UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of hydraulic characteristics on energy use and behaviour of adult upriver migrating sockeye (Oncorhynchus… Standen, Emily M. 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-0522.pdf [ 10.81MB ]
Metadata
JSON: 831-1.0090064.json
JSON-LD: 831-1.0090064-ld.json
RDF/XML (Pretty): 831-1.0090064-rdf.xml
RDF/JSON: 831-1.0090064-rdf.json
Turtle: 831-1.0090064-turtle.txt
N-Triples: 831-1.0090064-rdf-ntriples.txt
Original Record: 831-1.0090064-source.json
Full Text
831-1.0090064-fulltext.txt
Citation
831-1.0090064.ris

Full Text

EFFECTS OF HYDRAULIC CHARACTERISTICS ON ENERGY USE AND BEHAVIOUR OF ADULT UPRIVER MIGRATING SOCKEYE (ONCORHYNCHUS NERKA) AND PINK (O. GORBUSCHA) SALMON. by EMILY M . STANDEN B.Sc. Hon., The University of Kings College, Dalhousie University, 1995 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Faculty of Forestry (Department of Forest Sciences) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 2001 © Emily M . Standen, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Date Q(t 13. QtWi DE-6 (2/88) Abstract Adult Pacific salmon (Oncorhynchus sp.) undertake upriver migrations to reach their spawning grounds. Because these fish cease feeding upon entry to freshwater they depend entirely upon energy reserves to complete their upriver spawning migration. In the past in situ salmon migration energetics and behaviour have been studied independently. Body constituency analysis has been used to assess energy use of migrating fish on a spatial-temporal scale of hundreds of kilometers (Idler and Clemens 1959; Gilhousen 1980; Leonard and McCormick 1999) and radio telemetry as well as visual observations have been used to study animal behaviour over distances of tens to thousands of metres (Ellis 1966a; Fretwell 1981). The recent advent of radio telemetry and underwater stereo videography has allowed coupling of behaviour with energy use of upriver migrating salmon in situ (Boisclair 1992; McKinley and Power 1992; Hughes and Kelly 1996; Hinch and Rand 2000). Very few studies to date have used EMG telemetry to estimate energy use of moving fish in situ and only one of these was done on salmon (sockeye salmon Hinch and Rand 1998; sturgeon McKinley and Power 1992). A limited number of studies have used underwater videography to study in situ juvenile fish movement (Boisclair 1992; Hughes and Kelly 1996) however, to date only one study has used underwater videography to analyze adult fish migration again on sockeye salmon (Hinch and Rand 2000). The objective of this thesis was to assess the affect of sex, species and river features on in situ energy use and behaviour of upriver migrating pacific salmon (Oncorhynchus nerka and O. gorbuscha). The first part of the project took place in the fall of 1999 and used EMG telemetry to describe activity levels and estimate energy use in 12 adult pink salmon during migration through a 7 km section of the Fraser Canyon which had a diversity of flow conditions/This data was then compared with similar data collected for sockeye salmon and provided the first in situ comparison of its kind. Across all reaches sockeye salmon tended to use more energy (0.93 J-m" ') than pink salmon (0.26 J-m"1; P=0.0986). Sockeye salmon swimming speeds were twice as variable (mean CV; 54.78) as pink salmon swimming speeds (mean CV; 22.54). In all analysis reach was a significant factor determining swimming speed, ground speed and energy use (PO.0009 for all ANOVAs). Within sex species groups all fish increased their activity levels when they migrated through constricted reaches compared with non-constricted reaches. Between sex-species groups differences in behaviour depended upon reach. The second part of the project was conducted the summer of 2000 and used underwater stereovideography to assess differences in energy use and behaviour of adult sockeye salmon as they migrated through a variety of small scale flow fields (1-5 m2) within the Seton River, British Columbia. The objective of this study was to compliment the work conducted in Part I by more accurately defining river conditions as well as the energy use and behaviour of fish migrating through these conditions. On average fish encountered velocities (44 cm-s"1) which were less than the average site velocity (53 cm-s"1). In addition swimming speed and ground speed were similar in low encountered velocity sites (<0.13 m-s"1) with a ratio of 1:1. In mid to high encountered velocity sites (0.35-1.27 m-s"1) the ratio of swimming speed to ground speed stayed constant at 2:1 but both swimming speed and ground speed increased with increasing velocity. These projects suggest that although species and sex are relevant, river reach characteristics are the primary factors affecting energy use during upstream salmon migration in the Fraser River. Complex current patterns generated by river constrictions created confusing migration cues, which impeded the ability of salmon to locate appropriate pathways. When assessed with videography, at the majority of sites fish used low velocity flow fields to optimize migration. This thesis has been successful in confirming the importance of river reach characteristics in determining energy use and behaviour in upriver migrating salmon. These results suggest that river management, whose goal is to ensure successful fish passage, should focus on bank and bed characteristics that provide flow patterns and hydraulics that minimize energy use in migrating fish. iii Table of Contents Abstract ii Table of Contents iv List of Figures vi List of Appendices viii Acknowledgments . ix Chapter 1: Introduction 1 Rationale for Research on Salmon Migration 2 Salmon Migration and Life History 5 Chapter 2: Energetic costs of migration through the Fraser River canyon, British Columbia, in adult pink {Oncorhynchus gorbuscha) and sockeye (O. nerka) salmon as assessed by E M G telemetry 9 Abstract 9 Introduction 10 Methods 11 Swimming tunnel study '. 11 Telemetry study .13 Statistical analyses 15 Results 17 Swimming tunnel study.... 17 Telemetry study 17 Discussion 24 Chapter 3: Swimming patterns and behaviour of upriver migrating adult pink (Oncorhynchus gorbuscha) and sockeye (O. nerka) salmon as assessed by E M G telemetry in the Fraser River, British Columbia, Canada 29 Abstract : 29 Introduction 30 Description of study sites 31 Materials and methods 32 Study animals and telemetry approaches 32 Behavioural observations 33 Statistical analyses 34 Results 35 Statistical analyses 35 Qualitative description of migration path behaviours and swimming strategies 40 Discussion 41 Chapter 4: Examining the effects of hydraulic characteristics on energy use in upriver migrating salmon using underwater stereo videography 49 Abstract 49 Introduction 50 Methods ...53 Site description and selection 53 iv Video System. 53 Camera calibration and data collection 55 Data analysis 58 Results -58 Statistical Analysis 58 Discussion 62 Chapter 5: Conclusions and Future Research Direction 66 Literature Cited 69 APPENDIX A. The migratory paths of study fish through the Fraser River Canyon 78 APPENDIX B. The instantaneous swimming speeds of study fish through reaches 2, 4, 7 and 9 in the Fraser River Canyon 100 v List of Figures Figure 1. Map of Fraser River canyon indicating 9 reaches (numbered) that were studied by Hinch and Rand (1998). Inset is a map of British Columbia, Canada showing the Fraser River and its main tributaries, the locations of spawning destinations for the two stocks of salmon in this study, and the location of the study area 12 Figure. 2 Electromyogram Pulse Interval (ms) vs. Tailbeat Frequency (beats/min) for male (n=4) and female (n=4) pink salmon developed using a Brett-type respirometer. Fish tested in the tunnel showed similar range and spread of data points 18 Figure 3. Mean energy use (plus one SE) in each reach for each sex-species class of fish. The first row of digits on the x-axis represent the reach number, the second row describes the bank morphology of the reach, the third describes the main current direction as well as the surface velocity (m/s) and the fourth row is the coefficient of variation of energy use for each reach calculated using sex-species averages. Numbers above each bar represent sample size 21 Figure 4. Mean swimming speed (plus/minus the standard error) for pink salmon females (white bars), pink salmon males (dark grey bars), sockeye salmon females (light grey bars), and sockeye salmon males (black bars). Numbers on bars indicate sample size 36 Figure 5. Mean coefficient of variation in swimming speed (plus/minus the standard error) for pink salmon females (white bars), pink salmon males (dark grey bars), sockeye salmon females (light grey bars), and sockeye salmon males (black bars). Numbers on bars indicate sample size 38 Figure 6. Proportion of time spent swimming with sustained (white), prolonged (grey) and burst (black) speeds at each reach. These values were calculated by averaging proportions of total time spent swimming within each of the three speed categories for individual fish, within each sex-species class, and then averaging these values within reaches.. 39 Figure 7. a. Map of the study area showing the upstream migratory path of a male sockeye salmon (thin line). Times of entry, exit and intermediary points are indicated for each of the four study reaches 42 Figure 7. b. Time-series plots of instantaneous swimming speeds for the sockeye male shown in Figure 7.a. from time of reach entry to exit for each of the four reaches. The solid lines represent Ucrit .43 Figure 8. a. Map of the study area showing the upstream migratory path of a female pink salmon (thin line). See Figure 7.a. for details 44 vi Figure 8. b. Time-series plots of instantaneous swimming speeds for the female pink shown in Figure 8.a. from time of reach entry to exit for each of the four reaches. See Figure 7.b. for details 45 Figure 9. Map of the study area indicating the video sites. Inset map of British Columbia, Canada indicating location of the Fraser River and the Seton River. Sites SDI 2 and SD34 are located off of the map at the downstream end of Seton Lake at the start of the lower Seton River 54 Figure 10. Interpolated site velocity gradients. Each cube represents an individual site with site names listed below. Sites are varying sizes based on the topography of each site and field of view of each camera. Each square within each cube represents ten centimeters, also denoted by black bar in lower right corner of each site cube. Velocities (cm»s"') are noted in the legend to the right, negative values indicating flows moving in the upstream direction 57 Figure 11. Encountered velocity (white bars) and average site velocity (blackbars) (plus/minus the standard error) for each site. Sites are located in increasing average site velocity along the x-axis. The first row of digits on the x-axis represents site name, the second row represents site depth and the third row represent Froude number calculated for each site. The large letters at the top of the page are the results of the Tukey's Studentized Range Test where similar letters represent similar site velocities. Sample size is noted by small numbers above each bar 60 Figure 12. Swimming speed (white bars) and ground speed (grey bars) (plus/minus standard error) for each site. Sites are located in increasing encountered velocity along the x-axis. The thin solid line represents the swimming efficiency index (SEI) and the dotted line represents the optimal migration index (OMI) for each site. The thick solid line depicts an index value of 1.00. Sample sizes are listed along top of graph 61 vii List of Appendices APPENDIX A. The migratory paths of study fish through the Fraser River Canyon 78 APPENDIX B. The instantaneous swimming speeds of study fish through reaches 2, 4, 7 and 9 in the Fraser River Canyon 100 Acknowledgments I thank Scott Hinch for always having an open door and endless hours for discussion of ideas. I also thank my committee members Kathy Martin and Michael Healey for supplying alternative viewpoints and fresh perspectives. I am also grateful to Tony Farrell (Simon Fraser University), Pete Rand (North Carolina State University), Nick Hughes (University of Alaska Fairbanks) and Lon Kelly (United States Geological Survey) for their input and prompt response to a myriad of technical questions I have thrown at them. Invaluable field support was provided by Andrew Lotto, Shannon Machlaughlan and Yuho Okada. Thanks for your patients, competence and supportive laughter through any number of crazy ideas and adventures I suggested. In addition, thanks to my family who got me this far and despite everything are still right there .beside me. And to my interdisciplinary friends who over countless dinner conversations shared their wealth of knowledge, without your napkin scribbles the accuracy and expediency of these projects would never have been possible. Finally, thanks to the 57 pound lake trout mounted on the wall of the Wanapitei Chateau, Temagami, who, 24 years ago terrified me into a life long fascination with fish. DEO, LEGI, REGI, PESCI ix Chapter 1: Introduction Efficient use of energy is important to all species, particularly during energetically expensive behaviours such as migration. Migration is defined as the regular seasonal movement of bird and animal populations to and from different areas, often considerable distances apart (Dunster and Dunster 1996). Unfortunately, the relationship between energy use and animal behaviour is difficult to measure. Photography and video observations, data logging tags and telemetry allow us to make detailed records of animal movement and behaviour. However without an equally detailed description of the physical environment as well as a biomechanical and bioenergetic understanding of the study organism, animal behaviour cannot be linked to energy consumption. Biomechanics describes the mechanical processes of an organism whilst bioenergetics is concerned with energy use and can range from the study of energy used in the transport of ions across a cell membrane, to the transfer of energy between tropic levels in an ecosystem. This thesis studies migration bioenergetics by primarily focusing on locomotory, and to a small extent, reproductive and metabolic energy use. The physical environment influences the biomechanics of animal behaviour and locomotion. As a result the accuracy of knowledge about forces experienced in the environment will determine the accuracy of energy use calculations when dealing with animal movement. A detailed understanding of the physical forces exerted in an environment is essential to measuring energy use with respect to animal behaviour. Analysis of terrestrial systems often need only be concerned with animal movement and behaviour in relation to a constant gravitational force. The three-dimensional aspect of avian and aquatic behaviour, however, requires a more complex consideration of environmental forces. Much analysis of flight mechanics and aerodynamics has been done on bird species (Tucker and Parrott 1970; Spedding 1986; Spedding 1987; Tucker 1990). Birds such as the albatross reduce energy consumption in flight by using atmospheric movement (i.e. wind and thermals) for soaring (Pennycuick 1972; Costa and Prince 1987; Norberg 1996). The similarity between the physics of water and air suggests that fish utilize similar principles as birds to limit energy use during their migrations (Vogel 1994). Fish movement has been studied for several species, each species having different body form and swimming style to perform various functions and fill competitive niches in different environments (Videler 1993; Vogel 1994). 1 Most of these bird and fish studies have been done in laboratory wind tunnels and swimming tubes (Brett 1962; Brett 1967; Tucker and Parrott 1970; Brett 1973; Weihs 1973b; Brett 1982; Weihs and Webb 1983; Yates 1983; Spedding 1986; Bernatchez and Dodson 1987; Spedding 1987; Williams and Brett 1987; Tucker 1990; Jain et al. 1997; Jain et al. 1998). Laboratory studies allow for precise measurement of environmental conditions as well as the physiological processes of animals within them. These controlled conditions allow for the development of equations that describe the relationship between environment and energy use (reviewed in Brett 1995; Webb 1995; Norberg 1996; Leonard et al. 2000). Unfortunately these laboratory environments are designed to be simple in order to facilitate measurement and, as a result, do not reflect the natural environment. To complement laboratory studies, technologies have been used to record and measure animal behaviour in situ. Most of these technologies however, are limited in their ability to measure.forces within the environment, or precise environmental conditions. Engineers have been able to define and predict dynamic forces within the physical environment with precision and accuracy (Sahin et al. 1994; Beffa 1996; Afzalimehr and Anctil 1999). Again, most of these predictive models have been developed in controlled laboratory situations and it is difficult or impractical to apply them to the complex environmental conditions in situ. However, the common use of averages to describe what an animal is encountering (Hogan and Church 1989; Crowder and Diplas 2000; Hatfield and Bruce 2000) may be equally impractical and misleading (Kondolf et al. 2000) suggesting that a combined engineering and biological approach is necessary. This thesis attempts this combined approach in a study that focuses on the migration behaviour and energy use of two species of Pacific salmon. Rationale for Research on Salmon Migration Pacific salmon (Oncorhynchus spp) provide an important commercial industry on the Pacific Coast as well as an energetic link, bringing large stores of ocean nutrients to birds, mammals and oligotrophic stream ecosystems inland (Reimchen 2000). Salmon spend a large part of their lives in the rich ocean environment feeding and maturing before returning to their natal streams to spawn and die. The successful reproduction of salmon is directly affected by their ability to migrate upriver. Upon entering freshwater at the start of their upriver migration, salmon stop feeding. As a result, their migration, gamete production and spawning behaviour are completed using finite energy reserves. Overall energy conservation during migration is essential for salmon to reach their spawning grounds and to spawn successfully. 