UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The influence of surface currents on the dispersal of coho salmon (Oncorhynchus kisutch) from the Strait… Wagey, Gabriel Antonius 1995

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

Item Metadata

Download

Media
831-ubc_1995-0285.pdf [ 6.19MB ]
Metadata
JSON: 831-1.0053241.json
JSON-LD: 831-1.0053241-ld.json
RDF/XML (Pretty): 831-1.0053241-rdf.xml
RDF/JSON: 831-1.0053241-rdf.json
Turtle: 831-1.0053241-turtle.txt
N-Triples: 831-1.0053241-rdf-ntriples.txt
Original Record: 831-1.0053241-source.json
Full Text
831-1.0053241-fulltext.txt
Citation
831-1.0053241.ris

Full Text

The influence of surface currents on the dispersal of coho salmon (Oncorhynchus kisutch) from the Strait of Georgia by Gabriel Antonius Wagey B.Sc , Bogor Agriculture University, Indonesia, 1988 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Oceanography We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1995 © Gabriel Antonius Wagey, 1995 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. Department of O^^OG?-^^ The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract The dispersal pattern of coho salmon (Oncorhynchus kisutch) was studied to explore the extent to which coho distribution can be explained by random movement and advection by surface currents. A computer simulation model was constructed to simulate the dispersal of tagged coho salmon after 13 months of ocean residence. The simulation model is spatially organized into 310 habitat boxes representing the waters surrounding Vancouver Island. Simulated fish represented by number of particles were released and subject to pure random walk and biased random walk procedures. The biases in particle movements were proportional to the strength and direction of the surface currents. The behaviour of the fish also included a weak shore seeking tendency. The general pattern of surface currents of the waters around Vancouver Island was estimated using the results of oceanographic studies available from that area. Best estimates of surface current was based on the seasonal pattern in each major water bodies. Catch-effort analysis was performed to determine the nature of troll fisheries which are used to compare the results of the model. The troll fisheries are found to be more intensive in the outside region (in the west coast of Vancouver Island) than in the Strait of Georgia region. The distribution of tag recoveries was analyzed by using the abundance index of tag per unit effort (TPUE) from each recovery regions. From the TPUE distribution of Big Qualicum fish, it is significant that there are more fish in the inside area, especially in the Strait of Georgia region compared to the outside area. ii The simulation results show that the model is sensitive to current as compared to random movement. The general pattern of simulated fish distribution is found to be abundant in the Strait of Georgia region, followed by the Central troll region in the northern part of Vancouver Island, and the southwest and northwest regions, respectively. Trapping event by the surface current is speculated to cause the high proportion of fish in the Strait of Georgia compared to other regions. The comparison between the distribution resulted from the simulation models and the observed distribution from the tag return data shows a qualitative similarity which suggests that the influence of surface currents on fish dispersal is apparent. From this study it is apparent that the influence of surface currents alone is not enough to account for interannual variability in the distribution of tag recoveries. Several factors, such as directed migration of the fish, availability of food and oceanographic conditions, including anomalous flushing events, wind and sea surface temperature, were speculated to become the potential factors that might cause the interannual variations. iii Table of Contents Abstract ii List of Tables vi List of Figures vii Acknowledgement ix 1. Introduction 1.1. Coho distribution 2 1.2. Migratory behaviour 4 1.3. Influence of oceanographic factors on coho distribution 6 1.4. Modelling fish migration 9 1.5. Thesis objectives and overview 11 2. Background- Area of Interest 2.1. Coho life history • • • 13 2.2. Coho dispersal and distribution in the ocean 16 2.3. Shore seeking tendency 19 2.4. Study area 20 2.5. Hatchery production of coho 20 3. Oceanographic conditions 3.1. Geographical boundaries 28 3.2. General patterns of net surface currents 30 3.2.1. Strait of Georgia 30 3.2.2. Juan de Fuca Strait : 37 3.2.3. Johnstone Strait and Queen Charlotte Strait 38 3.2.4. West coast of Vancouver Island 39 4. The Model 4.1. Introduction 43 4.2. Structure of the model 43 4.2.1. Input data 43 4.2.2. Movement probability 46 (i) Pure random movement 46 (ii) Biased random movement 48 4.2.3. Particle distribution 51 4.3. Number of iterations 52 4.4. Simulation scenarios 54 iv 5. Mark Recovery Program and Coho fisheries 5.1. Introduction 56 5.2. Statistical areas 58 5.3. Troll fisheries 60 5.3.1. Catch and effort distribution of troll fishery . . . 65 (i) Catch 65 (ii) Effort 67 (iii) CPUE 69 5.4. Coded wire tag recoveries 71 5.5. Distribution of Big Qualicum fish from MRP database 72 6. Simulation results 6.1. Random movement 76 6.2. Biased random movement 79 6.3. Seasonal distributions 82 7. Discussions 7.1. General distribution of coho salmon from Big Qualicum hatchery 87 7.2. Interannual variations 93 Bibliography 103 v List of Tables Table 3.1. Dimensions of coastal waters surrounding Vancouver Island 28 Table 4.1. Probability table for particle moving from box # 5 to all adjacent boxes 48 Table 4.2. Probability table for particle moving from box # 5 given that box 4, and #7 are land 48 Table 4.3. Cumulative probability table for particle moving from box # 5 to any adjacent boxes in numerical order 51 Table 4.4. Season, coho size, swimming speed, distance swim and iteration of the model for standard runs of the model 54 Table 5.1. Classification of catch regions used in the study area 60 vi List of Figures Figure 2.1. Coho salmon life history, with emphasis on oceanic stages 14 Figure 2.2. Ocean distribution of British Columbia stocks of coho salmon 18 Figure 2.3. Geographical map of the study area 21 Figure 2.4. Location of major hatcheries for coho salmon 23 Figure 2.5. Number of coho smolts released from hatcheries located in the inside and outside areas 24 Figure 2.6. Catch proportion of coho salmon caught by trollers from the inside and outside areas 25 Figure 3.1. Major water bodies around Vancouver Island 29 Figure 3.2. Diagram of estimated net surface currents and velocities of the study area during the period of June to September 31 Figure 3.3. Diagram of estimated net surface currents and velocities of the study area during the period of October to November 32 Figure 3.4. Diagram of estimated net surface currents and velocities of the study area during the period of December to March 33 Figure 3.5. Diagram of estimated net surface currents and velocities of the study area during the period of April to May 34 Figure 4.1. Structure of the simulation model 44 Figure 4.2. Diagram map of the study area, showing the grid of habitat boxes 45 Figure 4.3. Diagram of 3 x 3 system 47 Figure 5.1. Monthly average catch from June to September of troll fishery operating in 4 recovery regions 58 Figure 5.2. British Columbia fisheries statistical areas 59 Figure 5.3. Catch proportion between inside and outside areas from 1975 - 1992 62 Figure 5.4. Sport and troll catch of coho salmon in the Strait ofGeorgia 64 Figure 5.5. Annual troll catch in the inside and outside areas 66 Figure 5.6. Annual troll effort in the inside and outside areas 68 Figure 5.7. Annual troll catch per unit effort (CPUE) in the inside and outside areas 70 Figure 5.8. The proportion of tag per unit effort of coho salmon caught by trollers 74 vii Figure 6.1. Distribution of simulated fish under random movements, with different swimming speeds 77 Figure 6.2. Sensitivity analysis graph of the simulated fish distribution under different currents and swimming speeds 80 Figure 6.3. Comparison of the distribution of simulated fish resulted from random and biased simulations 82 Figure 6.4. Seasonal distributions of simulated fish with standard swimming speed (iteration = 451) 84 Figure 6.5. Seasonal distributions of simulated fish with double swimming speed (iteration = 902) 85 Figure 7.1. Comparison between observed and simulated (random and biased) distributions 88 Figure 7.2. Variations of tagged fish per unit effort. Examples are from 1983, 1990 and 1991 94 Figure 7.3. Fraser River mean discharge (measured at Hope) 97 Figure 7.4. Exceedance diagram of the wind force from June 1982 to July 1983 99 Figure 7.5. Deviations of wind directions in 1983 as compared to the average wind pattern from 1981 - 1989 101 vi i i Acknowledgment This study was made possible through the support and assistance from many people whom I would like to thank truly. First of all, I sincerely thank my supervisor Professor Michael Healey, for his guidance, support and encouragement throughout my master program. I would also like to extend my gratitude to Professor Paul LeBlond, for his helpful guidance and suggestions. I thank Louis Lappi and Ron Kadowaki at the Pacific Biological Station in Nanaimo, for providing data for the Mark Recovery Analysis, and to Oscar and Joko for their assistance in computer programming. I am indebted to the Eastern Indonesian University Development Project (EIUDP) for the funding during my study, and the Pattimura University in Ambon, for giving me the opportunity to continue my study abroad. Loving acknowledgment goes to my wife, Venska, and my parents who encourage me through their love and prayer. ix 1. Introduction Study of fish migration is an important component to understand the life history of a species, since it may provide an opportunity to explore the interactions between fish and its environment, from the juvenile to adult stages. One interesting aspect of this study is the influence of environmental factors on the migration process itself, which could lead to better understanding the distribution of a specific during different stages in its life history. According to Harden Jones (1981), migratory behaviour is a feature in the life histories of fish directed towards reproductive success. Migration strategy is well developed in some anadromous species, such as the pacific salmon, so that the adult fish return for spawning to the exact same natal stream where they had hatched. In the migration process, the influence of the surrounding environment is important to the salmon species to accomplish their migratory goals. Like other juvenile fish that make their migration to the feeding ground in the ocean, most salmon are subject to the effects of surface currents. At the feeding areas, the effect of food supply distribution and some oceanographic features, such as water temperature, may also play an important part in the migration of these fish. Coho salmon in British Columbia, especially from the Strait of Georgia, provide a unique example to explore the dispersal and, hence, the distribution of fish. In 1950, Dr. J. Milne, a researcher from DFO in Nanaimo observed that the number of coho salmon in the Strait of Georgia fluctuated annually. Milne hypothesized that the reason for the fluctuation was the migration of the fish to the open ocean, mainly to the west coast of Vancouver Island. This idea 1 2 has raised the question of causes of this phenomenon, which still occurs. Beside this, another interesting reason for studying coho distribution, is the relatively simple pattern of dispersal and distribution shown by this species. Unlike the other pacific salmon, coho is regarded as a coastal species, meaning that usually this species does not make extensive migrations. The advantage of this situation is that one can explore the dispersal pattern of the whole ocean life of this species in a relatively small area. 1.1. Coho distribution North Pacific salmonids are among the most documented and well-studied species with regard to their distribution (Godfrey et al, 1975; Pearcy, 1992). Among the seven recognized species of the North Pacific salmonids, the coho salmon (Oncorhynchus kisutch) is known to be distributed from northern Japan through the Kamchatka peninsula, across the Bering sea to Alaska and south through all north American waters to California (Sandercock, 1991; Hartt and Dell, 1986). Although some Pacific salmon species make extensive oceanic migration, in the northeastern Pacific, Hartt and Dell (1986) and Pearcy (1992) speculated that coho do not make extensive off-shore migrations. Based on their work with Oregon and Washington coho, Pearcy and Fisher (1988) concluded that in the spring, this species tends to be carried southward by currents during the first weeks of ocean life. Later, during summer when the currents are weaker and the fish are able to swim northward against them, the movement of fish changes to the north. Pearcy and Fisher (1988) also stated that, despite these latitudinal movements, coho salmon from 3 Oregon and Washington do not migrate great distances. In the Strait of Georgia, Healey (1980) observed that coho smolts disperse quickly throughout the Strait after making their ocean entrance. However, like the Oregon and Washington coho, they do not make an extensive migration to the open ocean. Healey (1993) suggested that, in general, most coho in the North Pacific ocean are found in continental shelf waters throughout their ocean life. These studies suggest that coho distribution in the ocean is predominantly inshore, although some coho do move to the open ocean. In British Columbia, Milne (1950) has indicated that the population of coho salmon, based on historical observation, can be divided into two distinct types, the "inside" and "outside" coho. The inside type is found in the waters between Vancouver Island and the mainland, whereas the outside type is found in the area west of Vancouver Island. Milne (1950) believed that the outside coho spend most of their ocean residence off the west coast of Vancouver Island, and these fish grow more rapidly, whereas the inside coho remain within the Strait of Georgia and Puget Sound and grow more slowly. Godfrey et al. (1975), postulated that coho originating from British Columbia will spend time in the coastal area adjacent to their natal rivers. However, once they have migrated a considerable distance, and have found a rich feeding area such as the west coast of Vancouver Island, they will remain there until they have to migrate back to their natal tributaries to spawn. In a recent report, Department of Fisheries and Ocean (DFO, 1990), expressed concern about the high variation in the number of coho salmon in the Strait of Georgia. It has been suggested that the troll fisheries operating in the west of Vancouver Island affect the inside coho 4 population, especially those fish originating from the Strait of Georgia. This concern shows that effective management of coho salmon depends greatly on knowledge of the fish distribution. Since the migratory behavior of the North Pacific salmonids, as anadromous species, is a significant aspect in their life history, it is important to understand the dispersal pattern from which the oceanic distribution of this species is determined. A study of coho dispersal during their ocean life is, therefore, useful both to assist fisheries managers with information regarding the species distribution and to understand better the relationship between the fish and their environment.. 1.2. Migratory Behaviour Griffin (1955) defined three categories of orientation mechanisms in animals. The first category includes piloting, whereby the goal is reached by referring to familiar landmarks that may be identified by a variety of sensory modalities. Dodson (1988) mentioned that the piloting mechanism also implied that the animal possesses a familiar area map built on individual experience of the spatial distribution of particular features of the environment. In this case the animal may take short cuts or choose between alternative routes to a goal without having to rely on a particular sequence of locations. The second category is compass orientation, in which the goal is reached by orienting in a given compass direction without reference to local landmarks. In this category, the animal is believed to have the ability to move in a particular compass direction even in unfamiliar territory by means of celestial or other reference clues. This orientation is also called one-directional orientation. The third category includes true navigation in which the goal is reached by orienting in the appropriate direction. For coho salmon, Pearcy (1992) and Moser 5 et al. (1991) speculated that, despite the influence of ocean surface currents, the movement of this species in the ocean also involves directed migration described within these mechanisms. Healey and Groot (1987) mentioned that in general, juvenile or young salmon that are migrating away from the breeding grounds for the first time may use compass orientation to perform their migration. The third category, if used, is usually performed by older animals who are homing to their natal stream. According to Dodson (1988 ), these mechanisms are considered as the reactions to the stimuli received by the fish in establishing their migration process. In order to establish a spatial distribution during their ocean life, Pacific salmonids rely on migratory behavior, which is affected by both stimuli that are internal to the fish and external factors such as oceanographic conditions. Sandercock (1991), believed that the abundance of coho salmon inside the Strait of Georgia is largely dependent on the oceanographic features in this region. The next two sections of this chapter will focus on the effects of internal stimuli and the oceanographic factors that are important to the migratory behavior of this species. Several studies have discussed the importance of migratory behavior on the marine distribution of some species. Godfrey et al. (1975) postulated that the distribution of coho salmon in the North Pacific can be considered a product of the evolutionary process of the species after a considerable period of time. Similarly, Neill (1984) pointed out that the adaptive value of migration is for fish to optimize their surrounding environment. He argued that if the environment supports the requirement of the fish to grow and reproduce, then there is no need to migrate. Neill (1984) also believed that most migratory fishes have come to rely partly on sign stimuli to bias their movements in favor of the appropriate heading. In other words, the fishes try to maximize the comfort of the surrounding environment, through movements from one place to another. Although this strategy is a common behavior for most species, the process will vary among species, based partly on the behavior of the fish and factors that influence the movement itself. 