Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Experimental field manipulation of stream temperatures and suspended sediment concentrations : behavioural… Quigley, Jason Trevor 2003

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

Item Metadata

Download

Media
831-ubc_2003-0289.pdf [ 10.89MB ]
Metadata
JSON: 831-1.0075100.json
JSON-LD: 831-1.0075100-ld.json
RDF/XML (Pretty): 831-1.0075100-rdf.xml
RDF/JSON: 831-1.0075100-rdf.json
Turtle: 831-1.0075100-turtle.txt
N-Triples: 831-1.0075100-rdf-ntriples.txt
Original Record: 831-1.0075100-source.json
Full Text
831-1.0075100-fulltext.txt
Citation
831-1.0075100.ris

Full Text

EXPERIMENTAL FIELD MANIPULATIONS OF STREAM TEMPERATURES AND SUSPENDED SEDIMENT CONCENTRATIONS: BEHAVIOURAL AND PHYSIOLOGICAL EFFECTS TO JUVENILE CHINOOK SALMON. by JASON TREVOR QUIGLEY B.Sc, Simon Fraser University, 1994 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTERS OF SCIENCE in THE FACULTY OF GRADUATE STUDIES THE FACULTY OF FORESTRY Department of Forest Sciences We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2003 © Jason Trevor Quigley, 2003 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 The University of British Columbia Vancouver, Canada Date r lU2oo3> DE-6 (2/88) A B S T R A C T The early life history behaviour of juvenile chinook (Oncorhynchus tshawytscha) rearing in the Torpy River, British Columbia was investigated to assist in interpreting the significance of the effects of forest harvesting on juvenile chinook. Investigations of side-channels, tributaries and the mainstem Torpy River in 1997 and 1998 suggested that groundwater controlled, non-natal tributaries were the highest value habitats in all seasons (spring, summer, fall/winter). Small, non-natal tributaries contained both the greatest abundance and largest individuals in terms of length and mass for all age classes. Tributaries also provided the longest growing season and juveniles in these habitats displayed very high residency. Side-channels consistently provided the lowest value habitats. Juvenile chinook over-wintered in groundwater controlled non-natal tributaries, or emigrated from the mainstem in August. The exodus of juvenile chinook from the mainstem was likely triggered by sedimentation of substrate compartments coincident with summer low flows, precluding successful over-wintering. I simulated increases in suspended sediment concentrations and stream temperatures in side-channels and tributaries, to mimic those induced by forest harvesting and quantified their biological effects on juvenile chinook salmon. Experimentally elevated suspended sediment concentrations in un-logged tributaries (602 mg/L) and side-channels (2797 mg/L) resulted in physiological and behavioural sub-lethal stress responses in juvenile chinook. Even though tributary sedimentation trials were stressful to juvenile chinook, these individuals endured these conditions rather than exhibit an avoidance response. Side-channel sediment trials invoked upstream avoidance in juvenile chinook. Small increases in stream temperatures (0.97 - 3.87 °C), representative of post-forest harvesting conditions, also caused sub-lethal stress responses in juvenile chinook salmon. Temperature increases did not exceed the preferred range for the species indicating increased variation in temperature typical of post forest harvesting conditions could have profound effects ii on juvenile salmonids. Sub-lethally stressed fish in side-channel heating trials displayed upstream avoidance and aggregative behaviour. Avoidance was not part of the behavioural repetoire of tributary rearing individuals. Ecological motivation, due to the high value of tributaries, likely caused juvenile chinook to temporarily override adaptive avoidance responses and endure sub-lethally stressful conditions. The propensity for tributary rearing juvenile chinook to endure obviously unfavourable conditions makes them vulnerable to forest harvesting induced changes to water quality. The ecological relevance of sub-lethal stress and consequences of relatively small changes in water quality from land use activities, in the context of life history behaviour, are discussed. TABLE OF CONTENTS Page ABSTRACT : ii TABLE OF CONTENTS iv LIST OF FIGURES vii ACKNOWLEDGEMENTS ix CHAPTER 1: GENERAL INTRODUCTION AND THESIS OVERVIEW 1 CHAPTER 2: JUVENILE CHINOOK EARLY LIFE HISTORY BEHAVIOUR 10 INTRODUCTION 10 STUDY SITES 14 METHODS 15 Water Quality 15 Fish Sampling 16 DATA ANALYSIS 18 RESULTS 19 Age Class and Community Structure 19 Temperature : 20 Suspended Sediment Concentration 21 Post-Emergence Habitat Selection 21 Fork Length 21 Age 0+ 21 Age 1+ 22 Mass 23 Age 0+ 23 Age 1+ .' 24 Condition Factor 24 Age 0+ : 24 Agel+ 24 Catch Per Unit Effort '. 25 Age 0+ 25 Age 1+ ' . . . . J 25 Growth rates 26 Age 0+ 26 Recapture of Marked Fish 27 Biomass and Density 28 DISCUSSION ...28 Habitat Value 28 iv Alternate life history strategies 39 Forest Harvesting Implications 42 CHAPTER 3: EXPERIMENTAL FIELD MANIPULATIONS OF SUSPENDED SEDIMENT CONCENTRATIONS IN TRIBUTARIES AND SIDE-CHANNELS 50 INTRODUCTION 50 METHODS 54 Experimental Procedures 56 Data Extraction 57 DATA ANALYSIS 60 RESULTS 61 Suspended sediment additions 61 Stress 62 Behaviour 63 Movement 63 Coughing : 64 Feeding 64 Yawning 65 Aggression and Fleeing 65 Fishmin and Total Fish 66 DISCUSSION 66 Stress 66 Movement and Behaviour 68 Forest Harvesting Implications 74 CHAPTER 4: EXPERIMENTAL FIELD MANIPULATIONS OF TRIBUTARY AND SIDE-CHANNEL TEMPERATURES 84 INTRODUCTION 84 METHODS 88 Experimental Procedures 90 Data Extraction 92 DATA ANALYSIS 95 RESULTS 96 Temperature manipulations 96 Stress : 96 Behaviour 97 Movement 97 Feeding 98 Yawning 99 Aggression and Fleeing 100 Fishmin and Total Fish 101 v DISCUSSION 102 Stress 102 Movement and Behaviour 105 Forest Harvesting Implications I l l CHAPTER 5: THESIS SUMMARY 117 LITERATURE CITED 123 vi LIST OF FIGURES Page Figure 1.0. Location of the Torpy River Watershed in British Columbia, Canada 8 Figure 1.1. Location of the Torpy River Watershed in the Upper Fraser River 9 Figure 2.0. Diagram representing the three macrohabitats (mainstem, tributaries, side-channels) included in the fish sampling program 44 Figure 2.1. Mean density of predators and competitors of juvenile chinook in each macrohabitat (A) in 1998 and in the mainstem (B) from June 12 to October 23, 1998 (n =4). Error bars represent ± 1 SE. The SE for the density of chinook in tributaries that extends off the chart is ±0.19 45 Figure 2.2. Mean (daily) and maximum water temperatures in the mainstem and mean temperature of tributaries of the Torpy River, July 1996 - October 1998 46 Figure 2.3. Mean monthly suspended sediment concentration in each macrohabitat pooled for 1997 and 1998. Error bars represent ± 1SE. The SE for side-channels in September that extends off the graph is ± 153.3. Means are based on daily means in the mainstem and weekly means in tributaries and side-channels 47 Figure 2.4. Mean fork length of age 0+ (A) and 1+ (B), mean mass of age 0+ (C) and 1+ (D), and mean CPUE of age 0+ (E) and 1+ (F) juvenile chinook captured in 1997 and 1998, pooled seasonally and annually. Error bars represent ± 1 SE. Bars with similar letters did not differ. Note that "all dates" and "seasons" were analysed seperately and thus cannot be contrasted. Missing side-channel bars in the 1+ graphs are because this age class was never captured. All other missing bars indicate a zero value. Means are based on weekly values.. 48 Figure 2.5. Mean growth rates of age 0+juvenile chinook for 1997 and 1998, pooled annually and seasonally for different macrohabitats. Error bars represent ± 1 SE. The SE for the side-channel in fall/winter that extends off the graph is ± 2.2. Bars with similar letters did not differ (see figure 2.4. for details) 49 Figure 3.0. Diagram (A) representing the macrohabitats (tributaries, side-channels) in which experiments were conducted. Schematic (B) of general experimental design (not to scale). The location of sediment manipulations is indicated by the asterik 76 Figure 3.1. Temporal methods of behavioural data extraction from underwater video recordings. Each phase of the experiment was divided into 10 equally spaced intervals (e.g. 6 min. intervals in Pre-Sediment, 9 min. intervals in Sediment, 6 min. intervals in Post-Sediment) and random samples were extracted from each interval. The magnification of the Sediment phase illustrates the systematic random approach. Black bars are randomly selected intervals for data extraction 77 vii Figure 3.2. Mean Cortisol concentrations during tributary trials with the addition of differing suspended sediment concentrations (indicated in parentheses). Means are based on the number of fish collected (n = 8 for each) 78 Figure 3.3. The contrast estimates calculated from mean Cortisol concentrations plotted against suspended sediment concentration of each experiment. Note that the contrast estimate could not be calculated for the 97/09/30 trial because the Pre-Treatment value was missing. The regression model and coefficient of determination are presented 79 Figure 3.4. Mean Cortisol concentration concentrations (± 1 SE) for the 98/08/02 side-channel trial (A) and the 98/08/16 side-channel trial (B). Means are based on number of fish collected (n = 8 for each) 80 Figure 3.5. Mean total number of fish moving during tributary (A) and side-channel (B) sediment trials. Bars with similar letters did not differ. Error bars represent ± 1 SE. Means are based on the number of trials (n = 4 and n = 2 respectively) 81 Figure 3.6. Mean frequency of juvenile chinook coughing per minute during tributary sediment experiments. Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on the number of trials (n = 4) 82 Figure 3.7. Mean feeding attempts/minute for tributary sediment experiments. Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on the number of trials (n = 4) -. 83 Figure 4.0. Diagram (A) representing the macrohabitats (tributaries, side-channels) in which experiments were conducted. Schematic (B) of general experimental design (not to scale). The location of temperature manipulations is indicated by the asterik 113 Figure 4.1. Temporal methods of behavioural data extraction from underwater video recordings. Each phase of the experiment was divided into 10 equally spaced intervals (e.g. 6 min. intervals in Pre-Heat, 9 min. intervals in Heating, 6 min. intervals in Post-Heat) and random samples were extracted from each interval. The magnification of the Heating phase illustrates the systematic random approach. Black bars are randomly selected intervals for data extraction 114 Figure 4.2. Mean total movement of juvenile chinook during tributary (A) and side-channel (B) heating trials. Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on three and two trials respectively 115 Figure 4.3. Mean feeding attempts per minute during tributary trials (A), side-channel trials (B), mean aggression per minute during tributary trials (C), mean fleeing per minute during tributary trials (D), and mean fishmin during side-channel trials (E). Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on number of trials (sidechannels: n = 2; tributaries: n = 5) 116 viii Acknowledgements I am grateful to my supervisor, Dr. Scott Hinch, and the other members of my supervisory committee, Drs. Max Blouw, Michael Healey, and Steve Macdonald, for their support, patience and advice. Special thanks to my former managers at DFO, Bruce MacDonald and Otto Langer, who were supportive of my desire to undertake graduate studies on a part time basis while working full time. Field assistance was provided by Vesna Kontic, Paul Welch, and Laura Genn. Jason Bourgeois and Dr. Mark Shrimpton provided technical support. Lana Klassen assisted with data extraction of underwater video recordings. Thanks to the staff at CANFOR (Prince George) for their support of this research. Financial support was provided from Forest Renewal British Columbia. ix CHAPTER 1: GENERAL INTRODUCTION AND THESIS OVERVIEW The effects of forest harvesting on fish and fish habitat has been well documented. Of anthropogenic activities, the. impacts of forest harvesting on fish habitats can be considered one of the most widespread in British Columbia (Waldichuk 1993). Physical changes to fish habitat as a result of forest harvesting can include increased stream temperatures, increased light transmission, changes to nutrient regimes, changes to primary producers, altered macro-invertebrate densities, hydrological changes, changes in sediment inputs, altered large woody debris recruitment and morphological changes to the channel (Meehan 1991). What is less well documented is the relationship between these physical processes and their biological repercussions, such as the nature of physiological and behavioural impacts to juvenile salmonids, and whether they occur prior to, concurrent with, or after changes in habitat quality are perceptible. Past research regarding the effects of forest harvesting on fish populations has been mostly restricted to rainfall dominated coastal watersheds (Hartman and Scrivener 1990; Hetrick et al. 1998; Hall and Lantz 1969; Murphy et al. 1986; Beschta et al. 1987; Toews and Moore 1982a,b). Even though the future timber supply in the northern hemisphere is predicted to come from boreal and sub-boreal forests in interior regions (Bryant et al. 1997), the effects of forest harvesting on fish populations in snowpack dominated, interior watersheds are not well studied. The complexity and interconnections of forest harvesting related physical habitat changes and biological repercussions has challenged researchers abilities to draw causal connections. Many of the documented forest harvesting induced physical changes to fish habitat have compound effects (Hartman and Scrivener 1990), which can occur concurrently or asynchronously, and the resultant biological effects can operate on different time scales of years up to centuries (Meehan 1991). The compound effects of forest harvesting on the aquatic ecosystem is well illustrated by considering an aspect such as riparian removal which can alter 1 stream temperatures, light transmission and channel morphology. These multiple effects may act individually on aquatic biota on one or several trophic levels, either in a positive, negative, or neutral manner, or they may act synergistically. For example, Slaney et al. (1977) concluded that elevated water temperatures in clear-cut streams was the main cause of increased fish growth rates in those streams. However, it has been shown that increased light transmission from riparian removal can result in increased primary production (Hansmann 1969; Hetrick et al. 1998b; Lowe et al. 1986) which in turn can influence the next trophic level, resulting in enhanced macro-invertebrate densities (Bilby and Bisson 1992; Hetrick et al. 1998a; Newbold et al 1980), and can also improve salmonid prey capture efficiency (Wilzbach et al. 1986). Therefore, due to the many physical effects on aquatic ecosystems that co-vary with riparian removal (light, nutrients, etc.) and their multiple effects on lower trophic levels such as primary production and macro-invertebrate community structure, many of the conclusions drawn in the literature would benefit from experimental work. As a result of the many factors that co-vary with forest harvesting induced physical changes, and considering logistical constraints of experimental field work, many studies have relied upon laboratory studies or artificial streams to infer the biological effects from changes to components of fish habitat (Bradford and Taylor 1997; Chapman 1962; Hartman 1963; Bisson and Bilby 1982; Gregory 1993,1994; Redding et al. 1987; Berg and Northcote 1985; Sigler et al. 1984; Servizi and Martens 1992,1991; Servizi and Gordon 1990; McLeay et al. 1987; Brett 1952, 1970, Bisson and Davis 1976; Coutant 1973; Gibson 1988). Although laboratory studies and experiments conducted on artificial streams are valuable to provide insights to causal relationships between characteristics of fish habitat and effects on fish, they generally lack ecological realism (Adams 1990). 2 Of the studies that have had an experimental field approach, the majority have involved manipulations to riparian zones (Hall and Lantz 1969; Keith et al. 1998; Hetrick et al. 1998a, 1998b; Murphy et al. 1986; Scrivener 1988; Young et al. 1999), which were all confounded to a certain degree when attempting to elucidate causal mechanisms, by the compound effects that result from this experimental treatment. Holtby and Newcombe (1982) separated the confounding effects on stream temperature from forest harvesting and climatic variability with an empirical model. Few studies have involved experimental manipulations in the field to one physical effect, in isolation of co-variates (Wilzbach et al. 1986; Mason 1976; Lowe et al. 1986; Keith et al. 1998; Nislow et al. 1998; Meehan et al. 1987; Solazzi et al. 2000; Hillman et al. 1987). Determination of the linkages between physical changes induced by forest harvesting, and quantification of their biological repercussions in a natural field setting, are the primary focus of this study. This thesis is divided into three main chapters, each of which I hope to publish as separate papers. Consequently, some repetition exists among chapters with respect to site descriptions and methodologies. Experimental field manipulations to water quality characteristics were conducted on small tributary streams and side-channels to simulate changes induced by forest harvesting, and quantify their biological effects. Based on a review of the literature, the novel approach taken in my thesis represents the first time that experimental field manipulations to stream temperatures and suspended sediment concentrations have been conducted in a natural stream. The interconnected nature and compound effects of forest harvesting on the aquatic ecosystem, underscore the value in experimentally manipulating one factor in order to uncouple its effects from co-variates. In this thesis, conducting experimental manipulations to instream temperatures, suspended sediment concentrations, and turbidity levels enabled the isolation of the other variables that often covary with these factors due to climatic variation (discharge, cloud cover, precipitation, air temperature, 3 insolation, groundwater flows, etc.). This innovative research design also enabled the isolation of these variables from the other confounding effects, (including each other) often associated with forest harvesting activities, (i.e. increased light, altered hydrology, altered habitat diversity, etc.) on fish populations. As part of my thesis, (Chapters 3 and 4), I conducted experiments on both tributary streams and side-channels to the mainstem Torpy River, British Columbia to minimize confounding factors. Streams and side-channels were netted into experimental and control sections, separated by a series of weirs so that upstream and downstream movement of juvenile chinook could be recorded. Biological indicators of sub-lethal stress were recorded at multiple levels of biological organization, including physiological and ecological. Measurements included upstream and downstream avoidance, underwater video footage from which the following behaviours were extracted: feeding attempts, coughing, aggression, fleeing, and yawning, and finally whole body Cortisol was measured as an indicator of physiological stress. In order to be able to determine the impact of forest harvesting practices on juvenile chinook populations at both a local and landscape level of organization, I felt it was first essential to quantify the early life history behaviour of the chinook population in question (see Chapter 2). In North America, chinook salmon are the least abundant of the Pacific salmon (Healey 1991), yet display one of the most diverse life history patterns of all the Pacific salmon species (Miller and Brannon 1981). There are three general categories of early life history strategies displayed by chinook salmon, that collectively represent almost as much variation as that among different species of Pacific salmon. "Immediate-type" fry migrate directly to the estuary upon emergence, (e.g. the Harrison River stock) (Levy et al. 1979). "Ocean-type" fry rear in freshwater for 60 to 150 days prior to downstream migration to the ocean in the fall of their first year, (e.g. the 4 Thompson River stock (Fraser et al. 1982; Healey 1991). "Stream-type'" fry rear in freshwater for one to two years until smoltification occurs (Healey 1991). There is tremendous plasticity among different populations of chinook salmon with respect to the proportion of individuals that exhibit each general category of early life history strategies (Healey and Jordan 1982). More specific to this thesis, there is extensive phenotypic plasticity within a given category of early life history strategy, with respect to the ecology and behaviour of a particular population. Macro and microhabitat selection, freshwater juvenile migration patterns and timing, as well as the mechanisms that initiate these behavioural responses vary considerably both among different northern interior British Columbian watersheds as well as when compared to coastal systems, for a given category of early life history behaviour (Bradford and Higgins 2001; Healey 1991; Levings and Lauzier 1991; Swales et al. 1986). The juvenile chinook in this study are of the "stream type" category relying on freshwater habitat for one to t two years, and accordingly would be susceptible to forest harvesting induced changes in the aquatic environment for the duration of this life phase. Within the "stream type" category, juvenile chinook have been documented to have a diverse range of macrohabitat preferences and to display distinct seasonal habitat preferences that change over time based on ontogeny and environmental conditions (Anonymous 1987; Bradford and Higgins 2001; Emmet 1989; Healey 1991; Levings and Lauzier 1991; Murray et al. 1981; Rosberg et al. 1981; Reimers 1968; Russel et aL 1983; Swales et al. 1986). The majority of studies document "stream type" juvenile chinook to overwinter in the mainstem of large rivers (Anonymous 1987; Emmet 1989; Healey 1991; Levings and Lauzier 1991; Murray et al. 1981; Russell et al. 1983; Swales etal. 1986). However, lacustrine habitats (Healey 1991), side-channels (Bustard 1986; Sommer et al. 2001) and tributaries (Anonymous 1990) have also been reported as overwintering habitat. 5 Juvenile chinook early life history behaviour, seasonal habitat utilisation, and habitat values in the Torpy River were investigated in Chapter 2 to assist in interpreting the significance of the effects of forest harvesting in different macrohabitats on juvenile chinook, as simulated by experimental field manipulations in Chapters 3 and 4. STUDY LOCATION The Torpy River, located in the northern interior of British Columbia (Figure 1.0.), is approximately 100 kilometres in length. It flows into the north side of the Fraser River, just upstream of Dome Creek, which is approximately 75 kilometres northwest of McBride, British Columbia. The Torpy River has a drainage area of 128, 500 hectares (Mac Donald et al. 1996) and flows through the Rocky Mountain Trench/McGregor Plateau/Park Ranges physiographic area. This physiographic region is characterized by flat or gently rolling surfaces with poorly organized drainage patterns, and with mid-summer flows primarily a function of both groundwater seepage and rainfall (Rosberg and Aitken 1981). This watershed encompasses several biogeoclimatic zones, namely SubBoreal Spruce, primarily in the headwaters, and a mixture of Engelmann Spruce-SubAlpine Fir and Interior Western Hemlock in the lowlands. The lower Torpy River is a broad, U-shaped valley, with heavy glacial lacustrine deposits, resulting in highly erodible silts and clays. Investigations of the early life history behaviour of juvenile chinook and experimental manipulations to water quality were conducted in small tributaries and side-channels of the lower Torpy River (Figure 1.1.). The lower Torpy River consists of a series of tortuous meanders that flow through a broad floodplain composed of lacustrine deposits (Mac Donald et al. 1996). The lacustrine deposits in the valley bottom result in unstable streambanks, high suspended sediment levels in the mainstem and in general, low resiliency to environmental change (Mac Donald et al. 1996). This system is subject to low flows 6 3 * 3 in the summer (11.13 m /s) and the winter (3.65 m /s) and has a mean annual discharge of 53.7 m3/s and a mean flood discharge of 291.4 m3/s (Rood and Hamilton 1995). The Torpy River contains chinook salmon (Oncorhynchus tshawytscha), rainbow trout (Oncorhynchus mykiss), bull trout (Salvelinus confluentus), prickly sculpin (Cottus Asper), northern squawfish (Ptychocheilus oregonensis), burbot (Lota lota), white sucker (Catostomus commersoni), and rocky mountain whitefish (Prosopium williamsoni). The 1981-1992 average annual escapement of adult chinook salmon to the Torpy River watershed was approximately 2,136 with a historical maximum escapement of 4000 (Mac Donald et al. 1996). Chinook salmon arrive in the Torpy system in late July and spawn in late August primarily throughout the lower Torpy, and in some of the major tributaries, such as Walker Ck, Humbug Ck, and the West Torpy River (Rosberg and Aitken 1981). Since 1956, the Torpy River watershed has been extensively logged, with operations peaking during the 1970's (Rosberg et al. 1981). As of 1996, approximately 15% of the entire watershed, and 50% of the lower Torpy had been logged (Mac Donald et al. 1996). Forest harvesting in the lower Torpy river has concentrated almost exclusively in the low gradient floodplain and riparian areas, which typically contains some of the most valuable juvenile fish habitat. Logging roads and stream crossing networks are extensive. 7 Figure 1.0. Location of the Torpy River Watershed in British Columbia, Canada. 8 Torpy Drainage Basin Torpy Drainage Basin 6000000m.N. 10km Figure 1 . 1 . Location of the Torpy River Watershed in the Upper Fraser River drainage. 9 CHAPTER 2: JUVENILE CHINOOK EARLY LIFE HISTORY BEHAVIOUR INTRODUCTION . In North America, chinook salmon are the least abundant of the Pacific salmon (Healey 1991), yet display one of the most diverse life history patterns of all the Pacific salmon species (Miller and Brannon 1981). There are three general categories of early life history strategies displayed by chinook salmon, that collectively represent almost as much variation as that among different species of Pacific salmon. "Immediate-type" fry migrate directly to the estuary upon emergence, (e.g. the Harrison River stock) (Levy et al. 1979). "Ocean-type" fry rear in freshwater for 60 to 150 days prior to downstream migration to the ocean in the fall of their first year (e.g. the Thompson River stock) (Fraser et al. 1982; Healey 1991). "Stream-type" fry rear in freshwater for one to two years until smoltification occurs (Healey 1991). There is tremendous plasticity among different populations of chinook salmon with respect to the proportion of individuals that exhibit each category. For example, both the Nanaimo River and Shuswap Lake chinook stocks contain individuals, living in sympatry, that display "ocean-type" behaviour in addition to those that exhibit "stream-type" behaviour (Healey and Jordan 1982). Some intra-specific variation between life history categories has also been shown to be heritable (Carl and Healey 1984; Clarke etal. 1992, 1994; Taylor and Larkin 1986; Taylor 1988, 1990). Variation in life history strategies in salmonids has also been shown to be phenotypic in nature (Nordeng 1983; Randall et al. 1987). Salmonids can facultatively adjust anadromous migratory behaviour to varying levels of aquatic productivity; for example, manipulations to food supply resulted in variations in the proportions of resident salmonids and anadromous migrants (Nordeng 1983). There is extensive phenotypic plasticity within each category of chinook early life history strategy, with respect to the ecology and behaviour of a particular population. Macro and microhabitat selection, freshwater juvenile migration patterns and timing, as well as the 10 mechanisms that initiate these behavioural responses vary considerably both among different northern interior British Columbian watersheds as well as when compared to coastal systems, for a given category of early life history behaviour (Bradford and Higgins 2001; Healey 1991; Levings and Lauzier 1991; Swales et al. 1986). The juvenile chinook in this study are of the "stream-type" category relying on freshwater habitat for one to two years, and accordingly would be susceptible to forest harvesting induced changes in the aquatic environment for the duration of this life phase. Within the "stream-type" category, juvenile chinook have been documented to utilize a diverse range of macrohabitats and to display distinct seasonal habitat preferences that change over time based on ontogeny and environmental conditions (Anonymous 1987; Bradford and Higgins 2001; Emmet 1989; Healey 1991; Levings and Lauzier 1991; Murray etal. 1981; Rosberg et al. 1981; Reimers 1968; Russel et al. 1983; Swales et al. 1986). Seasonal habitat redistributions have been shown to confer benefits to those individuals that switch habitats, in terms of growth and survival (Bustard andNarver 1975; Tschaplinski and Hartman 1983). The majority of early life history studies on stream-type chinook have occurred during the summer period, likely due to ease of sampling and abundance of individuals (Anonymous 1987; Everest and Chapman 1972; Murphy et al. 1989; Lister et al. 1981; Murray et al. 1981; Reimers 1968; Rosberg et al. 1981). Most have focussed on large river systems such as the Fraser River (Levings and Lauzier 1991), Taku River (Murphy et al. 1989), Sacramento River (Sommer et al. 2001), andNechako River (Anonymous 1998, 1990, 1987; Emmet 1989; Lister etal. 1981; Russell et al. 1983), and have only investigated utilisation within one macrohabitat, often the natal mainstem (Anonymous 1998; Bradford and Higgins 2001; Everest and Chapman 1972; Levings and Lauzier 1991; Scrivener et al. 1994). These studies provide an incomplete picture of the early life history of the stream-type race, particularly considering the magnitude of variation within a 11 given life history strategy that is displayed by the species. The majority of these studies document that stream-type juvenile chinook overwinter in the mainstem of large rivers (Anonymous 1998, 1987; Chapman and Bjornn 1969; Emmet 1989; Everest and Chapman 1972; Healey 1991; Levings and Lauzier 1991; Murray et al. 1981; Russell et al. 1983; Swales et al. 1986), yet this finding may be biased by the greater proportion of studies that have focussed on this macrohabitat type. Lacustrine habitats (Healey 1991), side-channels (Bustard 1983; Sommer et al. 2001) and small non-natal tributaries (Keefe et al. 1995; Anonymous 1990) have also been reported as overwintering habitat. Few studies, purporting to document life history strategies of stream-type juvenile chinook, have considered all the macrohabitat opportunities available to juvenile chinook from the early spring post-emergence phase, through summer, as well as during the winter season [exceptions include research on Red River, Idaho (Hillman et al. 1987); the Nechako River, British Columbia (Russell et al. 1983); and the Grande Ronde River, Oregon (Keefe et al. 1995)]. In northern climates, such as the Upper Fraser River, not only do harsh environmental conditions of winter impose serious challenges to juvenile chinook, but the majority of forest harvesting operations occur during the winter, at a time when habitats are difficult to locate and protect, and the biological resiliency of fish is very low (Cunjak 1988; Cunjak et al. 1987; Cunjak and Power 1986). This chapter of my thesis focuses on documenting the early life history behaviour of juvenile chinook throughout the spring, summer, and fall/winter period and across multiple macrohabitats, to determine seasonal habitat utilisation and habitat values. The total habitat utilized by a species can be partitioned into preferred habitat and non-preferred areas, the latter affording sub-optimal conditions which translate into