2 Environmental conditions within the river influence the amount of energy a fish uses to travel a certain distance. River conditions that require the fish to expend more energy than it has available can result in pre-spawning mortality and decreased gamete viability (Gilhousen 1980; Brett 1983). Both natural river obstacles and fluctuations as well as man-made structures that alter the natural flow regime or temperature in rivers can affect salmon migration. Currently very little is known about the effects of hydraulic conditions on migrating salmon. If we learn more about how fish use particular river flow fields to aid their upriver migration we may be able to improve management practices which alter river flow conditions in favour of fish passage. This thesis explores the upriver energy consumption and migration behaviour in sockeye salmon (O. nerka), and pink salmon (O. gorbuscha), in relation to moderate (10-100's m long) and small (1-5 m long) scale flow fields. Research in the area of upriver migration is important for several reasons. Migrants use 75-95% of body fat and 40-60% of body protein migrating to their spawning rivers (Idler and Clemens 1959; Gilhousen 1980). If body energy stores are depleted prior to arrival at the spawning grounds or are too limited to develop gametes, salmon may not be able to spawn successfully. Difficult river conditions during migration which require excessive amounts of energy to navigate may be detrimental to the survival of fish and their success in spawning. Limiting this success jeopardizes the long-term survival of geographically and genetically distinct breeding populations (stocks) as well as the species themselves. Natural selective pressures exist on behaviours that affect the rate at which stored energy is consumed in fish (Ware 1982; Bernatchez and Dodson 1987; Hinch and Rand 2000). As a result, if behaviours differ between fish stocks different selective pressures may be exerted on each stock (Hinch and Rand 2000). Selective pressure in the form of sexual selection, sexual competition, migratory bioenergetics and behaviour may exist within different time scales (Priede 1985) making it important to study energy use over large and small spatial-temporal scales. Earlier salmon energetics work studied energy consumption over an entire migration (from ocean to upriver spawning grounds) using body constituency analysis of salmon at different points during migration (Idler and Clemens 1959; Gilhousen 1980). This work gave insight into energy use on an entire watershed scale but provided no information on migration behaviour. In contrast, Fretwell (1981) recorded some detailed behaviour patterns with standard radio tracking devices. These, however, were not capable of directly measuring energy use. Studies in which energy use has been linked to behaviour in salmonids has been conducted in laboratory situations in attempts to describe optimum swimming speeds that minimize the energy cost of locomotion (Brett 1982; Williams and Brett 1987). Recently, these laboratory-derived relationships and models have been combined with in situ measurements of electromyogram (EMG) radiotelemetry to generate detailed observations on migratory behaviour and energy expenditure on a scale of seconds to minutes (Hinch and Rand 1998; Rand and Hinch 1998; Geist et al. 2000). Underwater videography has also recently been used to assess very small-scale behaviours during migration through specific (1-2 m2 river area) hydraulic characteristics (Hughes and Kelly 1996; Hinch and Rand 2000). The goal of this thesis is to complement the work that has been carried out on long distance migrating sockeye salmon in the Fraser River Canyon (Hinch and Rand 2000). Using EMG telemetry and videography, Hinch and Rand (1998, 2000) have observed differences in energy consumption during migration between size, sex and stock of sockeye salmon. In addition, these techniques have been used to identify areas of difficult passage through the Fraser Canyon for migrating sockeye salmon. One component of the research was to use EMG telemetry to assess migration behaviour and energy consumption of adult pink salmon in the Fraser River Canyon. Very little work on bioenergetics has been conducted with pink salmon to date. This thesis also examines pink salmon bioenergetics by conducting a comparison in migration behaviour and energy expenditure of pink salmon with that already conducted on sockeye salmon. A comparison of the two species helps to generalize the understanding of energy use and behaviour during upriver Pacific salmon migration. Another component of the research was to use videography to relate sockeye salmon energy expenditure to small-scale hydraulic features. This study builds our knowledge of how small-scale current velocities and hydraulic features affect energy use in migrating salmon. Both the moderate scale EMG telemetry and the small-scale underwater stereovideography work within this thesis complement each other and should provide insight into the evolutionary and ecological aspects of anadromous fish migration. The thesis is structured in five chapters. The remainder of this first chapter provides a review of Pacific salmon life history and migration characteristics with an emphasis on pink and sockeye 4 salmon. Chapter two examines the differences in energy use between size, species, and sex of pink and sockeye salmon through hydraulically different reaches of the Fraser River Canyon based on EMG telemetry. Chapter three uses the same data set to examine the behaviour and instantaneous swimming speeds of these fish in relation to four hydraulically diverse reaches. Chapter four considers energy use of upstream migrating sockeye salmon in relation to specific (1 to 5 m2) hydraulic fields using underwater videography. Chapter five concludes the thesis by briefly outlining the direction for future research. Salmon Migration and Life History Pacific salmon demonstrate a remarkable life history. After incubating in fresh water streams and rivers, juvenile salmon hatch from eggs, move through the gravel substrate and into the water column. Some species rear in their natal rivers (0. tshawytscha; O. kisutch), while, some migrate downstream to freshwater lakes (0. nerka) or directly to the ocean (O. gorbuscha, O. keta) (Groot and Margolis 1991). After a freshwater rearing time, salmon carry out extensive migration patterns along the coast of BC and Alaska, as well as in the open ocean where they feed and grow until they return one to seven years later to their natal streams (Groot and Margolis 1991). Somatic growth takes place during the first phase of a salmon's ocean life. The final phase of marine life growth is primarily gonadal. Feeding conditions and growth rates in the open ocean alter the fecundity of salmon, and it has been suggested that the longer a fish waits to mature the greater its fecundity (Persov 1963; Grachev 1971). Research suggests that larger, heavier fish produce more eggs (Foerster and Pritchard 1941). The final maturation of gonads takes place in freshwater after feeding has ceased. Energy for final gonad maturation is derived entirely from stored body nutrients. Early studies by Eniutina (1972) in the Im River, Japan, documented gonadal weight as high as 15% and 25% of total body weight in male and female pink salmon respectively. In order to ensure healthy and viable gonadal production, body energy reserves are essential. In addition to gonadal formation, fish migrating upstream undergo considerable physiological changes (Heard 1991). The alimentary system of the non-feeding fish atrophies. The physical shape of the fish changes, particularly for males, who develop a large dorsal hump and pronounced kype. Overall energy conservation during migration is essential for salmon to 5 successfully reach their spawning grounds and spawn effectively. In a broader evolutionary and ecological context salmon face trade-offs between body size and shape versus fecundity when completing upriver migration and during sexual selection. For example, particular body forms such as well developed sexual characteristics of an increased hump and kype may be of benefit to the fish on the spawning grounds, however, they may not be efficient during upriver migration. The extent of these trade-offs should change between stocks and between sexes. Understanding the energetics of migrating salmon may help improve management strategies that maximize fish gonad maturation and reproductive success. Life history variation exists between and within species of Pacific salmon depending on spawning ground location, environmental conditions, and genetic differentiation. Salmon return to their natal stream to spawn, causing the development of genetic specificity within streams. These distinct groups of genetically similar fish are considered stocks. Pink salmon (O. gorbuscha) are the most abundant salmon in the Pacific Ocean. They comprise 40% by weight and 60% by number, of all Pacific salmon (Neave et al. 1967). Pink salmon are also the smallest of the Pacific salmon species with weight at maturity averaging 1.0-2.5 kg (Heard 1991). Pink salmon have a two-year life cycle, entering the estuary at 3 cm in length and returning to spawn two years later at a length of 45-55 cm (Ricker 1964). They are thought to migrate in the north-central Pacific, covering distances of 5,500-7,500 km. Some pink salmon spend time milling or gathering at the mouth of their natal river waiting until they mature before ascending (Davidson 1943). Generally they have less extensive upriver migrations than other salmonids, spawning in the lower reaches of most river systems (Heard 1991). Upriver migration in the Fraser River system occurs in late summer and fall and the fish return to their natal streams. Sockeye salmon (O. nerka) are the third-most abundant of the Pacific salmon (Burgner 1991). Sockeye salmon exhibit a diverse variety of life history patterns. Generally they have a four-year life cycle, lay their eggs in gravel streams, and spend their juvenile stages rearing in lakes for one to three years. Anadromous sockeye migrate to the ocean where they stay for one to four years. During this period they move continuously across the north Pacific in feeding migrations of over 3700 km (Royce et al. 1968). Distinct populations of sockeye have developed unique differences in migration length, spawning and rearing habitat choice and body morphology (Gilhousen 1980; Burgner 1991). One of the most dramatic distinctions of sockeye populations is seen in kokanee. These sockeye spend their entire life history in freshwater, never entering the sea. Sockeye salmon do not grow as rapidly as other species, and take four years to reach an average weight of 2.56-3.16 kg (Fraser River stocks averaging 2.73) (Burgner 1991). The majority of upriver migrations for pink salmon occur between Puget Sound Washington and Alaska (Vernon et al. 1964). In the Fraser River, BC, only the odd year runs have significant numbers (Neave 1952; Ricker 1964; Ricker 1989). Even year runs become dominant farther north (Aro and Shepard 1967). This even/odd year dominance is not completely understood but could be the result of particular environmental events during migration that acted as intensive selective pressures (Heard 1991). The Fraser River system has both an early and late run of pink salmon (Neave 1966). The upstream spawners enter the river two weeks ahead of the downstream spawners (Ward 1959). In the Fraser River the pink salmon which migrate the longest distance spawn in the Seton River at Lillooet, a distance of 250 km from the ocean (Ward 1959). Both the Skeena River (Northern BC) and Snake River (eastern Washington) have or have had pink salmon runs which migrate distances of greater than 480 km (Godfrey et al. 1954) and 694 km (Basham and Gilbreath 1978) respectively. Most runs however, are of considerably shorter distances and spawning can take place intertidally (Heard 1991). There are significant differences in the body sizes of adult pink salmon. These differences are seen between fish in different areas, between even and odd years, and between fish of different genetic lines or stocks within years (Neave 1966). There is some speculation that the size of the spawning population may influence the migration distance, with larger populations migrating farther distances (Vernon 1962). Odd year pink salmon are larger than even-year pinks (Godfrey 1959), suggesting a correlation between fish size and success of migration when compared with numbers of migrants. Major morphological changes in adult pink salmon may also affect migration behaviour and energy usage while swimming upstream, however little research has been conducted in this area. Populations of sockeye salmon spawn in the temperate and sub-arctic North Pacific Ocean as well as the Bering Sea and the Sea of Okhotsk. The majority of sockeye populations in North America spawn between the Columbia River and the Kuskokwim River in the Bering Sea (Burgner 1991). One of the largest spawning complexes of sockeye salmon is located in the 7 Fraser River drainage of British Columbia. There are many distinct stocks of sockeye which migrate up the Fraser mainstem to spawn in various tributaries upstream. This migration occurs from early July to early September depending on the stock (Gilhousen 1960). The variation in migration distances is enormous. The early arriving stocks in July travel up to 1,20.0 km to the Stuart Lake system, mid-summer stocks travel 550 to 885 km to the Chilko and Stellako rivers, and the late runs in September and October to the Adams River travel distances of only 386 km (Burgner 1991). The elevation each of these stocks has to climb during migration also varies, with the Chilko River fish reaching the highest elevation at 1200 m, and the Weaver Creek fish climbing just above sea level. The average weight and length of sockeye salmon in the Pacific Ocean varies depending on stock (Burgner 1991). Although the average weight of Fraser River fish is 2.73 kg, variation between stocks is documented. In addition, differences in condition factor (often measured as a combination of girth, weight and colour) are noted for various individual stocks of Fraser River sockeye (Gilhousen 1980). Because body shape and swimming behaviour are directly related to hydraulic drag (Yates 1983), the differences in required distance and elevation traveled by fish may have acted as a selective pressure affecting fish body shape and energy efficiency in swimming. Comparative study of interior and coastal juvenile coho suggests that swimming stamina is related to body form and is an adaptive trait (Taylor and McPhail 1985). Interior fish had a more fusiform body shape in contrast to the robust and deeper bodied coastal fish which is suggested to be due to the increased energetic demands of the freshwater migrations of the interior fish (Taylor and McPhail 1985). To date, little research has focussed on this question within adult populations. 8 Chapter 2: Energetic costs of migration through the Fraser River canyon, British Columbia, in adult pink (Oncorhynchus gorbuscha) and sockeye (O. nerka) salmon as assessed by EMG telemetry Abstract Adult Pacific salmon (Oncorhynchus sp.) depend on energy reserves to complete their upriver spawning migration. Little is known about how local river features such as flow patterns and bank characteristics affect salmon energetics nor how species differ in reach-specific energy use. In the fall of 1999, electromyogram (EMG) radio telemetry was used to describe activity levels and estimate energy use of 12 adult pink salmon (O. gorbuscha) during their upriver migration. Individuals were tracked continuously through a 7 km section of the Fraser Canyon. River reaches varied considerably in flow velocity and complexity. Similar data collected on sockeye salmon (O. nerka) in a previous study were reanalyzed to provide a species comparison of energy use. Across all reaches sockeye salmon tended to use more energy (0.93 J«m_1) than pink salmon (0.26 J*m"'; P=0.0986). In all analyses, reach was a significant factor determining swimming speed, ground speed and energy use (P<0.0009 for all ANOVAs). Within sex and species groups all fish increased their activity levels when they migrated through constricted reaches compared with non-constricted reaches. Differences in behaviour between sex-species groups depended upon reach. This study suggests that within constricted reaches the river area has higher velocity hydraulic conditions which result in two behavioural responses with similar energetic consequences. Fish either swim slowly and take more time finding lower velocity areas for migration thus increasing passage time, or they swim quickly through higher velocity flow fields, both behaviours are energetically expensive. This study concludes that although species and sex are relevant, river reach characteristics are the primary factors affecting energy use during upstream salmon migration in the Fraser River. 9 I n t r o d u c t i o n Upriver migration of anadromous fish can be energetically expensive particularly when travel distances are long and river environmental conditions adverse (Idler and Clemens 1959; Gilhousen 1980; Bernatchez and Dodson 1987; Brett 1995; Leonard and McCormick 1999). Body constituent analyses on Pacific salmon reveals that, depending on migratory conditions, more than 50% of body energy may be utilized to complete migration (reviewed by Brett 1995). Such high levels of energy depletion are thought to cause enroute or prespawning mortality (Cooper 1983; Hinch and Bratty 2000; Macdonald et al. 2000). However, very little is known about the mechanisms responsible for high levels of energy use. Constituent analyses provide only a coarse grained measure of how a stock uses its energy reserves and cannot be used to measure how specific habitat features influence energy use. Because fish must be killed to measure body energy, different individuals must be examined each time. Therefore, the influence of behaviour or morphology cannot be evaluated by this method. By contrast, electromyogram (EMG) telemetry enables one to quantify swimming speeds of individual migrating fish, and thus provide fine-scale information on energy-use patterns (McKinley and Power 1992; Hinch et al. 1996). EMG telemetry has been used for several years to investigate the factors responsible for variability in migratory energetics of Pacific salmon (Oncorhynchus spp.) (Hinch et al. 1996; Hinch and Rand 1998; Geist et al. 2000; Hinch and Bratty 2000). Most research has focussed on early Stuart sockeye salmon, the Fraser River sockeye stock that migrates the furthest upriver (ca 1200 km). Earlier research found that sex, size and individual swimming speed patterns are important factors affecting energy use in sockeye salmon (Hinch and Rand 1998; Hinch and Bratty 2000; Hinch and Rand 2000). Physical habitat features have also proven to significantly influence sockeye energy use. Specifically, river constrictions, and their associated multidirectional currents, cause relatively high passage costs (Hinch et al. 1996; Hinch and Rand 1998). Most of our detailed knowledge on salmon migration energetics and behaviour comes from the study of long distance migrating stocks of sockeye salmon (O. nerka); (Bernatchez and Dodson 1987; Hinch and Rand 1998; Hinch and Rand 2000). Using a modeling approach, Bernatchez and Dodson (1987) suggested that short distance migrants do not behave the same as long 10 distance ones, yet there has been little field investigation into the behaviour and energetics of short distance migrating stocks or of species other than sockeye. This study focuses entirely on long distance stocks to better isolate interspecies differences. In this study the first objective was to use EMG telemetry to estimate swimming speeds, ground speeds and energy use of adult Seton River pink salmon (O. gorbuscha) during their migration through the Fraser River canyon. I predict that, as previously found with sockeye, female pink salmon will be more energetically efficient than males, and that river constrictions will increase energy expenditures. The second objective was to compare pink salmon swimming energetics to that of early Stuart sockeye salmon previously described by Hinch and Rand (1998). To effectively make this comparison, a re-analysis of the early Stuart sockeye data was required. Although they travel the longest distance inland of any Fraser River pink salmon stock, Seton pink salmon only travel about one fourth the distance of early Stuart sockeye salmon. Previous swim tunnel experiments indicate that pink salmon are weaker swimmers than sockeye salmon possibly owing to their smaller size (Brett 1982; Williams and Brett 1987). This study predicts that Seton pink salmon are less efficient in their use of energy for swimming during migration compared to the early Stuart sockeye. Methods Swimming tunnel study In October of 1999, eight adult pink salmon, four males (fork length, 47.2-55.2 cm; mass, 1.25-2 kg) and four females (fork length, 49.6 to 55.0 cm; mass, 1-1.75 kg), were collected using dip nets from the fish ladder at the Seton Lake dam near Lillooet, British Columbia (Figure 1). Following the methodology of Hinch et al. (1996), EMG transmitters (Lotek Engineering, Newmarket Ontario) were surgically implanted into the abdomen of each fish. To generate the mathematical relationships between EMGPI and TBF for adult pink salmon, transmitter-carrying fish were individually swum in a large swimming tunnel (Farrell et al. 1990) which was set up at the dam. The tunnel had a volume of 471.2 L and was capable of creating velocities between 0 and 150 cm*s"'. The solid blocking effect of each fish (the fish body area relative to the cross-sectional area of the tunnel), which results in an increase in water velocity around the fish, was calculated according to Bell and Terhune (1970) and swimming velocity was adjusted accordingly. Speed of the pump was manipulated to obtain a given water velocity. 