1.3. Influence of oceanographic factors on coho distribution Leggett (1984) reviewed the literature on fish migration and noted that physical and biological features of the environment play an important role in the life history of aquatic migratory species. It is believed that oceanographic conditions such as temperature, salinity and water currents influence the migration and the distribution of coho in the North Pacific (Hoar, 1958; Sandercock, 1991). Godfrey et al. (1975), based on data from the Japanese gillnet and longline fisheries in the open ocean, found that coho salmon were present in waters having a relatively wide range of surface temperature. However, the best catch occurred where surface temperature ranged between 8 - 12 °C. Several studies conducted by other people, have shown a similar range of optimum temperature for coho salmon in the open ocean (Manzer et al. 1965 and Godfrey, 1965). The link between salinity and the ocean distribution of coho salmon may be related to the abundance of food organisms. Milne (1950) hypothesized that river run-off from the Fraser River during the summer period in the Strait of Georgia affects the availability of food organisms for coho salmon in this area. Furthermore, Healey (1980) noted that, in the summers of 1975 and 1976 in the Strait of Georgia, there were positive correlations between coho abundance and the 7 amount of food in their stomachs. French et al. (1976) and Burgner (1991) have studied the possible effects of salinity on sockeye salmon migration. Although they found that in the open ocean, salinity variation would probably not affect the migration route of this species, it could in coastal areas, where salinity variation is greater. Several studies have been done to examine the role of surface currents in fish migration. The best example for coho salmon is the work by Pearcy and Fisher (1988) who concluded that juvenile coho off Oregon and Washington coasts were advected southward after entering the ocean, but later would move northward against the current. McCleave and Wippelhauser (1987) have shown that larger ocean currents influence the distribution of juvenile American eel (Anguilla spp.). In addition, Power et al. (1978) mentioned that hatchery-reared Atlantic salmon smolts (Salmo salar) apparently used surface currents to move seaward by drifting passively. In the Strait of Georgia and the waters east of the Vancouver Island (see Figure 2.3 on page 21), winds and tides can have a great influence on surface water movement and can induce quite rapid currents. The area around Discovery Passage is known for its swift tidal currents which can reach 500 cm/ sec (Thomson, 1981). Similarly, strong tidal currents can be found in the southern part of the Strait of Georgia, around Haro Strait and the San Juan islands. Crean (1967) estimated that the surface current in this region has a maximum speed of 100 cm/ sec. These strong currents may have a significant effect on the movement of juvenile fish in this area. Fraser River sockeye smolts entering the Strait of Georgia can be transported west and south of the Fraser River by the river plume and tidal currents (Groot and Cooke, 1987). Recently, Peterman et al. (1994) concluded that surface currents, due to wind force within the Strait of Georgia, can affect the migratory route of sockeye salmon smolts. They found that sockeye salmon smolts tend to move northwestward through the eastern migration route (i.e. the coastal area on the mainland side). Adult fish, on the other hand, appear to be little influenced by the currents. Pasqual and Quinn (1991) tried to model the migration of adult Sockeye salmon in the Strait of Georgia by using telemetry data. They found that surface currents had little influence on the movement of mature fish inside the Strait. The time at which coho salmon enter the ocean and the size of coho smolts may impact the distribution of coho salmon in the Strait of Georgia. Bilton et al. (1984) and Morley et al. (1988) studied the distribution of coho salmon from Quinsam hatchery in Vancouver Island. They found that fish released early had a more extended northward distribution than fish released later. More recent study by Irvine and Ward (1989) on coho salmon leaving Keogh River in the northern part of Vancouver Island showed similar results. They speculated that early migrating smolts may have different oceanic migration routes than later migrating smolts. Despite their relatively limited dispersal and migrations, coho salmon show considerable variations and complications in their migrations. Modelling fish dispersal could be used as an approach to study coho dispersal, and hence the distribution, since it is difficult to examine the entire period of ocean life history of this species in the field. Historic approaches have shown that the use of computer modelling to analyze fish distribution has solved some issues regarding mechanisms of fish distribution (DeAngelis and Yeh, 1984; Thomson, et al., 1992; Peterman, et al., 1994). 9 1.4. Modelling fish migration The study of fish migration and orientation has been the main focus of several studies using computer simulation models. Although there are some concerns about the limitations and simplification implicit in this method, simulation models can contribute significantly to understanding some aspects of animal movements (DeAngelis and Yeh, 1984). These authors believe that mathematical models can play an important role in revealing the physiological and environmental factors involved in fish movements over long temporal and spatial scales. The study of salmon migration using computer simulation was pioneered by Saila and Shappy (1963), who showed that the open-ocean phase of the salmon's return to their natal streams could be achieved with much less orientation than had commonly been supposed. This work was criticized by Quinn and Groot (1984) and Thomson et al, (1992) who re-examined the assumptions that Saila and Shappy made. They found that the assumptions related to movements and orientations were unrealistic and that the movement of mature sockeye salmon was likely to be well directed. Other models have explored the distribution of sockeye salmon (O. nerka) from the underyearling stage (Simms and Larkin, 1977) to the mature sockeye on their way home to natal streams (Pasqual and Quinn, 1991). In a more general approach, Harden Jones (1968) and later van der Steen (1984), tested the methods that might be used by fish in their migration. Two basics methods have been considered. The first one is simple random movement. The second method is movement which contains elements of directed movement. The second method, known as biased random movement, has been widely used to simulate animal migration using computer programs 10 (DeAngelis and Yeh, 1984; Renshaw, 1991). For sockeye salmon, Simms and Larkin (1977) presented an excellent example of computer simulation of the distribution of juvenile sockeye salmon in Babine Lake, British Columbia. The model elements were a hypothetical grid environment representing the entire lake and a number of particles that represent the juvenile fish. Particles were moved from grid point to grid point according to random dispersal or directionally biased dispersal. The results of simulation under random dispersal gave a low correlation between simulated and observed dispersal. On the other hand, the best correlation was obtained when particle movement was simulated as a combination of random and biased movements. This result gives evidence of an important role for both internal (eg. migration behavior, swimming speed) and external (eg. currents, temperature) factors in the mechanism of fish dispersal. More useful information might have been obtained if the simulation model had used some environmental data, such as surface currents. Another example of computer simulation of fish migration is that of DeAngelis and Yeh (1984). In their simulation, the hypothetical heterogenous environment was designed to simulate the role of some environmental factors such as currents and chemical substances that are usually present in an estuarine ecosystem. The fish, represented by particles, were moved among the habitat compartments that had different conditions of such external factors. This model used biased random movement to move the particles. The movement was dependent on local environmental conditions. After a number of iterations, the particles in each compartment were counted and the results were analyzed to show the importance of environmental factors on fish distribution. 11 1.5. Thesis objectives and overview The main objective of this study is to use simulation modelling to explore the extent to which coho distribution can be explained by random movement and advection by currents. To attain this objective, a simple simulation model was constructed as a tool to explore the dispersal pattern of tagged coho salmon released from the Big Qualicum hatchery. The results of simulations involving different patterns of surface currents around the Vancouver Island and different swimming behaviours of coho salmon were compared with actual distribution of coho derived from tag returns. This study is designed: 1. To examine the distribution resulting from the dispersal behaviour of coho salmon released into the Strait of Georgia. 2. To model the effects of surface currents on coho dispersal in order to determine if these are sufficient to account for interannual variation in dispersal behaviour. The next chapter will discuss the background and the study area. This will include a description of some major coho hatcheries located on Vancouver Island and the mainland, particularly the Big Qualicum hatchery. Chapter 3 will provide a description of oceanographic features of the study area. Surface water currents, including their variability and magnitude, is the main topic to be discussed in this chapter. In chapter 4,1 will introduce the computer program that I used in the simulation model. The details of the program will also be explained. Chapter 5 describes the MRP (Mark Recovery Program) from which most data on the ocean distribution of coho salmon were obtained. The dispersal and distribution of coho salmon 12 originating from Big Qualicum hatchery will be described in this chapter. This chapter also discusses the existing troll fishery and the data derived from this fishery. The results and discussion of simulated coho distribution under various surface current scenarios are presented in chapter 6. This chapter includes the results of sensitivity analysis of the model, and as well, the results of the model using the 'best' estimate of surface currents. The final chapter presents the discussion as well as the concluding comments. 2. Background - Area of Interest 2 . 1 Coho Life History It is apparent that, in British Columbia, coho salmon are the most widespread of the five Pacific salmon species in terms of spawning locations. Among the 1500 streams in the province from which records of spawning salmon are available, coho are known to occur in at least 900 (i.e. about 64%). The locations of the main spawning sites are spread over a large number of moderate-sized streams rather than a few large ones (Aro and Shepard, 1967). To provide the ecological and biological context for this study I shall present a summary of the coho's life cycle. This summary will emphasize aspects of coho's life history and biology that are most relevant to understanding the marine distribution of this species. Unlike the other species of Pacific salmon, coho salmon spawn over prolonged periods very late in the year, in the channels of large rivers, small tributary or headwater streams and even drainage ditches. The spawning season for most coho populations in North America ranges from October to March (Sandercock, 1991). In British Columbia, Fraser et al. (1983) and Neave (1949) observed that the spawning season is primarily between November and December. Figure 2.1 summarizes the life history of coho salmon. 13 Age Habitat Phase F R E s H W A T E R Adult Spawner Eggs Smolts r Juvenile o c E A N Pre Adult "Jacks" — 3 Adult Figure 2.1. Coho salmon life history, with emphasis on oceanic stages. (Adapted from Botsford et al., 1989) 15 Almost all young coho spend a year or more in streams or lakes before they migrate to the sea as smolts (Godfrey, 1965). Between March of one year and May of the next year, this species will emerge from gravel nests, take up residence in a stream and remain there for at least one year. During this period, the small fish grow to become smolts. By this stage the fish are ready to make their downstream migration to the open ocean. The migration period for most British Columbian coho is primarily in May and takes only few days to two weeks to reach the ocean. In early June these fish are usually found in the estuary, at the mouth of the streams. Once in the estuary or adjacent marine waters, they will remain close to the shoreline for several weeks. Later, they move away from shore and over the next year disperse in coastal waters. With regard to juvenile coho, Healey (1978) reported that, from the 1974 - 1976 sampling of juvenile salmon in the Strait of Georgia, this species showed a clear distribution in the Strait during the first winter, especially in the area of Gulf Islands, where the fish found good feeding grounds. Furthermore, Healey (1980) hypothesized that the juvenile coho in the Strait of Georgia might move to other areas if they found themselves in an area of poor feeding, but stay in the Strait if they found a good feeding ground. Findings from tag returns suggest that coho salmon, unlike sockeye and chum salmon (O. keta), do not make long oceanic migrations. Sandercock (1991) summarized several studies, such as those of Clemens (1930); Foerster (1955); Godfrey (1965); Godfrey et al. (1975) and Hartt and Dell (1986) which, based on their analysis of catch data and recovery of tagged coho data in North America, confirm the short migration of coho. There are reports that mention the occurrence of coho salmon in the open Pacific ocean but the densities of the fish appear to be much less than in the coastal waters. 16 On average, coho spend 15 months in the ocean before migrating back to their natal streams to spawn. Usually they start this spawning migration in the late summer or fall of their second year in the ocean. There is a small but variable proportion of precocious males in most spawning populations, known as 'jacks', that return to the natal streams after only one summer in the ocean. In the subsequent section, the discussion will focus on the ocean phase of the life history, mainly the first year of ocean life, which is critical in the completion of the coho migration process. 2.2. Coho Dispersal and Distribution in the ocean Coho salmon are believed to disperse quickly as they enter the ocean. According to Healey (1980), this may be the result of some oceanographic factors. An important factor in establishing the distribution of juvenile coho salmon when they first enter the sea is surface currents. Pearcy and Fisher (1988) and later Pearcy (1992), showed that the strong currents that move southward in the coastal area of Washington and Oregon states, transported the juvenile coho coming out from coastal streams. This advection process, which usually occurs in May, carries the juvenile fish as far south as the California coast. In June, when the juveniles become larger and the currents weaker, they swim northward against the currents, up to the Juan de Fuca Strait in British Columbia. However, the majority of the tagged juveniles coho recovered in late summer were found close to the mouths of their home streams. Pearcy and Fisher (1988) concluded that juvenile coho from Washington and Oregon are not highly migratory. To further explore this idea, Pearcy and Fisher analyzed the results from coded-wire tag (CWT) data (1977 - 1983) for coho from different coastal streams along the west coast of North America. The results from this study suggested that the recovery of juvenile coho salmon during their first summer in the ocean were in the general region of their ocean entry locations. For British Columbia coho for instance, Pearcy and Fisher found that 95% of the recoveries of marked fish released in BC waters were landed in this province; only 2% landed in Alaska and the remaining fish moved down to Puget Sound and Washington coast. There were, however, biases reported from this study. Apparently, there were differences in management practices among sampling regions, such as the minimum legal size, time of opening fishing seasons and fishing effort in each region. For instance, the legal size limit of coho that were allowed to be harvested was smaller in British Columbia than in Washington or California. This difference could result in more recoveries of juvenile coho in the former area than in the latter. Previous studies by Godfrey (1965) as well as Hartt and Dell (1986) have shown a relatively similar distribution for coho in the open ocean, with most North Pacific coho found in the inshore area (Figure. 2.2). Godfrey et al. (1975) stated that the offshore distribution of coho in the eastern part of North Pacific needs to be studied further since they found that coho distributions were concentrated towards the inshore areas. In addition, Healey (1993) believed that most coho remain in continental shelf waters throughout their ocean life. 18 Figure 2.2. Ocean distribution of British Columbia stocks of coho salmon in the Pacific Ocean. Shaded area indicates the coastal distribution where the majority of coho were found. The dotted line represents the boundaries of the distribution in the open ocean, (adapted from Healey, 1993) With regard to the Big Qualicum river fish, which are the particular focus of this study, it has not been common for fish from this river to make extensive migrations either to the north of Vancouver island or to the south by entering Washington state waters (L. Lapi, pers. coram). Therefore, the fish are usually found in the waters surrounding Vancouver Island. Moreover, Cross et al., (1990) found that over 90% of coho originating from the hatcheries in British Columbia are caught by Canadian fisheries. It has been known for some time, however, that the distribution of coho around Vancouver Island has fluctuated between inside and outside waters. 2.3. Shore-seeking tendency There is a tendency for juvenile coho salmon to remain close to the shoreline. Godfrey (1965) noted that in British Columbia, it is quite frequent to find coho in the shallow coastal waters off island beaches for the first several months in their ocean life. However, as the fish grow, they may move off into deeper parts of the ocean. Nevertheless, this one year old coho will still remain in the coastal area before making the decision whether to move out to the open ocean or to stay in the coastal waters. From a survey of Pacific salmon in the Strait of Georgia, Healey (1980) found that coho tend to stay in the area close to the shore of Vancouver Island. This tendency suggests the fish prefer to be close to a shore line, instead of going to the off-shore area. 20 2.4 Study Area The area of interest in this study covers the waters surrounding Vancouver Island in British Columbia. This area is bounded by longitudes 123°00' W and 129°00' W and by latitudes 48°00' N and 51°00' N shown in Figure 2.3. Vancouver Island is separated from the B C mainland by the Queen Charlotte Strait in the northwestern part and the Strait of Georgia and Juan de Fuca Strait in the southeastern part. Both parts are connected by two narrow and sinuous channels, Johnstone Strait and Discovery Passage (Thomson, 1980). The West side of Vancouver island is basically a coastal area which is part of the Pacific Ocean. In this study, releases of tagged juvenile coho from Big Qualicum hatchery were chosen for detailed investigation and modelling, to examine the dispersal pattern of juvenile coho salmon in the waters around Vancouver Island. 