decreased growth, size, fitness and survival (Hearn 1987). I used juvenile chinook abundance (Heifetz et al. 1986; Murphy et al. 1989), growth and size as indicators of habitat value, due to the documented links between these 12 attributes and survival to adulthood (Beckman et al. 1998; Bilton et al. 1982; Bilton 1978, 1984; Cederholm and Scarlett 1981; Holtby and Hartman 1982; Swales et al. 1988). In general, newly emerged fry often cluster together in patchy aggregates, exhibit schooling behaviour, and are most abundant in shallow habitats (0.15- 0.3 m) with low velocity (< 0.15 m/s) and small substrate particle size, gradually shifting to deeper, faster water as they grow (Anonymous 1987; Chapman and Bjornn 1969; Everest and Chapman 1972; Healey 1991; Hillman et al. 1987; Rosberg et al. 1981; Russell et al. 1983). Dispersal of newly emerged fry is generally in the downstream direction (Bradford and Taylor 1997; Healey 1991; Lister et al. 1981; Taylor 1988); however, Taylor (1988) found that negative rheotaxis behaviour of stream-type juvenile chinook subsided after three months. Indeed, upstream migrations of age 0+ chinook fry shortly after emergence, presumably as a dispersal mechanism in search for suitable rearing habitat, has been documented in the Nechako River (Anonymous 1998; Russell et al. 1983). The underlying mechanism driving dispersal of post-emergent fry is not known, but it is generally considered to be either passive displacement due to peaks in the hydrograph (Healey 1991; Irvine 1986), and/or a function of social interactions whereby size governs the result of successful interactions (Chapman 1962; Hartman et al. 1982; Mason 1976; Reimers 1968). In either scenario, larger despotic individuals hold positions and smaller individuals are displaced downstream. Bradford and Taylor (1997) reported a contrasting finding in their experimental work. Emigrant chinook fry were significantly larger than resident individuals (Bradford and Taylor 1997). Further, emigrants were distributed in aggregates in the experimental channel rather than distributed evenly, as expected if downstream movement was stimulated by behavioural interactions (Bradford and Taylor 1997). The majority of stream-type chinook fry migrate at night (Healey 1991; Murray et al. 1981; Taylor et al. 1994). During downstream 13 migrations chinook fry prefer stream margins to the high velocity in the centre of the channel, and the water surface when river depths exceed 3m (Anonymous 1987; Healey 1991; Healey and Jordan 1982). The former points, coupled with the experimental findings of Bradford and Taylor (1997), support Healey's (1991) suggestion that post-emergent fry dispersal may involve active selection. My objective in this chapter was to determine the early life history behaviour of juvenile chinook, specifically the habitat value of tributaries, side-channels and the mainstem. Macrohabitat segregation is most pronounced in the spring and summer period when population densities are highest (Hartman 1965). Therefore, I used differences in the attributes of juvenile chinook populations utilizing different macrohabitats in the Torpy River drainage during these periods, as well as winter, to infer habitat value. The indicators of habitat value (size, growth rate, condition factor, abundance, residency) were investigated among macrohabitats, annually and seasonally, in relation to abiotic factors such as stream temperature and suspended sediment concentrations, and biotic factors including density of competitors and predators of juvenile chinook. I predict that indices of habitat value will be greater in tributaries than the mainstem and side-channels both seasonally and annually. STUDY SITES I sampled fish and environmental variables in three macrohabitats (side-channels, tributaries and the mainstem) (Figure 2.0.) located along a 10 km section of the Lower Torpy River, beginning approximately 40 km upstream from the confluence with the Fraser River. The mainstem of the Lower Torpy River had an average channel width of 30 m and a substrate composition of 30% gravels and 70% fines. Large woody debris complexes, as well as large boulders were infrequent. The mainstem was low gradient (< 1%) and characterized by deep 14 glides, separated by riffles. Gravel in the riffle sections was highly embedded with fines. Avian predators, including Belted Kingfishers (Megaceryle alcyon), Great Blue Herons (Ardea herodias) and Common Mergansers (Mergus serratof) were observed frequently in the mainstem and side-channels yet never in the tributaries. The abundance and efficiency of these piscivorous birds led Rosberg et al. (1981) to call them the most successful chinook fry predators in the Upper Fraser area. Side-channels were braids of the mainstem, having similar water quality characteristics albeit at substantially reduced flows. The two side-channels in the sampling program were approximately 3 - 4 m in width and 50 m in length, characterized by low velocity glides and deep pools with a substrate composition of 100% fines. Large logjams controlled the inlets of both side-channels. Tributaries were all low gradient (1 - 2%), approximately 1-2.5 m in width, with a substrate composition of approximately 70% fines and 30% gravels. Cover was prevalent in tributary streams in the form of undercut banks, over-hanging vegetation, deep pools and woody debris. Tributary streams drained valley slopes and flowed across the Lower Torpy Valley floodplain to the Torpy River. METHODS Water Quality I recorded water temperature in each macrohabitat for three years by remote, self contained temperature data loggers (Onset instruments). The loggers were programmed to measure and record temperature every five minutes. An automated, continuous water quality monitoring station, which consisted of a programmable data logger (Starlogger) linked to input sensors, was installed on the mainstem. Inputs included a submersible pressure transducer to measure stage, a nephelometric turbidity probe, and an automated pump sampler. The station was 15 programmed to record stage and turbidity data hourly, and to collect daily water samples from the pump sampler. One litre water samples were collected weekly to determine suspended sediment concentrations in the other macrohabitats (tributaries and side-channels). Suspended sediment concentrations (mg/L) were determined in the lab by filtration. Fish Sampling Fish sampling was conducted weekly from June 6th to December 9th in 1997, and May 6th to October 24th in 1998, using Gee type minnow traps baited with cat food. Minnow traps were selected as the most appropriate sampling technique due to their year-round functionality, low degree of disturbance, and effectiveness in providing relative estimates of juvenile salmonid abundance, for fish less than 130 mm in length (Bloom 1976; Bryant 2000; Swales et al. 1986). The fish sampling program consisted of 35 sampling sites partitioned between three macrohabitats. The mainstem had 14 sampling sites in varying locations including along the channel margin, associated with large woody debris, immediately upstream and downstream of tributary streams, and in the centre of the channel both on the stream substrate as well as suspended in the water column. There were six sites in two different side-channels, and 15 sites in three different tributaries, in which traps were located just upstream from the confluence with the Torpy River and extended approximately 1 km upstream. Minnow traps were set for 24 hours and captured fish were anaethesized with Alka Seltzer, identified, marked and enumerated. Fork length (FL) (± 1 mm) and wet weight (± 0.01 g), were recorded and Fulton's condition factor was calculated (weight/length3) for each fish. Al l fish were released after capture. Juvenile chinook were marked by a fin clip diagnostic for each tributary, side-channel, and the mainstem, thus enabling me to determine residency and movement among different macrohabitats. Data were separated into 0+ and 1+ age classes by length and 16 weight histograms. Age classes were confirmed by comparison to previously published data of Torpy River juvenile chinook lengths and weight at age, determined by scale analyses (Rosberg et al. 1981). Segregation of cohorts was necessary to eliminate the confounding influence of age-class differences upon indicators of habitat value (e.g. growth rates, length, and mass). Absolute growth rates for juvenile chinook captured in each macrohabitat were calculated using the following formula: G = (FL 2 - FLi) /At , where G is the absolute growth rate (mm/week), FL 2 is the mean fork length (mm) at time 2, and FLi is the mean fork length at time 1, and At is the time interval (weekly). Catch per unit effort (CPUE) was calculated by dividing the total number of juvenile chinook captured in each macrohabitat on each sampling date by the number of minnow traps set in that macrohabitat. Passive sampling techniques, such as minnow trapping, decrease in effectiveness as water temperatures decline, because fish metabolism and hence foraging activity declines. Active sampling techniques were employed on June 11 th, July 12th , August 22 n d and October 23 rd, 1998 in the macrohabitats to serve as a correction factor, and determine the effectiveness of passive techniques at representing actual population densities (i.e. cross-reference to CPUE). Tributaries were sampled by electroshocking (Smith Root Model 12c) a section of stream isolated by enclosure nets using a triple pass method (Seber and LeCren 1967). A minimum of two replicates (sections of tributaries) per sampling date were conducted in the tributaries. Captured fish were identified, enumerated, measured (fork length) and weighed (wet weight) after each sample pass. The area sampled was recorded to calculate densities and biomass of fish per unit area. Side-channels and the mainstem were sampled by three pass seining with a 15 x 2 m seine (3.2 mm mesh) and a 25 m x 2 m seine respectively. A minimum of two replicates (seine sites) per sampling date were conducted in both side-channels and the mainstem. Captured fish were 17 identified, enumerated, measured (fork length) and weighed (wet weight) after each sample pass. The area sampled was recorded to calculate densities and biomass of fish per unit area. Active sampling techniques were not size selective as the minnow traps were, and therefore it was possible to determine relative differences in fish community structure in different macrohabitats. Fish, other than juvenile chinook, that were captured in each macrohabitat by active sampling techniques, were categorized into potential competitors and predators, based on ecological relationships and size relative to juvenile chinook. Bull trout, northern squawfish and prickly sculpins were categorized as aquatic predators of juvenile chinook (Scott and Crossman 1973). Competitors of juvenile chinook included rainbow trout and rocky mountain whitefish (Scott and Crossman 1973). DATA ANALYSIS All data were visually inspected for normality and homogenous variances. A General Linear Model (GLM) was used to relate the response variables to the other factors in the study. I used one-way, univariate analyses of variance (ANOVA) to determine model significance for the pooled 1997 and 1998 data and two-way ANOVAs to test for interactions between macrohabitats and season. Least Square Difference (LSD) multiple range tests were used to determine which macrohabitats, or seasons within a given macrohabitat differed in fork length, mass, condition factor, CPUE, or growth rate. Data were pooled into the following temporal categories: a) Post-emergence (May 6, 13, 20, 27, 1998), b) Spring (June 6, 97 to July 25, 97 and May 6, 98 to July 18, 98), c) Summer (Aug 1, 97 to Sept 25, 97 and July 31, 98 to Sept 19, 98), d) Fall/Winter (Oct 3, 97 to Dec 9, 97 and Oct 4, 98 to Oct 24, 98), e) All (pooled 1997 and 1998) 18 Age 1 +juvenile chinook were never captured in side-channel habitats, therefore post-hoc multiple range tests were not required in analyses of these data. In the fall/winter season, 1+ juvenile chinook were not captured in the mainstem, and therefore in analysis of annual differences in mean indicators of habitat value, data from fall/winter tributary 1+ chinook were excluded to prevent biasing the analyses. All tests were considered to be significant to a P <0.05. RESULTS Age Class and Fish Community Structure In 1997 and 1998, a total of 1,893 juvenile chinook were collected during the weekly sampling program, of which 96.3% were 0+ and 3.7% (70) were 1+ in age. Mean density of predator fish species was highest in side-channels, where it was approximately 3.4 times greater than the density in the mainstem and 8.4 times the density in tributaries (Figure 2.1a.). Predator fish species in the mainstem were approximately 2.5 times greater in density than in tributaries (Figure 2.1a.). Aquatic predators in side-channels outnumbered juvenile chinook by 7.6:1 and in the mainstem were virtually at a 1:1 ratio, whereas in the tributaries, juvenile chinook outnumbered predators by approximately 40:1 (Figure 2.1a.). Mean density of competitor fish species was approximately 2.1 times greater in the tributaries compared to the mainstem (Figure 2.1a.). Competitors were never captured in side-channels. Mean density of juvenile chinook in tributaries was approximately 36.5 times greater than in side-channels, and 14.1 times greater than in the mainstem (Figure 2.1a.). The prevalence of agonistic interactions was high in tributary streams, approximately 40% of the juvenile chinook captured in these habitats had severely tattered and torn caudal fins. This condition was not observed in fish captured in other macrohabitats (mainstem, side-channel) and was also not observed in rainbow trout captured in tributary streams, presumably due to habitat segregation. In the mainstem the density of 19 competitors and juvenile chinook decreased over 10 and 51 times respectively between July and August (Figure 2.1b.)- The density of predators increased over 50 times in the mainstem during July and August (Figure 2.1b.). Temperature Both the monthly mean and maximum stream temperatures in the mainstem were higher than the tributaries from June to September 1997 and 1998 (Figure 2.2.). Between July 23 to August 5, 1998, the mainstem temperature exceeded 20 °C each day (14 days) for 12 - 16 hours per day. The opposite was true during the winter with the mainstem being colder than the tributaries from November to March in 1997 and 1998 (Figure 2.2.). Side-channel temperatures were only recorded from May to October 1998 and were not plotted to prevent obscuring the distinct patterns between tributaries and the mainstem. In general, side-channel temperatures closely matched that of the mainstem. On average, side-channels were slightly warmer than the mainstem during the months of May and October (+3.05 and +1.19 °C, respectively), and 2.35 °C colder than the mainstem during June to September. Suspended Sediment Concentration In both years the mean monthly suspended sediment concentration was approximately equivalent in the three macrohabitats in June. However, from July to October, the mean monthly sediment concentration was higher in the mainstem and side-channels compared to the tributaries (Figure 2.3.). Moreover, the monthly variability in suspended sediment concentration was also higher in the mainstem and side-channels compared to the tributaries (Figure 2.3.). Maximum monthly suspended sediment concentrations in the mainstem averaged 254.81 mg/L and peaked twice at 446.25 mg/L in July and 390.25 mg/L in October. Maximum concentrations in the side-20 channels remained below 34 mg/L except in September when there was a peak of 245.75 mg/L. Maximum concentrations in tributaries remained below 26.5 mg/L . Post-Emergence Habitat Association Early after emergence, during the month of May, juvenile chinook were observed in both tributary and mainstem habitats in shallow margin habitats with little or no velocity. Juvenile chinook (35-40 mm FL) were observed in high density aggregates (up to 25 fish/m2 ) displaying schooling behaviour. From June onwards, juvenile chinook (> 40 mm FL) that were observed rearing in the mainstem, side-channels and tributaries no longer formed loose aggregates, but held territories typical of stream dwelling salmonids. Mean fork length of juvenile chinook rearing in the mainstem did not differ from individuals rearing in side-channels or tributaries during the month of May (ANOVA, F = 2.038, df = 2,37, P = 0.145). I found no difference in mean mass (ANOVA, F = 0.518, df = 2,37, P = 0.6), condition factor (ANOVA, F = 0.493, df = 2,37, P = 0.615), or CPUE (ANOVA, F = 1.127, df = 2,17, P = 0.347) among juvenile chinook captured in different macrohabitats during May. Fork Length Age 0+ I found differences in mean fork length among juvenile chinook rearing in different macrohabitats (ANOVA, F = 43.007, df = 2,1820, P = 0.0005). Fork length differed among all macrohabitats (LSD, P = 0.0005, for all contrasts). Fish captured in tributaries had the largest fork length, mainstem fish were intermediate and side-channel individuals were the smallest (Figure 2.4a.). The two-way ANOVA, with season and macrohabitat as factors, revealed differences in fork length among macrohabitats (F = 3.234, df = 2,126, P = 0.043), and among seasons (F = 84.361, df = 2,126, P = 0.0005), but indicated there was no interaction between 21 season and macrohabitat (F = 0.551, df = 4,126, P = 0.699). In spring juveniles rearing in side-channels were smaller than individuals rearing in tributaries and those rearing in the mainstem (LSD, P = 0.001, for both contrasts) (Figure 2.4a.). There was no difference between mainstem and tributary rearing fish during spring (LSD, P = 0.487). In summer, juveniles rearing in side-channels were smaller than tributary and mainstem rearing individuals (LSD, P = 0.0005, for both contrasts) (Figure 2.4a.). Tributary and mainstem rearing fish did not differ in summer (LSD, P = 0.484). During fall/winter there were no differences in fork length among juveniles rearing in the macrohabitats (LSD, mainstem vs. side-channels P = 0.978, mainstem vs. tributaries P = 0.824, tributaries vs. side-channels P = 0.800). In the mainstem, fork length of juveniles was greater from spring to summer (LSD, P = 0.0005), and summer to fall/winter (LSD, P = 0.017)(Figure 2.4a.). In side-channels, fork length was greater from spring to summer and summer to fall/winter (LSD, P = 0.0005, for both contrasts)(Figure 2.4a.). In tributaries, fork length was greater from spring to summer as well as summer to fall/winter (LSD, P = 0.0005, for both contrasts)(Figure 2.4a.). Age 1+ Juvenile chinook rearing in tributaries had a greater fork length than those rearing in the mainstem (ANOVA, F = 7.396, df = 1,63, P = 0.008) (Figure 2.4b.). The two-way ANOVA, with season and macrohabitat as factors, revealed no differences in mean fork length among seasons (F .= 2.978, df = 2,37, P = 0.063), or among macrohabitats (F = 1.603, df = 1,37, P = 0.213). There was no interaction between season and macrohabitat (F = 0.26, df = 1,37, P = 0.613). 22 Mass Age 0+ I found differences in mean mass among macrohabitats (ANOVA, F = 35.083, df= 2,1820, P = 0.0005). Juveniles rearing in the tributaries had the greatest mass, followed by those rearing in the mainstem and side-channels (LSD, P = 0.0005, for all contrasts) (Figure 2.4c). The two-way ANOVA, with season and macrohabitat as factors, revealed differences in mass among seasons (F = 68.522, df = 2,126, P = 0.0005) and among macrohabitats (F = 4.380, df = 2,126, P = 0.014). There was no interaction between season and macrohabitat (ANOVA, F = 0.657, df = 4,126, P = 0.623). In spring, side-channels contained juvenile chinook with a smaller mass than the mainstem and tributaries (LSD, P = 0.001, for both contrasts) (Figure 2.4c). Mass did not differ between juveniles rearing in the mainstem and tributaries during spring (LSD, P = 0.537). In summer, fish rearing in side-channels had a smaller mass than those rearing in either the mainstem or tributaries (LSD, P = 0.0005, for all contrasts) (Figure 2.4c). Mass did not differ between juveniles rearing in the mainstem and tributaries during summer (LSD, P = 0.602). There was no difference in mass among macrohabitats during the fall/winter season (LSD, mainstem vs. side-channels P = 0.792, mainstem vs. tributaries P = 0.563, side-channels vs. tributaries P = 0.230). In the mainstem, mass of juveniles increased from spring to summer (LSD, P = 0.0005), and summer to fall/winter (LSD, P = 0.037)(Figure 2.4c). In side-channels, mass increased from spring to summer (LSD, P = 0.001) and summer to fall/winter (LSD, P = 0.0005)(Figure 2.4c). In tributaries, mass increased progressively through all seasons (LSD, P = 0.0005, for all contrasts)(Figure 2.4c). 23 Age 1 + Mean mass of age 1 +juveniles rearing in tributaries was greater than those rearing in the mainstem (ANOVA, F = 6.299, df = 1,63, P = 0.015)(Figure 2.4d.). The two-way ANOVA, with season and macrohabitat as factors, revealed no differences in mass among seasons (F = 2.555, df = 2,37, P = 0.091), nor among macrohabitats (F = 2.039, df = 1,37, P = 0.162). There was no interaction between season and macrohabitat (F = 0.368, df = 1,37, P = 0.548). Condition Factor Age 0+ I found no differences in mean condition factor among macrohabitats (F = 0.273, df = 2,1820, P = 0.761). The two-way ANOVA, with season and macrohabitat as factors, revealed no differences in condition factor among seasons (F = 0.206, df = 2,126, P = 0.814), nor among macrohabitats (F = 0.756, df - 2, 126, P = 0.472). There was no interaction between season and macrohabitat (F = 0.902, df = 4,126, P = 0.465). Age 1 + Mean condition factor of age 1+juvenile chinook did not differ among any of the macrohabitats (ANOVA, F = 0.006, df = 1,68, P = 0.941). The two-way ANOVA, with season and macrohabitat as factors, indicated there were no differences in condition factor among seasons (F = 0.818, df = 2,37, P = 0.449), nor among macrohabitats (F = 0.273, df = 1,37, P = 0.604). There was no interaction between season and macrohabitat (F = 0.017, df = 1,37, P = 0.897). 24 Catch Per Unit Effort Age 0+ I found differences in mean catch per unit effort (CPUE) of age 0+juvenile chinook among macrohabitats (ANOVA, F = 30.472, df = 2,165, P = 0.0005). CPUE was higher in tributaries compared to both the mainstem and side-channels (LSD, P = 0.0005, for all contrasts) (Figure 2.4e). CPUE did not differ between the mainstem and side-channels (LSD, P = 0.949). The two-way ANOVA, with season and macrohabitat as factors, revealed differences in CPUE among seasons (F = 5.484, df = 2,159, P = 0.005) and among macrohabitats (F = 40.61, df= 2,159, P = 0.0005). The interaction between season and macrohabitat was significant (F = 5.85, df = 4,159, P = 0.0005) (Figure 2.4e.). In spring, CPUE was greater in the tributaries compared to the mainstem (LSD, P = 0.013) and side-channels (LSD, P = 0.031) (Figure 2.4e). In summer, CPUE was greater in the tributaries compared to both the mainstem and side-channels (LSD, P = 0.0005, for all contrasts) (Figure 2.4e). In fall/winter, CPUE in tributaries was greater than both the mainstem and side-channels (LSD, P = 0.0005, for all contrasts) (Figure 2.4e). In the mainstem, CPUE did not differ from spring to summer (LSD, P = 0.057) or from spring to fall/winter (LSD, P = 0.238), yet declined from summer to fall/winter (LSD, P = 0.005) (Figure 2.4e). In tributaries, CPUE increased from spring to summer (LSD, P = 0.0005), and from spring to fall/winter (LSD, P = 0.018), and decreased from summer to fall/winter (LSD, P = 0.002) (Figure 2.4e.). There were no differences in CPUE in side-channels between spring to summer (LSD, P = 0.644), spring to fall/winter (LSD, P = 0.804), or summer to fall/winter (LSD, P = 0.884). Age 1+ I found differences in mean CPUE of age 1+juvenile chinook among macrohabitats (ANOVA, F = 10.882, df = 2,165, P = 0.0005). In tributaries, CPUE was greater than the 25 mainstem (LSD, P = 0.001) and the side-channels (LSD, P = 0.0005) (Figure 2.4f.). The CPUE did not differ between the mainstem and the side-channels (LSD, P = 0.659). The two-way ANOVA, with season and macrohabitat as factors, revealed differences in CPUE among macrohabitats (F = 9.305, df = 2,159, P = 0.0005) and no differences among seasons (F = 1.027, df=2,159, P = 0.360). There was no interaction between season and macrohabitat (F = 1.462, df=4,159, P = 0.216) (Figure 2.4f.). In spring, CPUE was greater in tributaries compared to side-channels (LSD, P = 0.010), but did not differ from the mainstem (LSD, P = 0.092) (Figure 2.4f). The CPUE did not differ between side-channels and the mainstem in spring (LSD, P = 0.440). In summer, CPUE was greater in tributaries compared to the mainstem (LSD, P = 0.007) and side-channels (LSD, P = 0.007) (Figure 2.4f). The CPUE did not differ between the mainstem and side-channels in summer (LSD, P = 0.948). In fall/winter, CPUE did not differ between tributaries and the mainstem (LSD, P = 0.114) or side-channels (LSD, P = 0.149) (Figure 2.4f). Growth rates Age 0+ I found no differences in mean growth rates of juvenile chinook among side-channels, tributaries or the mainstem (ANOVA, F = 1.572, df = 2,66, P = 0.215). The two-way ANOVA, with season and macrohabitat as factors, revealed differences in growth rates among seasons (F = 4.549, df = 2,60, P = 0.014), but indicated there was no differences among macrohabitats (F = 1.415, df = 2,60, P = 0.251). There was no interaction between season and macrohabitat (F = 0.924, df = 4,60, P = 0.456) (Figure 2.5.). There was a reduction in growth rates during the summer period. In tributaries, growth rates decreased from spring to summer (LSD, P = 0.026) and from spring to fall/winter (LSD, P = 0.023) (Figure 2.5.). In side-channels, growth rates did not differ between spring and summer (LSD, P = 0.205), spring and fall/winter (LSD, P = 0.926), 26 nor summer and fall/winter (LSD, P = 0.0.251). In the mainstem, growth rates differed between spring and fall/winter (LSD, P = 0.024), but did not differ between spring and summer (LSD, P = 0.158), nor summer and fall/winter (LSD, P = 0.179) (Figure 2.5.). Recapture of Marked Fish In 1997, the recapture of marked juvenile chinook from the tributaries increased progressively from 27% in August, 38% in September, 54% in October, 67% in November, and 70% in December. Fish marking was incorporated into the fish sampling program in August of 1997, which explains the absence of re-captures prior to this date. In 1997, recapture of marked juvenile chinook in the mainstem was zero in August and increased to 20% in September and 27% in October, and then dropped off to zero in November and December. In 1997, all re-captured fish were from the same macrohabitat they were originally marked in. No re-captured fish switched macrohabitats. Marked juvenile chinook were never re-captured from side-channels in 1997. In 1998, the recapture of marked juvenile chinook in the tributaries showed a similar pattern to 1997, increasing seasonally from 9% in June, to 7% in July, 16% in August, 40% in September and 39% in October. Macrohabitat switching was documented in June of 1998 because marked fish that had originally been captured and marked in tributaries were re-captured in different tributaries. Approximately 2% of the juvenile chinook captured in tributaries in June had switched habitats from a different tributary. The recapture of marked juveniles in the mainstem increased from 3% in June ofT998 to 6% in July, 20% in August and then, similar to 1997, dropped off to zero as the fall/winter season approached. Macrohabitat switching was documented in July of 1998 in the mainstem when approximately 4% of the juvenile chinook captured had been originally captured and marked in tributaries. Similar to 1997, there were no marked fish re-captured in side-channels in 1998. 27 Biomass and Density Mean biomass in tributaries (3.65 g/m ) (SE = 0.61, n = 226) was 28 times larger than the mainstem (0.13 g/m2) (SE = 0.05, n = 117) (ANOVA, F = 27.090, df = 2,8, P = 0.0005). Similarly, mean density in tributaries (1.13/m2) (SE = 0.25, n = 226) was 14 times greater than the mainstem (0.08/m2) (SE = 0.03, n = 117 ) (ANOVA, F = 13.821, df = 2,8, P = 0.003). DISCUSSION Habitat Value Small non-natal tributaries appeared to provide the highest habitat value to juvenile chinook. Tributaries contained both the greatest density and largest individuals in terms of length and mass for all age classes. Tributaries also provided the longest growing season and juveniles in these habitats displayed very high residency. Side-channels consistently provided the lowest value habitats. The tributaries provided a more stable year round thermal regime than either the mainstem or side-channels, which likely contributed to their high value. This trend has also been observed in the Stuart River, where tributaries have cooler summer temperatures compared to the mainstem which is frequently above optimum temperatures for growth (Lister et al. 1981). Groundwater and hyporheic exchange were likely the causative agents modifying stream temperatures of the headwater tributaries to the Torpy River. Indeed, a nearly uniform year round temperature, being cooler than other streams in the summer and warmer in the winter, is characteristic of groundwater controlled systems (Beschta et al. 1987). Not only did tributaries display less seasonal variation in temperature, but they also exhibited less daily variation in temperature than the mainstem Torpy River (Shrimpton et al. 1999b). The daily and annual mean and maximum temperature fluctuations in the mainstem Torpy River, relative to the thermal stability of the tributaries, were even more pronounced 28 considering that variations in temperatures are generally dampened in large systems due to the relatively large volumes of water contained in these systems (Beschta et al. 1987). The sensitivity of salmonids to differences in temperature is particularly acute, Brown and MacKay (1995) observed significant differences in abundance of over-wintering trout between habitats with temperature differences of only 1°C. I observed active recruitment of juvenile chinook from the Torpy River mainstem to tributaries on several occasions, when mainstem temperatures exceeded 20 °C. Groundwater controlled tributaries in the Torpy watershed likely provide thermal refugia from the mainstem for juvenile chinook, both during periods of high temperatures in the summer and low temperatures in the winter. Mean, maximum and variability in suspended sediment concentrations in the tributaries of the Torpy River were consistently lower than both the mainstem and side-channels. Thus, it appears that tributaries may also provide opportunities for refuge from high suspended sediment concentrations. Juvenile chinook use of non-natal tributaries, as refuge from high sediment concentrations, has been documented in other Upper Fraser River tributaries (Scrivener et al. 1994). Consistent with my findings, in the Grande Ronde River, Oregon, over-wintering juvenile chinook captured in tributaries during the fall/winter period were larger than those that emigrated to the mainstem (Keefe et al. 1995). However, the size advantage incurred by tributary overwintering individuals in the Torpy River appears to be more the exception, rather than the rule, for stream type chinook salmon. Juveniles rearing in tributaries of the Stuart River, British Columbia (Lister et al. 1981) and the Nechako River, British Columbia (Anonymous 1987,1990; Emmet 1989) have been reported to be smaller than their mainstem rearing cohorts due to warmer temperatures in the mainstem. In the Taku River, Alaska, juvenile chinook rearing in nine different macrohabitat types had no differences in mean fork lengths (Murphy et al. 1989). 