11 Figure 1. Map of Fraser River canyon indicating 9 reaches (numbered) that were studied by Hinch and Rand (1998). Inset is a map of British Columbia, Canada showing the Fraser River and its main tributaries, the locations of spawning destinations for the two stocks of salmon in this study, and the location of the study area. 12 EMG tagged fish were placed in the tunnel chamber and allowed to habituate for at least two hours. Fish were then exposed to different water velocities (increasing incrementally at 0.15 body lengths per second (BL«s"') intervals from 0.45 BL«s"' to 3.00 BL»s_1. The duration of each interval was five minutes until the fish was swimming at approximately 50% of its critical speed (U c rit). UCrit is defined here at the maximum speed a fish can swim for a period of 60 minutes before reaching exhaustion. Thereafter intervals were 20 minutes until the fish exhausted (Jain et al. 1997; Jain et al. 1998). A fish was exhausted when it lost its ability to control its swimming and remained on the electrified (6 V) shocking grid at the rear of the swim chamber for 30 consecutive seconds. At that point, the water velocity was reduced to 0.45 BL«s"' and the fish was allowed to recover. A stereo video camera system synchronously recorded onto a single video image the swimming fish and the digital face of the EMG radio receiver (model SRX 400, Lotek Engineering, Newmarket, Ontario). The EMG radio receiver also displayed every EMGPI that was transmitted from the fish, thus allowing each EMGPI value associated with the initiation of each tailbeat to be transcribed. Because the cameras record 60 images per second, TBF was determined by counting the number of video frames required to complete a tailbeat. Therefore, for each fish, a relationship was developed between EMGPI values and TBF. Pink salmon have limited time and energy stores for migration. Retaining fish for individual calibration prior to tracking risked causing increased exhaustion and mortality within the study sample. An independent set of fish were used to develop the calibration equation in order to prevent unnatural exhaustion and death in fish studied through the Fraser River canyon. Telemetry study The Fraser River canyon, situated in the southwest corner of British Columbia (Figure 1), exhibits a diversity of flow conditions and habitat features that create challenges for upriver passage of salmon (Cooper 1983; Hinch and Rand 1998; Hinch and Bratty 2000). The study area in the Fraser River canyon was from kilometer 150 to 157, the same as that used by Hinch and Rand (1998), and the same subreaches and bank morphology classifications were also used (constrictions, bends and straight banks). Reach length, width and gradient were determined using 1:4000 air photos. Distances were measured from the center of the river. Reaches 1 and 2, located immediately downstream of the canyon, had an average gradient of 75 cm^km"1. All other reaches were in the canyon and had an average gradient of 120 cm^km"1. Reaches ranged in length from 200 to 1100 m and in width from 50 to 500 m. Surface water speeds were 13 estimated by timing the passage of wooden blocks (20 x 20 x 5 cm) over a known distance in the main current (thalweg) of the river. Within each reach, surface currents were qualitatively described based on the surface velocity and direction of the main current. Main current direction was categorized as split, bent and linear. A split current occurred where the main current was divided by a large obstruction such as an island. A bent current was where the main current was deflected by an inriver obstacle, bank or gravel bar. And a linear current occurred where the main current ran roughly parallel to the banks of the river and did not change direction relative to the banks. Water temperatures (measured by Fisheries and Oceans Canada at Hell's Gate, situated approximately 10 km upstream of the study area) were 13.9-15.4, and 14.0-17.0 °C during the field seasons in 1999, and 1993 respectively. Daily discharge (measured by Water Survey of Canada at Hope BC, approximately 20 km downstream of the study area) ranged from 1900-4030 (mean 2758), and 3170-3770 (mean 3295) m3«s"' for 1999, and 1993, respectively. During September and October 1999, six male pink salmon (fork length, 47.5-56.6 cm; mass, 1.50-2.25 kg) and six female pink salmon (fork length, 47.5-50.6 cm; mass, 1.00-1.35kg) were collected from a fish wheel situated 200 m downstream of the study area and tags were implanted at that site. The wheel was operated by the Yale First Nations Band (Yale, BC). Fish were held in a pen near the fish wheel for at least one hour to recover from the surgery. Fish were released into a large slow water eddy at the edge of the river mainstem near the holding pen and individually tracked upstream using a hand-held directional three element Yagi antenna and EMG receiver. Only one fish was released per day. Fish were tracked by following them on foot along the river banks, and their positions could be ascertained to within 5 m. Migration time was measured as the amount of time it took a fish to swim through a reach from first entry to first exit. Because pink salmon did not migrate at night, time from dusk to dawn were not included in the migration time. Ground speed was calculated by dividing the migration time by the reach length. EMGPI values were recorded by the receivers at an interval of every 3-5 seconds. EMGPI values were converted to TBF in beats per second using the linear regression equations developed from the swimming tunnel trials (equations given below in results). TBF were converted to estimates of instantaneous swimming speeds using a relationship developed by Brett (1982) for adult pink salmon (TBF = 0.8685 (swimming speed) + 54.382). For an individual fish to pass through a specific reach, energy use was estimated from the average swimming speed through the reach, 14 the daily average river temperature, the fish mass and the migration time through the reach using a sockeye salmon bioenergetics model (Hinch and Rand 1998). This model was originally developed by Beauchamps et al. (1989). A sockeye model was used because bioenergetics models for adult pink salmon have not been developed. Using swimming tunnel experiments, Williams and Brett (1987) found that U c rit values for a long distance migrating stock of pink salmon were similar to those of sockeye salmon. Thus it is reasonable to assume that energetic model parameters may also be similar. Hinch et al. (1998) used EMG transmitters to track five male and five female early Stuart sockeye salmon in 1993 through the same reaches described above. Using results from laboratory swimming experiments and the other approaches described above, they collected ground and swimming speeds, and estimated reach specific energy expenditures. It is these data that are used for making comparisons with pink salmon. Preliminary analysis of the pink salmon data suggested that individual fish may be an important factor explaining variation in swimming and ground speeds as well as energy use. A re-analysis was conducted of Hinch and Rand's (1998) data because they did not consider 'individual fish' as a factor responsible for variation. Statistical analyses ANCOVA was used to compare the slopes and intercepts of the linear relationships between TBF and EMGPI from the swim tunnel trials to determine if separate predictive relationships were required for fish of specific sizes or sex. If error variance heteroscedasticity existed, TBF was log transformed prior to analysis. Differences in mass and length between the sexes offish used in the swimming tunnel were determined by ANOVA. To assess differences in condition between the sexes, sex-specific log-mass -length regression relationships were compared with ANCOVA. To examine the speeds and energy use of upriver migrating pink salmon, three separate split-plot repeated measures ANOVAs were used and then type III sums of squares were interpreted (SAS 1988). Separate analyses were performed using ground speed (cra»s"'), swimming speed (cm»s" '), and energy use (J*m_1) as dependent variables. In each analysis, individual fish was the repeated measure, and sex and reach were the main effects. Within this model, sex was tested using the fish nested in sex error term and all other dependent variables were tested using the full 15 model error (reach*fish(sex)). Al l models were tested for interaction between sex and reach. If an interaction was detected, further two-way ANOVAs with sex and reach as main effects were conducted to isolate the interactive terms. Differences among individuals, sexes, and reaches were determined a posteriori using least square means. Pearson's correlation between weight and swimming speed, ground speed and energy use were also examined in order to speculate on the role that fish size may have on the response variables. Differences in weight between fish was determined using two-way ANOVA and a Least Significant Difference t Test. Because Hinch and Rand (1998) did not consider individual fish as an independent variable in their analyses of sockeye salmon, their data on energy use, swimming speed and ground speed were re-analyzed using the same split plot repeated measures ANOVA as described above. To compare energy use of sockeye and pink salmon, species and sex classes were combined into one main effect variable that was termed species-sex class that had four levels (sockeye females, sockeye males, pink females and pink males). Because a sex by species interaction could not be tested for in these analyses, two additional split plot repeated measures ANOVAs were conducted with sex and then species replacing sex-year class as main effects. One-way ANOVAs were used to compare fish body mass among the sex and species groups. Least square means (LSM; SAS 1988) were used as the a posteriori approach for assessing differences among levels within the class variables. Following each ANOVA, plots of residual versus predicted values were examined to look for error variance heteroscedasticity. If present, log transformations were applied and the ANOVAs redone. Statistical significance was assessed at the 0.05 level. Significance levels for all a posteriori contrasts were adjusted using Bonferroni's method. The coefficient of variation (CV) of ground speed, swimming speed and energy use was calculated for each species-sex class. Two-way ANOVAs were then used to compare CV values among sex-species and reach. Sample sizes for all statistical procedures are listed in the text. Please refer to the relevant figures to determine how sample sizes were obtained. 16 Results Swimming tunnel study Pink salmon ranged in length from 47.2 to 55.2 cm (mean 52.0 cm, n=8) and in mass from 1.0 to 2.0 kgs (mean 1.6 kgs). There were no differences in mean length and mean mass between sexes (P=0.9642, P=0.4583, respectively). Furthermore, neither the slopes nor intercepts of the log-mass-length regression relationships differed between sexes (P=0.4991, P=O.5690, respectively). The TBF - EMGPI regression slopes differed among individuals (ANCOVA, P-0.0001). These differences were primarily due to sex. When separate ANCOVAs for males and females were conducted, the slopes of the male fish did not differ (P=0.3147), nor did those of the female fish (P=0.2081). Therefore, common regression relationships were developed for males and females by pooling data within sexes (Figure 2). Fish mass may also be an important independent variable helping to predict TBF in these regression relationships (Hinch and Rand 1998). Multiple regression revealed that mass contributed to explaining variation in TBF for male (P=0.0001) and female (P=0.0001) relationships. Thus, the regression models that were used for predicting TBF from EMGPI are; for males, log TBF = -0.0004186(EMGPI) + 0.1731621 (mass) + 0.5796416; and for females, log TBF = -0.0003026(EMGPI) -0.1646546(mass) + 0.9171931. Telemetry study Male pink salmon (mean mass 1.79 kg, n=6) were larger than female pink salmon (mean mass 1.18 kg, n=6, P=0.0090). Male sockeye salmon (mean mass 2.64 kg) were larger than female sockeye salmon (mean mass 1.75 kg; P=0.0001). In contrasting sex and species groups, only pink males and sockeye females were not different (P>0.05). The split-plot repeated measures ANOVA of pink salmon data revealed that swimming speeds varied among individual fish (Fio,66=77.48, P=0.0001), and among reaches (F8,66~10.38, P=0.0001) but did not vary between sexes (Fi;io=0.00, P=0.9625). Swimming speeds of pink salmon ranged from 51 to 183 cm»s"' (mean 114 cm»s"', n=93, SE=2.9) among fish and reaches. In paired comparisons (Bonferroni adjusted a=(0.05/9)), there were no differences in pink salmon swimming speeds between reaches 1, 2, 3, 4, 6, 7, and 9 (P>0.0230 for all contrasts), reaches l,2,3,5,7,and 9 (P>0.0119 for all contrasts) and reaches 1, 5, 7, 8, and 9 (P>0.0375 for all 17 o co • co +-> o <D cr J-H H 6 5 1 4 3 2 1 6 5 4 3 2 1 MALES r2=0.84 0 log TBF = -0.0004186(EMGPI) + 0.1731621(mass) + 0.5796416 0 500 1000 1500 2000 FEMALES r2=0.53 log TBF = -0.0003026(EMGPI) -0.1646546(mass) + 0.9171931 500 1000 1500 2000 EMGPI (ms) Figure. 2 Electromyogram Pulse Interval (ms) vs. Tailbeat Frequency (beats/min) for male (n=4) and female (n=4) pink salmon developed using a Brett-type respirometer. Fish tested in the tunnel showed similar range and spread of data points. 18 contrasts). Pink salmon, swam through reaches 4 and 6 faster than reaches 5 and 8 (LSM P<0.0016 for all contrasts). Swim speed was correlated with mass in males (r=0.91, n=6 fish, P=0.0113) but not in females (r=-0.40, n=6 fish, P=0,4325). Pink salmon ground speed data required log transformation. The split-plot repeated measures ANOVA revealed no differences among individual fish (Fio,64-l-27, P=0.2643), or between sexes (Fijc^O.09, P=0.7758) in pink salmon ground speeds, however, differences were evident among river reaches (Fg,64= 6.80, P=0.0001). Average ground speeds varied from 21.2 to 114.3 cm«s"' (mean=57.1 , n=9, SE=11.13) between reaches. In pairwise comparisons (Bonferroni adjusted ot=(0.05/9)), there were no differences in pink salmon ground speed between reaches 5, 1, 2, 6 (P >0.0372 all contrasts) or reaches 7 and 9. (P=0.3820). The fastest ground speeds were through reaches 1 and 5 which were significantly different from reaches 3, 4, 7, 8, and 9 (LSM P<0.0055 for all contrasts) and reaches 7 and 9 had significantly slower ground speeds than reaches 1, 2, 5, and 6 (LSM P<0.0033 for all contrasts). Ground speed was not correlated with mass for females (P=0.5822) or males (P=0.5999). Pink salmon energy use data required log transformation. The split-plot repeated measures ANOVA revealed that variation in energy use was attributed to differences among individual fish (Fi0)64=l 1.29, P=0.0001), and among reaches (F8,64=8.36, P=0.0001), but not between sexes (FU (p0.75, P=0.4058). Average energy use varied from 0.002 to 1.73 J»m"' (mean=0.26 , n=92, SE=0.032 ) among reaches and individuals. Although not significant, females tended to use less energy than males in all reaches (female mean=0.16 J«m"' male mean=0.36 J«m"'). In pairwise comparisons (Bonferroni adjusted a=(0.05/9)), there were no differences in energy use between reaches 3, 4, 6, 7, and 9 (P>0.1053 for all contrasts). There were also no differences between reaches 1 and 5 (P=0.0092). Pink salmon used the least amount of energy migrating through reaches 1 and 5 and the most migrating through reaches 4, 7, and 9 (P<0.0040 for all contrasts). Energy use was correlated with mass in males (r=0.96, n=6, P=0.0024) but not in females (P=0.5822). Analyses on the sockeye salmon data revealed differences in swimming speed between individual fish ^7,45=29.20, P=0.0001). Among fish, swimming speeds ranged from 50 to 179 cm»s"' (mean=101, n=9, SE=13.4). An interaction between sex and reach also existed in the 19 model (Fgi45=3.74, P=0.0020). In pairwise comparisons, there were no differences in swimming speed between the sexes in reach 1, 2, 3, 4, 6, 7, and 8 (LSM P>0.0169 for all contrasts). Females swam slower than males in reach 5 (LSM P=0.0001) and 9 (LSM P=0.0020). In pairwise comparisons of females there were no differences in swimming speed between reaches 1, 2, 3, 4, 6, and 7 (P>0.0112 for all contrasts), reaches 2, 5, 8, and 9 (P>0.0194 for all contrasts) or reaches 2, 4, 5, and 7 (P>0.0151 for all contrasts). Females swam faster through reaches 1, 3, 4, 6, and 7 than reaches 8, and 9 (P<0.0030 for all contrasts). In pairwise comparisons, males swam faster in reach 5 than reach 8 (P<0.0055). Al l other reaches showed no differences (P>0.0065 for all contrasts). Swimming speeds were weakly correlated with mass in females (r=0.88, n=5, P=0.0513) but not with males (P=0.2569). Ground speed data for sockeye salmon required log transformation. The split plot repeated measures ANOVA revealed no differences in ground speed between sexes (Fi>7=2.12, P=0.1885) or individual fish (1^7,45=2.01, P=0.0743). However, there were differences in ground speed among reaches (F8,45=4.39, P=0.0006). Average ground speed ranged from 19.2 to 61.2 cm»s"' (mean=38.3 cm»s'', n=9, SE=4.9) among reaches. In pairwise comparisons (Bonferroni adjustment a=0.05/9) there were no differences in ground speeds of sockeye salmon between reaches 4, 7, and 9 (P>0.1285 all contrasts). There were also no differences between reaches 1, 2, 3, 5, 6, 8, and 9 (P>0.0266 for all contrasts). Sockeye salmon had slower ground speeds in reaches 4 and 7 than reaches 5 and 2 (P<0.0039). There were no differences in energy use between male and female sockeye salmon (^ 1,45=0.83, P=0.3917). However, there were individual differences (F7,45=8.09, P=0.0001) and the interaction was significant between class and reach (Fg>45=2.58, P=0.0207). Energy use for sockeye salmon ranged from 0.02 to 10.05 J*m"' (mean=0.94 J«m_1, n=70, SE=0.24) among reaches and fish. A posteriori comparisons revealed that there were no differences in energy use between sexes in any of the reaches (P>0.0266 for all contrasts) with the exception of reach 5 where female fish used less energy than male fish (female mean=0.08 J»m_1, male mean=0.85 . J»m"'; P=0.0007). However, female fish tended to use less energy than male fish in reaches 2, 4, 7, 8 and 9 (Figure 3). In pairwise comparisons, female sockeye energy use did not differ between reaches 1, 2, 3, 4, 6, 7, and 9 (P>0.0067 for all contrasts), reaches 1, 2, 3, 4, 6, 8, and 9 (P>0.0149 for all contrasts) or reaches 1, 2, 5, 8, and 9 (P>0.0070 for all contrasts). Female 20 ro 00 I CT> CN U 1->/-> H T 3 S in ID *S a c o C w —< o _ -2 J H ."S3 >^  8 rtaU r- ro -B "*> ro U M 3 U •_ G _ u —i 2 £ cn (U ccJ .2 § « o * 2 i « § I 3 3 . S P _ ?3 "3 ' 3 u Q. 09 1 X u OB C • t-* cn •a cu -4—I a u sockeye used more energy migrating through reach 7 (mean=0.79 J»m_1) than reaches 5 (mean=0.08 J»m"'; P=0.0001) and 8 (mean=0.11 J*m"'; P=0.0022). In pairwise comparisons (Bonferroni adjusted a=0.05/9), male sockeye energy use did not differ between reaches 3, 4, 5, 6, 7, 8, and 9 (P>0.0089 for all contrasts) or between reaches 1, 2, 3, 5, 6, 7, 8, 9 (P>0.0171 for all contrasts). Male sockeye used more energy migrating through reach 4 (mean=4.78 J»m"3) than reaches 1 (mean=0.18 J-m"1; P=0.0029) and 2 (mean=0.18 J-m"1; P=0.0008). Energy use was correlated with mass in females (r=0.96, n=5, P=0.0.0097) but not with males (P=0.3425). The repeated measures A N O V A examining pink and sockeye salmon swimming speeds revealed significant differences between individual fish (Fi7>iii=6.53, P=0.0001), reach (Fs,in=6.53, P=0.0001) and an interaction between reach and species-sex class (F24,I'HF2.18, P=0.0035). The main effect of speciesrSex class was not significant (F3)i7=0.57, P=0.6392). Because of the highly conservative nature of Bonferroni correction the number of sites were reduced to more effectively interpret the interaction term (Bonferroni adjusted ct=0.05/16). Therefore reaches 2, 4, 7 and 9 were scrutinized because they have good data coverage and still represent the full, among reach, range of hydraulic conditions in this study. (These four sites will be further examined for specific migration behaviours in Chapter 3.) A Bonferroni adjustment of a=0.05/16 is a highly conservative interpretive tool, and resulted in no significant a posteriori differences being detected in any comparisons despite the significant interaction term. Pink females (pf) tended to swim faster than sockeye females (sf) in reaches 4 (pf mean=121 cm-s"1, sf mean=97 cm»s"', P=0.0613), 2 (pf mean=117 cm-s"1, sf mean=78 cm-s"1, P=0.0181), and 9 (pf mean=95 cm*s"', sf mean=63 cm»s"', P=0.0060), but not in reach 7 (P=0.5211). Pink females tended to swim faster than sockeye males (sm) in reach 2 (pf mean=l 17 cm«s'', sm mean=91 cm»s"', P=0.0608) but not in the other reaches (all P>0.26). Pink males (pm) tended to swim faster than sockeye males in reach 2 (pm mean=l 17 cm«s"', sm mean=91 cm«s"', P=0.0576) but not in reaches 4, 7, and 9 (P>0.2885). Pink males also tended to use less energy than sockeye females in reach 2 (pm mean=l 17 cm«s"', sf mean=78 cm«s"', P=0.0170), 4 (pm mean=125 cm«s"', sf mean=97 cm»s"', P=0.0254), and 9 (pm mean=138 cm«s"', sf mean=63 cm»s"', PO.0076) but not in reach 7 (P=0.5040). Ground speed data for all species-sex classes required log transformation. For ground speed, species-sex class (F3>i7=3.45, P=0.0402) and reach (Fgjio9=9.67, P<0.0001) were significant. 22 Individual was not significant (Fi7,io9=0.82, P=0.6642) and there was no interaction between species-sex class and reach (F24,io9=0.85, P=0.6702). Both pink females (mean=54.2 cm»s"', n=46, SE=6.3; P=0.0003) and pink males mean=63.4 cm-s"1, n=46, SE=10.1, P=0.0017) migrated faster than sockeye females (mean=33.9 cm-s"1, n=34, SE=3.8). Al l sex-species classes had slower ground speeds through reaches 4 (mean=31.6 cm-s"1, n=20, SE=4.9), 7 (mean=28.5 cm-s"1, n=18, SE=4.3), and 9 (mean=24.0 cm-s"1, n=12, SE=3.6), than reach 2 (mean=77.6 cm-s"1,n=20, SE=19.1). Energy use data for all species-sex classes required log transformation. Energy use varied among individuals (Fi7;io9=9.44, P<0.0001) and there was interaction between species-sex class and reach (F24l1io9=2.01, P=0.0082; Figure 3) again we used the reduced data set to interpret the 1 interaction as outlined above. Pink females used less energy than sockeye males (pf mean=0.16 J-m"1, sm mean=4.78 J-m"1) in reach 4 (P=0.0001). Pink males also used less energy (mean=0.61 J-m"1) than sockeye males in reach 4 (P=0.0017). Averaging sex-species classes reach 4 (mean=1.31 J-m"1, n=20, SE=0.59) and 7 (mean=l.l 1 J-m"1, n=18, SE=0.55) had higher energy use than reach 2 (mean=0.19 J-m"1, n=20, SE=0.05; P<0.0001 for all contrasts). When sex-species class was replaced with sex as a main effect in the same ANOVA, energy use varied among reaches (P=0.0001) but not between sexes (P=0.3294). Likewise when species was used as the main effect, energy use varied among reach (P=0.0001) but not between species (P=0.0986). Within group variation in swim speed differed among sex-species classes 11^23=22.26 P=0.0001). Pink salmon females (mean CV=13.2) were less variable than pink salmon males (mean CV=36.6; LSM P=0.0005). Sockeye salmon females (mean CV=22.6) did not differ in swim speed variation from pink salmon females (LSM P=0.0972) or pink salmon males (LSM P=0.0275). Sockeye salmon males (mean CV=55.0) were more variable than female pink salmon (P=0.0001), male pink salmon (P=0.0001) and female sockeye salmon (LSM P<0.0001). Within group variation in log-ground speed ^ 23=0.75 P=0.5361) or log-energy use ^ 23=0.99 P=0.4168) did not differ among sex-species classes. Within species-sex group variation in swimming speed (Fg,23=0.42, P=0.8987), log-energy use (Fg!23=0.99, P=0.4666) and log-ground speed (Fg,23=l .54, P=0.1964) did not differ between reaches. 