Characterization of Spawning Habitat, Incubation Environment and Early Growth
and Development in Bull Trout {Salvelinus confluentus) from Pristine Streams of
Northern British Columbia

Cory John Williamson
B.Sc. Simon Fraser University, 1996

Thesis Submitted in Partial Fulfillment of
The Requirements for The Degree of
Master of Science
in
Natural Resources and Environmental Studies

The University of Northern British Columbia
December, 2006

© Cory John Williamson, 2006

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Library and
Archives Canada

Bibliotheque et
Archives Canada

Published Heritage
Branch

Direction du
Patrimoine de I'edition

3 9 5 W ellington Street
Ottawa ON K1A 0N 4
C anada

395, rue W ellington
Ottawa ON K1A 0N 4
C anada
Your file Votre reference
ISBN: 978-0-494-28425-4
Our file Notre reference
ISBN: 978-0-494-28425-4

NOTICE:
The author has granted a non­
exclusive license allowing Library
and Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distribute and sell theses
worldwide, for commercial or non­
commercial purposes, in microform,
paper, electronic and/or any other
formats.

AVIS:
L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans
le monde, a des fins commerciales ou autres,
sur support microforme, papier, electronique
et/ou autres formats.

The author retains copyright
ownership and moral rights in
this thesis. Neither the thesis
nor substantial extracts from it
may be printed or otherwise
reproduced without the author's
permission.

L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these.
Ni la these ni des extraits substantiels de
celle-ci ne doivent etre imprimes ou autrement
reproduits sans son autorisation.

In compliance with the Canadian
Privacy Act some supporting
forms may have been removed
from this thesis.

Conformement a la loi canadienne
sur la protection de la vie privee,
quelques formulaires secondaires
ont ete enleves de cette these.

While these forms may be included
in the document page count,
their removal does not represent
any loss of content from the
thesis.

Bien que ces formulaires
aient inclus dans la pagination,
il n'y aura aucun contenu manquant.

i*i

Canada
R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Bull trout home to tributary streams in headwater areas for spawning. Site use
can be influenced by habitat quality, including physical features such as stream flow,
intergravel flow, hyporheic exchange, stream and intergravel-water chemistry,
temperature and cover. These variables can influence survival and growth during
incubation, however, the effect on bull trout spawning and incubation success is not
known. I used three studies to examine the importance o f spawning habitat for bull trout
spawning in northern B.C. My results show that temperature is likely a key determinant
o f site use at potential spawning sites as well as incubation success in northern
environments. Bull trout growth and yolk utilization were substantially altered by small
changes in incubation temperature in both the field and the laboratory. In this context,
fisheries and resource managers should take steps to prevent harmful alterations to bull
trout spawning areas as the result o f human activities.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table of Contents
ABSTRACT:.............................................................................................................................i
List of Tables...................................................................................................................... iv
List of Figures.................................................................................................................... vii
Acknowledgements.............................................................................................................xi
CHAPTER 1: INTRODUCTION............................................................................................1
CHAPTER 2: GROUNDWATER AND THE INCUBATION ENVIRONMENT FOR
BULL TROUT IN PRISTINE NORTHERN STREAMS..................................................... 6
Abstract................................................................................................................................. 6
Introduction...........................................................................................................................7
Materials and Methods.........................................................................................................9
Study A rea........................................................................................................................9
Stream Surveys...............................................................................................................12
Redd Habitat Characteristics..........................................................................................14
Intergravel Redd Characteristics................................................................................... 17
Hyporheic Exchange...................................................................................................... 18
Temperature....................................................................................................................19
Data Analysis................................................................................................................. 20
Results.................................................................................................................................21
Discussion...........................................................................................................................30
Habitat Use at the Reach Level..................................................................................... 32
Redd Habitat Characteristics......................................................................................... 36
Factors Moderating Site Use......................................................................................... 37
Lack of Habitat U se.......................................................................................................39
Summary.............................................................................................................................42
CHAPTER 3: FACTORS AFFECTING GROWTH AND SURVIVAL OF LARVAL
BULL TROUT IN PRISTINE NORTHERN STREAMS.................................................. 43
Abstract...............................................................................................................................43
Introduction........................................................................................................................ 44
Materials and Methods.......................................................................................................46
Study A rea..................................................................................................................... 46
Gamete Collection.........................................................................................................46
Egg Capsules.................................................................................................................. 47
Incubation Sites.............................................................................................................. 48
Alevin Survival, Size and Yolk Utilization................................................................. 49
Temperature, Hyporheic Flow and Artificial Redd Chemistry...................................50
Data Analysis................................................................................................................. 51
Survival and Size.......................................................................................................51
Habitat.........................................................................................................................52
Results................................................................................................................................. 54
Hatch and Survival.........................................................................................................54
Intergravel Environment................................................................................................60
Temperature............................................................................................................... 60

ii

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Hyporheic Flow and Intergravel Chemistry:........................................................... 64
Embryo Size and Yolk Utilization................................................................................67
Discussion...........................................................................................................................75
Site Selection Habitat and Groundwater Influences....................................................77
Habitat and Survival...................................................................................................... 80
Parental Effects on Survival.......................................................................................... 84
Temperature, Size and Development............................................................................85
Summary.............................................................................................................................87
CHAPTER 4 : THE EFFECT OF AN INCREASE IN INCUBATION TEMPERATURE
ON A COLD-WATER ADAPTED SALMONID, THE BULL TROUT........................... 88
Abstract...............................................................................................................................88
Introduction.........................................................................................................................89
M ethods..............................................................................................................................91
Experimental Setup........................................................................................................91
Fish.................................................................................................................................. 93
Survival, Embryo Size and Yolk Utilization............................................................... 93
Muscle Fibre Density.....................................................................................................94
Data Analysis................................................................................................................. 95
Results................................................................................................................................. 96
Discussion......................................................................................................................... 108
Rate of Development................................................................................................... 109
Embryonic Size and Yolk Utilization.........................................................................112
Muscle Cellularity at Hatch and Button-Up.............................................................. 113
Yolk Utilization........................................................................................................... 115
Summary........................................................................................................................... 116
CHAPTER 5: GENERAL DISCUSSION.......................................................................... 117
REFERENCES.....................................................................................................................122
APPENDIX A: ANOVA TABLES FOR ANALYSIS OF SURVIVAL............................131
APPENDIX B: ANOVA TABLES FOR ANALYSIS OF GROUNDWATER
CHEMISTRY.......................................................................................................................133
APPENDIX C: ANOVA TABLES FOR ANALYSIS OF EMBRYO SIZE AND YOLK
UTILIZATION...................................................................................................................... 135
APPENDIX D: ANOVA TABLES FOR ANALYSIS OF LABORATORY GROWTH. 137

iii

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

List of Tables
Table 1. Cover types associated with redd sites surveyed in the upper Davis River in
2001. Columns are tabulated by the percentage of redds with a score for each
category.................................................................................................................... 23
Table 2. Habitat and chemical data collected in the stream environment at redd sites in
the Upper Davis River in September 2001............................................................ 24
Table 3. Habitat and chemical data collected in the intergravel environment at redd sites
in the Upper Davis River in September 2001........................................................24
Table 4. Mean temperatures for selected natural redds from the Davis River Chowika
Creek and Swannell River collected in 2000-2002. Temperatures were
calculated over times to the eyed stage, hatch and button-up as determined in
Chapter 4. Where hydraulic gradient data were collected: U = upwelling; D =
downwelling. (n) indicated the number of redds present at a site. For the
periods: Pre-winter, n =1464 and 61 days; Pre-hatch, n = 2568 and 107 days;
Pre-emergence, n = 2135 and 89 days................................................................... 27
Table 5. Percent hatch and accumulated thermal units (ATU) for all capsule sites in both
experimental years. Sites are categorized by level of use and thermal profile.. 55
Table 6. Least squares means (ANOVA) for percent survival in experimental year 1
(2000-2001). All means are back transformed, thus standard errors are not
provided. A “*” signifies significant results (ANOVA; P ’s < 0.0005). Letters
within groups show overlap in Tukey’s simultaneous 95% confidence intervals.
Least squares means for parental effects are from a different ANOVA
(Appendix A Table 12) than results for the rest of the table (Appendix A Table
10)............................................................................................................................. 58
Table 7. Least squares means (ANOVA) for percent survival in experimental year 2
(2001-2002). All means are back transformed. A
signifies significant
results (ANOVA, P ’s < 0.0005). Letters within groups show overlap in Tukey’s
simultaneous 95% confidence intervals. Least squares means for male and
female effects are from a different ANOVA (Appendix A Tables 13 and 14
respectively) than results for the rest of the table (Appendix A Table 11)..........59
Table 8. Survival (%) at 100% hatch for each temperature regime by incubation tray,
family, female and male.......................................................................................... 99
Table 9. Date and number of days to 50 % hatch and button-up for three laboratory
temperature regimes and one natural redd from the Davis River (Davis River
Field). “Date” designates the estimated date for each stage; “Days” designates
the number of days from fertilization; and “ATU” represents the summed
average daily temperature for each period specified. The “Davis River Field”

iv

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

column is presented for comparison only and does not represent a specific
spawning or fertilization time in the field..............................................................99
Table 10. ANOVA table for year one analysis of survival, which includes the factor
family in the test.....................................................................................................131
Table 11. ANOVA table for year two analysis of survival, which includes the factor
family in the test.....................................................................................................131
Table 12. ANOVA table for year two analysis of survival, which includes the factors
male and female in the test....................................................................................131
Table 13. ANOVA table for year two analysis of survival, which includes the factor
male in the test....................................................................................................... 131
Table 14. ANOVA table for year two analysis of survival including the factor female in
the test..................................................................................................................... 132
Table 15. ANOVA tables for examination o f variance in specific discharge within three
collection periods (September, March and April, respectively)......................... 133
Table 16. ANOVA tables for examination o f variance in specific conductivity within
three collections periods (September, March and April, respectively).............. 133
Table 17. ANOVA tables for examination o f variance in oxygen concentration within
three collections periods (September, March and April, respectively)
134
Table 18. ANOVA tables for analysis of variance in specific discharge, specific
conductivity, and oxygen concentration respectively, between two collections
periods (September April)..................................................................................... 134
Table 19. ANOVA table for year one analyses of length with factors for male and female
parent included in the test......................................................................................135
Table 20. ANOVA table for year one analyses of total dry weight with factors for male
and female parent included in the test.................................................................. 135
Table 21. ANOVA tables for year one analyses of dry-body weight with factors for male
and female parent included in the test.................................................................. 135
Table 22. ANOVA tables for year one analyses of dry yolk weight with factors for male
and female parent included in the test.................................................................. 135
Table 23. ANOVA tables for year two analyses of length with the factor family included
in the test.................................................................................................................136

v

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 24. ANOVA tables for year two analyses of total dry weight with the factor family
included in the test................................................................................................. 136
Table 25. ANOVA tables for year two analyses of dry-body weight with the factor
family included in the test..................................................................................... 136
Table 26.ANOVA tables for year two analyses of dry yolk weight the factor family
included in the test................................................................................................. 136
Table 27.

ANOVA table for the

laboratory

analysis

of length........... 137

Table 28.

ANOVA table for the

laboratory

analysis

of dry-body weight.137

Table 29.

ANOVA table for the

laboratory

analysis

of dry-yolk weight.137

Table 30. Four ANOVA tables for the laboratory analysis of muscle fibres for myomeres
1-3 and 4 & 5 combined, respectively..................................................................137

vi

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

List of Figures
Figure 1. Map of study area showing the three streams used for the redd surveys (dark
blue stream lines) and streams surveyed but not used (green line). Artificial
incubation sites are plotted as
in Chowika Creek, Davis River and Swannell
River. High-use spawning areas are shown as a thick dark-blue line. The road
network is plotted in orange. The inset map shows the study area outlined inside
the red box................................................................................................................ 11
Figure 2. Map of redds surveyed in the Upper Davis River showing the survey limits for
aerial and ground-based redd surveys completed in fall 2001 and fall 2002. The
“plus” denotes the up and downstream limits of the detailed aerial surveys for
2001 only. The “star” denotes the up and downstream limits for both ground
surveys. Circles and triangles represent redds observed in fall 2001 and fall
2002 respectively. Red symbols represent 2001 surveys, Blue symbols represent
the 2002 surveys. Surveys completed in 2002 were not analyzed as part of this
thesis and are presented for comparison only........................................................ 13
Figure 3. Map of redd distributions in high use spawning area in Chowika Creek for
2001-2002. The “plus” denotes the up and downstream limits of the detailed
survey. Circles and triangles represent redds observed from the air in fall 2001
and fall 2002, respectively...................................................................................... 15
Figure 4. Plot of mean monthly temperatures for selected natural redds collected in the
winter 2000-2001 and 2001-2002. Where hydraulic gradient data were
collected, U = upwelling; D = downwelling; The number in brackets for each
site indicates the number o f redds present at a site............................................... 26
Figure 5. Plot of mean monthly temperatures for selected stream flow sites collected in
the winter 2001-2002. Open and dashed lines represent non-selected spawning
areas..........................................................................................................................28
Figure 6. Accumulated centigrade thermal unit (degree day) profiles calculated for
selected natural redds and stream flow sites collected in the winter 2000-2001
and 2001-2002. Where hydraulic gradient data were collected: U = upwelling;
D = downwelling. (n) indicated the number of redds present at a site; RB = river
bottom...................................................................................................................... 31
Figure 7. Boxplots of percent survival both experimental years. For each experimental
year the medians, interquartile ranges (box) , limits (whiskers) and outliers ( * )
are presented............................................................................................................ 5 6
Figures 8a-8c. Temperature profiles recorded for artificial egg incubation site for both
years. Each figure presents temperature data for up to three artificial redds (30cm intergravel), one deep (50-cm Intergravel) and the stream-bed for each
site.............................................................................................................................61

v ii

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Figure 9. Specific discharge and intergravel chemistry for artificial redds in year two
during the September and April collection periods. Data presented are data
means and standard errors for samples collected in the stream (chemistry only)
and at redd depth (30 cm. Letters following each reach name signify collection
dates where S = September and A = April. A
following reach name shows
sites where bull trout spawned...............................................................................65
Figure 10. Effect of reach on growth and yolk utilization for alevins collected in both
field experiments. Marginal means with standard errors are presented for total
dry weight, dry-body weight, dry-yolk weight and wet length except for drybody weight where only back transformed means are provided. Shared letters
represent overlap in Tukey’s simultaneous 95% confidence intervals. Lower
case letters are for year one and upper case letters are for year two. A
indicates sites where bull trout spawned................................................................68
Figure 11. Effect of parent on growth and yolk utilization for alevins collected in both
field experiments. Marginal means with standard errors are presented for total
dry weight, dry-body weight, dry-yolk weight and wet length, except for drybody weight where only back transformed means are provided, thus standard
errors are not provided for dry-body weight. Shared letters represent overlap in
Tukey’s simultaneous 95% confidence intervals. Uppercase and lowercase
letters differentiate males from females.................................................................69
Figure 12. Effect of family (year 2) on growth and yolk utilization for alevins collected
in both field experiments. Marginal means with standard errors are presented
for total dry weight, dry-body weight, dry-yolk weight and wet length except for
dry-body weight where only back transformed means are provided, thus
standard errors are not provided for dry-body weight. Shared letters represent
overlap in Tukey’s simultaneous 95% confidence intervals................................ 70
Figure 13. Means and standard errors of lengths for alevins from artificial redd sites
collected over both years plotted against accumulated thermal units (ATU).
Each point is an arithmetic mean for all alevins from one redd on a single date.
Individual reaches are differentiated in the legend by letters for each stream and
by numbers for each reach. (“S” for the Swannel River, “D” for the Davis River
and “C” for Chowika Creek). A
indicates sites where bull trout spawned. 71
Figure 14. Means and standard errors of dry-body weight for alevins from artificial redd
sites collected over both years plotted against accumulated thermal units (ATU).
Each point is an arithmetic mean for all alevins from one redd on a single date.
Individual reaches are differentiated in the legend by letters for each stream and
by numbers for each reach. (“S” for the Swannel River, “D” for the Davis River
and “C” for Chowika Creek). A
indicates sites where bull trout spawned. 72

viii

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Figure 15. Means and standard errors of dry yolk weights for alevins from artificial redd
sites collected over both years plotted against accumulated thermal units (ATU).
Each point is an arithmetic mean for all alevins from one redd on a single date.
Individual reaches are differentiated in the legend by letters for each stream and
by numbers for each reach. (“S” for the Swannel River, “D” for the Davis River
and “C” for Chowika Creek). A
indicates sites where bull trout spawned. 74
Figure 16. Temperature achieved for three incubation replicates in the laboratory plotted
with natural temperatures for a redd from the Davis River (Davis River Field).
Natural redd temperatures were collected in 2000-2001. A second order
polynomial was fit to this data to provide a general template to guide
temperature control of the Davis River treatment regime....................................97
Figure 17. Duration of development (hatch and button-up) plotted against thermal history
in accumulated thermal units (ATU). Data from the other authors were not
included in the models for our data...................................................................... 100
Figure 18. Thermal history of three alevins collected on March 18, 2002 under three
different temperature regimes. The arrow show the sample date. Lengths of
each individual alevin are presented for comparison and are representative of the
average alevin in each lot. The dashed line “Davis River Field” is the
temperature profile for a natural redd collected in winter 2000-2001. The arrow
shows the precise collection date..........................................................................102
Figure 19. Length plotted against accumulated temperature units (ATU;°C day) for each
temperature replicate..............................................................................................103
Figure 20. Dry-body weight plotted against accumulated temperature units
(ATU;°C-day) for each temperature replicate......................................................103
Figure 21. Dry yolk plotted against accumulated temperature units (ATU;°C-day) for
each temperature replicate.....................................................................................104
Figure 22. Length plotted against days from fertilization for each temperature
replicate..................................................................................................................105
Figure 23. Dry-body weight plotted against days from fertilization for each temperature
replicate................................................................................................................. 105
Figure 24. Dry yolk plotted against days from fertilization for each temperature
replicate...................................................................................................................106

ix

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Figure 25. Mean muscle fibre (myofibre) count from each muscle myomere for each
temperature replicate at hatch and at button-up. Some fish did not have a
distinguishable fifth myomere, so 4 and 5 were combined. Sample sizes are
provided for each temperature regime at hatch and button-up for each myomere.
“H” designates hatch and “B” designates button-up.......................................... 107

x

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Acknowledgements
These studies were generously funded by grants and support from the following
organizations: Peace Williston Fish and Wildlife Compensation Program; Fisheries
Renewal B.C.; the Northern Land Use Institute; Abitibi Consolidated Inc.; the British
Columbia Ministry of Environment, and the Pacific Western Brewery Ltd.
This project was made possible through the contributions o f many individuals
through all stages from the field to the final draft and my defence. Thanks especially to
my supervisor Dr. Mark Shrimpton who gave me the opportunity to enter graduate
studies, had a seemingly endless supply of patience on this very long journey and who
greatly improved the quality of my writing. Thanks also go to my committee, Dr. Ellen
Petticrew, Dr. Mike Gillingham and Dr. Ken Otter for their support and feedback.
Thanks also go to Dr. Steve Macdonald of the Department of Fisheries and Oceans for
his early support and encouragement that ultimately led me to eventually pursue graduate
studies. Thanks especially to Ted Zimmerman (Ministry of Environment) who conceived
of the need for this work and developed proposals for funding from the Peace Williston
Fish and Wildlife Program and sought a partnership between the University of Northern
British Columbia and the B.C. Ministry o f Environment.
Many people helped in the field and in the lab. Ted Zimmerman and Pierre Bock
(Vancouver Island Helicopters) helped with the long days in the field in all kinds of crazy
weather and places. They finished the field work in year 1 when I smashed my finger
and stepped up to the challenge o f winter field-work even when it meant chipping
through 30 cm of ice in -25°C, in frozen waders to collect samples. Pierre is unique and
un-matched as a helicopter pilot, not only does he get you there safely every-time, he is
always there to help get the job done in any way he can. Thanks go to Dr. Mark
Shrimpton, Don Cadden (Ministry of Environment) Ray Pillipow (Ministry of
Environment) who helped in the field collecting fish, digging redds and transporting eggs
back to Prince George. Thanks to the anonymous flight controllers in Vancouver who
saved me an additional year of work when they granted us permission to keep flying after
the terrible events on September 11, 2001. Aurora Sentlinger, Kevin Memickle and Gil
Earner assisted with the histology work in the laboratory. Thanks to Tom Lebo and
Michael Schlick from the Pacific Western Brewery in Prince George who provided space
and assistance for the laboratory study and allowed me to say that “I grew fish in a
brewery”.
Thanks to my extended family who have provided support and encouragement to me and
my wife Susanne throughout this long process. Finally to my wife Susanne, my best
friend and to my two boys Martin (3) and Liam (1) thanks most of all for your sacrifices,
keeping me grounded, taking care of me and feeding my soul when things seemed
impossible. I could not have done this without you. To Martin: My “book” is indeed
“smaller” now!