2.5. Hatchery production of coho Hatchery production of coho together with chinook salmon (O. tshawytscha) in British Columbia began in the late 1960's, but production on a routine basis dates from 1971, when Big Qualicum and Capilano hatcheries began operating (Cross et al., 1990). Many new facilities were built after 1977 when the government established the Salmonid Enhancement Program (SEP). Initially intended to double the production of salmon fisheries in British Columbia, SEP was designed to achieve this goal through the application of 3 major technologies: hatcheries, spawning channels and lake enrichment (Hilborn and Winton, 1993; Healey, 1993). In this section I will focus the discussion only on hatcheries, which were the principal technology for enhancing coho production. Figure 2.3. Geographical map of the study area 22 Currently there are more than one hundred hatcheries for coho in B C , many of which are small and run privately. However, most coho smolts are produced by large production hatcheries (Cross et al., 1990). The map on Figure 2.4 shows the distribution of the four production hatcheries in the areas of Vancouver island and the adjacent mainland. The facilities shown in this figure are estimated to produce over 1 million smolts per hatchery in 1990 (Cross et al., 1990). Most hatcheries enhance stocks from the watershed where they are located. Major hatcheries release primarily smolts (one year old) coho, but many also release fry. According to recent report (DFO, 1990), annual coho fry released from Strait of Georgia hatcheries increased from 800,000 in 1972 to 7.3 million in 1987. Coho hatcheries are considered an important asset for the fishing industry in British Columbia. Figure 2.5 shows a steady increase in number of juvenile coho released from hatcheries in the inside and outside waters of Vancouver Island. At the same time, the contribution of hatchery coho to the fisheries has increased markedly from 10% to 50% in the inside and 3% to 10% in the outside (Figure 2.6). The steepest increase was during the period of 1983 - 1987 and contribution decreased in the following year. 23 Figure 2.4. Location of major hatcheries for Coho Salmon (Adapted from: Cross et ai, 1990) 24 20 0 -I 1 1 1 1 1 1 1 1 1 1 1 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 Figure 2.5. Number of coho smolts released from hatcheries located in the inside and outside areas However, the apparent success of SEP has not been without its detractors (DFO, 1990; Hilborn, 1992; Healey, 1993; Hilborn and Winton, 1993). In general, critics have focused on 2 major disadvantages of SEP. The first is the marine survival rate. According to Hilborn and Winton (1993), the survival rate of hatchery coho salmon declined from 14.7 to 9.5% in the period of 1975 - 1985. 25 60 0 -I 1 1 1 1 1 1 1 1 1 1 1 1 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 Figure 2.6. Proportion of hatchery coho salmon, caught by trollers from the inside and outside areas Several possibilities have been postulated as the cause of this poor survival rate, including habitat loss, pollution, disease, low smolt quality from the hatchery and fishing activities (Pearcy, 1992; Hilborn and Winton, 1993). For Big Qualicum hatchery, DFO (1990) reported that it is likely the poor marine survival is due to declining smolt quality. The second and probably the most important criticism is that increased hatchery production is associated with the decline of wild stock populations. A recent report by DFO (1990) said that the contribution of wild coho in the total catch of this species has dropped substantially. It was reported that between 1976 to 1980 and 1985 to 1989, the average contribution of wild coho stocks to the sport and troll harvests in 26 the Strait of Georgia declined by 37% and 43%. Another example is from the Oregon Production Area (OPI). Pearcy (1992) reported that the contribution of wild coho to the total catch has shown a dramatic decline from 50% in 1960 to 10% in 1985. Hilborn and Winton (1993) hypothesized two key factors that cause the depletion of wild coho. The first factor is overfishing of wild coho populations. The initial success of hatchery-reared coho led to a large increase in the number of trollers fishing for coho. This build-up of fishing effort caused overharvesting of coho populations especially wild coho. Hilborn (1992) noted that, in the waters between Vancouver Island and the mainland, the harvest rates for coho salmon are very high, in some cases reaching as high as 95%. This harvest rate is sufficient to greatly reduce the wild coho stocks in this area. The second factor is man-induced habitat degradation. During their fresh water phase, coho salmon are highly dependent upon the quality of fresh water in their habitat. However, since their fresh water habitats are highly vulnerable to impacts from activities related to agriculture, domestic and industrial purposes, many of these habitats become degraded as rearing habitats for wild coho salmon. This situation will ultimately cause the decline of wild coho populations. As a result of these factors, Healey (1993) suggests that hatchery stocks have replaced the wild stocks in the fishery instead of supplementing them. This suggestion can be explained further by exploring the distribution pattern of both stocks. It is assumed that these two stocks have a similar distribution pattern, since they are still the same species. Furthermore, as a single species, these two stocks would have the same food preference and other environmental conditions that favour the achievement of higher survival rate. Since they are assumed to have similar 27 distributions, this situation may lead to a higher competition rate for the limited resources between these stocks. Even though the hatchery stock is believed to have lower survival rate than the wild stock, nevertheless, with the increased hatchery population, this population would have a significant impact on the competition of the resources with the wild stock. In this case, the high number of the hatchery stock could result in a less favourable situation for the wild stock, which may contribute to the depletion of this stock. A good knowledge of the distribution of coho salmon, including factors affecting the dispersal process, may therefore, be useful in solving this problem. It has been argued that coho distribution in the ocean is greatly affected by oceanographic conditions. The influence of surface currents on coho dispersal in the area surrounding Vancouver Island may be critical in determining the distribution pattern of Vancouver Island coho. The next chapter will discuss the oceanographic conditions, mainly the surface currents, of the study area. 3. Oceanographic Conditions 3.1. Geographical boundaries The geographical boundaries of the waters surrounding Vancouver Island are provided by the mainland area of British Columbia to the east and the Washington state coast to the south. There are four major bodies of water separating Vancouver island and the BC mainland. Queen Charlotte Strait and Johnstone Strait, including Discovery passages, in the northeastern part, the Strait of Georgia in the southeastern part and the Strait of Juan de Fuca in the South (Figure 3.1). The western part of the study area, which is the open Pacific Ocean, is limited to the continental shelf, known as the west coast of Vancouver island. The dimensions of the areas of concern are summarized in Table 3.1. Table 3.1. Dimensions of coastal waters surrounding Vancouver island. Locations Length (km) Width (km)1 Depth (m)1 Strait of Georgia 222 18.3 - 55 50 - 450 Juan de Fuca Strait 100 22-28 50 - 250 West coast of Vancouver Island 450 8-502 <200 Queen Charlotte Strait 90 13 -15 50 - 250 Johnstone Strait 100 3.5-4.5 50 - 400 Source: Thomson, R.E (1981); Tully and Dodimead (1957) 1 = maximum and minimum values 2 = continental shelf 28 29 30 3.2. General patterns of net surface-currents During the last decade or so, much has been written about the oceanography of the waters surrounding Vancouver Island. In particular, the physical oceanography of the Strait of Georgia, has been studied since 1930 (Le Blond, 1983). This chapter will discuss the pattern of net surface currents in the study area. Discussion will focus on the four main water bodies of the study area and the west coast of Vancouver Island (see Table 3.1). The description of the net surface-currents in these areas is taken from a number of sources, which are in turn compilations of information on the study area. Figures 3.2 to 3.5 show the diagrammatic pattern of the seasonal net surface-currents in the study area. These diagrams are the simplified maps of the study area, which are also used for the simulation model that will be discussed in chapter 5. 3.2.1. Strait of Georgia The Strait of Georgia lies on a northwest-southeast axis between Vancouver Island and the British Columbia mainland. Along the eastern side of the Strait, there are several inlets, sounds and passages, the result of repeated periods of glaciation. The western side, which is the eastern coast of Vancouver Island has a more regular coastline (Thomson, 1981). The oceanographic features of the Strait of Georgia have been studied intensively compared to other areas. The currents in the Strait of Georgia have shown a quite consistent cyclonic pattern driven by winds, tides and fresh water run-off (Waldichuk, 1957; Thomson, 1981). 31 Direction 4 i «- j « - ! « - ! «- «- «- t - r t 4 1 4 I «-!«-;«- 4 4 t 1 ' It , 1 : 4 t * / / ' • • J 4 4 t i t 4 t 4 4 t t 4 l t • ' ' , 4 . 4 t i t , 4 j t •4- ' ' • , 4 t x , 4 i l t 4 4 ; t 4 4 f 4 ! 4 t 4 4 t 4 : 4 It * i * i t 4 i t 4 4 i t 4 ! 4 t 4 4 t 4 4 t 4 * t * I* .t 4 : 4 t 4 4 t * ! 4 i t * t 4 I t 4 j t 4 t 4 ! t * i t 4 t 4 t 4 t x 4 i t :«-4 4 f t 4 4 i t t 4 4 t t 4 4 f t 4 4 t t 4 4 t t 4 4 t t 4 4 t t 4 i H t i t ' " t i X / / A • - * t X ' A Velocity 15 1 • 5 / 15 1 s 5 15 ! '0 ,0 15 j l . 10 IS I 1 0 10 15 | 10 10 15 I10 15 15 10, 15 15 ,0 ! 15 15 10 ; 20 15 ,0 | 20 15 10 i 20 15 10 ! 20 15 io i 25 15 10 | 25 15 ; 10 : 25 15 j 10 ; 25 15 : 10 ! i 25 15 ' 10 j » 15 | 10 I 30 15 j < s ! X JO 20 30 ; 20 JO 20 25 : so 30 : 20 • 10 I 20 i 15 ! 10 101 20 I ts i 10 ' « 20 15 ! 10 10 20 ; , j ] 10 10j 20 : 1 S ! 10 10 , 20 ! 15 ; 10 10 j 20 i IS ! 10 20 ; 20 j ts j 10 20 | 20 ; t s | 10 , J > A » i x i y / / / / / / 40 25: M r / A Directions Land x No Current Figure 3.2. Diagram of estimated net surface current direction and velocity from June -(summer period) 32 Direction Velocity 4 4- i 4- 1 ' 4-! 4-; 4-! 4-"—5 t • 4 !4- !•-!«-•«. ,' 4-\+-\ t t / t 4 -* "* ^ ~* 4 * t If ' y t : 4. t » y y 4 11 \ t t • ' / t -* t ' y 4 : 4 1 t jt ' / t • t / s ' y 4 4 t I f ' / t t 4- / ' y — 4 T X / t t t / ' i y 4 t ~} f ^ / t t t • ' y y 4 jt S ^ y ' y ' y ' . t t t • ' y y 4 j t / T t t ' y y mm— 4 ;! t S t : t ! t ' y y 4 j t y t ' t ! t y ' / y 4 ;t s . t : t t / ' / y 4 t t t / 4 i t t t t / * 4 ; t t t t • ' • X 4 : f j 4- y t T t / y / y 4 4 ; t ! t ' y t : t : t / 4 4 ) t ! t y t t t / 4 4 ! t i t y t ' t ; t / ' y ' y 4 4 ! f i r y t : t ; t / ' y ' s 4 4 t y T t r S y ' y 4 • 4 • t i t y t t r / ' y ' y 4 4 t ; ' y 4 t ; t / y / , 4 4 t ; t y -» 4 t / ' y / y 4 4- T ! t y y 4 : t / ' y ' • 4 7~? / t X y 4 : t 4- 4-1 4-; • - ' -> t • • > y y y t / y 10 j 10 ! 1 0 ! 10 j 10 j 10 1 1 0 | 10 10 ! 5 10 | i . 1 2 0 ! x l 20 ! 20 20 ; , o ' y 10 | , 0 i | 1 S l! 15 ! 15 10 | 10 15 I -• ' y 10 1 < 10 J I 10 I . ' r—: y t 5 15 | , 0 ' y 10 jio | 10 s s y l 5 i 1 0 15 i 10 ' / 10 s s ' 4 y 5 i " 15 ' / s ! 10 1 \ 5 s j / t 5 1 1 0 20 T x ' y t 5 10 j 5 s s y ' 3 y 1S 25 / r y 5 10 5 / s y ' 4 y IS 25 / ' y ' y * y ' y " y ' y ' y " / S 10 10 / s ' / y 15 25 / S 10 10 ' s • ' y  25 / s 10 10 J • ' > y 25 / 5 10 10 I / • ' y  ,5 25 / 5 10 ! 10 / • ' y • y " ! 1 5 25 s S 10 10 / • ' y • > ,5 25 4 / / 5 10 10 / • • / f 15 25 9 ,0 10 / • • X 15 15 y 10 10 10 / s y y 5 10 j 15 5 ' / ,0 10 15 / / ' y y 10 j 15 10 ' y 10 10 15 / / ' y y 5 10 j ,5 10 ' y 10 10 15 / / ' y y 10 10 I I 15 10 ' y 10 10 I 15 / / ' y y 10 10 ! 15 10 10 15 15 / / f y y 10 ,0 i 15 10 ' y 10 I 15 15 / s ' y / 10 10 j 15 10 -o ; 15 j 25 / s ' y 10 10 ! t 15 10 t 15 2 0 i 29 s ' y 10 , o j 15 10 / 25 25 / • y 25 25 X 25 20 25 25 25 25 [ 25 y y y y 25 25 » i 20 25 y f „ y * X r / x 5 i 5 • > Directions Land X No Current Figure 3.3. Diagram of estimated net surface current direction and velocity from October - November (fall transition) 33 Direction V e l o d f y 4 *~ 4- 1- ;«-•!«- • 4- i 4- | 4- ; 4- 4 15 i 1 ' 15 ; 15 ; 15 15 15 i , s ! 10 5 1 f 4- 1 4--;4- j 4-i 4- | 4- ! t ;t 10 ! 20 I " ! » • 20 20 20 i , S : 1 5 15 y t * -*|-*-»-» 1 3 » 4 ! t ' t it 4 10 j,0 j 20 i 20 20 20 1 ; , s l 1 5 15 * y t * t f 4 ; t ; t it 4 10 i 1 5 i 10 I 4 10 i 1 0 ; is 15 * y t t 4 : 4 t it t > 10 T i 1 0 110 S 4 / f 4 10 i i s IS y t t • | 4 ; t t / 10 j 10 • • 4 f ^ 10 I 10 ! , 5 15 f y ' t t t T 4- • S y -* 4 it X 10 | 10 i s • y 5 I 15 i 30 X y t t t / y 4 i t / 10 ! i o 10 4 / „ ' 7 20 y r I • / T / / / ' / 4 it / 10 10 10 • s / ' y 20 1 30 y y t t f s ' • 4 ! t s / / 10 I 10 10 s 4* / ' y 20 j 30 y / y t j t t • 4 : t / / / 10 10 • / s / y 20 i 30 / y t t t • ** • t / / / 10 ! i o 20 • s / ' y 20 ! 30 / y t t | t • • * it / * / 10 10 20 • s / ' y 20 J 1 1 30 * / y y t t t / / it ' / / 10 10 1 20 • / / ' y  J : 30 ' / y y t t t / y / + t ' J > 10 10 20 • s / ' / 20 1 i 30 ' / A* y t t t / 4 i t ' 4 10 ,0 20 / / ' y  I 3° ' / y t . t It / X 4 ; t 4- * 4> 10 10 20 • / / <— X 20 i 2 0 10 * y t t t / / / J 4 4 ' t t / 10 10 30 • y 10 10 j 20 10 y t t t / J + 4 | t t 4 10 10 30 • / / . 10 10 1 20 10 * y t t t / 4 4 i t t f / 10 10 30 / / J ,0 10 20 10 * y t t t / *f 4 t t f / 10 10 40 / / . y 10 10 r — 20 10 y t t t / / 4 - 4 t t f / 10 10 40 • / / . 10 10 25 10 y t t t / 4 4 t t t / 10 10 40 • / / . y 10 10 35  y t i t t / > 4 j 4 t : f / 10 30 • / / . y 10 15 25 10 t y * it i t / 4j 4 t i t t / 1 5 I 10 30 • / / . y 10 15 20 10 y -t * i t / *i 4- t j t .... y 20 3. | 30 • 4* S j y 10 15 20 10 y f j * j t / X —-> t / —7 t 30 X • 4* * , y — 1 25 X y t * i t , t 1 r ' / / y f 30 25 i 30 J 1 30 » I 25 25 y t / . y y 30 » l 30 30, 25 25! 25 / y S S y I t • • / y y x! 5 5 • / * y - Directions 0 Land X No Current Figure 3.4. Diagram of estimated net surface current direction and velocity from December to March (winter period) 34 Direction "5 f 4 4 - 4- I*- * - \ 4 - 4 - 4 - t t 4 4 4 - 4 - « - 4 - 4 4 t T f * 4 T s / s * s ' s 4 4 ; T t f > 4 -•> t / • / , 4 4 t t f y 4 t • / ' / ' / 4 4 t t f / 4 t 4 - / / • / , 4 t X f / 4 4 t / ' / • s y y 4 t —7 f / / y / 4 4 t / ' ' y ' ' S 4 t ' .y y 4 4 t y ' / ' ' y 4 t / 4 4 t y ' y] y 4 t yr '• y ' y ' / ' / / , > 4 1 4 !t y ' ' , t / * 4 t y ' • y 4 t / 4 4 t y • y y 4 t / 4 4 t y ' • , 4 t / 4 4 t y 4 t y f / 4 4 t y X 4 t 4 - t / 4 4 t y 4 4 t t t y 4 4 t y ' y 4 4 t t f y 4 4 t y f y 4 4 t t t y 4 4 t y x y / / — — 4 t t t y 4 4 t y ' y y / 4 4 t t t y 4 4 f y ' y 4 4 t t t y 4 4 t y 4 4 t t t y 4 4 t y ' y 4 4 t t t y -» 4 t y ' y 4 4 - t t t y y * 4 t y ' y 4 / / , > t X > * y y 4 t 4 - 4 - : 4 - «- t ""7 t y / t -* -» -> **:-• -+ t • / y y J / i X -» t / y Velocity 10 15 nr 15 10 10 10 , 0 * t 10 10 10 10 10 10 10 10 , 0 5 y 10 , 0 10 10 10 10 , 0 15 10 * t y 10 , 0 5 / t y / y y 10 15 10 5 f y 10 10 S y A y , J 10 10 5 t y 10 s y y ' y y / A 10 15 10 j 5 t y 10 5 s y ' y 10 15 X * y 10 5 10 y / y , '—* y 25 15 y y 10 ' 10 10 y y ' y ' y J 25 15 ' y y y y 10 10 10 y y ' y y , 1 5 y 10 10 10 y y ' y y , 25 IS ' y ' y y , y 10 10 10 y y ' y\ y s 25 15 y 10 10 10 y y ' y y / 25 15 y 10 10 10 y y ' y y s 25 15 y , y 10 10 10 y y ' y 25 15 y * y 10 10 10 y y y y y 25 15 y f y 10 10 20 y y y X 25 10 5 t y 10 10 20 y y y y 5 15 10 5 t y 10 10 20 y ' y y , 15 10 5 t y 10 10 20 y ' y y / 5 15 10 5 t y 10 10 20 y ' y / / 10 15 10 10 t y 10 10 20 y y / 10 15 10 10 t y 10 10 20 y ' y ' , 10 X 10 10 t y 10 10 30 y ' y 10 X 10 10 t y 10 10 30 y ' y y , 10 X 15 . 10 9 y 15 1 5 X y ' y ' y 10 X 1 5 10 t y —7 25 X y y y X y j / 25 X y 25 X X X X X 25 25 / * y 25 X 25 20 X 25 y y y / t X 5 5 y y • y Directions Land X No Current Figure 3.5. Diagram of estimated net surface current direction and velocity from April to May (spring transition) Wind is an important element for the surface currents in this region. As Tully and Dodimead (1957) reported, the effects of strong winds in the Strait of Georgia are more obvious in the surface water transport. Furthermore, Le Blond (1983) pointed out that the winds that blow over this Strait, tend to push the upper layer of water mass at the same direction as the wind, which is along the southeast -northwest axis of the strait. The wind-driven currents in the Strait of Georgia, when coupled with other factors such as tides and fresh water run-off, are then believed to establish the general circulation pattern in this area. The tides in the Strait of Georgia are forced by those in the Pacific Ocean which enter through both openings, Johnstone Strait to the north and Juan de Fuca Strait in the south (Fig. 3.1). This suggests that currents in the northern and southern parts of the Strait of Georgia are strongly influenced by tides. LeBlond (1983) argued that, due to the relatively small cross-sectional area of the northern passages, sea-level variations inside the Strait of Georgia are mostly due to the inflow via the Juan de Fuca Strait (see Fig. 3.1). The Fraser River run-off plays an important role in the surface current pattern in the Strait of Georgia (Pickard, 1956; Waldichuk, 1958; Thomson, 1981; Le Blond, 1983). This run-off is high during April through September, with its peak in June. Thomson (1981) stated that the importance of the run-off can be divided into two aspects. Firstly, the river flow is directed southwesterly toward the Gulf Islands due to its momentum. Secondly, as a source of light surface water, the fresh water from the Fraser River creates a strong near-surface stratification which enhances the wind-generated surface currents. Based on the observed currents, Thomson (1981) summarized the average current pattern in the Strait of Georgia as follows. 