23 Discussion The role of river hydraulics and habitat-specific current patterns on the energetics of upstream anadromous migrants is poorly understood. Body constituent analyses, which have previously addressed large spatial scale (distances of 10 to > 100 km) energy use patterns in upriver migrating fish (Idler and Clemens 1959; Gilhousen 1980; Brett 1995; Leonard and McCormick 1999) cannot resolve fish behaviour and energy use on small spatial scales (distances of 10 to > 100 m). To address the latter, EMG telemetry was used to increase understanding of the effects of hydraulic conditions on fish energy use and behaviour during migration. This study found that constricted reaches with relatively high surface velocities and splits or changes in direction of the main current were consistently more energetically expensive for migration than linear reaches, regardless of sex and species. Elevated energy use was associated with increased swimming speed and decreased ground speeds. One might expect that migrants would increase their swimming speed in high velocity sites in order to increase ground speed and hence decrease the amount of time spent within that site, since there is limited time to reach spawning areas. Hinch and Rand (2000) used underwater video to examine swimming speed, encountered water velocities and ground speeds of upriver migrating sockeye salmon and found that fish swam at metabolically optimal speeds through low encountered velocities. In other words, fish adjusted their swimming speeds to achieve certain ground speeds that resulted in energetically efficient migration. However, when they encountered currents that were relatively fast, fish did not alter swimming speeds to achieve efficient migration. This paper speculates, as did Hinch and Rand (2000), that fish may be unable to locate migration cues or take longer to find a desired path in areas of high velocity or turbulence. As a result, travel time would be increased leading to an increased energy use per metre. Variability (as measured by the CV) in energy use and swimming speed among sex-species groups also appeared to be greatest in reaches with constrictions and multidirectional currents (Figure 3). Reaches, which are energetically expensive to navigate, may have greater among-individual variation in energy use because fish are swimming near or at their individual optimal energy use rates. As,a result energy efficient swimming behaviours may be more prevalent in reaches with constricted and multidirectional currents. A closer examination of the migration path choice suggests that migrating fish cross the river and back-track more often in constricted sites (Chapter 3;Hinch and Bratty 2000). This may be evidence that the fish are being diverted by strong, turbulent flows or that they are attempting to locate small-scale low velocity fields for upriver migration. 24 Whichever the reason, small-scale hydraulic conditions appear to be of great importance in determining behaviour and resultant energy use in migrating salmon. Differences in energy use between sexes depended upon reach for both species (Figure 3). Within pink salmon, males tended to use more energy than females but this was most apparent at constricted reaches with multidirectional currents. This result for pink salmon is consistent with Williams et al. (1987) who, after swimming pink salmon in a laboratory swim tunnel, found adult male pink salmon used 15% more energy than female pink salmon. Hinch and Rand (1998) found that through all reaches males use more energy than females during migration. The present re-analysis of their data suggests that energy use between sexes depended upon reach. Although females used more energy than males in reach 1,3, and 6, males used more energy than females in the remainder of the reaches including all of the constricted reaches that had multidirectional currents. These results differ from those of Hinch and Rand (1998) for two reasons. First, they chose to use a two-way ANOVA to analyze their data, which, unlike the repeated measures split plot used in this analysis, did not take individual variation into consideration. Large individual variation may be masking stronger differences between sexes. Equally important, was that this analysis used only a sub-sample of their sites. They included three very low velocity sites on the Nechako River as well as the much more constricted Hell's Gate site on the Fraser. Hinch and Rand found strong differences between sexes at Hell's Gate, which supports the findings of this study that differences between the sexes become more apparent in areas of difficult passage. Several factors may explain increased energy use in males when compared with females. Increased selective pressure for migration efficiency may exist for females because they must allocate relatively more energy to egg production than males do for milt production (Idler and Clemens 1959; Gilhousen 1980; Jonsson et al. 1991; Brett 1995). Also males undergo a much more distinct morphological change. Males develop a large dorsal hump and extended kype (Brett 1995). This change in shape may increase their hydrodynamic drag (Weihs and Webb 1983; Videler 1993) and potentially force them to increase their swimming speed and energy expenditure to reach the spawning grounds at the same time as females. There may also be increased physiological differences in the cardiac and metabolic processes between males and females as they near reproductive maturity. During respirometer trials, males swam at consistently higher VO2 max than females of the same size, indicating that females are more efficient swimmers (unpublished data Christopher G. Lee 2000, Biology Department, Simon 25 Fraser University, Burnaby, B.C.). Webb's (1995) theory that fish can minimize energy use by swimming at metabolic and hydrodynamical optimal speeds suggests that differences in physical shape and metabolism between the sexes would cause them to have different energy use minima. Sockeye salmon tended to use more energy per metre than pink salmon. This trend was particularly evident in reaches 4, 5, 7, 8, and 9.1 speculate that interspecific differences in migration paths may be responsible for any differences in energy use between species. Using a split beam echosounder, Xie et al. (1997) found that pink salmon migrated in tight groups close to shore near the bottom of the river, while sockeye migrated in less dense aggregates further from shore and throughout the water column. River velocities generally increase with distance away from the bank and bed of the river (Henderson 1966), which would suggest that sockeye are encountering faster currents that could be more costly to migrate through. Migration distance may also play an important role in energy use between species. Although both study stocks have the greatest migration distances to travel for their species, Seton River pink salmon migrate only one third the distance of the Early Stuart sockeye salmon. In a comparison among species of Pacific salmon, Beacham and Murray (1993) found that distance of freshwater migration to the spawning grounds usually affected fecundity and egg weight. Observed fecundity and egg weight of upper river populations were less than lower river populations. However, when fecundity was standardized for fish size and individual stocks were studied, this relationship was not always apparent (Beacham and Murray 1993). Although the relationship between upriver migration distance and relative fecundity and egg size is not well established, it still may be worth considering as a potential influence on migration energetics. Female Early Stuart sockeye salmon may allocate less energy to egg production than female Seton River pink salmon, thus more energy could be available for migration activities (Fleming and Gross 1989) in the sockeye stock. Selection for energy efficiency in sockeye females may have expressed itself not through swimming but rather through energy allocation tactics. As a result the selective pressure for energy efficiency in swimming may be strongest in pink females who tend to use less energy for swimming than the males. Although similar selective pressures may exist for the sockeye females, it may be weaker and as a result differences between sexes in sockeye salmon appear only when very difficult migration sites are encountered. In both species, variation in swimming speed and energy use tends to be greater within males relative to females. Size and sex are correlated in this telemetry study, in that, within species, 26 males tend to be larger than females. However, if migration variability is regarded as a factor of size, two distinct patterns develop between species. Within pink salmon, it appears that smaller fish used slower and less variable speeds and used less energy than large fish. Because there were no differences in ground speed this would suggest that small fish are more efficient than large fish. The lower energy use of small fish may be due to differences in somatic energy content between large and small fish. Although not apparent in their anadromous populations, Jonsson and Jonsson (1997) found that the somatic energy content per unit mass of resident brown trout (Salmo trutta) increased with increasing body size. Small fish then, by overall mass, and possibly by mass of energy per unit mass, have fewer somatic body resources than large fish to travel the same distance. Because body energy is a limiting factor (Jonsson et al. 1997), fish with fewer body resources will have to swim more efficiently in order to succeed in migration. Within this paper it was found that, in contrast to the pink salmon, small male sockeye salmon swam faster than large males. Hinch and Rand (1998) explained this as a result of subtle shape differences and potentially weaker thrust generation in small fish. A smaller caudal peduncle may require that smaller fish increase their tailbeat frequency (swimming faster) in order to cover the same ground as the larger fish, and thereby cause them to be less efficient in their migration. Both arguments are deserving of further examination. The pink argument assumes that the strongest force of selection is on the energy reserves of the small fish migrating upriver, requiring them to choose a less expensive migration strategy to ensure arrival on the spawning grounds. The sockeye argument assumes that stronger selective forces exist in the competition between males for spawning ground space, with smaller males using migratory tactics that do not conserve energy but rather ensure their arrival at the grounds at the same time to compete with the larger males. Fleming and Gross (1989) found that increased competition in female salmon decreased egg size, supporting the suggestion that small sockeye males may allocate more energy towards aggression and other behaviours not directly linked to efficient migration, thereby decreasing available energy for gametes. This, however, is not the most plausible explanation. Upon arrival at the spawning grounds, Pacific salmon, in particular the long distance migrating stocks, approach their minimum somatic energy levels before death (Idler and Clemens 1959; Gilhousen 1980; Brett 1995). This would suggest that any behaviour that increased energy use during 27 migration would limit an individual's chance at successfully reaching the grounds before falling below the minimum amount of somatic energy required for life. As a result, energy efficiency enroute appears to be the stronger selective force on fish with limited energy stores. Two assumptions were made in order to compare the migration energetics of pink and sockeye salmon. First, although pink and sockeye salmon were studied through the same reaches in the Fraser River, they were studied in different years. However, temperature and river discharge were very similar during the periods of study of the two species, particularly when one considers how variable these features can be inter-annually (Macdonald et al. 2000). Second, a sockeye salmon bioenergetics model was used to predict energy use for pink salmon. Currently a bioenergetics model for pink salmon does not exist. Williams and Brett (1987) determined that the Ucrit for pink salmon was similar to that for sockeye suggesting that bioenergetics parameters may be similar between species, yet Brett (1982) found that pink salmon required higher amounts of energy to maintain U c r i t swimming. These types of conflicting results suggest that a bioenergetics equation for pink salmon needs to be developed and results compared to that of sockeye salmon. Lastly, although not an assumption, sample size was a general limitation to this study. A lack of strong significance in sex and species main effects may be due to small sample size and resulting large individual variation within each species-sex group. It appears that energy use during upriver migration is highly complex and dependent upon river morphology, as well as species, size and sex of salmon. Differences in energy use between sex and species groups during migration may be due to different selective pressures. Trade-offs in energy use versus path choice for different sexes and species of salmon seems to cause a diversity of migration behaviours which may be an example of the flexibility of the salmonids migration strategy to overcome diverse environmental conditions. 28 Chapter 3: Swimming patterns and behaviour of upriver migrating adult pink (Oncorhynchus gorbuscha) and sockeye (O. nerka) salmon as assessed by EMG telemetry in the Fraser River, British Columbia, Canada Abstract Little is known about the behaviours and swimming speed strategies of anadromous upriver migrating fish. Electromyogram telemetry was used to estimate instantaneous swimming speeds for individual sockeye (Oncorhynchus nerka) and pink salmon (O. gorbuscha) during their spawning migration through reaches that spanned a gradient in river hydraulic features in the Fraser River, British Columbia. The main objectives were to describe patterns of individual-specific swim speeds and behaviours, identify swimming speed strategies and contrast these between sexes, species and reaches. Although mean swimming speeds did not differ between pink salmon (2.21 BL«s~') and sockeye salmon (1.60 BL«s"'), sockeye salmon were nearly twice as variable (mean CV; 54.78) in swimming speeds as pink salmon (mean CV; 22.54). Using laboratory-derived criteria, swimming speeds were classified as sustained (<2.5 BL«s_1), prolonged (2.5-3.2 BL«s_1), or burst (>3.2 BL^s"1). No differences were found between sexes or species in the proportion of total time swimming in these categories: sustained (0.7,6), prolonged (0.18), and burst (0.06). Reaches with relatively complex hydraulics and fast surface currents had migrants with relatively high levels of swimming speed variation (e.g. high swimming speed CV, reduced proportions of sustained speeds, elevated proportions of burst speeds,"and high rates of bursts) and high frequency of river crossings. This paper speculates that complex current patterns generated by river constrictions created confusing migration cues, which impeded ability of salmon to locate appropriate pathways. 29 Introduction Adult Pacific salmon (Oncorhynchus spp.) do not feed during their upriver spawning migrations and must rely entirely on energy reserves to complete migration and spawning. Migration can be energetically expensive particularly when environmental conditions are adverse (e.g. caused by elevated discharge or temperature) or travel distances are long (Gilhousen 1980; Bernatchez and Dodson 1987; Brett 1995). Body constituent analyses on Pacific salmon reveal that more than 50% of total energy is often utilized to complete migration (reviewed in Brett, 1995). Relatively high levels of energy use are thought to cause mortality enroute or prespawning (Rand and Hinch 1998; Macdonald 2000). The advent of electromyogram (EMG) radio telemetry, has made it possible to examine activity and behaviour in migrating fish (McKinley and Power 1992; Hinch et al. 1996; Oakland et al. 1997; Hinch and Rand 1998; Hinch and Bratty 2000). Although still poorly understood, it is becoming apparent that swimming behaviour is an extremely important factor affecting energy use and migration success in Pacific salmon (Hinch and Rand 1998). The importance of swimming behaviour was recently demonstrated by Hinch and Bratty (2000) who observed with EMG telemetry that hyperactive swimming patterns in sockeye salmon were associated with passage failure and mortality at Hell's Gate, a notoriously difficult site for passage in the Fraser River, British Columbia. With the exception of Hinch and Bratty, there has been little description or quantification of individual-specific variation in swimming speeds of upriver migrating anadromous fish, nor any examination of swimming speed patterns that are attributable to specific sexes, species, or habitat features. In this study, EMG telemetry was used to estimate instantaneous swimming speeds for individual sockeye and pink salmon during their upriver migration through the Fraser River, British Columbia. The main objectives were to describe patterns of individual-specific swim speeds and behaviours, and contrast reach-specific swimming patterns between sexes and species. To address these issues, telemetry information that was previously collected as part of other studies was utilized (Hinch and Rand 1998; Chapter 2). However, these studies focussed on average energy use and only provided a brief summary description of average swim speeds. They did not report on swimming speed patterns, migration routes, or present any results or interpretation of species, sex or individual-specific activity levels. 30 Description of study sites The Fraser River drains approximately one third of British Columbia and is one of the largest producers of wild Pacific salmon in the world (Northcote and Burwash 1991). Although several of its tributaries are affected by hydro facilities, flows through the mainstem of the Fraser River are not regulated by humans. The research was carried out in the lower section of the Fraser River mainstem (Figure 1). This area exhibits a wide range of flow conditions and habitat features. Al l five species of Pacific salmon migrate through here (Groot and Margolis 1991), making it an ideal location for interspecific comparison of salmon migration behaviour. Our study area was the same as that used by Hinch and Rand (1998). The area is bounded between river kilometer 150 and river kilometer 157 (Figure 1). Hinch and Rand divided their study area into ten contiguous reaches and classified each based on bank morphology (e.g., constrictions, bends, or straight banks), the direction of the major surface flow patterns along their centreline (e.g. flows were primarily downstream, or flows were split and heading in . multiple directions), and centreline velocities. Four of their reaches were chosen (reaches 2,4, 7, and 9) for focus on because they reflect the range of natural variability in river habitat encountered in the Fraser River (see Figure 1) and because they had the most complete telemetry coverage out of all the reaches. Reach length, width and gradient were determined using 1:4,000 air photos; distances were measured from the center of the river. Reach 2, located immediately downstream of the Fraser River canyon, had an average gradient of 75 cm^km"1. Al l other reaches were in the canyon and had an average gradient of 120 cm«km"'. Reaches ranged in length from 200 to 973 m and in width from 117.5 to 192.5 m. Surface water speeds were estimated by timing the passage of floating wooden blocks (20 x 20 x 5 cm) over a known distance above the main thalweg of the river. Reach 2 has parallel banks, no constrictions, unidirectional centreline flows, and centreline surface water speeds of 147 cm»s"'. Reach 4 is constricted by an island, has two channels, multi-directional centreline flows, and centreline surface water speeds of 385 cm»s"'. Reach 7 has parallel banks with a large gravel bar creating a constriction, multidirectional centreline flows, and centreline surface water speeds of 246 cm»s''. Reach 9 is constricted by two large islands and 3 small ones, has several channels, multidirectional centreline flows, and centreline surface water speeds of 292 cm«s"'. Daily water temperature was measured by Fisheries and Oceans Canada at Hell's Gate, situated approximately 10 km upstream of the study area. Water temperatures were 13.9-15.4 °C during 31 the 1999 field season and 14.0-17.0 °C during the 1993 field season with mean temperatures of 15.7 and 14.9 °C, respectively. Daily discharge measurements were made by the Water Survey of Canada at Hope BC, situated approximately 20 km downstream of the study area. During the 1999 field season, the mean daily discharge was 2,758 mW 1 (range, 1,900-4,030 mS-s"1). During the 1993 field season, the mean daily discharge was 3,295 m V 1 (range, 3,170-3,770 mV 1 ) . Materials and methods Study animals and telemetry approaches The sockeye salmon are from the early Stuart stock, which