29 Overwinter survival increases with size (Holtby and Hartman 1982; Cederholm and Scarlett 1981). In fact, Brett (1952) demonstrated that in Pacific salmon of the same age there is a size influence on cold tolerance whereby larger individuals have greater survival. In juvenile chinook salmon, fish that differed in length by 25 mm had mortality rates that differed by a factor of 2 when exposed to temperatures between 0 to 3 °C (Brett 1952). Cold tolerance, and the adaptive significance of large size to overwinter survival, is strikingly evident in northern, interior watersheds subject to long, cold winters such as the Torpy River. Large size has many positive implications to juvenile chinook, including increased tolerance to suspended sediment (Sigler et al. 1984; Servizi and Martens 1991), increased cold tolerance (Brett 1952), and enhanced performance in social interactions (Chapman 1962; Hearn 1987). In fact, larger juvenile chinook have been documented to utilize non-natal rearing habitat in the Fraser River as refugia for shorter durations than their smaller cohorts (Scrivener et al. 1994). Smaller individuals are less able to cope with inhospitable environmental conditions. The linkages between larger size and enhanced survival to adulthood are well established (Bilton et al 1982; Bilton 1978, 1984; Cederholm and Scarlett 1981; Holtby and Hartman 1982; Swales et al. 1988). Therefore, a macrohabitat that consistently produces a greater number of larger juveniles, such as the tributaries in the case of the Torpy River, has clear benefits to the population. In the absence of active sampling, I would be unable to conclude that the tributaries were the highest value habitats based on a greater CPUE, because this could be an artifact of sampling effectiveness of passive techniques in different macrohabitats. The differences in biomass and density between the mainstem and the tributaries from active sampling were large and unequivocal, suggesting that the tributaries do provide the highest habitat value and that passive techniques were effective in representing relative differences in abundance between 30 macrohabitats. Extensive, year round utilisation of non-natal tributaries by juvenile chinook has also been demonstrated in the Yukon River system (Bradford et al. 2001). The greater density and biomass of juvenile chinook rearing in tributaries to the Torpy River relative to the mainstem is consistent with findings in the literature (Keefe et al. 1995; Lister et al. 1981; Scrivener et al. 1994); however, the values for the Torpy tributaries are quite high in comparison (Levings and Lauzier 1991; Hillman et al. 1987; Emmett 1989; Murphy et al. 1989). High densities of juvenile chinook rearing in the tributaries of the Stuart River have been attributed to more favourable temperatures and decreased competitors (Lister et al. 1981). Side-channels contained very low fry densities and did not contribute significantly to the rearing population (Lister et al. 1981), which is consistent with my study. Murray et al. (1981) reported a contrasting finding for juvenile chinook rearing in the Bowron and Willow Rivers; CPUE was lower in tributaries than the mainstem. The decline in CPUE in tributaries of the Bowron and Willow Rivers in September caused these authors to conclude that tributaries were unimportant as winter habitat in these systems (Murray et al. 1981), although sampling did not extend beyond September 30 th so this was not supported empirically. Disproportionate use of tributaries may be a function of a greater amount of edge habitat relative to wider systems, providing enhanced habitat diversity, allochthonous input, and cover. Strong association with margin habitat, gradually decreasing with increasing size, has been well documented for juvenile chinook (Rosberg et al. 1981; Healey 1991). A greater proportion of edge habitat was also suggested to explain the high density of juvenile cutthroat trout rearing in small streams in coastal British Columbia (Rosenfeld et al. 2000). Utilisation of tributaries may also be due to greater shading relative to the wider mainstem, which has been shown experimentally to be positively correlated to juvenile chinook biomass and abundance (Meehan et 31 al. 1987). Unquestionably, stochastic environmental factors, such as extremes in suspended sediment and temperature, render the mainstem a rather inhospitable habitat to occupy. Growth rates did not differ between any of the macrohabitats and the growth rate of juveniles rearing in tributaries appear similar, if not slower, than those reported in other systems (Healey 1991; Murphy et al. 1989; Weatherley and Gill 1995; Rosberg et al. 1981). Therefore, I would argue that food is not the primary driver attracting juvenile chinook to tributaries in the Torpy River. It is more likely that tributaries are high value habitats based on their role in providing refugia from extreme temperatures, suspended sediment levels, predators or a combination thereof, and that a longer growing season is a byproduct. The fact that growth rates did not differ between the mainstem and tributaries in the summer, even though temperatures were much higher in the mainstem, suggests that other factors may be limiting growth in the mainstem such as food availability, predation, suspended sediment, or extreme temperatures. The decline in growth rate of tributary and mainstem rearing chinook from spring to summer is somewhat surprising since summer is the season when growth rates would be expected to be highest due to increased light, temperature, primary production and food availability. Growth has been reported to be highest during the summer for juvenile chinook rearing in the Nechako River (Anonymous 1998; Russell et al. 1983). Perhaps elevated temperatures and suspended sediment loads increased metabolic demand, thus less energy could be allocated to somatic growth. However, this explanation seems insufficient because a decline in water quality in the tributaries in summer did not parallel that of the mainstem, yet growth rates similarly declined. The CPUE was highest in both of these macrohabitats during the summer, suggesting that density dependent factors (Chapman 1962, 1966; Mason 1976; Mason and Chapman 1965) are playing a role in influencing growth rates. The density of juveniles rearing in the tributaries far exceeded that in 32 the mainstem, which may explain why the decline in growth rates from spring to summer was more pronounced in comparison to the mainstem. The fact that condition factor was reasonably constant in all macrohabitats, over all the seasons, suggests that significant feeding must occur throughout the fall/winter season, otherwise a decline in condition factor would be apparent. Even though growth rates in tributaries declined from spring to fall/winter, they were still over 2.5 times higher than mainstem rates in the fall/winter season. Further, the decline in growth rates from summer to fall/winter was 73% in the mainstem, and only 14% in tributaries. In fact, other researchers have noted the difficulty in ageing juvenile chinook from the Upper Fraser River tributaries because a freshwater annulus is not readily apparent due to sustained growth throughout the winter (Rosberg et al. 1981; Tutty and Yole 1978). Further, juvenile chinook not displaying an annulus are larger than those that do (Rosberg et al. 1981). Rosberg et al. (1981) confirmed that juvenile chinook accrue length from November until the following spring in four Upper Fraser River tributaries (Slim, Torpy, Morkill, Holmes). If winter growth is in fact sustained, it is contradictory to the majority of life history studies on chinook salmon that report a cessation of growth (Bradford et al. 2001; Everest and Chapman 1972) and portray overwintering juveniles as huddling in substrate interstices in a near state of torpor (Swales et al. 1986; Levings and Lauzier 1991). The growth rates reported for the side-channel fish should be interpreted with caution. The recapture rate of marked individuals was zero for this macrohabitat and the CPUE was very low from side-channels suggesting that the residency rate was similarly low. The individuals captured in the side-channels cannot be considered a discrete population. Rather, it is more plausible that juvenile chinook utilise side-channels transiently, moving frequently between the mainstem and side-channels and for this reason no emphasis was placed on the growth rates reported for this macrohabitat. 33 The importance of growth to anadromous salmonids is particularly acute since it is correlated to the annual decision to smolt or to remain in freshwater (Thorpe 1987). Growth can be considered a controlling factor to large size, therefore, due to the established links between size and survival (Bilton et al. 1982; Bilton 1978, 1984; Beckman et al. 1998; Cederholm and Scarlett 1981; Holtby and Hartman 1982; Swales et al. 1988), any behaviour (tributary use) that sustains growth is likely to be favoured. I did not find any differences in the attributes of post-emergent juvenile chinook or from those captured during spring in the different macrohabitats, suggesting that distribution of post-emergent fry was a function of random events, rather than size selective displacement or active selection. Therefore, the indices of habitat value were representative of environmental conditions for the macrohabitat in question, since a possible confound would be if the individuals in each macrohabitat were not equivalent in all manners at the beginning of their life phase. The larger size of juvenile chinook in tributaries must be due to either higher growth rates and/or a longer growing season and not an artifact of selection by unique individuals. An assumption, fundamental to using abundance, size, and growth to infer habitat value, was that juvenile chinook were relatively static in terms of movement between macrohabitats, rather than transiently switching between macrohabitats. The degree of residency indicated by the recapture data suggests that the indices of habitat value from the mainstem and tributaries were a result of individuals rearing in those macrohabitats, and not confounded by fish moving between macrohabitats. Since marked fish only represented a fraction of the population, but progressively more of that population was marked each week of the sampling program, it follows that if the recapture rate did not continue to increase over time as the data showed, it would indicate immigration of un-marked individuals into macrohabitats. Marked and un-marked fish would be expected to emigrate equally, therefore emigration should not affect the recapture rate. The fact 34 that recapture of marked fish from other macrohabitats was rare, adds further support that the populations rearing in each macrohabitat were relatively constant over time. Therefore the assumption, that the indicators of habitat value were a function of conditions in a particular macrohabitat, is a reasonable one. Bradford et al. (2001) similarly found that over 80% of juvenile chinook marked in August were re-captured in the same non-natal tributary to the Yukon River in October suggesting that movements were relatively restricted. Stream-type juvenile chinook have been reported to display considerable permanence of station (Edmundson et al. 1968; Rosberg et al. 1981; Bustard 1983) which has been suggested to confer many benefits including familiarity with cover and food sources (Chapman and Bjornn 1969) providing obvious adaptive value considering the extended freshwater rearing phase of the stream type race. In both years, the mainstem recapture rate declined to zero as the fall/winter season approached, suggesting either emigration to the Fraser River, significant mortality or tributary recruitment. The recapture data supports the former two hypotheses based on the absence of any mainstem marked fish captured in the tributaries, coupled with the fact that the tributary recapture rate increased steadily throughout the fall/winter season, rather than declined which would be expected if substantial immigration of unmarked mainstem fish occurred. Further, during summer low flows (Aug - Sept) in the Torpy River the confluence of many tributaries with the mainstem was sufficiently shallow (<1 cm) that tributary recruitment at this time would be certainly restricted, if not impossible in many cases. Seasonal re-distributions of juvenile salmonids are a common avoidance response to unfavourable winter conditions in summer rearing areas (Swales et al 1988; Tschaplinski and Hartman 1983; Bustard andNarver 1975; Chapman and Bjornn 1969; Bjornn 1971; Cederholm and Scarlett 1981). A decline in abundance of mainstem rearing juvenile chinook as the fall/winter season approaches has been described for many systems (Keefe et al. 1995; Swales et 35 al. 1986; Chapman and Bjornn 1969; Hillman et al. 1987; Everest and Chapman 1972), including four Upper Fraser tributaries (Slim, Torpy, Morkill, Holmes) (Rosberg et al. 1981), the Bowron and Willow Rivers (Murray et al. 1981), the Stuart River (Lister et al. 1981) and the Nechako River (Russell et al. 1983). The triggers for this exodus have been speculated to include declining temperature (Swales et al. 1986; Chapman and Bjornn 1969), reduction in flows and corresponding rearing space resulting in intensified social interactions (Rosberg et al. 1981; Edmundson et al. 1968), and lack of suitable over-wintering substrate (Bjornn 1971). It is doubtful that the density dependent theory would explain the emigration from the mainstem Torpy River since the densities were 14 times greater in the tributaries, and an exodus did not occur even though increased aggression and social interactions were observed. Further, Hillman et al. (1987) demonstrated that the fall exodus of juvenile chinook from the Red River, Idaho was not size related as would be expected if governed by intensified social interactions. It is also not possible to discount mortality as a potential explanation for the decline in abundance in the mainstem. Over-winter mortality of juvenile chinook in the mainstem Grande Ronde River, Oregon has been reported to be as high as 75%, compared to 47% over-winter survival in the tributaries of that system (Keefe et al. 1995). However, the exodus of juvenile chinook from the mainstem Torpy River occurred in late August when mortality would presumably be low. The absence of suitable large cobble substrate for overwintering is correlated to downstream migration of juvenile salmonids in Idaho streams, with emigration coincident with declining stream temperatures (Bjornn 1971). Bjornn (1971) tested other variables including decreased food supply and temperature change and was unable to establish either as the causative agent for emigration. Declining stream temperatures trigger juvenile chinook to seek out substrate crevices as overwintering habitat (Hartman 1965). Therefore, in the absence of suitable 36 substrate habitat, temperature may indirectly elicit downstream migration (Bjornn 1971). Hillman et al. (1987) reported juvenile chinook densities to decline 73 and 84% from summer levels during two fall/winter seasons, in an Idaho stream heavily embedded with sediment. Juvenile chinook salmon were first observed emigrating in late August when temperatures dropped below 8 °C and peaked in October when stream temperatures were 4 °C (Hillman et al. 1987). This large population emigration was shown experimentally to be due to a lack of suitable substrate interstices for over-wintering due to extensive deposits of sediment, triggered by the onset of colder temperatures (Hillman et al. 1987). In the Torpy River, the most complete year of fish marking (1998) coupled with observations from snorkel surveys suggest that the peak emigration of mainstem individuals occurred prior to September, when the mean mainstem temperature, although decreasing towards the end of August, was well above 15 °C. This is consistent with findings from the Nechako and Stuart Rivers whereby the decline in abundance of juvenile chinook in the mainstem occurred from June to August (Lister et al. 1981; Russell et al. 1983). Water temperatures in the tributaries to the Torpy River were much colder than the mainstem, and were similarly declining in temperature, yet an exodus of juvenile chinook was not observed. Cold temperatures induce lethargy in fish, which can function as an impediment for fish escaping declining temperatures (Beitinger et al. 2000). The causal mechanism, triggering juvenile chinook to emigrate from mainstem rivers before the onset of winter conditions, likely occurs prior to declines in temperature, for obvious adaptive reasons, for this behaviour to have evolved. Unlike the findings of other researchers (Hillman et al. 1987; Chapman and Bjornn 1969), the exodus of juvenile chinook in the Torpy River was not coincident with temperatures less than 8 0 C. The primary trigger causing juvenile chinook to emigrate from the Torpy River mainstem is likely sedimentation of the gravel/cobble substrate at summer low flows, which is supported 37 experimentally (Hillman et al. 1987), and was the biophysical difference between tributaries and the mainstem in August that accounts for juvenile chinook residency in the tributaries. Such proactive emigration, due to the degradation of over-wintering habitat in the mainstem, in anticipation of extreme cold temperatures yet months prior to the onset of those temperatures, is a remarkable legacy of evolution. Proactive emigration confers additive adaptive value, considering from a competitive perspective the earlier the emigration, the greater the probability of securing available over-wintering habitat elsewhere. Also, water temperatures (15 0 C) during this early emigration are optimal for juvenile chinook swimming ability, of benefit for lengthy downstream migrations to suitable over-wintering habitat. The fall/winter exodus of stream type juvenile chinook from mainstem rivers has not been reported in coastal systems (Chapman and Bjornn 1969), which may be due to differences in hydrology, geology and climate, that collectively eliminate the trigger (sedimentation of substrate crevices) to downstream emigration. In general, coastal systems are higher energy and as a result of the higher gradient, larger substrate size, and frequent freshets, would not be subject to as extensive sedimentation. Further, coastal systems generally have a milder winter than interior systems, in terms of air temperature, which would cause juvenile chinook to have less dependence upon substrate crevices as refuge from anchor and frazil ice formation. The over-wintering period has been recognized as a limiting factor for salmonid production in many streams (Heifetz et al. 1986; Solazzi et al. 2000; Bustard and Narver 1975; Tschaplinski and Hartman 1983). In northern interior watersheds the severity of winters, in conjunction with the extended freshwater rearing phase would be expected to elicit greater evolutionary pressure for alternative life history strategies, such as emigration from mainstem rivers or selection of groundwater controlled non-natal tributaries. 38 Alternate life history strategies The approach of the fall/winter season highlights three distinct life history strategies for juvenile chinook in the Torpy River watershed; over-winter in groundwater controlled tributaries, emigrate from the mainstem, or over-winter in the mainstem. The latter strategy was so unpopular that I will focus on the first two, which I will term "non-natal" strategy and "emigrate" strategy respectively. The survival benefits for juvenile salmonids over-wintering in non-natal tributaries has been well documented in other systems (Keefe et al. 1995; Bustard and Narver 1975; Hartman and Holtby 1982). Non-natal tributaries appeared to provide the highest habitat value for juvenile chinook in the Torpy River. Furthermore, individuals exhibiting the "non-natal" strategy may have a survival advantage, since behavioural disposition for selection of these habitats has adaptive value during downstream migrations, demonstrable by the fact that non-natal tributaries to the Fraser River provide critical refuge habitat to juvenile chinook from high suspended sediment levels (Scrivener et al. 1994). However, the fitness implications for juvenile chinook that emigrate downstream from their natal mainstem is largely unknown. Data from the Stuart River indicate that the majority (87%) of the juvenile chinook in the mainstem of this system emigrate with the onset of fall (Lister et al. 1981). It is reasonable to assume that a large proportion of the juvenile chinook in the Torpy River system similarly emigrate. Individuals that emigrate to the Fraser River, over-winter along the channel margins in substrate interstices, and do not exhibit growth if rearing above Hope (Levings and Lauzier 1991). Yet this scenario proposes an apparent paradox: If a significant proportion of juvenile chinook emigrate from their natal Upper Fraser tributaries, how do they manage to incur sustained growth throughout the winter such that a freshwater annulus in 39 returning adults is barely discernible (Tutty and Yole 1978; Rosberg et al. 1981)? I propose three possible explanations. Firstly, it is possible juveniles in the "emigrate" life history strategy sustain very high mortality, such that even though fewer select the "non-natal" life history strategy, these individuals disproportionately comprise the majority of returning adults that, upon scale analysis, indicate sustained growth occurred throughout the winter (Tutty and Yole 1978; Rosberg et al. 1981). Alternatively, perhaps juveniles that emigrate incur sustained growth throughout the winter through selection of ground water controlled, non-natal tributaries of the Fraser River that confer the same benefits as their non-natal rearing cohorts enjoy in the Torpy River tributaries. Thus, in the Upper Fraser River, juveniles in the "emigrate" life history category may simply represent those individuals unable to secure stable groundwater controlled overwintering habitat in their natal stream and consequently emigrate downstream in search of groundwater controlled non-natal over-wintering opportunities. Considering that of the Pacific salmon, juvenile chinook are the least tolerant of cold temperatures (Brett 1952), and the positive correlation between greater size and cold tolerance (Brett 1952), sustained winter growth via selection of groundwater controlled tributaries has obvious adaptive value to the stream type race in northern interior systems subjected to long cold winters. Indeed, utilisation of non-natal tributaries of the Fraser River by juvenile chinook for refuge habitat has been documented (Scrivener et al. 1994; Murray and Rosenau 1989) and this strategy may be significantly more prevalent and critical to survival of the stream type race than previously thought. A third possibility is that juvenile chinook emigrating from Upper Fraser River tributaries incur sustained over-winter growth by extensive fall/winter migrations to the lower Fraser River, below Hope, where over-winter growth has been documented (Levings and Lauzier 1991). 40 However, this theory seems insufficient to account for the substantial numbers emigrating from the Upper Fraser River, considering that during winter, juvenile chinook density in the Fraser River increases progressively with distance upstream (Levings and Lauzier 1991). Levings and Lauzier (1991) theorized that the reason juvenile chinook density increases with distance upstream in the Fraser River during winter, was due to either improved rearing conditions or closer proximity to the spawning grounds. I suggest these authors latter theory is more applicable, yet for reasons contradictory to the former. As distance along the Fraser River increases, primarily in the northward direction, the moderating oceanic influence on climate decreases (Swales et al, 1986), elevation increases, and winter temperatures become colder, with greater frazil and anchor ice formation; conditions which appear worse for over-winter survival rather than improved. As distance increases along the Fraser River, natal streams are exposed to increasingly harsh winters, and become less hospitable to over-wintering juvenile chinook. Emigration from natal systems likely evolved in direct response to severity of winter conditions, which results in greater densities of juvenile chinook in the Fraser mainstem as distance increases upstream. Corroborating evidence is provided in the fact that the "emigration" over-wintering strategy is only described in northern, interior systems subject to long, cold winters and not coastal watersheds (Bjornn 1971; Everest and Chapman 1972; Chapman and Bjornn 1969; Hillman et al. 1987; Rosberg etal. 1981; Swales et al. 1986; Murray etal. 1981; Lister et al. 1981; Russell et al. 1983). Chinook display one of the most diverse life history patterns of all the Pacific salmon species (Miller and Brannon 1981). The "non-natal" over-wintering strategy may represent yet another layer of variation exhibited by the stream type race. Forest Harvesting Implications Non-natal tributaries to the Torpy River are among the smallest of streams that have been assessed for juvenile chinook habitat utilisation (Bradford et al. 2001). Small streams have the 41 least resiliency to human impacts and are also afforded the least protection from forestry activities (FPC 1995). The fact that these same small streams provide the highest habitat value to juvenile chinook underscores the importance for natural resource managers to significantly elevate habitat protection of these systems. Other researchers have documented that long term physiological indices of growth were highest in juvenile chinook rearing in unlogged tributaries than those captured from the mainstem or from logged tributaries in the Torpy River, even though logged tributaries were warmer (Shrimpton et al. 1999b). This provides additional evidence that un-logged tributaries are the highest value habitats. In contrast, rainbow trout in logged tributaries in northern, interior British Columbia displayed greater growth than those in un-logged tributaries (Mellina 2002). Rainbow trout have a wider thermal tolerance than juvenile chinook (Reiser and Bjornn 1979; Lee and Rinne 1980), which may explain this difference. Removal of riparian vegetation on streams in areas with average winter air temperatures < 0° C can result in a depression of winter stream temperatures (Mitchell 1999; Brownlee et al. 1988; Beschta et al. 1987; Anderson et al. 1997; Rishel et al. 1982), although exceptions have been documented (Macdonald et al. 2003). Post-forest harvesting stream temperatures declined up to 2° C in Slim Creek, northern interior British Columbia (Brownlee et al. 1988). In fact, experimental manipulations to riparian zones in the Torpy River watershed demonstrated that forest harvesting resulted in decreased mean stream temperatures relative to un-logged streams (Shrimpton et al. 1999a&b). This is of particular concern in the Torpy River, since one of the reasons tributaries offer the highest habitat value is because they are cooler in the summer and warmer in the winter than the mainstem. Alterations to this thermal regime by even a couple of degrees could eliminate one of the advantages of rearing in non-natal tributaries, compromise the 42 survival of populations selecting this "non-natal" over-wintering strategy, and perhaps jeopardize the persistence of this life history trait. 43 Study Sites Torpy River * tributary sidechannel Figure 2.0. Diagram representing the three macrohabitats (mainstem, tributaries, side-channels) in the fish sampling program. 44 Figure 2 . 1 . Mean density of predators and competitors of juvenile chinook in each macrohabitat (A) in 1998 and in the mainstem (B) from June 12 to October 23, 1998 (n =4). Error bars represent ± 1 SE. The SE for the density of chinook in tributaries that extends off the chart is ±0.19. 45 25 Mainstem mean Tributary mean Mainstem max Figure 2.2. Mean (daily) and maximum water temperatures in the mainstem and mean temperature of tributaries of the Torpy River, July 1996 - October 1998. 46 June July August September October November Figure 2.3. Mean monthly suspended sediment concentration in each macrohabitat pooled for 1997 and 1998. Error bars represent ± 1SE. The SE for side-channels in September that extends off the graph is ± 153.3. Means are based on daily means in the mainstem and weekly means in tributaries and side-channels. 4 7 Figure 2.4. Mean fork length of age 0+ (A) and 1+ (B), mean mass of age 0+ (C) and 1+ (D), and mean CPUE of age 0+ (E) and 1+ (F) juvenile chinook captured in 1997 and 1998, pooled seasonally and annually. Error bars represent ± 1 SE. Bars with similar letters did not differ. Note that "all dates" and "seasons" were analysed seperately and thus cannot be contrasted. Missing side-channel bars in the 1+ graphs are because this age class was never captured. Al l other missing bars indicate a zero value. Means are based on weekly values. 48 Summer Fall/Winter • Mainstem • Tributary • Side-channel Figure 2.5. Mean growth rates of age 0+juvenile chinook for 1997 and 1998, pooled annually and seasonally for different macrohabitats. Error bars represent ± 1 SE. The SE for the side-channel in fall/winter that extends off the graph is ± 2.2. Bars with similar letters did not differ (see figure 2.4. for details). 4 9 CHAPTER 3: EXPERIMENTAL FIELD MANIPULATIONS OF SUSPENDED SEDIMENT CONCENTRATIONS IN TRIBUTARIES AND SIDE-CHANNELS INTRODUCTION Suspended sediment in streams has been identified as the "greatest single water pollutant" by the U.S. Environmental Protection Agency (Downing 1980). Elevated concentrations can cause serious impacts to all trophic levels in aquatic ecosystems. The United Nation's European Inland Fisheries Advisory Commission (EIFAC 1964) concluded fish production would be diminished in waters containing 25 - 80 mg/L of suspended sediment and waters in excess of 80 mg/L were unlikely to support good fisheries. Suspended sediment is a broad spectrum pollutant (Langer 1980) that can directly affect juvenile salmonids, either behaviourally or physiologically, or indirectly through alterations of habitat. Suspended sediment can affect fish habitat through decreased light attenuation (Lloyd et al. 1987), and smothering or scouring of periphyton (Langer 1980), both of which result in reduced photosynthesis causing a decrease in primary production (Anderson et al. 1996). As a result, less food may be available for primary consumers such as macro-invertebrates. Macro-invertebrate abundance can also be reduced as a direct result of suspended sediment, through scouring, or smothering food availability for salmonids (Cordone and Kelly 1961; Langer 1980; Lloyd 1987; Lloyd et al. 1987). Invertebrate drift patterns may also be altered which can affect their survival as well (Cordone and Kelly 1961; Langer 1980). Fish can be directly affected by suspended sediment as substrate interstices get filled eliminating substrate cover habitat (Langer 1980). Juvenile chinook survive harsh winter conditions by entering substrate crevices (Bjornn 1971; Edmundson et al. 1968). Bjornn (1971) reported that the absence of cobble substrate resulted in emigration of juvenile salmonids whereas they over-wintered when cobble substrate was available. The effect of substrate embeddedness on densities of juvenile chinook has been 50 examined experimentally during winter (Hillman et al. 1987). Additions of clean substrate to an Idaho stream heavily embedded with sediment, resulted in a nine-fold increase in densities of juvenile chinook. Hillman et al. (1987) concluded that juvenile chinook were forced to either emigrate or utilize over-wintering habitat that was prone to ice-scouring, as fine sediments had filled the interstitial spaces of the substrate. In a more general sense, suspended sediment can cause a reduction in rearing habitat availability by gradually filling in pools and reducing the area for feeding and refuging (Langer 1980). Lastly, substrate embeddedment can also negatively impact eggs and alevins by trapping emerging fry and/or impairing gas exchange (Langer 1980). Avoidance is one of the first behavioural responses exhibited by salmonids to suspended sediment (Servizi and Martens 1992; Bisson and Bilby 1982; McLeay et al. 1987; Birtwell 1999). Avoidance of unfavourable conditions has obvious adaptive value; however, if it leads to displacement to sub-optimal habitats it can result in fitness consequences for juvenile salmonids such as decreased growth or increased predation. Suspended sediment can be a major contributor to turbidity resulting in a loss of light transmission due to absorption or scattering (Waters 1995). Salmonids are visual predators, thus changes in turbidity can affect salmonid feeding ability (McLeay et al. 1987; Noggle 1978; Berg and Northcote 1985). Feeding is a fundamental requirement for energy acquisition and growth for fish, therefore impaired feeding has direct fitness implications. Pulses of suspended sediment have been documented to disrupt dominance hierarchies in juvenile coho salmon (Berg and Northcote 1985). Linkages between territoriality, densities and size of stream dwelling salmonids are well established (Chapman and Bjorrn 1969; Kalleberg 1958; Allen 1969; Chapman 1962; Hartman 1965; Dill 1978; Puckett and Dill 1985) therefore chronic disruptions to social hierarchies may have consequences to fish abundance, distributions and fitness (Berg and Northcote 1985). 