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

CHAPTER 1: INTRODUCTION
Bull trout (Salvelinus confluentus) are a large piscivorous char (Family
S a l m o n i d a e ) found in cold rivers and lakes in western North America from Northern

California to Alaska (Cavender 1978; Haas and McPhail 1991). Several characteristics of
this species are thought to make it vulnerable to environmental change and ecological
disturbance. Bull trout are known to exhibit aggressive feeding behaviours, have a
propensity to hybridize with introduced eastern brook trout, (Salvelinus fontinalis)
(Markle 1992; Kitano 1994) and show a distinct preference for relatively cold-water
habitats (Baxter and McPhail 1996). As a result, bull trout are sensitive to overfishing,
competitive exclusion and habitat disruption or loss (Fraley and Shepard 1989). Until
recently, bull trout were not managed on the basis of species-specific requirements and as
a result many populations are in decline and are considered endangered, at risk or
vulnerable (Goetz 1989; Rieman and McIntyre 1995; McCart 1997; Swanberg 1997;
Thurow 1997).
Bull trout exhibit homing behaviours and spawning-site fidelity (Pillipow and
Williamson 2003; Bahr and Shrimpton 2004). They also appear to select specific
microhabitats within particular stream reaches for reproduction (Baxter and McPhail
1999; Baxter and Hauer 2000). It is not clear, however, what advantages and features the
selected sites confer in terms of survival and growth for eggs and alevins. Habitat
features such as in-stream cover and groundwater influence may be important for
spawning-site use. Groundwater and thermal habitat that is stable may be especially
important for success of spawning and rearing (Goetz 1989; Baxter and McPhail 1999).
In one northern British Columbia (B.C.) stream (Chowade River), Baxter and McPhail

1

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

(1999) found that bull trout used a single groundwater upwelling area for spawning
despite the presence of other numerous potential spawning sites in an adjacent section of
river. Microhabitat factors including stream flow, depth and substrate size were
qualitatively similar between the selected and non-selected areas. Groundwater influence
was cited as the factor explaining differences between habitats which included the
thermal regime, vertical hydraulic gradient and intergravel flow.
The broad use of the term 'groundwater' in many studies has made it unclear,
however, which attributes of groundwater are important in spawning areas. Several
possibilities exist such as, absolute temperature, temperature differences from surface
waters, flow, chemical constituents (including specific ions) and dissolved gases,
particularly oxygen. These constituents can vary widely between sources of
groundwater, whether it is deep sub-surface water (phreatic) or hyporheic in origin
(Freeze and Cherry 1979). Phreatic groundwater is often devoid of oxygen; thus sources
of pure phreatic water would be avoided by spawning fish. Hyporheic water, which is the
zone of mixing between surface and ground waters, in contrast, can provide a thermally
stable as well as an oxygen-rich environment ideal for spawning and incubation and is a
zone of substantial biological activity. Mixing of the two sources is normal, with the
depth of the hyporheic zone dependant on local sub-surficial stratigraphy (sediment types
and sizes), phreatic groundwater flow and surface discharge. For bull trout it is poorly
understood at which habitat scale, as well as which chemical and physical attributes of
groundwater are important for site use and successful incubation. Baxter and Hauer
(2000) analysed the relationship between hyporheic groundwater and redds (spawnng
sites) for bull trout at several scales in Montana, USA. They found a strong link with

2

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

upwelling and redd distributions among reaches. Reaches are stream sections containing
relatively homogenous habitat (e.g. pools and riffles) of similar gradients. In contrast
with other studies, however, their data did not support the conclusion that upwelling is
important at the habitat scale. Baxter and Hauer (2000) concluded that bull trout in their
study area spawn in wide alluvial valleys within stream reaches that have a complex
hyporheic interface characterized by extensive zones of upwelling sub-surface water.
Specific redds sites were most often located in transitional zones between habitat units
(e.g., pool-riffle transitions) with strong localized down-welling of surface water and high
intergravel flow. Other features may also be important for spawning site use. Bull trout
often spawn in clear, shallow headwater areas adjacent to cover and redds are often found
adjacent to stream margins.
This specificity may make bull trout particularly susceptible to anthropogenic
alterations in their environment. The decline in populations of bull trout in many
jurisdictions appears to be linked to habitat changes brought about by industrial
development as well as increased access to watersheds (Fraley and Shepard 1989; Ripley
2004). The threat to fish populations due to changes in the thermal environment caused
by resource extraction and global warming, is not however, limited to bull trout.
Rombough (1997) reviewed the effects of temperature changes on the development and
growth of temperate fishes and concluded that temperate fish are generally sensitive to
thermal changes in their environment. In another study, DeMarch (1995) demonstrated
that Arctic char {Salvelinus alpinus), a coldwater specialist, incubated above 6°C had
lower survival and growth compared to char incubated at 3°C. Thus, changes to streams
that alter the thermal regime, particularly where fish demonstrate local adaptations and

3

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

low plasticity to large changes, have the potential to alter or reduce species distributions.
At present, many B.C. populations of bull trout remain unstudied and
management decisions are made based on extrapolation of information collected from
other districts. In addition, studies of bull trout habitat have been largely undertaken in
the southern portion of their natural geographic range. Thus, the ecology and habitat
requirements of northern bull trout populations are virtually unknown. Recent studies
from outside B.C. have indicated that bull trout abundance can be linked to habitat
variables in a hierarchical scale (e.g., Watershed, Reach, Site, Habitat) (Watson and
Hillman 1997; Baxter and Hauer 2000). I suggest that extrapolation of these results be
done with caution. Watson and Hillman (1997), for example, found that combinations of
habitat variables that predicted bull trout abundance varied substantially between
different watersheds over a comparatively small geographic range. Extrapolating
knowledge of bull trout life-history and habitat requirements from southern areas of their
range may lead to misquided management decisions for northern populations. In some
areas of B.C. (particularly remote northern districts) bull trout are relatively abundant and
opportunities exist to study their ecology in relatively pristine conditions. Comparatively,
bull trout have not been studied as intensively as other salmonids (e.g., Pacific salmon
and trout) and there are little data describing the habitat requirements and conditions for
successful spawning, incubation and rearing. Furthermore, information describing these
life-stages has been identified as a priority for future research (McPhail and Baxter 1996).
My objectives in these studies were to investigate how habitat influences
spawning-site use, incubation success as well as, growth and development of embryonic
bull trout from tributaries of the Finlay Arm of the Williston Reservoir, B.C. I used a

4

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

combination of field and laboratory studies to examine bull trout spawning and
incubation habitat requirements. In my first study (Chapter 2), I described bull trout
spawning habitat at the scale of individual redds. I focussed on stream and intergravel
chemistry, sub-surface flow, temperature and stream features such as large woody debris
and channel bed-forms. In my second study (Chapter 3), I examined growth and
development of larval bull trout in the context o f habitat in the field. I recorded size
(length, body weight) and yolk utilization at several sites used and not used for spawning.
I also recorded stream and intergravel chemistry, temperature and intergravel flow to
describe factors contributing to variation in size within and among sites. Finally, to
evaluate the effect of temperature on the growth and development of larval bull trout
(Chapter 4), I incubated eggs and alevins under controlled laboratory conditions in a
natural and two altered temperature regimes. For this study, I measured survival, size and
the timing of development under the three temperature treatments. Conclusions derived
from data collected in Chapter 2-4 are presented in Chapter 5 and were used to generate a
better understanding of the reproductive and early life history habitat requirements of
northern populations of bull trout. This information can support the effective
management of these populations.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

CHAPTER 2: GROUNDWATER AND THE INCUBATION ENVIRONMENT
FOR BULL TROUT IN PRISTINE NORTHERN STREAMS1.

Abstract
Bull trout are known to home to, and spawn in, specific stream reaches in headwaters. It
is not well understood, however, what habitat features are important for successful
incubation and survival of incubating bull trout in northern environments. To assess
distribution and habitat characteristics of bull trout redds, we employed a synoptic study
of the physical habitat of bull trout redds in high-use spawning tributaries within the
Finlay reach of the Williston watershed of B.C. We described bull trout spawning sites
located in headwater depositional areas influenced by relatively warm groundwater and
cold winter temperatures approaching freezing. Individual redds were often associated
with at least one type of security cover, high levels of intergravel flow and oxygen, as
well as cold thermal regimes intermediate to what was available. Individual redds were
often clumped with other bull trout redds. Clumped redds were associated with warmer
areas of up-welling, however, in general, redds were located in areas that exhibited both
up- and down-welling.

1 Throughout this chapter I use the first person plural to acknowledge the contribution o f others to this
work, which will be submitted for publication with the authorship of C.J. Williamson, J.T. Zimmerman and
J.M. Shrimpton.

6

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Introduction
Large migratory bull trout home to smaller streams and headwater areas to spawn
(Goetz 1989; Baxter and McPhail 1996; Swanberg 1997; McPhail and Baxter 1999;
Baxter and Hauer 2000; Pillipow and Williamson 2003; Bahr and Shrimpton 2004;
Scholz 2005). Migrations to and from spawning habitat can be extensive and can exceed
distances over 500 km (e.g., Pillipow and Williamson 2004). Migrations over long
distances to spawning areas in other salmonids are well known and have been particularly
well studied in Pacific salmon (Oncorhynchus sp.) (e.g., Groot and Margolis 1991).
Homing involves behaviours that enable adult salmonids to return to high quality
spawning habitats with accuracy (Ditman and Quinn 1996), despite long-distance
migration through sometimes complex habitats.
Based on genetic evidence (Thomas et al. 2001; Spruel et al. 2003) and movement
studies (Carson 2001; Bahr and Shrimpton 2004; Pillipow and Williamson 2004; and see
Scholz et al. 2005 for a thorough review) migratory bull trout appear to home with
precision to natal areas to spawn, where spawners show a high degree of fidelity to
spawning sites within stream reaches. Collectively, these studies suggest that bull trout
are using an adaptive mechanism to seek and select high-quality, reproductive habitats
possibly as a means of optimizing juvenile survival or growth. The specific mechanism
for homing in bull trout has not been identified; however, in Pacific salmon
{Oncorhynchus sp.), behaviours involved in homing are dependant on olfactory and
visual imprinting at an early life stage. Imprinting can sometimes occur before hatch, but
usually takes place during significant phases of development including outmigration from
natal to rearing habitats (reviewed by Dittman and Quinn 1996). Environmental and

7

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

chemical gradients provide the basis for imprinting and subsequent discrimination
between habitats during spawning migrations. Successful imprinting by juveniles and
subsequent homing of mature adults to natal habitat for spawning allows salmonids to
return to specific rivers and stream reaches where site selection, based on microhabitat
variables, can potentially occur. For bull trout, the behaviours that are involved in
homing to natal rivers, to tributaries and then finally to spawning sites are likely
dependant on gradients in habitat quality at multiple spatial scales from the river or basin
scale down to the scale of local habitat (Baxter and Hauer 2000).
In northern latitudes, air and surface-water temperatures below 0°C can limit the
distribution of fishes (Power et al. 1999). Groundwater, however, is often warmer in
winter than surface water, is less subject to diurnal, seasonal and even annual fluctuations
compared with surface water; thus it can provide a stable thermal regime or refuge for
different life stages including incubation. Predictability in environmental conditions
afforded by stream systems mediated by groundwater (including stable or warm thermal
regimes) may provide a basis upon which salmonids have developed adaptive
mechanisms to maximize survival and growth through and beyond the embryonic stages.
Areas of groundwater may also provide a consistent and distinct chemical signature that
aids homing by returning adults.
Groundwater may be especially important in environments where low
temperatures result in the extensive formation o f ice on the stream-bed (anchor ice) and it
has been commonly suggested that upwelling groundwater is important for successful
spawning and rearing of larval and juvenile bull trout as well (see reviews by Goetz
1989; Baxter and McPhail 1999). Anchor ice formation in redds results freezing of eggs

8

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

and larvae. In one northern B.C. stream (Chowade River), Baxter and McPhail (1999)
found that bull trout concentrated at a single groundwater upwelling area for spawning
despite the apparent presence of other numerous potential spawning sites in an adjacent
section of river. Groundwater influence, particularly upwelling at the site level, may not
be the only determinant for spawning site use by bull trout. Several of the studies that
have examined groundwater influence on spawning-site use by bull trout have noted the
relative contribution of upwelling subsurface water, but have not examined the influence
of security cover, hyporheic exchange, vertical hydraulic gradient and local thermal
regime on site selection. Each of these features is an important indicator of habitat
quality that may dictate site use by spawners.
There are few studies describing the natural incubation environment experienced
by salmonids in northern environments during winter. Our objectives for this study were
to identify bull trout-spawning sites in tributaries to the Finlay Reach of the Williston
Reservoir B.C., characterize stream habitat where redds were located, and describe the
physical and chemical environment experienced by incubating eggs and larval bull trout.
Streams surveyed for this work are within the northern part o f the natural range for bull
trout (Haas and McPhail 1991). Information on bull trout spawning habitat is important
for the management of spawning areas and will be generally applicable for the protection
and management of fish habitats.

Materials and Methods
Study Area
The Finlay reach o f the Williston Reservoir in north central B.C. supports several
robust bull trout runs that spawn in streams draining the Rocky and Omineca mountains.

9

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Prior to the construction of the W.A.C. Bennett dam on the Peace River, bull trout in this
watershed were presumed to have exhibited a primarily fluvial life history where
spawning took place in the smaller tributaries with rearing in large rivers, although it is
possible that spawning could have taken place in the large rivers as well. At present there
does not appear to be spawning habitat in the Williston Reservoir and directed migration
followed by spawning has been observed in headwaters of tributaries throughout the
watershed (e.g., O ’Brien and Zimmerman 2000). Three o f these systems, Davis River,
Chowika Creek and Swannell River were used in this study (Fig. 1). The Davis River
was the primary location for this study and subsequent chapters and has been the subject
of ongoing bull trout monitoring and research activities (O’Brien and Zimmerman 2000)
since 1998.
Spawning activity by bull trout in the Davis River peaks in early September and
spans several weeks (O’Brien and Zimmerman 2000). The Davis River is a third-order
stream with a drainage area of approximately 49,000 ha. Chowika Creek is located
directly north of the Davis River and has a similar aspect, size and drainage area to the
Davis River. Both streams flow westward from the front ranges of the Muskwa
Mountains (Northern Rocky Mountains) into the northern end of the Finlay Reach of the
Williston Reservoir. The Swannell River drains 104,000 ha from the eastern slopes of
the Swannell Ranges of the Omineca Mountains. The Swannell River flows into the
Ingenika River, which is the second largest tributary flowing into the Finlay reach of the
Williston Reservoir. The study reaches used in all three streams are in pristine condition
with no direct industrial development (e.g., forestry, mining) and there are no roads in the
headwater spawning areas. All sites were accessed by helicopter.

10

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Reproduced
with permission

B ritish C olum bia -

of the copyright owner. Further reproduction

ilay River

.'■oPesika
Creek

Chowil
Creek

C3

C1-C2

Ingenika River

D2
D1

Wrede Creek

prohibited without perm ission.

Davi^
,River
^

Swannell
Factor Rot

V * iver
NSw .

Creek'"'0 25 5

10

15

20

25

^ Km

Figure 1. Map of study area showing the three streams used for the redd surveys (dark blue stream lines) and streams surveyed but not
used (green line). Artificial incubation sites are plotted as
in Chowika Creek, Davis River and Swannell River. High-use
spawning areas are shown as a thick dark-blue line. The road network is plotted in orange. The inset map shows the study area
outlined inside the red box.

Stream Surveys
To assess distribution and habitat characteristics of bull trout redds, we employed a
synoptic approach to find streams that could be considered high-use spawning areas for
bull trout. One reach, we selected was located in the upper Davis River watershed (Figs.
1 and 2). The importance of this spawning location was established following the
completion of a radio telemetry study that determined that a large proportion of the radio­
tagged bull trout returning to the Davis River spawned in 5-km long reach (Zimmerman
and O’ Brian 2000). Additionally, since 1998, the B.C. Ministry of Environment and the
Peace Williston Fish and Wildlife Compensation Program have completed aerial redd
counts for bull trout in this section of the upper Davis River on an annual basis to index
population trends in the Williston watershed.
In July and August of 2000, we examined seven streams from the vicinity of the
Finlay Arm of the Williston Reservoir that were of similar magnitude and order to the
upper Davis, including Chowika Creek, Pesika Creek, Factor Ross Creek, Swannell
River, West Zygadine Creek, Wrede Creek and Cut-bank Creek for potential as spawning
areas. We used several criteria to qualitatively establish whether a site had potential as a
bull trout spawning area and could be used in the habitat study. These included: aerial
over-flights to determine ease of site access for completing our investigation and to assess
spawning-habitat potential; visual ground-based surveys to establish presence of bull
trout juveniles in typical juvenile habitat in pools and back-channels; snorkel surveys to
locate concentrations of adult and juvenile bull trout; minnow trapping to collect
juveniles; and placement of temperature data-loggers to establish whether stream
temperatures were within the range used by bull trout.

12

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Reproduced with permission
of the copyright owner. Further reproduction

Groundwati
Source
Tributi

;ast Tribute]

Confluence
Area

.Mai

prohibited without p erm ission.
0

0.25

0.5

1.5

2

Figure 2. Map of redds surveyed in the Upper Davis River showing the survey limits for aerial and ground-based redd surveys completed in fall 2001 and fall 2002. The “plus” denotes the up and downstream limits
of the detailed aerial surveys for 2001 only. The “star” denotes the up and downstream limits for both ground surveys. Circles and triangles represent redds observed in fall 2001 and fall 2002 respectively. Red
symbols represent 2001 surveys, Blue symbols represent the 2002 surveys. Surveys completed in 2002 were not analyzed as part o f this thesis and are presented for comparison only.

13

Following these assessments, two additional stream reaches were selected; one in the
upper Swannell River (Fig. 1) and one in Chowika Creek (Figs. 1 and 3). We selected
both the Swannell River and Chowika Creek as sites for this study because both streams
were similar in size, both supported juvenile bull trout and both had areas that were
similar to the high-use spawning area in the Davis River. The upper Swannell River,
however, was only examined in fall 2000 due to a lack of bull trout redds, in addition to
the difficulty in visually identifying potential redds or disturbed areas at this site.
Redd Habitat Characteristics
We completed redd surveys on the ground in fall 2001 in the upper Davis River
to assess habitat conditions within redds. The surveys were completed between river km
42 and 47 (Fig. 2). Over a two-day period, a sub-set of the total redds counted on the
ground were physically surveyed to document redds through a greater length of the study
reach. We assessed habitat within the intergravel environment as well as habitat adjacent
to the redds. Bull trout are the only large salmonid known to spawn in fall in the upper
Davis, Chowika and Swannell watersheds (Unpublished Data, Ministry of Environment,
Prince George, B.C.). Bull trout redds, therefore, were identified by areas o f “cleaned”
gravel with a distinct depression and tail-spill area. “Cleaned” areas on the streambed
were created when periphyton bearing substrate was moved downstream as the result of
digging by the female. Redd digging resulted in patches of “clean looking” gravel
surrounded by olive- or brown-stained substrate.
We surveyed redds on foot, working downstream from the upper part of the study
area. A UTM grid coordinate was recorded with a global-positioning system unit
(Garmin Inc. GPS Map 76S) for each spawning site. A series of habitat measures that

14

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

/// /

w

0.5 0.25 0

0.5

1

1.5

2

Km

Figure 3. Map of redd distributions in high use spawning area in Chowika Creek for
2001-2002. The “plus” denotes the up and downstream limits of the detailed survey.
Circles and triangles represent redds observed from the air in fall 2001 and fall 2002,
respectively.

15

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

included redd dimensions and area, stream depth, stream velocity, stream temperature,
oxygen, pH and conductivity were recorded at each site. To save time during the ground
survey and to extend the distance covered through the spawning reach, we sub-sampled
redds for physical and chemical measures on the basis of aggregations. Where there were
redds separated by only a few metres, we typically sampled only one redd. We noted
redd position relative to the stream margin and prominent cover features such as pools,
under cut-banks, overhead vegetation, and woody debris. We recorded redd dimensions
as the spatial extent o f the disturbed area at redd locations. The disturbed area at redd
sites was typically elongated along the direction o f stream flow forming an ellipse shaped
redd. Where spawning sites were superimposed, we attempted to delineate each
individual redd. Area was approximated from the formula for calculating the area of an
ellipse ABn-4'1, where A is the length of the long axis and B is the width of the redd.
We measured stream depth to the nearest mm at the leading edge of each redd with a
metre stick. We also recorded stream velocity (m-s'1) at the leading edge of each redd.
Stream velocity was collected at a depth 15-cm off the stream-bed to approximate the
nose velocity of a fish holding in the current over the redd site. We used a current meter
counter with a Price AA type current meter head (CMC-20, Hydrological Services Ltd,
AU) to measure flow velocities. We recorded temperature and oxygen concentration
(mg I'1and percent saturation) directly from the stream at each site using portable
electronic meters (Handy M KII, Oxyguard International A/S, DK.; Multiline P4, WTW
GmbH, DE). We calibrated oxygen meters to local elevation and barometric pressure
according to the manufacturer’s procedures. Water samples were also collected from
each site and conductivity as well as pH were measured in the laboratory (Model 744,

16

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Metrohm Ltd, CH). Conductivity and pH meters were calibrated to known standard
solutions before the samples were analyzed.
Intergravel Redd Characteristics
We constructed mini-piezometers (Lee and Cherry 1978) to measure intergravel
temperature, water chemistry, vertical hydraulic gradient and to calculate hydraulic
conductivity at redd sites. Our mini-piezometers followed the design used by Geist et al.
(2000), incorporating the following modifications as suggested by D. R. Geist (pers.
comm; Battelle, Pacific Northwest National Laboratory, Richland, Wa.). We constructed
our mini-piezometers from a 1.5- to 2-m length of 3.1-mm inside diameter, 6.4-mm
outside diameter poly-vinyl-chloride (PVC) tubing attached to a 10-cm section of
perforated polyethylene tubing (6.4-mm inside diameter, 9.5-mm outside diameter)
containing 12, 3.2-mm- diameter holes. The lower end of the perforated well tip was
attached to a machined, brass drive-point and then was wrapped with an 1 lx 10-cm
section of 240-pm Nitex ® screening that was fastened with hot-melt glue.
To collect water adjacent to egg pockets within redds, we inserted our mini­
piezometers into the streambed using the methods described by Lee and Cherry (1978) to
a depth of approximately 30 cm with the first well openings situated at a depth of
approximately 20 cm. Each piezometer was installed as close to the edge of a redd site as
practicable but were not installed directly in a redd tail-spill (location o f egg pockets) to
minimize direct disturbance to bull trout eggs. We collected water samples from the
piezometers with a 125 ml, metal ear syringe with an elongated tip that fit easily inside
the PVC tubing. Where possible, we extracted two volumes of the syringe from each
piezometer before collecting samples in order to clear water stored in the piezometer, as

17

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

well as to clear water from the disturbed area immediately surrounding the piezometer
tip. We collected measures of temperature, oxygen, conductivity and pH as described
above.
Hyporheic Exchange
We collected two measures of hyporheic flow at each site. First, we used vertical
hydraulic gradient (VHG; unit-less measure) as a measure of the extent of upwelling or
downwelling water within each redd that was calculated using the following equation:
VHG = dh/dl
where dh is the difference in pieziometric head between the water level in the piezometer
and the surface of the stream and dl is the difference in height between the stream bed
surface and the first well opening in the piezometer tip. We measured head differential
with the aid of a manometer (Lee and Cherry 1978) mounted to half (50 cm) of an
aluminum metre stick. We mounted a rubber photographic lens-cleaning bulb to one end
of the manometer as a suction device for drawing water from the piezometer and stream
surface into the manometer.
We used a falling head test (Lee and Cherry 1978) to approximate hydraulic
conductivity (K) of the sediment around redds. For the falling-head tests, we used a
graduated, 60-ml plastic syringe as a reservoir. The locking tip was bored out o f the
syringe such that the tubing leading from the piezometer could be inserted inside the
outer sheath of the syringe tip without constricting the inside diameter of the piezometer
tubing. For each test, a timed 50-ml volume of water was allowed to pass though the
through the syringe into the piezometer tube (Lee and Cherry, 1978).
We calculated hydraulic conductivity using the following equation (Hvorslev