spawn in the Stuart Lake system (Figure 1) and migrate through the study area in July. This is the longest distance migrating sockeye stock in the Fraser River system, traveling a distance of approximately 1,200 km upriver. In 1993, 9 sockeye salmon were studied (four males, mass 2.0-2.8 kg, fork length 57.9-61.7 cm; five females, mass 1.4-2.0 kg, fork length 50.7-57.2 cm). The pink salmon are from the Seton River stock, which spawn in the Seton River system (Figure 1) and migrate through the study area in September and October. This is the longest distance migrating pink stock in the Fraser system, traveling approximately 250 km upriver. In 1999, 12 pink salmon were studied (six males, mass 1.50- 2.25 kg, fork length 47.5- 56.6 cm; six females, mass 1.00-1.35 kg, fork length 47.5-50.6 cm). Fish were collected in the Fraser River canyon and EMG radio transmitters (Lotek Engineering Inc., Newmarket, Ontario) were implanted on site. One fish was released per day downstream of reach 2 and tracked upstream on foot using a hand-held directional three element Yagi antenna. EMG pulse interval data were recorded on a 3-5 second interval by hand-held radio receivers (Model SRX 400; Lotek Engineering Inc, Newmarket, Ontario). Fish positions could be ascertained to within 5 m. Details on fish collection, release and position-finding methodologies are given in Hinch and Rand (1998) and Chapter 2. A description of the EMG radio transmitters and details about their surgical implantation are outlined in Hinch et al. (1996). Hinch and Rand (1998) studied volitionally swimming adult sockeye containing EMG transmitters in the laboratory and showed that EMG pulse interval signals were strongly correlated to tailbeat frequency. They developed predictive relationships for sockeye salmon 32 between EMG pulse intervals and instantaneous swim speeds (in body lengths per second, BL«s" Chapter two conducted similar trials on adult pink salmon and developed relationships for the prediction of instantaneous swim speeds from EMG data. These two sets of predictive relationships were used to estimate swimming speeds associated with EMG pulse intervals recorded in the field from sockeye and pink salmon. Behavioural observations Travel time through a reach was determined for each fish by subtracting time of reach first exit from time of reach first entry. Travel rate was calculated (also termed ground speed) by dividing reach length by travel time. To explore swimming patterns, the mean and coefficient of variation (CV) for swimming speeds were calculated for each individual at each reach. The proportion of total within-reach time that each individual spent swimming under sustained speeds, prolonged speeds, and burst speeds were also calculated. The three swimming speed categories are defined from laboratory results of swimming performance and fatigue trials and are described below. Although each category encompasses a broad range of swimming speeds, they provide a convenient and well-understood convention for describing swimming patterns. The number of bursts per second (i.e., rate of burst swimming) that were elicited by each individual at each reach were also calculated because burst swimming could be a very important behaviour enabling migrating salmon to pass through relatively fast currents. Defined as speeds which can be maintained without fatigue for 200 minutes or more (Beamish 1978), sustained swimming is the slowest class of speeds but may be the most common ones elicited by fish during non-migratory phases of life (Beamish 1978). Prolonged swimming reflects the fastest class of speeds that can be elicited while still performing predominantly under aerobic metabolism, although high prolonged speeds are not usually attained without some component of anaerobic metabolism (Burgetz et al. 1998). Prolonged speeds may be the ones most commonly elicited during migration (Beamish 1978). Burst swimming is the fastest class, representing speeds generated by swimming entirely under anaerobic metabolism. These speeds are often defined as those that cannot be continuously maintained for more than 20-60 seconds (Beamish 1978). Based on swimming performance and fatigue information provided in Brett (1967), Brett and Glass (1973), and Beamish (1978) for adult sockeye, sustained speeds were classified as those under 2.5 BL»s"', prolonged speeds as 2.5-3.2 BL»s"', and burst speeds as 33 greater than 3.2 BL«s" . Detailed swimming performance experiments have not been performed on adult pink salmon. Williams and Brett (1987) determined that the critical swimming speed (the Ucrit) for adult pink salmon was very similar to that for adult sockeye salmon. Therefore, the same swimming criteria was assumed for both species. We quantified three characteristics of the pathways used by individual fish as they migrated through each reach and examined these values for strong and consistent trends between species and sexes. First, because fish tended to migrate near riverbanks, the percentage of fish exiting a reach along a different bank than was entered (termed 'bank infidelity') was determined. Banks associated with islands (e.g. in reaches 4 and 9), were also considered. Second, the number of times within a reach that individuals crossed.from one bank to another (termed 'crossings') was counted. Third, the number of times individuals backtracked to a previous downstream position (termed 'backtracking') was counted. Backtracking involved relocation downstream directly in a straight-line path, or indirectly in a circular path that may have resulted from a failed crossing. Statistical analyses Split-plot repeated-measures ANOVAs (SAS 1988) were used to assess how variability in each salmon swimming behaviour (ground speed, mean swimming speed, CV of swimming speed, proportion of time in sustained swimming, proportion of time in prolonged swimming, proportion of time in burst swimming and, number of burst swimming speeds) was accounted for by between-sex, between-species and among-reach variation. For each analysis, individual fish was the repeated measure and had 17 degrees of freedom (df). As type III sums of squares were used for calculating F-statistics, and individuals were nested both within a sex and a species, there was inadequate degrees of freedom to consider sex and species as separate main effects. Therefore it was necessary to consider sex and species as one variable, but with four separate classes (male pink, female pink, male sockeye, and female sockeye); henceforth called sex-species class (this main effect had 3 df in each analysis). Reach was the other main effect (3 df in each analysis). This ANOVA design enabled us to test for interactions between reach and sex-species class (9 df in each analysis). Statistical significance was assessed at the 0.05 level. Least square means (LSM;SAS 1988) were used to assess, a posteriori, differences among the levels within the class variables. Bonferroni's method was used to adjust the a posteriori significance levels when making multiple simultaneous contrasts of the LSMs. To minimize error variance heteroscedasticity variables were log (x+1) transformed (or arcsine square root in the case of 34 variables that were proportions) prior to conducting the ANOVAs. For sake of visual clarity and to prevent making a posteriori back-transformations, all LSM values presented in the text will be based on non-transformed variables. We generated Pearson correlation's between fish mass and each swimming behaviour, within-species and -reach to assess the role that fish size may have in contributing to observed swimming patterns. Fish mass was unable to be incorporated as an additional variable into the split-plot ANOVAs because of inadequate degrees of freedom. Statistical significance was assessed at the 0.05 level. Results Statistical analyses . The split-plot repeated-measures ANOVA accounted for 64% (P=0.0258) of the variation in ground speed. Neither individual fish, sex-species class nor the interaction of reach and sex-species accounted for variation (P>0.26 for each). Only reach significantly explained variation in ground speed (P=0.0004). Specifically, ground speed through reach 2 (LSM 44.64 cm»s"') was nearly three times faster (P<0.003 for all contrasts) than through the other reaches. The other reaches did not differ in ground speed (for each P>0.63; LSMs: reach 4 17.83 crn«s"', reach 7 17.43 cra»s"', reach 9 15.55 cm»s"'). The split-plot repeated-measures ANOVA accounted for 97% (PO.0001) of the variation in mean swimming speed. Individual fish and reach both accounted for significant variation in mean swim speed (P<0.0001 for each). Speeds through reach 9 were slower than through the other reaches (P<0.002 for each contrast), and speeds through reach 4 were faster than through the other reaches (P<0.006; Figure 4). Speeds through reaches 2 and 7 did not differ (P=0.2007). Sex-species class was not significant (P=0.1863). The interaction between reach and sex-species class was significant (P=0.0361) and was caused by a relatively low average swim speed by female sockeye at reach 9 (see Figure 4). The split-plot repeated-measures ANOVA accounted for 94% (P<0.0001) of the variation in swimming speed CV. Individual fish, reach, and sex-species class accounted for significant 35 I 1 43 O cd •18) ps^ds SUTUIUIIAVS iresjAT c3 JO >> u H 5) * . 03 u "O N u u B £ IS •5 CO O 03 09 (D •a JJ £ B w g II 5 ^ 0 o 6 C3 l e S a <a o Q cj 5 8 PS ^ c o « to Vi Vi B b c3 5 1 2 . S P 0> r-1 (U w I I i i S3 C c3 . ^ "* JJ B a 3 amounts of variation (P<0.0001 for each). Fish passing through reach 9 had a larger CV than when they passed through the other three reaches (P<0.008 for each of the three contrasts, Figure 5). Fish passing through reach 2 had a smaller CV than when they passed through the other reaches (P<0.01 for each of the three contrasts). CV did not differ for passage through reaches 4 and 7. There is a species effect on CV. Pink salmon males and females had lower values than sockeye salmon males and females (PO.0001 for each contrast; Figure 5). Female pink salmon had lower CV values than male pink salmon at each reach (PO.0001) and a similar among-reach trend was noted for female and male sockeye salmon (Figure 5), but was not statistically significant (P=0.20). There was no interaction between reach and sex-species class (P=0.4750). The split-plot analysis explained 91% (PO.0001) of the variation in the proportion of time spent swimming at sustained speeds. Individual fish (PO.0001) and reach (P=0.0004) accounted for significant amounts of variation. There was no significant effect of sex-species class (P=0.6648). Sustained speeds were elicited for higher proportions of time at reach 9 relative to reaches 4 and 7 (PO.0035 for both; Figure 6), and higher proportions of time at reach 2 relative to reach 4 (P=0.002; Figure 6). Reaches 2 and 9 did not differ (P=0.0640) nor did reaches 4 and 7 (P=0.1101). A weak interaction was detected between reach and sex-species class (P=0.0659) which arose because female sockeye salmon at reach 9 had a disproportionately elevated duration of sustained speeds (LSM 98% of total time). The split-plot analysis explained 86% (PO.0001) of the variation in the proportion of time spent swimming at prolonged speeds. As with sustained speeds, individual fish (PO.0001) and reach (P=0.0163) explained significant amounts of variation, and there was no effect of sex-species class (P=0.3963). Prolonged speeds were elicited for lower proportions of time at reach 9 relative to reaches 4 and 7 (PO.0027 for both; Figure 6). There were no differences among reaches 2, 4 and 7 (P>0.0650). A weak interaction was detected between reach and sex-species class (P=0.0561) which arose because female pink salmon at reach 4 had a disproportionately elevated duration of prolonged speeds (LSM 45% of total time). The split-plot analysis explained 83% (PO.0001) of the variation in the proportion of time spent swimming at burst speeds. As with sustained and prolonged speeds, individual fish (PO.0001) 37 as O CN d o £ to CO •&.s oo CO CO d o X a 2 £ x "3 Z £ ^ ,<P co p39dS SUTUIUIIAVS UT UOTJBTJBA JO 1U9I0TIJ300 UB9]AT ON m ON O a CD 4 = o CD o CD CN 4 = CD CD o o in o in CN d o o 9UI IX p j o x jo uopjodojj co CU _ > W U —J co _ u d H > -d ^ • d o CD u - c2 45 cu c *- o rt (50 <2 & t s co CU _ (50 CO d 55 cu rt CO rt M CO _ CU <u <-> cu i « >^  D, Q jD ep •2 I 8 - -3 -d o d o n,_> rt cu o o d g £ * d bp and reach (P=0.0410) explained significant amounts of variation, and there was no effect of sex-species class (P=0.2359). Burst speeds were elicited for higher proportions of time at reach 4 relative to reach 2. There were no differences among any other combinations of reaches (P>0.0583 for each contrast). There was no interaction between sex-species class and reach (P=0.2822). The split-plot analysis explained 85% (P<0.0001) of the variation in number of bursts elicited per second. Individual fish (PO.0001) and reach (P=0.0113) explained significant amounts of variation. There was no effect of sex-species class (P=0.1487). Burst rates were twice as high at reaches 4 and 7 (both LSMs 0.011 bursts«s"') relative to that at reach 2 (LSM 0.005 bursters"1; PO.0054 for both contrasts). Burst rates at reach 9 (LSM 0.008 bursts»s"') did not differ from that at the other reaches (P<0.2283 for all contrasts). A significant interaction between sex-species class and reach (P=0.0220) was caused primarily by male sockeye salmon having relatively high burst rates at reaches 4, 7 and 9 (LSMs 0.020, 0.027 and 0.021 bursts»s_1, respectively) compared to all other sex-species and reach combinations (LSMs range from 0.001 to 0.010 bursts»s"'). Within-species, data was pooled from both sexes and Pearson correlations were calculated for each reach, between fish mass and each of the seven transformed swimming behaviour variables. Positive correlations existed for pink salmon at reaches 2 and 4 between mass and proportion of time spent burst swimming (r=0.7388, n=12, P-0.0061; r=0.6521, n=l 1, P=0.0297, respectively) and between mass and burst swimming rate (r=0.8469, n=12, P=0.0005; r=0.6215, n=l 1, P=0.0412). No other significant correlations were found for pink salmon and none for sockeye salmon (P>0.05 in each case). Qualitative description of migration path behaviours and swimming strategies Of the 38 times that pink salmon individuals entered the study reaches, 84% were along north and west banks. In contrast, of the 29 times that sockeye salmon entered the reaches, 83% were along south and east banks. Al l study animals were released downstream of reach 1 (Figure 1) at the identical locale on the north-west riverbank. At each reach, sockeye salmon consistently exhibited higher levels of bank infidelity than pink salmon (sockeye salmon: among-reach mean 39.5% and range 13-50%; pink salmon: among-reach mean 21.8% and range 0-33%). Sockeye salmon tended to cross reaches (crossings per individual) more often than pink salmon (sockeye 40 salmon: among-reach mean 1.14 and range 0.50-1.75; pink salmon: among-reach mean 0.75 and range 0.20-1.20. There were no clear between-species differences in regards to number of backtrackings per individual (sockeye salmon: among-reach mean 0.77 and range 0-1.75; pink salmon: among-reach mean 0.84 and range 0-1.44). To help illustrate the general differences between species that were observed in swimming patterns (Figure 4) and migration path characteristics, two individuals were selected (a male sockeye salmon and a female pink salmon) for a detailed reach-specific presentation of travel paths (Figures 7a and 8a) and of temporal patterns of instantaneous swim speeds (Figures 7b and 8b). Figure 7 illustrates results for a typical sockeye salmon. This male displayed backtracking and crossing in reaches 7 and 9 (Figure 7a). Bouts of burst swimming were elicited in all reaches but were prevalent in reaches 7 and 9, and through the first half of reach 4 (Figure 7b). Bursts were frequent in these reaches but usually brief in duration, lasting only a few seconds at a time. Burst speeds were relatively high (often 6 to 9 BL«s_1) and were usually immediately followed by periods of low levels of sustained speeds (often 0 to 1 BL»s_1). Periods of sustained and prolonged swimming did occur in parts of reaches 2 and 4. However, never more than 3 minutes passed without burst swimming interrupting these slower swimming speeds. Prolonged speeds were not consistently elicited and seemed to result primarily as a transition between burst and sustained speeds. The swimming patterns of sockeye salmon were termed a "burst-coast and burst-sustained" strategy. Figure 8 demonstrates results for a typical pink salmon. This individual exhibited no backtracking and few crossings (Figure 8a). As with sockeye, burst swimming was elicited in all reaches, however unlike sockeye, burst speeds never exceeded 4 BL«s"' (Figure 8b). As with sockeye, speeds oscillated among sustained, prolonged and burst levels but in contrast to sockeye, most speeds were within a relatively narrow range (1 to 3 BL»s_1) and only rarely were very slow speeds of 0 to 0.5 BL«s"' elicited. The swimming patterns of pink salmon were termed a "burst-prolonged and sustained-prolonged" strategy. Discussion This study details the swimming behaviours of two species of salmon migrating through the same reaches in the Fraser River. In this study mean swimming speeds did not differ between 41 pink and sockeye salmon, nor did they differ between sexes within species. Further, groundspeeds did not differ between sexes and species. One might conclude from this that migration swimming strategies are relatively conservative. However, the level of variability in swimming speeds, the patterns of this variability, and a consideration of migration pathways, revealed clear differences in swimming behaviours and strategies between species. Relative to pink salmon, sockeye salmon had highly variable swimming speeds and changed migration paths frequently. It is suspected that sockeye explore alternative migration paths more frequently than pink salmon and in the process, encounter much higher variability in head currents. Using hydroacoustics at one locale in the lower Fraser River, Xie et al., (1997) found that sockeye salmon tended to migrate farther from shore in deeper water and in much looser aggregations than pink salmon, the latter supporting the notion that sockeye may utilize more diverse migration paths. Adult sockeye salmon, being larger than adult pink salmon (in this study 58% heavier, 13% longer), should be able to generate more power for a given tailbeat and thus could make forward progress against faster and more diverse currents than pink salmon. Thus, one might expect sockeye to roam more freely. However, the role of size is not clear. Sockeye females are the same size as pink males yet their respective swimming patterns are different based on the CV analysis. The two species were studied in different years, yet the differences in temperature and discharge between study periods were quite small, particularly in light of the extensive natural among-year variability that exists in hydrologic conditions in the Fraser River (Macdonald et al. 2000). Thus, it is likely that river conditions did not play a strong role in causing the observed interspecific differences. River reach was also an important determinant of swimming patterns. The four reaches spanned a gradient in river hydraulic features from reach 2 with no constrictions, a single channel and unidirectional currents to reach 9 with multiple constrictions, several distinct channels and multidirectional currents. Reaches 4 and 7 were intermediate in their hydraulic character, both had constrictions and multidirectional currents, but the former had two channels versus the latter which had only one. A general relationship between hydraulic complexity and swimming speed variation existed in this study. The CV was lowest at reach 2, intermediate for reaches 4 and 7, and highest at reach 9. It was also observed that, compared to reach 2, the number of river crossings were 2-3 times higher at reaches 4, 7 and 9. High levels of crossings and increased swimming speed variability at sites of multidirectional currents and multiple channels suggests that individuals had difficulty locating appropriate migration paths, possibly because these sites 46 generated contusing migration cues (Hinch et al. 1996; Hinch and Bratty 2000). The reach with the fastest surface currents (reach 4) required the highest mean swimming speeds, and the highest proportion of time swimming at burst level, for passage. Interestingly, reach 4 also had the lowest bank infidelity out of all reaches. Taken together, these results suggest that salmon can readily locate migration paths in fast water, but elevated swimming speeds are required to overcome the encountered currents. Females allocate significantly more energy to gonad development during migration than do males (reviewed in Brett 1995). As energy reserves to complete migration are limited and females are relatively smaller than males, it has been suggested that energy-saving swimming behaviours should be more strongly selected for in females than in males (Hinch and Rand 1998). This should be especially true for long distance migrating stocks (Bernatchez and Dodson 1987) such as those in this study. Although the study found no general differences between male and female swimming behaviours at all reaches, it was evident that at the reaches with complex hydraulics where passage is presumably most difficult, females exhibited some energy conserving behaviours. For instance, at reach 9, female sockeye salmon exhibited disproportionately high levels of sustained swimming and as a result, very low average swimming speeds. At reach 4, female pink salmon exhibited extremely high levels of prolonged swimming. Females of both species exhibited low burst rates, whereas sockeye salmon males exhibited disproportionately high burst rates (2-10 times higher) at the reaches with complex hydraulics. Bioenergetics modeling has shown that these types of elevated rates of swimming speeds by male sockeye make them energetically less efficient migrators than females (Hinch and Rand 1998). Consistently, individual fish explained significant amounts of variation in swimming speed behaviours. Was variation in fish size a contributing factor? Based on body constituent analyses of upriver migrating Atlantic salmon (Salmo salar, Jonsson et al. 1997) and American shad (Alosa sapidissima, Leonard and McCormick 1999), within the same sex, large fish used more energy than small fish. At reaches 2 and 7, large pink salmon exhibited higher proportions of burst swimming and higher burst rates, and like the examples above, presumably had higher energetic costs than small pink salmon. However, no other size-related relationships were detected. Thus the results provide only weak evidence to support a size-based explanation for observed levels of individual variability. Regardless of weight, sex or species, there could also 47 be a genetic propensity for certain swimming behaviours in some individuals. Hinch and Bratty (2000) could find no clear explanation for why some individual upstream sockeye migrants elicited hyperactive swimming patterns. The causes of individual variability in swimming patterns require much further study. In conclusion, sockeye salmon and pink salmon "solve" their migration challenges by swimming' with very different strategies. Although mean swimming speeds were not different, sockeye salmon were much more variable in swimming speeds than pink salmon, a point that was clearly illustrated in Figures 7 and 8 where the sockeye swimming pattern was identified as "burst-coast and burst-sustained" and the pink swimming pattern as "burst-prolonged and sustained-prolonged". Interestingly, there were no broad differences between species in terms of proportions of total time spent at sustained, prolonged or burst speeds, implying that both species (and sexes) used aerobic and anaerobic metabolic pathways approximately the same amounts of time. As the criteria defining these swimming speed categories are developed in confined laboratory respirometer tunnels under steady swimming conditions, conclusions about the actual amounts of aerobic versus anaerobic costs in field studies such as ours must be made cautiously until such criteria can be assessed under more natural swimming conditions. 