51 Exposure to suspended sediment can elicit sub-lethal stress responses in fish, manifested in multiple physiological effects including impaired growth, histological changes to gill tissue, alterations in blood chemistry and an overall decrease in resistance to disease and additional stressors (Anderson et al. 1996; Redding et al. 1987; Servizi and Martens 1987). Suspended sediment can cause abrasions to sensitive gill tissue triggering fish to respond with increased mucus secretion (Langer 1980). At high suspended sediment concentrations, particles can stick to the mucus in sufficient quantity causing respiratory distress and even suffocation (Langer 1980), and this gill irritation induces a coughing response (Noggle 1978). Sub-lethal effects of suspended sediment on juvenile salmonids have never been examined in field experiments. Extensive field research has been conducted investigating the impacts of sediment on egg and fry survivorship (Hartman and Scrivener 1990; Cederholm et al. 1981; Meehan 1991; Everest et al. 1987; Chapman 1988). Research focusing on the sublethal effects of suspended sediment and turbidity on fish behaviour and physiology has primarily concentrated on laboratory experiments and artificial streams (Bisson and Bilby 1982; Gregory 1993,1994; Waters 1995; Redding et al. 1987; Berg and Northcote 1985; Ginetz and Larkin 1976; Sigler et al. 1984; Noggle 1978; Servizi and Martens 1992,1991; Servizi and Gordon 1990; McLeay etal. 1987). Although valuable for providing insights to causal relationships between characteristics of fish habitat and effects on fish, they generally lack ecological realism (Adams 1990). In this chapter, I conducted experimental field manipulations to suspended sediment concentrations on small tributary streams and side-channels of the Torpy River, a snowpack dominated interior system in northern British Columbia (see Chapter 1 for study site description). Forest harvesting activities are prevalent in the Torpy River. Forest harvesting can cause changes to suspended sediment inputs (Meehan 1991). I simulated increases in suspended sediment concentrations, to mimic those induced by forest harvesting activities and quantified their 52 biological effects on juvenile chinook salmon. These experiments enabled the isolation of other variables associated with forest harvesting which often generate confounding effects (i.e. increased light, altered hydrology, altered habitat diversity, etc.) on fish populations. I predict that sediment additions will cause increases in stress levels in juvenile chinook as assessed by whole body tissue Cortisol concentrations, and that these responses will be most pronounced in tributary rearing individuals, since acclimation to elevated suspended sediment levels is likely greater in side-channel fish because of naturally higher levels of suspended sediments. I further expect that sediment additions will result in increased movement indicating an avoidance response by juvenile chinook. Because side-channels are the lowest value habitats (see Chapter 2), avoidance may be most pronounced in side-channels. I also predict that sediment additions will result in decreased foraging in tributaries but increased foraging in side-channels. Predators are considerably more prevalent in side-channels compared to tributaries (see Chapter 2). Therefore, I postulate that under the cover of turbidity, anti-predator behaviours would subside, and side-channel fish would forage more (Gregory and Northcote 1993; Gregory and Levings 1998; Gregory 1993,1994). Sediment additions may result in disruptions to social systems (assessed by levels of aggression, fleeing and yawning) and I expect the greatest effects to be seen in tributaries due to the higher density of juvenile chinook rearing there (Chapter 2). 53 METHODS I used three tributaries and one side-channel as sites for the experiments. The tributary experiment consisted of four trials which were conducted in 1997 and 1998; the side-channel experiment consisted of two trials which were conducted in 1998 (see Figure 3.0a for an example of a side-channel and tributary). Sediment additions were pulsed to simulate the type of exposure that fish in watersheds subjected to forest harvesting are most likely to encounter (Berg and Northcote 1985; Scrivener 1988; Scrivener and Brownlee 1989; Waldichuk 1993). The duration of each experiment was approximately three hours and 30 minutes, divided into a pre-sediment phase (1 hr.), a sediment phase (1.5 hrs.) and a post-sediment phase (1 hr.). Discharge was measured by flow meter and measuring tape according to standard methodology (RIC 1997; Schuett-Hames et al. 1994) prior to each trial. Channel widths of the tributaries ranged from 1 to 2.5m and share similar habitat characteristics. Tributaries flowed in a southerly direction, had a low gradient (1-2%), had a substrate composition of 70% fines and 30% gravels, and abundant cover in the form of undercut banks, over-hanging vegetation, deep pools and woody debris. Channel width of the side-channel was approximately 3.5 m. It connected to the mainstem and flowed freely at the time of both experiments. The side-channel was characterized by low velocity glides and deep pools with a substrate composition of 100% fines. Cover was limited to overhanging vegetation, deep pools and the occasional piece of woody debris. The tributary and side-channel riparian had never been logged. Homogeneous sections of tributaries and side-channels, approximately 100 metres in length, were selected each containing a minimum of two sets of pools, riffles, and glides. Enclosure nets (1 mm mesh) were installed at the downstream and upstream end of each section as well as one in the centre, resulting in two, 50 m, isolated sections. The upstream section was 54 identified as the control and the downstream section was the treatment (Figure 3.0b). The fact that control sections were located immediately upstream from experimental sections ensured similar physical stream reach characteristics and fish population characteristics (species abundance, distribution, etc.). A series of three weirs that extended approximately 90% across the wetted channel on a downstream angle of 30 degrees were installed at each end of the control and treatment sections (Figure 3.0b). Two-way weirs trapped fish that moved into these areas enabling the assessment of upstream and downstream movement direction. Twelve temperature data loggers (Onset instruments) were deployed at 10 m intervals in the side-channel and tributary sites. Loggers recorded temperature every five minutes. Four video cameras (Citizen model JSS 1012C) housed in dark green, waterproof, aluminum housings (approx. dimensions: 50mm x 105mm) with Lexan lids for optical clarity were used to observe fish behaviour. In both the treatment and control sections one camera was positioned at mid-channel and aimed downstream and another was positioned at the stream margin, aimed towards mid-channel, perpendicular to flow (Figure 3.0b). Sections of 1 cm diametre PVC pipe, marked at 5 cm intervals, were installed upright in the substrate in the field of view of the underwater cameras. The pipes functioned as reference points to ensure that the field of view was the same area (generally lm x lm), in both side-view and frontal view cameras, as well as between treatment and control sections. Two, 25.4 cm, black and white monitors (Citizen model JSS 1012M) contained in rainproof Lexan housings, tinted to eliminate sun glare, were used to display images which were recorded by two VHS video recorders (Sanyo model VHR-H607), also in Lexan housings. The underwater video cameras and recording system were powered by a Honda variable speed generator. 55 Experimental Procedures One litre water samples were taken every 15 minutes in both the control and treatment sections during each phase of the experiment and analysed in the lab for suspended sediment concentration by filtration (mg/L). Turbidity was measured in the treatment section during all phases of the experiments by a nephelometric turbidity probe (Analite NEP 190) set to take readings at five second intervals and averages every five minutes. Turbidity data were recorded in units of NTU by a Starlogger data logger (Model 3.08). The relationship in natural streams between the concentration of suspended sediment (mg/L) to turbidity (NTU) varies depending on the nature of the material in suspension, but is generally considered to be approximately 3:1 (Birtwell 1999). However, caution must be exercised in converting NTU to mg/L, since this ratio has been shown to vary as widely as 1:1 to 5:1 (Lloyd 1987; Lloyd et al. 1987), I present my results in mg/L; however, the conversion ratio serves comparative value when discussing my results with those in the literature that have used NTU. Fish numbers (as an indicator of movement) were recorded at each of the four weirs (upstream and downstream treatment, upstream and downstream control) every 15 minutes during each of the pre-sediment, sediment, and post-sediment phases. The number of fish moving both upstream and downstream in each section (treatment and control) were pooled within each experimental phase and termed "total number moving". Simultaneous underwater video recordings of fish behaviour in the treatment and control sections were taken continuously during all phases of the experiments. Samples of juvenile chinook (n=6-10 from each site) were taken by electroshocking, during the pre-sediment phase in the control section (netted), and upstream of the control section (un-netted) to determine the influence of the net on acute stress. Fish were snap frozen on dry ice (< 60 sec.) for laboratory whole body Cortisol analyses. 56 The pre-sediment phase lasted approximately one hour. Sediment additions were made in a pulsed fashion (5 min on, 5 min off) by disturbing the substrate with a length of rebar at a constant rate (see Figure 3.0b for the location of sediment additions). By pulsing sediment, I was able to observe behaviour during the clear water phase between the pulses. Additions of sediment were made in the middle of the treatment section so that upstream and downstream movements would be perceptible. The sediment phase lasted approximately 1.5 hours, at which time lethal sampling of juvenile chinook (n=6-10/site) occurred in the control and treatment. During the one hour post-sediment phase, suspended sediment, turbidity, fish movement and behaviour data were collected. In the 98/08/02 side-channel trial, blood plasma (ng/ml) was extracted from caudal arteries in microhematocrit tubes and centrifuged (5 min) in the field as an alternative means of assessing Cortisol levels due to a shortage of dry ice. Because of the different techniques and experimental units, acute levels of juvenile chinook Cortisol concentrations could not be statistically compared between side-channel experiments. Data Extraction To save time during field experiments, I froze fish whole rather than extract blood plasma for analyses of Cortisol levels. Whole body Cortisol levels from lethally sampled juvenile chinook were assessed in the laboratory. After capture, fish (ranged in mass from 0.41 to 6.36 g) were transferred from dry ice to a -80 °C freezer. Frozen fish were ground into powder using a mortar and pestle in a slurry with liquid nitrogen which was then evaporated off. Distilled water, 300 ul, was added to each sample which was then disrupted by ultrasonication for one minute (B. Braun Biotech, Braun-Sonic L, Allentown, PA, 18103). Five millilitres of diethyl ether was added to samples which were mixed then centrifuged at 1000 g for 5 min. The water component was 57 frozen in liquid nitrogen and the ether decanted. Samples were then re-extracted and the ether fractions combined. Ether was evaporated under a stream of nitrogen gas. Following complete evaporation of ether, 2 ml of phosphate buffered saline (1.5 mM KH2PO4, 2.7 mM Na2HPO"4, 150 mM NaCl, pH 7.4) containing 2% bSA (bovine serum albumin) was added to samples which were kept at 4 °C and vortexed frequently for 2 h. Cortisol concentrations were determined by radioimmunoassay using 1 2 5I (Diasorin, Stillwater Minnesota, 55082-0285). Efficiency of extraction was determined from spiking paired samples with Cortisol and was found to be 92 to 95%. Whole body Cortisol is expressed as ng/g tissue. Information on the following behaviours were transcribed from the video tapes: feeding attempts, coughing, aggression, yawning, fleeing, fish numbers, and fish minutes. These behaviours, which are described in detail below, were selected because they were quantifiable, repeated across treatments, were readily identifiable, and were biologically relevant for sediment exposure studies. In all cases, frequencies of behavioural activities were recorded. Time constraints limited the ability to extract duration, intensity or pattern of behavioural activities. Feeding attempts were defined as the number of foraging attempts per minute. The high resolution of the cameras enabled the observation of individual prey items. Feeding attempts consisted of a rapid opening and closing of the mouth for a potential prey item and often coincided with a change in orientation in the water column, or small movement to capture a prey item. Foraging success was not quantified. Coughing was defined as the backward propulsion of water across the gills to displace obstructions (e.g. sediment), and was readily identified by excessive flaring of the operculum, gaping mouth and rapid spasm of the buccal cavity. Displaced items were often observed during this activity, as was a small forward movement of the fish (Berg and Northcote 1985). 58 Aggression was defined as any agonistic behaviour directed toward another fish, which included chasing, nipping, and charging (Hartman 1965; Chapman 1962). Yawning was defined as the protracted and excessive gaping of the mouth, and was easily distinguishable from feeding due to an absence of a potential prey item, a lack of any change in position or orientation in the water column, and a very protracted and excessive gaping of the mouth. Fleeing was defined as evasive behaviour, most often displayed in response to another aggressive individual. Fleeing fish were displaced from their territory or previously occupied position either temporarily or permanently. The number of fish observed per sampling period was recorded. Fish moving in and out of the field of view were not considered to be different fish; however, this variable provides an indication of activity level. Fish were not marked due to their small size for many of the experiments (<5.0 cm), and to prevent artificially altering behaviour as a result of marking. The frequencies of extracted behaviours were divided by the time observed to provide a frequency of behaviour per minute that could be compared between different sections and phases of experiments. The total time that each fish was in the field of view in each section (treatment, control) was summed for each phase of the experiment (pre-sediment, sediment, post-sediment) to provide an accumulated time. This was divided by the sample period time to provide an estimate of equivalent number of fish viewed per minute, termed "fishmin". A systematic random approach was used to extract behavioural data from the underwater video recordings. Each phase of every experiment was divided into ten equal time intervals for a total of 30 intervals in the control recording(s) per experiment and 30 intervals in the treatment recording(s) per experiment (Figure 3.1.). Considerable deliberation and effort was taken to determine the appropriate sampling approach to extract data from each interval. A test was completed whereby, behavioural data were extracted continuously from a 16.3 minute segment of 59 video footage, and the frequencies of behaviour were calculated and treated as true frequencies. I extracted replicate samples from this segment of different fixed times (1 min 38 seconds, 2 min 30 seconds, and 4 min 8 seconds) starting from randomly generated times within the segment. The percentage deviation of the behavioural means extracted from each different fixed sampling time was compared to the true values. All of the means were well within 1 SD of the true values, and since the 1 min 38 second fixed sample time was >97% of the true values it was decided to proceed with that sample time which would allow more replicates. Due to the time consuming nature of data extraction it was estimated that the number of intervals would have to be reduced from 10 per phase, to 5 per phase, if a longer sample time was chosen. Indeed, Zinner et al. (1997) similarly documented that many short sampling periods are better than a few long ones, to ensure data accuracy in behavioural sampling. Fish behaviours were only recorded if they were completely contained within the sample period. Behaviours that began prior to, or ended after a particular sample period, were not recorded. Data extraction during the sediment phase required many randomly generated times to result in a sample that occurred between sediment pulses. Sample periods that occurred during a pulse were not recorded and another random time was generated, unless the PVC rulers defining the field of view were visible. DATA ANALYSIS All data were visually inspected for normality and homogeneous variances. Cortisol and behavioural data were natural log transformed to minimize the effects of heterogenous variances. Trials were treated as replicates thus data were consolidated within each of the experiments to provide single response variables. For each trial that stress was measured I calculated a contrast estimate (CE) from the log transformed mean Cortisol concentrations as follows: CE = (Control -60 Treatment). The contrast estimate was regressed against the suspended sediment concentration. I used univariate ANOVAs to analyse Cortisol data and two-way, repeated measures ANOVAs for movement data. The factors in the models included the effect of time (pre-sediment, sediment, post-sediment), location (control, treatment), a statistical block for the effect of individual streams, and the effect of the day on the response variable; The effect of the day of the experiment is used to control for many different effects that vary over different experiment dates and cannot be considered constant over all dates, such as ambient water temperature, ambient air temperature, UV radiation, etc. Bonferroni multiple range tests were used to determine which response variable means differed (i.e. pre-sediment, sediment, post-sediment). Least square means (LS Means) were used to calculate means for a posterior contrasts. To determine the directionality of movement (upstream or downstream), Chi-Square tests were used to test the independence of the upstream and downstream variables in the control and treatment sections. Behavioural data were analysed using a Poisson process to compare response variables among experiments because these data were not normally distributed due to the large number of zeros in the variables. Al l tests were considered to be significant to a P < .05. Descriptive statistics (mean, +/- 1 SE) were used to describe the side-channel Cortisol results due to lack of replicates. RESULTS Suspended Sediment additions Suspended sediment additions differed on each experimental date. In tributaries, sediment addition concentrations ranged from 209.97 mg/L to 1335.96 mg/L and averaged 601.75 mg/L (SE = 252.76), 531.55 mg/L (SE = 207.99) and 611.95 mg/L (SE = 247.64) for stress, movement and behaviour variables respectively. Mean suspended sediment concentrations varied for each response variable since not all variables were measured in each experiment. The mean 61 background suspended sediment concentration for all tributary experiments was 4.12 mg/L (SE = 0.59) in the control sections. Side-channel sediment additions resulted in a mean concentration of 2797.72 mg/L (SE = 676.37) compared to a mean background suspended sediment concentration of 11.5 mg/L (SE = 2.69) in the controls. The mean discharge during tributary experiments was 0.0512 m3/s (SE = 0.0166) and 0.0077 m3/s (SE = 0.0007) in side-channel experiments. Stress In the tributary sediment experiments I found no differences in mean Cortisol concentrations between netted and un-netted sections (ANOVA, F = 0.07; df = 1,1.92; P = 0.8131) or between control and treatment phases (F = 1.53; df = 1,3; P = 0.3045). The variation attributed to the date factor (0.4529) was very high compared to the residual (0.1719) in the control versus treatment analysis. Date is inextricably confounded with suspended sediment concentration and therefore I could not incorporate sediment concentration as a covariate. The considerable variation attributed to date suggests that strength of the sediment addition may be influencing the response variables. Indeed, the trial with the highest suspended sediment concentration resulted in the largest relative increase in Cortisol concentration from control to treatment (Figure 3.2.). Trials with the second and third highest suspended sediment concentrations had the second and third largest Cortisol responses respectively (Figure 3.2.). I regressed the contrast estimate against the sediment concentration and found a positive linear relationship (Figure 3.3.) that explained 92% of the variation (ANOVA; df = 1,2; F = 22.84; P = 0.0411). Increasing sediment concentrations reveals a decline in the contrast estimate, or simply stated larger sediment additions resulted in larger stress responses. 62 The 98/08/02 side-channel trial had a mean suspended sediment concentration of 1544.67 mg/1 (SE = 179.13) during the sediment phase, compared to 5.00 mg/1 (SE = 0.83) in the control section. Ambient water temperature was 23 °C during this trial. Lethal samples were not taken from the netted control section prior to the experiment because of a shortage of capillary tubes and cryotite in the field (Figure 3.4a.). The 98/08/16 side-channel trial had a mean suspended sediment concentration of 4050.78 mg/L (SE = 827.75) in the treatment section; the control section had a mean suspended sediment concentration of 15.21 mg/L (SE = 3.52). Ambient water temperature was 17 °C. Lethal sampling of juvenile chinook in the 98/08/16 side-channel trial indicated elevated mean concentrations of whole body Cortisol in the treatment section (Figure 3.4b.). Behaviour Movement In tributary trials I found no differences in total number moving among time periods (Repeated Measures ANOVA, F = 0.45, df = 2, 10, P = 0.6513), or between locations (F = 4.76, df = 1,10, P = 0.0541). There was no interaction between time and location (F = 0.54, df = 2,10; P = 0.5985) (Figure 3.5a.). Total number moving in side-channel trials differed among time periods (Repeated Measures ANOVA, F = 102.92, df = 2,2.01, P = 0.0095) and between locations (F = 325.04, df = 1, 2.09, P = 0.0025). The interaction between time and location was significant (F = 115.91, df = 2,2.03; P = 0.0082). Compared to pre-sediment movement, total number moving was elevated over five times in the sediment phase (Bonferroni, P = 0.0045) and over eight times in the post-sediment phase in the treatment sections (Bonferroni, P = 0.0013) (Figure 3.5b.). Total number moving also differed between the sediment phase and post-sediment phase in treatment sections (Bonferroni, P = 0.0175). In control sections, total number moving did not 63 differ among any of the experimental phases (Bonferroni, pre-sediment vs. sediment P = 0.6734; pre-sediment vs. post-sediment P = 0.9476) (Figure 3.5b.). In terms of directionality, in side-channel trials approximately 72% of the juvenile chinook (n = 74) in the treatment sections moved in the upstream direction and 28% moved downstream whereas 50% (n = 32) in the control sections moved upstream and 50% moved downstream (%2 = 4.596, df = 1, P = 0.032). Coughing In tributary trials, the mean frequency of coughing per minute differed between locations (Poisson, x 2 = 18.99, df = 1, P < 0.0001) and among time periods ( x 2 = 11.06, df = 2, P = 0.004). There was no interaction between location and time ( x 2 = 0.04, df = 2, P = 0.9783). Coughing was greater in treatment sections compared to control sections ( x 2 = 14.80, df = 1, P = 0.0001) (Figure 3.6.). Coughing increased over four times between pre-sediment and sediment phases ( x 2 = 6.75, df = 1, P = 0.0094), but did not differ between pre-sediment and post-sediment phases ( x 2 = 0.47, df = 1, P = 0.4946), nor between sediment and post-sediment phases ( x 2 =2.17, df = 1, P = 0.1406) (Figure 3.6.). The variation attributed to date was several times larger than the residual variation. This suggests that the coughing response varied depending on the suspended sediment concentration which ranged widely on the different experimental dates. Feeding In tributary trials, mean feeding attempts per minute differed between locations (Poisson, X 2 = 7.35, df = 1, P = 0.0067) but did not differ among time periods ( x 2 = 4.43, df = 2, P = 0.1091). The interaction between location and time was significant ( x 2 = 20.20, df = 2, P < 0.0001) (Figure 3.7.). In control sections, feeding did not differ between pre-sediment and 2 2 sediment phases ( x = 2.35, df = 1, P = 0.1255), sediment and post-sediment phases ( x = 0.27, 64 df = 1, P = 0.6041), or pre-sediment and post-sediment phases (%2 = 0.47, df = 1, P = 0.4927). In treatment sections, feeding increased nearly two times between pre-sediment and sediment phases ( X = 15.77, df = 1, P < 0.0001) and two times between pre-sediment and post-sediment phases 2 2 ( X = 8.58, df = 1, P = 0.0034), but did not differ between sediment and post-sediment phases ( x = 0.07, df = 1, P = 0.79) (Figure 3.7.). Yawning In tributary trials, mean yawning per minute did not differ between locations (Poisson, %2 = 0.00, df = 1, P = 1.00) but did differ among time periods ( x 2 = 7.12, df = 1, P = 0.0285). There was no interaction between time and location ( j 2 - 4.03, df = 1, P = 0.1335). Yawning differed between pre-sediment (0.036/minute) (SE = 0.032) and sediment phases (0/minute) (SE = 0) ( x 2 =131.01, df = 1, P O.0001), but did not differ between sediment to post-sediment phases (0/minute) (SE = 0.022) ( x = 0, df = 1, P = 1.000), or between pre-sediment and post-sediment phases ( x 2 = 0, df = 1, P = 0.9999). Aggression and Fleeing In tributary trials, I found no differences in aggression per minute between locations (Poisson, x 2 = 0.01, df = 1, P = 0.9298), or among time periods ( x 2 =4.17, df = 2, P = 0.1244). There was no interaction between time and location ( x = 2.25, df = 2, P = 0.3252). Similarly, fleeing per minute did not differ between locations ( x 2 = 0.08, df = 1, P = 0.7816), or among time 2 2 periods ( x = 3.21, df = 2, P = 0.2008). There was no interaction between location and time ( x -2.17, df= 2, P = 0.3372). 