuals for each hatchery were not significantly different (χ 2, all p > 0.05). The highest prevalence of vaterite otoliths were identified in steelhead from the Salmon River, NY hatchery (47%), followed by Fairview, PA (40%), Wolf Lake, MI (26.7%), Tionesta, PA (21.4%), and Castalia, OH (13.3%). Steelhead from the five Lake Erie hatcheries showed a large range of elemental concentrations in their otoliths (Table 9). When the three outer regions of vaterite otoliths were pooled (85-100%) the average elemental concentrations of all elements except Ba were significantly different among hatcheries (Wilcoxon; Ba, p = 0.08; all others p < 0.05). Salmon River fish had higher concentrations of Mg in their 85-90% and 90-95% regions compared to Pennsylvania hatchery fish and in the 90-95% region compared to Wolf Lake individuals. Salmon River fish also had higher levels of Mn in all outer regions compared to Tionesta individuals and in the 90-95% region compared to Fairview individuals. In Wolf Lake fish, higher concentrations of Sr were identified in all outer regions compared to Fairview and Salmon River individuals. Although the pooled average concentrations of Ba in the % region were not significantly different among hatcheries, Ba levels were higher in Wolf Lake compared to Salmon River in the 90-95% regions, and in the 85-90% regions Ba levels were lower in Salmon River individuals when compared to all other hatcheries, except Castalia. QDFA was very successful in discriminating among hatchery sources (Figure 9) and cross-validated classification accuracy was 100% for all sources (Table 10). Canonical axes accounted for 99% (CV1; p < 0.05; eigen value = 400.9), 0.6% (CV2; eigen value = 2.54), and 0.4% (CV3 eigen value = 1.58) of the total variances (all p < 0.05). The first canonical axis was

63 52 mainly driven by Ba (85-90% region), and Sr also contributed (85-90% and % regions). Loadings on the second and third canonical axes were again driven by Ba (85-90% region), and to a lesser extent Mn (95-100% region) (Table 11). Discussion The prevalence of vaterite in Lake Erie hatchery steelhead otoliths was low (30%) compared to farm-raised rainbow trout in Chili (48%) and other farmed salmonid species throughout the world (48-60%; Reimer et al., 2016). Conversely, the prevalence was high when compared to percentages reported by Boehler et al. (20%; 2012), who identified vaterite otoliths in adult and juvenile steelhead in Lake Erie. Although differences in the proportion of individuals from each hatchery that contained vaterite otoliths were not significantly different, sample sizes were limited. It is entirely possible that due to the large range of vaterite proportions found in the Lake Erie hatcheries (13-47%), fish in certain hatcheries (specifically, Salmon River, NY and Fairview, PA) may be more susceptible to vaterite transformation, possibly due to increased stress. Some investigators have hypothesized that exposure to acute or chronic stress deriving from mechanical trauma (Strong et al., 1986), starvation (Payan et al., 2004), density effects (Casselman, 1990), or temperature fluctuation (Johansson, 1966) can promote the development vaterite otoliths. Thus, records of temperature, feeding, and the number of individuals produced from each hatchery could provide insight into potential vaterite formation mechanisms. While sagittal otolith chemistry has been used extensively to discriminate fish stocks (Campana, 1999; Elsdon et al., 2008; Pangle et al., 2010; Reichert et al., 2010; Hayden et al., 2011), including fish from Great Lakes hatcheries (Marklevitz et al., 2011; Boehler et al., 2012;

64 53 Chapter I), we are unaware of any studies that have used vateritic sagittal otoliths for stock discrimination. Our analyses revealed variation in the elemental signatures of steelhead from the various Lake Erie hatcheries, even when only the outer 15% of otoliths were analyzed. Of the four elements used for hatchery source discrimination, Ba had the greatest among hatchery variation (when split into 5% regions) and proved to be most useful for QDFA. Ba has shown to be useful in discriminating Lake Erie hatchery stocks previously (Boehler et al., 2012; Chapter I), especially when discriminating between Pennsylvania hatchery fish and those from the other Lake Erie hatcheries. Sr also proved useful in our analyses, which is not surprising as Sr has shown to be an important discriminator for Lake Erie fish stocks, due to a west (high) to east (low) Sr gradient that exists in Lake Erie waters (Pangle et al., 2010). In Castalia hatchery steelhead otoliths, levels of Sr 2.5 to 6 times higher than other Lake Erie sources have been identified (Boehler et al., 2012; Chapter I). Although average Sr concentrations were highest in Castalia steelhead otoliths, the small number of Castalia individuals analyzed (n = 2) likely limited our ability to determine if these Sr concentrations were significantly higher when compared to the other hatchery sources. Mg and Mn were also significantly different among some hatcheries, but did not weigh heavily into the QDFA. Using vateritic otoliths to produce unique otolith chemical signatures for Lake Erie hatchery steelhead was successful and could likely be used in any system where otolith chemistry is an effective method for stock discrimination. The ability to make use of all study animals is certainly an advantage as large amounts of data are often lost when fish with vateritic sagittal otoliths are removed from analysis. Additionally, the identification of the source population of individuals with vaterite otoliths, using the procedures described here, could

65 54 potentially provide opportunities to investigate the causes of vaterite replacement and its effect on fish survival and fitness.