48 Chapter 4: Examining the effects of hydraulic characteristics on energy use in upriver migrating salmon using underwater stereo videography Abstract The natural hydraulic regime and flow patterns in many rivers in British Columbia have been altered by human-made structures. These alterations change the migration environment of adult salmon and could decrease fish passage success. Underwater videography was used to observe sockeye salmon (Oncorhynchus nerka) during their spawning migration through a range of natural velocity flow fields in the Seton River, British Columbia. The main objectives were to determine if the velocities fish encounter during migration are similar to the average velocites in the migration area, as well as if individual swimming and ground speeds differ among different hydraulic velocity flow fields. On average, encountered velocity (44 cm«s"') was less than the average site velocity (53 cra«s"') indicating that fish were selecting low velocity flow fields during migration. Swimming speed and ground speed were similar in low encountered velocity (<0.13 m«s"') sites with a ratio of 1:1 indicating that fish were moving as if they were in still water. In mid to high encountered velocity (0.35-1.27 rn^s"1) sites the ratio of swimming speed to ground speed remained constant at 2:1 but both speeds increased with increasing encountered velocity. An optimal migration index (theoretical minimum energy expenditure divided by observed energy expenditure) indicated that fish migrated optimally (OMI values ranged from 0.80-1.15, with 1 being optimal, <1 being sub-optimal and >1 superoptimal) through low and moderate encountered velocity sites (0.05-0.67 m«s"'). The three sites with low OMI values were the shallowest site (OMI=0.61), the highest velocity site (OMI=0.22), and the lowest velocity site (OMI=0.34). This paper proposes that, in the majority of sites fish are saving energy by using low velocity flow fields to optimize migration. 49 Introduction Over the past hundred years, many rivers and streams in British Columbia have been altered by hydroelectric installation, irrigation weirs, and flood and diversion dams. These alterations often have changed the natural hydraulic regime that has affected the migration environment of salmonjds and other species. (Baxter 1977; Fretwell 1989; Quinn and Adams 1996). Because most species of salmon stop feeding upon entering freshwater, they have limited energy stores to complete their upriver migrations. Understanding the impact of changing hydraulic regimes is therefore important in order to manage water flow and ensure successful fish passage. Studies into fish biomechanics state that fish locomotion is determined by the body muscle activity interaction with surrounding water (Weihs 1973a; Videler 1993; Vogel 1994; Altringham and Ellerby 1999; Pedley and Hill 1999; and others). Much like migratory birds, who use flying patterns to reduce wind resistance, fish may use the physical properties of water to increase their efficiency. Sea bass appear to school to increase efficiency, fish swimming at the back of the school using 9-23% less oxygen swimming compared with the fish at the front (Herskin and Steffensen 1998). Differences in body shape, size and rigidity have also been proven to affect the drag coefficient or hydrodynamic potential of fish. Thus different fish experience different costs for a given swimming style (Muller et al. 2000; Sagnes et al. 2000). Detailed laboratory biomechanics studies not only suggest that different fish have different forces exerted on them by their environment but that animals use subtle body movements and environmental conditions to facilitate locomotion (Wardle et al. 1995; Muller et al. 2001). Differences in swimming style and shape among fish have been proven in lab to affect the drag coefficient and hydrodynamic ability of fish (Gray 1936; Lighthill 1969; Lighthill 1975; Herskin and Steffensen 1998; Drucker and Lauder 2000; Sagnes et al. 2000) Passage studies have also shown that hydraulic features associated with dams and fishways affect the ability of adult fish to move upstream. In addition, passage success at these facilities appears to differ for different fish species. Hydroacoustic sampling in the Fraser River determined that pink salmon migrate near to shore in close proximity to one another versus sockeye salmon which migrate further offshore in a more dispersed arrangement (Xie et al. 1997). Electromyogram (EMG) radio telemetry studies in the same system have shown that pink salmon tend to use less energy than sockeye salmon (Chapter 2). Sockeye salmon appear to cross the main current of the river more 50 frequently than pink salmon during migration (Chapter 3). These behavioural path choices and swimming strategies may determine not only fish passage success, but also the amount of energy they require to pass through different hydraulic conditions (Hinch and Bratty 2000; Chapter 2). Increasing our understanding of fish migratory behaviour and energy use across a range of hydraulic conditions may improve management of river flows ensuring hydraulic conditions appropriate to facilitate salmon migration. Studies examining salmon passage success and upriver migration have been conducted across very different spatial scales. Studies examining fish biomechanics combined with in situ telemetry results have made it clear that during migration fish are experiencing velocity environments on a much smaller scale than we are capable of describing with EMG technology. To better our understanding of among reach differences in energy use, we must turn to a technology that allows us to more closely examine the interaction of the migrant with its environment (Weihs and Webb 1983; Videler 1993). Underwater stereo videography can be used to assess small-scale flow fields and determine fish location and activity level within a 1-5 m" field of view (Boisclair 1992; Hughes and Kelly 1996; Trudel and Boisclair 1996; Tang et al. 2000). Hinch and Rand (2000) applied the three dimensional video methodology developed in Hughes and Kelly (1996), and assessed the in situ swimming strategies and migration energetics of three long distance migrating stocks of sockeye salmon and were able to demonstrate that migrants indeed were affected by, and responded to, very small flow fields. There has been considerable research into understanding energetic efficiency and optimality in many migrating fish species including several salmon stocks (Brett 1973; Weihs 1973b; Weihs 1974a; Brett 1983; Weihs and Webb 1983; Bernatchez and Dodson 1987; Beauchamp et al. 1989; Brett 1995; Webb 1995; Hinch and Rand 1998; Hinch and Rand 2000). For the purpose of this paper efficiency is measured as the relationship between swimming speed and ground speed and is relative rather than absolute. For instance, a fish is migrating relatively efficiently if its ground speed to swimming speed ratio is relatively high. Optimization, however, involves 'adjusting swimming speeds up or down so that minimum amounts of energy are used per metre of forward distance traveled. In other terms, optimization is the theoretical minimum amount of energy a fish should use under particular conditions divided by the observed energy usage. Bernatchez and Dodson (1987) determined that long distance migrators were more efficient than short distance ones. They made two assumptions in their estimation of efficiency, both of which 51 may have led to an over estimation of energy use during migration. First, they assumed fish migrate at mean river velocities. Hydroacoustic surveys (Xie et al. 1997), as well as visual and telemetry observations, (Ellis 1966a; Fretwell 1981; Chapter 3) have shown that fish migrate near the bottom or banks of river systems where water velocities are often less than the mean river velocity (Henderson 1966). Second, they assumed fish swim at a steady state, without altering speed over time. EMG studies and videography have shown that fish adopt a more dynamic swimming strategy, altering their swimming speed depending on the type of reach they move through, ultimately reducing drag (Weihs 1974a; Hinch and Bratty 2000; Chapter 3). A fish, when coasting, has 3-5 times less drag than during a powered swimming burst (Weihs 1973b; Weihs 1974b; Lighthill 1975; Webb 1995). Weih's (1974a) predicts that burst coast swimming can reduce transport costs up to 60%, resulting in increased swimming efficiency. Hinch and Rand (2000) determined that long distance migrating sockeye salmon swam according to an optimal swimming speed model at low encountered velocities. However, most stocks became less efficient in high velocity sites. They tested their assumptions by applying their optimal migration index (OMI) to ocean migration data taken from (Madison et al. 1972) as well as a short distance migrating stock studied by Ellis (1966b) and found in both cases that fish migrate at optimal speeds. The general applicability of their model and of optimal swimming speed theory to other stocks needs to be assessed. This study assesses the effects of small-scale hydraulic features on the migration tactics and energy use in Gates Creek sockeye salmon, a middle distance migrating stock (relative to those reported in Hinch and Rand 2000). Underwater stereovideography was used to observe the upriver migration but the flow description and camera calibration methodology was altered from Hinch and Rand (2000) in order to increase the precision and accuracy of measuring flow velocities and fish movement within the field of view. In addition, a wider range of sites was included in the video analysis to ensure data collection over a large range of encountered velocities. This study had two main objectives. The first was to determine whether migrants use low velocity flow fields to decrease energy use during migration. The second was to determine whether fish minimize transport costs by swimming at metabolically optimal speeds, and whether these activities depend upon encountered velocity. 52 Methods Site description and selection I studied a Fraser River sockeye stock that migrate 400 km to their spawning channels in a tributary of the Fraser River (Figure 9). They gain 300 m in elevation from sea level during this migration. This stock of fish was chosen for video observation because it was relatively abundant, accessible and provided a contrast to the long distance migrating stocks studied in Hinch and Rand (2000). Ten sites were chosen for video observations based on velocity, accessibility, the presence of migrating fish, and water clarity. Sites were selected to represent a wide range in velocities. Video System We used four black and white charged couple device (CCD) cameras. Two were Panasonic WV-BP312 with 570 lines of horizontal resolution and two were Cohu Model 2100 with 570 lines of horizontal resolution and were fitted with aspherical high-speed aperture lens (Panasonic WV-LA408C3; 4.5-mm focal length and Cosmicar/Pentax 3.7-mm focal length, respectively). Each pair of cameras was oriented in parallel (optical axis separation of 11.3 and 8.2 cm, respectively) and fixed to a stand located 1-2 m away from the desired field of view. Each camera was enclosed in a cylindrical anodized aluminum housing fitted with flat Plexiglas port. Video signals were sent to a multiplexer-quad unit (Panasonic WJ-420) and then to a VHS time-lapse recorder (Panasonic AG-6124) where images from all cameras were recorded simultaneously on the same video frame at 60 frames*s"'. Thus the recorded video image was divided into four independent images, one for each camera. A black and white LCD monitor (OmniVision Inc.; 1024x768 pixels) was used to assist with focus, calibration, camera orientation, and determination of water and fish speeds. Al l components of the system were powered by two 12-V deep cycle batteries. At all sites, the cameras were set up underwater facing the center of the river, with their optical axis parallel to the surface of the water. Cameras were able to clearly monitor between five to ten metres into the river. However, data were only extracted from video images within a few metres of the cameras (the reasoning for this will be explained later in the methodology). Fish movement appeared to be concentrated along the interface of substrate and water with the 53 114 * British Columbia 60" 200 km Figure 9. Map of the study area indicating the video sites. Inset map of British Columbia, Canada indicating location of the Fraser River and the Seton River. Sites SDI 2 and SD34 are located off of the map at the downstream end of Seton Lake at the start of the lower Seton River. 54 majority of fish moving in distinct paths along the banks. Hinch and Rand (2000) termed this . shoreline migration as 'line swimming' and similar behaviour has been described by visual observations (Ellis 1966a) and hydroacoustic sampling (Xie et al. 1997; Macdonald et al. 2000). As a result, the cameras were situated such that they recorded the majority of the fish as they migrated through the system. Camera calibration and data collection Cameras were calibrated using a 0.3 x 0.5 x 0.5 m aluminum cube. The cube resembled a cage with horizontal and vertical aluminum bars (each 1 cm diameter) that crossed each other at exactly 10 cm intervals. Each pair of cameras was set up in a determined location and secured so they were immobile during the observations. The cube was placed in the field of view of each pair of cameras to ensure that at least nine nodes (e.g., aluminum bar cross-joints) from both the near and far face of the cube were visible in both cameras. In order to reduce calibration error the cube was placed within the path of the majority of fish as seen from shore. After the cube's image was recorded, it was removed from the water. In the laboratory, video images were replayed and a single frame, clear video image of the cube was digitized. The nodes of the near and far face were digitized for each camera in order to get two sets of x-y coordinates of the cube. Because the video image was divided into four quadrants on the screen, the x-y coordinates of each camera view had to be transformed so that the lower left corner of each quadrant (or camera view) was set to 0,0. The x-y cube coordinates were then fed through a Java Script program (Kelly 1999), which created parameter coordinates from the individual near and far face coordinates. These parameter coordinates, along with the transformed x-y digital coordinates, were then processed through a series of Excel macros (Kelly 1999), which transformed the data into three-dimensional x-y-z coordinates complete with an error estimate. Detailed information regarding the Java Script program as well as the Excel macros can be found in Kelly (1999). Water velocity within the field of view was measured using a CMC-2 current metre (Hydrological Services Ltd. Australia). Measurements were taken in a three-dimensional grid (3 x 3 x 3 arrangement) within the field of view of the camera. Depth measurements were recorded at each velocity measurement (on the 3x3 grid). The propeller of the current metre was videotaped and digitized in order to determine the three dimensional location of the measured 55 flow points. These flow points, as well as their location, were then subjected to an interpolation program within Matlab (Version 5.3.0.14912a(Rl 1), The Mathworks, Inc.) which used distance-based averaging with a smoothing factor to predict water velocities between measured points (Figure 10). In this manner, encountered velocities along the fish path through the video image could be predicted. Water velocity and depth measures were used to determine the Froude number (Fr=(v2)»(g»d)~l; where v=water velocity, ^ gravitational constant 9.81 m«s"2, and J=depth) of each site. This number is a dimensionless hydraulic index that describes the hydraulic condition in open channel flow (Henderson 1966). In a hydraulic sense the Fr number is the velocity of the water at a given site divided by the wave, speed at that site (the speed at which wave would propogate if a pebble was tossed into the river). At large Fr numbers the water is supercritical, meaning water is shallow but moving fast, at low Fr numbers the water is subcritical, meaning deep and slow. The transition between sub and supercritical flow occurs at a Fr number equal to one. I will use this index in analysis to understand how fish behaviour is affected by hydraulic conditions. Fish were selected from the video image based on their clarity in the video image and their distance from the camera calibration cube. To decrease error estimates, fish which were greater than 50 cm from the nearest cube or velocity measurement point were excluded from analysis. Cameras recorded fish for two to eight hours depending on the frequency of fish passage at a particular site. At least 30 fish at each site were recorded, twenty of which were later digitized for analysis. Al l digitized measurements were converted to real world three-dimensional points using the Excel macros file (Kelly 1999). Fish dimensions were recorded by digitizing rostrum and tail locations as well as dorsal ventral points while the fish was fully extended in the field of view. Fish path was digitized based on nose location at the point of first entry and first exit as well as every 15/60ths of a second in-between. Ground speed for each fish was determined by measuring distance traveled in number of video frames (each frame measured l/60 t h of a second). Swimming speed for each fish was determined by adding encountered water velocity at each rostrum point to their calculated ground speed. This paper uses the Swimming Efficiency Index (SEI) as well as the Optimal Migration Index, (OMI) both developed by Hinch and Rand (2000), to compare energetic efficiency and energetic 56 Figure 10. Interpolated site velocity gradients. Each cube represents an individual site with site names listed below. Sites are varying sizes based on the topography of each site and field of view of each camera. Each square within each cube represents ten centimeters, also denoted by black bar in lower right corner of each site cube. Velocities ( c m » s _ 1 ) are noted in the legend to the right, negative values indicating flows moving in the upstream direction. 