65 Fishmin and Total fish In tributary trials, I found no differences in fishmin between locations (ANOVA, F = 1.87, df = 1,14.1, P = 0.1934), or among time periods (F = 0.34, df = 2,14.1, P = 0.7191). There was no interaction between time and location (F = 0.31, df = 2,14.1, P = 0.7395). Total fish observed in tributary trials did not differ between locations (F = 2.01, df = 1,14.3, P = 0.1775) or among time periods (F = 1.18, df = 2,14.3, P = 0.3362). There was no interaction between location and time (F = 0.03, df = 2,14.3, P = 0.9713). Analyses of behavioural data were not completed for side-channel trials because the high suspended sediment concentrations precluded adequate data extraction from the sediment and post-sediment phases. DISCUSSION Stress The stress response in the tributary sediment experiment was consistent with my prediction and other studies. Suspended sediment concentrations ranging between 500 to 1500 mg/L have caused alterations in blood chemistry of salmonids, indicative of sub-lethal stress (Servizi and Martens 1992, 1987; Redding et al. 1987). Sub-lethal acute stress responses in fish can also occur at very low concentrations of suspended sediment. Arctic Grayling exposed to 50 mg/L of suspended sediment for 1 - 4 days exhibit a stress response (McLeay et al. 1987). My study suggests that short exposures of suspended sediment can also induce stress responses. Fish in the first side-channel trial were very stressed. The treatment plasma Cortisol concentration was several times higher than juvenile coho subjected to severe, prolonged handling stress (Sumpter et al. 1986; Olla et al. 1992). However, it is impossible to say how much was attributable to the sediment manipulation versus the net design. In addition, stress levels were high prior to any influence of the experiment (net or sediment), since the un-netted control plasma 66 concentration was over eight times greater than typical basal rates reported for juvenile salmonids (Sumpter et al. 1986; Fagerlund et al. 1995). High water temperature likely contributed to this stress response, clouding the influence of the sediment additions. Indeed, in the Fraser River system, sediment sensitivity in salmonids has been reported to be temperature dependent (Servizi and Martens 1991). The second side-channel trial had a lower ambient water temperature which likely did not similarly confound this experiment. The sediment concentration was greater in this trial, and resulted in approximately a doubling of the tissue Cortisol concentration compared to control values. Stress has been defined as "the sum of all the physiological responses by which an animal tries to maintain or re-establish a normal metabolism in the face of a physical or chemical force" (Selye 1950). When a biotic or abiotic challenge extends the homeostatic process of fish beyond the point of coping, stress occurs (Anderson et al. 1996). Elevated Cortisol, indicative of stress, is an endocrine response invoked by trauma (Barton and Toth 1980), that in turn evokes secondary effects on metabolic processes. Elevated Cortisol is beneficial to fish during times of trauma, since it facilitates increased swimming ability and other behavioural processes (Olla et al. 1992). However, chronic exposure to sub-lethal stress can result in substantial metabolic costs, among other consequences. The sediment experiments simulated a one time sedimentation event typical of forest harvesting activities (Berg and Northcote 1985; Scrivener 1988; Scrivener and Brownlee 1989; Waldichuk 1993). Many of the acute stress responses I measured could become chronic because increases in suspended sediment concentrations from forest harvesting are generally not a one time event (Brownlee et al. 1988). Prior exposure to a stressor can result in increased tolerance, but it does not ensure that this tolerance will be sustained (Fagerlund et al. 1995). Natural and forest harvesting induced occurrences of suspended sediment are by nature, episodic, which may 67 limit the development of tolerance. Further, the physiological effects of chronic sub-lethal stresses, such as multiple sedimentation episodes, are cumulative and can result in a stepwise pattern of increased plasma Cortisol (Barton et al. 1986). However, fish populations rearing in northern interior systems with highly erodible, glaciolacustrine soils such as the Torpy watershed may have evolved some genetic resistance to high suspended sediment concentrations providing increased tolerance to natural and forest harvesting-induced elevations of suspended sediment. Short exposures to relatively low concentrations of suspended sediment, particularly multiple episodes, and the coincident increased levels of Cortisol, could have negative consequences for juvenile chinook rearing in northern, interior systems. For example, multiple exposure to sub-lethal stress during the summer and fall can cause lipid depletion (Lemly 1996), which would compromise over-winter survival. Elevated plasma Cortisol can affect social behaviour, sensory abilities, and the ability to learn behaviours (Olla et al. 1992). The ecological implications of cumulative sub-lethal effects are increased susceptibility to predation (Olla and Davis 1989; Kruzynski and Birtwell 1993; Korstrom et al. 1998), decreased fitness, increased susceptibility to disease and other stressors (Fagerlund et al. 1995; McLeay et al. 1987; Herbert and Merkens 1961; Redding et al. 1987; Servizi and Martens 1991) as well as decreased ability to compete for food and space. Movement and Behaviour Contrary to my prediction, tributary chinook were reluctant to leave their territories during sediment additions, displaying lateral avoidance or holding position instead. Even though tributary trials were stressful to juvenile chinook (based on the Cortisol results), these individuals endured these conditions rather than exhibit a strong avoidance response. This reluctance to avoid unfavourable conditions may be because tributaries are preferred habitat. Other researchers have 68 also shown that ecological motivation can cause juvenile salmonids to temporarily override adaptive avoidance responses and endure sub-lethally stressful environmental conditions (Birtwell et al. 1999). Indeed, stream-type juvenile chinook have been reported to display considerable permanence of station (Edmundson et al. 1968; Rosberg et al. 1981; Bustard 1983), which has been suggested to confer many benefits including familiarity with cover and food sources (Chapman and Bjornn, 1969), providing clear adaptive benefits considering the relative value of tributary habitats compared to side-channel habitats (Chapter 2). Suspended sediment is known to trigger lateral or vertical avoidance behaviour in salmonids (Bisson and Bilby 1982; Berg and Northcote 1985; Servizi and Martens 1992). In their laboratory study, Servizi and Martens (1992) found that juvenile coho were more prone to exhibit lateral rather than vertical avoidance. In my research, juvenile chinook that moved laterally in the water column were able to avoid the majority of the sediment plumes because the stream margins generally remained clear. Lateral avoidance in response to elevated turbidity would also permit juvenile chinook to retain visual and tactile contact within their physical environment, because fish use visual contact with the stream bank and substrate to maintain position in flowing water (Berg and Northcote 1985). The ability of juvenile chinook to cope with elevated suspended sediment via lateral avoidance is greater in tributaries likely because of the greater relative proportion of stream margin habitat available. Indeed, highest densities of cutthroat trout and coho salmon in small streams has been suggested to be attributable to the greater relative proportion of edge habitat in smaller streams (Rosenfeld et al. 2000). Lateral avoidance was not observed in side-channel trials. As predicted, sediment additions in side-channel experiments triggered a strong avoidance response in juvenile chinook. Turbidity (hence suspended sediments) has been shown to reduce the perceived risk of predation in juvenile chinook (Gregory 1993; Gregory and Northcote 1993) 69 resulting in elevated activity (Gregory 1993; Gradall and Swenson 1982). Because predators are considerably more prevalent in side-channels compared to tributaries (Chapter 2), a reduction in perceived risk of predation may contribute to increased movement in side-channel trials. However, juvenile chinook moved in aggregations during side-channel trials, huddling together and moving in rapid, erratic darting movements, characteristic of stressed fish (Mason and Chapman 1965). Avoidance of stressful conditions is more likely the causal agent, rather than elevated activity due to the cover of turbidity. Suspended sediment concentrations were higher during side-channel experiments compared to tributaries. However, since juvenile chinook rearing in side-channels would be more acclimatised to high suspended sediment concentrations (Chapter 2), this difference alone seems insufficient to explain the different movement responses. Results from other components of my research, (Chapter 2), have shown that side-channels are the lowest value habitats and that tributaries are the highest value habitats for juvenile chinook rearing in the Torpy River watershed. Considering the disparity in value of these two different macrohabitats, to juvenile chinook, the pronounced avoidance response in side-channels and reluctance of juvenile chinook to evacuate tributaries, is not surprising. The mass exodus of mainstem and side-channel rearing juvenile chinook in the Torpy River occurs during late summer, most likely triggered by low flows and the coincident extensive sedimentation (Chapter 2) that precludes over-winter survival. In fact, additions of sediment have been demonstrated to be negatively correlated to abundance of over-wintering juvenile chinook (Hillman et al. 1987; Bjornn et al. 1977). Elevated suspended sediment during side-channel experiments may have elicited a similar innate migration response, which may not have evolved as part of the behavioural repertoire of tributary individuals, due to the presence of warm groundwater flows providing over-wintering opportunities. 70 In laboratory and artificial stream experiments, avoidance responses to suspended sediment seems to vary among fish species, fish size, water temperature, previous exposure to suspended sediment and other stressors, and sediment particle size (Birtwell 1999; McLeay et al. 1987; Servizi and Martens 1992 ). The reluctance of tributary rearing juveniles to avoid pulses of suspended sediments in my study may be partly due to the more realistic and natural conditions of my experiments. Indeed, fish are generally able to tolerate a much higher degree of pollution in nature because of motivation (Sprague 1971) and ecological realism (Adams 1990), both of which are often lacking in laboratory studies. Avoidance of unfavourable conditions has obvious adaptive value; however, if it leads to displacement to sub-optimal habitats it can result in fitness consequences for juvenile salmonids. McLeay et al. (1987) reported downstream displacement of Arctic grayling exposed for one hour to suspended sediment concentrations of 300 mg/L. Downstream displacement has also been reported for juvenile coho salmon exposed to 25 NTU (Sigler et al. 1984). Downstream displacement can be quite harmful to juvenile chinook considering the high habitat value of tributaries. Displacement from preferred habitat would likely result in reduced growth and survival, as well as increased susceptibility to predation in other sub-optimal habitats (mainstem and side-channels). The frequency of coughing increased in response to sediment additions. This is consistent with findings in the literature (Berg and Northcote 1985; Noggle 1978; Servizi and Martens 1992). Elevated coughing indicates a difficulty with respiration and impaired oxygen exchange. In order to keep the gills clear of sediment for oxygen exchange, fish must exert energy to perform the cough reflex (Servizi and Martens 1991), which in turn increases metabolic oxygen demand. In tests of lethal concentrations of suspended sediment on juvenile salmonids, Servizi and Gordon (1990) reported buccal cavities of dead fish to be filled with sediment because fish 71 became too exhausted to continue clearing sediment via the cough reflex. Thus, the implications of elevated coughing are that fish may have less energy to allocate to feeding, growth and survival. The rates of coughing reported in my research are substantially lower than those observed in other studies (e.g. Servizi and Martens 1992). This could be due to their use of hatchery specimens, longer exposures (96 hr), but may largely be an artifact of most previous experiments not being conducted in situ. Because chinook were able to express lateral avoidance and use margin habitat, this may have mitigated the effects of sediment on the gill surface to a certain degree. Servizi and Martens (1992) confined coho to small cages (15x3x4 cm) which were submerged at a fixed depth, thus limiting behavioural responses. Pulses of suspended sediment (30 - 60 NTU) have been documented to disrupt dominance hierarchies in juvenile coho salmon (Berg and Northcote 1985) and reduce the ability of juvenile coho to hold territories (Hartman and Holtby 1982). The mechanisms involved in this breakdown in social behaviour have been suggested to include increased food (Mason 1976) and reduced visual ability at elevated turbidities, since visual isolation results in declines in territorial interactions (Kalleberg 1958; Chapman 1966; Berg and Northcote 1985; Harvey et al. 1999). Since the presence of reference objects induces territorial behaviour in juvenile salmonids (Hartman 1963), it follows that the inverse should hold true as well. Contrary to my prediction, aggression and fleeing did not decline during tributary experiments. Reductions in feeding have been reported for juvenile chinook exposed to 100 mg/L (Birtwell 1999), Arctic grayling exposed to 100 mg/L (McLeay et al. 1987), for coho salmon exposed to 100 mg/L (Noggle 1978), 2000-3000 mg/L (Redding et al. 1987), and 11 - 55 mg/L (11 NTU converted from Berg and Northcote, 1985), and a cessation of feeding at 300 mg/L (Noggle, 1978). Suspended sediment can be a major contributor to turbidity (Waters 1995). 72 Salmonids are visual predators, thus I predicted changes in turbidity to decrease the rate of feeding; however, the inverse was observed. During tributary experiments, mean foraging rate increased in response to sediment additions which is consistent with findings from others (Mason 1976; Gregory and Northcote 1993; Gregory 1993, 1994). Increased food, in the form of drift, is available during sediment pulses (Gammon 1970; Culp et al. 1986; Anderson et al. 1996; Birtwell 1999; Shaw and Richardson 2001; Rosenberg and Wiens 1978). Although I did not quantify foraging success, juvenile chinook were observed to increase foraging attempts at non-prey items such as bits of detritus, sticks and leaves during sediment additions. Thus, although the rate of feeding increased in response to sediment additions, I suspect that the quality of foraging declined. Under turbid conditions fish may take in less energy due to the ingestion of non-prey items. Indeed foraging efficiency has been demonstrated experimentally to decline with decreasing light (Wilzbach et al. 1986). In laboratory and artificial stream experiments, prey reaction distance and prey capture efficiency in juvenile salmonids has been found to decline in response to pulses of sediment (Berg and Northcote 1985; Hartman and Holtby 1982; Gregory and Northcote 1993). Others have noted positive effects of suspended sediment on juvenile chinook (57-69 mm FL), including increased feeding at intermediate turbidity levels likely caused by enhanced cover from predation (18-150 NTU, Gregory and Northcote 1993; Gregory 1993,1994; Gregory and Levings 1998). However, for smaller individuals (49 - 55 mm FL), chinook feeding rates are reported to decline with increasing turbidity (Gregory 1994). The mean lengths of juvenile chinook in tributary trials were less than 57 mm, yet feeding rates increased during sediment pulses. Since avian and aquatic predators are rare in tributaries compared to other macrohabitats (see Chapter 2), cover from predators did not likely cause the observed increase in feeding. Rather, I suspect the elevated availability of potential food items was responsible. 73 Forest Harvesting Implications Elevated suspended sediment and subsequent deposition in aquatic habitats is one of the most pervasive problems facing habitat managers (Newcombe and MacDonald 1991). My research suggests that experimental additions of suspended sediment at relatively low concentrations and short exposures cause sub-lethal stress in juvenile chinook. Of the Pacific salmon species, juvenile chinook are the most tolerant to suspended sediment concentrations (Servizi and Gordon 1990; Servizi and Martens 1991). Thus, other salmonid species may be even more vulnerable to suspended sediment caused by forest harvesting activities. Although the future timber supply in the northern hemisphere will come from boreal and subboreal forests (Bryant et al. 1997), the effects of forest harvesting on fish populations in snowpack dominated, interior watersheds is not well studied (Mellina 2002). Past research regarding the effects of forest harvesting on fish populations has been mostly restricted to rainfall dominated coastal watersheds (Hartman and Scrivener 1990; Hetrick et al. 1998a&b; Hall and Lantz 1969; Murphy et al. 1986; Keith et al. 1998; Beschta et al. 1987; Toews and Moore 1982a&b). Furthermore, field experiments that manipulate one effect of forest harvesting in natural streams have not been conducted in any region. The highly erodible, glaciolacustrine soils in the Torpy watershed result in unstable streambanks and high suspended sediment loads indicative of northern interior systems (MacDonald et al. 1996). Thus, my study results have application to other boreal and subboreal watersheds. The timing of peak suspended sediment events in the Torpy River mainstem are associated with June snowmelt periods and rainfall events in October, which is consistent with findings from other researchers in the Upper Fraser River (Brownlee et al. 1988). The majority of sediment is mobilized into suspension during the rising limb of the hydrograph (Everest et al. 1987), which for the Torpy River would occur in early June 74 and October. Smaller juvenile salmonids are less able to tolerate elevated suspended sediment than their larger cohorts (Sigler et al. 1984). Indeed, Servizi and Martens (1991) reported that small juvenile coho salmon (<40 mm) have only 35% of the tolerance of suspended sediment compared to larger (>46 mm) individuals, possibly due to a limited ability to clear their gills and buccal cavities via the coughing reflex (Servizi and Martens, 1991). Considering the size dependent nature of juvenile salmonid tolerance to suspended sediment, the timing of the early June peaks of suspended sediment in the Torpy River is of particular concern since they coincide with an early period of juvenile chinook ontogeny, underscoring the importance of refuge habitats at this time. 75 Study Sites Experimental Design Torpy River I tributary sidechannel downstream camera Flow sideview camera * experimental manipulation JJ downstream camera PS sideview camera A B Figure 3.0. Diagram (A) representing the macrohabitats (tributaries, side-channels) in which experiments were conducted. Schematic (B) of general experimental design (not to scale). The location of sediment manipulations is indicated by the asterik. 76 Pre-Sediment phase Sed iment phase Post-Sediment phase 1 1 I . 1 1 1 1 1 1 J 90 minutes Figure 3.1. Temporal methods of behavioural data extraction from underwater video recordings. Each phase of the experiment was divided into 10 equally spaced intervals (e.g. 6 min. intervals in Pre-Sediment, 9 min. intervals in Sediment, 6 min. intervals in Post-Sediment) and random samples were extracted from each interval. The magnification of the Sediment phase illustrates the systematic random approach. Black bars are randomly selected intervals for data extraction. 77 10 1 7 c •2 6 e S 5 o c ° 4 = 3 O CO •e o o 518.5 mg/L 1335 mg/L 342.6 mg/L 209 mg/L Un-netted Netted Control Treatment Figure 3.2. Mean Cortisol concentrations during tributary trials with the addition of differing suspended sediment concentrations (indicated in parentheses). Means are based on the number of fish collected (n = 8 for each). 78 0.4 -1.4 1 0 200 400 600 800 1000 1200 1400 1600 Sediment Concentration (mg/L) Figure 3.3. The contrast estimates calculated from mean Cortisol concentrations plotted against suspended sediment concentration of each experiment. The regression model and coefficient of determination are presented. 79 300 _ 250 E 200 cn c__ c o •*= CO = 150 9 O o 100 e o O 50 18 16 D3 14 12 B O I 10 c 0) g 8 o o 0 6 w 1 4 O - L . B Un-netted Netted Control Treatment Figure 3.4. Mean Cortisol concentration concentrations (± 1 SE) for the 98/08/02 side-channel trial (A) and the 98/08/16 side-channel trial (B). Means are based on number offish collected (n = 8 for each). 80 30 25 | 2 0 O E i_ o -Q 15 E 3 C f 10 a a a a 30 25 O) c > o E Im <D E | 10 o 20 15 B a a a Pre-sediment Sediment • Contro l • Treatment Post-sediment Figure 3.5. Mean total number of fish moving during tributary (A) and side-channel (B) sediment trials. Bars with similar letters did not differ. Error bars represent ± 1 SE. Means based on the number of trials (n = 4 and n = 2 respectively). 81 0.16 CD 3 C 0.12 O a 0.08 D) t C D) O 0.04 o Control Treatment Pre- Sediment Post-sediment sediment Figure 3.6. Mean frequency of juvenile chinook coughing per minute during tributary sediment experiments. Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on the number of trials (n = 4). 82 • Control • Treatment Pre-Sediment Sediment Post-Sediment Figure 3.7. Mean feeding attempts/minute for tributary sediment experiments. Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on the number of trials (n = 4). 83 CHAPTER 4: EXPERIMENTAL FIELD MANIPULATIONS OF TRIBUTARY AND SIDE-CHANNEL TEMPERATURES INTRODUCTION Stream temperature plays a central role in aquatic ecosystems because it influences metabolic rates, growth, behaviour, and survival of fish. Of all abiotic factors contended with, temperature has been referred to as the "master ecological factor" for juvenile fish (Brett 1971). As water temperature increases, the solubility of dissolved oxygen decreases, diminishing its availability to aquatic organisms. This creates a double jeopardy scenario for poikilotherms, since during periods of intensified metabolic and respiratory demands, less oxygen is available. High temperatures can be lethal to fish as a result and act as stressors at sublethal levels. Oxygen demand approximately doubles for fish when temperatures rise by ten degrees Celsius (Theurer et al. 1985), thus at high temperatures oxygen demand could exceed supply. Temperature influences fish distributions (Brett 1971; Levy 1992; Glova and Mason 1977) and fish behaviour (Beschta et al. 1987; Glova and Mason 1977; Hartman 1963). For poikilotherms, the primary mechanism for body temperature regulation is through behaviour (Levy 1992). Indeed, salmonids display considerable behavioural plasticity when faced with stream temperatures that deviate from their preferred range and have been documented to avoid and/or select habitats based on thermal regimes (Beschta et al. 1987) and to compete for favourable temperatures (Beitinger et al. 2000). The harvest of riparian vegetation can result in elevated summer stream temperatures due to increased solar radiation (Hartman and Scrivener 1990; Ringler and Hall 1975; Beschta et al. 1987; Holtby 1988; Brown and Krygier 1970; Brownlee et al. 1988; Hetrick et al. 1998; Rishel et al. 1982; Hall and Lantz 1969; Shrimpton et al. 1999a&b). For example, Holtby and Newcombe (1982) found a 7° C increase in the mean temperature of Carnation Creek, British Columbia 84 between May and October after 39% of the watershed had been clearcut. Small tributary streams are highly susceptible to changes in water temperature due to their inherent large surface area to volume ratio. Riparian forest harvesting also results in increased maximum and minimum stream temperatures, as well as an increase in diurnal temperature fluctuations (Hartman and Scrivener 1990; Rishel et al. 1982; Brownlee et al. 1988; Hall and Lantz 1969). This may have a more profound influence on fish than increases in mean temperatures. Mean stream temperature in three northern interior British Columbian streams increased 1 to 3° C after forest harvesting, yet maximum temperatures were up to 9° C greater (Brownlee et al. 1988). These authors also found that diurnal temperature fluctuations more than doubled in downstream logged reaches in comparison to upstream control sections (Brownlee et al. 1988). In fact, this was demonstrated experimentally in the Torpy River watershed; manipulations to riparian zones resulted in greater diel fluctuations in temperature in sections of groundwater controlled streams that were logged than in un-logged reaches (Shrimpton et al. 1999a,b). Because the stream's canopy can insulate the stream, riparian forest harvesting can result in colder winter stream temperatures. Minimum stream temperatures declined by 3.9° C after forest harvesting during fall and winter in a northeastern United States stream (Rishel et al. 1982), and by 2° C in northern interior British Columbia (Brownlee etal. 1988). Past research regarding the effects of forest harvesting on fish populations has been mostly restricted to rainfall dominated coastal watersheds (Hartman and Scrivener 1990; Hetrick et al. 1998a&b; Hall and Lantz 1969; Murphy et al. 1986; Keith et al. 1998; Beschta et al. 1987; Toews and Moore 1982a&b). The effects of forest harvesting on fish populations in snowpack dominated, interior watersheds are less published (Mellina 2002). However, the future timber 85 supply for the northern hemishpere is expected to come from boreal and sub-boreal forests in temperate interior regions (Bryant et al. 1997) such as those in northern, interior British Columbia. Research focusing on the lethal and sublethal effects of elevated temperatures on fish behaviour and physiology has primarily concentrated on laboratory experiments and artificial streams (Brett 1952, 1970; Bisson and Davis 1976; Coutant 1973; Linton et al. 1998a&b, 1997, 1999; Sylvester 1972; Glova and Mason 1976; Reid et al. 1995; Birtwell et al. 1999; Gibson 1988), or has been synoptic and comparative in design (Filbert and Hawkins 1995; Kaeding 1996; Brown and McMahon 1988; Welsh etal. 2001; Brown 1999; Brown and MacKay 1995). Synoptic studies suffer from the difficulty in isolating causal factors. For instance, Brown and McMahon (1988) observed higher growth of coho juveniles rearing in off-channel habitats compared to the mainstem Carnation Creek, but were unable to determine whether this difference was due to temperature, food availability, or differential energy expenditure from dissimilar water velocities. Although laboratory studies and experiments conducted on artificial streams are valuable to provide insights to causal relationships between characteristics of fish habitat and effects on fish, they generally lack ecological realism (Adams 1990; Waldichuk 1993). Indeed, fish are generally able to tolerate a much higher degree of pollution in nature because of motivation (Sprague 1971), which is often lacking in laboratory studies. Furthermore, fish are generally held at static temperatures in laboratory experiments, whereas diurnal fluctuations predominate in natural conditions (Beschta et al. 1987). Conclusions drawn from laboratory experiments may not apply to the spatial and temporal complexities of thermal regimes in natural systems. The homogeneous laboratory environment is relatively sensory deprived; intuitively the response of salmonids to manipulated temperatures, including aspects such as microhabitat selection and avoidance, would be most natural in a field setting. 86 In this chapter, I report the response of chinook salmon to experimental field manipulations of temperatures in small tributary streams and side-channels of the Torpy River, a snowpack dominated interior system in northern British Columbia (see Chapter 1 for study site description). Forest harvesting activities are prevalent in the Torpy River. I simulated increases in water temperatures, to mimic those induced by forest harvesting and quantified their biological effects on juvenile chinook salmon. These experiments enabled the isolation of other variables associated with forest harvesting which often generate confounding effects (i.e. increased light, altered hydrology, altered habitat diversity, etc.) on fish populations. Elevated temperatures can be beneficial to fish. Increased metabolic rates and growth can occur in response to warmer temperatures. Extremely high temperatures can cause physiological stress. Therefore, I expect to find elevated levels of whole body tissue Cortisol, indicative of sub-lethal stress, when temperatures exceed 15° C - a level which starts to exceed Chinook's preferred range (Brett 1952). I also expect that temperature increases will result in elevated movement, indicative of avoidance, if manipulated temperatures exceed 15° C. Because side-channels provide the lowest habitat value for juvenile chinook, and have higher ambient temperatures (Chapter 2), I expect that avoidance responses will be more pronounced in side-channels. Avoidance of unfavourable conditions has obvious adaptive value, but if it leads to displacement to sub-optimal habitats it could have fitness consequences for juvenile salmonids such as decreased growth, increased predation or increased susceptibility to disease. Because increasing temperature causes increases to fish metabolism (Brett 1995), I lastly expect that temperature increases will result in elevated frequencies of feeding, as well as increased levels of social interactions among individuals (leading to higher levels of aggression, fleeing, and yawning). METHODS 87 I used three tributaries and one side-channel as sites for the experiments. The tributary experiment consisted of five trials which were conducted in 1997 and 1998; the side-channel experiment consisted of two trials which were conducted in 1998 (see Figure 4.0a for an example of a side-channel and tributary). The duration of each experiment was approximately three hours and 30 minutes, divided into a pre-heating phase (1 hr.), a heating phase (1.5 hrs.) and a post-heating phase (1 hr.). Each heating experiment was initiated at 11:00 am. Discharge was measured by flow meter and measuring tape according to standard methodology (RIC 1997; Schuett-Hames et al. 1994) prior to each trial. Channel widths of the tributaries ranged from 1 to 2.5 m and share similar habitat characteristics. Tributaries flowed in a southerly direction, had a low gradient (1-2%), had a substrate composition of 70% fines and 30% gravels, and abundant cover in the form of undercut banks, over-hanging vegetation, deep pools and woody debris. Channel width of the side-channel was approximately 3.5 m. It connected to the mainstem and flowed freely at the time of both experiments. The side-channel was characterized by low velocity glides and deep pools with a substrate composition of 100% fines. Cover was limited to overhanging vegetation, deep pools and the occasional piece of woody debris. The tributary and side-channel riparian had never been logged. Homogeneous sections of tributaries and side-channels, approximately 100 metres in length, were selected each containing a minimum of two sets of pools, riffles, and glides. Enclosure nets (1 mm mesh) were installed at the downstream and upstream end of each section as well as one in the centre, resulting in two, 50 m, isolated sections. The upstream section was identified as the control and the downstream section was the treatment (Figure 4.0b). The fact that control sections were located immediately upstream from experimental sections ensured similar physical stream reach characteristics and fish population characteristics (species 88 abundance, distribution, etc.). A series of three weirs that extended approximately 90% across the wetted channel on a downstream angle of 30 degrees were installed at each end of the treatment and control sections (Figure 4.0b). Two-way weirs trapped fish that moved into these areas enabling the assessment of upstream and downstream movement direction. Movement data was only collected in 1998. Streams were heated by pumping water from the stream to a series of heaters, using a variable speed Honda portable pump (50 gpm) (Model WB15) connected to a non-collapsible hose. The intake consisted of a filter basket covered by fine netting to prevent damage to the pump by sediment. Water was pumped at a rate of approximately 16 L/min through a series of four propane powered instantaneous hot water heaters (Bosch Booster Pressure Wash Model W400K5, 117 000 btu), and reintroduced into the stream 25°C above ambient stream temperature through a diffuser system to ensure mixing. Each diffuser unit was constructed from five household showerheads in a manifold type configuration. The water intake was located downstream of the diffuser unit to ensure maximum heating of the stream or side-channel water temperature. The heaters were powered by two, 100 lb propane tanks which provided approximately 12 hours of use. The propane powered, instantaneous water heaters were only ignited during the heating phase of the experiment, yet the pump was operated during all phases to ensure that responses exhibited by juvenile chinook were a result of the elevated temperature, and not due to a subtle flow alteration. Twelve temperature data loggers (Onset instruments) were deployed at 10 m intervals in the side-channel and tributary sites. Two loggers were placed immediately upstream from the heating apparatus and the rest were placed downstream. Loggers recorded temperature every five minutes during experiments and for several days before and after experiments. Upstream loggers were used for the reference data. 89 Four video cameras (Citizen model JSS 1012C) housed in dark green, waterproof, aluminum housings (approx. dimensions: 50 mm x 105mm) with Lexan lids for optical clarity were used to observe fish behaviour. In both the treatment and control sections one camera was positioned at mid-channel and aimed downstream and one other was positioned at the stream margin aimed towards mid-channel, perpendicular to flow, (Figure 4.0b.). Sections of 1 cm diametre PVC pipe, marked at 5 cm intervals, were installed upright in the substrate in the field of view of the underwater cameras. The pipes functioned as reference points to ensure that the field of view was the same area (generally lmxlm), in both side-view and frontal view cameras, as well as between treatment and control sections. Two, 25.4 cm, black and white monitors (Citizen model JSS 1012M) contained in rainproof Lexan housings, tinted to eliminate sun glare, were used to display images which were recorded by two VHS video recorders (Sanyo model VHR-H607), also in Lexan housings. The underwater video cameras and recording system were powered by a Honda variable speed generator. Experimental Procedures One litre water samples were taken every 15 minutes in both the control and treatment sections during each phase of the experiment and analysed in the lab for suspended sediment concentration by filtration (mg/L). Turbidity was measured in the treatment section during all phases of the experiments by a nephelometric turbidity probe (Analite NEP 190) set to take readings at five second intervals and averages every five minutes. Turbidity data were recorded in units of NTU by a Starlogger data logger (Model 3.08). Temperature data was recorded every five minutes. Fish numbers (as an indicator of movement) were recorded at each of the four weirs (upstream and downstream treatment, upstream and downstream control) every 15 minutes during 90 each of the pre-heating, heating, and post-heating phases. The number of fish moving both upstream and downstream in each section (treatment and control) were pooled within each experimental phase and termed "total number moving". Simultaneous underwater video recordings offish behaviour in the treatment and control sections were taken continuously during all phases of the experiments. Samples of juvenile chinook (n=6-10 from each site) were taken by electroshocking, during the pre-heating phase in the control section (netted), and upstream of the control section (un-netted) to determine the influence of the net on acute stress. Simultaneous underwater video recordings of fish behaviour in the