66 55 CHAPTER V: SUMMARY AND CONCLUSIONS The Lake Erie steelhead fishery is an invaluable resource for local communities in the Great Lakes region. Because of the immense time and effort put into sustaining this fishery, research that can inform decision makers on best management practices (including the work within this dissertation) is vitally important. Our otolith chemical analyses presented in Chapter I, proved effective in discriminating hatchery stocks of steelhead, even when fish from multiple year-cohorts were used. The low occurrence of strays identified in 16-mile Creek, PA and the Vermilion River, OH suggest that fish stocked by OH and PA are exhibiting high stocking sitefidelity. Even so, the large proportion of OH and PA fish present in Chautauqua and Cattaraugus Creeks, NY suggests that potentially large numbers of individuals from OH and PA are choosing to spawn in NY tributaries. This may partially be a symptom of the large number of individuals stocked by PA and OH but our results from back-calculation of age at stocking, reported in Chapter II, confirm that return rates for fish stocked at a size <140 mm by NY are poor. The results obtained from using DIDSON, and reported in Chapter III, provide additional support for stocking at large sizes, as juvenile steelhead stocked below 150 mm were more likely to remain in Lake Erie tributaries for an additional year where they are exposed to a high mortality risk due to high summer water temperatures. Other results presented in Chapter III, suggest that within-tributary mortality is minimal when steelhead are stocked into small Lake Erie tributaries and that date of stocking, specifically in relation to photoperiod, should be considered when trying to optimize release timing. In the final chapter of this dissertation a technique to perform stock discrimination for steelhead in Lake Erie using vateritic sagittal otoliths is presented. This procedure allows the successful discrimination of steelhead with vaterite otoliths allowing managers to understand the

67 56 prevelance of vaterite otoliths in Lake Erie steelhead and the contribution of these individuals to the fishery. Using this procedure also prevents the loss of valuable data from fish that have vaterite otoliths. The method presented here could also help provide insights into potential negative fitness consequences and the causes of vaterite deposition, which are still unclear. As demonstrated by the research within this dissertation, straying within steelhead is complex and can be highly variable when multiple state agencies, using different methods, stock into the same water body. We suggest that future research focus on long-term patterns over a larger spatial scale to test the robustness of results. The findings presented here provide evidence for some of the mechanisms of straying of steelhead in Lake Erie and should improve our understanding of this economically important fishery.

68 57 REFERENCES Able, K. W., Grothues, T. M., Rackovan, J. L., and Buderman, F. E Application of mobile dual-frequency identification sonar (DIDSON) to fish estuarine habitats. Northeast. Nat. 21: Adkison, M. D Population differentiation in Pacific salmon: local adaptation, genetic drift, or the environment? Can. J. Fish. Aquat. Sci. 52: Bagenal, T. B. and Tesch, F. W Age and growth. In: Methods for assessment of fish production in freshwater, vol. 3. (T. B. Bagenal, ed). Blackwell Scientific Publications, Oxford. Pp Barber, W. E., and McFarlane, G. A Evaluation of three techniques to age Arctic char from Alaskan and Canadian waters. Trans. Am. Fish. Soc. 116: Barnett-Johnson, R., Teel, D. J., and Casillas, E Genetic and otolith isotopic markers identify salmon populations in the Columbia River at broad and fine geographic scales. Environ. Biol. Fish. 89: Beamish, R. J., Sweeting, R. M., Neville, C. M., Beacham, T. D., and Preikshot, D Wild chinook salmon survive better than hatchery salmon in a period of poor production. Environ. Biol. Fishes 94: Biette, R. M., Dodge, D. P., Hassinger, R. L., and Stauffer, T. M Life history and timing of migrations and spawning behavior of rainbow trout (Salmo gairdneri) populations of the Great Lakes. Can. J. Fish. Aquat. Sci. 38: Bjornn, T.C Trout and salmon movements in two Idaho streams as related to temperature, food, stream flow, cover, and population density. Trans. Am. Fish. Soc. 100:

69 58 Bjornn, T.C Survival, production, and yield of trout and chinook salmon in the Lemhi River, Idaho: A Final Report for Federal Aid to Fish Restoration, Project F-49-R, Salmon and Steelhead Investigations. Idaho Cooperative Fishery Research Unit, College of Forestry, Wildlife and Range Sciences, University of Idaho. Moscow, Idaho. Boehler, C. T., Miner, J. G., Farver, J. R., and Fryer, B. J Within-stream release-site fidelity of steelhead from Lake Erie hatchery stocks. J. Great Lakes Res. 38: Boswell K. M., Miller, M. W., and Wilson, C. A A lightweight transducer platform for use in stationary shallow water horizontal-aspect acoustic surveys. Fish. Res. 85: Bowen, C. A. II, Bronte, C. R., Argyle, R. L., Adams, J. V., and Johnson, J. E Vateritic sagitta in wild and stocked lake trout: Applicability to stock origin. Trans. Am. Fish. Soc. 128: Brazil, C. E., and Buck, G. B Sonar enumeration of Pacific salmon escapement into the Nushagak River, Alaska Department of Fish and Game, Fishery Data Series No , Anchorage, Alaska. Brazner, J. C., Campana, S. E., Tanner, K. T., and Schram, S. T Reconstructing habitat use and wetland nursery origin of yellow perch from Lake Superior using otolith elemental analysis. J. Gt. Lakes Res. 30(4): Brown, G.E., and Brown, J. A Kin discrimination in salmonids. Rev. Fish Biol. Fish. 6: Burger, C. V., Finn, J. E., and Holland-Bartels, L Patterns of shoreline spawning by sockeye salmon in a glacially turbid lake: evidence for subpopulation differentiation. Trans. Am. Fish. Soc. 124: 1 15.