57 optimization among reaches. Based on a bioenergetics model designed by Webb (199.5) and further developed by Beauchamp et al. (1989), the theoretical minimum energy expenditure can be determined for a given encountered velocity (Hinch and Rand 2000). Using the same bioenergetics model, energy use can be estimated for each fish based on known swimming and ground speeds. The Optimal Migration Index is determined by dividing the theoretical minimum energy expenditure with the observed energy expenditure. Therefore an OMI value of 1 indicates optimal migration, a value less than one indicates suboptimal swimming behaviours. In contrast, a value of greater than one indicate superoptimal, or highly effective swimming behaviours. The SEI is simply the ratio of ground speed to swimming speed of a migrating fish and provides an index of activity per unit distance for each site. Thus SEI is defined in this chapter as efficiency. These types of indices allow a comparison of migration efficiency and optimality in varying encountered velocities and site conditions that may provide insight into how different hydraulic conditions affect fish migration. Data analysis To identify site differences, velocities between sites were compared using a Tukey's Studentized Range Test. We used analysis of variance (ANOVA) and least squared means, as the a posteriori method, to compare ground speed, swimming speed, tailbeat frequency and SEI among sites. Bonferroni's correction was made to control for Type I errors when making multiple comparisons (adjusted value oc=0.05/10). If error variance heteroscedasticity was present in the raw data, variables were log transformed prior to analysis. Paired t-tests were used to compare encountered water velocity of the fish with average water velocity at each site in order to examine if fish were selecting particular velocity fields for migration. In addition, Pearson's correlations were calculated between encountered water velocity and each of the fish variables (ground speed, swimming speed and SEI). Pearson's correlations were also calculated between site Froude number and each of the above fish variables with the addition of OMI. Results Statistical Analysis When compared with a Tukey's Studentized Range Test sites fell into three main groups depending on average site velocity. Sites SDI2, SD34 and B12 were the lowest velocity sites, 58 sites E12, E34, 012, D212, J34 and J12 were middle velocity sites and site V34 was the highest velocity site. Al l differences were based on a P<0.05 and are represented by different letters in Figure 11. Ground speed ranged among sites and fish from 24.4 to 341.9 cm«s"' (overall mean 82.3 cm»s"', n=150, SE=4.12; site averages shown in Figure 12). Ground speed was log transformed prior to ANOVA. ANOVA revealed that ground speed differed among sites (F9;i40=37.13, P=0.0001). In pairwise comparisons, there were no differences between sites E12, J34, B12 and J12 (P>0.0314 for all contrasts), sites J34, B12, J12, 012 and D212 (P>0.0131 for all contrasts), sites B12, J12, 012, D212 and SD34 (P>0.0333 for all contrasts), or sites V34 and SDI2 (P=0.5749). Ground speeds were faster in sites SDI2 and V34 than all other sites (P<0.0001 for all contrasts) and . slower in sites E34 than all other sites (P<0.0015 for all contrasts). Swimming speed varied from 54.2 to 396 cm*s_1 (overall mean 129.3 cm»s"', n=150, SE=5.32; site averages shown in Figure 12). The one way ANOVA revealed that swimming speed varied among sites (F9ii4o=41.05, P=0.0001). In pairwise comparisons, there were no differences in swimming speed between sites E34, B12, SD34, E12 and J34 (P>0.0834 for all contrasts), sites B12, SD34, El2, J34 and J12 (P>0.0158 for all contrasts), or sites J12, D212, 012 and SDI2 (P=0.0076 for all contrasts). Swimming speeds were faster in site V34 than all other sites (P=0.0001 for all contrasts) and slower in sites E34 than J12, D212, 012, Sdl2 and V34 (P<0.0005 for all contrasts). Swimming efficiency index was log transformed prior to analysis. ANOVA revealed that SEI varied among sites (F9,i5o=l 12.44, P=0.0001; Figure 12). In pairwise comparisons, there were no differences in SEI between sites E34, V34, D212, El2, 012, J12 and J34 (P>0.0900 for all contrasts), or sites B12 and SD34 (P=0.0346). Site SD12 had a higher SEI value than all other sites (P<0.0001 for all contrasts) and sites B12 and SD34 had higher SEI values than all sites but SD12 (P<0.0006). Encountered water velocities between fish, ranged from 3.9 to 141 cm«s"' (mean 47cm»s"', n=150, SE=2.96) whereas site velocities ranged from —4 to 179 cm»s-1 (range calculated from 27 59 p £ - o3 C s s K 8 • I U ( U cn ( U 43 o 03 < U Ut a 0 3 e co >, 43 O 03 u 3 T 3 < U G CO 0 3 p C U H ~ I < U CO cO <u '3 £ a to X .5 43 co O. 3 to I U 43 -*-» ( U l-H 0 3 ( U bO 03 a, 43 3^ +3 iu ° § "S 'so CO & -*-> 03 ju «J DO .a Ul -£> — o « -2 43 0> H > - . ( U w .13 .13 CO OT U, 43 J2 O -I W i -N g 3 CO "O 0 3 » H 0 3 ( U I-I CO c 0 3 ( U u co <u tu X .ti 0 3 2 * 0 3 o I? "to1 'o 0 3 42 u > ^ I U CJO 0 3 t-c 5 S 0 3 Ul $3 cu t3 <u .5 co p »-, I H C U 43 3 * Ui to £ o3 2 * <u > CU CO (T-S«uio) pssds C U P C3 o 3 5 e « -g .. _ J +3 C U Q tu O 60 T3 ^C < U •3 N 43 • — ~ 3 •£ T3 -2 u A M a 0 3 co X U H C U -13 cu o « 4 3 4 3 8 <« ® s 3 3 C H co 3cu C U 2 u Ui C (T.s.uio) pssds L_< X V^ J. OH cn 6J) cn independent velocity measurements taken at each of 10 sites, negative values indicate flow fields moving in the opposite direction from the main current; site averages shown on Figure 11). At most sites, encountered water velocity was similar to or less than average site velocity. In particular, encountered water velocities were less than average site velocities at sites E34 (P=0.0276), J12 (P=0.0218), J34 (P=0.0004), SD12 (P=0.0072) and SD34 (P=0.0035). There were no differences between encountered velocity and site velocity for sites B12 (P=0.1115), D212 (P=0.1955), E12 (P=0.1327), 012 (P=0.6579), and V34 (P=0.2892). Encountered water velocity was correlated to ground speed (r=0.16, P=0.0546, n=150 fish and sites), swimming speed (r=0.64, P=0.0001, n=150 fish and sites) and SEI (r=-0.67, P=0.0001, n=150 fish and sites). Site OMI was correlated with average swimming speed (r=-0.78, P=0.0075, n=10 sites) and average ground speed (r=-0.90, P=0.0004, n=10. sites) but was not correlated with average SEI (r=-0.31, P=0.3760, n=10 sites) or average encountered velocity (r=-0.32, P=0.3708, n=10 sites). Site Froude number was correlated with average swimming speed (r=0.84, P=0.0022, n=10 sites) and average encountered velocity (r=0.91, P=0.0003, n=10 sites) but was not correlated with average ground speed (r=0.41, P=0.2388, n=10 sites) or average SEI (r=-0.46, P=0.1860, n=10 sites). Both the OMI values and Fr numbers for each site are listed on Figure 12. Discussion Gates Creek sockeye salmon appear to use very small-scale flow fields during upstream migration. Through most sites, encountered water velocities were less than the average site velocity. Fish only encountered velocities similar to the site average in shallow sites (<0.67m). This paper proposes that fish are taking advantage of small-scale flow fields and boundary layers to minimize their encountered velocity. One of the physical properties of fluids is termed the 'no-slip condition' (Vogel 1994), meaning that fluids are affected strongly by friction and do not slip when they come into contact with solid bodies. In a river context this means that where the water flow encounters obstacles a boundary layer develops where water velocities increase from zero (at the contact point of the solid obstruction) to the maximum velocity of the river flow. Based on this physical property, water hydraulics and patterns of flow such as eddies, vortecies and von karmen streets develop around in stream obstacles and increase the hydraulic complexity within the system (Henderson 1966; Sahin et al. 1994; Vogel 1994; Triantafyllou and Triantafyllou 1995; Beffa 1996; Afzalimehr and Anctil 1999). In deeper sites or sites with 62 greater hydraulic complexity, fish are able to swim in this boundary layer, experiencing lower current velocities than the main stream (Webb 1995; Hinch and Rand 2000). Within similar velocity shallow sites, uniform flow or few obstructions may cause this boundary layer to be too small for large fish to benefit from. As a result large fish experience the average site velocities, and require more energy for upriver migration (Webb 1995). Certainly further research into more exact hydraulic modeling is necessary to understand how fish use different inriver conditions. Migration efficiency is defined by the swimming efficiency index (SEI). The SEI is the ratio of ground speed to swimming speed and, as a result, is a behaviour estimate that relates to the amount of energy a fish uses per metre (Hinch and Rand 2000). With the exception of one site, swimming speed in low velocity sites was relatively constant andlow and all sites had SEI values of hear 1.0. It appears fish swim with little effort in velocities less than 0.25 cm»s ; l. This result is similar to that of Hinch and Rand (2000) with long distance migrants. This paper proposes, as did Hinch and Rand (2000), that fish use existing water vortices and flow fields to propel themselves forward with little exertion. In moderate and high velocity sites both, swimming and ground speed increased with increasing encountered velocity (Figure 12). However, in high velocity sites, the SEI remains fairly constant (roughly 0.5) indicating the relationship between swimming speed and ground speed to be constant (Figure 12). This suggests fish may increase swimming speed until a ratio of activity to forward progress is achieved (roughly 2:1). In this way, fish may increase energy use per metre to minimize travel delays through higher velocity areas (Hinch and Rand 2000). We also compared swimming speed and ground speeds with encountered velocity to predict OMI (Hinch and Rand 2000). As mentioned earlier, optimization involves 'adjusting' swimming speeds to achieve minimum energy expenditure per metre traveled. On theoretical grounds it has been predicted that in still water the optimum speed for a fish swimming is roughly 1 BL»s"'. At these speeds, energy demand for propulsion equals the fishes standard metabolism (J«s_1) (Weihs 1973b). Quinn (1988) has demonstrated that sockeye salmon that migrate in the ocean swam at 1 BL«s_1. If an 'optimal' activity level exists in still water one would assume the same 'optimization' would occur in some form in moving water. In fact, Hinch and Rand (2000) showed that under slow flow conditions (<0.25 m»s"'), sockeye do swim at metabolically optimal speeds. Within this study, sites with low velocity (<0.25m«s"') have OMI levels that are 63 slightly below optimal (Figure 12). This may be due to fish taking advantage of lower velocity sites to gain ground quickly at minimal expense. Fish did, however, have optimal or super optimal migration indices through moderate velocity sites of moderate depth (Figure 12). Again, it is predicted that fish achieve these indices by utilizing small-scale velocity fields and boundary layers that have slower velocities than the average encountered velocities of the site (Weihs and Webb 1983; Videler 1993; Webb 1995; Hinch and Rand 1998; Hinch and Rand 2000). In addition, it appears fish determine the appropriate swimming speed to ground speed ratio for a given encountered velocity in order to minimize energy use per unit distance. There were three sites where fish showed sub-optimal migration indices. Two of these sites were among the high velocity, shallow, depth sites (0.53-0.49m). Once again high velocities combined with shallow depth result hydrodynamically in an environment with a very steep velocity gradient within the boundary layer. The reduction in availability of boundary layers to large migratory fish limits their ability to optimize migration efficiency. Again, their response to this environment appears to be reflected in an increase in swimming and ground speed to potentially decrease the amount of time spent in expensive delays at high velocity sites (Chapter 2; Chapter 3; Hinch and Bratty 2000) Fish also show a particularly low OMI at the lowest velocity, site which is not expected and is contrary to findings in Hinch and Rand (2000). At this site ground speed and swimming speed are near to equal but, as mentioned above, fish swim at a much faster speed. Two explanations for this occurrence can be proposed. First, this site is the first near-zero velocity area just upstream of a long riffle. The fish, therefore, may be taking advantage of the sudden low velocity area to cover ground rapidly to make up for potential delays experienced downstream. Second, the site is located directly downstream of a large hydro spillway. After attempting to enter the fishway, fish were seen milling and circling back along this low velocity section of river. It may be that the fish are disoriented from the turbulent tailrace water and, as a result, are not migrating efficiently. These results suggest that specific hydraulic features may influence swimming efficiency and optimal migration behaviour. From our data it appears depth and encountered velocity may be important factors influencing fish migration. However, although this study measured flow fields 64 more accurately than in the past, further research using improved velocity profiling techniques would be helpful to more clearly define differences in site environments. In addition a project which includes a larger sample of sites where velocity, depth and substrate size were controlled for in a factorial design would provide more insight into potential causes of efficiency and optimality differences between sites. f 65 Chapter 5: Conclusions and Future Research Direction Salmon are an important biological and commercial species on the Pacific coast and the success of their upriver spawning migration is essential for ensuring species survival. At present, little is known about the effects of hydraulic conditions on migrating fish. If this knowledge is improved, management practices such as defining bank shape and water flow through rivers may benefit fish passage. The goal of this thesis was to develop an increased understanding of the upriver energy consumption and migration behaviour in sockeye salmon and pink salmon in relation to in situ hydraulic features and velocity fields. The fields of biomechanics, physics and physiology must be combined when studying energy use patterns of any type of animal behaviour. The density and three-dimensional complexity of the aquatic environment demands particular attention. Past studies in Pacific salmon migration have provided valuable information on energy use and, in some cases, behaviour (Idler and Clemens 1959; Gilhousen 1980; Fretwell 1981; Brett 1995; Webb 1995; Leonard and McCormick 1999; Hinch and Rand 2000). The more recent EMG telemetry literature examines both energy use and behaviour (Hinch et al. 1996; Hinch and Rand 1998; Geist et al. 2000; Hinch and Bratty 2000). This paper proposed that there is a critical gap in the spectrum of species studied, as well as the temporal-spatial scale of existing work. To help fill the gap, this study is the first to take a detailed look at in situ migratory energetics of pink salmon. The design of the project allowed for an inter-species comparison of new pink salmon data with existing sockeye salmon data (Hinch and Rand 1998). It was found that although differences between sex and species existed in particular situations, the primary determination of migration behaviour and energy use in both species was reach or river environment. These result support earlier studies (Hinch and Rand 1998; Hinch and Bratty 2000; Hinch and Rand 2000). The moderate scale measurements of the river reaches within the EMG telemetry study created, but could not address questions regarding how fish use particular hydraulic features during migration. To address these questions of hydraulics and migration, the project was expanded and underwater stereovideography was used to more accurately describe fish movement through defined hydraulic features. Studies in biomechanics and physiology suggest that size, shape and 66 swimming style of fish as well as the condition of the water surrounding the fish affect the animals thrust generation and drag coefficient (Gray 1936; Lighthill 1969; Lighthill 1975; Weihs and Webb 1983; Videler 1993; Vogel 1994; Herskin and Steffensen 1998; Altringham and Ellerby 1999; Muller et al. 2000; Sagnes et al. 2000; Muller et al. 2001). By incorporating some of the detailed observational techniques of biomechanical studies in situ, the underwater video project measured and interpolated flow velocities and fish movement on a smaller scale than has been done in the past. Furthermore it allowed for increased spatial-temporal resolution within the field of fish migration bioenergetics (Hinch et al. 1996; Hinch and Rand 1998; Geist et al. 2000; Hinch and Rand 2000). The initial results of this study suggest fish are in fact isolating low velocity flow fields on a very small-scale. Supporting work done by Hinch and Rand (2000) our data found fish to be highly efficient in low velocity sites (<0.25 cm«s"'). However, their efficiency decreased but remained constant at velocities equal to, or above 50 enns"1. Contrary to Hinch and Rand (2000), this work determined that fish migrated optimally in sites up to 60 cm»s" \ This suggests that middle distance fish as well as long distance migrants are capable of optimizing migration behaviour, contrary to the findings of Bernatchez and Dodson (1987). As a result continued work on middle distance and shorter distance migrating stocks as well as different species may help to clarify optimization pressures on migrating salmon. ' Both the EMG telemetry and the underwater videography projects have shown that hydraulic characteristics are important in determining energy use and behaviour in upriver migrating salmon. In order to further our understanding of fish migration energetics, one might adopt an interdisciplinary approach incorporating expertise from the fields of biomechanics, physics and physiology. In this way we might better understand how fish use river hydraulics and dynamic flow gradients. The limitations of the spatial scale of the EMG technology in situ points to future work using underwater video. Although the underwater video system has enabled fisheries scientists to describe flow fields and fish interactions on a scale smaller than ever before, even finer resolution is needed for a fuller understanding of these subjects. The first step for future studies might be to develop new methods for three-dimensional digitizing of video images and accurate measurement of velocity fields in situ. Once these technological issues are addressed, a combined research project using laboratory and in situ methods could be conducted. Controlled laboratory experiments examining fish movement through quantified naturally occurring hydraulic conditions would allow us to better understand 67 the locomotory effect of subtle body movements in fish. Once these movements are quantified in situ, hydraulic features using underwater video could be more accurately described in terms of fish locomotion. The development of this combined methodology could then be applied to any number of fish species. This thesis found that inriver hydraulic features determine energy use and behaviour in adult salmon during upriver migration. Practically this has purpose in addressing questions regarding the management offish passage. In disturbed systems we can use structural bed and bank modifications as well as flow discharge regimes to create hydraulic features such as eddies, vortices and reverse flow fields which benefit fish passage. In conclusion continued research to further define the relationship between hydraulic features and fish energy use is a worthy pursuit, not only in a management context but in the interest of science itself. 68 Literature Cited Afzalimehr, H. and Anctil, F. 1999. Velocity distribution and shear velocity behaviour of decelerating flows over a gravel bed. Canadian Journal of Civil Engineering. 26:468-475. Altringham, J. D. and Ellerby, D. J. 1999. Fish swimming: Patterns in muscle function. Journal of Experimental Biology. 202:3397-3403. Aro, K. V. and Shepard, M . P. 1967. Pacific Salmon in Canada. North Pacific Fisheries Commission. Bull. 23. Basham, L. and Gilbreath, L. 1978. Unusual occurrence of pink salmon (Oncorhynchus gorbuscha) in the Snake River of southeastern Washington. Northwest Science. 52:32-34. Baxter, R. M . 1977. Environmental effects of dams and impoundments. Annual Review of Ecological Systems. 8:255-283. Beacham, T. D. and Murray, C. B. 1993. Fecundity and egg size variation in North American Pacific salmon (Oncorhynchus). Journal of Fish Biology. 42:485-508. Beamish, F. W. H. 1978. Swimming capacity. In Fish Physiology. Edited by Academic Press, New York, pp. 101-186. Beauchamp, D. A., Stewart, D. J. and Thomas, G. L. 1989. Corroboration of a bioenergetics model for sockeye salmon. Transactions of the American Fisheries Society. 118:597-607. Beffa, C. 1996. Backwater computation for transcritical river flows. Journal of Hydraulic Engineering. December:745-748. Bell, W. H. and Terhune, L. D. B. 1970. Water tunnel design for fisheries research. Journal of the Fisheries Research Board of Canada. 