treatment and control sections were taken continuously during all phases of the experiments. Fish were snap frozen on dry ice (< 60 sec) for laboratory whole body Cortisol analyses. The pre-heating phase lasted approximately one hour at which time water heaters were ignited and heating manipulations initiated. Heated water was added in the middle of the treatment section. The heating phase lasted approximately 1.5 hours, at which time lethal sampling of juvenile chinook (n=6-10/site) occurred in both the control and treatment sections. During the one hour post-heating phase, suspended sediment, turbidity, fish movement, and behaviour were collected. In the 98/08/01 side-channel trial, blood plasma (ng/ml) was extracted and centrifuged in the field as an alternative means of assessing Cortisol stress due to a shortage of dry ice. Because of the different techniques and experimental units, acute levels of juvenile chinook Cortisol concentrations could not be statistically compared between side-channel trials. Data Extraction To save time during field experiments, I froze fish whole rather than extract blood plasma for analyses of Cortisol levels. Whole body Cortisol levels from lethally sampled juvenile chinook were assessed in the laboratory. After capture, fish (ranged in mass from 0.41 to 6.36 g) were 91 transferred from dry ice to a -80 °C freezer. Frozen fish were ground into powder using a mortar and pestle in a slurry with liquid nitrogen which was then evaporated off. Distilled water, 300 ul, was added to each sample which was then disrupted by ultrasonication for one minute (B. Braun Biotech, Braun-Sonic L, Allentown, PA, 18103). Five millilitres of diethyl ether was added to samples which were mixed then centrifuged at 1000 g for 5 min. The water component was frozen in liquid nitrogen and the ether decanted. Samples were then re-extracted and the ether fractions combined. Ether was evaporated under a stream of nitrogen gas. Following complete evaporation of ether, 2 ml of phosphate buffered saline (1.5 mM KH2PO4, 2.7 mM Na2HPC>4, 150 mM NaCl, pH 7.4) containing 2% bSA (bovine serum albumin) was added to samples which were kept at 4 °C and vortexed frequently for 2 h. Cortisol concentrations were determined by 125 radioimmunoassay using I (Diasorin, Stillwater Minnesota, 55082-0285). Efficiency of extraction was determined from spiking paired samples with Cortisol and was found to be 92 to 95%. Whole body Cortisol is expressed as ng/g tissue. Information on the following behaviours were transcribed from the video tapes: feeding attempts, coughing, aggression, yawning, fleeing, fish numbers, and fish minutes. These behaviours, which are described in detail below, were selected because they were quantifiable, repeated across treatments, were readily identifiable, and were biologically relevant for temperature manipulation studies. In all cases, frequencies of behavioural activities were recorded. Time constraints limited the ability to extract duration, intensity or pattern of behavioural activities. Feeding attempts were defined as the number of foraging attempts per minute. The high resolution of the cameras enabled the observation of individual prey items. Feeding attempts consisted of a rapid opening and closing of the mouth for a potential prey item and often 92 coincided with a change in orientation in the water column, or small movement to capture a prey item. Foraging success was not quantified. Coughing was defined as the backward propulsion of water across the gills to displace obstructions (e.g. sediment), and was readily identified by excessive flaring of the operculum, gaping mouth and rapid spasm of the buccal cavity. Displaced items were often observed during this activity, as was a small forward movement of the fish (Berg and Northcote 1985). Aggression was defined as any agonistic behaviour directed toward another fish, which included chasing, nipping, and charging (Hartman 1965; Chapman 1962). Yawning was defined as the protracted and excessive gaping of the mouth, and was easily distinguishable from feeding due to an absence of a potential prey item, a lack of any change in position or orientation in the water column, and a very protracted and excessive gaping of the mouth. Fleeing was defined as evasive behaviour, most often displayed in response to another aggressive individual. Fleeing fish were displaced from their territory or previously occupied position either temporarily or permanently. The number of fish observed per sampling period was recorded. Fish moving in and out of the field of view were not considered to be different fish; however, this variable provides an indication of activity level. Fish were not marked due to their small size for many of the experiments (<5.0 cm), and to prevent artificially altering behaviour as a result of marking. The frequencies of extracted behaviours were divided by the time observed to provide a frequency of behaviour per minute that could be compared between different sections and phases of experiments. The total time that each fish was in the field of view in each section (treatment, control) was summed for each phase of the experiment (pre-heating, heating, post-heating) to provide an accumulated time. This was divided by the sample period time to provide an estimate of equivalent number of fish viewed per minute, termed "fishmin". 93 A systematic random approach was used to extract behavioural data from the underwater video recordings. Each phase of every experiment was divided into ten equal time intervals for a total of 30 intervals in the control recordings per experiment and 30 intervals in the treatment recordings per experiment (Figure 4.1.). Considerable deliberation and effort was taken to determine the appropriate sampling approach to extract data from each interval. A test was completed whereby, behavioural data were extracted continuously from a 16.3 minute segment of video footage, and the frequencies of behaviour were calculated and treated as true frequencies. I extracted replicate samples from this segment of different fixed times (1 min 38 seconds, 2 min 30 seconds, and 4 min 8 seconds) starting from randomly generated times within the segment. The percentage deviation of the behavioural means extracted from each different fixed sampling time was compared to the true values. All of the means were well within 1 SD of the true values and since the 1 min 38 second fixed sample time was >97% of the true values, it was decided to proceed with that sample time which would allow more replicates. Due to the time consuming nature of data extraction it was estimated that the number of intervals would have to be reduced from 10 per phase, to 5 per phase, if a longer sample time was chosen. Indeed, Zinner et al. (1997) similarly documented that many short sampling periods are better than a few long ones, to ensure data accuracy in behavioural sampling. Fish behaviours were only recorded that were completely contained within the sample period. Behaviours that began prior to or ended after a particular sample period were not recorded. DATA ANALYSIS All data were visually inspected for normality and homogeneous variances. Cortisol and behavioural data were log transformed to minimize the effects of heterogenous variances. Trials 94 were treated as replicates thus data were consolidated within each of the experiments to provide single response variables. I used univariate ANOVAs to analyse Cortisol data and two-way, repeated measures ANOVAs for movement data. The factors in the models included the effect of time (Pre-Heat, Heat, Post-Heat), location (control, treatment), a statistical block for the effect of individual streams, and the effect of the day on the response variable. The effect of the day of the experiment is used to control for many different effects that vary over different experiment dates and cannot be considered constant over all dates, such as ambient water temperature, ambient air temperature, UV radiation, etc. Behavioural data were analysed using a Poisson process to compare response variables among experiments because these data were not normally distributed due to the large number of zeros in the variables. Bonferroni multiple range tests were used to determine which response variable means differed (i.e. Pre-Heat, Heat, Post-Heat). Least square means (LS Means) were used to calculate means for a posteriori tests. To determine the directionality of movement (upstream or downstream), Chi-Square tests were used to test the independence of the upstream and downstream variables in the control and treatment sections. Al l tests were considered to be significant to a P < 0.05. Descriptive statistics (mean, ± 1 SE) were used to describe the temperature data and the side-channel Cortisol results due to lack of replicates. Temperature data from the two upstream reference data loggers (control) and the downstream data loggers (treatment) were pooled for the pre-heating phase, and categorized into upstream ( t\ ) and downstream ( /2 )• Temperature means calculated from the pre-heating phase from the upstream reference means were subtracted from the downstream sites. This served as a correction factor to ensure that any variation in temperature recorded during the heating phase was a function of the heating apparatus, and not due to differences in calibration. The mean 95 temperature from the upstream loggers during the heating phase ( t\ Heating) was subtracted from the mean temperature from the downstream loggers during the heating phase ( ^ Heating)- The correction factor calculated from the pre-heating data was subtracted from this value to determine the mean temperature increase during the heating phase: A H = ( ^ H e a t i n g - H e a t i n g ) - ( h Pre-Heating - t\ Pre-Heating ) RESULTS Temperature manipulations The tributary experiment resulted in a mean stream temperature increase (A H) of 1.12 °C (SE = 0.53). The corresponding mean ambient stream-temperature for the tributary trials was 10.73 °C (SE = 1.77). Mean stream temperature increase in tributary trials in 1998 was 1.67 °C (SE = 0.75). The corresponding mean ambient temperature was 12.72 °C (SE = 2.35). Mean temperature increase in side-channel trials was 3.09 °C (SE = 0.79). The corresponding mean ambient water temperature was 21.48 °C (SE = 2.25). The mean discharge during tributary trials was 0.0512 m3/s (SE = 0.0166), and 0.0077 m3/s (SE = 0.0007) in side-channel trials. Dissolved oxygen concentrations ranged from 8.3 mg/L to 10.4 mg/L in both experiments. Stress In the pooled 1997 and 1998 tributary heating trials I found no differences in mean Cortisol concentrations between un-netted (1.77 ± 0.26 ng/g) and netted (3.08 ± 0.45 ng/g) sections (ANOVA, F = 4.74, df = 1, 8; P = 0.0612). Cortisol concentrations differed between control (2.29 ± 0.33 ng/g) and treatment (4.07 ± 0.59 ng/g) sections (F = 16.84; df = 1,8; P = 0.0034). In the 1998 tributary heating trials I found no differences in mean Cortisol concentrations between un-96 netted and netted sections (F = 2.75, df = 1,2; P = 0.2390) or between control and treatment sections (F = 7.01; df = 1,4; P - 0.0571). The 98/08/01 side-channel trial resulted in a mean increase (AH) during the heating phase of 3.87 °C (SE = 0.09), to a mean ambient water temperature of 23.73 °C (SE = 0.04). The un-netted mean plasma Cortisol concentration was 54.08 ng/ml (SE =8.78). The netted, control and treatment mean plasma Cortisol concentrations were 314.35 ng/ml (SE = 54.28), 175.20 ng/ml (SE = 55.08), and 167.80 ng/ml (SE = 46.54), respectively. Some mortality of juvenile chinook occurred in the treatment section during heating. No mortality occurred in the control section. Absolute numbers of fish rearing in the treatment sections were not quantified, but a conservative estimate of observed dead fish is approximately 10%, which does not include latent stress induced mortality or other causes (e.g. predation or disease). The 98/08/15 side-channel trial had a mean increase in water temperature of 2.30 °C (SE = 0.07). The mean ambient water temperature was 19.23 °C (SE = 0.09). Pre-Control data were not collected due to a shortage of dry ice. Lethal sampling of juvenile chinook yielded the following mean concentrations of whole body Cortisol: netted 10.00 ng/g (SE = 1.67), control 18.34 ng/g (SE = 2.09), and treatment 12.43 ng/g (SE = 1.83). Behaviour Movement In tributary trials I found no differences in total number moving among time periods (Repeated Measures ANOVA, F = 1.46, df = 1,6; P = 0.2728). Total number moving differed between locations (F = 8.65, df = 1,6; P = 0.0259). There was no interaction between location and time (F = 0.56, df = 1,6; P = 0.4837). Total number moving in treatment sections (3.87 ± 1.46) was greater than control sections (2.38 ± 1.46) (Bonferroni, P = 0.0259) (Figure 4.2a). 9 7 In side-channel trials I found no differences in total number moving among times (Repeated Measures ANOVA, F = 6.16, df = 1,4; P = 0.0681) or between locations (F = 1.67, df-1,4; P = 0.2659). The interaction between location and time was significant (F = 9.41, df = 1,4; P = 0.0374). Total number moving increased over 64 times in the treatment sections, during the heating phases compared to the pre-heating phases (Bonferroni, P = 0.0172) (Figure 4.2b.). There was no difference in total number moving in the control sections (Bonferroni, P = 0.6998). In the treatment sections, direction of fish movements during the heating phases of side-channel trials was primarily upstream, compared to a tendency for downstream movement in the control sections. Approximately 68% of the juvenile chinook in the treatment sections (n = 140) moved upstream and 32% moved downstream as a result of the temperature manipulations, compared to 28% in the control sections (n = 64) moving upstream and 71.9% moving downstream ( x 2 = 28.061, df = 1, P = 0.001). Feeding In tributary trials I found differences in the mean frequency of feeding attempts per minute between locations (Poisson, x 2 = 15.45, df = 1, P < 0.0001) and among time periods (Poisson, x 2 = 9.95, df = 2, P = 0.0069). The interaction between location and time was significant (Poisson, X = 26.25, df = 2, P < 0.0001) (Figure 4.3a.). In control sections, feeding decreased significantly from pre-heating to heating phases ( x 2 = 12.05, df = 1, P = 0.0005), but did not differ between heating and post-heating phases ( x 2 = 1.06, df = 1, P = 0.3023), or pre-heating and post-heating phases ( x 2 = 3.09, df = 1, P = 0.0786). In treatment sections, feeding did not differ between pre-heating and heating phases ( x 2 = 0.16, df = 1, P = 0.6896), but increased two times from heating to post-heating phases ( x 2 = 14.93, df = 1, P = 0.0001), and also increased between pre-heating and post-heating phases ( x 2 = 13.84, df = 1, P = 0.0002) (Figure 4.3a.). 98 In side-channel trials I found no differences in mean feeding attempts per minute between locations (Poisson, % = 0.28, df = 1, P = 0.5992), but found differences among time periods (Poisson, x = 9.51, df = 2, P = 0.0086). The interaction between location and time was significant (Poisson, %2 = 13.36, df = 2, P = 0.0013). In control sections, similar to the tributary experiment, feeding decreased significantly from pre-heating to heating phases (%2 = 5.25, df = 1, P = 0.0219), and from pre-heating to post-heating phases ( x 2 = 4.56, df = 1, P = 0.0328). Feeding did not differ between heating and post-heating phases in control sections ( x 2 - 0.06, df = 1, P = 0.8122). In treatment sections, feeding did not differ between pre-heating and heating phases ( x 2 = 0.02, df = 1, P = 0.8943), but increased 56 times from the heating to post-heating phases ( x 2 = 10.34, df = 1, P = 0.0013). Feeding also increased between pre-heating and post-heating phases ( X 2 = 26.17, df = 1, P < 0.0001) (Figure 4.3b.). It is interesting to note that in all trials, the increase in the frequency of feeding did not occur simultaneously with increased temperature, but rather lagged behind temperature manipulations, occurring instead during the post-heating phase. Yawning In tributary trials I found differences in mean yawning per minute between locations * 2 2 (Poisson, x = 8.55, df = 1, P - 0.0035), but no differences among time periods (Poisson, x = 0.61, df = 2, P = 0.7368). There was no interaction between location and time (Poisson, x 2 = 0.26, df = 2, P = 0.8762). Yawning was significantly greater in control sections (0.061/minute +/-0.023) than treatment sections (0.014/minute +/- 0.008) ( x 2 = 6.52, df = 1, P = 0.0107). Aggression and Fleeing In tributary trials I found differences in mean aggression per minute between locations 2 y (Poisson, x = 12.05, df = 1, P = 0.0005), but no differences among time periods (Poisson, x = 99 2 1.73, df = 2, P = 0.4209). The interaction between location and time was significant (Poisson, x = 11.78, df = 2, P = 0.0028). In control sections, aggression decreased from pre-heating to heating phases ( x 2 = 8.33, df = 1, P = 0.0039), and increased from heating to post-heating phases (X2 = 11-61, df = 1, P = 0.0007), but did not differ between pre-heating and post-heating phases (X2 = 0.62, df = 1, P = 0.4323) (Figure 4.3c). In treatment sections, aggresion did not differ between pre-heating and heating phases ( x 2 = 0.54, df = 1, P = 0.4642), heating and post-heating phases ( x 2 = 2.30, df = 1, P = 0.1293), or pre-heating and post-heating phases ( x 2 = 0.85, df = 1, P = 0.3556). Aggression was greater in control sections during pre-heating phases ( x 2 = 8.28, df = 1, P = 0.0040), as well as during post-heating phases compared to treatment sections ( x -12.34, df = 1, P = 0.0004). Aggression did not differ between control and treatment sections during heating phases ( x 2 = 0.48, df = 1, P = 0.4872) (Figure 4.3c). In tributary trials, I found differences in mean fleeing per minute between locations (Poisson, x 2 = 12.98, df = 1, P = 0.0003), but there were no differences among time periods (Poisson, x 2 = 2.20, df = 2, P = 0.3337). The interaction between location and time was significant (Poisson, x 2 = 9.82, df = 2, P = 0.0074). In control sections, fleeing decreased from pre-heating to heating phases ( x 2 =10.57, df = 1, P = 0.0012), and increased from heating to post-heating phases ( x 2 =11.02, df = 1, P = 0.0009), but did not differ between pre-heating and post-heating ( x 2 = 0.06, df = 1, P = 0.8145) (Figure 4.3d.). In treatment sections, fleeing did not differ between pre-heating and heating phases ( x 2 = 0.52, df = 1, P = 0.4712), heating and post-heating phases ( x = 1.27, df = 1, P = 0.2598), or pre-heating and post-heating phases ( x = 0.25, df = 1, P = 0.6174). Fleeing was greater in control sections during pre-heating phases ( x 2 = 10.92, df= 100 1, P = 0.0010), and post-heating phases compared to treatment sections ( x 2 = 10.87, df = 1, P = 0.0010), but did not differ during heating phases ( x 2 = 0.26, df = 1, P = 0.6114) (Figure 4.3d.) Statistical analyses were not possible in side-channel trials because the models did not converge (one factor in the model had a variance of zero). In the treatment sections, mean aggression and fleeing per minute declined to zero during the heating and post-heating phases. In contrast, mean aggression and fleeing per minute had a natural tendency to increase during heating and post-heating phases in control sections. Fishmin and Total fish In tributary heating trials I found no differences in mean fishmin between locations (ANOVA, F = 0.24, df = 1,10, P = 0.6336), or among time periods (F = 0.23, df = 2,10, P = 0.7980). There was no interaction between location and time (F = 0.13, df = 2,10, P = 0.8815). I found no differences in total fish observed between locations (F = 3.22, df = 1,10, P = 0.1031), or among time periods (F = 0.71, df = 2,10, P = 0.5159). There was no interaction between location and time (F = 0.04, df = 2,10, P = 0.9590). In side-channel trials I found differences in mean fishmin between locations (ANOVA, F = 16.58, df = 1,5, P = 0.0096) and among time periods (F = 12.64, df = 2,5, P = 0.0111). The interaction between time and location was significant (F = 11.08, df = 2,5, P = 0.0146). In control sections, fishmin did not differ among any of the phases (Bonferonni, P = 1.00, for all contrasts). In treatment sections, fishmin did not differ between pre-heating and heating phases (Bonferonni, P = 0.0521), or heating and post-heating phases (Bonferonni, P = 1.00). Fishmin was greater during pre-heating compared to post-heating phases (Bonferonni, P = 0.0205) (Figure 4.3e.). Analyses of total fish observed in side-channel trials revealed no differences between locations (F 101 = 0.17, df = 1,5, P = 0.6963), or among time periods (F = 4.15, df = 2,5, P = 0.0865). There was no interaction between location and time (F = 4.59, df = 2,5, P = 0.0738). DISCUSSION Experimental field manipulations to tributary and side-channel temperatures resulted in a range of increased temperatures (0.97 - 3.87 °C) that are representative of post-forest harvesting temperature increases reported in the literature for Pacific Northwest streams. For example, water temperatures increased an average of 3 - 4 °C in response to riparian forest harvesting in southeast Alaska (Keith et al. 1998). In Carnation Creek, forest harvesting related temperature increases ranged from 0.7 - 3.25°C (Hartman et al. 1987), and in the northern interior of British Columbia, forest harvesting related temperature mean increases ranged from 1 - 3°C (Brownlee et al. 1988). Natural fluctuations in stream temperatures can also be quite rapid in this region. Daily temperature increases of 2°C per hour having been reported for the Thompson River, British Columbia (Walthers and Nener 1997). Therefore, the relatively abrupt elevation in stream temperatures during experimental manipulations in this thesis is also reasonably representative of conditions in nature. Stress Juvenile chinook rearing in tributaries with a mean ambient temperature of 10.73 °C, exhibited significant stress responses to experimental elevations in temperature of about 1 °C. The enclosure nets also appeared to elicit a stress response in juvenile chinook. However, in each experiment stress levels declined over time in the enclosure nets suggesting partial recovery. 102 This Cortisol response suggests that even small changes (1 °C) in stream temperatures are stressful to juvenile chinook, even if they do not exceed the preferred temperature range for the species. In fact, contrary to my prediction, increases in temperature to the preferred range were stressful. Subtle changes in stream temperature are not stressful to rainbow trout (Mellina 2002). Rainbow trout can tolerate a wider range of temperatures (Reiser and Bjornn 1979; Lee and Rinne 1980) which may explain this contradictory finding. The streams studied by Mellina et al. (2002) were lake-headed which likely moderated forest harvesting induced temperature changes and corresponding physiological responses by fish. Furthermore, these workers' (Mellina et al. 2002) comparative study design would not have excluded other factors that co-vary with streamside forest harvesting treatments (e.g. increased light) possibly mitigating stress responses. Ambient temperatures during side-channel trials were high and likely confounded the stress results. In the 98/08/15 side-channel trial, the netted Cortisol concentration was approximately 2.5 times greater than the treatment concentration from the pooled tributary trials, suggesting that the ambient temperature of 19 °C was stressful to juvenile chinook prior to initiation of this trial. The 98/08/01 side-channel trial was conducted during a mean ambient water temperature of 24 °C. Similar to the other side-channel trial, juvenile chinook captured upstream from the enclosure nets exhibited a stress response in the un-netted section, five times greater than typical basal rates reported for juvenile salmonids (Sumpter et al. 1986; Fagerlund et al. 1995), prior to the initiation of the trial. The netted fish exhibited even higher plasma Cortisol levels, elevated approximately six times from the un-netted phase, undoubtedly due to the combined stress invoked from high ambient temperatures and enclosure in the net sections. This fact confounded interpretation of all subsequent stress values. However, heating the water by an additional 3.87 103 °C, created further stress to juvenile chinook because mortality occurred to approximately 10% of the treatment population compared with no mortality in the control sections. The fact that small changes in temperature are stressful to juvenile chinook is disconcerting not only because forest harvesting can cause an increase in summer stream temperatures (Hartman and Scrivener 1990; Ringler and Hall 1975; Beschta et al. 1987; Holtby 1988; Brownlee et al. 1988; Hetrick et al. 1998; Rishel et al. 1982; Hall and Lantz 1969), but also because the increase in diurnal fluctuations associated with forest harvesting (Hartman and Scrivener 1990; Rishel et al. 1982; Ringler and Hall 1975; Beschta et al. 1987; Hetrick et al. 1998; Brownlee et al. 1988; Hall and Lantz 1969) may be additionally stressful. Considering that streamside forest harvesting can cause increases in diurnal temperature fluctuation, following forest harvesting salmonids could be subjected to amplified fluctuations in temperature, which my results suggest may cause chronic physiological stress. Indeed, fluctuations in diurnal temperatures have been documented to be stressful to juvenile coho, and negatively impact energy reserves (Thomas et al. 1986). A strong correlation between diel temperature fluctuation and summer mortality of juvenile coho has been shown in field studies (Martin etal. 1986). A suite of tertiary effects are induced in response to chronic elevations of Cortisol, including reductions in disease resistance, growth, reproduction, and tolerance of additional stressors (McLeay et al. 1987; Thomas 1990; Fagerlund et al. 1995). Chronic stress due to increased diurnal fluctuations in temperature can depress growth in salmonids (Beschta et al. 1987). Further, the physiological effects of chronic sub-lethal stresses, such as increased diurnal temperature fluctuations, are cumulative and result in a stepwise pattern of increased plasma Cortisol (Barton et al. 1986). Chronic sub-lethal stress can impact juvenile chinook at future life 104 history stages. For example, increased diurnal temperature fluctuations during the summer and fall can cause lipid depletion (Lemly 1996), which could compromise over-winter survival. Sub-lethal stress can impair the ability of juvenile salmonids to avoid predators (Sylvester 1972; Olla et al. 1992; 011a and Davis 1989) and it is well established that predators capture substandard prey disproportionately from prey populations (Mesa et al. 1994; Temple 1987). The mechanisms causing these phenomena may include decreased learning and sensory abilities resulting in lapses in decision-making (Olla et al. 1992) or decreased swimming performance (Mesa et al. 1994). Coutant (1973) described an abnormal posture exhibited by thermally shocked salmonids and suggested that this behaviour was responsible for elevated levels of predation on these fish because it increased the conspicuousness of them. Interestingly, I also observed this posture with fish oriented on an angle of 30° towards the surface in my side-channel trials. Thus, elevated temperatures, particularly diurnal fluctuations, and the coincident increased levels of Cortisol, can have serious consequences (e.g. potentially higher predation rates) for juvenile chinook even with only modest temperatures increases such as the ones examined in the present study. Movement and Behaviour Although the Cortisol results suggest that fish were stressed, contrary to my prediction tributary rearing juvenile chinook did not display an avoidance response to elevated temperatures, even when they exceeded 16°C. Therefore, although temperatures in some trials did exceed the preferred range for juvenile chinook (12 - 14°C; Brett 1952), these individuals tolerated this elevated temperature, much like the reluctance of juvenile chinook to evacuate tributaries during sediment pulses (Chapter 3). The fact that tributary rearing juvenile chinook found increased temperatures stressful, yet endured rather than emigrate when temperatures increased, underscores 105 my previous discussions regarding the preferred nature of tributary habitat (Chapter 2). Other researchers have also shown that ecological motivation can cause juvenile salmonids to temporarily override adaptive avoidance responses and endure sub-lethally stressful environmental conditions (Birtwell et al. 1999). Indeed, stream-type juvenile chinook have been reported to display considerable permanence of station (Edmundson et al. 1968; Rosberg et al. 1981; Bustard 1983), which has been suggested to confer many benefits including familiarity with cover and food sources (Chapman and Bjornn 1969), providing obvious adaptive benefits considering the relative value of tributary habitats compared to side-channel habitats (Chapter 2). Brett (1952) noted juvenile chinook avoided water that exceeded 15 °C, but his study was conducted in the laboratory, which would not capture the ecological reality and disparity in habitat value encountered by juvenile chinook rearing in side-channels and tributaries of the Torpy River. Tributaries to the Torpy River are the highest value habitat (Chapter 2); thus, motivation to remain may override the impulse to flee from acute temperature fluctuations. In side-channel trials, however, juvenile chinook exhibited a pronounced avoidance response to elevated stream temperatures, as predicted. This avoidance response was primarily in the upstream direction indicating active selection of upstream habitats approximately three degrees Celcius cooler than the heating manipulations. This is contrary to the literature that suggests fish have a natural tendency to move downstream when exposed to danger (Brown 1999). In fact, Keenleyside and Hoar (1954) reported increased rates of downstream migration for both juvenile chum and coho salmon as water temperature increased. The direction of juvenile chinook avoidance, in my research, is even more striking considering that the natural tendency of juvenile chinook in the control sections of side-channel trials was to move in the downstream direction. Upstream movements in search of favourable temperatures have been 106 reported in cutthroat trout (Brown and MacKay 1995) and for juvenile chinook in the Nechako River (Anonymous 1998; Russell et al. 1983). The experiments were conducted at the mid-point of the Lower Torpy River and upstream habitat opportunities for side-channel rearing individuals are ample. As temperatures are cooler upstream, movements in that direction would presumably have adaptive value. Other differences in movement behaviour were noted between side-channel and tributary individuals during the heating phases. Juvenile chinook behaved territorially in tributary trials, whereas in side-channels movement was in aggregations in treatment sections, huddling together in groups of 10 - 20, and moving in erratic, rapid darting movements, characteristic of stressed fish (Mason and Chapman 1965). Temperature-invoked changes in social behaviour have been noted in other salmonid species. Atlantic salmon abandon territories and form aggregations in response to high stream temperatures (>22 °C; Gibson 1988; Gibson and Cunjak 1986). Aggregative behaviour occurs in response to extremes in temperature as a mechanism to escape unfavourable stress (Brett 1970), and has been commonly reported in response to cold winter temperatures in salmonids (Brown 1999; Cunjak and Power 1986; Brown and MacKay 1995; Hartman 1965). The transition to gregarious behaviour in response to harsh winter conditions is gradual with declining temperatures (Brown 1999). However, metabolic rate likely plays a pivotal role in the rate of change to aggregative behaviour, because cold temperatures induce lethargy and high temperatures invoke frantic activity (Beitinger et al. 2000). In the side-channels, juvenile chinook formed aggregates and shifted to gregarious behaviour almost instantaneously with application of heating manipulations. During the heating phase of side-channel trials, juvenile chinook aggregations were always in association with some form of structure such as large woody debris. In control sections 107 aggregations and association with large woody debris did not occur. Higher temperatures are favourable to warmwater piscivores, and these predators will be more likely to be hungry, rather than satiated, because of the profound influence of higher temperatures on metabolism. Large woody debris may provide cover from predators for individuals that are stressed and potentially more vulnerable to predation. Predators are considerably more prevalent in side-channels compared to tributaries (Chapter 2) so aggregative behaviours may be good predator avoidance tactics providing safety in numbers, rapid transfer of information and confuse predators and/or reduce capture efficiency (Brown 1999). However, this explanation alone seems insufficient since aggregations were also observed in side-channel sediment trials (Chapter 3), when the perceived risk of predation would be reduced due to the cover provided by turbidity (Gregory 1993; Gregory and Northcote 1993). Thus, it seems plausible that aggregative behaviour is a response to stressful conditions, and during side-channel heating trials aggregations of fish selected microhabitats that provided the additional predator avoidance benefit of cover. Foraging appeared to increase in treatment sections in response to heating which was consistent with my prediction yet, in direct contrast to a tendency for foraging behaviour to decline in control sections of both experiments. Increased feeding in response to elevated temperatures has been reported by other researchers (Glova and Mason 1977). The strength of the signal was more pronounced in side-channel trials; the rate of feeding increased over 56 times during the post-heating phase compared to a doubling of the feeding rate in tributary trials. This difference is probably because of the greater increase in temperature during these trials, compared to tributaries and the higher ambient temperatures. Metabolic demand and ability for activity (such as feeding) increases almost exponentially with temperature (Brett 1995). In this thesis, juvenile chinook may have facultatively adjusted their foraging rate in an attempt to meet 108 heightened metabolic demands from elevated temperatures. Salmonids are opportunistic foragers, quite adept at adjusting behaviour to meet energy demands (Nislow et al. 1998). Undoubtedly juvenile chinook subjected to the high temperatures during side-channel heating trials would be allocating the majority of their energy to intense metabolic demands rather than somatic growth, since the preferred temperature range (12-14 °C, Brett 1952), where metabolic processes are optimized (Brett 1971), was far exceeded. Temperature increases of 4 °C have been shown to decrease production of juvenile chinook mass due to elevated metabolic requirements (Bisson and Davis 1976) and decreased food conversion efficiency (Reid et al. 1995). In the side-channel trials, temperatures were elevated well above the optimal range, thus even though feeding rates increased, juvenile chinook may lose weight in the long term and this could have negative consequences for fitness. The deposition of energy into lipid is considered one of the most important components to fish survival (Adams and Breck 1990). Thus, elevated temperatures and associated lipid depletion may have negative consequences to juvenile chinook over-winter survival. Increased activity in response to higher temperatures is a common response of salmonids (Gibson 1988; Beschta et al. 1987). The fishmin variable represents the mean equivalent number of fish viewed per minute. By definition, for this variable to decline as it did during side-channel trials, either less fish were observed, which would be indicated by a decline in frequency, and/or fish were observed for a shorter period of time, which would be indicated by an increase in movement. In side-channels high temperatures amplified total movement and caused territorial behaviour to cease altogether. Accordingly less fish were seen in the field of view as the trials progressed because they were moving upstream in aggregates to avoid the thermal stress, rather than remaining stationary, resulting in a decline in the fishmin variable. 