18

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

1951; from Lee and Cherry 1978):
q*\n[(mL + D) + l + (mL + D )2]5

K-i, = ----------------------------------------------------------------------h

2 jjL H c

where
D = diameter of intake
L —length to intake
He = pieziometric head (cm)
q = flow of water (cm s' )
m = transformation ratio ( K h/ K v) assumed to be 1 as suggested by Lee and Cherry
(1978).
Estimates of the horizontal hydraulic conductivity Kh were used to calculate the
•7

1

vertical component of water flow (v; specific discharge cm s'1) through the substrate
surrounding bull trout redds using Darcy’s equation (v = K vAh/ A /) where Ah / AI is
VHG measured as the pieziometric head at each redd site (Freeze and Cherry 1979). For
calculations of specific discharge we assumed isotropic flow where the vertical and
horizontal components of hydraulic conductivity are equal.
Temperature
Following aerial redd surveys in 2000 and aerial and ground surveys in fall 2001,
we installed data loggers that automatically recorded temperature (Optic Stowaway Temp
Logger, Onset Computer, Bourne, MA) in the intergravel environment immediately
adjacent to a selection of bull trout redds to monitor incubation temperatures. We placed
the data loggers at approximately 25- to 30-cm depth in the gravel to record temperature
at the depth of the egg pocket. We installed four data loggers at redd sites in the fall of
2000 in the upper Davis River; three were installed near a high-use spawning area, and
one was installed in the Upper Swannell River at the only redd site we located in that
stream that year. In fall 2001, we installed six dataloggers adjacent to redds in the upper

19

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Davis River and two in the upper Chowika Creek for comparison. We programmed
temperature data loggers to record hourly temperatures. For both years of the study, we
installed all loggers at natural redd sites at the conclusion of ground-based surveys in the
last week of September and after spawning was complete. The data loggers were left to
record data until the following summer.
Data Analysis
To examine the relative distribution of redds in the Davis River, we plotted each
spawning location at a scale of 1:20 000, using a B.C. provincial government GIS system
(ArcMap 9.1, ESRI, Redlands, CA). After mapping, we grouped redds into sites
according to their proximity to other redds, where a site was defined as a collection of
redds that were no more than 50-m distant from any redd in the group. We then
calculated descriptive statistics for the mean number of redds at each site. We also
calculated descriptive statistics for habitat measures at each redd, including means,
standard deviation and minimum as well as maximums (SPSS v.l 1.5, Chicago, IL; Excel
2003, Microsoft Corporation, Redmond, Wa.) for all measured and calculated variables.
Stream and intergravel temperatures were collected over the course of two days at
different times of the day for each redd site. As a result these measurements may not be
directly comparable between sites due to daytime heating of the stream. To account for
daytime heating during the surveys, we corrected stream temperatures to a fixed time and
day using hourly stream temperatures recorded by a data logger installed by the B.C.
Ministry of Environment in the upper portion of the high-use spawning area. We used
the corrected measurements for the analysis of redd results. We examined the corrected
stream temperatures statistically for normality and compared redd temperatures to

20

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

corrected stream temperatures using a paired-sample t-test (SPSS). We also used single­
sample t-tests to compare pH, specific conductance with known values in the stream to
test for differences between the stream and intergravel waters. We also tested the
hypothesis that absolute specific discharge within the redds was significantly different
from zero.
To compare thermal regimes at individual redd sites, we tabulated the descriptive
statistics for electronically logged temperature data for each site by month and by season.
We then summed mean daily temperatures cumulatively for each site to calculate
accumulated thermal units (ATU) during the putative incubation period. We used
temperatures from October 30 and the estimated date of emergence of June 14 (see
Chapter 3) for this comparison. These values potentially underestimate the actual total
ATU as the loggers were installed approximately two weeks after the peak spawning
period estimated by Zimmerman and O ’Brien (2000), and the actual date for egg
deposition at each site is unknown.

Results
We observed the highest densities of bull trout redds in the headwaters of both the
Davis River and Chowika Creek (Fig. 2 and Fig. 3). Our results are based on formal
ground-based surveys as well as visual observation made from the air during flights to
and from the study area. We only observed one redd in the Swannell River. The reaches
in the Davis River and Chowika Creek are hereafter described as high-use reaches. The
reach examined in the Swannell River is hereafter described as a low-use reach. Thirtynine bull trout redds were observed on the ground in the high-use spawning area in the
upper Davis River in 2001 (Fig. 2), however, redds were clumped together in a number of

21

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

sites. Sixteen sites were identified containing from one to six redds (x = 2.4; 5 = 1.9).
Within these sites, redds were typically aggregated, where 79% (30) of redds were
located at only 50% o f the sites. The mean number o f redds at sites with more than one
redd was ( x =3.8; s = 1.8). At two sites super-imposition of redds was observed; each of
these locations contained six redds.
Habitat information, including cover type and location relative to the stream
margins was recorded for 37 redds. Redds were categorized as stream edge (< 3 m from
the stream edge) or stream center (> 3 m from the stream margin). All but three of 37
redds (91.7 %) were located near the stream margin. O f the habitat cover types, woody
debris and stable under-cut banks were numerically the most common security cover
types (Table 1). Only four redds (11%) were not associated with obvious cover and all
four were found in open, relatively shallow areas. In contrast, almost 90% of redds were
associated with at least one of three in-stream cover types and 43% of these redds were
associated with two or more cover types.
Depth and stream velocity were collected at 27 redds. All redds were located in
< 1 m of water and the mean depth o f water was 0.43 m (Table 2). Redd area averaged
0.23 m2 (Table 2). Eighteen (47%) redds were sampled for chemical and physical
attributes; descriptive statistics, including sample sizes for each measure are provided in
Table 3. Within each redd, corrected stream and intergravel temperatures were
significantly different (tn = 9.51, P < 0.0001). Mean redd temperature (8.13 °C) was on
average 2.95°C warmer than the corrected mean stream temperature (Tables 2 and 3).
Specific conductance of intergravel water in redds averaged 271.0 pS-cm'1 and was
significantly different from a stream water sample (241.1 pS-cm'1) collected the previous

22

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 1. Cover types associated with redd sites surveyed in the upper Davis River in
2001. Columns are tabulated by the percentage of redds with a score for each category.

Redd
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
31
32
33
34
35
36
37
38
39
Total %

Stable
Undercutbank
1

Over-head
Vegetation
1

Woody
Debris

Pool
1

1
1
1
1
1
1

1

1
1
1
1

1

Instream
Cover
1
1
1
1
1
1
1
1

No
Obvious
Cover

1
1
1
1
1
1
1
1

1
1
1
1
1
1
1

1

1
1
1

1

1

1

1
1
1

1
1
1
1
1

1
1
1
1
1
1
1
1

1
1
1
1
1
1
1
1
1
1
1

1
1
1
1
1
1
1
62.2

1
1
1
1
1
1
1
89.2

1

1
1

1

1

1
1

1
1

51.4

29.7

13.5

10.8

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Reproduced
with permission

Table 2. Habitat and chemical data collected in the stream environment at redd sites in the Upper Davis River in September 2001.

of the copyright owner. Further reproduction

Stream Environment
Stream Stream
Depth Velocity Area temp.
temp.
Site
(m)
(m-s'1) (m2) (°C)
(°C)
0.40
0.23 5.9
5.2
X
0.43
s
0.7
0.15
0.15
0.11
0.6
Min. 0.25
0.04
0.08 4.5
4.4
0.59 7.2
Max. 0.90
0.70
6.6
n
27
27
23
18
18
Table 3. Habitat and chemical data collected in the intergravel environment at redd sites in the Upper Davis River in September 2001.

prohibited without perm ission.

ReddOxygen
Redd Stream
Oxygen
Temp. Temp.
%
Site (°C) Difference Saturation m glI’1
9.4
8.1
2.3
80
X
s
1.4
1.2
9.5
1.1
62
7.4
Min. 4.9
0.0
Max. 10.8
95
11.1
5.3
n
18
18
18
18

K>

Absolute
Absolute
Vertical Vertical
Specific
Specific
Conductivity hydraulic hydraulic
(pS/cm
gradient gradient Discharge Discharge
pH
at 25 °C)
VHG
VHG
v (cm -s1) v (cm-s'1)
8.1
271.0
-0.17
0.25
-0.12
0.30
0.2
0.45
0.41
18.9
0.30
0.10
7.6
247.9
-1.51
0.01
-0.47
0.12
8.4
319.6
0.24
1.51
0.36
0.47
18
18
18
18
18
18

day at the upper end of the high use spawning area (67 = 6.70; P < 0.001). The average
pH of the intergravel environment was less basic (pH = 8.1) than that of the stream
(pH = 8.4) (f17 = -6.22, P < 0.001). The mean concentration of intergravel oxygen was
also significantly less than the stream sample, at 9.4 m gT 1 (tn = -3.156; P = 0.006), but
all redds had concentrations greater than 7.4 m gT1. In contrast the mean oxygen
saturation (80%) in the gravel was not significantly different from that of the stream
sample (tn; = -0.873; P - 0.395). The vertical hydraulic gradient (VHG) was indicative
of down welling at 12 out of 18 measured redds, with a range of 0.24 to -1.5.
Extrapolating to redds adjacent to the sites where intergravel water was sampled (n = 30)
and assuming similar VHG at redds that were immediately adjacent to measured redds,
13 redds exhibited upwelling compared with 17 redds with down-welling. In addition,
the mean absolute specific discharge was high and was significantly greater than zero (tn
= 12.32, P < 0.0001) (Table 3).
Over the course of the two-year study, we collected incubation temperatures
successfully from 11 natural redds (Fig. 4). Mean daily incubation temperatures for all
natural redds during the incubation period (end date June 14) ranged from almost 0°C to
a maximum of 8.3°C (Table 4). Hourly-temperature data are presented as monthly means
(smoothed data) to simplify presentation and comparison o f trends (Figs. 4 and 5).
Mean-monthly-incubation temperatures for all natural redd sites during the incubation
period (end date June 14) ranged from almost 0°C to a maximum of 4.9°C (Fig. 4). In
mid September in both years, redd temperatures ranged from 3 to 5°C. By early
December the temperature of most redds had declined to near 0°C with the exception of

25

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Reproduced
with permission

•

Davis M iddle (1)

of the copyright owner. Further reproduction

-■----- D avis (U,5)
■*

Davis (U,3)
Davis (U,7)

A

Davis (D ,l)

jfc.....D avis (D, 3)
•

Davis East (2)

•

Davis M ain (2)

■

Swannell (1)
- C how ika
- C how ika 2

prohibited without perm ission.

Figure 4. Plot of mean monthly temperatures for selected natural redds collected in the winter 2000-2001 and 2001-2002. Where
hydraulic gradient data were collected, U = upwelling; D = downwelling; The number in brackets for each site indicates the number of
redds present at a site.

K>
O\

Reproduced
with permission
of the copyright owner. Further reproduction

Table 4. Mean temperatures for selected natural redds from the Davis River Chowika Creek and Swannell River collected in 20002002. Temperatures were calculated over times to the eyed stage, hatch and button-up as determined in Chapter 4. Where hydraulic
gradient data were collected: U = upwelling; D = downwelling. (n) indicated the number of redds present at a site. For the periods:
Pre-winter, n =1464 and 61 days; Pre-hatch, n = 2568 and 107 days; Pre-emergence, w = 2135 and 89 days.

prohibited without perm ission.

Chowika

C how ika

(1)
1.19
0.02

(2)
1.62
0.02

78.18
0.01
4.47

52.00

0.03
75.24

-0.05
3.76

-0.01
5.16

X

0.01

0.19

SE

0.00
184.44
0.01
0.17

0.00
49.45
-0.05
0.42

0.03
0.00

1.22
0.03
100.92
0.01
5.25

1.07
0.02
88.70
0.1
4.22

1.54
0.03
100.00
-0.01
7.01

80.81
0.06
5.71

182

215

225

242

to Nov 30
(Pre-W inter)
X

SE
CV
M in.
M ax.

D avis
(D ,l)
1.39

Davis
(D ,3 )
1.48
0.02

D avis

R ed d L ocations
D avis
D avis

(U,7)
1.79
0.03
54.73

(U ,3)
1.89
0.02

D avis
E ast (2)

D avis
M ain (2)

D avis
M iddle

0.88
0.02

1.28
0.02

2.32

1.46

0.02

2.1
5.36

80.73
0.14
3.32

76.40
0.01
4.28

25.13
0.5
3.98

0.03
43.80

0.02
0.01

(U ,5)
3.44
0.01
14.74

0.06
5.12

31.99
0.59
4.34

0.18

0.36

0.79

1.36

0.22

0.00
116.66
0.06
1.17

0.01
142.73
0.06
2.14

0.01
52.31

0.01
36.13
0.68
2.73

0.00
61.52

1.50
0.03

1.64
0.03
88.63
0.06

1.41
0.02
63.93
0.11

1.35
0.01

5.89

5.11

37.88
0.68
5.67

1.08
0.02
92.50
0.14
5.04

294

326

476

174

63.16
0.06
4.78

Sw annell (1)

-0.01
4.57

To March 17
(Pre-Hatch)

CV
M in.
M ax.

209.96
-0.01
0.31

-0.36
1.53

-0.01
0.47

1.15

1.13

0.03
65.00
0.01
2.56

1.73
0.02

2.42
0.03

92.66
-0.14
7.39

44.67
0.01
6.01

235

480

377.16
0.01

-0.28
0.00
-474.64
-4.64
2.69

To Ju n e 14 (PreEm ergence)
X

SE
CV
M in.
M ax.
A T U (O ct 1 to
E m ergence)

K>
^4

1.74
0.03
88.71
-0.01
8.3

215

Reproduced
with permission
of the copyright owner. Further reproduction

6

5

A

Davis Main

□ ■ Davis Middle
A

4

Davis East
Swannell
Chowika

3

2

■n-

prohibited without perm ission.

l
o
l
&

£

&

&

3^

*

Figure 5. Plot of mean monthly temperatures for selected stream flow sites collected in the winter 2001-2002. Open and dashed lines
represent non-selected spawning areas.

N>
00

three sites in the Davis River main stem that also exhibited superimposed or multiple
redds, and one site in the middle tributary o f the Davis River, that contained one redd
(Fig. 4). By early March temperatures began to increase and were in the range of 2.53.5°C by mid June. Where we measured vertical hydraulic gradient in 2001, the warmest
sites (n = 2) (i.e., those above 1.0°C) exhibited up-welling, an indicator of groundwater
influence; whereas the coldest sites were associated with down-welling (n = 2 ), an
indicator of stream water influence. The warmest sites were also associated with the
highest number of redds at a given site (Figs. 4 and 5). In addition to upwelling, each of
the warm sites influenced by groundwater, demonstrated more gradual cooling during the
incubation period and exhibited a more gradual warming trend in spring compared to the
cool sites (Fig. 4). One redd site located in the Swannell River exhibited a variable
temperature regime where hourly temperatures well below 0°C were recorded and where
ice formation would have been lethal to incubating embryos for long periods during
winter. O f four other sites with minimum recorded temperatures below 0°C (Table 4)
three recorded temperatures within 0.05°C below 0°C for long periods of time (days to
weeks) and one recorded temperatures less than -0.05°C to as low as -0.36°C for 15 h.
The warmest average monthly temperatures were encountered in the fall towards the end
o f November and in spring (approximately mid April) after hatch would have been
complete (see Chapter 4). Average redd temperature for the period, winter to putative
hatch, ranged from 0.01°C to 1.36°C with two sites exhibiting average temperatures over
this period of up to 1.15°C and 1.35°C (Table 4). Estimates for total accumulated
thermal units demonstrated a relatively wide range of values from 174 to 480 ATU (Fig.
5). Four sites associated with groundwater discharge demonstrated thermal regimes with

29

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

higher total ATU ( x = 394) than did the remaining sites ( x = 212) (tg = 4.40; P = 0.0023).
The Swannell River site was excluded from this comparison due to prolonged periods of
below freezing temperatures. With the exception of one high use redd site in the Davis
River (D U,5) the other redd locations had substantially lower total ATU’s than stream
bed temperature, indicating an important difference between stream and intergravel
temperatures within sites (Fig 6 ).

Discussion
Bull trout home to specific areas to spawn that are often found in headwater
reaches (Fraley and Shepard 1989; Rieman and McIntyre 1995; McPhail and Baxter
1996; Baxter and McPhail 1999; Bahr and Shrimpton 2004) and our results are consistent
with these observations from other watersheds. The determinants of site use at the reach
scale and site scale are, however, less clear. For the purpose of this discussion we
differentiate between habitat use versus habitat selection as it is unclear whether bull
trout are returning to specific sites as the result o f precise homing to a site where
spawning was previously successful or whether bull trout are using specific indicators of
habitat quality to select good nest sites.
In other studies, groundwater has often been described as a driver of spawning
selection (e.g., McPhail and Baxter 1996; Baxter and McPhail 1999) and this has been
observed for other char species and salmonids in general. These species include: chinook
(Oncorhynchus tschawytcha), (Geist 2000); chum (O. keta) (Leman 1993); brown trout
(Salmo trutta) (Hansen 1975); eastern brook char (Salvelinus fontinalis)-, (Webster and
Eriksdottir 1976; Curry et al. 1995; Blanchfield and Ridgeway 1997; Curry et al. 1995);
Arctic char (S. alpinus) (Cunjak 1986); and bull trout (Baxter and McPhail 1999; Baxter

30

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Reproduced
with permission

Davis Middle (1)
Davis (U,5)

of the copyright owner. Further reproduction

Davis (U,3)
Davis (U,7)
Davis (D ,l)
Davis (D, 3)
Davis Main (2)
Davis East (2)
Swannell (1)
Chowika
Chowika 2
Davis River (RB)
Trib B (RB)

prohibited without perm ission.

Figure 6 . Accumulated centigrade thermal unit (degree day) profiles calculated for selected natural redds and stream flow sites
collected in the winter 2000-2001 and 2001-2002. Where hydraulic gradient data were collected: U = upwelling; D = downwelling.
(n) indicated the number of redds present at a site; RB = river bottom.
u>

and Hauer 2000). Groundwater influence is, however, only one component o f habitat and
many other physical factors such as substrate, oxygen, and gravel permeability are also
important to salmonid incubation success (Chapman 1988). It is important, therefore, to
understand at which scale habitat features are important and at what level they affect site
selection and use (Baxter and Hauer 2000).
We began our study by assessing the distribution of bull trout within several large
watersheds of the Williston Reservoir. Our results confirm previous work (O’Brien and
Zimmerman 2000) that large migratory bull trout move to and spawn in smaller tributary
streams. At the sub-basin or reach scale we found that bull trout were using headwater
depositional areas influenced by relatively warm groundwater during the incubation
period. Within reaches, several micro habitat features including high hyphorheic flow
rates, relatively warm temperatures and security cover appeared to be important features
that were often associated with spawning sites.
Habitat Use at the Reach Level
In the Davis River and Chowika Creek, redds were only observed at high
densities in two specific reaches near headwater areas (Figs. 3 and 4). Habitat features at
the reach-scale in both high-use areas are consistent with high-quality spawning habitat
that has high levels of hyporheic exchange and thermal signatures moderated by
groundwater (McPhail and Baxter 1996; Baxter and McPhail 1999; Baxter and Hauer
2000). The lack of spawning in a wider range of habitats in our study area suggests that
bull trout were preferentially using specific reaches with specific characteristics such as
warm thermal profiles and high rates of hyporheic exchange. Considering the limitations
of our investigation, it is possible that some bull trout were using other portions of both

32

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

watersheds as we did not examine in detail all small tributaries and many lower reaches
using ground-based surveys. Based on aerial and ground-based redd counts, results from
a previous telemetry study (O’Brien and Zimmerman 2000), and redd surveys in the
Davis River subsequent to this project (Ministry of Environment and Peace Williston
Fish and Wildlife Compensation Program, unpublished data, Prince George, B.C.), we
are confident, however, that the majority of spawners were using the reaches examined.
Within the Davis River, spawning sites were located mainly in a 5-km long valley
segment downstream of the confluence of three similar sized tributaries (confluence area)
(Fig. 3). The valley bottom in this mainstem reach is up to 1200-m wide, the channel is
sinuous and low gradient (Fig. 3), and the substrate is composed primarily of gravel and
cobble with little fine sediment present. Stream temperatures in this reach were relatively
warm (Figs. 4 and 5) during the winter incubation period suggesting a strong
groundwater influence. Another spawning area with a high density of redds, in the
Chowika Creek drainage, also had similar geomorphic characteristics to that o f the highuse area in the Davis River (Fig. 2). Stream temperatures in the Chowika Creek reach
were on average cooler than the main spawning area in Davis River (Fig. 2) and were
similar to temperature measurements collected in the east tributary reach in the Davis
River (Figs. 5 and 6 ). This observation is consistent with the idea that absolute
temperature is not the only factor influencing habitat use at the reach scale.
In the Davis River east tributary, aerial redd counts were intermediate to counts in
the middle tributary and the main-stem reach (Fig. 3). Differences in habitat use within
spawning areas in the Davis River and between spawning areas in the Davis River and
Chowika Creek may be explained by differences in habitat quality and habitat

33

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

availability. Habitat quality could limit habitat use where variables such as temperature
are adequate for incubation but not optimal for survival or growth. For example, the east
tributary within the Davis River is much more laterally confined and is higher in gradient
than the main-stem Davis River and both of these features together could limit hyporheic
exchange with the effect of reducing potential groundwater influence on spawning sites
(Baxter and Hauer 2000). Reduced groundwater influence could explain differences in
the level of spawning between reaches in the Davis River if bull trout are indeed selecting
warmer groundwater-fed areas. Thus, differences in use between reaches in different
stream systems may be explained by the availability o f quality habitats within a given
watershed. The high-use reach in Chowika Creek is to our knowledge the only highquality site used for spawning in that system. Chowika Creek is generally colder than the
Davis River, however, and the Chowika Creek spawning site is higher in elevation (1120
m vs. 940 m). If warmer incubation habitat were available elsewhere in the Chowika
Creek drainage as is the case in the Davis River system, lower levels of use may
otherwise have been observed at the site we studied.
Other authors have indicated that bull trout appear to select groundwater-fed areas
for spawning that are relatively warm (e.g., Baxter and McPhail 1999; Baxter and Hauer
2000; and for a review see McPhail and Baxter 1996) and one account noted selection for
a spring-fed site (Boag and Hvenegaard 1997). Based on patterns of use, our results are
generally consistent with these observations, however, two of our study reaches were not
well used despite both having relatively warm thermal profiles indicative of groundwater
influence. In the Davis River middle tributary (Fig. 3), we observed an area with a direct
groundwater input to the stream that remained ice-free even at very low winter

34

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

temperatures (-25°C). The lack of spawning within this reach seems counterintuitive;
however, our observations may be explained by higher levels of fine sediments,
compared with the other nearby reaches. The deleterious effects o f fine sediments on
salmonid eggs and incubation are well known (reviewed in Chapman 1988).
Accumulations of fine sediment in redds can reduce the amount of interstitial space
available for incubating eggs and can result in lower egg and embryo survival through
reduced oxygen uptake and reduced rates of metabolite transport.
Habitat characteristics at the reach scale, including stream sinuosity, gradient and
bed characteristics in both the Davis River and Chowika Creek were qualitatively similar
to other spawning areas used by bull trout in Northern B.C. (e.g., Pillipow and
Williamson 2004, C. Williamson 2006, unpublished observations, Ministry of
Environment, Prince George, B.C.). Baxter and Hauer (2000) also recorded similar
spawning areas in Montana, USA, where bull trout typically used bounded alluvial valley
segments constrained by downstream geomorphic knick points. These areas can provide
zones of high hyporheic exchange (Baxter and Hauer 2000) that in turn moderate the
thermal regimes of incubation habitat. Warmer intergravel temperatures during winter
could also increase embryo survival by preventing the formation of anchor ice, which
leads to high embryo mortality rates (Baxter and McPhail 1999; Chapter 3). In our study,
we observed that areas of alluvial gravel deposits (including both spawning and non­
spawning areas) characteristically had free flowing water beneath surface ice, during cold
periods when air temperature approached -25°C, whereas laterally confined reaches in
downstream areas often had extensive formations of anchor ice and little or no surface
ice.