195. Bernatchez, L. and Dodson, J. J. 1987. Relationship between bioenergetics and behaviour in anadromous fish migrations. Canadian Journal of Fisheries and Aquatic Science. 44:399-407. Boisclair, D. 1992. An evaluation of the stereocinematographic method to estimate fish swimming speed. Canadian Journal of Fisheries and Aquatic Science. 49:523-531. Brett, J. R. 1962. Some considerations in the study of respirometry metabolism in fish, particularly salmon. Journal of the Fisheries Research Board of Canada. 19:1025-1038. Brett, J. R. 1967. Swimming performance of sockeye salmon (Oncorhynchus nerka) in relation to fatigue time and temperature. Journal of the Fisheries Research Board of Canada. 24:1732-1741. Brett, J. R. 1973. Energy expenditure of sockeye salmon, Oncorhynchus nerka, during sustained performance. Journal of the Fisheries Research Board of Canada. 30:1799-1809. 69 Brett, J. R. 1982. The swimming speed of adult pink salmon, Oncorhynchus gorbuscha at 20° C and a comparison with sockeye salmon, O. nerka. Canadian Technical Report of Fisheries and Aquatic Sciences. 1143. Brett, J. R. 1983. Life energetics of sockeye salmon, Oncorhynchus nerka. In Behavioural energetics: the cost of survival in vertebrates. Edited by Ohio State University Press, Columbus, pp.29-63. Brett, R. J. 1995. Energetics. In Physiological Ecology of Pacific Salmon. Edited by UBC Press, Vancouver, pp. 1-68. Burgetz, I. J., Rojas-Vargas, A., Hinch, S. G. and Randall, D. J. 1998. Initial recruitment of anareobic metabolism during sub-maximal swimming in rainbow trout (Oncorhynchus mykiss). Journal of Experimental Biology. 201:2711-2721. Burgner, R. L. 1991. Life history of sockeye salmon (Oncorhynchus nerka). In Pacific salmon life histories. Edited by UBC Press, Vancouver, pp.1-117. Cooper, A. C. 1983. Examination of possible causes of the loss of early Stuart sockeye in 1982 during the migration up the Fraser River system to the spawning grounds. International Pacific Salmon Fisheries Commission Engineering. April 21, 1983. Costa, D. P. and Prince, P. A. 1987. Foraging energetics of grey-headed albatrosses Diomedea chrysostoma at Bird Island, South Georgia. Ibis. 129:149-158. Crowder, D. W. and Diplas, P. 2000. Evaluating spatially explicit metrics of stream energy gradients using hydrodynamic model simulations. Can. J. Fish. Aquat. Sci. 57:1497-1507. Davidson, F. A., E. Vaughan, S.J. Hutchinson, and A.L. Pritchard. 1943. Factors influencing the upstream migration of the pink salmon (Oncorhynchus gorbuscha). Ecology. 24:149-168. Drucker, E. G. and Lauder, G. V. 2000. A hydrodynamic analysis offish swimming speed: wake structure and locomotor force in slow and fast labriform swimmers. Journal of Experimental Biology. 203:2379-2393. Dunster, J. and Dunster, K. 1996. Dictionary of Natural Resource Management. UBC Press. p.363 Ellis, D. V. 1966a. A survey of the behaviour of salmon on spawning migration through a large river system. Fisheries Research Board of Canada. 876. Ellis, D. V. 1966b. Swimming speeds of sockeye and coho salmon on spawning migration. Journal of the Fisheries Research Board of Canada. 23:181-187. Eniutina, R. I. 1972. The Amur pink salmon (Oncorhynchus gorbuscha): a commercial and biological survey. Izv. Tikhookean. Nauchno- Issled. Inst. Rybn. Khoz. Okeanogr. Transl. From Russian; Fish. Res. Board Can. Transl. Ser. 3160. 77:3-126. 70 Farrell, A. P., Johansen, J. A., Steffensen, J. F., Moyes, C. D., West, T. G. and Suarez, R. K. 1990. Effects of exercise-training and coronary ablation on swimming performance, heart size and cardiac enzymes in rainbow trout, Oncorhynchus mykiss. Canadian Journal of Zoology. 68:1174-1179. Fleming, I. A. and Gross, M. R. 1989. Evolution of adult female life history and morphology in a Pacific salmon (coho: Oncorhynchus kisutch). Evolution. 43:141-157. Foerster, R. F. and Pritchard, A. I. 1941. Observations on the relation of egg content to total length and weight in the sockeye salmon (Oncorhynchus nerka) and the pink salmon (O. gorbuscha). Proc. Trans. R. Soc. Can. Ser. 3. 35(5):51-60. Fretwell, M . R. 1981. Migration of adult sockeye salmon in the Nechako River system in 1980. International Pacific Salmon Commission. Unpublished report. Fretwell, M . R. 1989. Homing behaviour of adult sockeye salmon in response to a hydroelectric diversion of homestream waters at Seton Creek. International Pacific Salmon Fisheries Commission. Bull. X X V . Geist, D. R., Abernethy, C. S., Blanton, S. L. and Cullinan, V. I. 2000. The use of electromyogram telemetry to estimate energy expenditure of adult fall chinook salmon. Transactions of the American Fisheries Society. 129:126-135. Gilhousen, P. 1960. Migratory behavior of adult Fraser River sockeye. International Pacific Salmon Fisheries Commission Progress Report. 7. Gilhousen, P. 1980. Energy sources and expenditures in Fraser River sockeye salmon during their spawning migration. International Pacific Salmon Fisheries Commission Report. XXII. Godfrey, H. 1959. Variations in annual average wieghts of British Columbia pink salmon, 1944-1958. Journal of the Fisheries Research Board of Canada. 16:329-337. Godfrey, H., Hourston, W. R., Stokes, J. W. and Withler, F. C. 1954. Effects of a rock slide on Babine River salmon. Fisheries Research Board of Canada. Bull. 106. Grachev, L. Y. 1971. Alteration in the number of oocytes in the pink salmon (Oncorhynchus gorbuscha (Walbuam)) in the marine period of life. J. Ichthyology. 11:199-206. Gray, J. 1936. Studies in Animal Locomotion. VI. The propulsive powers of the dolphin. J. Exp. Biol. 13:391-400. Groot, C. and Margolis, M . 1991. Pacific Salmon Life Histories. UBC Press. 564. Hatfield, T. and Bruce, J. 2000. Predicting salmonid habitat-flow relationships for streams from western North America. N . Am. J. Fish. Manage. 20:1005-1015. Heard, W. R. 1991. Life history of pink salmon (Oncorhynchus gorbuscha). In Pacific salmon life histories. Edited by UBC Press, Vancouver, pp. 119-230. 71 Henderson, F. M . 1966. Open Channel Flow. Collier Macmillan Canada Ltd. p.522 Herskin, J. and Steffensen, J. F. 1998. Energy savings in sea bass swimming in a school: measurements of tail beat frequency and oxygen consumption at different swimming speeds. Journal of Fish Biology. 53:366-376. Hinch, S. G. and Bratty, J. 2000. Effects of swim speed and activity pattern on success of adult sockeye salmon migration through an area of difficult passage. Transactions of the American Fisheries Society. 129:598-606. Hinch, S. G., Diewert, R. E., Lissimore, T. J., Prince, A. M . J., Healey, M . C. and Henderson, M . A. 1996. Use of electromyogram telemetry to assess difficult passage areas for river-migrating adult sockeye salmon. Transactions of the American Fisheries Society. 125:253-260. Hinch, S. G. and Rand, P. S. 1998. Swim speeds and energy use of upriver-migrating sockeye salmon (Oncorhynchus nerks): role of local environment and fish characteristics. Canadian Journal of Fisheries and Aquatic Science. 55:1821-1831. Hinch, S. G. and Rand, P. S. 2000. Optimal swimming speeds and forward-assisted propulsion: energy-conserving behaviours of up-river migrating adult salmon. Canadian Journal of Fisheries and Aquatic Science. 57:2470-2478. Hogan, D. L. and Church, M. 1989. Hydraulic geometry in small, coastal streams: Progress toward quantification of salmonid habitat. Canadian Journal of Fisheries and Aquatic Science. 46:844-852. Hughes, N . F. and Kelly, L. H. 1996. New techniques for 3-D video tracking offish swimming movements in still or flowing water. Canadian Journal of Fisheries and Aquatic Science. 53:2473-2483. Idler, D. R. and Clemens, W. A. 1959. The energy expenditures of Fraser River sockeye salmon during the spawning migration to Chilko and Stuart Lakes. International Pacific Salmon Fisheries Commission Progress Report. 6. Jain, K. E., Birtwell, I. K. and Farrell, A. P. 1998. Repeat swimming performance of mature sockeye salmon following a bried recovery period: a proposed measure of fish health and water quality. Canadian Journal of Zoology. 76:1488-1496. Jain, K. E., Hamilton, J. C. and Farrell, A. P. 1997. Use of a ramp velocity test to measure critical swimming speed in rainbow trout (Onchorhynchus mykiss). Comparative Biochemistry and Physiology. 117A :441-444. Jonsson, N. , and B. Jonsson. 1997. Energy allocation in polymorphic Brown Trout. Functional Ecology. 11:310-317. Jonsson, N. , Jonsson, B. and Hansen, L. P. 1991. Energetic cost of spaening in male and female Atlantic salmon (Salmo salar L.). Journal of Fish Biology. 39:739-744. 72 Jonsson, N., Jonsson, B. and Hansen, L. P. 1997. Changes in proximate composition and estimates of energetic costs during upstream migration and spawning in Atlantic salmon Salmo salar. Journal of Animal Ecology. 66:425-436. Jowet, I. G. 1993. A method for objectively identifying pool, run and riffle habitats from physical measurements. New Zealand Journal of Marine and Freshwater Research. 27:241-248. Kaseloo, P. A., Weatherley, A. H., Lotimer, J. and Farina, M . D. 1992. A biotelemetry system recording fish activity. Journal of Fish Biology. 40:165-179. Kelly, L. H. 1999. How to calculate 3D coordinates with two cameras, a calibration object, a java program, and a lot of MS Excel macros. Lon Kelly. 1-9. Kondolf, M . G., Larsen, E. W. and Williams, J. G. 2000. Measuring and modeling the hydraulic environment for assessing instream flows. N . Am. J. Fish. Manage. 20:1016-1028. Leonard, J. B. K., Leonard, D. R. and Ueda, H. 2000. Active metabolic rate of masu salmon determined by repirometry. Fisheries Science. 66:481-484. Leonard, J. B. K. and McCormick, S. D. 1999. Effects of migration distance on whole-body and tissue-specific energy use in American shad (Alosa sapidissima). Canadian Journal of Fisheries and Aquatic Science. 56:1159-1171. Lighthill, M . J. 1969. Hydromechanics of aquatic animal propulsion. An. Rev. Revs. Fluid Mech. 1:413-446. Lighthill, M. J. 1975. Mathematical biofluiddynamics. Society for Industrial and Applied Mathematics, p.281 Macdonald, J. S. 2000. Mortality during the migration of Fraser River sockeye salmon (Oncorhynchus Nerka): a study of the effect of ocean and river enviromental conditions in 1997. Canadian Technical Report of Fisheries and Aquatic Science. 2315:120. Macdonald, J. S., and eleven others and . 2000. The influence of extreme water temperatures on migrating Fraser River sockeye salmon (Oncorhynchus nerka) during the 1998 spawning season. Canadian Technical Report of Fisheries and Aquatic Science. 2326. Macdonald, J. S., Williams, I. V. and Woodey, J. C. 2000. The effects of in-river conditions on migrating sockeye salmon (Oncorhynchus Nerka). Canadian Technical Report of Fisheries and Aquatic Science. 2315:39-58. Madison, D. M. , Horrall, R. M. , Stasko, A. B. and Hasler, A. D. 1972. Migratory movements of adult sockeye salmon (Oncorhynchus nerka) in coastal British Columbia as revealed by ultrasonic tracking. Journal of the Fisheries Research Board of Canada. 29:1025-1033. McKinley, R. S. and Power, G. 1992. Measurement of activity and oxygen consumption for adult lake sturgeon (Acipenser fulvescens) in the wild using radio-transmitted EMG signals. In Wildlife telemetry: remote monitoring and tracking of animals. Edited by Ellis Horwood, West Sussex, UK, pp.307-318. 73 Muller, U. K., Smit, J., Stamhuis, E. J. and Videler, J. J. 2001. How the body contributes to the wake in undulatory fish swimming: flow fields of a swimming eel (Anguilla anguilla). Journal of Experimental Biology. 204:2751-2762. Muller, U. K., Stamhuis, E. J. and Videler, J. J. 2000. Hydrodynamics of unsteady fish swimming and the effects of body size: Comparing the flow fields of fish larvae and adults. Journal of Experimental Biology. 203:193-206. Neave, F. 1952. 'Even-year' and 'odd-year' pink salmon populations. Proceedings of the Transactions of the Roayal Society of Canada. 46(5):55-70. Neave, F. 1966. Pink Salmon in British Columbia. International North Pacific Fisheries Commission. Bull. 18. Neave, F., Ishida, T. and Murai, S. 1967. Salmon of the North Pacific Ocean. Part VI. Pink salmon in offshore waters. International North Pacific Fisheries Commission. Norberg, U. M . 1996! Energetics of Flight. In Avian energetics and nutritional ecology. Edited by Chapman and Hall, New York, pp. 199-249. Northcote, T. G. and Burwash, M . D. 1991. Fish and fish habitats of the Fraser River basin. In Water in sustainable development: Exploring our common future in the Fraser River basin. Edited by Westwater Research Center, University of British Columbia, Vancouver, B.C., Oakland, F., Finstad, B., McKinley, R. S., Thorstad, E. B. and Booth, R. K. 1997. Radio-transmitted electromyogram signals as indicators of physical activity in Atlantic salmon. Journal of Fish Biology. 51:476-488. Pedley, T. J. and Hill, S. J. 1999. Large-amplitude undulatory fish swimming: fluid mechanics coupled to internal mechanics. Journal of Experimental Biology. 202:3431-3438. Pennycuick, C. J. 1972. Soaring behaviour and performance of some East African birds, observed from a motor-glider. Ibis. 114:178-218. Persov, G. M . 1963. The 'potential' and 'final' fecundity of fish through the example of the pink salmon Oncorhynchus gorbuscha (Walb.) acclimatixed in White and Barents Sea Basins. 3:490-496. Priede, I. G. 1985. Metabolic scope in fishes. In Fish Energetics: New Perspectives. Edited by Croom Helm Australia Pty Ltd., Sydney, pp.33-64. Quinn, T. P. 1988. Estimated swimming speeds of migrating adult sockeye salmon. Canadian Journal of Zoology. 66:2160-2163. Quinn, T. P. and Adams, D. J. 1996. Environmental changes affecting the migratory timing of american shad and sockeye salmon. Ecology. 77:1151-1162. Rand, P. S. and Hinch, S. G. 1998. Swim speeds and energy use of upriver-migrating sockeye salmon (Oncorhynchus nerka): simulating metabolic power and assessing risk of energy depletion. Canadian Journal of Fisheries and Aquatic Science. 55:1832-1841. 74 Reimchen, T. E. 2000. Some ecological and evolutionary aspects of bear-salmon interactions in coastal British Columbia. Canadian Journal of Zoology. 78:448 - 457. Ricker, W. F. 1964. Ocean growth and mortality of pink and chum salmon. Journal of the Fisheries Research Board of Canada. 21:905-931. Ricker, W. F. 1989. History and present state of the odd-year pink salmon runs of the Fraser River region. Canadian Technical Report of Fisheries and Aquatic Sciences. 1702. Royce, W. F., Smith, L. S. and Hartt, A. C. 1968. Models of oceanic migrations of Pacific salmon and comments on guidance mechanisms. Fisheries Bulletin (U.S.). 66:441-462. Sagnes, P., Champagne, J.-Y. and Morel, R. 2000. Shifts in drag and swimming potential during grayling ontogenesis: relations with habitat use. Journal of Fish Biology. 57:52-68. Sahin, I., Hyman, M . C. and Nguyen, T. C. 1994. Three-dimensional flow around a submerged body in finite-depth water. AppI. Math. Modelling. 18:611-619. SAS, 1.1. 1988. SAS/STAT User's Guide, Release 6.03 Edition. SAS Institute Inc. p.1028 pp. Spedding, G. R. 1986. The wake of a jackdaw (corvus monedula) in slow flight. Journal of Experimental Biology. 125:287-307. Spedding, G. R. 1987. The wake of a kestrel ^Falco Tinnunculus) in flapping flight. Journal of Experimental Biology. 127:59-78. z Tang, M. , Boisclair, D., Menard, C. and Downing, J. A. 2000. Influence of body weight, swimming characteristics, and water temperature on the cost of swimming in brook trout (Salvelinus fontinalis). Canadian Journal of Fisheries and Aquatic Science. 57:1482-1488. Taylor, E. B. and McPhail, J. D. 1985. Variation in burst and prolonged swimming performance amoung British Columbia populations of coho salmon, (Oncorhynchus kisutch). Canadian Journal of Fisheries and Aquatic Science. 42:2029-2033. Triantafyllou, M . S. and Triantafyllou, G. S. 1995. An efficient swimming machine. Scientific American. March:64-70. Trudel, M . and Boisclair, D. 1996. Estimation offish activity consts using underwater video cameras. Journal of Fish Biology. 48:40-53. Tucker, V. A. 1990. Body drag, feather drag and interference drag of the mounting strut in a peregrine falcon, Falco Peregrinus. Journal of Experimental Biology. 149:449-468. Tucker, V. A. and Parrott, C. G. 1970. Aerodynamics of gliding flight in a falcon and other birds. Journal of Experimental Biology. 52:345-367. 75 Vernon, F. H. 1962. Pink salmon populations of the Fraser River system. In Syposium of Pink Salmon. H.R. MacMillian Lectures in Fisheries. Edited by Institute of Fisheries, University of British Columbia, Vancouver, BC, pp.53-58. Vernon, F. H., Hourston, A. C. and Holland, G. A. 1964. The migration and exploitation of pink salmon runs in and adjacent to the Fraser River Convention Area in 1959. Videler, J. J. 1993. Fish swimming. Chapman and Hall. p.260 Vogel, S. 1994. Life in Moving Fluids. Princeton University Press, p.467 Ward, F. J. 1959. Character of the migration of pink salmon to Fraser River spawning grounds in 1957. International Pacific Salmon Fisheries Commission. Bull. 10. Wardle, C. S., Videler, J. J. and Altringham, J. D. 1995. Review tuning in to fish swimming waves: body form, swimming mode and muscle function. Journal of Experimental Biology. 198:1629-1636. Ware, D. M . 1982. Power and Evolutionary Fitness of Teleosts. Canadian Journal of Fisheries and Aquatic Science. 39:3-13. Webb, W. 1995. Locomotion. In Physiological Ecology of Paific Salmon. Edited by UBC Press, Vancouver, pp.69-100. Weihs, D. 1973a. The mechanism of rapid starting in the slender fish. Biorheology: 10:343-350. Weihs, D. 1973b. Optimal fish cruising speed. Nature. 241:48-50. Weihs, D. 1974a. Energetic advantages of burst swimming of fish. Journal of Theoretical Biology. 48:215-229. Weihs, D. 1974b. Some hydromechanical aspects of fish schooling. In Swimming and Flying in Nature. Edited by Plenum, New York, pp.203-208. Weihs, D. and Webb, P. W. 1983. Optimization of Locomotion. In Fish Biomechanics. Edited by Praeger Publishers CBS Educational and Professional Publishing, New York, pp.339-371. Williams, I. V. and Brett, J. R. 1987. Critical swimming speed of Fraser and Thompson River pink salmon (Oncorhynchus gorbuscha). Canadian Journal of Fisheries and Aquatic Science. 44:348-356. Williams, I. V., Brett, J. R., Bell, G. R., Traxler, G. S., Bagshaw, J., McBride, J. R., Dye, H. M. , Sumpter, J. P., Donaldson, E. M. , Billinski, E., Tauyuki, H., Peters, M . D., Choromanski, E. M. , Cheng, J. H. Y. and Coleridge, W. L. 1986. The 1983 early run Fraser and Thompson River pink salmon; morphology, energetics and fish health. International Pacific Salmon Commission. Bull. XXIII. Xie, Y., Cronkite, G. and Mulligan, T. J. 1997. A split beam echosounder perspective on migratory salmon in the Fraser River: A progress report on the split-beam experiment at Mission, B.C., in 1995. Pacific Salmon Commission. 8. 76 Yates, G. T. 1983. Hydromechanics of body and caudal fin propulsion. In Fish Biomechanics. Edited by Praeger Publishers, New York, pp. 182-195. 77 APPENDIX A. The migratory paths of study fish through the Fraser River Canyon. 78 oo 90 oo so o ON CN O cn ON ON ON ON u H be 00 ON ON APPENDIX B. The instantaneous swimming speeds of study fish through reaches 2, 4, 7 and 9 in the Fraser River Canyon. 100 era" o l-t a CO ON ^ 3 cs ct> o CO CO t r o St 3 5 SS. CD P co 3 3 O O a £ S i 3 09 co ct> CD Cu CO P T 3 CO p 3 O 3 3 » cV co' 3 CD O 1-1 CD P O t r CD 3 CD X Swimming Speed (BL*s_ 1) M M w ^ U I - O P — i (S> W 4^  ^ - O i — Is) U ) J i . cyi O O O O Os I—»• o B s CD . . 4^  IS) VD 4^  4^  O K ) OO 4^  00 to 5! C) H a> o =r vO O O o o o Os 4^  vo Is) O 4-. 4^  Ui VO C>) 4^  4^  O IS) OO 4^  00 is) u> 2 C) > CD o or M o o o o o Os 4-. vo IS) o 4^  IS) Ui VO 4^  Ui 4-. O IS> 00 4^  00 is) o > 2 o o © o o Os ts) U ) 4^. <-fi 4^  VO O 4^  4^  Is) Ui VO 4^  O J 4^  O 4^  Is) oo 4^  00 K ) CD o or t o J .3' o CD P o 3" ts) < £ • i-i CD « ' H fa 3 3 cx CD VO on a> H CD' cr cn CD ^ a a i—. on ^ O ST. ^ s i' oo oo i-t P ^ 03 CD 3 SS « 2 O rt 00 3 ere oo >a CD CD o-00 c? l-l 03 Swimming Speed (BL»s_1) O O H - rO U ) 4> U i d o K3 U ) -fc. U i 13 o o © d o 4^  p d ON oo 3 o 05 3 oo" cr D o 3 o l-l CD 03 o cr CD S3 CD 3> CD o O K> U i 1>J O U) ON o 2 o u > > CO pa O =r vO vo to > ) CO fa o c r P Q d o d o o o K> LtJ U l O d ON o O K> U i o ON o 4S> VO CO 11 CD 03 O cr TO' c l-l 03 oo i i ' P- f t vo co 3 CD C r cn CD 2^ ° a — • cya 8 i ' cn cn i i ps 1- i P CD CS 2 o 2- e " cn C cn 3. 3 O i-b I - I CD » o cr CD 3 Swimming Speed (BL #s _ 1) > i—> to O J 4> Ui to H - to O J 4^. ( O i - M W to O O d o o to o to 4^ O O J OS to 00 4^ oo O J ON o o 4^ to Ul o to 4^ Ul ^4 O J ON o 4^ OO -O i - K) W J i Ui o o d o p d to ON O to Ul O J o 4 ^ NO o Ul 4^ ON o 1^ cn s» o CD X 3> l-l l-l CD 03 O cr to o oo Swimming Speed (BL #s _ 1) l-l a> CO Swimming Speed (BL #s _ 1) CD to i-t OS CD • . P H S-B' (vj CD CO 4*. CD P O o-V£> cn co o* CO O i-4 as oo co CO ^ CO • • P tya Swimming Speed (BL»s_1) o o o o o p o is) Os O Is) lo O 4-. VO o OS o o oo oo o O O W 4> C \ 00 o O O W J i O \ 00 o 21 O 7* o or o o © o © o OS o H to o 4*-SO o OS o ^1 o oo 00 o io > o -f o O © P o o o r 4> © © -io Os O © • Is) © © -4=. VD O O • Ut *» OS p © -K ) © © -00 i>) 00 © o © © CD P O or IsJ CD o o CD X oo 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090064/manifest

Comment

Related Items