109 Contrary to my predictions, the frequency of aggression, fleeing, and yawning did not increase in response to elevated temperatures in tributary or side-channel trials. In fact, in side-channel trials temperatures elevated up to 24.57 °C were so stressful to juvenile chinook that there was a breakdown in territorial behaviour altogether. A similar breakdown in social interactions has been documented for juvenile salmonids in response to pulses of suspended sediment (Berg and Northcote 1985), and crowding stress (Chapman and Bjorrn 1969). The rapid transition from territorial to gregarious behaviour occurred concurrently with application of the heating manipulation and was not observed in control sections. Aggregate behaviour is a response to extremely stressful conditions (Brett 1970), and has been described for low temperatures (Brown 1999; Cunjak and Power 1986; Brown and MacKay 1995; Hartman 1965), as well as for high temperatures (Gibson 1988; Coutant 1973; Gibson and Cunjak 1986). Environmental conditions influence the degree of social aggression in animals (Huntingford 1976), which for elevated temperature, was apparent in this research and others (Gibson 1988; Gibson and Cunjak 1986). When salmonid populations are limited by abiotic conditions, competition is minimal or non-existent (Hearn 1987). Indisputably, territory defense and avoidance in aggregations, are two behaviours that cannot co-occur. Typically, behavioural interactions, including aggression increase with higher temperatures (Glova and Mason 1977; Gibson 1988), therefore the cessation of social behaviour observed during side-channel trials adds further evidence that these juvenile chinook were under stress. The cessation of territorial behaviour (aggression and fleeing) in side-channel trials supports my earlier suggestion that increased movement during heating and post-heating phases is an avoidance response, and does not reflect the displacement of fish arising from unsuccessful social interactions (Chapman 1962). If the increase in movement during side-channel heating trials was a result of increased social 110 interactions and subsequent displacement through competitive exclusion, the rate of aggression would be expected to increase rather than cease. Forest Harvesting Implications Recent research in northern, interior British Columbia has indicated that thermal effects of forest harvesting on lake-headed streams with groundwater influence are largely benign (Mellina et al. 2002). In that study, streamside forest harvesting did not appear to cause acute or chronic stress in rainbow trout, possibly due to the combined moderating influence of groundwater and lakes (Mellina 2002). These findings suggest that the effects of forest harvesting are relatively subtle and potentially less severe in northern, interior British Columbia compared to coastal systems (Mellina 2002). However, rainbow trout have a much wider tolerance for temperature changes than juvenile chinook (Reiser and Bjornn 1979; Lee and Rinne 1980) so it is possible that some species like chinook salmon would be more affected. In the Torpy River, streamside forest harvesting of groundwater-controlled tributaries resulted in depressed winter temperatures and elevations of both mean and variance in summer temperatures, as well as decreases in growth of juvenile chinook (Shrimpton et al. 1999a,b). It appears that the presence of groundwater alone may be insufficient to mitigate changes to the thermal regime of small streams and prevent stress to fish in northern, interior British Columbia. This also suggests that the influence of lakes (sensu Mellina 2002) is paramount for moderating potential stressful impacts to salmonids caused by forest harvesting. Small changes in stream temperature (0.97 - 3.87 °C), well within the realm of those associated with streamside forest harvesting (Brownlee et al. 1988; Keith et al. 1998; Hartman et al. 1987), caused behavioural and physiological responses in juvenile chinook indicative of sub-lethal stress. Long term display of these responses could lead to decreased fitness and survival. 111 Juvenile chinook are the most tolerant of the Pacific salmon species to elevated temperatures (Brett 1952); thus other salmonid species may be even more susceptible to warming caused by streamside forest harvesting. Natural resource managers should afford the highest protection possible to the riparian zones of small streams to prevent the deleterious impacts resulting from subtle changes in thermal regimes. Elevated temperatures can have positive effects on juvenile salmonids such as increased growth (Hartman and Scrivener 1990; Slaney et al. 1977b; Brownlee et al. 1988). However, positive effects at one life stage have been shown to have negative and indirect latent consequences to fish production due to alteration in the timing of life history events (Hartman and Scrivener 1990; Thedinga and Koski 1984). Northern, interior chinook populations may be vulnerable to this type of affect, warranting further research in this area. 112 Study Sites Experimental Design Torpy River tributary sidechannel downstream camera Flow •*- Q - sideview camera stop net experimental manipulation jHJ downstream camera •JTIi sideview camera A B Figure 4.0. Diagram (A) representing an example of the macrohabitats (tributaries, side-channels) in which experiments were conducted. Schematic (B) of general experimental design (not to scale). The location of temperature manipulations is indicated by the asterik. 113 Pre-Heat phase Heat ing phase Post-Heat phase Minutes 1 1 I 1 1 1 1 1 1 ; I 90 minutes Figure 4.1. Temporal methods of behavioural data extraction from underwater video recordings. Each phase of the experiment was divided into 10 equally spaced intervals (e.g. 6 min. intervals in Pre-Heat, 9 min. intervals in Heating, 6 min. intervals in Post-Heat) and random samples were extracted from each interval. The magnification of the Heating phase illustrates the systematic random approach. Black bars are randomly selected intervals for data extraction 114 B Contro l Treatment • Pre-heat • Heat Figure 4.2. Mean total movement of juvenile chinook during tributary (A) and side-channel (B) heating trials. Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on three and two trials respectively. 115 Pre-Heating Heating Post-Heating [•Control • Treatment| Figure 4.3. Mean feeding attempts per minute during tributary trials (A), side-channel trials (B), mean aggression per minute during tributary trials (C), mean fleeing per minute during tributary' trials (D), and mean fishmin during side-channel trials (E). Error bars represent ± 1 SE. Bars with similar letters did not differ. Means are based on number of trials (sidechannels: n = 2; tributaries: n = 5). 116 CHAPTER 5: THESIS SUMMARY Juvenile chinook early life history studies that focus on only one season and/or one macrohabitat provide an incomplete picture as to the early life history of the stream-type race, particularly considering the magnitude of variation within a given life history strategy that is displayed by the species. In this thesis, the seasonal habitat preferences of juvenile chinook were investigated in three macrohabitats, and it was determined that non-natal tributaries were the highest value habitats in all seasons (spring, summer, fall/winter). Tributaries proffer thermal stability both during the summer and winter, and provide refuge from suspended sediment, high/low temperatures, and predators in the mainstem. Furthermore, individuals rearing in non-natal tributaries may have a survival advantage. Behavioural disposition for selection of these habitats has adaptive value during downstream migrations, demonstrable by the fact that non-natal tributaries to the Fraser River provide critical refuge habitat to juvenile chinook from high temperatures and suspended sediment levels (Scrivener et al. 1994). The high utilisation of tributaries in the Torpy River watershed is of concern from a habitat protection perspective because the majority of forest harvesting has concentrated almost exclusively in the low gradient floodplain area of the lower valley, which contains some of the most valuable juvenile fish habitat. The winter period is a particular sensitive time for juvenile chinook, which unfortunately is the season when the majority of forest harvesting takes place in the Upper Fraser River. The unique over-wintering behaviour documented in this thesis is contrary to the majority of studies in the literature that document "stream type" juvenile chinook to over-winter in the mainstem of large rivers (Anonymous 1987; Emmet 1989; Healey 1991; Levings and Lauzier 1991; Murray etal. 1981; Russell et al. 1983; Swales et al. 1986). In the Torpy River, juvenile 117 chinook either over-winter in groundwater controlled non-natal tributaries, actively feeding and exhibiting positive growth throughout the winter, or they emigrate from the mainstem in August. Juvenile chinook have been reported by others to display a similar exodus from the mainstem of large rivers in northern interior systems with the onset of the fall/winter season (Bjornn 1971; Everest and Chapman 1972; Chapman and Bjornn 1969; Hillman et al. 1987; Rosberg et al. 1981; Swales et al. 1986; Murray et al. 1981; Lister et al. 1981; Russell et al. 1983); however, extensive tributary over-wintering has not been documented before. Unlike the findings of other researchers (Hillman et al. 1987; Chapman and Bjornn 1969 ), the environmental cues for this seasonal habitat re-distribution of juvenile chinook in the Torpy River was not coincident with declining, cold temperatures. Temperatures during peak emigration in the Torpy River were > 15° C, well above the 4-8° C threshold reported by Hillman et al. (1987), as they are during summer in other Upper Fraser systems exhibiting an exodus of juvenile chinook (Lister et al. 1981, Russell et al. 1983). In the Torpy River, the primary trigger causing juvenile chinook to emigrate from the Torpy River mainstem is likely sedimentation of the gravel/cobble substrate, which is coincident with late summer low flows. Such proactive emigration, due to the degradation of over-wintering habitat in the mainstem, in anticipation of extreme cold temperatures yet months prior to the onset of those temperatures, is a remarkable product of evolution. Proactive emigration confers additive adaptive value, considering from a competitive perspective the earlier the emigration, the greater the probability of securing available over-wintering habitat elsewhere. Also, water temperatures (15 0 C) during this early emigration are optimal for juvenile chinook swimming ability, of obvious benefit for lengthy downstream migrations to suitable over-wintering habitat. It also appears that juvenile chinook that emigrate incur sustained growth throughout the winter through selection of ground water controlled, non-natal tributaries of the Fraser River, that 118 confer the same benefits as their non-natal rearing cohorts enjoy in the Torpy River tributaries. Thus, in the Upper Fraser River, juvenile chinook in the "emigrate" life history category may simply represent those individuals unable to secure stable groundwater controlled over-wintering habitat in their natal stream and consequently emigrate downstream in search of groundwater controlled non-natal over-wintering opportunities. Considering that juvenile chinook salmon are the least tolerant of cold temperatures (Brett 1952), and the positive correlation between greater size and cold tolerance, sustained winter growth via selection of groundwater controlled tributaries has clear adaptive value to the stream type race in northern interior systems subjected to harsh, cold winters. Indeed, utilisation of non-natal tributaries of the Fraser River by juvenile chinook for refuge habitat has been documented (Scrivener et al. 1994; Murray and Rosenau 1989) and this strategy may be significantly more prevalent and critical to survival of the stream type race than previously thought. The interconnected nature and compound effects of forest harvesting on the aquatic ecosystem, underscore the value in experimentally manipulating one factor in order to uncouple its effects from co-variates. In this thesis, conducting experimental manipulations to stream temperatures and suspended sediment concentrations in a natural un-logged stream, enabled the isolation of the other variables that often covary with these factors due to forest harvesting and/or climatic variation. Of all the Pacific salmon, chinook, are the most tolerant of high temperatures (Brett 1952) and suspended sediment concentrations (Servizi and Gordon 1990; Servizi and Martens 1991). In this thesis, relatively small changes in these variables, representative of post-forest harvesting conditions, caused sub-lethal stress responses; thus other salmonid species may be even more susceptible to warming and sedimentation caused by streamside forest harvesting. 119 The magnitude and type of response exhibited by juvenile chinook depended not only on ambient conditions, but also on the value of the habitat occupied. Avoidance was not a prominent part of the behavioural repertoire of tributary fish even though they appeared stressed in terms of elevated Cortisol concentrations, to experimental manipulations to temperatures and suspended sediment concentrations. Tributary rearing individuals suffered through obviously unfavourable conditions, choosing to endure rather than avoid, because of the gross disparity in value of tributaries relative to other habitat opportunities. Indeed, fish are generally able to tolerate a much higher degree of pollution in nature because of motivation (Sprague 1971) and ecological realism (Adams 1990), which are often lacking in laboratory studies. Other researchers have also shown that ecological motivation can cause juvenile salmonids to temporarily override adaptive avoidance responses and endure sub-lethally stressful environmental conditions (Birtwell et al. 1999). The potential for exposure to chronic sub-lethal stress is high for tributary rearing juvenile chinook because of their reluctance to evacuate preferred habitat. Multiple acute stressors have cumulative impacts on juvenile chinook salmon, resulting in a stepwise pattern of increased plasma Cortisol (Barton et al. 1986), causing additive performance deficits. Land use activities, such as forest harvesting, that result in multiple episodes of increased suspended sediment concentrations and temperatures, can be exceptionally damaging to tributary rearing juvenile chinook. Chronic stress can evoke many effects on juvenile salmonids including decreased growth and survival, and increased susceptibility to predation and disease, among others. Altered suspended sediment and thermal regimes can have latent, and potentially indirect, negative consequences to juvenile chinook through effects on life history behaviour. Life histories, particularly their timing and uniqueness, are a function of adaptation to local conditions. Therefore, any changes to these local conditions, induced by land use activities or otherwise can 120 reduce the future fitness of affected populations. For example, the annual decision to emigrate or over-winter in the Torpy River, including the timing thereof, is likely triggered by elevated sediment loads. Thus, sedimentation of clear, groundwater-controlled tributaries could place the perpetuation of the "non-natal" over-wintering strategy in peril. The Torpy River mainstem is an inhospitable habitat during the winter, so much so, that an exodus of mainstem rearing chinook occurs in the late summer. Displacement of tributary rearing chinook in the winter could have negative consequences, considering that the Pacific salmon species with the lowest cold tolerance (Brett 1952), is being displaced to a colder habitat. The temperature differential between groundwater controlled tributaries and the mainstem in the winter was approximately 1.5 °C. This delicate difference in thermal regimes, which provides refuge from harsh temperatures in the mainstem, and has likely fostered the evolution of the "non-natal" over-wintering strategy, may be susceptible to disruption from forest harvesting. A reduction in post-forest harvesting winter stream temperatures (Beschta et al. 1987; Anderson et al. 1997; Rishel et al. 1982; Brownlee et al. 1988), could render the "non-natal" over-wintering strategy non-adaptive, compromise the survival of populations selecting this strategy, and perhaps jeopardize the persistence of this life history trait. There is unequivocal evidence for global warming (Levy 1992; Linton et al. 1997, 1998; Schindler 1997, 2001) and this phenomenon may have far reaching implications to stream type juvenile chinook, such as the Torpy River population. Temperatures in the mainstem Torpy River already exceed 20° C, the point at which sub-lethal effects to juvenile chinook can occur, (Walthers and Nener 2000) and come precipitously close to the lethal limits in summer. In a global warming scenario of even a few degrees Celsius, the critical nature of non-natal groundwater controlled tributaries is even more crucial. Groundwater sources are predicted to 121 decline in response to global climate change (Schindler 1997), so rarity of preferred habitat in the Torpy watershed will likely become more pronounced. Phenotypic diversity, as illustrated by the juvenile chinook rearing in the Torpy River, is the outcome of the interaction between an individual's genetic underpinning with its unique environment (Healey and Prince 1995). Therefore, protection of these unique, and likely increasingly rare, groundwater controlled tributaries is critical to conservation. 122 Literature Cited Adams, S.M. 1990. Biological indicators of stress in fish. American Fisheries Society Symposium 8. Adams, S.M., and J.E. Breck. 1990. Bioenergetics. Pages 389-415 in C B . Schreck and P.B. Moyle, (eds.). Methods for fish biology. American Fisheries Society, Bethesda, Maryland. Allen, K.R. 1969. Limitation on production in salmonid populations in streams. In T.G. Northcote (ed.). Symposium on Salmon and Trout in Streams, p 3-20. H.R. MacMillan Lectures in Fisheries, University of British Columbia, Vancouver, B.C. Anderson, P.G., B.R. Taylor, and G.C. Balch. 1996. Quantifying the effects of sediment release on fish and their habitats. Can. MS. Rep. Fish. Aquat. Sci. No. 2346. Anderson, B.C., J.S. Macdonald, L.B. MacDonald. 1997. Temperatures in the Nadina River and Slim Creek watersheds, 1993-1996. Can. Dat. Rep. of Fish, and Aquat. Sci. 1017. pp. 102. Anonymous. 1987. Studies of Juvenile Chinook Salmon in the Nechako River, British Columbia 1985 and 1986. Can. MS. Rep. Fish. Aquat. Sci. No. 1954: 152 p. Anonymous. 1990. Overwintering of juvenile chinook in the Nechako River. Nechako Fisheries Conservation Program, NFCP-056. pp. 47. Anonymous. 1998. Size distribution and abundance of juvenile chinook salmon of the Nechako River, 1994. Nechako Fisheries Conservation Program. M94-3. pp. 26. Barton, B.A., and L.T. Toth. 1980. Physiological stress in fish: A literature review with emphasis on blood Cortisol dynamics. Fisheries Research Report No. 21. Alberta Department of Energy and Natural Resources, Calgary, Alberta, pp 18. Barton, B.A., C B . Schreck, and L.A. Sigismondi. 1986. Multiple acute disturbances evoke cumulative physiological stress responses in juvenile chinook salmon. Trans. Am. Fish. Soc. 115: 245-251. 123 Beckman, B.R., D.A. Larsen, B.Lee-Pawlak, and W.W. Dickhoff. 1998. Relation of fish size and growth rate to migration of spring chinook salmon smolts. N . Am. J. Fish. Manage. 18:537-546. Beitinger, T., W. Bennet, and R. McCauley. 2000. Temperature tolerances of North American freshwater fishes exposed to dynamic changes in temperature. Environmental Biology of Fishes 58: 237-275. Berg, L. and T.G. Northcote. 1985. Changes in territorial, gill-flaring, and feeding behavior in juvenile coho salmon (Oncorhynchus kisutch) following short-term pulses of suspended sediment. Can. J. Fish. Aquat. Sci. 42:1410-1417. Beschta, R.L., R.E. Bilby, G.W. Brown, L.B. Holtby, and T.D. Hofstra. 1987. Stream temperature and aquatic habitat: fisheries and forestry interactions. In Streamside management: forestry and fishery interactions. E.O. Salo and T.W. Cundy. (eds.) College of Forest Resources, University of Washington, Seattle, Wash. pp. 191-232. Bilby, R.E., and P.A. Bisson. 1992. Allochthonous versus autochthonous organic matter contributions to the trophic support of fish populations in clearcut and old-growth forested streams. Can. J. Fish. Aquat. Sci. 49:540-551. Bilton, H.T. 1978. Returns of adult coho salmon in relation to mean size and time at release of juveniles. Fisheries and Marine Service Technical Report No. 832., Pacific Biological Station, Nanaimo, British Columbia. Bilton, H.T. 1984. Returns of chinook salmon in relation to juvenile size at release. Can. Tech. Rep. Fish. Aquat. Sci. 1245. Bilton, H.T., D.F. Alderdice, and J.T, Schnute. 1982. Influence of time and size at release of juvenile coho salmon (Oncorhynchus kisutch) on returns at maturity. Can. J. Fish. Aquat. Sci. 39: 426-447. 124 Birtwell, LK. 1999. The effects of sediment on fish and their habitat. Canadian Stock Assessment Secretariat Research Document 99/139. Birtwell, I.K., J.S. Korstrom, R.P. Fink, J.A. Tanaka, D.I. Tiessen, and B.J. Fink. 1999. The effects of thermal change on juvenile chum salmon behavior, distribution and feeding: laboratory and field studies. In: Hawkins, S., C. Kondzela, R. Wilmot, C. Guthrie, J. Pohl, and H. Nguyen (eds.) Proceeding of the 19th Northeast Pacific Pink and Chum salmon workshop. March 3-5, 1999. Juneau, Alaska, pp. 87-92. Bisson, P. A. and R.E. Bilby. 1982. Avoidance of Suspended Sediment by Juvenile Coho Salmon. North Amer. J. Fish. Management 4: 371-374. Bisson, P. A. and G.E. Davis. 1976. Production of juvenile chinook salmon, (Oncorhynchus tshawytscha), in a heated stream. NOAA Fishery Bull. 74:763-774. Bjornn, T.C. 1971. Trout and salmon movements in two Idaho streams as related to temperature, food, stream flow, cover, and population density. Trans. Amer. Fish. Soc. 100: 423-438. Bloom, A.M. 1976. Evaluation of minnow traps for estimating populations of juvenile coho salmon and Dolly Varden. Prog. Fish. Cult. 38: 99-101. Bradford, M.J., J.A. Grout, and S. Moodie. 2001. Ecology of juvenile chinook salmon in a small non-natao stream of the Yukon River drainage and the role of ice conditions on their distribution and survival. Can. J. Zool. 79: 2043-2054. Bradford, M.J., and P.S. Higgins. 2001. Habitat-, season-, and size-specific variation in diel activity patterns of juvenile chinook salmon (Oncorhynchus tshawytscha) and steelhead trout (Oncorhynchus mykiss). Can. J. Fish. Aquat. Sci. 58:365-374. Bradford, M.J., and G.C. Taylor. 1997. Individual variation in dispersal behaviour of newly emerged chinook salmon (Oncorhynchus tshawytscha) from the Upper Fraser River, British Columbia. Can. J. Fish. Aquat. Sci. 54: 1585-1592. Brett, J.R., 1952. Temperature tolerance in young Pacific salmon, genus Oncorhynchus. J. Fish. Res. Board Can. 9: 265-323. Brett, J.R., 1970. Temperature - Fishes. In O. Kinne, (ed.) Marine Ecology, Vol . 1. Environmental Factors, Part 1, Wiley-Interscience, London, pp. 513-560. Brett, J.R., 1971. Energetic responses of salmon to temperature. A study of some thermal relations in the physiology and freshwater ecology of sockeye salmon {Oncorhynchus nerkci). Am. Zoologist 11,99-113. Brett, J.R., 1995. Energetics, pp.1-68. In C. Groot, L. Margolis, and W.C. Clarke, (eds.) Physiolodical Ecology of Pacific Salmon. U B C Press, Vancouver. Brown, G.W. and J.T. Krygier. 1970. Effects of clearcutting on stream temperature. Wat. Resour. Res. 6:1133-1140. Brown, R.S. 1999. Fall and early winter movements of cutthroat trout, Oncorhynchus clarki, in relation to water temperature and ice conditions in Dutch Creek, Alberta. Envt. Biol, of Fishes 55:359-368. Brown, R.S., and W.C. MacKay. 1995. Fall and winter movements of and habitat use by cutthroat trout in the Ram River, Alberta. Trans. Amer. Fish. Soc. 124: 873-885. Brown, T.G. and T.McMahon. 1988. Winter ecology of juvenile coho salmon in Carnation Creek: Summary of findings and management implications, p. 108-117. In T.W. Chamberlin (ed.) Proceedings of the Workshop: Applying 15 years of Carnation Creek results. Pacific Biological Station, Nanaimo, British Columbia. 239 p. Brownlee, M.J. , B .G. Shepherd, and D.R. Bustard. 1988. Some effects of forest harvesting on water quality in the Slim Creek watershed in the central interior of British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 1613: 41 p. 126 Bryant, D., D. Nielsen, and L. Tangley. 1997. Last frontier forests: ecosystems and economies on the edge. WRI Publications, Baltimore, Md. Bryant, M.D. 2000. Estimating fish populations by removal methods with minnow traps in southeast Alaska streams. North American Journal of Fisheries Management. 20: 923-930. Bustard, D.R. 1983. Juvenile salmonid winter ecology in a northern British Columbia River - A new perspective. American Fisheries Society (NPI Chapter), Bellingham, Washington, pp. 11. Bustard, D.R. 1986. Some differences between coastal and interior stream ecosystems and the implications to juvenile fish production, p. 117-126. In J.H. Patterson (ed.). Proceedings of the workshop on habitat improvements, Whistler, BC, May 8-10, 1984. Can. Tech. Rep. Fish. Aquat. Sci. 1483: 219 p. Bustard, D.R. and D.W. Narver. 1975. Aspects of winter ecology of juvenile coho salmon (Oncorhynchus kisutch) and steelhead trout (Salmo gairdneri). J. Fish. Res. Board Can. 32:667-680. Carl, L .M. and M.C. Healey. 1984. Differences in enzyme frequency and body morphology among three juvenile life history types of chinook salmon (Oncorhynchus tshawytscha) in the Nanaimo River, British Columbia. Can. J. Fish. Aquat. Sci. 41: 1070-1077. Cederholm, C.J., and W.J. Scarlett. 1981. Seasonal immigrations of juvenile salmonids into four small tributaries of the Clearwater River, Washington, 1977-1981. In: E.L. Brannon and E. Salo, (eds.). Proceedings of the salmon and trout migratory symposium. Univ. of Washington, School of Fisheries, Seattle: pp. 98-110. Chapman, D.W. 1962. Aggressive behaviour in juvenile coho salmon as a cause of emigration. J. Fish. Res. Board Can. 19:1047-1080. 127 Chapman, D.W. 1966. Food and space as regulators of salmonid populations in streams. Amer. Nat. 100:345-357. Chapman, D.W. 1988. Critical Review of the Variables Used to Define Effects of Fines in Redds of Large Salmonids. Trans. Amer. Fish. Soc. 117:1-21. Chapman, D.W. and T.C. Bjornn. 1969. Distribution of salmonids in streams with special reference to food and feeding, p. 153-176. In: T.G. Northcote (ed.) Symposium on salmon and trout in streams. H.R. MacMillan Lectures in Fisheries, University of British Columbia, Vancouver, B.C. Clarke, W . C , Withler, R.E., and J.E. Shelbourn. 1992. Genetic control of juvenile life history pattern in chinook salmon (Oncorhynchus tshawytscha). Can. J. Fish. Aquat. Sci. 49:2300-2306. Clarke, W . C , Withler, R.E., and J.E. Shelbourn. 1994. Inheritance of smolting phenotypes in backcrosses of hybrid stream type x ocean-type chinook salmon (Oncorhynchus tshawytscha). Estuaries 17:13-25. Cordone, A.J . , and D.W. Kelly. 1961. The influences of inorganic sediment on the aquatic life of streams. California Fish and Game 47:189-228. Coutant, C.C. 1973. Effect of Thermal Shock on vulnerability of juvenile salmonids to predation. J. Fish. Res. Board Can. 30:965-973. Culp, J.M., F.J. Wrona, and R.W. Davies. 1986. Response of stream benthos and drift to fine sediment deposition versus transport. Can. J. Zool. 64: 1345-1351. Cunjak, R.A. 1988. Physiological consequences of overwintering in streams: the cost of acclimatization? Can. J. Fish. Aquat. Sci. 45:443-452. Cunjak, R.A. and G. Power. 1986. Seasonal changes in the physiology of brook trout, (Salvelinus fontinalis) (Mitchell), in a sub-Arctic river system. J. Fish. Biol. 29:279-288. 