36 For all seasons, there seems to be an anti-clockwise, cyclonic, circulation of the surface water in the Strait of Georgia. During the months of June and July, the current velocity on the Vancouver Island side is expected to be greater than the velocity on the mainland side due to the wind force. On the Vancouver Island side, the near-shore currents from the vicinity of Comox down to Haro Strait are estimated to have a net velocity of 25 - 30 cm/sec to the southeast. Further off-shore on the Vancouver Island side, the net current velocity is slower and ranges from 10-15 cm/sec to the southeast. This lower velocity may be caused by the Fraser River during its peak run-off season (Thomson, 1981). The large fresh water volume that comes out from the River mouth flows west towards the Vancouver Island side. However, this fresh water outflow is turned northwest due to the tidal currents and the Coriolis force. This northward flow of the Fraser River plume is believed to restrict to the southeast movements of water off Vancouver Island. On the mainland side, the net surface water flows at an average speed of 15 - 20 cm/sec to the northwest, except at the area of San Juan Island and Rosario Strait, where the currents are estimated at about 10-20 cm/sec to the north. In the months from October through March, the influence of Fraser River run-off is low. The wind is believed to have more important effects in producing wind-driven currents especially at the surface (Waldichuk, 1957). The southeasterly wind, combined with the tidal currents, tends to push the water on the mainland side causing higher velocities compared to Vancouver Island side. In the southern part of the mainland side, the net current is estimated at 20 - 30 cm/sec to the north west, whereas in the northern part (i.e. before the entrance of Johnstone Strait), the net current velocity decreases to 15 - 20 cm/sec to the northwest (Thomson, 1981). On the Vancouver Island side, the net currents during fall and winter seasons are relatively stable at 15 -37 20 cm/sec to the southeast, except for the northern part of the strait, which has net current velocity of 10 cm/sec to the southeast. 3.2.2. Juan de Fuca Strait As in the Strait of Georgia, there are numerous mechanisms affecting the surface currents in the Strait of Juan de Fuca, including tides, winds and river run-off (Thomson, 1981). Herlinveaux (1954) and later Thomson (1981) concluded that the currents acting in the Strait of Juan de Fuca are mainly tidal currents, which in some locations may reach a velocity of up to 3 m/s. Other, non-tidal currents tend to move seaward toward the open ocean along the northern margin of the Strait. These currents have seasonal fluctuations following the amount of fresh water coming out from the rivers in the Strait of Georgia. The main source however, is the Fraser river. These currents are considered to be the source currents for the Vancouver Island Coastal Current (VICC) that flows northward along the west coast of the Vancouver Island. Cannon (1978) noted that the speed of the net surface current in the Strait of Juan de Fuca fluctuates seasonally. In fall and winter, when the southeasterly wind blows, the net surface current's speed, especially in the Northern part of the Strait, is about 25 - 30 cm/s. With regard to VICC, Thomson et al, (1989) reported that its velocity is low in winter due to less fresh water from the Fraser River. During the spring and summer, the surface circulation in the Strait of Juan de Fuca becomes more complicated. VICC is stronger and, coupled with wind and tidal currents, produce a net surface current that flows westward in the northern side of the strait. A compensating flow to the east develops in the southern part of the strait (Holbrook et al.,1980). 38 Furthermore, Thomson (1981) and Thomson et al, (1989), estimated that on the northern margin, the net current speed in the strait of Juan de Fuca, in early summer, is about 30 to 40 cm/sec to the west. 3.2.3. Johnstone Strait and Queen Charlotte Strait Among the major water bodies of the study area, Johnstone Strait and Queen Charlotte Strait are the least intensively studied. Oceanographic observations were usually conducted together with fisheries research. A few reports are available, however, regarding the oceanographic properties of these areas. In this discussion, I shall often refer to the reports of Thomson (1976, 1977). The net currents in these straits show a 2-way directional pattern. On the Vancouver Island side, the currents tend to flow eastward whereas on the mainland side the currents flow westward toward the open ocean (Thomson, 1981). This suggests an anti-clockwise circulation of the water in these straits (Thomson, 1976, 1977). The water body of Johnstone Strait, including Discovery Passage, is the narrowest of all the major channels in British Columbia (Thomson, 1981). Due to the relative narrowness of Johnstone Strait, the currents in this area are swift and rectilinear and can exceed 1 m/s. The net surface currents range from 25 to 30 cm/sec, depending on the wind direction and strength. In the summer, the surface current on the Vancouver Island side would have greater velocity due to the prevailing northwesterly wind which causes the surface water to move faster to the southeast. On the other hand, during winter, winds are primarily from the southeast, and therefore, the net surface current on the mainland side 39 would have the faster flow to the northwest. Queen Charlotte Strait is approximately 90 km long and has a width ranging from 13 km at its eastern end to more than 26 km at its midlength, then narrows to less than 15 km at its western end. As in the other straits, the wind strength and direction play an important part in determining the strength of the currents in Queen Charlotte Strait. During May through August, the prevailing northwesterly wind results in a strong southeasterly current on the Vancouver Island side, with an estimated speed of 30 cm/sec. On the mainland side, the current is estimated as 15 - 20 cm/sec to the northwest. On the Vancouver Island side, the current is stronger, with estimated range of 25 - 30 cm/sec. The southwesterly wind has its influence during fall and winter seasons, causing stronger currents on the mainland side with net current velocity at about 30 cm/sec to the northwest and weaker currents on the Vancouver Island side, with velocity ranging from 10-20 cm/sec to the southeast. The part facing the open ocean still maintains the current at speed 25 - 30 cm/sec to the west. 3.2.4. West coast of Vancouver Island The coastal area on the western part of Vancouver Island is characterized by a continental shelf of varying width. The continental shelf is approximately 25 km wide in the south and becomes narrower toward the north. The region off Brook's Peninsula in the northern part of Vancouver Island, has the narrowest continental shelf, which is about 8 km wide. The steep topography in this area drops to over 1000 m deep within about 30 km of the coast. The surface currents on the inner continental shelf on the west coast of Vancouver Island is known locally as the Vancouver Island Coastal Current. This buoyancy-driven current is persistent throughout the year and is approximately 15 to 25 km wide. This current flows northward over the shelf with a speed of 40 to 50 cm/s during winter near the entrance to Juan de Fuca Strait (Hickey et al. 1991). However, as the current moves northward, the speed decreases to about 15 to 20 cm/s about half way up Vancouver island. Thomson et al., (1989) estimated that during summer, the current in the vicinity of La Perouse bank is about 30 to 40 cm/s. The current normally extends to the bottom, and its core is centered along the 50-m isobath (Thomson etal. 1989). Run-off is credited as being the primary forcing mechanism for this current. The maximum and minimum inputs from estuarine discharge occur at different times of the year, with a high discharge during the freshet period in spring and early summer and low in winter. This condition suggests an annual fluctuation in the coastal current flow with a high in spring - summer and low in winter and fall (Freeland et al., 1984; Le Blond et al., 1986). In addition, Thomson et al. (1989) described seasonal fluctuations in the seaward extent of the coastal current in response to large scale circulation pattern changes. In winter, the current tends to remain close to the coast; whereas in summer, the current appears to move offshore, especially in the southern part. On the west side of the Vancouver Island, there is a southward water movement on the outer continental shelf during the Summer season. This current, known as the Shelf Break current, develops near the shelf break in May and is believed to be a response to a shift to Northerly winds in summer (Freeland et al. 1984). The location of this current is thought to be just above the continental shelf break at the 200 m isobath. However, it might get closer to the coast, especially in the narrow continental shelf regions. This current has an average speed of 15 -20 cm/sec to the southeast. At the northern end of the Vancouver Island, there is a flow westward and then southeastward past Cape Scott. This coastal flow is then deflected seaward as is approaches Brook Peninsula. The velocity at this point is approximately 15-20 cm/sec. The Vancouver Island Coastal Current, approaching from the south, is also believed to be deflected by Brooks Peninsula. The net circulation pattern, assumed to be the best description of surface water movements in the study area, may be summarized as follows. The general pattern of net surface currents comprises several circulation cells that each has a cyclonic pattern (anti-clockwise). The waters between Vancouver Island and the mainland have three of these cells, starting from the Queen Charlotte Strait in the north, the Johnstone Strait and Discovery Passage in the middle, and the Strait of Georgia in the south. In all of these regions, the current direction is to the south on the Vancouver Island side and to the north on the mainland side. In the Juan de Fuca Strait, the water tends to move seaward on the northern part of the Strait, which is south of the Vancouver Island. A reverse direction is assumed to take place in the southern part of this Strait. Although this is the area with less available information with regard to the surface currents, the northern part of Vancouver Island, including the Queen Charlotte Strait, is assumed to have water movement that flows seaward (i.e. westward) and then southeastward, as it enters the open ocean to join the shelf break current that also moves to the southeast. In addition, during the months of October through March, the surface currents at the northern tip of 42 Vancouver Island, are believed to move landward, (eastward) toward the Queen Charlotte and Johnstone Straits. A fourth circulation cell is found on the west coast of Vancouver Island during spring and summer. Vancouver Island Coastal Current, moves northwestward along the coast. During April through September, this current, as it approaches Brook's Peninsula, gets deflected to the west and then turned southeastward with the shelf break current coming from the north. However, during the Fall and Winter months (i.e. October to March), when winds from the southeast are strong, the southeastward currents along the shelf break is replaced by a net northwestward movement. This complex pattern of surface currents is a major feature of the physical environment of coho salmon. The model to simulate the interactions of dispersal behaviour of the fish with these currents is described in the next chapter. 4. The Model 4.1. Introduction In this chapter I will discuss the structure of the simulation model that explores the dispersal of coho salmon from the Strait of Georgia. The model is designed to test the effects of surface currents on the dispersal and hence to examine the distribution of the fish after one year of ocean life. To examine the dispersal pattern, I have used the case of juvenile coho released from Big Qualicum hatchery as they enter the ocean. The model, written in Q-Basic (Microsoft Inc., 1992), is spatially organized as a number of compartments (boxes) representing the study area. In general, the model can be divided into 3 main parts (Figure 4.1). In the subsequent sections these parts will be discussed in more detail. 4.2. Structure of the Model 4.2.1 Input Data The first part is the Input submodel that consists of two components. The first component is the map, in which a simple geographical map is used to represent the water surrounding Vancouver island including the Strait of Georgia, Juan de Fuca Strait, Johnstone Strait and the coastal water west of Vancouver island. This is the habitat in which the dispersal of coho will be modelled. The area is divided into 310 such habitat boxes. Each has a dimension of about 5 x 5 km (Figure 4.2). The fish, represented by particles, were released from Big Qualicum hatchery into the habitat box, shown as point P. Movements of the particles were restricted to the water area. The second component is the best-estimate of surface current data for each habitat 43 44 Map D a t a — I N P U T I B l a s j j ! Part Seasonal Currents Fish Grow Calculate Probability Probability and Ace. tables (Summer) Particle dlstr. (summer) Probability and Acc. tables (Fall) Particle dlstr. (Fall) Probability and Acc. tables (winter) Particle dlstr. (winter) O U T P U T One year distribution Probability and Ace. tables (spring) 7 Particle dlstr. (spring) II III Figure 4.1. Structure of the simulation model, consisting of 3 parts; Input Data, Movement Probabilities and Particle Distributions. 45 West coast of Vancouver Island / S f t\ V" Land Water P Releasing point SCTR * South Central Troll *) GSTR » Georgia Strait Troll *) SWTR = South West Troll *) NWTR * North West Troll *) .*) see explanation in text I NWTR "7—7 7\ S / SWTR SCTR ' ' , . s i I ! GSTR Queen Charlotte Strait Johnstone Strait Georgia Strait Juan de Fuca Strait Figure 4.2. Diagram map of the study area showing the grid habitat boxes (total = 310) Main water bodies includes Queen Charlotte Strait, Johnstone Strait. Strait of Georgia Juan de Fuca Strait, and the west coast of Vancouver Island. Salmon tags recovery areas (SCTR, GSTR. SWTR and NWTR), were also shown 46 box. These values were obtained from surface current estimation, explained in the previous chapter. Since the current speed may be the same in some areas, several habitat boxes will have the same current estimation values. 4.2.2. Movement Probability (i) Pure Random Movement The second part of the model is the movement probability that determines the movement of particles (fish) in certain directions. This probability is first calculated for each particle by assuming that particles move in a purely random way (i.e. no biased probability). Figure 4.3 shows the diagram of a 3 X 3 system of habitat boxes that can be used to illustrate the probability that a particle in one box will move to any adjacent box if pure random dispersal is the only factor affecting its movement. Suppose a particle is in box #5 and is free to move to adjacent boxes. Pure random walk is applied to calculate the movement probability by considering the distance from box #5 to the center of all adjacent boxes. Since there is an unequal distance between diagonal and two horizontal movements, the probability of moving diagonally is different from that of moving horizontally. Assuming that the distance to move horizontally is B and the distance to move diagonally is b, then the probability for a particle to move from box #5 to box #2 is proportional to 1/a and, the probability to move from box #5 to box #1 is proportional to 1/b. If a is 1, then b is if 2. Consequently, the probabilities for particles to move from box #5 to adjacent boxes are: 47 Figure 4.3. Diagram of 3 x 3 system, showing the illustration of particle movements in 9 habitat boxes • — — 1 Table 4.1. Probability table for particle moving from box #5 to all adjacent boxes Box# 1. 2 3 4 5 6 7 8 9 T, Prob 0.104 0.146 0.104 0.146 0.000 0.146 0.104 0.146 0.104 1.000 The same procedures of probability calculation is applied for other boxes over the entire study area. Note that the probability in box #5 is zero, indicating that particles would have to move to adjacent boxes. The program also takes into account the possibility that some of the habitat boxes are considered as "land". In this case, the probability for such boxes will be assigned as zero. For example, if box # 1, #4 and #7 in Figure 4.3 are assigned as land, the probability table in Table 4.1 will be changed to the values put in Table 4.2. Table 4.2. Probability table for particle moving from box #5 given that box #1, #4 and #7 are land. Box# 1 2 3 4 5 6 7 8 9 Prob. 0.000 0.227 0.160 0.000 0.000 0.227 0.000 0.227 0.160 (ii) Biased Random Movement The probability description above dealt with pure random movement. In order to take into account the influence of currents on particle movement, a bias in the direction of movement induced by the current is applied. Bias in the direction of movement (induced by currents) is 49 calculated by taking the ratio between current speed and swimming speed following the formula: Bias = V c/V s (4.1) where: V c = current speed (cm sec"1) V s = swimming speed (cm sec"1) The data on swimming speed takes into account the relationship between growth rate and swimming capability of the fish. As the fish increases in size, its swimming capability also increases and the influence of currents on fish movement decreases as indicated by equation 4.1 above. The assumption I used for growth rate is that the fish grow linearly during their 13 months of ocean life. The smolts are assumed to enter the ocean at 10 cm of body length with an increment of 0.11 cm per day during the 13 months period. The relationship between swimming speed and body size follows the equation: V s = BL * C (4.2) where: V s = Swimming speed (cm / sec ) B L = Body Length (cm) C = constant To introduce the bias on the probability, the calculated probability values from pure random movements will have to be changed according to the direction of the bias. For example, a 10% bias to the north means the total probabilities of moving southward will be reduced 10%, and this amount will be added evenly to the probabilities of moving northward, including the northeast and northwest boxes. Similarly, a 10% bias to the west means a reduction of 10% probability of moving eastward. Since the bias introduced is related to the seasonal surface currents, the results will vary both spatially and seasonally. 50 Another bias that was applied is the shore-seeking tendency. This bias is to take into account an apparent tendency of the fish to seek shorelines. In this procedure, the position of each particle is the criterion that decides whether a particle will experience this bias. If a particle is two habitat boxes away from land then a constant small bias will be applied to the probability calculation so that the particle will have a tendency to move toward the closest land. The procedure to calculate the shore-seeking tendency is the same as the procedure to calculate north-south bias, by subtracting the bias from the probability of moving to boxes away from shore and evenly distributing the "extra" amount of probability in the boxes closest adjacent land. Having calculated all the probabilities, the program will store the results in a set of libraries for movement procedure. These libraries provide the probability of movement maps for particle, in any habitat box. Each library represents the probabilities associated with a specific set of habitat and behaviour conditions (eg. currents, swimming speed and shore-seeking tendency). The procedure used to move a particle from one box to another involves comparing cumulative probability values for all possible movement choices by a particle with random numbers generated by the Q-Basic program. Table 4.3 shows the cumulative probabilities, taken from the results in Table 4.1, for movement from box #5. For each iteration of movement for each particle a random number between 0 and 1.0 is chosen. If, for instance, the random generator gives a value of 0.35, then the result is that the particle will move to box #3. If on the other hand, the random number generator gives a value of 0.68, then the result is the particle will move to box #7. 