35

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Redd Habitat Characteristics
As described above, many populations of salmonids apparently seek specific
habitats for spawning. Consequently, redd distributions are often clumped and limited to
specific sections of available stream habitat. For bull trout the selection of spawning
habitat has been linked to groundwater inputs at or near the spawning sites (Boag and
Hvenegaard 1997) and differences in habitat use may be categorized and explained by
factors related to habitat abundance, habitat quality and behavioural traits related to
habitat selection (Essington 1998). In the Davis River, redds were typically aggregated
within reaches despite apparently high levels of alternate available habitat (Figs. 3 and 4)
and we observed two sites in the Davis River where redds were superimposed. Gradients
in habitat quality could explain differences in habitat use where quality has an effect on
survival or growth of either adults or their progeny.
In the Davis River strong hyporheic exchange was evident and intergravel flow
and oxygen all appeared uniformly high at all redd sites. Intergravel flow rates
(measured as absolute specific discharge) ranged from 0.12 to 0.47 cm-s'1(Table 3).
These observations are within the ranges provided by Freeze and Cherry (1979) for clean
sand and gravel substrates. In the Davis River oxygen levels within redds were not
always at saturation, but all sites had concentrations above 7.4 g T 1 (x = 9.4 m gT 1)
(Table 3), which is within the range of oxygen levels considered optimal for salmonids
(Barton and Taylor 1996). Oxygen levels remained high in a related field incubation
experiment (Chapter 3) at nearby sites where we were able to collect additional oxygen
measurements during winter.
Based on thermal and chemical measures, groundwater influence was evident at

36

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

redds examined in this study (Table 3). Oxygen levels, pH and conductivity all indicated
that surface waters were mixing with groundwater in the hyporheic zone (Freeze and
Cherry 1979). Interstitial oxygen levels were lower and pH and conductivity were on
average higher than surface waters adjacent to each redd. Similarly, intergravel
temperatures were warmer than those of the stream for each redd measured. Measures of
vertical hydraulic gradient, however, did not always indicate upwelling. In fact, only 13
redds exhibited upwelling compared with 17 redds where down-welling was measured.
Three of the warmest upwelling sites (Figs. 5 and 6 ) in the Davis River also had the
highest number of redds which suggests that some bull trout may be more successful in
habitats with these characteristics. In contrast to Baxter and McPhail (1999) who
observed exclusive use of an upwelling site, our results suggest that bull trout can use a
wider range of habitats for spawning where localized upwelling may not be the most
important factor for site use. This hypothesis is supported by the work o f Baxter and
Hauer (2000) who found that bull trout in Montana generally spawn in areas of extensive
upwelling at the reach scale. At the habitat scale spawning sites were predominantly
located in transitional bedforms that exhibited localized downwelling. Both our results
and those of Baxter and Hauer (2000) support the hypothesis that groundwater influence
at the reach scale scale is a key factor influencing the use of bull trout spawning habitats.
Factors Moderating Site Use
In addition to site suitability, predation risk is likely an important factor
influencing use of spawning microhabitats, particularly in northern environments that
lack anadromous sources of protein for large carnivores. In the Davis River and Chowika
Creek, redds were most often situated near stream margins and the majority of redds in

37

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

the Davis River (89%) were associated with at least one type of in-stream security cover
(Table 1).
Bull trout that reproduce in clear headwater streams in northern B.C. and
particularly spawners that use stream margin habitat can be highly visible to predators.
The use of sites associated with cover such as woody debris or undercut banks may help
to reduce predation risk. Considering the relatively shallow average redd depth (0.43cm) in the Davis River and the relative proximity to the stream margins for most redds,
predation risk in the absence of cover may be considerable. Although we did not observe
any direct predation in our study, grizzly bears (Ursus arctos horribilus) were observed
near the spawning areas on several occasions and several adult bull trout captured for a
related incubation study (Chapter 3 and 4) had large dorsal wounds and scars consistent
with bites from large predators. Predation on bull trout during spawning is not unique to
our study. Bahr (2002) also observed mortality during spawning in a bull trout
population in the Morice River, B.C. watershed during a telemetry study. It was unclear,
however, whether these fish died as the result o f spawning and were consumed dead or
were captured live by predators. In another example, Chandler et al. (2001) reported
80% mortality of small bull trout that were radio-tagged in two streams in Oregon during
the spawning period where mammalian predators were an important factor. Predation
intensity can be high for other salmonids during spawning and it is thought to influence
habitat use and body morphology in these circumstances. Quinn et al. (2001) examined
the relationship between age at maturity, body size, water depth in spawning streams and
predation intensity for several populations of sockeye salmon. Their results show a
positive relationship between mean water depth among streams and body depth where

38

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

predation by bears was substantial.
Lack of Habitat Use
It is notable that bull trout spawning activity was observed at very low levels (< 1
redd per km of stream) in two of the warmest groundwater influenced reaches we studied.
This result was unexpected considering the potential benefits of warmer temperatures
associated with upwelling groundwater which includes accelerated growth and enhanced
survival for incubating fish (e.g., Garett et al. 1998). The seasonal temperature profile of
the Davis River middle tributary is indicative of a stream that is heavily influenced by a
relatively stable source o f groundwater (Fig. 6 ). In winter, stream-flow in this reach (Fig.
2) can be more than 2°C warmer and in summer is typically cooler than adjacent
tributaries as the main-stem section of the Davis River. It is important to note that this
reach also had higher levels of coarse woody debris than other reaches which resulted in
more potential cover. It is not clear why fish are not spawning at higher densities in this
reach. One explanation for the lack of use may be that bull trout are avoiding the higher
winter temperatures associated with sites in this reach. At warmer sites, embryonic
development may proceed at too rapid a rate for successful incubation. Brannon (1987)
postulated that spawning date and incubation period for several species of Pacific
salmonids is timed to take advantage of increases in spring primary production. The
range of suitable incubation temperature may be limited for northern populations of bull
trout. Relatively high incubation temperatures may result in premature emergence from
the gravel or premature depletion of yolk. Depletion of yolk combined with reduced food
availability prior to spring could reduce juvenile survival thus reducing the fitness of
parents that spawned in the warmer habitat.

39

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Specific habitat features, such as low intergravel flow rates and lower oxygen
availability due to higher levels of fine sediments, may also be important factors affecting
the level of use. One of the few redds observed in this reach was located on top of a
disturbed area that was used for a field incubation experiment in the previous year
(Chapter 3). It is interesting to note that during a walking survey of this reach in the fall
of 2000 , we did not locate any redds, but many juveniles were plainly visible which
suggests that this type of reach and access to it may be important for other life stages.
Similarly, we observed low levels of spawning activity in the upper Swannell
River where the lack of use may be related to the influence of groundwater, however, this
is more difficult to explain than for the Davis River and use may be influenced by factors
other than habitat quality. The upper Swannell River (Fig. 1) has many similar
characteristic to that of the high-use spawning area in the Davis River including: stream
size and width, high sinuosity, low gradient, clean gravel, substrates and groundwater
influence, however, only one redd was observed. Unlike the other study reaches used for
spawning, we observed few mature adults dining our reconnaissance surveys in the
Swannell River and we found these fish holding at the downstream end o f the study reach
(J. T. Zimmerman and C. J. Williamson 2006 unpublished observations, Ministry of
Environment, Prince George, B.C.). Both groundwater and stream temperatures in the
Upper Swannell were cooler than the Davis River during spawning season and winter
temperatures were nearly as warm as the most selected sites in the Davis River. Thus,
cool temperature during spawning may also be a factor in the lack of use.
It is also possible that the upper Swannell has been used more extensively for
spawning than we observed. First, we noted that redd identification was more difficult in

40

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

the Swannell drainage due the substrate colour as well and the lack o f periphyton growth.
To confirm this general observation we disturbed areas of gravel and found these patches
very difficult to see compared with the other watersheds where redds and disturbed areas
were highly visible. It is, therefore, possible that we did not observe some redds that
were actually present. Second, as part of our initial work to identify study sites, we
electro-shocked a number of areas in the upper Swannell reach while looking for juvenile
bull trout. We were able to find small numbers of bull trout fry at one site, suggesting
spawning had taken place in that reach, however, we failed to find fry at other locations.
Lack of spawning at sites influenced by groundwater in this reach may also be linked to a
potentially smaller population with fewer spawners. Unlike the Davis River and
Chowika Creek, there is substantially more development related to forestry in the valley
bottom of lower Swannell River which includes a relatively high density of roads (Fig.
1). This development could have resulted in better access to bull trout staging areas for
anglers. Bull trout are aggressive-piscivores that are highly vulnerable to capture with
fishing gear; thus road access, due to forest development as well as proximity to forestry
camps and a small community, may have led to impacts on the population of spawners
resulting in fewer redds. For example, Ripley et al. (2005) have linked access to declines
in bull trout abundance in the Kakwa River drainage in Alberta, Canada. In the Swannell
drainage, bull trout could have been using other sites, with more suitable habitat
characteristics within the drainage. During our initial examination of the Upper Swannell
many of the smallest tributaries were not surveyed.

41

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Summary
The use of specific areas influenced by groundwater with specific thermal
signatures can help to maximize reproductive success of fish by limiting the effects of
stochastic environmental process that act to lower survival during incubation. At the
reach scale, our results are consistent with previous work that indicates bull trout spawn
in valley segments influenced by groundwater that have zones that exhibit high rates of
hyporheic exchange. We found evidence of localized groundwater influence at the
habitat scale, where bull trout were spawning in a variety o f habitats with localized warm
upwelling groundwater or downwelling streamflow. Finally we observed several habitat
features common to site used for spawning, including high intergravel-flow rates, high
levels of interstitial oxygen and proximity to security cover.

42

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

CHAPTER 3: FACTORS AFFECTING GROWTH AND SURVIVAL OF
LARVAL BULL TROUT IN PRISTINE NORTHERN STREAMS.2

Abstract
The importance of the yearly temperature regime as well as other habitat variables for
incubation and development of bull trout is not well known. It is also not known how
small deviations from the thermal and chemical regime characteristic of bull trout
spawning habitat may affect incubation success and ultimately juvenile survival. To
understand these relationships, we examined the influence of intergravel temperature, pH,
conductivity, intergravel flow, and oxygen on bull trout survival, development, embryo
size and yolk utilization. We assessed incubation success and growth using fertilized bull
trout eggs contained within perforated plastic capsules. The capsules were buried at
typical redd depths in three tributaries draining into the Williston Reservoir, B.C., where
bull trout are known to spawn. Survival during incubation was independent of all
physical and chemical measures we examined at both used and non-used sites and habitat
quality was relatively high at all sites. In contrast, thermal regime was an important
determinant of the rate of bull trout development and size. Warmer temperatures
associated with groundwater influenced sites that were not used for spawning
significantly advanced development, size and yolk utilization relative to used sites. Thus,
differences in thermal regime between sites are a likely explanation for observed patterns
of use.

2 Throughout this chapter I use the first person plural to acknowledge the contribution of others to this
work, which will be submitted for publication with the authorship of C.J. Williamson, J.T. Zimmerman,
and J.M. Shrimpton.

43

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Introduction
Salmonids typically reproduce in tributary streams with clean gravel substrates,
often after a substantial migration from feeding or rearing areas. During migration, adults
can encounter a wide variety of potentially suitable spawning habitats. Site selection
tends to be precise with year-after-year fidelity for specific areas or habitats within
watersheds, either by individuals, successive migrants or successive generations (Fleming
and Reynolds 2004). Behaviours involved in site selection have likely evolved as an
adaptive response to maximize reproductive fitness in areas with variable habitat
conditions (Brannon 1987; Hendry et al. 1998; Kinnison et al. 1998; Fleming and
Reynolds 2004). Site selection can involve homing to specific locations within a
watershed (e.g., Dittman and Quinn 1996), selection of specific channel locations
(Chapter 2), or selection of specific habitat conditions such as intergravel flow and
temperature (Baxter and McPhail 1999). Additionally, localized site selection may be
influenced by environmental or chemical cues such as oxygen levels, water temperature
or groundwater inputs (Chapter 2). Habitat and other factors influencing survival and
growth during incubation have been particularly well studied in salmonids (reviewed in
Chapmanl988). Investigations of habitat quality for incubation have historically focused
on stream sediments, stream flows and related land management practices (Reiser and
Bjomn 1979; Chapman 1988). More recent research has integrated the role of fluvial
processes and geomorphic structure as mediators of biological processes (Baxter and
McPhail 1999; Baxter and Hauer 2000; Geist 2000) and has highlighted the importance
of sub-surface groundwater, including surface water chemistry and hyporheic flow, and
hyporheic exchange for incubating salmonids.

44

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

The quality of incubation habitat can affect the survival and growth of juvenile
fish (see Chapman 1988). Due to the difficulty of winter field research, however, few
studies have investigated in situ the effects of habitat on the survival and growth of
incubating salmonids in cold environments. To better understand the role of subsurface
fluvial processes on the incubation of salmonids in a cold northern environment, we
investigated the role of groundwater and five related habitat parameters on the incubation
success of bull trout in three northern British Columbia streams. Bull trout appear to
select specific microhabitats within particular stream reaches for reproduction (Baxter
and McPhail 1999; Baxter and Hauer 2000); however, it is not clear what advantages and
features the selected sites confer in terms of survival and growth of eggs and alevins. In
this study, survival, somatic growth and yolk utilisation of bull trout eggs and alevins
were examined in two field incubation experiments during the fall and winter incubation
periods of 2000-2001 and 2001-2002. We examined differences in survival and embryo
size in reaches that were used frequently and sites not used or infrequently used by bull
trout for spawning to better understand the importance of habitats selected. We measured
habitat parameters including intergravel flow (specific discharge and vertical hydraulic
gradient), temperature, oxygen, pH and conductivity as each of these factors could
influence survival, embryo size and yolk utilization, and may be used as cues by
spawners for redd site selection. Comparison of size and survival among reaches
exhibiting different habitat characteristics may help to explain patterns of site use or site
selection. Identification of these features could help in the management and protection of
spawning habitat.

45

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Materials and Methods
Study Area
Over the two-year study, experimental sites were located at six reaches of three tributary
streams to the Finlay Reach of the Williston Reservoir (Fig. 1). For this experiment, sites
were located in the Davis River, Chowika Creek and Swannell River. A more detailed
description of the study location is provided in Chapter 2.
Gamete Collection
We collected eggs and milt from adult bull trout we captured by angling in the
Davis River. None of the fish captured showed any interest in, or aggression to lures or
bait, therefore all the fish were caught in shallow water, by “snagging” of a pectoral fin.
With the exception of one male, all spawners were from the upper portion of the Davis
River watershed near a spawning area with relatively high levels of use. In 2001, we
captured one male in the middle reaches of the Davis River. In 2000, we captured two
females and two males and in 2001, we captured three males and three females. To test
for local adaptation including differences in size and survival relative to specific
environmental conditions in spawning streams, we attempted to collect additional
spawners from Chowika Creek and the Swannell River for the second experiment in
2001. Logistical constraints associated with flight bans following the events of September
11, 2001, however, prevented us from collecting additional gametes. We collected
several fish in Chowika Creek, but the females were near the end of spawning and had
few eggs.
Prior to artificial spawning, we anaesthetized each fish with a 10-ppm solution of
clove oil and ethanol (1:9 clove oil to ethanol) until they exhibited a loss of equilibrium

46

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

and slow opercular movement (Anderson et al. 1997; Prince and Powell 2000). To
prevent activation of sperm or eggs, we removed excess water on the outside of each fish
with a dry towel, before gametes were removed. The gametes were gently massaged by
hand from the fish into cool, clean plastic containers. Following gamete collection, we
allowed each fish to fully recover in river water before release. We mixed eggs and milt
for each family cross separately in clean plastic buckets, without water. After
approximately 2 min, we added river water to the eggs, and then left them for 5 min. We
then washed excess milt from the eggs and then left them to water harden for 30 min.
Egg Capsules
We placed the fertilized eggs into plastic incubation capsules, and then deployed
the capsules by helicopter to the incubation sites within 24 h after fertilization. Our
capsule design (including size) was similar to that used by Baxter and McPhail (1999)
and Cope and Macdonald (1998). We used plastic, a less thermally conductive material,
instead of stainless steel to inhibit and arrest the unnatural formation of anchor ice in
redds. We constructed the capsules (13-cm long by 5.5-cm diameter) from perforated,
extruded food-grade polyethylene tubing ( 1-mm by 3-mm slots) that were covered at
each end by a perforated, polyethylene-plastic cap (T-series CAPLUG®). We manually
perforated each cap with at least 60, 1,5-mm diameter holes. We added clean, sieved
gravel < 3-cm diameter to each capsule to provide an internal substrate consistent with
natural redds. In 2000, we placed approximately 30 eggs (counted by weight) into each
capsule. In 2001, we placed 40 or 45 eggs, counted in batches with a counting paddle,
into each capsule where sample size was dependant on the number of eggs available
within each family group.

47

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Incubation Sites
We used a nested experimental design for the incubation sites. In experiment one
(fall-winter 2000-2001), we used three experimental reaches, two in the Davis River and
one in Chowika Creek for incubating eggs (Fig. 1). In the second experiment (fall-winter
2001-2002), we used six reaches, three in the Davis River, two in Chowika Creek and
one in the Swannell River (Fig. 1). We selected reaches based on two broad categories of
use by bull trout-spawners for both years: high-use and low-use. High-use sites were
located among redds in the Davis River and Chowika Creek at sites that were determined
to be important spawning areas following reconnaissance work for our synoptic study of
redd habitat (see Chapter 2). Low-use sites were selected in areas not used for spawning
where the habitat appeared qualitatively similar to highly-used sites. For example, one
low-use site in the Davis River (Davis River middle tributary) was selected as it was
located in the vicinity of the main spawning area. Three sites were located in reaches that
were used in year one or were adjacent to sites used in year one.
In each reach, we excavated three artificial redds that were within 3 m of each
other (i.e., redds nested within reaches). We excavated the artificial redds using a shovel
so that fine sediments were carried downstream with the current and each redd had an
upstream depression followed by a tail-spill mound, similar to a natural redd. We buried
the capsules within the tail-spill mound so that a minimum of four capsules containing
the eggs from each of up to five families (i.e., up to 20 capsules) were buried in clusters
at approximately 25-cm depth in each artificial redd (i.e., capsules nested within redds).
We selected this depth to be representative of eggs within a natural egg pocket within a
redd and this depth is within the range of redd depths observed for large-bodied (65 to 85

48

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

cm) salmonids (Devries 1997; Steen and Quinn 1999). We collected two incubation
capsules for each family cross at each of two sampling intervals. In experimental year
one, we selected the first sampling period (December) to correspond with the period
shortly after hatch. We selected the second date to roughly correspond to the end of the
pre-emergence period, based on information from Gould (1987). The sampling dates
were adjusted in year two to better correspond with the goals to sample alevins just post­
hatch and well past post hatch as outlined for year one. Following collection, we
transported the capsules by helicopter in water filled tubs (to prevent freezing) to a field
laboratory where we removed the eggs and alevins from the capsules and recorded
survival. Once counted, we rapidly froze the live eggs and alevins on dry ice to
simultaneously euthanize and preserve them for growth analysis in the laboratory.
Alevin Survival, Size and Yolk Utilization
We enumerated live and dead eggs or alevins in each capsule on each collection
date. To compare embryo sizes among stream sites under a range of physical conditions
adjacent to natural spawning sites and areas that appeared suitable, but where bull trout
were not found to spawn, we collected measurements of body size and yolk utilization
from all alevin and embryo samples. Before collecting each measurement, we dissected
embryos that had not hatched from the eggs. We recorded embryo and alevin standard
lengths to the nearest 0.01 mm under a dissecting microscope using Vernier callipers.
We used an analytical balance that measured to 0.0001 g to determine wet and dry
weights. Embryos and alevins were re-frozen after we collected lengths and weights to
keep the alevin and the yolk intact, while the yolk material was dissected away from the
alevin. We dried tissue samples individually in aluminum weighing boats for a minimum