128 Cunjak, R.A., R.A. Curry, and G. Power. 1987. Seasonal energy budget of brook trout in streams: implications of a possible deficit in early winter. Trans. Amer. Fish. Soc. 116:817-828. Dill, L .M. 1978. An energy-based model of optimal feeding-territory size. Theoretical Population Biology, 14: 396-429. Downing, W.L. (ed.). 1980. Proceedings, National conference on urban erosion and sediment control: institutions & technology. U.S. Environmental Protection Agency, EPA Report 905/9-80-002, Washington, DC. Edmundson, E.,F.H..Everest, and D.W. Chapman. 1968. Permanence of station in juvenile chinook salmon and steelhead trout. J. Fish. Res. Bd. Canada 25: 1453-1464. Emmett, B. 1989. Overwintering of juvenile chinook in the Nechako River: 1988-1989 studies. Report prepared by Archipelago Marine Research Ltd., Victoria, B.C., November 1989. Available from Department of Fisheries and Oceans, 401 Burrard Street, Vancouver, BC. European Inland Fisheries Advisory Commission (EIFAC). 1964. Water quality criteria for European freshwater fish. Report on Finely Divided Solids and Inland Fisheries. European Inland Fisheries Advisory Commission. Food and Agriculture Organization of the United Nations. Rome. EIFAC/1.21p. Everest, F.H., and D.W. Chapman. 1972. Habitat selection and spatial interaction by juvenile chinook salmon and steelhead trout in two Idaho streams. J. Fish. Res. Board Can. 29: 91-100. Everest, F.H., R.L. Beschta, J.C. Scrivener, K.V. Koski, J.R. Sedell, and C.J. Cederholm. 1987. Fine sediment and salmonid production: a paradox. USDA Forest Service, Corvallis, OR. 104p. 129 Fagerlund, U.H.M., McBride, J.R., and Williams, I.V. 1995. Stress and Tolerance, pp. 461-503. In: C. Groot, L. Margolis, and W.C. Clarke, (eds.). Physiological ecology of Pacific salmon. UBC Press, Vancouver. Fraser, F.J., P.J. Starr, and A.Y. Fedorenko. 1982. A review of the chinook and coho salmon of the Fraser River. Can. Tech. Rep. Fish. Aquat. Sci. No. 1126. Filbert, R.B. and C P . Hawkins. 1995. Variation in Condition of Rainbow trout in relation to food, temperature, and individual length in the Green River, Utah. Trans. Am. Fish. Soc. 124:824-835. Gammon, J.R. 1970. The effect of inorganic sediment on stream biota. Water Pollution Control Research Series. Environmental Protection Agency, Water Quality Office, Washington, DC. 142 p. Gibson, J.R. 1988. Mechanisms regulating species composition, population structure, and production of stream salmonids; a review. Pol. Arch. Hydrobiol. 35:469-495. Gibson, J.R. and R.A. Cunjak. 1986. An investigation of competitive interactions between brown trout (Salmo trutta L.) and juvenile Atlantic salmon (Salmo salar L.) in rivers of the Avalon Peninsula, Newfoundland. Can. Tech. Rep. Fish. Aquat. Sci. 1462, pp. 82. Ginetz, R.M. and P. A. Larkin. 1976. Factors affecting rainbow trout (Salmo gairdneri) predation on migrant fry of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board Can. 33: 19-24. Glova, G.J. and J.C. Mason. 1977. Interactions for food and space between sympatric populations of juvenile coho salmon and coastal cutthroat trout in a stream simulator during winter and spring. Fish. Mar. Serv. MS Rep. 1429: 31 p. Gradall, K.S., and W.A. Swenson. 1982. Responses of brook trout and creek chubs to turbidity. Trans. Amer. Fish. Soc. 111:392-395. Gregory, R. S. 1993. Effect of turbidity on the predator avoidance behaviour of juvenile chinook salmon (Oncorhynchus tshawytscha). Can. J. Fish. Aquat. Sci. 50:241-246. Gregory, R. S. and C. D. Levings. 1998. Turbidity reduces predation on migrating juvenile Pacific salmon. Trans. Amer. Fish. Soc. 127:275-285 Gregory, R.S. 1994. The influence of ontogeny, perceived risk of predation, and visual ability on the foraging behavior of juvenile chinook salmon. In: Stouder D.J., K.L. Fresh, and R.F. Feller, (eds.) Theory and application in fish feeding ecology. Belle W. Baruch Libr. Mar. Sci. No. 18. University of South Carolina Press, Charleston, S.C. pp. 271-284. Gregory, R.S. and J.A. Servizi and D.Martens. 1993. Comment: Utility of the stress index for predicting suspended sediment effects. North American Journal of Fisheries Management 13:868-873. Gregory, R.S. and T.G. Northcote. 1993. Surface, planktonic, and benthic foraging by juvenile chinook salmon (Oncorhynchus tshawytscha) in turbid laboratory conditions. Can. J. Fish. Aquat. Sci. 50:233-240. Hall, J.D., and R.L. Lantz. 1969. Effects of logging on the habitat of coho salmon and cutthroat trout in coastal streams, p. 355-375. In: T.G. Northcote (ed.), Symposium on Salmon and Trout in Streams. H.R. MacMillan Lectures in Fisheries. University of British Columbia, Vancouver. Hansmann, E.W. 1969. The effects of logging on periphyton communities of coastal streams. Ph.D. thesis. Oregon State University. 89 p. Hartman, G.F. 1963. Observations of behavior of juvenile brown trout in a stream aquarium during winter and spring. J. Fish. Res. Board. Can. 20:769-787. 131 Hartman, G.F. 1965. The role of behaviour in the ecology and interaction of underyearling coho salmon {Oncorhynchus kisutch) and steelhead trout (Salmo gairdneri) J. Fish. Res. Board Can. 22:1035-1081. Hartman, G.F., B.C. Andersen, and J.C. Scrivener. 1982. Seaward movement of coho salmon (Oncorhynchus kisutch) fry in Carnation Creek, an unstable coastal stream in British Columbia. Can. J. Fish. Aquat. Sci. 39: 588-597. Hartman, G.F., L .B . Holtby, and J.C. Scrivener. 1984. Some effects of natural and logging-related winter stream temperature changes on the early life history of coho salmon (Oncorhynchus kisutch) in Carnation Creek, British Columbia. In: W.R. Meehan, T.R. Merrell, Jr., and T.A. Hanley (eds.), Proceedings of Fish and Wildlife Relationships in Old-Growth Forests, American Institute of Fisheries Research Biologists' Symposium, April 1982, Juneau, Alaska. Bookmasters, Ashland, OH. p. 141-149. Hartman, G.F. and L .B. Holtby. 1982. An overview of some biophysical determinants of fish production and fish population responses to logging in Carnation Creek, British Columbia, p. 348-374. In G.F. Hartman (ed.). Proceedings of the Carnation Creek Workshop, a 10 year review. Pacific Biological Station, Nanaimo, British Columbia. 404p. Hartman, G.F. and J.C. Scrivener. 1990. Impacts of forestry practices on a coastal stream ecosystem, Carnation Creek, British Columbia. Can. Bull. Fish. Aquat. Sci. 223: 148 p. Hartman, G.F., J.C. Scrivener, L.B.Holtby, and L.Powell. 1987. Some effects of different streamside treatments on physical conditions and fish population processes in Carnation Creek, a coastal rain forest stream in British Columbia. In: E.O. Salo and T.W. Cundy (eds.) Streamside management: forestry and fishery interactions. University of Washington, Seattle, pp. 330-372. 132 Hartman, G.F., J.C. Scrivener, and M.J. Miles. 1996. Impacts of logging in Carnation Creek, a high energy coastal stream in British Columbia, and their implication for restoring fish habitat. Can. J. Fish. Aquat. Sci. 53: 237-251. Harvey, B.C., R.J. Nakamoto, and J.L. White. 1999. Influence of large woody debris and a bankfull flood on movement of adult resident coastal cutthroat trout (Oncorhynchus clarki) during fall and winter. Can. J. Fish. Aquat. Sci. 56:2161-2166. Healey, M.C. 1991. Life history of chinook salmon, pp. 311-393. In: C. Groot and L. Margolis (eds.), Pacific Salmon Life Histories. UBC Press, Vancouver. Healey, M.C. and F.P. Jordan. 1982. Observations on juvenile chum and chinook and spawning chinook in the Nanaimo River, British Columbia. Can. Manuscr. Rep. Fish. Aquat. Sci. 1659:35 pp. Healey, M.C. and U. Reinhardt. 1995. Predator avoidance in naive and experienced juvenile chinook and coho salmon. Can. J. Fish. Aquat. Sci. 52: 614-622. Healey, M . C , and A. Prince. 1995. Scales of Variation in Life History Tactics of Pacific Salmon and the Conservation of Phenotype and Genotype. American Fisheries Society Symposium 17: 176-184. Hearn, W.E. 1987. Interspecific Competition and Habitat Segregation Among Stream-Dwelling Trout and Salmon: A Review. Fisheries. Vol. 12, No. 5, 24-31. Heifetz, J.,M.L. Murphy, and K.V. Koski. 1986. Effects of logging on winter habitat of juvenile salmonids in Alaskan streams. North Amer. J. Fish. Management. 6:52-58. Herbert, D.W.M., and J.C. Merkens. 1961. The effect of suspended mineral solids on the survival of trout. Int. J. Air Water Pollut.5:56-75. Hetrick, N.J., M.A. Brusven, W.R. Meehan, and T.C. Bjornn. 1998a. Changes in solar input, water temperature, periphyton accumulation, and allochthonous input and storage after canopy removal along two small salmon streams in southeast Alaska. Trans. Amer. Fish. Soc. 127:859-875. Hetrick, N.J., M.A. Brusven, T.C. Bjornn, R.N. Keith, and W.R. Meehan. 1998b. Effects of canopy removal on invertebrates and diet of juvenile coho salmon in a small stream in southeast Alaska. Trans. Amer. Fish. Soc. 127:876-888. Hillman, T.W., J.S. Griffith. And W.S. Platts. 1987. Summer and winter habitat selection by juvenile chinook salmon in a highly sedimented Idaho stream. Trans. Amer. Fish. Soc. 116(2):185-195. Holtby, L.B. 1988. Effects of logging on stream temperatures in Carnation Creek, British Columbia and associated impacts on coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 45: 502-515. Holtby, L.B. 1987. The effects of logging on stream temperatures in Carnation Creek. In T.W. Chamberlain (ed.), Applying 15 years of Carnation Creek Results. Proceedings of workshop held in January 1987 in Nanaimo, B.C. Carnation Creek Steering Committee, Pacific Biological Station, p. 118-122. Holtby, L.B., and CP . Newcombe. 1982. A preliminary analysis of logging-related temperature changes in Carnation Creek, British Columbia. In: G.Hartman (ed.) Proceedings of the Carnation Creek Workshop, A 10-Year Review, Pacific Biological Station, Nanaimo, BC. Holtby, L.B., and G.F. Hartman. 1982. The population dynamics of coho salmon (Oncorhynchus kisutch) in a west coast rain forest stream subjected to logging. Pp 308-347, In: G. Hartman (ed.) Proceedings of the Carnation Creek workshop, a 10-year review. Pacific Biological Station, Nanaimo, BC. Huntingford, F.A. 1976. The relationship between inter- and intra-specific aggression. Anim. Behav. 24:485-497. 134 Irvine, J.R. 1986. Effects of varying discharge on the downstream movement of salmon fry, Oncorhynchus tshawytscha Walbaum. J. Fish. Biol. 28: 17-28. Kaeding, L.R. 1996. Summer use of coolwater tributaries of a geothermally heated stream by rainbow and brown trout, Oncorhynchus mykiss and Salmo trutta. Am. Midi. Natl. 135:283-292. Kalleberg, H. 1958. Observations in a stream tank of territoriality and competition in juvenile salmon and trout (Salmo solar and S. trutta). Inst. Freshw. Res. Drottningholm 39: 55-98. Keefe, M. , D.J. Anderson, R.W. Carmichael, and B.C. Jonasson. 1995. Early life history study of Grande Ronde River basin chinook salmon. U.S. Department of Energy. Portland, Oregon, pp. 39. Keenleyside, M.H.A., and W.S. Hoar. 1954. Effects of temperature on the response of young salmon to water currents. Behavior 7:77-87. Keith, R.M., T.C. Bjornn, W.R. Meehan, N.J. Hetrick, and M.A. Brusven. 1998. Response of juvenile salmonids to riparian and instream cover modifications in small streams flowing through second-growth forests of southeast Alaska. Trans. Amer. Fish. Soc. 127:889-907. Korstrom, J.S., R.P. Fink, J.A. Tanaka, D.I. Tiessen, B.J. Fink, and I.K. Birtwell. 1998. The influence of lethal temperature and food on the behavior of juvenile chum salmon (Oncorhynchus keta) under simulated marine conditions. Can. Data Rep. Fish. Aquat. Sci. 1040: 70 p. Kruzynski, G.M. and I.K. Birtwell. 1993. A predation bioassay to quantify the ecological significance of sublethal responses of juvenile chinook salmon (Oncorhynchus tshawytscha) to the Antisapstain fungicide TCMTB. Can. J. Fish. Aquat. Sci. 51:1780-1790. 135 Langer, O.E. 1980. Effects of sedimentation on salmonid stream life. In K. Weagle, editor. Report on the technical workshop on suspended solids and the aquatic environment. Department of Indian Affairs and Northern Development (Contract Ott-80-019), Whitehorse, Canada. Lee, R.M. and J.N. Rinne. 1980. Critical thermal maxima of five trout species in the southwestern United States. Trans. Am. Fish. Soc. 109:632-635. Lemly, D.A. 1996. Winter Stress Syndrome: An important consideration for hazard assessment of aquatic pollutants. Ecotoxicology and Environmental Safety. 34:223-227. Levings, CD. , and R.B. Lauzier. 1991. Extensive use of the Fraser River basin as winter habitat by juvenile chinook salmon (Oncorhynchus tshawytscha). Can. J. Zool. 69: 1759-1767. Levy, D.A. 1992. Potential impacts of global warming on salmon production in the Fraser watershed. Can. Tech. Rep. Fish. Aquat. Sci. 1889: 96 p. Levy, D.A., T.G. Northcote, and G.J. Birch. 1979. Juvenile salmon utilization of tidal channels in the Fraser River Estuary, British Columbia, Tech. Rep. Westwater Res. Centre. University of British Columbia, 23: pp. 81. Linton, T.K., S.D. Reid, C M . Wood. 1997. The metabolic costs and physiological consequences to juvenile rainbow trout of a simulated summer warming scenario in the presence and absence of sublethal ammonia. Trans. Am. Fish. Soc. 126:259-272. Linton, T.K., I.J. Morgan, S.D. Reid, and C M . Wood. 1998a. Long-term exposure to small temperature increase and sublethal ammonia in hardwater acclimated rainbow trout: does acclimation occur? Aquatic Toxicology 40: 171-191. Linton, T.K., S.D. Reid, and C M . Wood. 1998b. The metabolic costs and physiological consequences to juvenile rainbow trout of a simulated winter warming scenario in the presence and absence of sublethal ammonia. Trans. Am. Fish. Soc. 127: 611-619. 136 Linton, T.K., S.D. Reid, and C M . Wood. 1999. Effects of restricted ration on the growth and energetics of juvenile rainbow trout exposed to a summer of simulated warming and sublethal ammonia. Trans. Am. Fish. Soc. 128:758-763. Lister, D.B., I. Wallace, and D.G. Hickey. 1981. Salmonid Enhancement Baseline Investigations at Stuart River, British Columbia. D.S.S. Contract No. 07SB.FP501-9-1288. pp. 65. Lloyd, D.S. 1987. Turbidity as a water quality standard for salmonid habitats in Alaska. North Amer. J. FishManagement. 7:18-33. Lloyd, D.S., J.P. Koenings, and J.D. LaPerriere. 1987. Effects of turbidity in fresh waters of Alaska. North Amer. J. Fish. Management. 7:18-33. Lowe, R.L., S.W. Golladay, and J.R. Webster. 1986. Periphyton response to nutrient manipulation in streams draining clearcut and forested watersheds. J. N. Amer. Benthol. Soc. 5(3): 221-229. Macdonal, J.S., E.A. Maclsaac and H.E. Herunter. In press. The effect of variable retention riparian buffers on water temperatures in small headwater streams in boreal ecosystems of British Columbia. Can. J. For. Res. Accepted Dec. 2002. Mac Donald, L.B., F.N. Leone, and D.E. Rowland. 1996. Salmon Watershed Planning Profiles for the Fraser River Basin within the Prince George Land and Resource Management Plan. Prepared for Department of Fisheries and Oceans Fraser River Action Plan. pp. 153. Martin, D.J., L.J. Wasserman, and V.H. Dale. 1986. Influence of riparian vegetation on post-eruption survival of coho salmon fingerlings on the west-side streams of Mount St. Helens, Washington. N . Am. J. Fish. Manag. (1)6: 1-8. Mason, J.C. 1976. Response of underyearling coho salmon to supplemental feeding in a natural stream. J. Wildl. Manage. 40: 775-788. 137 Mason, J . C , and D . W . Chapman. 1965. Significance o f early emergence, environmental rearing capacity, and behavioural ecology of juvenile coho salmon in stream channels. J. Fish. Res. Board Can. 22:173-190. McLeay, D.J . , Birtwell , I .K., G.F. Hartman and G . L . Ennis. 1987. Responses of Arctic Grayling (Thymallus arcticus) to acute and prolonged exposure to Yukon Placer mining sediment. Can. J. Fish. Aquat. Sci. 44:658-673. Meehan, W.R. , M . A . Brusven, and J.F. Ward. 1987. Effects o f artificial shading on distribution and abundance of juvenile chinook salmon (Oncorhynchus tshawytscha). Great Basin Naturalist, 47(1):22-31. Meehan, W.R. (ed.). 1991. Influences of Forest and Rangeland Management on Salmonid Fishes and Their Habitats. American Fisheries Society Special Publication 19. Mell ina, E . 2002. Stream habitat, and rainbow trout (Oncorhynchus mykiss) behavioural, physiological, and population responses to streamside timber harvesting in the central interior of British Columbia. Ph.D. thesis. U B C . pp. 127. Mell ina, E . , R . D . Moore, S.G. Hinch, J.S. MacDonald, and G.Pearson. 2002. Stream temperature responses to clearcut logging in British Columbia: the moderating influences of groundwater and headwater lakes. Can. J. Fish. Aquat. Sci. 59: 1886-1900. Mesa, M . G . , Poe, T.P., Gadomski, D . M . and J .H. Petersen. 1994. Are all prey created equal? A review and synthesis of differential predation on prey in substandard condition. Journal of Fish Biology 45 (Supplement A ) , 81-96. Mi l le r , R .J . and E . L . Brannon. 1981. The origin and development of life history patterns in pacific salmonids. In Salmon and Trout Migratory Behavior Symposium, E . L . Brannon and E .O . Salo, (eds.), 296-309. 138 Mitchell, S. 1999. A simple model for estimating mean monthly stream temperatures after riparian canopy removal. Environmental Management, 24: 77-83. Murphy, M.L., J. Heifetz, S.W. Johnson, K.V. Koski, and J.F. Thedinga. 1986. Effects of clear-cut logging with and without buffer strips on juvenile salmonids in Alaskan streams. Can. J. Fish. Aquat. Sci. 43: 1521-1533. Murphy, M.L.,J. Heifetz, J.F. Thedinga, S.W. Johnson, and K.V. Koski. 1989. Habitat utilization by juvenile Pacific salmon (Oncorhynchus) in the glacial Taku River, southeast Alaska. Can. J. Fish. Aquat. Sci. 46:1677-1685. Murray, C.B., and M.L. Rosenau. 1989. Rearing of juvenile chinook salmon in nonnatal tributaries of the lower Fraser River, British Columbia. Trans. Am. Fish. Soc. 118: 284-289. Murray, P.R., S.R. Hamilton, and G.O. Stewart. 1981. Studies on juvenile chinook salmon (Oncorhynchus tshawytscha) in the Bowron and Willow rivers, B.C., during 1980. Report prepared for DFO by Envirocon Ltd. 86 p. Newbold, J.D., D.C. Erman, and K.B. Roby. 1980. Effects of logging on macroinvertebrates in streams with and without buffer strips. Can. J. Fish. Aquat. Sci. 37:1076-1085. Newcombe, CP . and D.D. MacDonald. 1991. Effects of suspended sediments on aquatic ecosystems. North American Journal of Fisheries Management, 11: 72-82. Nislow, K.H., C. Folt, and M . Seandel. 1998. Food and foraging behavior in relation to microhabitat use and survival of age-0 Atlantic salmon. Can. J. Fish. Aquat. Sci. 55:116-127. Noggle, C.C. 1978. Behavioral, physiological and lethal effects of suspended sediment on juvenile salmonids. A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science, University of Washington. 87 p. 139 Nordeng, H. 1983. Solution to the "charr problem" based on Arctic charr (Salvelinus alpinus) in Norway. Can. J. Fish. Aquat. Sci. 40: 1372-1387. Olla, B.L. , and M.W. Davis 1989. The role of learning and stress in predator avoidance of hatchery-reared coho salmon (Oncorhynchus kisutch) juveniles. Aquaculture 76: 209-214. Olla, B.L. , M.W. Davis and C B . Schreck. 1992. Comparison of predator avoidance capabilities with corticosteroid levels induced by stress in juvenile coho salmon. Trans. Am. Fish. Soc. 121: 544-547. Poulin, V . A . 1986. Extent and severity of mass wasting on the Queen Charlotte Islands and impact on fish habitat and forest sites. Fish/forestry interaction program. British Columbia Ministry of Forests and Lands, Land Manage. Rep. 51: 33 p. Puckett, K.J . , and L . M . Di l l . 1985. The energetics of feeding territoriality in juvenile coho salmon (Oncorhynchus kisutch). Behaviour. 92:97-111. Randall, R.G., M.C. Healey, and J.B. Dempson. 1987. Variability in length of freshwater residence of salmon, trout, and char. American Fisheries Society Symposium. 1: 27-41. Redding, J.M., Schreck, C B . and F.H. Everest. 1987. Physiological effects on coho salmon and steelhead of exposure to suspended solids. Trans. Amer. Fish. Sock. 116:737-744. Reid, L . M . , 1998. Cumulative Watershed Effects: Caspar Creek and Beyond. USDA Forest Service Gen. Tech. Rep. PSW-GTR-168. 1998. Reid, S.D., J.J. Dockray, T.K. Linton, D.G. McDonald and C M . Wood. 1995. Effects of a summer temperature regime representative of a global warming scenario on growth and protein synthesis in hardwater and softwater-acclimated juvenile rainbow trout (Oncorhynchus mykiss). J. therm. Biol. 20:231-244. . 140 Reiser, D.W. and T.C. Bjornn. 1979. Habitat requirements of anadromous salmonids. Pages 1-54 In W.R. Meehan, (ed.). Influence of forest and rangeland management on anadromous fish habitat in the Western United States and Canada. U.S. For. Serv. Gen. Tech. Rep. PNW-96. Portland, Oregon. Ringler, N.H., and J.D. Hall. 1975. Effects of logging on water temperature and dissolved oxygen in spawning beds. Trans. Am. Fish. Soc. 104:111-121. Rishel, G.B., J.A. Lynch, and E.S. Corbett. 1982. Seasonal stream temperature changes following forest harvesting. J. Environ. Qual. 11: 112-116. Reimers, P.E. 1968. Social interaction among juvenile fall chinook salmon. J. Fish. Res. Board. Can. 25: 2005-2008. Resources Inventory Committee (RIC). 1997. Freshwater Biological Sampling Manual. Government of BC. http://www.for.gov.bc.ca/ric/Pubs/Aquatic/freshwaterbio/index.htm> Rood, K . M . and R.E. Hamilton. 1995. Hydrology and water use for salmon streams in the Upper Fraser Habitat Management Area, British Columbia. Can. Manuscr. Rep. Fish. Aquat. Sci. 2294: 170 p. Rosberg, G.E. and D. Aitken. 1981. Adult chinook salmon studies in four tributaries to the upper Fraser River. DSS Contract No: 05SB.FP501-0-SP124. pp. 140. Rosberg, G.E., D. Aitken, and E. Oguss. 1981. Juvenile chinook salmon studies in four tributaries to the upper Fraser River, 1981. Prepared by Beak Consultants Ltd. for Department of Fisheries and Oceans, 401 Burrard Street, Vancouver, BC. DSS Contract No: 05SB.FP501-0-SP124.pp. 158. Rosenberg, D.M. and A.P. Wiens. 1978. Effects of sediment addition on macrobenthic invertebrates in a northern Canadian River. Water. Res. 12:753-763. 141 Rosenfeld, J., M . Porter, and E. Parkinson. 2000. Habitat factors affecting the abundance and distribution of juvenile cutthroat trout (Oncorhynchus clarki) and coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 57: 766-774. Russell, L.R., K.R. Conlin, O.K. Johansen, and U. Orr. 1983. Chinook salmon studies in the Nechako River: 1980, 1981, 1982. Can. MS Rep. Aquat. Sci. 1728: 185 p. Schindler, D.W. 1997. Widespread Effects of Climatic Warming on Freshwater Ecosystems in North America. Hydrol. Process., Vol. 11: 1043-1067. Schindler, D.W. 2001. The cumulative effects of climate warming and other human stresses on Canadian freshwaters in the new millennium. Can. J. Fish. Aquat. Sci. 58: 18-29. Schuett-Hames, D., A. Pleus, L. Bullchild, S. Hall. 1994. Ambient Monitoring Program Manual. Timber-Fish-Wildlife. TFW-AM9-94-001. Northwest Indian Fisheries Commission. Olympia, Washington, USA. 128pp. Scott, W.B., and E.J. Crossman. 1973. Freshwater fishes of Canada. Fish. Res. Board Can. Bull. 184: 1-966. Scrivener, J.C. 1988. Changes in composition of the streambed between 1973 and 1985 and the impacts on salmonids in Carnation Creek, p. 59-65. In T.W. Chamberlin (ed.). Proceedings of the workshop: Applying 15 years of Carnation Creek results. Pacific Biological Station, Nanaimo, British Columbia. 239 p. Scrivener, J.C, and M.J. Brownlee. 1989. Effects of forest harvesting on spawning gravel and incubation survival of chum (Oncorhynchus keta) and coho (Oncorhynchus kisutch) in Carnation Creek, British Columbia. Can. J. Fish. Aquat. Sci. 46: 681-696. Scrivener, J.C, T.G. Brown, and B.C. Andersen. 1994. Juvenile chinook salmon (Oncorhynchus tshawytscha) utilisation of Hawks Creek, a small and non-natal tributary of the upper Fraser River. Can. J. Fish. Aquat. Sci. 51:1139-1146. Seber, G.A.F., and E.D. Le Cren. 1967. Estimating population parameters from catches large relative to the population. J. Anim. Ecol. 36: 631-643. Selye, H. 1950. Stress and the general adaptation syndrome. British Medical Journal 1950(1):1383-1392. Servizi, J.A. and D. W. Martens. 1991. Effect of temperature, season, and fish size on acute lethality of suspended sediments to coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 48:493-497. Servizi, J.A. and D.W. Martens. 1987. Some effects of suspended Fraser River sediments on sockeye salmon (Oncorhynchus nerka). In H.D. Smith, L . Margolis and C C . Wood (ed.) Sockeye salmon (Oncorhynchus nerka) population biology and future management. Can. spec. Publ. Fish. Aquat. Sci. 96, p. 254-264. Servizi, J.A. and D.W. Martens. 1992. Sublethal responses of coho salmon (Oncorhynchus kisutch) to suspended sediments. Can. J. Fish. Aquat.Sci. 49:1389-1395. Servizi, J.A. and R.W. Gordon. 1990. Acute lethal toxicity of ammonia and suspended sediment mixtures to chinook salmon (Oncorhynchus tshawytscha). Bull. Environ. Contam. Toxicol. 44:650-656. Shaw, E.A. and J.S. Richardson. 2001. Direct and indirect effects of sediment pulse duration on stream invertebrate assemblages and rainbow trout (Oncorhynchus mykiss) growth and survival. Can. J. Fish. Aquat. Sci. 58: 2213-2221. Shrimpton, J .M., J.F. Bourgeois, J.T. Quigley and D . M . Blouw. 1999a. Removal of the riparian zone during forest harvesting increases stream temperature: are the effects cumulative downstream? In: Darling, L . M . (ed.) Proc. Biology and Management of Species and Habitats At Risk, Kamloops, B.C., 15-19 Feb. 1999. Volume 2 pp. 531-535. 1 4 3 Shrimpton, J.M., J.F. Bourgeois, J.T. Quigley and D.M. Blouw. 1999b. Manipulation of Riparian Reserves and Stream Temperatures to Evaluate the Effects of Forest Harvest Practices on Fish Populations. Science Council British Columbia Final Report, FRBC # OP97254-1 RE, 24pp. Sigler, J.W., T.C. Bjornn, and F.H. Everest. 1984. Effect of chronic turbidity on density and growth of steelheads and coho salmon. Trans. Amer. Fish. Soc. 113:142-150. Slaney, P.A., T.G. Halsey, and A.F. Tautz. 1977. Effects of forest harvesting practices on spawning habitats of stream salmonids in the Centennial Creek watershed. British Columbia Ministry of Environment, Fish and Wildlife Branch, Fisheries Management Report 73, Vancouver. Solazzi, M.F., T.E. Nickelson, S.L. Johnson, and J.D. Rodgers. 2000. Effects of increasing winter rearing habitat on abundance of salmonids in two coastal Oregon streams. Can. J. Fish. Aquat. Sci. 57: 906-914. Sommer, T.R., M.L. Nobriga, W.C. Harrell, W. Batham, and W.J. Kimmerer. 2001. Floodplain rearing of juvenile chinook salmon: evidence of enhanced growth and survival. Can. J. Fish. Aquat. Sci. 58:325-333. Sprague, J.B. 1971. Measurement of pollutant toxicity to fish. III. Sub-lethal effects and "safe" concentrations. Water Res. 5: 245-266. Sumpter, J.P., H.M. Dye and T.J. Benfey. 1986. The effects of stress on plasma ACTH, -MSH, and C o r t i s o l levels in salmonid fishes. Gen. Comp. Endocrinol. 62: 377-385. Swales, S., F. Caron, J.R. Irvine, and C D . Levings. 1988. Overwintering habitats of coho salmon (Oncorhynchus kisutch) and other salmonids in the Keogh River system, British Columbia. Can. J. Zool. 66:254-261. 144 Swales, S., R.B. Lauzier and C D . Levings. 1986. Winter habitat preferences of juvenile salmonids in two interior rivers in British Columbia. Can. J. Zool. 64:1506-1514. Sylvester, J.R. 1972. Effect of thermal stress on predator avoidance in sockeye salmon. J. Fish. Res. Bd. Canada 29: 601-603. Taylor, E.B. 1988. Adaptive variation in rheotactic and agonistic behaviour in newly emerged fry of chinook salmon, Oncorhynchus tshawytscha, from ocean- and stream-type populations. Can. J. Fish. Aquat. Sci. 45: 237-243. Taylor, E.B. 1990. Phenotypic correlates of life history variation in juvenile chinook salmon, Oncorhynchus tshawytscha. J. Anim. Ecol. 59: 455-468. Taylor, E.B. and P.A. Larkin. 1986. Current response and agonistic behavior in newly emerged fry of chinook salmon, Oncorhynchus tshawytscha, from ocean- and stream-type populations. Can. J. Fish. Aquat. Sci. 43:565-573. Taylor, G.C, J.A. Allan, and M.J. Bradford. 1994. Juvenile chinook sampling data, Slim Creek, British Columbia, 1993. Can. Data Rep. Fish. Aquat. Sci. No. 942. Temple, S.A. 1987. Do predators always capture substandard individuals disproportionately from prey populations? Ecology: 68(3) pp.669-674. Thedinga, J.F., and K.V. Koski.1984. A stream ecosystem in an old-growth forest in southeast Alaska. Part VI: The production of coho salmon (Oncorhynchus kisutch) smolts and adults from Porcupine Creek, pp. 99-108. In: W.R. Meehan, T.R. Merrell, Jr., and T.A. Hanly (eds.). Fish and wildlife relationships in old-growth forests, proceedings of a symposium held April 1982, Juneau, AK. Am. Inst. Fish. Res. Biol., Juneau, AK. Theurer, F.D., I.Lines, and T. Nelson. 1985. Interaction between riparian vegetation, water temperature, and salmonid habitat in the Tucannon River. Water Resour. Bull. 21:53-64. 145 Thomas, R.E., J.A. Gharrett, M.G.Carls, S.D. Rice, A. Moles, and S. Korn. 1986. Effects of fluctuating temperature on mortality, stress, and energy reserves of juvenile coho salmon. Trans. Am. Fish. Soc. 115: 52-59. Thomas, P. 1990. Molecular and biochemical responses of fish to stressors and their potential use in environmental monitoring. Am. Fisheries Soc. Symp. 8: 9-28. Thorpe, J.E. 1987. Smolting versus residency: developmental conflict in salmonids. American Fisheries Society Symposium. 1: 244-252. Titcomb, J.W. 1926. Forests in relation to freshwater fishes. Trans. Amer. Fish. Soc. 56:122-129. Toews, D.A.A. and M.K. Moore. 1982a. The effects of streamside logging on large organic debris in Carnation Creek. B.C. Ministry of Forests Land Manage. Rep. 11. 29 p. Toews, D.A.A. and M.K. Moore. 1982b. The effects of three streamside logging treatments on organic debris and channel morphology of Carnation Creek, In: G.Hartman (ed.), Proceedings of the Carnation Creek workshop, a 10 year review, Feb 24-26, 1982. Pacific Biological Station, Nanaimo, B.C. p. 129-153. Tschaplinski, P.J., and G.F. Hartman. 1983. Winter distribution of juvenile coho salmon (Oncorhynchus kisutch) before and after logging in Carnation Creek, British Columbia, and some implications for over-winter survival. Can. J. Aquat. Sci. 400: 452-461. Tutty, B.D., and F.Y.E. Yole. 1978. Overwintering chinook salmon in the upper Fraser River system. Can. Fish. Mar. Serv. MS Rep. 1460: 23 p. Waldichuk, M . 1993. Fish habitat and the impact of human activity with particular reference to Pacific salmon, p. 295-337. In L.S. Parsons and W.H. Lear (eds.) Perspectives on Canadian marine fisheries management. Can. Bull. Fish. Aquat. Sci. 226. 146 Walthers, L.C. and J.C. Nener. 1997. Continuous water temperature monitoring in the Nicola River, B.C., 1994: implications of measured temperatures for anadromous salmonids. Can. Tech. Rep. Fish. Aquat. Sci. 2158: 65 p. Walthers, L.C. and J.C. Nener. 2000. Water temperature monitoring in selected Thompson River tributaries, B.C., 1996: implications of measured temperatures for anadromous salmonids. Can. Tech. Rep. Fish. Aquat. Sci. 2306: 69 p. Waters, T.F. 1995. Sediment in streams: sources, biological effects and control. Amer. Fish. Soc. Bethesda, Maryland. 251 p. Weatherley, A.H., and H.S. Gill. 1995. Growth, pp.70-101. In C. Groot, L. Margolis, and W.C. Clarke, (eds.) Physiolodical Ecology of Pacific Salmon. UBC Press, Vancouver. Welsh, H.H. Jr., G.R. Hodgson, B. Harvey, and M.F. Roche. 2001. Distribution of Juvenile Coho Salmon in Relation to Water Temperatures in Tributaries of the Mattole River, California. North Amer. J. Fish. Manage. 21:464-470. Wilzbach, M.A., K.W. Cummins, and J.D. Hall. 1986. Influence of habitat manipulations on interactions between cutthroat trout and invertebrate drift. Ecology. 67(4): 898-911. Young, K.A., S.G. Hinch, and T.G. Northcote. 1999. Status of resident coastal cutthroat trout and their habitat twenty-five years after riparian logging. North Amer. J. Fish. Manage. 19: 901-911. Zinner, D., J. Hindahl, and M . Schwibbe. 1997. Effects of temporal sampling patterns of all occurrence recording in behavioural studies: many short sampling periods are better than 147 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0075100/manifest

Comment

Related Items