51 Table 4.3. Cumulative probability table for particle moving from box #5 to any adjacent box in numerical order Box# 1 2 3 4 5 6 7 8 9 I*) 0.104 0.250 0.354 0.500 0.000 0.646 0.750 0.896 1.000 Prob. 0.104 0.146 0.104 0.146 0.000 0.146 0.104 0.146 0.104 Range s 0.104 0.105 -0.250 0.251 -0.354 0.355 -0.500 - 0.501 -0.646 0.645 -0.750 0.751 -0.896 0.897 -1.000 *) : cumulative probability **): Interval of random number that assigns particle to move to correspondence box 4.2.3. Particle distribution The third part of the model is the final distribution of particles after one year of ocean life. After running the program for 13 months, it is terminated and the distribution of particles is counted within the major fishing regions, displayed in Figure 4.1. These regions have been used as pooling areas for the simulated particles after the 13 months of simulation period. A l l regions are associated with the troll fishing activities taking place around the Vancouver Island. Details of these regions will be discussed in the next chapter. There are two reasons for setting the limit up to 13 months. Firstly, maturing coho salmon that were released one year earlier, are now subject to the troll fishery. Since some of the tagged fish may also be caught by fishermen, this activity will reduce the number of tagged fish available for tag-recovery analysis. Secondly, there are some stocks of coho salmon that start their spawning migration after spending a little more than one year in the ocean (Sandercock, 1992). Therefore, movements of this species after one year in the ocean become more oriented toward specific directions. After 13 months, therefore, it becomes progressively more unlikely that random movement and surface currents alone, could 52 account for coho dispersal. 4.3. Number of iterations Particles are required to move on every iteration and each iteration is equivalent to the length of time required for the fish to travel the distance between habitat boxes. The duration of the total number of iterations is then equal to the period during which the fish disperse in the real situation (ie. 13 months). Several assumptions were used to calculate the number of iterations for 13 months of program running. (1) For each iteration, a particle can only move from one box to the next box. The distance between box centres is 5 km. (2) Fish enter the ocean at a length of 10 cm. Growth rate of the fish is 1.1 mm/ day during 13 months. The growth increment is calculated every month. (3) Swimming speed is calculated as a function of body length, but varies seasonally. In first summer swimming speed is 1/2 body length (BL)/ sec. In fall and winter, swimming speed is 1/4 B L / sec. The next spring, swimming speed is 1/2 B L / sec. and the final summer in the ocean, the fish swim at the 1 B L / sec. Fish swim 12 hrs/ day. (4) In the model, swimming speed is converted from B L / sec to cm/ sec. Using the values of swimming speed (V) combined with the total period of time (t), the total distance (Sp) can be calculated as: Sp = V * t V = speed (cm/ sec); t = time (second); Distance in cm is converted to km. The next step is to convert the total distance swum (column 6 in Table 4.3.) into number of iterations by using the equation: N = S p/S i where: N = number of iterations Sp = total distance Sj = 5 km (i.e. distance of iteration) Summary of relevant data on size, swimming speed and iterations per season are given in Table 4.4. The total of 451 iterations which is assumed to represent 13 months in the real situation, is then used to run the simulation program. The results are evaluated based on the classification of Troll Fishery regions in Table 5.1. 54 Table 4.4. Season, coho size, swimming speed, distance swim and iteration of the model for standard runs of the model. Seasons Month Total days Body Length (BL) (cm) Standard Swimming Speed' (BL/sec) Total Distance Covered (km)/ Iterations # (Standard Swimming Speed) Summer Jun Jul Aug Sep 122 10 13.3 16.6 19.9 1/2 BL/sec 395 79 Fall Oct Nov 61 23.2 26.5 1/4 BL / sec 165 33 Winter Dec Jan Feb Mar 121 29.8 33.1 36.4 39.7 1/4 BL / sec 455 91 Spring Apr May 61 43 46.3 1/2 BL / sec 590 118 Summer Jun 30 49.6 1 BL/sec 650 130 total : 451 1 4.4. Simulation Scenarios The simulation was run under different scenarios. Each scenario represents a different surface current regime in the study area or swimming speed of the fish. The standard values of surface current are based on the "best estimation" of the real surface currents provided in chapter 3. This is called the default scenario. Two other scenarios were established by reducing and increasing the current speeds by 50%. These two scenarios will be referred as 1/2 Default and 2 Default, respectively. A fourth scenario was established to test the results of the model when the swimming speed was doubled. Doubling the swimming speed requires more iterations, since the 55 iteration numbers are a direct function of swimming speeds. In the last scenario the surface currents were set to zero. In this scenario particles disperse randomly without any bias in their movements. Note, however, that current directions are the same for each scenario except for the last. The program was run 15 times for each scenario. The results are plotted as average values (in percent) ± standard deviation (X ± SD). The results of the simulations are compared to the observed distribution of coho. The observed distribution was derived from recovery of tagged Big Qualicum coho in the troll fishery. The next chapter will discuss in more detail the fisheries conditions in the study area as well as the tagging data. 5. Mark Recovery Program and Coho Fisheries 5.1. Introduction In 1977, in connection with the much expanded hatchery program under SEP (Salmonid Enhancement Program), DFO began an extensive program of tagging coho smolts released from hatcheries. The program was intended to provide data on the contribution of hatchery coho to the fisheries. The core of this program is the Mark Recovery Program (MRP) which samples net fisheries to recover tagged fish in the Strait of Georgia. The Mark Recovery Program database, developed by the Canadian Department of Fisheries and Ocean (DFO), has been used for several studies on the distribution and migration pattern of Pacific salmonids ( Quinn and Fresh 1984; Labelle, 1992; DFO, 1989). These works demonstrate the utility of this database to study the distribution of salmonids over space and time as early as the juvenile stage. Since the MRP database is dependent on the recovery of tagged fish from the catch by commercial fishermen or anglers, and only a portion of the catch is sampled, the distribution of catch and fishing effort in a particular fishery is critical in interpreting the overall recovery estimates. For this study, the MRP database and its estimates of tags recovered (Kuhn et al., 1988) was used as a source of data from which to determine the ocean distribution of coho salmon released from Big Qualicum hatchery. The following sections will describe the main fishing activity for coho including its catch and effort trends. 56 57 In this analysis the July data from the MRP database and troll fishery were used to examine the distribution of tagged fish. According to Argue et al. (1987) the opening season for commercial troll fisheries for coho salmon has been standardized to July 1 each year, except for the South Central Troll Region in the northern part of Vancouver Island, where the season is usually opened two days earlier, June 28th. Figure 5.1 shows that the July catch is the predominant catch from the period of 1975 to 1992 and contribute more than 50 % of the total catch from each region. It is apparent that after July, catch declines each month until the fishery is closed in September. It should be noted however, that although July 1 is the official opening day for troll season, obviously, there were catch recorded prior to the opening day. This can be seen by the presence of catches in June that were found in the MRP database. The percentage of catch during this month is very low, indicating that these catches could be considered as bias or error associated with the database. There were no explanations could be presented to explain the presence of the data in this month. Another reason for using the July data is associated with the life history of coho salmon which usually live in ocean waters for 1.5 years (Sandercock, 1991). The spawning migration however, generally starts after the fish spend about 15 months in their ocean life. After that time, the movement of the fish is considered directed and the influence of environmental factors is less crucial. Since the objective of this study is to examine the ocean distribution of coho salmon as influenced by the ocean currents, I chose to exclude the period during which directed swimming occurs. 58 100 SCTR GSTR SWTR NWTR Figure 5.1. Troll Average monthly catch from June to September in 4 recovery regions. Data were taken from 1975 -1992. 5.2. Statistical Areas Currently, over 30 statistical areas are defined by the Department of Fisheries and Ocean (DFO) along the B.C. coast to facilitate management of salmon fisheries in this province (Figure 5.2.). Catch and effort data have been collected routinely from each statistical area through the fish-sales-slip system (Healey, 1993). Kuhn et al. (1988) noted that commercial fishing vessels often land fish caught in several adjacent statistical areas. This creates difficulties in sampling at the landing sites, since the catch from different areas is often mixed. To minimize the difficulties, catch and sampling statistics are usually associated with catch regions, which are the combinations of several adjacent statistical areas and time strata for the same fishing gear. To distinguish the spatial distribution of Big 59 Figure 5.2. Fisheries statistical areas in British Columbia, where catch and effort data are classified. Bold lines indicate four recovery regions of coho salmon caught by trollers in the areas around the Vancouver Island. 60 Qualicum coho, 4 different catch regions have been used to represent the area surrounding Vancouver Island. These catch regions have been included in the study area, as explained in chapter 3. Table 5.1 summarizes the catch regions used in this study, with the associated statistical areas. These regions can be further divided into the outside area which consists of SWTR and NWTR, and the inside area which consists of SCTR and GSTR. Illustrations of these regions are provided in figures 4.2 and 5.2. Table 5.1. Classification of Catch Regions used in the study Catch regions Statistical areas Geographical locations South Central Troll (SCTR) 10, 11 and 12 Johnstone Strait, Queen Charlotte Strait Strait of Georgia Troll (GSTR) 13, 14, 15, 16, 17, 18 and all sub areas in 29 Strait of Georgia Juan de Fuca Strait Southwest Vancouver Island Troll (SWTR) 21,22, 23 and 24 Southern part of West coast of Vancouver isl. Northwest Vancouver Island Troll (NWTR) 25, 26 and 27 Northern part of West coast of Vancouver isl. 5.3. Troll Fisheries The troll fishery in British Columbia is one of the primary fisheries in this province, especially for chinook {O. tshawytscha) and coho salmon (Argue et al., 1987, Healey, 1986). Troll vessels have evolved from the traditional hand lining vessels into highly sophisticated mechanized vessels. In British Columbia, trollers vary greatly in size, although the majority are between 6 -15 m in overall length. Most troll vessels have an average length between 11 - 14 m 61 and carry six trolling lines run by hydraulic power. Vessels operating on the west coast of Vancouver Island are larger than the vessel operating in the Strait of Georgia. The average troll vessels in the Strait of Georgia is about 10 - 11 m in overall length. Trollers are mobile and, therefore, have a wide range of distribution and can pursue fish over a relatively large geographic area (Healey, 1986). Boats that preserve their catch in ice normally stay at sea 10-12 days between landings whereas boats that freeze their catch usually deliver their catch after 60 days or more. Trollers fishing in the Strait of Georgia land their catch each day (Ron Kadowaki, Pers. Comm.). Regarding the fishing grounds for the troll fishery, there are two major areas for the BC trollers to fish. The first is the west coast of Vancouver Island and the second is the Strait of Georgia. Figure 5.3 shows the catch trend from 1975 to 1992 for the two areas, respectively called outside and inside. It is obvious that the majority of the catch is off the west coast of Vancouver Island. In this graph, it is apparent that the catch in the west coast side was several times higher than that in the Strait of Georgia. 62 Figure 5.3. Catch proportion from troll fishery, between inside and outside areas (data from 1975 -1992) Coho salmon usually dominates the catch of trollers until approximately the end of August, except in years when pink and sockeye salmon are abundant. In those years, pink and sockeye salmon dominate the catch from late July to late August (Argue et al., 1987). Healey (1993) reported that based on the 1981 - 1985 production data, trollers caught more than 65% of total coho landings in British Columbia. This is followed by Sport, Net and finally native fisheries. 63 The management of troll fishery in British Columbia has changed over the years. Between 1966 - 1979, the opening for the coho fishery was from July 1 to September 30 in the Strait of Georgia, but extended to October 31 for the rest of southern British Columbia (Argue et al, 1987). In 1981, coho opening and closing dates were standardized in all areas from July 1 to September 30. In the same year, all trolling vessels were restricted to six fishing lines whereas formerly they were permitted to fish with eight lines. Nineteen eighty-one was also marked with the implementation of the new license system called "two-area" troll regulation. Within this system, fishermen have to decide either to fish in the Strait of Georgia or in the rest of the coast. However, there is one area in the Johnstone Strait, called the overlapping area, where vessels with both licenses share the same fishing ground. In 1985, the US and Canada ratified the Pacific Salmon treaty which placed the quota for coho in the west coast catch at 1.75 million fish. In the Strait of Georgia, troll fisheries also coexist with the sport fisheries, in which a relatively higher number of coho were harvested. The sport fishery is the second largest type of fishing activity for coho after the commercial troll fishery. In contrast to the troll fishery, harvest of the sport fishery is centered in the Strait of Georgia (Argue et al, 1987). There are some places on the west coast of Vancouver Island, where sport fishing occurs, but the catch is small. In the Strait of Georgia, it is significant that the sport catch is always higher than that of the troll fishery between 1981 -1992 (Figure 5.4). The 1980 data was used for the sport fishing since there were no appreciable data prior to that year. 64 3.5 Figure 5.4. Sport and Troll Catch in the Strait of Georgia (1981 - 1992) Due to the limited spatial distribution of the sport fishing activities, tag returns from this fishery were not included in this study. It is quite apparent that the M R P data for sport fishery could only be applied in limited areas, mainly the Strait of Georgia. This situation may have impact on the analysis of tag returns in all regions, especially the distribution in GSTR. The number of tag returns, available for the troll fishery in all regions, especially in GSTR, could be reduced, because of the presence of sport fishery. Hence, the estimated tag recoveries for the troll fishery could be underestimated. 5.3.1. Catch and effort distribution of troll fishery Time series data of catch and effort are available from the Pacific Biological Station (PBS) database in Nanaimo. Data from 1975-1992 were chosen for analysis for the following reasons. First, 1975 marked the first year of coded wire tag (CWT) recoveries conducted by the Department of Fisheries and Ocean, Canada. Having the complete data of the coho recovery program will provide the best overall picture of the distribution of tag recoveries of hatchery fish. Secondly, in 1981, the government introduced the "two-area" troll regulation which affects the distribution of troll vessels in the inside and outside waters at the same time. The implementation of this regulation may change the number of tagged coho caught by troll vessels, because of the reduced number of vessels in the regions. However, it is less likely that it may affect the spatial distribution of fishing effort, since the trollers can still fish at the same location (i) Catch Coho salmon catches in the inside and outside areas show a considerable difference. The average catch in the outside area is 3 - 6 times greater than that in the inside (Figure 5.5). The annual average catch for both SCTR and GSTR is around one hundred thousand fish. SWTR has the highest mean catch of 600 thousand fish / year and NWTR has the average of 300 thousand fish / year. Catch variations in the inside are relatively higher than those in the outside area. The measurement of Coefficient of Variation (CV = (Standard Deviation / Mean) * 100%) for the inside regions shows that C V for SCTR and GSTR are 50.73% and 59.35%, respectively. For 66 350 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 Year 1200 0 -I ! • i 1 1 1 i i 1 1 1 1 i 1 1 1 1 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 Year Figure 5.5. Annual troll catch in the inside and outside areas 67 the outside regions, the C V is 39.39% for SWTR and 47.76% for NWTR. The high variations associated with the data create difficulties in interpreting the trends in catch numbers. Regression plots of the number of fish caught on the catch years were made to analyze the data trends. The results, shown in Figure 5.5, indicate that no significant trend was found in all regions except for NWTR (R 2 = 0.712; P r = 0 < 0.001). It is obvious that catches in this region increased over the 1975 - 1992 period, during which the catch trend of coho in NWTR is positive and significant at 99.9%. For the other regions, except for SWTR, the trends showed negative tendency, however, the regression coefficients were not statistically significant from zero. (ii) Effort The general pattern of effort distribution is quite different between inside and outside areas (Fig. 5.6). The outside area dominates the amount of fishing effort (given by days of fishing). SWTR has the highest effort averaging 13,100 fishing days / year. The second highest fishing effort was in NWTR, which averaged 7,400 fishing days / year. The mean value of effort in the inside area is 2 - 3 times smaller than that in the outside, averaging 7-13 thousands fishing days / year. The coefficient of variation of the effort data from 1975 - 1992 suggests less variations than those of the catch data, in all regions. The inside regions have higher C V ranging from 30 -35%, compared to outside, which ranges from 19.3 - 22.7%. This result, in general, indicates that there was a steady pattern of effort put by trollers in all regions throughout the period of 1975 - 1992. 