49

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

of 24 h at 60°C. After drying, we stored the tissue in a sealed dessicator containing
CaS04 until dry weights could be collected.
Temperature, Hyporheic Flow and Artificial Redd Chemistry
In the field, we installed data loggers to record temperature (Optic Stowaway
temp logger, Onset Computer, Bourne, Ma) within the hyporheic zone at each artificial
redd at a depth of approximately 25 to 30 cm (capsule depth) to continuously monitor
incubation temperatures. We also placed a single data logger on the river bottom at each
site to monitor stream temperature. We programmed each data logger to record
temperature on an hourly basis for the duration of the incubation period.
In 2000, we attempted to measure intergravel chemistry using rigid standpipes
with screened tips installed in the center of each redd, however, this method proved
unsatisfactory due to cold winter conditions (-20 to -25°C in December, 2000) and
shifting river ice. Several of the standpipes were knocked down by ice while many others
were difficult or impossible to sample as ice-plugs had formed within the tubes at the
level of the stream surface. Following the first egg collection period in 2000, we
removed the remaining standpipes to prevent uncontrolled disturbances to the incubating
eggs. For the second egg collection period in 2000, we attempted to install mini­
piezometers using Lee and Cherry’s (1978) design. This design, however, also proved
problematic under cold conditions and data collection was again limited.
In 2001, we successfully installed two mini-piezometers(Geist 2000) (see Chapter 2
methods) within each artificial redd, one at 30 cm, and one at approximately 50 cm below
the redd. The use of flexible vinyl mini-piezometers facilitated storage of the
piezometers below the water surface near the stream bed between field trips; this

50

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

prevented the piezometer tube from freezing into surface ice during extended periods of
cold weather. To monitor intergravel temperatures below redds and to aid the
identification of strong groundwater influence in the vicinity of the artificial redds, we
installed one additional temperature logger at a depth of 50 cm within each artificial redd
cluster.
At each artificial redd, we recorded intergravel water temperature, pH, specific
conductance, oxygen, vertical hydraulic gradient and hydraulic conductivity (see Chapter
2 ) on three occasions: two weeks after egg capsules were buried, when egg capsules were

retrieved in mid-winter, and when egg capsules were retrieved in late winter. We also
recorded stream water chemistry for the same chemical parameters during each site visit
to allow comparison with hyporheic water.
Data Analysis
Survival and Size
To examine estimates of survival and measures of size for each field experiment
(i.e., each year), we completed separate ANOVA’s for each experimental year, and for
each dependant variable (i.e., survival, length, dry-yolk weight, dry-body weight and total
dry weight) (Sokal and Rolf 1995). Factors in the models included sample date, family or
parent, reach, redd and capsule. Capsule was nested within redd and redd was nested
within reach. Within this group of analyses, to understand the effect of parent or family
on each dependant variable, we completed separate ANOVA’s with family or with parent
u sin g the nested designs d escrib ed above. In y e a r one each fam ily w as u sed in each reach

(not nested), in year two family was nested within reaches due to limited number of egg
relative to the number of sites. To understand the effect of parent on size, we tested the
interaction between male and female for year one of our experiment in ANOVA’s that
51

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

included both male and female as factors in the tests. This test was not possible for year
two as parent was nested within reach (i.e., not all combinations of parent and reach were
possible). For analysis of survival and size, we completed all ANOVAs using the general
linear model (GLM) procedure in Minitab (v.14.1, State College, PA) and each factor
was treated as a fixed variable, as each alevin was sampled destructively. We used the
adjusted sum of squares to test significance. To examine the differences between
dependant variables (survival, length, yolk and tissues weights) at each experimental
level (date, reach, redd, capsule and family) for each test, post hoc comparisons were
completed when significance was found for any given factor, using the Tukey-Kramer
multiple comparison procedure in Minitab. We used the Tukey-Kramer procedure as this
method for multiple comparisons is suitable for un-balanced experimental designs and is
more conservative than other similar procedures (Sokal and Rolf 1995). We screened all
data before attempting any analysis. Screening involved examination of outliers,
erroneous values and transcription errors. As a result of the low sample sizes for some
experimental factors, following screening only data that were obviously erroneous were
deleted from the datasets. Thus, we only excluded data that were transcription errors. As
part of each procedure, we visually checked all ANOVA models for normality and
heterogeneity using the output provided by Minitab, which included: normal probability
plots, histograms and scatter plots of the residuals for each model. Percent survival data
were arcsin a / x transformed for all tests (Tabachnick and Fidell 2001). All ANOVA
tests met the assumptions of normality and heteroscedasticity.
Habitat
Chemical parameters (specific conductivity, oxygen and pH) and specific

52

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

discharge were compared among sample units using nested ANOVA including use, reach
and piezometer as fixed factors within each ANOVA model. Use was used as a
categorical factor based on either sites used or not used for spawning. Reach was nested
within use and piezometer was nested within reach. Piezometer measurements were
collected at two different substrate depths below one-of-three redds within each reach.
Piezometer measurements collected on any given date were therefore treated as replicates
within a reach. Our sample sizes for measurement of our dependant factors was too low
to support a repeated measures ANOVA design; therefore, to better understand
groundwater influence and chemistry within the intergravel environment in each study
reach, we assessed differences between surface waters and piezometers at two intergravel
depths using a constructed variable. Thus the factor “piezometer” was calculated as the
difference between each piezometer measure and the surface which allowed us to better
understand within reach variation as the differences between hyporheic and surface
waters. For each dependant variable we completed separate ANOVAs for each of the
three sample dates to better understand how chemistry and specific discharge varied
within reaches and between sites used and not used for spawning. Next, we used a
similar approach and design described above to examine differences among factors,
including site use, reaches and piezometer depth through time (September and April).
Thus, for comparisons through time we subtracted the last measurement collected for
each site (April) from the first measurements collected (September). For both within
reach and between time tests, this approach resulted in non-normal data that were
successfully transformed (except pH) using natural log-skew transformations that we
completed using Stata 9.2 (Stata Corp., College Station, TX). We used Kolmogorov-

53

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Smimov goodness of fit tests on ungrouped data to test for normality following
transformation (Sokal and Rolf 1995). For simplicity of presentation, we used data
means to report our results for comparisons between reaches and through time. To
understand temperature differences between redds and among reaches, we plotted
temperature data for each redd as a time-series and analysis was restricted to qualitative
description and qualitative contrasts of the profiles. All ANOVA tests met the
assumptions of normality and heteroscedasticity. We used the adjusted sum of squares to
test significance. All data were screened for erroneous values, outliers and as part o f each
procedure, we visually checked all ANOVA models for normality and heterogeneity
using the output provide by Minitab, which included: normal probability plots,
histograms and scatter plots of the residuals for each model.

Results
Hatch and Survival
In both years, hatch occurred during the month of March for reaches that were used for
spawning or had similar temperature regimes to the sites used for spawning (Table 5). At
two warmer sites that were not used for spawning (Swannell River- SI and Davis River D2), hatch occurred earlier than all other sites (Table 5). Due to experiment-wide
differences in survival between year one and year two, we first present descriptive
statistics for survival for both years. In year one survival was generally high (Fig. 7) and
individual egg capsule survival ranged from 0 to 100 % with a mean survival rate of
64.0% (n = 122, s = 23.7). In year two survival was low (Fig. 7). Individual capsule
survival rates ranged from 0% to a maximum o f 52.5% and the mean survival rate was
9.6% (n = 193 s = 11.5).

54

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Reproduced
with permission

Table 5. Percent hatch and accumulated thermal units (ATU) for all capsule sites in both experimental years. Sites are categorized by
level of use and thermal profile.

of the copyright owner. Further reproduction

Reach

Year Two

Year One
Days Post
Fertilization

Dec 12, 2000

Mar 22, 2001

95

195

Mar 01, 2002
169

Apr 29, 2002
228

prohibited without perm ission.

% Hatch

ATU

% Hatch

ATU

% Hatch

ATU

% Hatch

ATU

0

186.3

100.0

225.8

4.4

198.0

100.0

230.5

Davis River 2
(Low Use-Warm)
Davis River 3
(High Use)
Chowika 1
(High Use)
Chowika Creek 2
(High Use)

0

239.5

100.0

356.8

100.0

343.9

100.0

433.0

0.9

176.9

100.0

189.5

0.0

149.3

100.0

167.1

Chowika Creek 3
(Low Use- Cold)
Swannell River 1
(Low UseWarm)

-

-

n/a

-

0.0

151.9

100.0

155.1

-

-

n/a

-

100.0

470.3

100.0

601.7

Davis River 1
(High Use)

Ui

n/a
0

152.3

80.0

147.6

n/a

100

-

75 -

50 -

25 -

1

2000-2001

1-------

2001-2002

Figure 7. Boxplots of percent survival both experimental years. For each experimental
year the medians, interquartile ranges (box), limits (whiskers) and outliers ( * ) are
presented.

56

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

We observed highly significant differences (P ’s > 6.51; P ’s < 0.001) in survival
between collection dates, among reaches and among families in both years with the
exception of ‘reach’ in year one where there was no difference (P 2,99 =1.21; P = 0.303)
in survival (Appendix A, Tables 10 and 11). In both years, survival was significantly
lower on the second collection date (Tables 6 and 7) (Y1 P | ,9 9 =12.61; P - 0.001 and Y2
Fi,i 48 = 11.17; P < 0.0005) (Appendix A Tables 10 and 11). Pairwise comparisons
between families for year one revealed significant differences between families one and
two, and families one and three (Tukey-Kramer P < 0.001 and P < 0.005). Family one
demonstrated the lowest mean survival (47.74%) next to family four (62.14%, Table 6).
Individual parental effects on survival in year one were not significant for males ( Y I P
1,124 = 2.75; P = 0.100) or females (Y1 F 1,124 = 1.07; P = 0.303) (Appendix A, Table 12).

In year two, family two typically had the highest mean value for survival, which was
more than double that of all the other families, however post hoc tests indicated no
significant difference for many paired comparisons (Table 7). We were unable to test
individual parental effects simultaneously due to limitations of the experimental design
related to the nesting of families within reach. When tested separately, however, male
effects on survival were significant, however females were not (Y2 Males, F 5 ,1 9 7 = 15.53;
P < 0.0005 and Y2 Females, P 5,197 = 2.06; P = 0.073) (Appendix A Tables 13 and 14).
Pair-wise comparisons among all males were highly significant (Tukey-Kramer; P ’s <
0.0005). Survival for reaches is presented for comparison with survival at the similar
reaches in year one (Table 7), however, no additional statistical treatment of survival data
for year two was considered due to the low and variable survival that was likely the result
of poor fertilization success rather than an experimental effect (see discussion).

57

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 6. Least squares means (ANOVA) for percent survival in experimental year 1
(2000-2001). All means are back transformed, thus standard errors are not provided. A
signifies significant results (ANOVA; P ’s < 0.0005). Letters within groups show
overlap in Tukey’s simultaneous 95% confidence intervals. Least squares means for
parental effects are from a different ANOVA (Appendix A Table 12) than results for the
rest of the table (Appendix A Table 10).
X
Collection Date*
December 1,2000
March 22, 2001

74.80
53.34

Family*
1
2
3
4

47.74
74.29
72.40
62.14

Reach
Davis River 1 (High Use )
Davis River 2 (Warm)
Chowika Creek 1 (High Use)

70.66
60.77
61.68

Parent
Male 1
Male 2

(Families 2 &4)
(Families 1 &3)

67.98
59.29

Female 1 (Families 1 &2)
Female 2 (Families 3 &4)

60.94
66.39

post hoc

A
B
B
B

58

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 7. Least squares means (ANOYA) for percent survival in experimental year 2
(2001-2002). All means are back transformed. A “*” signifies significant results
(ANOVA, P ’s < 0.0005). Letters within groups show overlap in Tukey’s simultaneous
95% confidence intervals. Least squares means for male and female effects are from a
different ANOVA (Appendix A Tables 13 and 14 respectively) than results for the rest of
the table (Appendix A Table 11).
_____________________________ X ________________ post hoc Groups
Collection Date*
March 1, 2002
April 29, 2002

10.69
3.68

Reach*
Chowika Creek 2 (High Use)
Chowika Creek 3 (Low Use)
Davis River 1 (High Use)
Davis River 2 (Low Use Warm
Davis River 3 (High Use)
Swannell River 1 (Low Use Ws

6.48
11.44
5.69
2.37
3.04
24.20

A

Family(Reach)*
Chowika Creek2-Family-1
Chowika Creek3-Family-1
Davis Riverl-Family-1
Davis River 1-Family-2
Davis Riverl-Family-3
Davis Riverl-Family-4
Davis Riverl-Family-5
Davis River2-Family-1
Davis River2-Family-2
Davis River2-Family-4
Davis River2-Family-5
Davis River3-Family-1
Davis River3-Family-2
Davis River3 -Family-3
Davis River3-Family-4
Davis River3-Family-5
Swanell Riverl-Family-1

6.48
11.44
9.74
16.34
5.16
2.28
0.82
2.22
12.48
0.30
0.28
4.89
14.21
3.73
0.69
0.01
24.20

A
A B
A B C
A B C D
A B C D E
A B C
E F
A B
E F G
A B C
E F G H
A B C D E F
I
A
E F G H
A
E F G H
A B C D E F G H I
A B C D E F
I
A B C D E
G H I
A
E F G H
A
E F G H
B C D E
I

4.73
1.11

A

B
B C
B C D
B C D
B

J
J K
J K L
L M
J K L MN
J K L
N
J K L
N
M

M ale (Reach)*
M ale 2 (Fam ily 3)
M ale 3 (Fam ilies 4/5)
M ale 5 (Fam ilies 1/2)

11.46

B
C

Fem ale (Reach)
Fem ale 1 (Fam ily 3)
Fem ale 2 (Fam ilies 1/5)
Fem ale 3 (Fam ilies 2/4)

9.50
6.52
12.54

59

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Additional analysis of survival in the context o f physical habitat was also not attempted
due to the lack of physical differences (except temperature) between sites in year one and
the low survival in year two.
Intergravel Environment
Temperature
Intergravel temperatures ranged from a maximum of 7.0°C to minimum of -1.7°C
across all artificial incubation sites. All but three artificial redds remained above freezing
during the incubation period in both years (Figs. 8a-8i). Temperatures at sites naturally
used for spawning (Davis mainstem, Davis east tributary and Chowika 2; Figs. 8a-8e)
were intermediate to sites not used for spawning (Swannell, Chowika 3 and Davis
(middle tributary); Figs. 8f-8i). Thermographs measured from all but two incubation
reaches followed a predictable pattern for northern streams where temperatures declined
to near 0°C at the onset of winter, remained low and stable for several months with a
warming trend again in spring at the onset of snowmelt and freshet (Figs. 8a-8i). Two
exceptions to general trends were found: the Davis River middle tributary (Figs. 8g and
8h) and the upper Swannell (Fig. 8i). The Davis River, middle tributary, site exhibited a
relatively variable thermograph in both years, with comparatively sharp temperature
changes of up to 2.5°C over days; a relatively short period of time (Figs. 8g and 8h).
Temperature loggers in both of these reaches exhibited temperatures that were above 1°C
throughout much of the winter.
W ith in reaches, stream a n d in terg rav el tem p eratu res w ere n ea rly identical for

most sites except for the Davis River mainstem (Figs. 8a and 8b), and the upper Swannell
River (Fig. 8i). Daily mean stream temperatures from September 12 to March 28, in the
Davis River mainstem reach in 2001 were on average 0.62°C (range 0.37 - 0.89°C)
60

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

7
Davis River (Dl- Main Spawning Area) 2000-2001

6

5

■Stream Bed
- Artificial Redd 1 (Left)
-Artificial Redd 2 (Middle)
-ArtificialRedd 3 (Right)

4
3
2

0
c?

S

J<

o?

Kf

A /

A '

f f

.cf

5? 5$v sSy J$v S?v .sSv 5Ss'5Sv 5Sv sSv 5^ ^

ef

Davis River (Dl- Main Spawning Area) 2001-2002
Stream Bed
Deep Intergravel (50 cm)
Artificial Redd 1 (Left)
Artificial Redd 2 (Middle)
Artificial Redd 3 (Right)

Davis River- (D3- East Tributary- Moderate Spawning) 2001-2002

Streambed
Artificial Redd 3 (Right)

Figures 8a-8c. Temperature profiles recorded for artificial egg incubation site for both
years. Each figure presents temperature data for up to three artificial redds (30-cm
intergravel), one deep (50-cm Intergravel) and the stream-bed for each site.

61

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

7i

D

Chowika Creek (C2-Main spawning area) 2001-2002
6 -

5

- - • Stream Bed

4

- ■Deep Intergravel (50 cm)
Artificial Redd 1 (Left)

3

Artificial Redd 2 (Middle)
Artificial Redd 3 (Right)

2 -

10-1
^
<0&

<41 cP o6"A0*1

jS'

<$' & ^
<<?
4 ? 4&

^

E

Chowika Creek (C2-Main spawning area) 2000-2001

Streambed
Artificial Redd 1 Downstream
Artificial Redd 2 Middle
Artificial Redd 3 Upstream

I ~ -A

J? J ?

4 *'V/c,<

^ 4f

S?

5?

<§> SS®

J?

S5"

J5S SSN ,£ N ,$ N

^ N°4?nja4?V<rV<r ^ or

'V

S5S ,^ N

»N

JSN $5N

rf

«r ^

V ' ^Y'

7
Chowika Creek (C3- Lower Main Stem- No Spawning) 2001-2002

J7

6

5

Streambed
Deep Intergravel (50 cm)
Artificial Redd 1 (Left)
Artificial Redd 2 (Middle)
Artificial Redd 3 (Right)

4
3
2

1
0
-1
<$>■
<4

>r 4?

Figures 8d-8f. continued.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Davis River (D2- Middle Tributary- No Spawning) 2000-2001
,
|
1

7

Streambed
. . . . Artificial Redd 1 (Left)
---------Artificial Redd 2 (Middle)
Artificial Redd 3 (Right)

Davis River (D2- Middle Tributary- No Spawning) 2001-2002

6

H

— Stream Bed
- Deep Intergravel (50 cm)
- Artificial Redd 1 (Left)
■Artificial Redd 2 (Middle)
■Artificial Redd 3 (Middle)

5
4
3
2 -

10
-1
°r
Swannell River (SI- No spawning) 2001-2002
Stream Bed
Deep Intergravel (50 cm)
Artificial Redd 1 (Left)
Artificial Redd 2 (Middle)
Artificial Redd 3 (Right)

V

v V

^

0
-1
cS& cS
°r^r
Figures 8g- 8i. continued.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

warmer than daily mean intergravel temperatures. In year 2, the mean daily difference
from September 11 to April 29 was 0.85°C (range 0.26-1.99). Similarly, deep intergravel
temperatures (50-cm depth) were within the range of temperatures collected at 30-cm
depth at all sites except the Davis River main-stem and the upper Swannell River site. At
the Davis River main-stem site, the deep intergravel logger was warmer than the shallow
loggers and had accumulated more thermal units by the end of the experiment (Table 5).
The Swannell River site thermograph was relatively unique in comparison to
other sites. Differences in temperature between the stream surface, the shallow and the
deep buried loggers were not constant through time and appeared to vary as groundwater
had an increasingly dominant influence as winter progressed (Fig. 8 i). Following late
September, a relatively warm ground water influence was evident, starting with the deep
buried logger, which then progressed from the deepest part of the channel (right logger)
to the shallowest part of the channel (left logger) over the winter. The Swannell River
site was the warmest overall during the incubation period and had accumulated the most
thermal units by the conclusion of the experiment (Table 5; Fig. 8 i).
Hyporheic Flow and Intergravel Chemistry:
ANOVA tables for analysis of hyporheic flow and chemistry are provided in
Appendix B. Three sites used naturally for spawning and one site not used (Swannell
River) had the highest mean values for specific discharge which often differed from
piezometers at 50-cm depth (Fig. 9a). In September absolute specific discharge was
h ig h est at sites u sed fo r spaw ning ( P i ;22 = 6.32; P = 0.020) b u t there w ere n o significant

differences between used and un-used sites in March and April, (F \ t22 = 6.32; P = 0.343;
P i ,22 = 2.84; P = 0.109) (Appendix B Table 15). Differences between reaches were not
significant (P’s < 2.65; P ’s > 0.061) during all dates, however, intergravel flow was
64

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

350

I0D 300
1

250

1?
a £ 200
o3
%

150
100

IwMliL
C2-S*

C2-A*

Dl-S*

Dl-A*

D3-S* D3-A*

C3-S*

C3-A*

D2-S

D2-A

Sl-S SI-A

■ Redd Depth
■ 50 cm
□ Surface

C2-S*

C2-A*

Dl-S*

Dl-A*

D3-S* D3-A*

C3-S

C3-A

D2-S

D2-A

Sl-S

Sl-A

Figure 9. Specific discharge and intergravel chemistry for artificial redds in year two
during the September and April collection periods. Data presented are data means and
standard errors for samples collected in the stream (chemistry only) and at redd depth (30
cm. Letters following each reach name signify collection dates where S = September and
A = A pril. A

follow ing reach n am e show s sites w h ere b u ll tro u t spaw ned.

65

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

significantly higher in the shallow piezometers within reaches over each of the three
sample periods (Fig 9a) (P ’s > 4.28; P ’s < 0.01) (Appendix B Table 15a - 15c). There
were also differences between sites used and not used and between reaches from
September to April (P’s > 12.26; P ’s < 0.0005) (Appendix B Table 17). Specific
conductance was high at all sites in the Muskwa Mountains (Northern Rocky Mountains)
and differences between reaches were small but highly significant (F ’s > 22.16; P <
0.0005) (Appendix B Table 16) for all three time periods. The upper Swannell River
(Omineca Mountains) site had the lowest conductivity, which was almost half that of the
streams on the opposite site of the Rocky Mountain Trench (Fig. 9b). Comparison of
specific conductance between surface waters and the shallow and deep piezometers
showed highly significant differences among sites (P ’s > 22.16; P ’s < 0.0005) for
September and March, but not April at the level of site use ( P 1 3 = 0.83; P = 0.374)
(Appendix B Tables 16a-16c). Similarly, differences between the surface and
piezometers at depth within reaches were not significant (P ’s < 1.51; P 's > 2.18)
(Appendix B Tables 16a-16c). We did however observe significant differences in
conductivity from September to April between sites used and not used (P 1,22 = 26.52; P <
0.0005) but not among reaches (P 1,22 = 1-03; P = 0.459) (Appendix B Table 18b). These
results indicate differences in hyporheic mixing through time for sites used and not used
for spawning but not among reaches.
Oxygen concentrations were high at almost all sites at the redd level, however,
deep piezometers from two sites (Site D3 and D2) had relatively low concentrations
< 6.5 mg l"1 (Fig. 9c). Differences in oxygen concentration between piezometers and
surface waters at sites used for spawning were significantly less than at sites not used for

66

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

samples September and March (F\j ,2 = 27.27; P < 0.0005 and F \ji = 22.32, P < 0.0005)
but not for April (Fig. 9) (F^is = 0.84, P = 0.553) (Appendix B Table 17) and F ’s >2.67,
P ’s < 0.0005) (Appendix B Table 18c), indicating less hyporheic mixing at sites not used
for spawning. We observed the lowest oxygen levels in September at the two warmest,
groundwater-influenced sites; Davis (middle tributary) and Upper Swannell River (Fig.
9c) at the 50-cm depth. We were unable to complete the analysis of pH among sites and
between collection dates to non-normal distribution o f the data, however, pH at all sites
was within a narrow range (7.14-8.44) and differences are unlikely to be of biological
significance. Numerically, Swannell River had the highest pH levels compared to the rest
of the sites.
Embryo Size and Yolk Utilization
In both years measurements of growth and yolk utilization varied across experimental
units and differences in measures of embryo size and yolk utilization were highly
significant for most experimental factors including; date, reach (Fig. lOa-d), redd,
capsules, family (Fig. 1la-d), and parent (Fig. 12a-d; Table 8 ). However, post hoc
comparisons between redds within reaches and between capsules within redds were
generally not significant. Thus, growth differences were largely explained by variation
occurring at the level of the reach and our presentation of results is confined to factors at
the level of the reach and above.
In general, the longest and heaviest alevins were found within the warmest
reaches (D2 and SI) (Figs. 10a, b, d) that also had the highest ATU’s (Figs. 13 and 14).
For example, by the end of the experiment in year 2 alevins at site SI were on average
23.6 mm whereas alevins collected on the same date from site D1 were 20.1 mm.