70 59 Burnham, K.P., Anderson, D. R., Model selection and multi-model inference: a practical information-theoretic approach, second ed. Springer-Verlag, New York. Burwen, D. L., Fleischman, S. J., and Miller, J. D Accuracy and precision of salmon length estimates taken from DIDSON sonar images. Trans. Am. Fish. Soc. 139: Campana, S.E Feeding periodicity and the production of daily growth increments in otoliths of steelhead trout (Salmo gairdneri) and starry flounder (Platichthys stellatus). Can. J. Zool. 61: Campana, S.E Chemistry and composition of fish otoliths: pathways, mechanisms and applications. Mar. Ecol. Prog. Ser. 188: Campana, S. E., and Thorrold, S. R Otoliths, increments, and elements: keys to a comprehensive understanding of fish populations? Can. J. Fish. Aquat. Sci. 58: Carlsen, K.T., Beg, O. K., Finstad, B., and Heggberget, T. G Diel periodicity and environmental influence on the smolt migration of Arctic char, Salvelinus alpinus, Atlantic salmon, Salmo salar, and sea trout, Salmo trutta, in northern Norway. Environ. Biol. Fishes 70: Casselman, J.M Growth and relative size of calcified structures in fish. Trans. Am. Fish. Soc. 119: Chapman, E. D., Hearn, A. R., Michel, C. J., Ammann, A. J., Lindley, S. T., Thomas, M. J., Sandstrom, P. T., Singer, G. P., Peterson, M. L., MacFarlane, R. B., and Klimley, A. P., Diel movements of out-migrating Chinook salmon (Oncorhynchus tsawytscha) and steelhead trout (Oncorhynchus mykiss) smolts in Sacramento/San Joaquin watershed. Environ. Biol. Fish 96:

71 60 Chittaro, P. M., Hogan, J. D., Gagnon, J., Fryer, B. J., and Sale, P. F In situ experiment of ontogenetic variability in the otolith chemistry of Stegastes partitus. Mar Biol. 149(5): Chilton, D. E., Beamish, R. J Age determination methods for the fishes studied by the groundfish program at the Pacific Biological Station. Can. Spec. Publ. Fish. Aquat. Sci. 60. Chrisp, E. Y., and Bjornn, T. C Parr-smolt transformation and seaward migration of wild and hatchery steelhead trout in Idaho. University of Idaho, Idaho Cooperative Fisheries Research Unit, Federal Project F-49-12, Final Report, Moscow. Connor, W.P., Steinhort, R. K., and Burge, H. L Migrational behavior and seaward movement of wild subyearling fall chinook salmon in the Snake River. N. Am. J. Fish Manage. 23: Courtenay, S.C., Quinn, T. P., Dupuis, H. M. C., Groot, C., and Larkin, P. A Discrimination of family-specific odours by juvenile coho salmon: roles of learning and odour concentration. J. Fish Biol. 58: Cowx, I.G., editor Stocking and Introduction of Fish. Fishing News Books, Blackwell Science, Oxford. Cronkite, G. M. W., Enzenhofer, H. J., Ridley, T., Holmes, J., Lilja, J., and Benner, K Use of high-frequency imaging sonar to estimate adult sockeye salmon escapement in the Horsefly River, British Columbia. Canadian Technical Report of Fisheries and Aquatic Sciences 2647, Nanaimo, British Columbia.

72 61 Curtis, J.M., Stunz, G. W., Overath, R. D., and Vega, R. R Otolith chemistry can discriminate signatures of hatchery-reared and wild spotted seatrout. Fish. Res. 153: CWTG (Coldwater Task Group) Report of the Lake Erie Coldwater Task Group, March Presented to the Standing Technical Committee, Lake Erie Committee of the Great Lakes Fishery Commission. Ann Arbor, Michigan, USA. CWTG (Coldwater Task Group) Report of the Lake Erie Coldwater Task Group, March Presented to the Standing Technical Committee, Lake Erie Committee of the Great Lakes Fishery Commission. Ann Arbor, Michigan, USA. CWTG (Coldwater Task Group) Report of the Lake Erie Coldwater Task Group, March Presented to the Standing Technical Committee, Lake Erie Committee of the Great Lakes Fishery Commission. Ann Arbor, Michigan, USA. CWTG (Coldwater Task Group) Report of the Lake Erie Coldwater Task Group, March Presented to the Standing Technical Committee, Lake Erie Committee of the Great Lakes Fishery Commission. Ann Arbor, Michigan, USA. Dale, T The labyrinthine mechanoreceptor organs of the cod Gadus morhua L. (Teleostei: Gadidae). Norw. J. Zool. 24: Daugherty, D.J., Sutton, T. M, and Greil, R. W Life-history characteristics, population structure, and contribution of hatchery and wild steelhead in a Lake Huron tributary. J. Great Lakes Res. 29: Dittman A. H., Quinn, T. P., and Nevitt, G. A Timing of imprinting to natural and artificial odors by coho salmon, Oncorhynchus kisutch. Can. J. Fish and Aquat. Sci. 53:

73 62 Dittman, A.H., May, D., Larsen, D. A., Moser, M. L., Johnston, M., and Fast, D Homing and spawning site selection by supplemented hatchery- and natural-origin Yakima River spring Chinook salmon. Trans. Am. Fish. Soc. 139: Doehring, K., Young. R. G., Hay, J., and Quarterman, A. J Suitability of Dual-frequency Identification Sonar (DIDSON) to identify juvenile fish movement at floodgates. New Zeal. J. Mar. Fresh. Res. 45: Elsdon, T.S., Wells, B. K., Campana, S. E., Gillanders, B. M., Jones, C. M., Limburg, K. E., Secor, D. H., Thorrold, S. R., and Walther, B. D Otolith chemistry to describe movements and life-history parameters of fishes: hypotheses, assumptions, limitations and inferences. Ocean. Mar. Biol. 46: Emlen, J Heterosis and outbreeding depression: a multi-locus model and an application to salmon production. Fish. Res. 12: Felsenstein, J Population differentiation and evolutionary processes. In: Genetic effects of straying of non-native hatchery fish into natural populations: proceedings of the w orkshop (W. S. Grant, ed). U.S. Department of Commerce, National Marine Fisheries Service, NOAA Technical Memorandum NMFS-MWFSC 30, Seattle, Washington. Pp Falini, G., Fermani, S., Vanzo, S., Miletic, M., and Zaffino, G Influence on the formation of aragonite or vaterite by otolith macromolecules. Eur. J. Inorg. Chem. 2005: Fay, R.R The goldfish ear codes the axis of acoustic particle rotation in three dimensions. Science 225:

74 63 Fleming, I. A., Lamberg, A., and Jonsson, B Effects of early experience on the reproductive performance of Atlantic salmon. Behav. Ecol. 8: Francis R. I. C. C The analysis of otolith data. A mathematician s perspective (what, precisely, is your model?). In. Recent developments in fish otolith research (D. H. Secor, J. M. Dean, and S. E. Campana, eds) University of South Carolina Press, Columbia, South Carolina. Pp Gagliano, M., Depczynski, M., Simpson, S. D. and Moore, J. A Dispersal without errors: symmetrical ears tune into the right frequency for survival. Proc. R. Soc. B Biol. Sci. 275: Galbreath, P. F., and Barber, P. E Validation of a long-range dual frequency identification sonar (DIDSON-LR) for fish passage enumeration in the Methow River. Columbia River Inter-Tribal Fish Commission, Technical Report 05-4, Portland, Oregon. Gauldie, R.W Vaterite otoliths from Chinook salmon (Oncorhynchus tshawytscha). New Zeal. J. Mar. Fresh. Res. 20: Gauldie, R.W Effects of temperature and vaterite replacement on the chemistry of metal ions in the otoliths of Oncorhynchus tshawytscha. Can. J. Fish. Aquat. Sci. 53: Gauldie, R.W., Sharma, S. K., and Volk, E Micro-raman spectral study of vaterite and aragonite otoliths of the coho salmon, Oncorhynchus kisutch. Comp. Biochem Physiol. 118A: Goehle, M.A Assessment of natural recruitment in the mixed rainbow/steelhead fishery of Cattaraugus Creek and tributaries. Thesis. State University of New York at Buffalo.

75 64 Gibson-Reinemer, D.K., Johnson, B. M., Martinez, P. J., Winkelman, D. L., Koeing, A. E. and Woodhead, J. D Elemental signatures in otoliths of hatchery rainbow trout (Oncorhynchus mykiss): distinctiveness and utility for detecting origins and movement. Can. J. Fish. Aquat. Sci. 66: Greenstreet, S. P. R Migration of hatchery reared juvenile Atlantic salmon, Salmo salar L., smolts down a release ladder. 1. Environmental effects on migratory activity. J. Fish Biol. 40: Gregory R. S., and Levings, C. D Turbidity reduces predation on migrating juvenile pacific salmon. Trans. Am. Fish. Soc. 127: Groot, C., Quinn, T. P., and Hara, T. J Responses of migrating adult sockeye salmon (Oncorhynchus nerka) to population-specific odours. Can. J. Zool. 64: Hand, C. P., Ludsin, S. A., Fryer, B. J., and Marsden, J. E Statolith microchemistry as a technique for discriminating among Great Lakes sea lamprey (Petromyzon marinus) spawning tributaries. Can. J. Fish. Aquat. Sci. 65: Hanski, I. A., and M. E. Gilpin. editors Meta-population biology: ecology, genetics, and evolution. Academic Press, San Diego, California. Hayden, T. A, Miner, J. G., Farver, J. R., and Fryer, B. J Philopatry and vagrancy of white bass (Morone chrysops) spawning in the Sandusky River: Evidence of metapopulation structure in western Lake Erie using otolith chemistry. J. Great Lakes Res. 37: Hayes, J.W Comparative stream residence of juvenile brown and rainbow trout in a small lake inlet tributary, Scotts Creek, New Zealand. New Zeal. J. Mar. Fresh. Res. 22:

76 65 Hendry, A. P., V. Castric, M. T. Kinnison, and T. P. Quinn The evolution of philopatry and dispersal: homing versus straying in salmonids. In:Evolution illuminated: salmon and their relatives (A. P. Hendry and S. C. Stearns, eds). Oxford University Press, Oxford. Pp Hill, M.S., Zydlewski, G. B., and Gale, W. L Comparisons between hatchery and wild steelhead trout (Oncorhynchus mykiss) smolts: physiology and habitat use. Can. J. Fish. Aquat. Sci. 63: Holmes, J. A., Cronkite, G. M. W., Enzenhofer, H. J., and Mulligan, T. J Accuracy and precision of fish-count data from a dual-frequency identification sonar (DIDSON) imaging system. ICES J. Mar. Sci. 63: Holmes, R., Hayes, J. W., Jiang, W., Quarterman, A., and Davey, L. N Emigration and mortality of juvenile brown trout in a New Zealand headwater tributary. Eco. Fresh. Fish 23: Horrall, R. M Behavioral stock-isolating mechanisms in Great Lakes fishes with special reference to homing and site imprinting. Can. J. Fish. Aquat. Sci. 38: Hvidsten, N.A., Jensen, A. J., Vivas, H., Bakke, Ø., and Heggberget, T. G Downstream migration of Atlantic salmon smolts in relation to water flow, water temperature, moon phase and social interaction. Nord. J. of Fresh. Res. 74: Johansson, G Contribution to the biology of the dab (Limanda limanda L.) in Icelandic waters. Rit Fiskid. 4: Johnson, N. L, Kotz, S., and Balakrishnan, N Continuous Univariate Distributions. Wiley, New York. Jonsson, N Influence of water flow, water temperature, and light on fish migration in