68 Figure 5.6. Annual troll effort in the inside and outside areas 69 The results of linear regressions show that in all regions, the trends are significant, except for SWTR. In the inside regions, the trend lines show a decreasing tendency, with significant results at P r = 0 = 0.01. Both SCTR and GSTR have the values of R 2 = 0.4. In the outside regions the regression results give R 2 values for SWTR = 0.3 with no significant trend, however, for NWTR the result is quite significant (R2= 0.4; P r = 0 =0.01). The results indicate, that the data from NWTR show a positive tendency within the period of 1975 - 1992. (iii) CPUE The CPUE in this study was derived by dividing catch (in piece) by effort (in days fishing) for each region, and the results are presented in Figure 5.7. Mean CPUE values for the outside area, during the 1975 - 1992 period, ranged from 40 to 50 pieces / day fishing, whereas the inside area has mean of 25 - 33 pieces / day fishing. The coefficients of variation are relatively similar in all regions (37 - 39%) except in GSTR, which has higher C V of 54%. The CPUE trends in the inside area indicate low significant correlations between catch numbers per unit effort and the catch year. The coefficient of correlation for the inside regions are R 2 = 0.2; P r = 0 = 0.05 and R 2 =0.1; P r = 0 = 0.2 for SCTR and GSTR, respectively. The CPUE for the outside area, on the other hand, shows a positive correlation between CPUE and the observation years. In both SWTR and NWTR, the CPUE show a significant increasing tendency for 1975 - 1992 period (see Fig. 5.7), with correlation coefficients R 2 = 0.5; P r = 0 = 0.001 for 70 Figure 5.7. Annual troll catch per unit effort (CPUE) in the inside and outside areas. The strait lines indicate the trend lines resulted from regression analysis. 71 SWTR and R 2 = 0.52; P r = 0 = 0.001 for NWTR. The catch and effort analyses were performed to provide general background about the distribution of coho Salmon throughout different regions in the study area. The results indicate that more fish are found in the outside area. The low effort in the inside area seems to have resulted in low catch of coho salmon. The increasing trend of CPUE in the outside area may be used as indicator of increasing fish abundance in this area. However, as Hilborn and Walters (1992) suggested, the information of CPUE should be treated carefully. Factors such as high efficiency of fish searching and sharing information on fishing grounds among fishermen, could mislead the interpretation of CPUE results because CPUE has become less proportional to fish abundance. In connection to the CPUE analysis, the fish abundance could probably be overestimated. However, the effect of this estimation is small with regard to the distribution of fish in specific regions. Therefore, it is assumed that the CPUE index is still adequate to indicate fish distribution. 5.4. Coded Wire Tag Recoveries Information on the number of tags recovered is obtained through sampling of commercial and sport fisheries in the U.S. and Canada. In British Columbia, commercial catch estimates are based mainly on records of fish sales by fishermen (Kuhn et al., 1988). Commercial catches are examined and sampled at various landing sites along the coast throughout the fishing season in order to estimate the catch of tagged fish. According to Kuhn et al, (1988), sampling of the catch to recover tagged fish has been designed to ensure that approximately 20% of the weekly 72 catch is examined for tags. This is accomplished by sampling all fish from 20% of vessels, trucks or packers. 5.5. Distribution of Big Qualicum Fish From MRP Database Data from the M R P database and troll fisheries were analyzed in order to examine the spatial distribution of the Coded Wire Tagged (CWT) Big Qualicum fish. In order to determine the relative distribution of the tagged fish caught by one unit of effort of the troll fisheries, I have used the ratio between the number of estimated recoveries in each region to the total effort in that particular region. Values of the estimated recoveries for each year were adjusted by taking into account the number of tagged coho released from Big Qualicum hatchery each year. Since the numbers of tags released were not the same for each brood year, the annual recoveries should then be adjusted as though they were released with the same number each year. The adjustment factor F j ; was calculated by taking into account the numbers of tagged coho released annually from Big Qualicum hatchery, through equation: Fj = # of tag released in one brood yearj / average release over all years The index of recoveries distribution is called Tagged per Unit Effort (TPUE) calculated as: TPUEy = Rjj / Ej where: = Estimated recoveries in (region i , year j) * F F = adjustment factor Ej = total effort from Troll fisheries in region i 73 Figure 5.8 shows that between 1975 - 1992 CWT fish from Big Qualicum were generally more abundant in the inside areas than outside except in 1991 when TPUE was higher in the outside regions. GSTR is the area where most fish (average = 49%) were recovered by troll vessels although the TPUE of this region varied markedly. SCTR has the second highest TPUE proportion with an average of 24.6% over all years. A distinct peak in 1983 was also noted in this region when over 70% of CWT fish from Big Qualicum were found in it. The two regions on the outside of Vancouver Island had the same pattern of TPUE over the years, with average TPUE values of 10% and 16.4% for SWTR and NWTR, respectively. Although the general distribution is concentrated in the inside areas, there were several years in which a large proportion of tagged fish were captured in the outside areas. In 1989, Big Qualicum fish were less abundant in SCTR compared to NWTR and SWTR had almost the same proportion as GSTR. Similarly in 1991, over 60% of Big Qualicum fish were captured in NWTR, and the rest were found in the other regions with relatively equal proportions. There is an apparent tendency in the recent years for more fish in the outside waters and more year to year variations of the TPUE estimations. 74 Figure 5.8. The proportion of tagged coho salmon caught in July by trollers in different recovery regions. Unit is given in tagged fish per unit effort (TPUE) A caution should be raised when comparing both inside and outside areas, especially when dealing with the amount of effort. Most trollers in the outside area are larger, and may have more fishing lines, than those in the inside. Therefore, it can be assumed that the outside trollers will have more unit of effort per boat as compared to inside trollers. This makes standardizing the amount of effort between those areas difficult since the effort in the inside is rather underestimated. However, given this condition and the fact that there are more tagged fish 75 recovered from the inside area, it is fair to assume that even more fish would have been recovered in the inside area, had there been one standard unit of effort for both areas. In summary, the results from catch-effort data show that coho salmon are relatively more abundant in the outside area. However, the calculation of TPUE for Big Qualicum fish shows that they are relatively more abundant in the inside area. The distribution of Big Qualicum coho also varies from year to year, with more Big Qualicum fish recovered in outside waters in some years. In the next chapter I shall discuss the results of the computer model that simulates coho salmon dispersal influenced by net surface-currents. The average distribution of fish predicted by the model will be compared with the catch distribution of coho salmon discussed in this chapter. The results from the simulation will also be compared to the annual variations in distribution of coho salmon. 6. Simulation results Previous chapters have provided some background on fisheries biology, oceanography and modelling. The fisheries aspects consist of two parts. The first part is the background information about the biology of coho salmon, dealing primarily with the marine phase of this species. The second part contains information on the distribution of tag recoveries, as well as the distribution of the troll fisheries that have been the main source of tag recovery data. The oceanographic aspects consist primarily of information on net surface circulation around Vancouver Island. This includes the best estimates of speeds and directions of surface currents in that area. The modelling part explains the structure of the model which was used to simulate the dispersal of coho salmon from Big Qualicum hatchery under different scenarios of current velocities and swimming speeds. In the present section, the detailed results of the simulation model are presented according to the scenarios that were mentioned in chapter 4. 6.1. Random movement Random movement, in this simulation model, is defined as the isotropic movement of particles without bias resulting from surface current movements. In this scenario, particles move without directional bias and the results will then provide a reference to evaluate the effects of surface currents on particle distribution. Figure 6.1 shows the results of random movement, using the assumptions that fish swim normally (standard swimming speed) or double their swimming speed (double swimming speed). The simulation was run 15 times (n = 15), and the 76 results were plotted as the average values ± SD. 100 n= 15 Standard 80 + s o ••a £ 60 •c Double SCTR GSTR SWTR NWTR Figure 6.1. Distribution of simulated fish under random movements, with different swimming speeds It should be noted that different swimming speeds affect the number of iterations used in the simulation procedures, as the number of iterations is proportional to the swimming speed. In simulation, the iteration number for standard and double swimming speed are 451 and 902, respectively. Explanation of the relationship between swimming speed and number of model iterations was presented in chapter 4. The distribution patterns of the simulated fish are, in general, relatively similar for both swimming speed scenarios. A large proportion of particles (average 37% - 45%) is still in Georgia Strait region after 13 months. The region with the second greatest abundance when running the simulation with standard swimming speed, is SCTR, with an average percentage of 20%. The outside regions show almost the same proportion, which is less than 20%, although the scenario with double swimming speed gave relatively more fish in both SWTR and NWTR. When the swimming speed is doubled, the particles are more evenly distributed among all regions (see Fig 6.1), although GSTR still has the greatest proportion (average 38%). This can be explained as follows. As indicated previously, the number of iterations is a function of swimming speed. Due to the random movements used in this scenario, the more iterations put in the simulation, the more dispersed the particles will be at the end of the simulation. With enough iterations, the distribution will eventually be equal in each region. The results from both standard and double swimming speeds indicate that the dispersal pattern expands seaward through both northern or southern openings of the Strait of Georgia: Johnstone Strait and Juan de Fuca Strait. The small percentage in the NWTR (17%) when the standard swimming speed was used, is due to the greater distance that particles have to travel in order to reach that location. 79 6.2. Biased Random movement Although pure random movement provides an important reference point in interpreting simulation results, in the real ocean the fish are always subject to the influence of currents. The purely random movements scenarios are, therefore, unrealistic when applied to the real distribution of fish. To explore the effects of currents, fish dispersal was simulated under three scenarios of surface current patterns. Each scenario corresponds to different swimming speeds. In the default pattern, the speeds and directions of surface currents used are the best estimates from available current data. The other two scenarios are the '1/2-default' which reflects the condition as current speeds are reduced to half of the default condition and the '2 X default', which is the condition when currents are doubled from the default pattern. Swimming speeds were the standard and double speeds. The particle distribution shows, in general, a similar pattern among the three scenarios, regardless of the swimming speed assigned to the fish. Figure 6.2 shows that the Strait of Georgia region has the highest proportion of the simulated fish distribution, ranging 61 - 79%, and dominates the other regions. For the standard swimming speed (Fig. 6.2a), the percentages of simulated fish distributed in this region were 62%, at the minimum, and 76% at maximum. Increasing the swimming speed gave more variation in the proportion of the simulated fish in the Strait of Georgia, which was 64% when the currents were 1/2-default, but reaches close to 80% in the Strait of Georgia when the current is doubled (Fig. 6.2b). The South Central Region (SCTR) had the second greatest abundance when the fish were simulated with standard swimming speed, regardless of current speed. The same was true at double swimming, when currents were at default or 1/2-default speed. With double swimming speed and double current speed, however, SWTR and NWTR had more fish than SCTR. The West coast regions (SWTR and NWTR), in general, had smaller proportions of fish in most 80 100 80 standard swimming speed n = 15 60 Q I 40 20 -f '** -1 - I mm S C T R • 1/2 Def ffl Default • 2 def G S T R S W T R N W T R 100 double swimming speed n = 15 80 60 | 40 U EL 20 4-m S C T R G S T R B • 1/2 Def • Default • 2 def S W T R N W T R Figure 6.2. Graph of sensitivity analysis, showing the distribution of simulated fish under different currents and simming speeds 81 simulations, ranging 5 - 15%. Obvious results from the simulation are that the simulated fish are less abundantly distributed in the Northern part (SCTR and NWTR) when currents are increased and there are fewer particles in the outside, especially in SWTR, when currents are reduced. The exception is for SWTR, when using 1/2 default current speeds and the double swimming speed. This scenario resulted in a higher percentage of fish in SWTR compared to the default and 2 default current patterns. Another feature of this sensitivity test is that the modelled distribution is similar to the genera] distribution pattern resulting from random movement. This indicates that varying current speeds will not significantly change the general pattern of particle distribution. However, it is apparent that surface current has an important effect on the distribution of fish from the Strait of Georgia in that the proportion of fish remaining in the Strait of Georgia was higher under all surface currents scenarios than under purely random movement. This can be illustrated by comparing the distribution patterns of random movement and biased movement using the default current and the standard swimming speed. Based on figure 6.3, the simulation results with biased movement showed a higher percentage of simulated fish in GSTR, but fewer in the outside area, compared to the patterns with purely random movement. In SCTR, there was relatively similar percentage of fish, about 20%, regardless of the surface current influence. These results clearly indicate that the simulated distribution of fish, in some areas, is sensitive to biases caused by the surface currents. However, as mentioned earlier, the model is less sensitive to changes in the magnitude of the currents within the range tested, or the swimming speed. 82 100 80 + 60 + = 40 BM Random Biased 20 - - -I j | H - -n J—I H H I 1 ru SCTR GSTR SWTR NWTR Figure 6.3. Distributions of simulated fish resulting from random and biased simulations 6.3. Seasonal distributions Seasonal distributions of the simulated fish were plotted by using the percentage of fish at different seasons during the 13 months of the simulation period. Figure 6.4 and 6.5 depict the distribution patterns of simulated fish with standard and double swimming speeds, respectively. The general results show a similar seasonal pattern for both swimming speeds. In the scenario with random movement, as expected, the fish showed a tendency to become equally distributed in all four regions, as the iterations increased, suggesting that there was a steady movement of fish out of the Strait of Georgia. Since the number of iterations for double swimming speed is greater than the standard swimming speed, the percentage of fish that moved to the outside regions was 83 relatively higher for the double swimming speed. The results of the seasonal distributions with bias due to surface current, show a pattern corresponding with the seasonal surface currents that were used in the simulation program. It is quite obvious that a northward movement was experienced by the fish during fall and winter seasons, regardless of the swimming speeds, which caused the fish to move to SCTR. This can be seen by the increased number of fish in SCTR, and the decreased number in GSTR. In the spring and second summer in the ocean, however, the fish were pushed back to GSTR by the net southward current during those seasons, resulting in return in more fish in the Strait of Georgia. A small portion of the fish apparently moved to the outside regions, once they reached SCTR or via Juan de Fuca Strait in the south. A variation of the general results was shown in the winter distribution when the fish were simulated with 2-default current speeds under standard swimming speed (Figure 6.4). The proportion of simulated fish in SCTR and GSTR shows that the fish were distributed in relatively equal proportion, indicating a different distribution pattern from that in other simulations. The explanation for this can be given by taking into account the calculation procedures of the probability that would determine the bias in movement direction, which is the ratio of surface current speed and the fish swimming speed (detailed explanations of the calculation procedures were provided in chapter 4). In this calculation, any bias value that is greater than 1.0 will be set to a maximum value of 1.0. In the 2-default scenario with standard swimming speed, due to the large bias values resulting from the large current speed, the bias during the fall and winter will be mostly at the maximum, which is 1.0. This condition applies to both north and south current 84 Summer I Fall Winter Spring Summer n Summer I F«ll Wfater Spring Summer D Figure 6.4. Seasonal distribution of simulated fish with standard swimming speed (number of iterations = 451) 85 Summer I F.D W » « , Sp r t a , Summer II Summer I FiB Winter Spring Summer • Figure 6.5. Seasonal distribution of simulated fish with standard swimming speed (number of iterations = 902) 86 directions in Johnstone Strait, meaning that there will be a balance regarding the directions through Johnstone Strait, as long as the bias value is greater than 1.0. As the fish get larger, the ratio becomes smaller and consequently, there would be clear bias of what direction the fish might take. In the 2-default scenario with standard swimming speed, the number of iterations put in the simulation was not enough to show the northward bias, as was the case with double swimming speed. Therefore, the distribution pattern in Fig. 6.4 gives the relatively similar proportion between SCTR and GSTR in the scenario with 2 X default currents and standard swimming speed. 