67

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

0.0300
0.0275
w 0.0250
B
o 0.0225
H
& 0.0200
Q 0.0175
0.0150

A

I

0.0100
00

B

0.0080

0.0060
o'
« 0.0040
&
Q 0.0020 H
0.0000

ABE

0.0300 -|
00

0.0250 0.0200 0.0150 -

Q 0.0100
0.0050
21.00
18.00 ■oo
5

2002

8

0 2001

Reach
Figure 10. Effect of reach on growth and yolk utilization for alevins collected in both
field experiments. Marginal means with standard errors are presented for total dry
weight, dry-body weight, dry-yolk weight and wet length except for dry-body weight
where only back transformed means are provided. Shared letters represent overlap in
Tukey’s simultaneous 95% confidence intervals. Lower case letters are for year one and
upper case letters are for year two. A
indicates sites where bull trout spawned.

68

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

0.02300

B D

0.02200
£o 0.02100
H
& 0.02000

DC

B

"
-t-

A

A

0.01900
0.00550 n

B

£
3

0.00500

o

b
Q

0.00450
0.00400
0.01800

^

B

0.01600
A

BC

C D

D

-i-

"f-

O 0.01400 H
>«
&
Q 0.01200 H
0.01000
c

18.00

D

17.50 -|

ABD

AB

17.00
su
h-l

16.50 -|
16.00
a

b

c
Family

d

e

Figure 11. Effect of parent on growth and yolk utilization for alevins collected in both
field experiments. Marginal means with standard errors are presented for total dry
weight, dry-body weight, dry-yolk weight and wet length, except for dry-body weight
where only back transformed means are provided, thus standard errors are not provided
for dry-body weight. Shared letters represent overlap in Tukey’s simultaneous 95%
confidence intervals. Uppercase and lowercase letters differentiate males from females.

69

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

0.0270

A

B

@ 0.0268
o 0.0266
H
& 0.0264

*

■f

E

■f

0.0262
0.0041
0.0039
3 0.0037
o 0.0035
DQ 0.0033
b
Q 0.0031
0.0029
0.0235
'bO
# 0.0230

B

B

'i ■

A

A
-±~

o

a
-i-

b

^ 0.0225
C1
° 0.0220

-ir

13.70
13.60
13.50
nu
h-l

13.40 13.30
1

2

Female

1

2

Male

Figure 12. Effect of family (year 2) on growth and yolk utilization for alevins collected
in both field experiments. Marginal means with standard errors are presented for total
dry weight, dry-body weight, dry-yolk weight and wet length except for dry-body weight
where only back transformed means are provided, thus standard errors are not provided
for dry-body weight. Shared letters represent overlap in Tukey’s simultaneous 95%
confidence intervals.

70

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Figure 13. Means and standard errors of lengths for alevins from artificial redd sites
collected over both years plotted against accumulated thermal units (ATU). Each point is
an arithmetic mean for all alevins from one redd on a single date. Individual reaches are
differentiated in the legend by letters for each stream and by numbers for each reach. (“S”
for the Swannel River, “D” for the Davis River and “C” for Chowika Creek). A
indicates sites where bull trout spawned.

71

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

0.0140

0.0120 -

0.0100

•5*
o
m

0.0080

• SI

fc> 0.0060

Q

0.0040

* D2

i

■ Dl*
♦ D3*

sz

+ C3
* C2*

0.0020

0.0000
.Cl
\v

♦ Cl*

cvbb

>5)
©cSS

b
©
b®

b
A*

ATU

Figure 14. Means and standard errors of dry-body weight for alevins from artificial redd
sites collected over both years plotted against accumulated thermal units (ATU). Each
point is an arithmetic mean for all alevins from one redd on a single date. Individual
reaches are differentiated in the legend by letters for each stream and by numbers for each
reach. (“S” for the Swannel River, “D” for the Davis River and “C” for Chowika Creek).
A
indicates sites where bull trout spawned.

72

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Alevins from the coldest site (C2) were 16.9 mm. Dry-body weight increased
significantly over time at the warmer sites (F’s > 3158.99; F ’s <0.0005) (Appendix C
Table 21 and Table 25) and the heaviest alevins were found at the warmest sites (Fig.
10b). Dry-body weight was similar among the coolest sites (C1-C3, D3) in year 2
(Fig.lOb). Yolk utilization was also highest at the warmest locations (Fig. 10c) and
alevins collected in April 2002 at site SI had nearly exhausted their yolk reserves and
only had on average 0.0027 mg of dry yolk remaining, whereas alevins from site C2 had
0.014 mg of yolk. Total dry weight (combined dry yolk and dry-body weight) was lowest
at the two warmest sites (Fig. 10a and 15). Over the course of the experiment, there was
little difference in ATU’s between collection dates for low temperature sites (C1-C3, D3)
(Table 5) and growth and yolk metabolism proceeded despite the low temperatures (Figs.
13, H a n d 15).
The magnitude of differences for dry weights and length due to individual
parents was generally small (Figs. 1la-1 Id), however, female parental effects on length,
total dry weight and dry-body weight were significant in year one (F’s > 4.08; F ’s < 0.05)
(Appendix C Tables 19, 20 and 21) (Table 8; Figs. 11a, l ib and 1Id). In all cases female
2 was linked to higher growth and a larger total weight, but not higher yolk reserves
(Figs. 1la -1Id). In contrast, ANOVA showed male parental effects for length, dry-body
and dry-yolk weight (Appendix C Table 19-21) (Figs. 1Id, 1lb and 11c). In year two we
were unable to test individual parental effects on growth factors due to limitations
imposed by low survival rates and the nested study design; therefore we substituted
family as a combined factor in place of individual parents. In year two, there were highly
significant differences in length, total dry weight and dry yolk that were apparent

73

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

0.0350 i

0.0300

3*
0.0250
~

0.0200

"o

• SI
h

* D2

f'C

■ Dl*

I I s

♦ D3*

*s

+ C3

& 0.0150

Q

a C2*

♦ Cl*

0.0100

0.0050

0.0000
c©

$

$
A'

ATU

Figure 15. Means and standard errors of dry yolk weights for alevins from artificial redd
sites collected over both years plotted against accumulated thermal units (ATU). Each
point is an arithmetic mean for all alevins from one redd on a single date. Individual
reaches are differentiated in the legend by letters for each stream and by numbers for each
reach. (“S” for the Swannel River, “D” for the Davis River and “C” for Chowika Creek).
A
indicates sites where bull trout spawned.

74

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

between families (Table 8, Fig. 12 a, 12c and 12d) (F ’s > 8.97; F ’s < 0.0005) (Appendix
C Tables 23, 24, 26), however a family effect was not apparent for dry-body weight (Fig.
12b) (F 4,7i 8 = 1-28; P = 0.342) (Appendix C Tables 25). In general, the magnitude of
differences in yolk reserves between families was small where the differences were on
average less than 0.005 mg. Differences in length between families were more striking
where families C and E (linked by a common male) were almost 0.8 mm longer on
average than the other three families (Fig. 12d).

Discussion
Many studies have emphasized the importance of the incubation environment,
particularly temperature, for salmonids and it has been suggested that Pacific salmon
have adapted their migration timing and patterns to exploit different habitats across
species ranges to maximize egg-to-fry survival (Brannon 1987). Groundwater and
hyporheic exchange have frequently been cited as important factors influencing spawning
site selection and potentially incubation success in several salmonid families, including
chinook salmon (Oncorhynchus tshawytscha) (Geist 2000), chum salmon (O. keta),
(Leman 1993), brown trout (Salmo trutta) (Hansen 1975), eastern brook char (Salvelinus
fontinalis), (Curry et al. 1995; Blanchfield and Ridgeway 1997; Curry et al. 1995), Arctic
char (S. alpinus) (Cunjak et al. 1986), and bull trout (Baxter and McPhail 1999; Baxter
and Hauer 2000; Chapter 2). Patterns of use vary between species, for example brook
trout (Webster and Eriksdottir 1976) and Arctic char (Cunjak et al. 1986) in some
locations, appear to have strong preferences for upwelling water. This preference is not
as clear with other species, such as chum salmon (Leman 1993) and bull trout (Baxter
and Hauer 2000, Chapter 2) and there appears to be intra-specific variation between

75

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

populations. Leman (1993) observed that two populations of chum salmon occupying the
same river system appeared to select spawning sites differently where one had a
preference for sites with groundwater up-welling into the hyporheic zone and the other
preferred sites with strong hyporheic exchange (mixture of both up- and downwelling).
Differences in preferences within and between species are likely representative of a range
of adaptive responses to a range of available habitats that are suitable. The adaptive
response to this variation (i.e., site selection and its various mechanisms) is likely driven
by combinations of habitat features that affect juvenile growth and survival and
ultimately incubation success at a population level. Few studies however, have attempted
to link attributes of groundwater to incubation success within a population of fish.
Our aim was to understand how variation in the incubation environment of bull
trout affected size and survival of eggs and alevins in reaches selected and not selected
for spawning. Our results have shown that all reaches examined were of relatively high
quality based on intergravel flow (specific discharge) and measures of chemistry and did
not affect survival and size, despite instances of significant differences between sites.
Temperature regimes differed substantially between some sites, but did not appear to
affect survival during incubation. Temperature, however, had a strong effect on size and
development and the temperatures we observed at sites used for spawning were
intermediate to sites not used. In particular, we found that small changes in temperatures
had relatively large effects on the size of embryos and alevins as well as the timing of
hatch within and among the stream reaches examined. Considering that the habitats we
examined appeared to be of high quality, our results suggest that temperature is a key
habitat attribute for incubation success and that bull trout are selecting spawning sites

76

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

with stable, intermediate incubation temperatures relative to what is available.
Site Selection Habitat and Groundwater Influences
Chemical and physical measures (i.e., oxygen, pH, intergravel flow, conductivity
and temperature) at all reaches appeared to be within a suitable range for bull trout
survival. Within our study areas, we observed temporal and spatial differences among
sites and between deep and shallow piezometers for some locations but not all (Fig. 9a9c). Spawning site selection did not appear to be influenced by any one factor measured
and no clear patterns of use, based on measurements collected at the six sites were
observed. Several factors in combination may determine site selection, but for ease of
presentation the factors are described individually below. Temperature may be a key
determinant of site use, as sites intermediate to the range of available temperature
regimes were used for spawning to a greater extent. Gravel permeability and interstitial
oxygen levels may also partially explain spawning site use in the context of site quality;
lower intergravel flow and lower oxygen concentrations were linked to sites that were
less used, however these effects did not appear to be strong. Finally, we found evidence
that water chemistry (including pH, conductivity and oxygen concentration) varied
among and within sites and combined with temperature, these features may be important
indicators of groundwater that could act to influence spawning site selection.
We used specific discharge as a measure of gravel permeability, an indicator of
site quality and a potential indicator of groundwater discharge (Freeze and Cherry 1978;
Baxter and McPhail 1999; Baxter and Hauer 2000). Values of specific discharge were
highest at high-use reaches in the Davis River and Chowika Creek (D1 and C2) and at
one non-selected reach in the Swannell River (SI) (Fig. 9a). Specific discharge was

77

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

lowest at two of the non-selected reaches in the Davis River and Chowika Creek (sites D2
and C3; Fig. 9). Oxygen concentrations, another indicator of habitat quality, were well
within the ranges suitable for incubating salmonids (Barton and Taylor 1996) (Fig. 9c).
Oxygen concentrations were, however, lowest in two sites not used for spawning (D2 and
SI) and were more variable at these sites compared to the other sites used for spawning.
We also recorded measurable differences within reaches, between redds and between
deep and shallow piezometers, and surface waters, however, these differences were not
statistically significant (Fig. 9c). Given our relatively small sample size (three measures
per redd, 12 per reach spread over three dates), it is possible that differences that would
have been of biological significance were masked by low statistical power. Our
observations for lower individual measurements of oxygen concentration and lower
specific discharge correspond well with visual observations of fine sediment levels within
each of the sites we used for the incubation experiment and may be indicators of site
quality. During construction of the artificial redds in reaches C3 and D2, we noted (but
did not specifically measure) higher levels of fine sediments relative to the other reaches
with reach D2 having the highest levels. High levels of fine sediments are known to
reduce interstitial spaces within the gravel nests of salmonids resulting in lower oxygen
levels and suffocation of incubating salmonid embryos leading to reduced emergence
success (Chapman 1988; Bjomn and Reiser 1991).
Measurements of pH and specific conductance (Fig. 9b) were relatively
h o m o g en o u s w ith in an d b etw een s reaches o n the eastern side o f the R o ck y M o u n tain

trench where relatively high conductivity and alkaline pH reflected the prevalence of
calcareous sedimentary rock in the Muskwa Mountains. Differences between Swannell

78

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

River and the remaining sites were likely the result o f differences in geology, where the
Swannell Ranges are primarily igneous rock of either volcanic or deep origin which
would have less propensity to leach minerals into stream and groundwater compared to
rock containing calcite. For both pH and conductivity, we did not observe differences
between surface waters and, deep and shallow piezometers that would likely be
biologically significant, with the exception of site D2 (Fig. 9b). Site D2 had significantly
lower conductivity in surface waters which could be related to a strong groundwater
influence at this site. Higher conductivity and lower relative pH due to increased levels
of dissolved carbon dioxide and increased levels of dissolved organic carbon along with
increased dissolved solids are characteristics of subsurface or groundwater (Kaplan and
Baker et al. 2000; Newbold 2000). We also found further evidence of a strong
groundwater influence at this site. In winter we noted a nearby seepage area that
remained relatively warm and ice free even when air temperature was below -25 °C. In
some circumstances, localized discharge of deep origin groundwater devoid o f oxygen
may have negative influence on site selection (Dent et al. 2000; Geist 2000) and this may
partially explain lack of use at site D2.
Temperature patterns within and between stream reaches in our study area varied
substantially during the winter incubation period and the influence of relatively warm
groundwater intrusions was evident (Fig. 8). The two warmest sites (SI and D2),
remained well above 0°C even during the coldest months, accumulated the most thermal
units (Table 5) and were also the least used sites for spawning. Site SI also demonstrated
a large-degree of spatial heterogeneity within the site between temperature loggers during
any given time period. As base-flows declined during winter a thermally-stable

79

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

groundwater intrusion was evident at depth within the stream-bed, where the intrusion
progressed from the logger nearest the thalwag to the logger near the stream margin
during winter (Fig. 8i). Based on these observations, including the lack of use at sites SI
and D2, we suggest that in thermally heterogeneous environments, bull trout may be
avoiding sites with relatively high or temporally variable intergravel temperatures.
Temperatures above an optimum range may produce rapid growth and advanced size, but
may induce pre-mature exhaustion of yolk reserves leading to excessive mortality in the
natural environment. Thus, predictable thermal habitat, both spatially and temporally, in
cold northern environments could form the basis upon which northern stocks of bull trout
use sites for spawning. Brannon (1987) observed that several species of Pacific
salmonids have adapted their spawning timing and probably spawning site selection on
the basis of thermal regimes in an effort to enhance success of their progeny. The timing
of spawning and duration of incubation in these populations was thought to optimize the
timing of emergence to precisely coincide with an increase in stream productivity in
spring. In our study area, relatively high or variable temperatures may therefore have a
significant negative influence on the rate of development of incubating bull trout if the
timing of emergence was suboptimal. Variable or low intergravel flow rates and oxygen
concentrations or inappropriate temperatures at potential spawning sites may, therefore,
partially explain patterns of use in our study area. Although, site selection could be
dependant on several factors, including groundwater influence, water chemistry, substrate
and temperature regime, temperature is likely the strongest link to site selection.
Habitat and Survival
Habitat parameters, including incubation temperature, specific discharge and

80

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

chemical measures collected within reaches were all sufficient to maintain life at all sites
and likely had little effect on survival during incubation at the sites we used in year one
of our study. Year one mean survival was relatively high (Fig. 7) at all sites and was
within the range reported for experiments using similar artificial incubation vessels in
stream gravels: for bull trout 76.1-88.6 % (Baxter and McPhail 1999); for sockeye, 2751% (Cope and Macdonald 1998) and for chinook 22-58% and steelhead 61%; (Merz et
al. 2004). We did not make similar comparisons for year two due to low and variable
survival likely due to poor fertilization success (see below), although one groundwater
fed site in the Swannell drainage demonstrated a relatively high survival possibly
indicative of suitable site conditions during incubation. Further study at this site would
be required, however, to establish this relationship.
Differences in survival between two reaches naturally selected for spawning and
one low-use site for year one were not apparent (Tables 7 and 8) despite some significant
differences in measure of intergravel flow rates, intergravel chemistry and incubation
temperatures (Figs. 8-9). Habitat quality was therefore sufficiently high to support
relatively high levels of survival at all sites. In fact, virtually all of the habitat measures
we collected for our incubation experiment were within the ranges that we recorded in a
survey of natural redds for a related field study (Chapter 2). Our observations, therefore,
suggest that all the habitats we measured, including those chosen and not chosen by bull
trout, appeared to have been suitable for adequate survival during incubation. In other
w ords, at a m in im u m th resh o ld h ab itat quality, survival is un affected a n d site selection

may be beyond that threshold may be influenced by factors unrelated to survival during
incubation.

81

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Although not supported by our study, small differences in habitat quality have
been observed to affect habitat preferences or survival in other studies. Geist et al. (2000)
observed that chinook spawning in the Hanford reach of the Columbia River selected
sites with groundwater influence that also had oxygen levels above 7 ± 0.9 mg f 1 and
selection for this attribute was presumed to enhance survival. Baxter and McPhail (1999)
compared two field sites, one used for spawning and one that was not used and they
found that survival was enhanced and less variable for bull trout that were incubated in
slightly-warmer areas with upwelling groundwater compared to an adjacent area that was
not used for spawning. Although the differences in survival were small, survival at the
selected spawning area was 88.6 percent compared to 76.1 percent for the non-selected
site. Cope and Macdonald (1998) compared in situ survival of sockeye salmon embryos
among incubation sites of different apparent habitat qualities, where quality was based on
observations of relative use and measures of physical habitat, including gravel
permeability. As with our results, they did not find differences in survival among sites
and they argued that their initial assumptions of habitat quality, based upon high and low
site use by spawners, were not valid and that truly marginal areas were not used for
spawning at all.
Similar to the approach of Cope and Macdonald (1998), we selected our field sites
on the basis of relative levels of use by spawners to test for the effects of habitat quality
on survival and growth. In our experiment, the middle tributary (low-use) reach in the
Davis River was not used for spawning and we assumed this site to be marginal due to
the lack of use; likewise the two other sites in the Davis River and Chowika Creek that
were used in year one were selected on the basis of relatively high use by spawners

82

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

(Chapter Two) and habitat quality was assumed to be high. Given that survival was
similar at all sites (Table 8) the high intergravel flow (Fig. 9a) and high levels of oxygen
(Fig. 9c), it is apparent that the Davis River low-use site was not marginal based on these
measures. It is notable that different temperature regimes did not affect survival despite
temperature differences between sites within the Davis River of more than 2°C during the
winter incubation period (Figs. 8g and 8h). These differences resulted in values for
accumulated thermal units (ATU) for the warm site that were more than 200 ATU greater
than the cooler site by the second collection date in year one(Table 5). Temperature
differences were even more pronounced for other sites used in year 2 (Table 5) where
difference in ATU’s between the Swannell River site and the high-use site in Chowika
Creek approached 440 ATU over approximately the same time period.
In both the Davis River and Chowika Creek, natural incubation sites were
generally less than 1°C once winter set in (late Nov - Chapter 2). Thus, it is possible that
temperature differences much greater than 1°C or 2°C could have stronger negative
effects on survival during incubation on populations adapted to spawning in cold
environments. In more southern stocks, cold temperatures may have the opposite effect
for fish adapted to spawn in warmer locations. For a southern population of bull trout,
McPhail (1979) noted slightly lower survival for fish incubated at 2°C compared with
4°C, however, at both temperatures survival was greater than 80 percent. Adaptation to
local temperature conditions appears to be a common feature among stocks of different
salmonids (Brannon 1987; Murray and McPhail 1988; Beacham and Murray 1989;
Baxter and McPhail 1996).
One limitation of our study is that we recorded survival well before the normal

83

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

emergence period from the gravel, thus survival may have lower after emergence 1.5
months later (near the middle of June; Chapter 4). Survival at these sites could also have
been influenced in part by the methods we used, as artificial egg capsules may have had a
positive bias on survival. The capsule may act to reduce infiltration of fine sediments to
the eggs during incubation, thus promoting survival. Second, emergence from the gravel
is normally a period o f high natural mortality and our methodology did not test for
survival post-emergence, which could have been more than a month before natural
emergence. Considering the low levels of use by spawners at sites C3 and D2 and SI
(Chapter 2) and our methods for measuring survival, it is possible that bull trout may
have been avoiding these areas due to the fine sediment loads and lower oxygen levels
despite our observations that the these sites were adequate for survival. Further
examination of survival at a wider range of sites and with a variety of incubation methods
could be used to test these effects.
Parental Effects on Survival
For both of our experiments, variation in survival for both years was partially
explained by family or parental effects. Inadequate fertilization or poor gamete quality
may be largely responsible for the observed effects in year two, including low overall
survival. Family differences in survival in our experiment were apparent for both years
(Tables 6 and 7), however, individual parental effects were not significant in year one
when survival was generally high. Parental effects on year two survival were strong for
males but not for females (Table 6). Male three for example was linked to the lowest
mean survival across all families (Table 7). In contrast, male five sired families one and
two which had the highest mean survival. It is possible that the observed effects reflected