77 66 rivers. Nord. J. Fresh. Res. 66: Jonsson, B., and Jonsson, N Ecology of Atlantic salmon and brown trout. Springer. Dordrecht. Keefer, M. L., and C. C. Caudill Homing and straying by anadromous salmonids: a review of mechanisms and rates. Rev. Fish. Biol. Fish 24: Keefer, M. L., Caudill, C. C., Peery, C. A., and Boggs, C. T Non-direct homing behaviours by adults Chinook salmon in large, multi-stock river system. J. Fish Biol. 72: Keeley, E.R Demographic responses to food and space competition by juvenile steelhead trout. Ecology 82: Kelch, D., Lichtkoppler, F., Sohngen, B., and Daigneault, A The value of steelhead (Onchorhynchus mykiss) angling in Lake Erie Tributaries. J. Great Lakes Res. 32: Kucera, P. A., and Faurot, D Chinook salmon (Oncorhynchus tshawytscha) adult abundance monitoring in Lake Creek and Secesh River, Idaho, Annual Report to the Bonneville Power Administration, Project , Portland, Oregon. Kucera, P. A., and Orme, R. W Chinook salmon (Oncorhynchus tshawytscha) adult abundance monitoring in Lake Creek and Secesh River, Idaho, Annual Report to the Bonneville Power Administration, Project , Portland, Oregon. Labelle, M Straying patterns of Coho salmon (Oncorhynchus kisutch) stocks from southeast Vancouver Island, British Columbia. Can. J. Fish. Aquat. Sci. 49: Landsman, S.J., Nguyen, V. M., Gutowsky, L. F. G., Gobin, J., Cook, K. V., Binder, T. R., Lower, N., McLaughlin, R. L., and Cooke, S. J Fish movement and migration studies

78 67 in the Laurentian Great Lakes: Research trends and knowledge gaps. J. Great Lakes Res. 37: Leduc, A. O. C. H., Kim, J. W., Macnoughton, C. J., and Brown, G. E Sensory complement model helps to predict diel alarm response patterns in juvenile Atlantic salmon (Salmo salar) under natural conditions. Can. J. Zool. 88: Leider, S.A., Increased straying by adult steelhead trout, Salmo gairdneri, following the eruption of Mount St. Helens. Envrion. Biol. Fish. 24: Lirette, M. G., and Hooton, R. S Coded wire tag recoveries from Vancouver Island sport caught steelhead, British Columbia Ministry of Environment, Fisheries Branch, Fisheries Management Report 92, Vancouver. Lockwood, Roger N. and Schneider, J. C Stream fish population estimates by mark and recapture and depletion methods (Chapter 7). In. Manual of fisheries survey methods II: with periodic updates. (J. C. Schneider, editor). Michigan Department of Natural Resources, Fisheries Special Report 25, Ann Arbor. Longerich, H.P., Jackson, S. E., and Gunther, D Laser ablation inductively coupled plasma mass spectrometric transient signal data acquisition and analyte concentration calculation. J. Anal. Spectrom. 11: Ludsin, S.A., Fryer, B.J., Gagnon, J.E Comparison of solution-based versus laser ablation inductively coupled plasma mass spectrometry for analysis of larval fish otolith micro-elemental composition. Trans. Amer. Fish. Soc. 135: Macdonald, J. I., Shelley, J. M. G., and Crook, D. A A method for improving the estimation of natal chemical signatures in otoliths. Trans. Am. Fish. Soc. 137(6):

79 68 MacCrimmon, H.R., and Gots, B. L Rainbow Trout in the Great Lakes. Ontario Ministry of Natural Resources, Toronto, Ontario. Marklevitz, S. A. C., Fryer, B. J., Gonder, D., Yang, Z., Johnson, J., Moerke, A., and Morbey. Y. E Use of otolith chemistry to discriminate juvenile Chinook salmon (Oncorhynchus tshawytscha) from different wild populations and hatcheries in Lake Huron. J. Great Lakes. Res. 37: Martin, J. T Racial identification of chinook salmon (Oncorhynchus tshawytscha) and juvenile steelhead trout (Salmo gairdneri). Thesis. Oregon State University. Maxwell, S. L., and Gove, N. E The feasibility of estimating migrating salmon passage rates in turbid rivers using a dual frequency identification sonar (DIDSON), Alaska Department of Fish and Game, Regional Information Report 2A04-05, Anchorage, Alaska. McGarigal, K., Cushman, S., and Stafford, S Multivariate Statistics for Wildlife and Ecology Research. Springer, New York. McKern, J. L Steelhead trout otoliths for age, race, and stock analysis. Thesis. Oregon State University. McKern, J. L., Horton, H. F., and Koski, K. V Development of steelhead (Salmo gairdneri) otoliths and their use for age analysis and for separating summer from winter races and wild from hatchery stocks. J. Fish. Res. Board Can. 31: Melancon, S., Fryer, B. J., Ludsin, S. A., Gagnon, J. E., and Yang, Z Effects of crystal structure on the uptake of metals by lake trout (Salvelinus namaycush) otoliths. Can. J. Fish. Aquat. Sci. 62:

80 69 Melancon, S., Fryer, B. J., Gagnon, J. E., and Ludsion. S. A Mineralogical approaches to the study of biomineralization in fish otoliths. Mineral. Mag. 72(2): Melnychuk, M. C., Welch, D. W., Walters, C. J., and Christensen, V Riverine and early ocean migration and mortality patterns of juvenile steelhead trout (Oncorhynchus mykiss) from the Cheakamus River, British Columbia. Hydrobiol. 582: Melnychuk, M.C., Welch, D. W., and Walters, C. J Spatio-temporal migration patterns of Pacific salmon smolts in rivers and coastal marine waters. PLoS One 5:e Milner, A. M., and Bailey, R. G Salmonid colonization of new streams in Glacier Bay National Park, Alaska. Aqua. Fish. Manag. 20: Moore, M., Berejikian, B. A., and Tezak, E. P Variation in the early marine survival and behavior of natural and hatchery-reared Hood Canal steelhead. PLoS One 7:e Morat, F., Betoulle, S., Robert, M., Thailly, A. F., Biaginati-Rosbourg, S., and Lecomte-Finiger. R What can otolith examination tell us about the level of perturbations of Salmonid fish from the Kerguelen Islands? Ecol. Fresh. Fish 17: Moyle PB, and Cech Jr JJ Fishes, an Introduction to Ichthyology (3rd edition). Prentic Hall.Upper Saddle River, New Jersey. Murray, C. and M. Shields Creel analysis and economic impact of Pennsylvania s Lake Erie tributary fisheries in Erie County, Pennsylvania, with special emphasis on landlocked steelhead trout (Oncorhynchus mykiss). Pennsylvania Fish and Boat Commission, Project F-71-R-14, Fairview, Pennsylvania. Neter, J., Wasserman, W., and Kutner, M. H Applied linear statistical models (2nd edition). Irwin, Homewood, Illinois. Nevitt G. A., Dittman, A. H., Quinn, T. P., and Moody, W. J Evidence for peripheral

81 70 olfactory memory in imprinted salmon. PNAS 91: Oxman, D. S., Barnett-Johnson, R., Smith, M. E., Coffin, A., Miller, D. L., Josephson, R. and Popper, A. N The effect of vaterite deposition on sound reception, otolith morphology, and inner ear sensory epithelia in hatchery-reared Chinook salmon (Oncorhynchus tshawythscha). Can. J. Fish. Aquat. Sci. 64: Pangle, K. L., Ludsin, S. A., and Fryer, B. J Otolith microchemistry as a stock identification tool for freshwater fishes: testing its limits in Lake Erie. Can. J. Fish. Aquat. Sci. 67: Pannella, G Fish otoliths: daily growth layers and periodical patterns. Science 173: Pannella, G Otolith growth patterns: an aid in age determination in temperate and tropical fishes. In. Ageing of fish (T. B. Bagenal, ed) Unwin Brothers, London, England. Pp Payan, P., De Pontual, H., Edeyer, A., Borelli, G., Boeuf, G., and Mayer-Gostan, N Effects of stress on plasma homeostasis, endolymph chemistry, and check formation during otolith growth in rainbow trout (Oncorhynchus mykiss). Can. J. Fish. Aquat. Sci. 61: Peck, T. H Differentiation of hatchery and stream juvenile coho salmon (Oncorhynchus kisutch) from Washington and Oregon by the use of scales and otoliths. Thesis. University of Washington. Peven, C.M., Whitney, R. R., and Williams, K. R Age and length of steelhead smolts from the Mid-Columbia River Basin, Washington. N. Am. J. Fish. Manage.14: Pipal, K., Jessop, M., Boughton, D., and Adams, P Using dual-frequency identification

82 71 sonar (DIDSON) to estimate adult steelhead escapement in the San Lorenzo River, California. Calif. Fish Game 96: Pipal, K. A., Notch, J. J., Hayes, S. A., and Adams, P. B Estimating escapement for a lowabundance steelhead population using dual-frequency identification sonar (DIDSON). N. Am. J. Fish. Manage. 32: Poe, T. P., Hansel, H. C., Vigg, S., Palmer, D. E., and Prendergast, L. A Feeding of predaceous fishes on out-migrating juvenile salmonids in John Day Reservoir, Columbia River. Trans. Am. Fish. Soc. 120: Popper, A. N. and Lu, Z Structure-function relationships in fish otolith organs. Fish. Res. 46: Quinn, T.P A review of homing and straying of wild and hatchery-produced salmon. Fish. Res. 18: Quinn, T. P., The behavior and ecology of Pacific salmon and trout. University of Washington Press, Seattle, Washington. Quinn, T. P., and Tolson, G. M Evidence of chemically mediated population recognition in coho salmon (Oncorhynchus kisutch). Can. J. Zool. 64: Quinn, T.P., Brannon, E. L., and Dittman, A. H Spatial aspects of imprinting and homing in Coho salmon, Oncorhyncus kisutch. Fish. Bull. 87: Raymond, B.L Migration rates of yearling chinook salmon in relation to flow and impoundments in the Columbia and Snake Rivers. Trans. Am. Fish. Soc. 97: Reichert, J. M., Fryer, B. J., Pangle, K. L., Johnson, T. B., Tyson, J. T., Drelich, A. B., and Ludsin, S. A River-plume use during the pelagic larval stage benefits recruitment of a lentic fish. Can. J. Fish. Aquat. Sci. 67:

83 72 Reimer, T., Dempster, T., Warren-Myers, F., Jensen, A. J., and Swearer, S. E High prevalence of vaterite in sagittal otoliths causes hearing impairment in farmed fish. Sci. Rep. 6: Reisenbichler, R. R., McIntyre, J. D., Solazzi, M. F., and Landino, S. W Genetic variation in steelhead of Oregon and northern California. Trans. Am. Fish. Soc. 121: Rempel, M.H., Land, R. W., Mitchell, L. W., and Ward, B. R Wild and hatchery steelhead and cutthroat trout populations of the Little Campbell River, B.C., Province of British Columbia Fisheries Technical Circular 62, Victoria, British Columbia. Rieman, B. E., Beamesderfer, R. C., Vigg, S., and Poe, T. E Estimated loss of juvenile salmonids to predation by northern squaw-fish, walleyes, and smallmouth bass in John Day Reservoir, Columbia River. Trans. Am. Fish. Soc. 120: Romer, J. D., C. A. Leblanc, Clements, S., Ferguson, J. A., Kent, M. L., Noakes, D., and Schreck, C. B., Survival and behavior of juvenile steelhead trout (Oncorhynchus mykiss) in two estuaries in Oregon, USA. Environ. Biol. Fish 96: Rybock. J. T., Horton, H. F., and Fessler, J. L Use of otoliths to separate juvenile steelhead trout from juvenile rainbow trout. Fish. Bull. 73: Schirripa, M. J An evaluation of back-calculation methodology using simulated otolith data. Fish. Bull. 100: Schroeder, R. K., Lindsay, R. B., and Kenaston. K. R Origin and straying of hatchery winter steelhead in Oregon coastal rivers. Trans. Am. Fish. Soc. 130: Secor, D. H., Dean, J. M., and Laban, E. H Manual for otolith removal and preparation for microstructural examination. Baruch Institute Technical Report University of South Carolina, Columbia, South Carolina.

84 73 Seelbach, P. W Smolting success of hatchery-raised steelhead planted in a Michigan tributary of Lake Michigan. N. Am. J. Fish. Manage. 7: Seelbach, P. W Population biology of steelhead in a stable-flow, low-gradient tributary of Lake Michigan. Trans. Am. Fish. Soc. 122: Seelbach, P.W., Dexter, J. L., and Ledet, N. D Performance of steelhead smolts stocked in southern Michigan warmwater rivers. Fisheries Research Report Number Michigan Department of Natural Resources, Ann Arbor, Michigan. Slaney, P.A., and Harrower, W. L Experimental culture and release of steelhead trout reared in netpens at O Connor Lake in British Columbia. In: Proceedings: propagation, enhancement, and rehabilitation of anadromous salmonid populations and habitat symposium. (T. J. Hassler, ed). Humboldt State University, Arcata, California. Pp Slaney, P. A., Berg, L, and Tautz, A. F Returns of hatchery steelhead to site of release below an upper-river hatchery. N. Am. J. Fish. Manage. 13: Smith, S.G., Muir, W. D., Williams, J. G. and Skalski, J. R Factors associated with travel time and survival of migrant yearling chinook salmon and steelhead in the Lower Snake River. N. Am. J. Fish. Manage. 22: Stabell, O.B Intraspecific pheromone discrimination and substrate marking by Atlantic salmon parr. J. Chem. Ecol. 13: Stabell, O.B Olfactory control of homing behaviour in salmonids. In: Fish chemoreception (T. J. Hara, ed) Chapman and Hall, London. Pp Strong, M.B., Neilson, J. D., and Hunt, J. J Aberrant crystal-lization of pollock (Pollachius virens) otoliths. Can. J. Fish. Aquat. Sci. 43:

85 74 Sweeting, R. M., Beamish, R. J., Noakes, D. J., and Neville, C. M Replacement of Wild Coho Salmon by Hatchery-Reared Coho Salmon in the Strait of Georgia over the past Three Decades. N. Am. J. Fish. Manage. 23: Thompson, B. E., and Ferreri. C. P Population biology of steelhead spawning runs in three Pennsylvania tributaries to Lake Erie. J. Great Lakes Res. 28: Thorrold, S.R., Jones, C. M., Campana, S. E., McLaren, J. W., Lam, J. W. H Trace element signatures in otoliths record natal river of juvenile American shad (Alosa sapidissima). Limno. Oceanogr. 43: Tzeng, W.N Effects of salinity and ontogenetic movements on strontium:calcium ratios in the otoliths of the Japanese eel, Anguilla japonica Temminck and Schlegel. J. Exp. Mar. Biol. Ecol. 199: Ueda, H Physiological mechanisms of imprinting and homing migration in Pacific salmon Onchorhynchus spp. J. Fish Biol. 81: USDOI (United States Department of the Interior), United States Fish and Wildlife Service, United States Department of Commerce, and United States Census Bureau National survey of fishing, hunting, and wildlife-associated recreation. Washington, D. C. Vargason, C. Pennsylvania Fish and Boat Commission Lake Erie Steelhead: A Success Story. Pennsylvania Angler and Boater 82: Wagner, H. H Effect of stocking time on survival of Steelhead Trout, Salmo gairdnerii, in Oregon. Trans. Am. Fish. Soc. 97: Wagner, H. H Photoperiod and temperature regulation of smolting in steelhead trout (Salmo gairdneri). Can. J. Zool. 52: Wagner, H.H., Wallace, R. L., and Campbell, H. J The seaward migration and return of

86 75 hatchery-reared steelhead trout, Salmo gairdneri, in the Alsea River, Oregon. Trans. Am. Fish Soc. 92: Walther, B.D., Thorrold, S. R., and Olney, J. E Geochemical signatures in otoliths record natal origins of American shad. Trans. Am. Fish. Soc. 137: Waples, R. S Genetic interactions between hatchery and wild salmonids: lessons from the Pacific Northwest. Can. J. Fish. Aquat. Sci. 48(Supplement 1): Ward, B.R., and Slaney, P. A Returns of penreared steelhead from riverine, estuarine, and marine releases. Trans. Am. Fish. Soc. 119: Ward, B. R., Slaney, P. A., Facchin, A. R. and Land, R. W Size-biased survival in steelhead trout (Oncorhynchus mykiss): back-calculated lengths from adult s scales compared to migrating smolts at the Keogh River, British Columbia. Can. J. Fish. Aquat. Sci. 46: Watson, J. E Determining the age of young herring from their otoliths. Trans. Am. Fish. Soc. 93: Wedemeyer, G. A., Saunders, R. L., and Clarke, W. C., Environmental factors affecting smoltification and early marine survival of anadromous salmonids. Mar. Fish. Rev. 42: Welch, D.W., Rechisky, E. L., Melnychuk, M. C., Porter, A. D., Walters, C. J., Clements, S., Clemens, B. J., McKinley, R. S., and Schreck, C Survival of migrating salmon smolts in large rivers with and without dams. PLoS Biol. 6(10): Welch, D.W., Ward, B. R., and Batten, S. D Early ocean survival and marine movements of hatchery and wild steelhead trout (O. mykiss) determined by an acoustic array: Queen Charlotte Strait, British Columbia. Deep-Sea Res. 51:

87 76 Wells, B.K., Rieman, B. E., Clayton, J. L., Horan, D. L., and Jones, C. M Relationships between water, otoliths, and scale chemistries of Westslope Cutthroat trout from the Coeur d'alene River, Idaho: the potential application of hard-part chemistry to describe movements in freshwater. Trans. Am. Fish. Soc. 132: Westley, P. A. H., Quinn, T. P., Dittman, A. H., Rates of straying by hatchery-produced Pacific salmon (Oncorhynchus spp.) and steelhead (Oncorhynchus mykiss) differ among species, life history types, and populations. Can. J. Fish. Aqua. Sci. 70(5): Wolff, B. A., Johnson, B. M., Breton, A. R., Martinez, P. J. and Winkelman, D. L Origins of invasive piscivores determined from the strontium isotope ration ( 87 Sr/ 86 Sr) of otoliths. Can. J. Fish. Aquat. Sci. 69: Yamamoto, Y., Hino, H., and Ueda, H Olfactory imprinting of amino acids in lacustrine sockeye salmon. PLoS ONE 5(1): e8633. Yang, Z LA-ICPMS Data Reduction Program. Great Lakes Institute for Environmental Research, University of Windsor, Windsor, Ontario. Yokota, T., Mitamura, H., Arai, N., Masuda, R., Mitsunaga, Y., Itani, M., Takeuchi, H., and Tsuzaki, T Comparison of behavioral characteristics of hatchery- reared and wild red tilefish Branchiostegus japonicus released in Maizuru Bay by using acoustic biotelemetry. Fish. Sci. 72: Zimmerman, C.E., and Nielson, R. L Effect of analytical conditions in wavelength dispersive electron microprobe analysis on the measurement of strontium-to-calcium (Sr/Ca) ratios in otoliths of anadromous salmonids. Fish. Bull. 101:

88 Fig.1 Adult Steelhead Collection Sites: Location of Lake Erie tributaries where adult steelhead were collected in 2009/2010 and

89 78 TL (mm) OR (mm) Fig.2 Regression of Steelhead Total Length vs Otolith Radius: Regression analysis of TL (mm) and OR (mm) of steelhead. Juvenile lengths (n = 52) were obtained from hatchery fish from each Lake Erie hatchery before stocking and adult lengths (n = 54) were obtained from adult steelhead collected in Cattaraugus and Chautauqua Creeks, NY (n = 52). A quartic function was used and resulted in a regression equation of y = x x x x and a r² of Because otoliths are formed early in the development of steelhead embryos the regression was forced through the origin.

90 Fig.3 Map of Trout Run, PA: Black circles ( ) indicate DIDSON location, location of upstream barrier, and point of stocking by PA Fish and Boat Commission. Shaded gray areas highlight tributary sections used for population density estimate. 79

91 Fig.4 DIDSON Configuration at Trout Run, PA: DIDSON positioning in Trout Run for most sampling in 2014 and Early exceptions are explained in text. 80

92 Fig.5 Steelhead Emigration Out of Trout Run in 2014 and 2015: Abundance and timing of steelhead emigrating from Trout Run in 2014 and White sections of histogram bars represent fish that emigrated during daylight hours and black sections of bars represent fish that emigrated during night hours. Dates when DIDSON data was not collected for a full 24 hours have gray bars to represent extrapolation of the number of steelhead that emigrated during hours not sampled. Areas shaded with gray-dot pattern represent dates when no DIDSON footage was recorded. Dates when steelhead were stocked are denoted by an S and dates when electrofishing occurred are denoted with an EF. 81

93 Fig.6 Lengths of Juvenile Steelhead in Trout Run: Box plots of the lengths of (a) emigrating juvenile steelhead trout over time and (b) fish measured prior to stocking (hatchery), and emigrated versus non-emigrated fish. In (b) the * represents fish that had their lengths estimated using DIDSON measuring tool. The end of each box defines the 25 th and 75 th percentiles, the line within the box is the median, and the error bars (whiskers) denote the 10 th and 90 th percentiles, with the outlier represented by ( ). Those with the same upper case letters were not significantly different (p > 0.05). 82

94 Fig.7 Lorentzian Peak Model for Steelhead Emigration: Lorentzian peak regression of diel emigration rate (# emigrants/day) with photoperiod as the independent variable. ( ) represents data points from 2014 and ( ) represents data points from

95 Fig.8 Plot of Elemental Concentrations in Vaterite Otolith: Otolith elemental signature from edge (100% region) to core (0% region) of an individual from the Salmon River Hatchery. Region labels are as follows: A = aragonite, V = vaterite, T = transition. Otolith sections were classified as aragonite when Sr and Ba were high and Mg was low (0-25% regions), vaterite when Sr and Ba were low and Mg was high (35-100% regions), and transition during periods when shifts in elemental concentrations were in progress (25-35% regions). 84