7. Discussion 7.1. General distribution of coho salmon from Big Qualicum hatchery The results of the coho dispersal model with biased movements due to surface currents, shows a general distribution pattern with strong qualitative similarities to the actual distribution of coho based on tag recovery data (i.e. the observed distribution). Both indicate that more coho salmon are found in the Strait of Georgia (GSTR), followed by SCTR, NWTR and finally SWTR. With regard to the distribution in GSTR, however, the model overestimated the proportion of fish that stay in that region. This can be seen from the high percentage of the fish in the modelled distribution that remains in GSTR, which showed approximately 20% higher than the observed distribution. This condition is changed when the comparison is made between the random movement (i.e. non-bias model) and the observed distribution. The percentage of fish in GSTR is relatively similar between random and observed distribution, although there were differences in the other regions, such as fewer fish in SCTR from the random model, and relatively more fish in the SWTR. Comparison between observed and simulated distributions from random and biased model, was made by comparing the average distribution of tag recoveries from 1975 - 1992 with the distribution provided by the computer simulation using the best current estimates in the study area. Figure 7.1 shows the patterns, generated from three distribution models. In general, the simulated distribution shows a similar pattern with that from the observed pattern. However, the proportion of fish in the Georgia Strait region is considerably higher in simulated distribution, 87 conversely, the computer simulation gives fewer simulated fish in the other regions. 88 SCTR GSTR SWTR NWTR Figure 7.1. Comparison between observed, modelled and random distributions Despite the similarities between random and observed distributions, as compared to between modelled and observed distributions, shown in the above figure, surface currents are still important factor in determining fish dispersal and hence the ocean distribution of Big Qualicum coho. This statement could be explained as follows. Firstly, the distribution from the random model is produced when the fish movements are simulated through non-moving ocean waters and therefore the random swimming behaviour of the fish, represented by particles, is the only factor that influences the fish movements. As mentioned in section 6.1, given enough iterations, the 89 random model is expected to result in an equal distribution for all regions. This argument suggests that there is a condition of both random and observed distributions that is coincidentally similar after certain period. This means that the random model will eventually give a relatively similar distribution in all regions over a long period of simulation. The biased model (ie. influenced by surface currents) reveals a rather constant distribution at the end of 13 months of simulation time, with GSTR is the dominant place. This relatively similar pattern under various scenarios of currents and swimming speed from the biased model, is due to the trapping mechanisms resulting from the permanent circulation in the Strait of Georgia that causes organisms to stay in this Strait. Secondly, since the presence of surface currents is real, the comparison should be made between the distributions derived from the biased model (i.e. bias due to surface currents) and the observed distribution. In this case, the results of the model represent one possible outcome of the conditions as found in the nature. In the following section, I shall discuss the seasonal pattern of coho salmon dispersal that resulted from simulating the model with bias due to net surface currents. During the May and June period, when fish enter the ocean, their movements will be mostly driven by advection process caused by the currents flowing southward. The degree of surface current influence is perhaps the main difference between Big Qualicum coho and those from Washington and Oregon (Pearcy and Fisher, 1988). The coho originating from Big Qualicum experience currents that are not as strong as those on the Washington and Oregon coasts, which have average flow of 20 - 40 cm/sec (Huyer et al, 1979), compared to mean flow of 10 - 20 cm/sec in the Strait of Georgia (Thomson, 1980). Thus, the swimming behaviour of 90 the fish is probably more important in the Strait of Georgia than off Washington and Oregon. Although there was a difference in the speed of surface currents, the net result is still a southward movement of the juvenile coho in the Strait of Georgia. During the simulation, however, a few particles move northward in the summer period. This small percentage of northward moving fish were those that crossed the Strait of Georgia and were advected by the relatively slow net surface-currents, which move northward along the mainland coast of B C during the summer. This surface current is part of the permanent circulation observed in the Strait of Georgia (Thomson, 1980). By fall or early winter, coho from Big Qualicum occupy most areas in the central and southern part of the Georgia Strait. These areas, as Healey (1978) describes, appear to be favourite nursery grounds for juvenile Salmon in the Strait of Georgia. The Gulf Islands area, especially along the Vancouver Island shore, was considered a good feeding ground for the juvenile salmonids. During winter, however, due to the swift surface currents that move northward, fish were carried to the north toward Johnstone and Queen Charlotte Straits. It is noticeable, however, that some fish remained in the Georgia Strait, although the majority moved along with the current. A proportion of fish that were able to reach SCTR, were then carried toward the open ocean and the NWTR region. Also during the winter and spring some of the fish in the GSTR are carried through the Juan de Fuca Strait to SWTR in the West coast of Vancouver Island. However, as indicated in section 6.1, the proportion that move to the outside regions is less than 20% of the total number of coho that were initially released at the beginning of the simulation. The net surface-currents in spring and early summer would cause the fish that had been advected to the north during the winter to move southward again into the GSTR. It is 91 apparent however, that the influence of the net surface-current at this period was reduced because of the increased ability of the fish to swim against the current. The obvious effect on the simulation program is that the bias on the particles will also decrease. Another aspect of the simulation results is that both simulation and observed distributions provide evidence that Big Qualicum coho, as part of the inside coho populations, tend to stay in the coastal waters. Big Qualicum coho are mainly distributed in the region of the Georgia Strait followed by the South Central region in the Northern part of Vancouver Island. Apparently, there are only a few Big Qualicum fish that moved outside to the open ocean. This can be seen from the low percentage of fish in Southwest and Northwest regions. Following Milne's (1950) hypothesis of two separate 'groups' of coho population in British Columbia, it is obvious that Big Qualicum coho would be classified as an 'inside' coho population. The influence of surface current alone is important. However, it is not sufficient to explain the whole distribution pattern of Big Qualicum coho during their ocean life. The distribution pattern of coho salmon explained above is a result of the distribution of simulated fish under the assumption that the fishing activities are equally distributed around the coast of Vancouver Island. However, based on the troll fishing analysis provided in chapter 5, the nature of fishing activities in the area around Vancouver Island (i.e. inside and outside areas) is relatively different, in terms of the fisheries actions between these two areas. It is obvious that the troll fishing activities in the outside area are more intense than those in the inside area. This can be seen from the amount of effort, and subsequently, the number of fish caught from this area. During the 1975 to 1992 period, the effort by outside trollers was, on average, two or 92 three times higher than that by inside trollers. The same numbers are shown by the catch data of these areas. Another difference between the inside and outside areas is the presence of sport fishing in the inside area, which is mainly in the Strait of Georgia. In this region, it is quite significant that the sport fishery is more dominant, and has higher catch than the troll fishery. This means that the troll fishery has to compete with the sport fishery to fish the coho salmon available in this region. These differences indicate that there is bias associated with the nature of troll fishing between inside and outside areas. This condition can be translated that if the same troll fishery were to operate and catch the available fish resulted from the simulated fish distribution, then, there would be bias, in terms of the number of fish caught in these two areas. It is expected that the number of tagged coho salmon caught in the inside area, especially GSTR, would not be as high as the predicted number resulting from the simulated fish distribution. The reason for this is the presence of sport fishery that would compete with the troll fishery in exploiting the same number of fish. As a result of this situation is the reduction of the available fish that can be caught by the troll fishery. Since troll fishery has been used, in the present study, as the type of activity to examine the distribution of coho salmon, it would be an adequate reason to expect that the number of fish caught by troll fishery in the inside area, especially, in GSTR, would be lower than the number predicted by the model. In the outside area, on the other hand, the troll fishing is the only major activity in coho fishery. Therefore, it would be expected that the number of fish caught in both SWTR and NWTR, is the representative number of fish available in the outside area. The argument presented here is based on the assumption that the condition of the coho fishery (i.e. sport and troll fisheries) in both inside and outside areas would operate normally, similar to the 93 condition in the period of 1975 to 1992, which was used to examine the fisheries data. In summary, therefore, the fish catch data probably underestimate the abundance of Big Qualicum coho in the Strait of Georgia region and overestimate their abundance in other regions 7.2. Interannual variations Although the results of the simulation model show that the general distribution of Big Qualicum coho is qualitatively similar to that of the Mark Recovery Program (MRP) data, there are variations from the MRP analysis, such as the 1983 and 1991 distributions, that show a different pattern. A sensitivity analysis, aimed to test the effects of different current speeds, was performed to examine if variations in current speeds could account for the variations of tag recoveries. The results, as discussed in section 6.2 (Fig. 6.2), show a similar distribution pattern, regardless of the different current speeds, which indicate that the model is relatively insensitive to current changes. Apparently, there are factors other than average surface currents that influence the dispersal of juvenile coho around the Vancouver Island. One possibility is that changes in water circulation around the study area may have resulted in the variations of tag returns of coho salmon in the Strait of Georgia. The variation of tag returns in different years of Big Qualicum coho were selected to show the different extremes of the fish distribution. The results of 1983, 1990 and 1991 distribution were chosen as examples. Figure 7.2 depicts the percent distribution of tag recoveries in these years. The case in 1983 was an example when the majority of the Big Qualicum fish were recovered in 94 S C T R G S T R S W T R NWTR Figure 7.2. Variations of tagged per unit effort. The examples used to show the variations are 1983,1990 and 1991. the north, in SCTR and NWTR. 1990 shows a typical but rather extreme example of the distribution of Big Qualicum fish whose ocean life was mostly spent in the Strait of Georgia. In this particular year, there was a very high proportion (63%) of Big Qualicum fish in the Strait of Georgia, resulting in fewer fish in the other regions. Another example is the distribution pattern in 1991. In that year the Big Qualicum fish were found outside, in SWTR and NWTR. Since the fish were more abundant in SWTR in 1991, it is speculated that the southern route via Juan de Fuca strait was the route taken by the fish in order to move outside. The 1983 and 1991 examples clearly indicate movements of this species, away from the Strait of Georgia, where they can be found in a large numbers. Several possibilities could explain this variation. 95 Directed migration, undertaken by the fish, is one possibility to explain the interannual variation. Through this process, the fish choose a migratory direction based on factors such as the celestial or other reference cues. Compass orientation movement, following Griffin's (1955) type 2 classification, is speculated to be responsible for coho movement outside the Strait of Georgia. Through this directed movement, the influence of surface currents is minimized due to cues that fish might use in their migration process. A similar observation was reported by Pearcy and Fisher (1988) in their study of Oregon and Washington coho, which showed a tendency for the fish to move north against the southward current on the coastal area of Washington state. This directed movement is regarded as governing the behaviour of the fish when they first enter the ocean. According to Harden Jones (1981), Dodson (1988), and Neill (1984), migration process in fish is undertaken to maximize the comfort of their surroundings. This means that when a group of coho salmon enter the ocean, and experience a less favourable condition, these fish might migrate to other places. By doing this, fish are using their capability to move under directed migration to find a better place. The most common factor that cause the fish to migrate is a limitation of food supply. Another factor is predator avoidance. The directed movement is, in my opinion, related to the food abundance in the entrance point of the fish to the ocean. Another factor that might possibly cause the interannual variation is the food supply in the ocean. From the directed movement explanation, it is clear that the limited amount of food supply in a certain region where the fish are present, could result in the movement of those fish to other regions. In the case of coho salmon in the Strait of Georgia, Healey (1980) concluded that if the fish entered the Strait and found little food, then the fish were most likely to move to the 96 -•' outside area. Whereas fish that found themselves in a good feeding ground in the Strait of Georgia will stay in the inside waters. It should be noted, however, that studies on food abundance are limited in this study area, causing difficulties to interpret the relationship between food supply and the distribution pattern of coho salmon. An environmental factor that could possibly cause this interannual variation is an anomalous flushing event, first hypothesized by Favorite (1961) and later Wickett (1977) who pointed out that sea level differences between inside and outside areas may be responsible for changes in the relatively permanent circulation of surface water in the Strait of Georgia. In this Strait, sea level difference might possibly occur if there is a large input of water entering the Strait. One possible source of such input is the Fraser River, which has an annual discharge pattern from the period of 1975 to 1991 as shown in Figure 7.3. There are three obvious peaks, in 1976, 1982 and 1990-1991, associated with the high amount of fresh water carried by the Fraser River flows. These years of high fresh water discharge are speculated to have a link with the coho movement, out from the Strait of Georgia. 97 4000 2000 -I 1 1 1 1 1 1 1 1 1975 1977 1979 1981 1983 1985 1987 1989 1991 Figure 7.3. Fraser River mean discharge (measured at Hope) According to the graph of tagged fish per unit effort (TPUE) in Figure 5.8, in 1976 there were net migrations of Big Qualicum coho to the outside of Strait of Georgia. In this year, there were at minimum 40% of fish moved to the SCTR in the north, and about 25% moved to SWTR and NWTR in the west coast of Vancouver Island. Similar condition of high fresh water discharge was observed during 1991, when the majority of the Big Qualicum coho moved to SWTR and NWTR. The relatively high amount of Fraser run-off in 1976 and 1991, could have resulted in a stronger buoyancy driven current that flows to the outside area, through the Juan de Fuca Strait. This phenomenon, then, has the potential to affect the distribution of the organisms in the Strait of Georgia through advection processes, in which the organisms could be carried to the outside area. For two of the years of high discharge, 1982 and 1990, however, coho salmon 98 preferred to stay in the Strait of Georgia. Thus, discharge from the Fraser River appears not to be a satisfactory explanation for high numbers of coho moving out of the Strait of Georgia. Another possibility that might contribute to the variation of fish distribution is the presence of strong wind that causes the surface water to move in a certain direction, in which the fish could be carried out. To examine the influence of wind in a certain area, exceedance diagrams of wind force ((km/h)2) were used to plot the frequency of occurrence (in days) of the wind directions to the northwest and southeast, respectively. The exceedance diagram is based on the hourly data of wind's speed and direction, measured at Sand Heads station (located in the Strait of Georgia). In this analysis, the 24-hour data were averaged into daily data , and then plotted according the axis of the Strait, which is the northwest and southeast axis. Exceedance diagram is used to determine the cumulative frequency of the number of days when the wind moved to either northwest or southeast direction. Figure 7.4 shows the exceedance diagram of wind pattern in June 1982 - July 1983. The reason for plotting the diagram by using this period is to explore the influence of wind force to the surface currents during the same period, which might have resulted in the northward distribution of coho salmon, caught in 1983. 