84

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

natural variation in survival among family crosses as the result o f a paternal effect on
early development. Withler (1987) investigated differences in a paternal effect on the
survival and development of a hatchery stock of chinook salmon. She noted that the
effect was heritable and affected early embryonic development. Pattemal effects, both
heritable and environmental are generally uncommon (Heath et al. 1999). In our study,
we strongly suspect that the male effect for year two was related to gamete quality and
subsequent fertilization success. Stress due to an extended holding time for adults
captured on September 9 and 10, 2001 or our egg handling methods are likely
mechanisms for poor fertilization success or poor gamete quality. Prior to spawning in
2001, adults were retained in holding tubes in the river for several days following the
events o f September 11, 2001 as flight bans prevented us from returning to the field sites
to spawn these fish until September 13, 2001. It was further observed that during
spawning in year two, the volume of milt obtained appeared to be less than the previous
year, although we did not measure this directly. An assessment of fertilization success
would have enabled us to determine whether or not the poor survival was an artefact of
the experiment or a natural occurrence. Unfortunately, we were unable to determine this
in the lab with confidence and confirmation was therefore not possible. Eggs that were
examined a few days after fertilization were very light in colour and we found it difficult
to identify any evidence of cell division.
Temperature, Size and Development
Small temperature differences between sites over the course of incubation had
profound effects on the size and development o f bull trout at different sites within our
study area. Higher incubation temperatures advanced the timing of hatch, increased yolk

85

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

utilization (Fig. 15) and resulted in heavier and longer alevins at warmer sites by the
conclusion of our experiment (Tables 6, Figs. 13 and 14) despite relatively small
temperature differences between sites (Table 5, Fig. 8). Large sizes and advanced
development at higher temperatures are not un-expected for salmonids (Combs 1965;
Peterson 1976; Heming 1982; Humpesch 1985; Tang et al. 1987; Garrett 1988; Murray
and McPhail 1988; Beacham and Murray 1989; Baxter and McPhail 1999; Huuskonen
2003), however, considering the relatively low range of incubation temperatures we
observed across sites used for our experiment (Figs. 8a-f), the magnitude of differences in
sizes and decreases in the time to hatch at the warmest sites were surprising. For example
hatch likely took place in late January through the month of February for the warmest
sites based on the results o f a related laboratory study (Chapter 4), whereas hatch
occurred in March or later for the cooler sites.
Many other incubation studies for salmonids have noted substantial increases in
mortality and relatively slow or decreased size for temperatures similar to those
characteristic of our study. Few o f these experiments, however, have explored low
temperature conditions that would be encountered naturally by any given stock and thus
offer limited inference about adaptation (or lack thereof) to cold conditions. Much of this
experimentation has employed constant temperature regimes to test temperature effects
on southern or coastal stocks that are likely adapted for more temperate conditions. Very
low temperatures (i.e., < 2°C) during early incubation are unrealistic for most species and
have been observed to increase mortality for many stocks, including those adapted to cold
temperatures (e.g., Murray and McPhail 1987; Tang et al.1987). Cold temperatures
during early incubation can disrupt development prior to organogenesis, (e.g., Peterson

86

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

1976) and few salmonids (including northern stocks) could be expected to spawn at these
temperatures. In our study area for example spawning was largely complete before
temperature declined below 4-5°C.

Summary
We found relatively large significant differences in size (lengths and weights),
yolk utilization, as well as differences in the timing of developmental stages across
habitats (reaches) that were used and not used for spawning. Significant differences in
size were also noted at lower factor levels than the reach, as we also observed differences
between artificial redds and between families and individual sires and dams. We found
that physical and chemical habitat quality appeared to be high at all sites, and with the
exception of temperature all features we measured appeared to be poor predictors of
survival and size during incubation. Bull trout in our study area appeared to be well
adapted to temperatures below 1°C after the onset of winter and for majority of the
incubation period. Finally, we found relatively high survival at all sites, and size was
relatively constant among the sites naturally selected for spawning. Small increases in
incubation temperature, however, resulted in large sizes and higher yolk utilization; thus,
based on the observations in this study and a related synoptic study of habitat at redd sites
(Chapter Two) it appears that the range of preferred temperatures is relatively narrow.
Temperatures outside this range may result in pre-mature yolk exhaustion and advanced
developmental timing that could lower survival post emergence.

87

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

CHAPTER 4 : THE EFFECT OF AN INCREASE IN INCUBATION
TEMPERATURE ON A COLD-WATER ADAPTED SALMONID, THE BULL
TROUT.3

Abstract
Fish are most sensitive to changes in temperature during early life-stages, including
incubation. As fish grow and develop, the thermal tolerance of most species becomes
greater. We examined the effect of small increases in temperature above a natural
temperature regime on stages of development for larval bull trout until button-up; the
stage when all yolk has been utilized and the end o f the larval phase. We incubated eggs
in a laboratory environment where temperature was controlled. Our objective was to
assess body size (length and weights), yolk utilization and the timing of developmental
stages under natural and two elevated thermal regimes (+1.5°C and +3°C). We observed
substantial differences in alevin length and yolk utilization rates among temperature
replicates within both the field (Chapter 3) and laboratory experiments. The timing of
hatch, button-up as well as size and yolk utilization were all advanced relative to the
control temperature for both of the experimental temperature replicates. We also
observed significant increases in mean white muscle fibres in the warmest temperature
treatment.

3 Throughout this chapter I use the first person plural to acknowledge the contribution of others to this
work, which will be submitted for publication with the authorship of C.J. Williamson and J.M. Shrimpton.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Introduction
Temperature has a critical role in determining behaviour, growth, development
and ultimately survival in fish. Reproductive behaviours in many fish, including
salmonids, are adapted to take advantage of a wide variety of thermal regimes in lakes,
rivers and streams (e.g., Brannon 1987; Leman 1993; Power et al. 1999). Once eggs are
deposited by females into gravel nests and are covered, parental care is essentially
complete and the development of eggs to emergence proceeds largely as a function of the
temperature of the incubation habitat (Fleming and Reynolds 2004). In northern and
interior environments, there is evidence that salmonids can be highly selective when
choosing spawning locations, and the locations of redds are often linked to areas of
groundwater influence with distinct temperature regimes (Hansen 1975; Leman 1993;
Cunjak et al. 1986; Curry et al. 1995; Blanchfield and Ridgeway 1997; Geist and Dauble
1998; Ridgeway and Blanchfield 1998; Power et al. 1999; Baxter and McPhail 1999;
Baxter and Hauer 2000; Geist 2000). Groundwater moderates both warm summer
temperatures and cold winter temperatures where water can often decline to temperatures
near 0°C in winter (Power et al. 1999).
In general, the zone of thermal tolerance during embryonic development is
considerably more narrow than it is later in life (Rombough 1997) and the rate of
embryonic and larval development is highly dependent on temperature. Optimum
temperatures for incubation vary inter- and intra-specifically (Brannon 1987; Beacham
and Murray 1988) and temperatures outside of the optimum can alter growth and
development rates, influencing both the timing of emergence and first feeding in
juveniles as yolk reserves are finite (Heming et al. 1982). Above thermal optima the

89

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

conversion efficiency of yolk is inversely related to temperature and yolk reserves are
less efficiently converted at warmer temperatures (Heming 1982; Ojanguren et al. 1999).
Increases in temperature may, therefore, lead to precocial emergence of smaller fry with a
lower potential for survival for salmon trout and char; this is particularly important for
interior (versus coastal) populations which tend to experience relatively cool incubation
temperatures. Sublethal effects can also include changes in muscle fibre density and
cellularity, which have implications for ultimate growth potential (Johnston 2001). Thus,
changes to the temperature regime can disrupt the timing and rate of development during
incubation and can ultimately lead to changes in reproductive success or output
(Markevich and Bilenskaya 1992).
A number of factors have been shown to alter temperature of aquatic systems.
Timber harvest can increase stream temperature approximately 1°C (Holtby 1988;
Shrimpton et al. 1998). Additionally, some scenarios of global warming predict increases
in air temperatures of 2-4°C (Hengeveld 1990; Wood and Macdonald 1997) which may
lead to changes in stream hydrology and changes in temperatures in north temperate
streams. A stable incubation environment may be particularly important for embryos that
are immobile for long periods after spawning until hatch. Eggs and alevins that over­
winter in the hyporheic zone in stream reaches benefit from the high oxygen content of
the surface water and thermal stability of the groundwater (Baxter and McPhail 1999;
Power et al. 1999). Changes in surface water temperature resulting from land-use
practices or global warming, therefore, have the potential to impact rate of development,
growth, and yolk conversion efficiency of incubating salmonid eggs and alevins.
In B.C., bull trout are a species of concern and are blue listed by the B.C.

90

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Conservation Data Centre (B.C. CDC, 2006). This species has received considerable
interest as it has been shown to have a low tolerance to habitat disturbances and has a
preference for cooler water temperatures (Goetz 1997). Currently, quantitative
descriptions of the thermal tolerance or temperature optima for developing bull trout from
northern streams are not available. Our objectives for this study were to measure
development and changes in body size (length and weight) and yolk utilization during
incubation for bull trout from the Davis River, B.C., under a natural temperature regime
and two elevated temperature regimes. The results of this study will provide fisheries
resource managers a better understanding of potential impacts to bull trout where landuse practices or global warming could disrupt incubation habitats.

Methods
Experimental Setup
To determine the effect of temperature on incubation o f bull trout we used three
thermal regimes; one similar to a site heavily used by bull trout for spawning in the Davis
River, B.C. (Fig. 1), the second approximately 1.5°C warmer than the Davis River
temperature, and similar to that of non-spawning streams in the Davis watershed, and the
third approximately 3°C warmer than the Davis River temperature (Fig. 16). We
modeled the temperature regimes on intergravel temperatures recorded from a natural
redd in the upper Davis River during the incubation period in 2000-2001. To reduce
noise in the natural stream data and to provide a guide for controlling temperatures in the
laboratory, we fit a second order polynomial regression curve (MS Excel 1997, Microsoft
Corporation, Redmond, Wa.) to a plot depicting mean daily temperature for this site (Fig.
16). We used this curve as guide to control the Davis River regime, however this curve

91

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

only approximates the actual temperatures achieved. The other two regimes were
elevated by 1.5°C and 3°C from the Davis regime in the laboratory.
Our experiments were conducted at the Pacific Western Brewery located in Prince
George, B.C. Inlet water for each temperature regime was unchlorinated groundwater.
The well water was chilled down from 9°C using water chillers (1 Hp and 1/3 Hp
Models, Frigid Units, Toledo, Ohio). Each water-chilling unit was placed in a
continuously aerated and insulated 170 1plastic reservoir tank. Chilled water from each
reservoir flowed to a stack of vertical incubation trays. Outflow from the incubation
trays was pumped back into the reservoir from a sump with an overflow outlet. The
water reservoir continuously received fresh water at a flow rate of approximately
1 1-min'1 and the same volume was allowed to drain from the overflow in the sump. To
achieve temperatures close to 0°C in the Davis River temperature replicate, we modified
a 1 Hp chiller with 110-v timer, so that the compressor cycled on and off (2 min on and 5
min off), while the circulation impeller ran continuously. This modification prevented
the build up of ice on the chilling coils that would have otherwise reduced the chilling
efficiency of the unit. Without modification, the coldest temperature we could achieve
was approximately 1.5°C. The actual temperature regimes that were achieved for the
laboratory experiment were an approximation of Davis River thermograph and were
subject to the design limitations of the water chillers and equipment used. For example,
diurnal and weather related fluctuations in stream temperature were not reproduced in the
laboratory experiment. To record temperature we used three submersible temperature
data loggers (Optic Stowaway Temp, Onset Corporation, Bourne, MA) within each
temperature replicate. We programmed the data-loggers to recorded hourly temperatures

92

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

for the duration of the experiment.
Fish
We captured four adult bull trout (two males and two females) from the upper
Davis River during the spawning period in September 2001. We collected gametes from
the Davis River fish in the field and placed them in storage containers that were kept cool
in ice-chilled coolers. We retained approximately 1200 eggs from both females which
were transported unfertilized to the laboratory by plane within 6 h. We fertilized the eggs
immediately upon arrival at the laboratory facilities. The remaining eggs from these fish
were used for our field experiment (Chapter 3).
At the brewery, we created four family groups by dividing the gametes from the
two females and two males and subsequently recombining them to form a two by two
breeding design. We mixed eggs and milt for each cross separately in clean plastic
buckets, without water. After approximately 2 min, we added water to the eggs, which
were then left to sit for 5 min. We then washed excess milt from the eggs which were
then left to water harden for 30 min. After fertilization, we divided each family group
into 12 portions and placed them in incubation trays (MariSource, Milton, Wa.) under
three temperature regimes. We replicated each family group within each temperature
regime; four families, three temperatures and two family replicates per temperature.
Each tray was divided into four lanes, and one to three family group replicates were
placed in each tray. To reduce incident light levels we wrapped each stack in black
polyethylene plastic sheeting.
Survival, Embryo Size and Yolk Utilization
Starting at approximately two months after fertilization, we monitored trays for

93

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

survival on a weekly basis and we removed all dead eggs. Once the eggs had reached the
eyed stage, we mechanically shocked each tray to aid in the identification and removal of
unfertilized dead eggs. Thereafter, on at least a weekly basis for the duration of the
experiment, we removed and counted dead eggs and alevins, counted live alevins and
counted and recorded hatched eggs. We cleaned individual trays periodically for the
duration of the experiment. Starting at 50% hatch and then at approximately one-month
intervals we collected samples of five to 20 live alevins from each family replicate for
analysis of size (length and weight) and yolk utilization. Individual sample sizes were
dependent on total family survival to the eyed stage. We simultaneously euthanized and
preserved eggs and alevins for later analysis by freezing them instantly at -80°C. Using
the methods of Chapter 3, we dissected eggs and alevins into yolk and tissue portions.
For each embryo, we recorded standard length to the nearest 0.01 mm using Vernier
callipers under a dissecting microscope and then collected weights on an analytical
balance (0.0001 g).
Muscle Fibre Density
For each family within each temperature replicate, samples of three to six alevins
were collected for analysis of muscle fibre density. Samples for the muscle analysis were
collected at 50 percent hatch and at button-up. Fish were preserved in formalin for 48 h
and then stored in phosphate buffered saline. To quantify muscle fibre density, alevin
samples were mounted in paraffin, sectioned, mounted on microscope slides and then
stained. Briefly, after storage each alevin was sectioned with a scalpel just anterior to the
dorsal fin insertion and just posterior to the trailing edge of the anal fin. The mid­
sections of each alevin for each treatment replicate were transferred in lots to distilled

94

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

water and then dehydrated through a graded ethanol series from 50% ethanol to 100%
ethanol using six different solutions over a 31-h period, after which they were washed in
three, 7-min baths of fresh toluene. Each lot was then bathed in 60°C paraffin for 24 h
before being mounted for sectioning with a microtome. Individual fish were then
sectioned transversely to the longitudinal body axis in 5- to 6-pm increments to a point at
least 15 to 20 pm anterior to the external opening of the anus. Sections were floated on a
warm (approximately 45 °C) water bath and mounted on gelatin coated microscope slides.
Using the procedure outlined by Kieman (1999) slides for each fish containing the
appropriate landmark (opening of the anus) were stained with haematoxylin-eosin.
Digital photographs of one section nearest to the external opening of the anus were
collected with a Powershot G5 digital camera (Canon Canada Inc.) mounted to an Axio
Star Plus microscope (Carl Zeiss Microimaging, Inc.Thomwood, NY). For our analysis
we used one half of each myotome section. Muscle myofibres and myomere crosssectional areas from myotome sections were then counted and measured using ImageJ
image analysis software (Rasband 1997-2006).
Data Analysis
Measures of alevin size (dry-body weight, dry-yolk weight, length and total
body weight) as well as muscle fibre counts for each myomere were examined using
nested ANOVA where factors in the model included, temperature treatments, parents,
and tray and the interaction of parent (male crossed with female). Tray was nested within
temperature treatments. All treatment levels were treated as fixed factors. All ANOVA’s
were completed using the general linear model (GLM) procedure in Minitab (v. 14.1,
State College, PA). We used the adjusted sum of squares to test significance. Post-hoc

95

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

comparisons were completed, using the Tukey-Kramer multiple comparison procedure in
Minitab. The Tukey-Kramer method was used as this method for multiple comparisons
was suitable for un-balanced experimental designs and is more conservative than other
similar procedures.
All data were screened for erroneous values, outliers and as part of each
procedure, we visually checked all ANOVA models for normality and heterogeneity
using the output provide by Minitab, which included: normal probability plots,
histograms and scatter plots o f the residuals for each model. All ANOVA procedures
met the assumptions of for normality and homogeneity of variance. To better understand
growth trajectories within each temperature replicate, von Bertalanffy growth curves
were fit to the length data using a non-linear least squares procedure (LevenbergMarquardt) (Statistica v.6.1, StatSoft Inc. Tulsa, OK). Estimates of the asymptotic length
were computed to compare growth trajectories.

Results
We were able to achieve an approximation of the three temperature treatments for our
experiments as described in our methods (Fig. 16). The temperature of the Davis River
was 5.6°C in September when the bull trout were spawned. The temperature of the coldwater treatment simulated the natural decline in temperature seen in the Davis system.
The mid-temperature group targeted approximately 1.5°C above the Davis River
temperature, and the warm water group target temperature was approximately 3°C above
the Davis River temperature. At the end o f November, the lowest temperatures for each
group were achieved; 4°C, 2°C, and 0.6°C for the warm, mid, and cold temperature
groups, respectively. We kept temperatures approximately constant until mid-March

96

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Davis River
+1.5
+3

Davis River Field
Poly. (Davis River Field)

cd
u
&
u
H
L-i

T

CS^

CS*

CS*

cO1
’

cO^

cO'

CS^

5

cO^

cn^

cO'

cO'

cO'

cO'

////# /# //# //////
N f i^

cO^ cO^ eC^ C V ^

1----------1

cO'

C \^

D ate

Figure 16. Temperature achieved for three incubation replicates in the laboratory plotted
with natural temperatures for a redd from the Davis River (Davis River Field). Natural
redd temperatures were collected in 2000-2001. A second order polynomial was fit to
this data to provide a general template to guide temperature control of the Davis River
treatment regime.

97

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

2002. At this time of the year, water temperatures in the Davis River begin to warm (Fig.
16). Temperature spikes (Fig. 16) reflect some of challenges in conducting the
experiment including power and equipment failures or the need to clean the incubation
apparatus.
On November 8, 56-d after fertilization, eggs for each experimental regime had
reached the eyed-stage. Mortality estimates before 100% hatch were not calculated due
to poor and variable fertilization success for some families. Poor survival was likely in
part the result of poor gamete quality due to an extended holding time for adults after
capture (see Chapter 3). Survival was, therefore, not examined in detail for this study.
Overall, survival to 100% hatch ranged from 12 to 95% (Table 8). The first samples
were collected at hatch for each family group (Table 9). The timing of hatch, button-up
as well as size and yolk utilization were all significantly advanced relative to the control
temperature for both of the experimental temperature replicates (Table 9, Fig. 17).
Compared to control fish, 50% hatch was advanced by 56 and 76 days for the
experimental mid temperature and warm temperature groups, respectively (Table 9).
Differences between hatch and time-to-button-up increased as temperature increased.
Button-up occurred at approximately 77, 86 and 93 days following hatch for the warmest
to the coldest group respectively (Fig. 17). These differences show that the effect of
warmer temperatures on hatch was stronger than it was on time to button-up.
Differences in length, dry-body and dry-yolk weights among dates,
temperatures, females and trays were all highly significant (F’s >3.32; P ’s < 0.001)
except for dry-body weights where there was no significant female effect (F i;6i8 = 2.39;
P = 0.123) (Appendix D Tables 27-29). The interaction of male on female was not

98

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 8. Survival (%) at 100% hatch for each temperature regime by incubation tray,
family, female and male.
Davis River + 1,5°C + 3°C
51.5
11.5
23.8
75.0
16.8
90.5
75.0
52.6

79.0
24.0
94.7
93.9
23.1
88.1
62.0
56.1

77.5
29.0
84.6
87.9
23.1
81.0
51.0
57.9

Tray Family Female Male
1
1
2
2
3
3
3
4

A
B
C
A
B
C
D
D

1
1
2
1
1
2
2
2

2
3
2
2
3
2
3
3

Table 9. Date and number of days to 50 % hatch and button-up for three laboratory
temperature regimes and one natural redd from the Davis River (Davis River Field).
“Date” designates the estimated date for each stage; “Days” designates the number of
days from fertilization; and “ATU” represents the summed average daily temperature for
each period specified. The “Davis River Field” column is presented for comparison only
and does not represent a specific spawning or fertilization time in the field.

Stage
50 % Hatch Date
Days
ATU
Button-up Date
Days
ATU

+ 3°C
+ 1.5°C
Davis River Davis River Field
31-Dec-01 20-Jan-02 17-Mar-02
17-Mar-02
110
130
187
187
493.3
434.1
285.2
218.3
27-Mar-02 23-Apr-02 02-Jun-02
02-Jun-02
196
223
263
263
767.9
604.4
390.3
334.8

R ep ro d u ced with p erm ission o f the copyright ow ner. Further reproduction prohibited w ithout perm ission.