99 Figure 7.4. Exceedance diagram of wind pattern using the data from June 1982 to July 1983. The strong northwestward movement is believed to become one of the major reasons of northward distribution of coho salmon from Big Qualicum hatchery in 1983. The interpretation derived from this analysis is that there was a strong possibility for coho salmon originating from the Strait of Georgia to be carried northward due to relatively strong southeasterly winds. The graph in Figure 7.2, shows that the majority of coho salmon recovered in 1983 were caught in SCTR and NWTR, which are the regions in the northern part of Vancouver Island. 100 To compare the wind pattern in 1983 with the average pattern, the graph in figure 7.5 shows the distribution of anomaly between the wind pattern in 1983 and the average pattern from 1981 - 1989. The positive values to the northwest indicate that in 1983 there were net southeasterly wind, which is believed to have influence on the water circulation patterns in the Strait of Georgia in 1983. Furthermore, due to this significant northwest movement in 1983, organisms that occupy the surface water might have migrated northward, as a possible result of this condition. Another factor that may have affected the distribution of coho salmon in 1983, besides the flushing event hypothesis and the strong northerly wind, is the high Sea Surface Temperature (SST). The increased SST during the 1982 - 1983 period was found by others to have an influence on salmonids distribution in the Northeast Pacific region (Mysak, 1986; Xie and Hsieh, 1989; Pearcy and Fisher, 1988). Studies on the distribution of Juvenile coho salmon in Oregon and Washington (Pearcy and Fisher, 1988) and on the landfall latitude of returning Sockeye Salmon (Thomson et al, 1992), all indicate a strong northward movement of the species in 1983. In the Strait of Georgia, coho salmon that enter the ocean in summer might be carried to the outside area by the flushing currents. However, due to the unfavourable warm water in the southern part of Vancouver Island, these fish may then have moved northward to NWTR and some of them moved to SCTR. 101 Figure 7.5. Anomaly graph of wind directions (northwestward and southeastward) in 1983, compared to the average pattern from 1981 -1989, at different wind force. Solid lines is the smoothed line derived from the data. Data were taken from Sand Heads station, courtesy of Ian Jardine, Dept. of Oceanography, UBC. A l l factors that are believed to contribute to interannual variation of coho distribution in the inside and outside areas are thought to 'work' together as a combination process. For instance, the 1983 variation to the north, might be the result of directed movement of the fish because there was no or less food in the Strait of Georgia. Moreover, the presence of strong northwest variation on oceanic condition due to prevailing southeasterly wind, may also have contributed to this variation. Similarly, in 1991, due to relatively high possibility of a flushing event in the Strait of Georgia, the fish might have felt less comfortable in this area and therefore, 102 decided to move outside. The anomalous flushing hypothesis through high fresh water discharge, however, was contradicted by the fish distributions in 1982 and 1990. In these years, coho salmon were found to distributed more in the Strait of Georgia region. In my opinion, although the fresh water discharge was high during these years, the fish still preferred to stay in GSTR probably because the environment was still supplying enough food for them to stay in the Strait. These conditions were, then, changed in 1983 and 1991, which resulted in the movement of the fish to the northern regions (SCTR) and outside regions (SWTR and NWTR), respectively. I speculate that the effect of high fresh water discharge in the previous years (i.e. 1982 and 1990), coupled with the environmental conditions in 1983 and 1991, might have resulted in declining food abundance, which caused the fish to leave the Strait of Georgia. The northward distribution of fish in 1983 is also related to strong southeasterly winds, which might cause the organisms living in the surface water column to be pushed to the north. The hypotheses of coho distribution, however, in particular the speculations regarding the effects of other environmental factors, need further analysis and testing, which could be regarded as an extension of the present study. An individual-based model, such as the one used in this study, provides a tool to better understand the interactions among variables that might influence the distribution of organisms in a rather large geographical area. Other variables that might be incorporated in to the model and that would possibly affect coho dispersal, include the spatial and temporal distribution of food organisms and predators in the study area. The compartment model (illustrated in habitat boxes), used in this computer program allows for future modification, to incorporate other environmental variables that needed to be tested. Bibliography Argue, A.W., R. Hilborn, R. M . Peterman, M . J. Staley and C. J. Walters. 1983. Strait of Georgia chinook and coho fishery. Can. Bull. Aquat. Sci. 21: 91 p. Argue, A.W., M . P. Shepard, T.F. Shardlow and A.D. Anderson. 1987. Review of salmon troll fisheries in southern British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 1502: 150 p. Aro, K . V . and M . P. Shepard. 1967. Pacific salmon in Canada, p. 225-327. In: Salmon of the north Pacific ocean. Part IV: spawning populations of north Pacific salmon. Int. North. Pac. Fish. Comm. Bull. 23 Bilton, H.T., R.B. Morley, A.S. Coburn, and J. Van Tyne. 1984. The influence of time and size at release of juvenile coho salmon (Oncorhynchus kisutch)ox\ returns at maturity: results of releases from Quinsam River hatchery, BC. in 1980. Can. Tech. Rep. Fish. Aquat. Sci. 1306: 98 p. Botsford, L.W., D.A. Armstrong, and J.M. Shenker. 1989. Oceanographic influences on the dynamics of commercially fished populations. In :Coastal oceanography of Washington and Oregon. M.R. Landry and B . M . Hickey (eds.). Elsevier science publications, Amsterdam, pp: 511 - 565 Burgner, R.L. 1991. Life history of sockeye salmon (Oncorhynchus nerka). In: Pacific salmon life histories. C. Groot and L. margolis (eds.). University of British Columbia Press, Vancouver, pp: 1 - 117 Cannon, G. A. 1978. (Ed.), Circulation in the Strait of Juan de Fuca. Some recent oceanographic observations. N O A A Technical Report E R L 399-PMEL 29. 49 pp Clemens, W.A. 1930. Pacific salmon migration: the tagging of coho salmon off the east coast of Vancouver Island in 1927 and 1928. Bull. Biol. Board Can. 15:1-19 Crean, P.B. 1967. Physical oceanography of Dixon Entrance British Columbia. Bull. Fish. Res. Board Can. 156: 66p Cross, C.L., L . Lapi, and E.A. Perry. 1990. Production of chinook and coho salmon from British Columbia hatcheries, 1971 through 1989. Can. Tech. rep. Fish. Aquat. Sci. 1816: 48 p. 103 104 DeAngelis, D.L., and G.T. Yeh, 1984. An introduction to modeling migratory behaviour of fishes. In : Mechanisms of migration in fishes. J.D. McCleave, G.P. Arnold, J.J. Dodson and W.H. Neill (eds.) Plenum Press, New York, pp: 445-469 DFO, 1990. Strait of Georgia coho salmon: resource status and management planning process. Department of Fisheries and Ocean, 22 p Dodson, 1988. The nature and role of learning in the orientation and migratory behaviour of fishes. Environmental biology of fishes Vol . 23, No. 3, pp: 161-1182 Favorite, F. 1961. Surface temperature and salinity off the Washington and British Columbia coasts, August, 1958 and 1959. J. Fish. Res. Board Can. 18: 311-319 Foerster, R.E. 1955. The pacific Salmon (genus Oncorhynchus ) of the Canadian Pacific coast with particular reference to their occurance in or near fresh water. Int. North. Pac. Fish. Comm. Bull. 1: 1-56. Fraser, F.J., E.A. Perry, and D.T. Lightly. 1983. Big Qualicum river salmon development project. Volume 1: a biological assessment 1959-1972. Can Tech. Rep. Fish. Aquat. Sci. 1189: 198 p. Freeland, H.J., W.R. Crawford, and R.E. Thomson. 1984. Currents along the pacific coast of Canada. Atmosphere-Ocean 22: 151-172 French, R.R., T. Bilton, M . Osaka, and A.C. Hartt. 1976. Distribution and origin of sockeye salmon in offshore waters of the North Pacific ocean. International North Pacific Fisheries Commission report 34, 113 pp Godfrey, H. 1965. Coho salmon in offshore waters. In: Salmon of the North Pacific Ocean, part IX. Coho, chinook and masu salmon in offshore waters. Int. North. Pac. Fish. Comm. Bull. 16 : pp 1 - 39 Godfrey, H. , K . A . Henry, and S. Machidori. 1975. Distribution and abundanve of coho salmon in offshore waters of the North Pacific ocean. International North Pacific Fisheries Commission report 31, 80 pp Griffin, D.R. 1955. Bird navigation. In: A. Wolfson (ed.). Recent studies in avian biology. University of Illinois. Urbana, USA. p: 154-197 Groot, C , and T. Quinn. 1987. Homing migration of sockeye salmon (Oncorhynchus nerka), to the Fraser River. Fish. Bull. (US) 85: 455-469. 105 Groot, C. and K. Cooke. 1987. Are the migrations of juvenile and adult Fraser River sockeye salmon (Oncorhynchus nerka) in near-shore water related ?, In: sockeye salmon Population biology and future management. H.D. Smith , L . Margolis and C C Wood (eds.), Can. Spec. Publ. Fish. Aquat. Sci. 96: 56-60 Hartt, A .C and M . B . Dell, 1986. Early oceanic migrations and growth of juvenile pacific salmon and steelhead trout. International North Pacific Fisheries Commission report 46, 105 pp Harden Jones, F.R. 1968. Fish Migration. Edward Arnold Ltd., London. 325 pp Harden Jones, F.R. 1981. Fish migration: Strategy and tactics. In: D.J. Adley (ed.). Animal Migration. Cambridge University Press. Cambridge, England, p: 139-165. Harlinveaux, R.H. 1954. Surface tidal currents in Juan de Fuca Strait. J. Fish. Res. Board Can. 11: 14-31 Healey, M.C. 1978. Distribution, abundance, and feeding habits of juvenile Pacific salmon in Georgia Strait, British Columbia. Fish. Mar. Serv. Tech. Rep. 788: 49 p. Healey, M.C. 1980. The ecology of juvenile salmon in Georgia Strait, British Columbia. In: W.J. McNeil and D.C. Himsworth (Eds.) Salmonid ecosystem of the north Pacific. Oregon State University Press, Corvalis. pp: 203-229 Healey, M.C. 1986. Regional and seasonal attributes of catch in the British Columbia troll fishery. Can. Tech. Rep. Fish. Aquat. Sci. 1494: 65p. Healey, M.C. 1993. The management of Pacific salmon fisheries in British Columbia. In: L.S. Parsons and W.H. lear (Eds.) Perspectives on Canadian marine fisheries management. Can. Bull. Fish. Aquat. Sci. 226 Healey, M.C. and C. Groot, 1987. Marine migration and orientation of ocean-type chinook and sockeye salmon. In: M.J. Dodswell., R.J. Klauda., C M . Moffitt., R.L. Saunders., R.A. Rulifson., and J.E. Cooper (Eds.) Common strategies of anadromous and catadromous fishes. American Fisheries Society symposia (AFS., Bethesda, Maryland), vol. l ,pp. 298 - 312 Hickey, B . M . , R.E. Thomson, H. Yih, and P.H. LeBlond. 1991. Velocity and temperature fluctuations in a buoyancy-driven current off Vancouver Island. J. Geophys. Res. 96: 10,507-10,538. Hilborn, R. 1992. Hatcheries and the future of salmon in the northwest. Fisheries 17(1): 5-8 106 Hilborn R, and J. Winton, 1993. Learning to enhance salmon production: lessons from the Salmonids Enhancement Program. Can. J. Fish. Aquat. Sci. 50: 2043-2056 Hoar, W.S. 1958. The evolution of migratory behaviour among juvenile salmon of the genus Oncorhynchus. J. Fish. Res. Board Can. 15: 391-428 Holbrook, J.R., R.D. Muench, D.G. Kachel, and C. Wright. 1980. Circulation in the Strait of Juan de Fuca. Recent oceanographic observations in the eastern basin. N O A A Technical Report E R L 412-PMEL 33 Huyer, A. , E.J.C. Sobey, and R.L. Smith. 1979. The spring transition in currents over the Oregon continental shelf. J. Geophysical Res. 84: 6995 - 7011 Irvine, J.R., and B.R. Ward, 1989. Patterns of timing and size of wild coho salmon (Oncorhynchus kisutch) smolts migrating from Keogh River watershed on northern Vancouver Island. Can. J. Fish. Aquat. Sci. 46: 1086-1094 Kuhn, B.R., L . lapi, and J.M. Hamer. 1988. An introduction to the C a n a d i a n database on marked pacific salmonids. Can. Tech. rep. Fish. Aquat. Sci. 1649: viii + 56 p. Labelle, M . 1992. Straying patterns of coho salmon (Oncorhynchus kisutch) stocks from southeast Vancouver Island, British Columbia. Can. J. Fish. Aquat. Sci. 49: 1843-1855 LeBlond, P.H. 1983. The Strait of Georgia: functional anatomy of a coastal sea. Can. J. Fish. Aquat. Sci. 40(7): 1033-1063 LeBlond, P.H., B . M . Hickey, and R.E. Thomson. 1986. Runoff driven coastal flow off British Columbia. In: S. Skreslet (ed.) The role of freshwater outflow in coastal marine ecosystems. N A T O ASI Series G7, Springer-Verlag, Berlin, pp: 309-317 Leggett, W.C. 1984. Fish migration in coastal and estuarine environments: A call for new approaches to the study of an old problem. In : Mechanisms of migration in fishes. J.D. McCleave, G.P. Arnold, J.J. Dodson and W.H. Neill (eds.) Plenum Press, New York, pp: 159-178. Manzer, J.L, T. Ishida, A .E . Peterson, and M . G . Hanavan. 1965. Salmon of the north pacific ocean, part V . Offshore distribution of salmon. Int. North pac. Fish. Comm. Bull. 15: 452 p. McCleave, J.D., and G.S. Wippelhauser. 1987. Behavioral aspects of selective tidal stream transport in juvenile american eels. In: M.J. Dodswell., R.J. Klauda., C M . Moffitt., R.L. Saunders., R.A. Rulifson., and J.E. Cooper (Eds.) Common strategies of anadromous and catadromous fishes. American Fisheries Society symposia (AFS., Bethesda, Maryland), vol. 1, pp. 138 - 150 107 Milne, D.J. 1950. The difference in the growth of coho salmon on the east and west coasts of Vancouver Island in 1950. Fish. Res. Board Can. Prog. Rep. Pac. Coast Stn. 85: 80-82. Microsoft Q-basic. 1992. Microsoft Q-basic compiler for DOS, version 4.5. Microsoft Coorporation. Redmond, Washington, USA. Morley, R.B., H.T. Bilton, A.S. Coburn, D. Brouwer, and J. van Tyne. 1988. The influence of time and size at release of juvenile coho salmon (Oncorhynchus kisutch) on returns at maturity; results of three studies at Quinsham Hatchery, B C . Can. Tech. Rep. Fish. Aquat. Sci. 1620: 120 p. Moser, M.L . , A.F. Olson, and T.P. Quinn. 1991. Riverine and estuarine migratory behavior of coho salmon (Oncorhynchus kisutch) smolts. Can. J. Fish Aquat. Sci. 48: 1670-1678. Mysak, L . A . 1986. El Nino, interannual variability and fisheries in the northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 43: 464-497 Neave, F. 1949. Game fish population of the Cowichan River. Bull. Fish. Res. Board Can. 84: 1-32 Neill, W.H. 1984. Behavioural enviroregulation's role in fish migration. In: Mechanisms of migration in fishes. J.D. McCleave, G.P. Arnold, J.J. Dodson and W.H. Neill (eds.) Plenum Press, New York, pp: 61-66. Neill, W.H. 1979. Mechanisms of fish distribution in heterothermal environments. Amer. Zool., 19: 305-317 « Pearcy, William G. 1992. Ocean ecology of North Pacific salmonids. University of Washington Press, Seattle. 179 p. Pearcy, W.G., and J.P. Fisher, 1988. Migrations of coho salmon (Oncorhynchus kisutch), during their first summer in the ocean. Fish. Bull. (U.S). 86: 173-195 Pasqual, M.A. , and T. P. Quinn, 1991. Evaluation of alternative models of the coastal migration of adult Fraser River sockeye salmon (Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. 48:799-810 Peterman, R. M . , S.G. Marinone, K. A. Thomson, I. D. Jardine, R.N. Crittenden, P.H. LeBlond, and C.J. Walters. 1994. Simulation of juvenile sockeye salmon (Oncorhynchus nerka) migrations in the Strait of Georgia, British Columbia. Fisheries Oceanography, 3: 221-235 . 108 Pickard, G.L. 1956. Surface and bottom currents in the Strait of Georgia. J. Fish. Res. Board Can. 13(4): 581-590. Power, G., M . V . Power, R. Dumas, and A. Gordon. 1987. Marine migration of Atlantic salmon from Rivers in Ungava bay, Quebec. Am. Fish. Soc. Symp. 1: 364-376. Quinn, T.P., and C.Groot. 1984. Pacific salmon (Oncorhynchus) migrations: Orientation versus random movement. Can. J. Fish. Aquat. Sci. 40: 1319-1324. Quinn, T.P, and K. Fresh. 1984. Homing and straying in chinook salmon (Oncorhynchus tshawytscha ) from Cowlitz River hatchery, Washington. Can. J. Fish. Aquat. Sci. 41: 1078-1082 Renshaw, E. 1991. Modelling biological population in space and time. Cambridge University Press. Cambridge, England. 403 pp. Saila, S.B., and R.A. Shappy, 1963. Random movement and orientation in salmon migration. J. Cons. Perm. Int. Explor. Mer. 28: 153-166 Sandercock, F.K. 1991. Life history of coho salmon (Oncorhynchus kisutch). In. Pacific salmon life histories. C. Groot and L. Margolis (eds.). University of British Columbia Press, Vancouver, pp: 395-446 Simms, S.E., and P. A. Larkin, 1977. Simulation of dispersal of sockeye salmon (Oncorhynchus nerka) underyearlings in Babine Lake. J. Fish. Res. Board Can. 34: 1379-1388 Thomson K . A . , W.J. Ingraham, M.C. Healey, P.H. LeBlond, C. Croot, and C.G. Healey. 1992. The influence of ocean currentson latitude of landfall and migration speed of sockeye salmon returning to the Fraser River. Fisheries Oceanography 1(2): 163-179 Thomson, R.E. 1976. Tidal currents and estuarine-type circulation in Johnstone Strait, British Columbia. J. Fish. Res. Board. Canada, 33: 2242 - 2264 Thomson, R.E. 1977. Currents in Johnstone Strait, British Columbia: supplemental data on the Vancouver Island side. J. Fish. Res. Board. Canada, 34: 697-703 Thomson, R.E. 1981. Oceanography of the British Columbia coast. Can. Spec. Publ. Fish. Aquat. Sci. 56: 291 p. 109 Thomson, R.E., B . M . Hickey, and P.H. LeBlond. 1989. The Vancouver Island coastal current: fisheries barrier and conduit, p. 265-296. In: R.J. Beamish and G.A. McFarlane (eds.) Effect of the ocean variability on recruitment and evaluation of parameters used in stock assessment models. Can. Spec. Publ. Fish. Aquat. Sci. 108 Tully, J. P., and A.J . Dodimead, 1957. Properties of the water in the Strait of Georgia, British Columbia, and influencing factors. J. Fish. Res. Board. Can. 14(3): 241 - 319 Van der Steen, W.J. 1984. Methodological aspects of migration and orientation in fishes. In: Mechanisms of migration in fishes. J.D. McCleave, G.P. Arnold, J.J. Dodson and W.H. Neill (eds.) Plenum Press, New York, pp: 421-444. Waldichuk, M . 1957. Physical oceanography of the Strait of Georgia, British Columbia. J. Fish. Res. Board. Can. 14(3): 321 - 486 Wickett, W.P. 1977. Relationship of coastal ocenographic factors to the migration of Fraser River sockeye salmon {Oncorhynchus nerka). Int. Counc. Explor. Sea. C M 1977/M26: 18 p. Xie, L., and W.W. Hsieh. 1989. Predicting the return migration routes of the Fraser River sockeye salmon {Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. 46: 1287-1292 

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.831.1-0053241/manifest

Comment

Related Items