Emergence/
y = -98.08Ln(x) + 848.99
Button-up
R2 = 0.9969

275 n
250
225
200

cs

175

tJ
u
Pi
o

150

«

100 H ♦ Davis

50% Hatch y = -136.1 lLn(x) + 956.82

N

S'
Q

R2 = 0.9889
▲

125

♦ +1.5
♦ +3
■ Gould (1987)
▲McPhail and Murray (1980)
♦ Weaver and White (1985)

75
50
25

0
240

290

340

390

440

490

540

590

640

690

740

790

ATU

Figure 17. Duration of development (hatch and button-up) plotted against thermal
history in accumulated thermal units (ATU). Data from the other authors were not
included in the models for our data.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

significant (F ’s < 3.32; P 's > 0.05) (Appendix D Tables 26-28). Lengths and dry-body
weights were greatest at the warmest temperatures (Figs. 18-20) and at the highest
ATU’s; although dry-body weights declined once yolk reserves were depleted in the
warmest group. Maximum dry-body weights in the Davis and +1.5°C temperature
groups were similar, despite differences in the number of ATU’s (Figs. 19 and 20). Our
estimates of the asymptotic length (length versus ATU from fertilization) for fish
incubated in the cold (3c = 24.41-mm; 95% Cl 24.77- to 25.68-mm) were higher than for
the +1.5°C (3c = 23.11; 95% Cl 23.96- to 23.25-mm) but were similar to the warmest
group (3c = 25.74; 95%CI 25.4- to 26.07-mm ) (Fig. 22). The relationship between yolk
utilization and ATU was also similar between the Davis and the +1.5°C temperature
regimes (Fig. 21), however, at higher ATU’s the rate of yolk utilization was more rapid
for the +3°C regime. Higher temperatures also resulted in larger sizes and yolk
utilization (Figs. 22-24). Maximum size was achieved in the warmest group by
approximately 196 days after fertilization, while this did not occur until 263 days after
fertilization for the Davis group. Yolk exhaustion occurred near day 223 in the warmest
group, whereas yolk was still present by day 263 (button-up) for the Davis regime.
Mean white muscle fibre counts were highest in the heated groups (F’s > 6.84;
P 's < 0.002) for each myofibre although there were no differences explained by parent or
by tray or by developmental stage (Fig. 25) (Appendix D Tables30a-d). Post hoc
examination between treatments revealed significant differences between the Davis River
regime and warmest treatment (+3°C) (Tukey P - 0.0010); there were, however, no
significant differences in myofibre density between the Davis and +1.5°C and between
the +1.5°C and +3°C regimes (Tukey P’s > 0.05).

101

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

800 i

Davis River
"+1.5"

29.0 mm

Davis River Field

18.9 mm

Date
Figure 18. Thermal history o f three alevins collected on March 18, 2002 under three
different temperature regimes. The arrow show the sample date. Lengths o f each
individual alevin are presented for comparison and are representative o f the average
alevin in each lot. The dashed line “Davis River Field” is the temperature profile for a
natural redd collected in winter 2000-2001. The arrow shows the precise collection date.

102

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

26.00

♦ Davis
+1.5
* +3

24.00

22.00 ^

y= 25.91 ■1 -exp(-0.014 ( x - 447.80 ))
20.00
60

8 18.00 16.00
y= 23.13 • 1 -exp(-0.049 ( x - 409.49 ))
14.00
y= 24.41 • 1 -exp(-0.018 ( x - 220.5 ))
12.00
200

300

400

500

600

700

800

900

1000

ATU

Figure 19. Length plotted against accumulated temperature units (ATU;°C-day) for each
temperature replicate.
0.016 0.014 -

♦ Davis
+ 1.5
a+ 3

0.012 ^

s

♦
♦
♦

0.01-

-g 0.008 -

♦

b

♦
%

Q 0.006
0.004 -

♦♦
V
♦
*

0.002
0 -I

200

1--------------- ,--------------- 1--------------- 1--------------- 1--------------- 1--------------- 1--------------- 1

300

400

500

600

700

800

900

1000

ATU

Figure 20. Dry-body weight plotted against accumulated temperature units

103

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

(ATU;°C day) for each temperature replicate.

0.026
♦ Davis
+ 1.5
* +3

0.024
0.022

♦

♦

0.02 H
♦

0.018

A
A

♦ ♦
♦

a

0.016

"§ 0.014
& 0.012
Q

0.01

0.008
0.006
0.004
0.002 -i

0
200

300

400

500

600

700

800

900

1000

ATU

Figure 21. Dry yolk plotted against accumulated temperature units (ATU;°C day) for
each temperature replicate.

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

y= 25.59 • 1 -exp(-0.074 ( x - 104.13 ))

26.00

♦ Davis
+1.5
a +3

24.00

y= 23.07 • 1 -exp(-0.101 ( x - 118.66))

22.00

20.00 -

n>
8 18.00
hJ

y= 26.91 • 1 -exp(-0.015 ( x - 129.4 ))
♦

16.00 -

♦

14.00

Pre-hatch
12.00
150

100

200

250

300

Days from Fertilization

Figure 22. Length plotted against days from fertilization for each temperature replicate.

0.016 n

♦ Davis
0.014

+ 1.5
* +3

0.012
<3

0.01

B
• f 0.008 H

£>

o 0.006 0.004 -

/

0.002

Pre-hatch

0
100

150

200

250

300

Days from Fetilization

Figure 23. Dry-body weight plotted against days from fertilization for each temperature
replicate

105

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

0.026 n
0.024

Pre-hatch

0.022

♦ Davis
+ 1.5

0.02

4+3

Dry yolk (mg)

0.018
0.016 0.014 0.012 0.01

0.008
0.006
0.004
0.002

0
100

150

200

250

300

Days from Fertilization

Figure 24. Dry yolk plotted against days from fertilization for each temperature replicate.

106

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

1475 i
♦

1275 -

♦ Cold (H)
♦ Middle (H)
♦ Warm (H)
■ Cold (B)
« Middle (B)
■ Warm (B)

1075

n =33-36
«=24
n =20-21
«=13
n= 1
n = 25

o
♦

♦

475

275 -

I

»
»

♦

«
*

75

2

3

4&5

Total

Myomere #

Figure 25. Mean muscle fibre (myofibre) count from each muscle myomere for each
temperature replicate at hatch and at button-up. Some fish did not have a distinguishable
fifth myomere, so 4 and 5 were combined. Sample sizes are provided for each
temperature regime at hatch and button-up for each myomere. “H” designates hatch and
“B” designates button-up.

107

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Discussion
Temperature dominates the rate of early development in salmonids. Our results
confirm and expand on existing work examining the effects of temperature on
development in embryonic and larval salmonids in general and for bull trout in particular.
The acceleration of development as well as the large differences in the size and yolk
reserves resulting from small increases in temperature in our study were dramatic; similar
effects have seldom been demonstrated previously (but see Rombough 1997). We also
confirmed that bull trout can successfully tolerate incubation regimes with temperatures
less than 1°C during incubation and may in fact require these temperatures for long
periods during incubation to prevent premature exhaustion o f yolk in northern streams.
A surprising result of our work is the total length of time required to reach
developmental stages for these fish. Hatch and emergence were both substantially
delayed compared to more southern populations of bull trout (Fig. 17; McPhail and
Murray 1980; Weaver and White 1985; Gould 1987). This is the first reported account of
bull trout incubation at controlled temperatures below 1°C. Weaver and White (1985)
reported field incubation temperatures of 2°C to 5.4°C, although their thermographs were
not used to record data during the coldest part o f winter. Cope and Macdonald (1998)
reported cold incubation temperatures approaching freezing in a sockeye salmon stream
that is also known to be used by bull trout for spawning.
Additionally, time to hatch and emergence were not constant between temperature
rep licates, w h ich suggests th at b u ll tro u t show g ro w th co m p en satio n (B rannon 1987)

based on temperature. Longer incubation times likely represent adaptation to local
thermal regimes and a longer winter period compared to populations in the south.
Alteration of thermal regime during incubation may also lead to changes in growth, and
108

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

ultimately survival, of juvenile bull trout.
Rate of Development
Time-to-hatch ranged from 122-187 days and time to button-up ranged from 196to 263 days (Table 9, Fig. 17). The duration of incubation for the natural temperature
regime was approximately nine months. An example of the effect of the warmer
temperatures on rate of development is seen in the time of button-up compared to hatch.
Time of development differed considerably among the temperature treatments such that
hatch for the control fish occurred within nine days of button-up for the warmest group
(Table 8, Fig. 18). The large temporal differences in developmental stages we observed
are similar to other studies. Time-to-hatch for Davis River fish was consistently two
months longer than for bull trout from southern B.C. and the north-western United States,
which incubated over a similar range of accumulated thermal units (ATU, measured in
degree days) (Fig. 17; McPhail and Murray 1980; Weaver and White 1985; Gould 1987).
Gould (1987) reported time-to-emergence was also about 2-months earlier than the time
for button-up we recorded for Davis River fish. In contrast, Weaver and White (1985)
reported a time-to-emergence (223 days) that was similar to that of our 1.5°C temperature
treatment (Fig. 17). It was, however, unclear whether winter incubation temperature were
recorded or estimated in their study. The other authors did not report button-up or
emergence times. In our experiment we noted, but did not specifically record, that
button-up and behaviours related to emergence such as positive or neutral photo-taxis
(versus negative phototaxis prior to emergence) were roughly coincident. It is possible,
therefore, that the differences in time to emergence we have noted between northern and
southern populations could in part be explained by a difference between the timing of

109

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

button-up and the time of emergence from gravel nests.
Time-to-hatch is not necessarily a precise indicator o f development and may be
subject to external selective pressures such as changes in habitat quality that alter survival
during incubation. Once hatch occurs, alevins may be mobile and can potentially avoid
suboptimal or lethal conditions such as anchor ice formation. Cope and Macdonald
(1998) found that after hatch, sockeye salmon alevins from an interior population had the
capability of avoiding ice by moving through the gravel whereas a coastal population did
not exhibit this behaviour. A relatively rapid time-to-hatch that occurred before freezeup was also a characteristic o f this interior stock. Bull trout in our study area do not
appear to exhibit rapid development to hatch and hatch occurs well into the winter
months (Fig. 17) when ice-formation in natural redds is possible. Interestingly, the late
timing of hatch may be one of the reasons bull trout in particular seem to have a strong
affinity for groundwater fed influenced spawning areas. Groundwater intrusions can
provide stable thermal regimes and can prevent the formation of anchor ice (Power et al.
1999).
Time-to-emergence is likely subject to strong selective pressures that act to
optimize juvenile survival and growth (Brannon 1987). Compared to hatch, however,
there is likely less flexibility for emergence dates as yolk reserves are finite and
emergence in salmonids often coincides with button-up, first feeding and increases in
food availability during spring (Dill 1967). We observed times to hatch and potentially
emergence in our study that were longer than for southern populations. The timing of
emergence approximately coincides with spring freshet (mid-June) based on our field
observations during 2001 and 2002. Both of these observations support the hypothesis

110

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

that this population of bull trout is adapted to local temperature regimes in order to
maximize the growth and survival of juveniles. We observed distinct differences in the
timing of specific life stages between our data and southern populations of bull trout (Fig.
17). Such adaptations have also been reported for populations of Pacific salmon
(Brannon 1987; Beacham and Murray 1988; Leman 1993), brook trout (Curry et al.
1995) and Atlantic salmon (Salmon salar) (Webb and McLay 1996).
The number of degree days to achieve a specific developmental stage decreased
with increases in temperature (Fig. 17). In the warmest temperature group, 768 degree
days were required for complete button-up. In the mid temperature group, 604 degree
days were required for complete button-up and 390 degree days were required in the
natural regime. Thus, the relationship between ATU and development is not constant for
larval development at very low temperatures for bull trout. A negative relationship
between growth and time-to-development was described as growth compensation by
Brannon (1987). Brannon speculated that compensation in growth is an adaptive
mechanism that acts to stabilize emergence timing for fry under variable environmental
conditions during spawning through emergence. Growth compensation has been
observed intra- and inter-specifically in the development rates for four species of Pacific
salmon (Brannon 1987; Beacham and Murray 1988; Murray and McPhail 1988). Our
results show that bull trout may also be able to growth compensate; however, further
study over a wider range of temperatures as well as at constant temperatures will be
required for this comparison.

Ill

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Embryonic Size and Yolk Utilization
Substantial differences in alevin size and yolk utilization rates were evident
between each temperature regime throughout the experiment (Figs. 19-21), which if
experienced in the natural environment may have lead to depletion of yolk reserve and
pre-mature emergence. Warmer temperatures significantly increased lengths and weights
of alevins among dates and even among trays within incubation stacks where there were
small amounts of warming from top to bottom. We also found small maternal and some
paternal effects. Female effects are common in salmonids as egg size has a direct bearing
on growth until yolk reserves are exhausted (Heath et al. 1999). Alevins were collected
on March 18, 2002, within a few days of the time of button-up for the warmest regime
(+3°C regime) and 50% hatch for the natural regime (Davis River regime). A difference
of >10-mm was observed for the warmest experimental group compared to the control
(Fig. 18). The warmest temperature regime resulted in the longest and heaviest alevins,
but this occurred earlier in time at the warmest temperature. Our observation contrasts
with those previously reported. Heming (1982) noted that chinook salmon incubated at
higher temperatures were smaller and had reduced yolks reserves at hatching and
emergence than fish reared at cooler temperatures. Murray and McPhail (1989) made
similar observations on five species of Pacific salmon, however, weights were lower at
temperatures near 5°C for pink salmon compared to warmer temperatures. Hamor and
Garside (1997) and Ojanguren et al. (1999) also found an inverse relationship between
temperature and size in Atlantic salmon. Huuskonen et al. (2003) observed the same
relationship for Arctic char for both length and weight. Considering that bull trout often
incubate in cold water, it could be expected that low-temperature incubation should

112

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

produce large fry, however, it is unclear why our results differed from these reported
studies. The coldest thermal regimes used in our study was close to the freezing
temperature of water and this may be a common phenomenon of tissue growth and
development at cold temperatures. It is possible, therefore, that bull trout embryonic
growth is more efficient and that a growth optima exists at slightly warmer temperatures
than we used in our experiment.
A possible confounding factor in the study was that both the middle and natural
temperature replicate experienced relatively high levels of blue-sac disease. Blue-sac
disease is characterized by increased fluid retention in the yolk sac as well as decreased
growth and increased mortality (Peterson et al. 1977; Ihssen 1978). Peterson et al. (1977)
observed that rapid decreases in temperature at the eyed and hatch stages increase the
incidence of edema of the yolk sac in Atlantic salmon. In our study, both of these
temperature treatments experienced equipment failures that resulted in temperature spikes
(sharp increase and decrease; Fig. 16) during the course of the experiment which may
have contributed to this effect.
Muscle Cellularity at Hatch and Button-Up
Previous work has demonstrated that changes to the thermal incubation
environment of salmonids can alter muscle development in fish (reviewed in Johnston
2001). Nathanailides et al. (1995) found that Atlantic salmon alevins incubated at
ambient temperatures had similar white muscle cross-sectional areas to fish incubated at
higher temperatures, but fibre number was higher for the ambient group. In a similar
study involving two different stocks of Atlantic salmon incubated under natural
temperature regimes and under the regime of the other respective stock, Johnston et al.

113

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

(2000a) examined muscle cellularity at hatch and at button-up. They noted significant
effects on both white muscle cross-sectional and fibre area that remained until the time of
first feeding. In an extension of the previous study Johnston et al. (2000b) found that
these effects were maintained into the juvenile stage for one population but not the other.
Based in part on these results, Johnston (2001) speculated that different responses may
reflect local adaptation in these fish. He further added that alterations to thermal regime
or other factors affecting growth of muscle can cause effects that may be maintained until
later in life; such effects could include changes in ultimate body size and swimming
performance. To better understand how an altered temperature regime could affect
development in bull trout, we examined differences in white muscle fibre density at the
three different temperature regimes. We found highly significant differences in the
number of muscle fibres existed between the temperature groups (Fig. 25); however, the
results for white muscle area did not differ significantly. Post hoc examination of our
results revealed that difference in muscle fibre number existed only between the warmest
and natural temperature replicates.
Previous work on Atlantic salmon has shown the cross-sectional area o f white
muscle fibers and fiber number (myofibre hypertrophy) to be greater for embryos reared
at lower temperatures (Johnston 2001). As a consequence, larval fish reared at lower
temperature are not only larger at emergence, but also have a greater number of muscle
fibers. The ecological implication of this finding could be that fish reared at lower
temperatures have a greater chance of survival due to greater size and greater growth
potential (Rombough 1987). Within the range of temperatures we examined, myofibre
hypertrophy increased at elevated temperatures. This result contrasts with other

114

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

researchers who demonstrated decreases in muscle hypertrophy at higher temperatures.
Differences may reflect adaptation to naturally low temperature incubation environments.
We suggest that our observation of significant differences in white muscle fibre number
in combination with our clear observation on growth indicate that incubation at
temperatures near freezing has a profound effect on development in larval salmonids.
Yolk Utilization
Early exhaustion of yolk reserves in the warmest temperature group was another
important finding of our study. Under the natural temperature regime, the depletion of
yolk was initially slow in comparison to the two warmer groups and accelerated through
time. We presume that the absorption rate would have decelerated after button-up, as the
yolk depletion appears to be described well by a sigmoid function (Joworski and Kamler
2002). In the two warmer groups, yolk absorption was rapid in the period following
hatch and the amount of yolk remaining at the time of emergence was inversely related to
temperature (Figs. 21 and 24). Near the time of complete depletion of yolk in the
warmest replicate (Figs. 21 and 24), the weight o f these alevins had started to decline.
The depletion of yolk represents an end-point for endogenous growth. The alevin must
begin exogenous feeding to survive once yolk reserves are depleted. Exhaustion or
depletion of yolk reserves leads to high mortality in salmonids in the natural environment
due to early emergence and can also result in reduced body weight at emergence (Heming
1982; Ojanguren et al. 1999). The implications of small increases in temperature during
incubation, therefore, are profound.

115

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Summary
Our results confirm and expand on existing work examining the effect of temperature on
larval development in salmonids. For bull trout, even apparently small differences in
incubation temperature can have substantial impacts on growth and development from
emergence to hatch. We have demonstrated that time-to-hatch, time-to-emergence, size,
yolk utilization and potentially muscle fibre numbers are all affected significantly by
relatively small changes to the incubation regime. As shown by previous researchers (see
Rombough 1997) early developmental stages in fish are particularly sensitive to
temperature changes, with the acceleration of development in salmonids being especially
dramatic. Alterations in the stream environment related to industrial development or
changes in global climate can cause stream warming or changes in groundwater flow
patterns. The resulting changes may cause precocial exhaustion of yolk reserves and
early emergence. Thus, changes to the incubation environment of bull trout, a cold water
specialist may disrupt the timing of development, duration of incubation and size at
emergence and ultimately long term reproductive success.

116

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

CHAPTER 5: GENERAL DISCUSSION
Bull trout are coldwater specialists that home to specific areas to spawn that are
commonly influenced by groundwater (Baxter and McPhail 1996; Baxter and McPhail
1999; Baxter and Hauer 2000; Bahr and Shrimpton 2004). At the reach scale, my results
demonstrate that bull trout spawn in groundwater-influenced valleys with extensive
gravel deposits with high rates of hyporheic exchange. I also found evidence of localized
groundwater influence at the habitat scale, where bull trout were spawning in a variety of
habitats with both relatively warm upwelling groundwater or downwelling streamflow.
This confirms the findings of Baxter and Hauer (2000) from a Montana population of bull
trout, where the specific site chosen appeared to dictate the direction of the vertical
hydraulic gradient. Sites located at the start of the transition between stream features
exhibited downwelling whereas areas adjacent to banks or downstream of pool riffle
sequences exhibited upwelling. The use of specific areas influenced by groundwater with
specific thermal signatures may help to maximize reproductive success of fish by limiting
the effects of stochastic environmental processes that lower survival during incubation.
The presence of upwelling did not necessarily explain all the variation in site use as other
authors have suggested (Baxter and McPhail 1996; Baxter and McPhail 2000); however,
warm up-welling water appeared to have a positive bias on the sites used. Three of the
warmest naturally-used sites in my study had the highest number of redds. In general,
however, the warmest available sites appear to be avoided by spawners based on our
observations. Other physical features linked to habitat quality may moderate site
selection. I found that high intergravel-flow, high levels of oxygen and proximity to
security cover all were features found at sites used for spawning.

117

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

In an effort to better understand the importance of habitat features at redd sites
and the implications of habitat alteration on spawning and incubation success; I used a
field incubation experiment to explore the effects of different habitat parameters on
incubation success in an ecological context. I attempted to define the range of natural
variation in habitat within my study area to determine the potential effects on bull trout
survival, development size and yolk utilization.. I found that physical and chemical
habitat quality appeared to be high at all sites, and, with the exception of temperature, all
features I measured were poor predictors of survival, size and yolk utilization during
incubation. In terms of thermal habitat, bull trout in my study area appeared to be well
adapted to temperatures below 1°C after the onset of winter and for the majority of the
incubation period. I also found relatively high survival at all sites, and growth was
relatively constant at the sites naturally selected for spawning. Surprisingly, warmincubation temperatures at non-used sites relative to used sites resulted in increased
growth rates and advanced timing of hatch and button-up, despite the fact that the
differences were relatively small. Thus the growth response as well as differences in the
timing of developmental stages across habitats (reaches) that were used and not used for
spawning were relatively large and significant.
Comparing growth between reaches, I found average lengths and body weights
were highest at the two warmest sites that were also not used for spawning.
Temperatures at these sites were more than 2°C warmer than three sites that were used
ex ten sively b y spaw ning bull trout. In this context, y o lk reserves w ere lo w est at the

warmest sites and were less than half of the reserves for alevins collected within the
selected sites. Considering that spring freshet, normally a period of natural fry

118

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

emergence, was still at least a month and a half away, survival of the alevins at the
warmest sites was unlikely as exogenous food resources are likely scarce prior to spring
freshet. Furthermore, the digestibility of prey items is lower at low temperatures
(Salvanes et al. 1995) and even if food resources were found, low temperature may limit
growth from exogenous feeding. I found that bull trout selected spawning sites with low
absolute temperature. The incubation temperatures at these sites throughout the winter
were, however, intermediate within the range that was surveyed at all sites. The narrow
range of temperatures throughout the winter for selected sites, therefore, suggests that a
selective pressure for a specific thermal regime exists within these populations of bull
trout. It is likely, that the incubation thermal regimes are directly related to survival post
emergence and is suggestive that the mechanism for selection is through habitat
preference for temperature differences among reaches. Based on my observation of an
apparent preference for sites with low absolute temperature that were intermediate within
the range I surveyed, as well as a pessimistic outlook for survival to emergence for
alevins that were incubating in the warmest habitats, I hypothesize that for bull trout in
my study area, habitat preference is strongly linked to temperature differences between
reaches.
I also observed significant differences in length, weight and yolk utilization at
lower factor levels than the reach, including differences between artificial redds and
between families and individual sires and dams. Heterogeneity in temperatures within
reaches and measurable differences in size at this scale suggests the possibility that site
selection and adaptation to incubation temperature could be operating at a very fine scale
in bull trout. Considering these results in the context of the synopsis of natural redd

119

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

habitat characteristics, it appears that the range of preferred temperatures is relatively
narrow.
In the laboratory experiment, I examined the effects of small increases in
temperature on size and development of bull trout that normally incubate at low
temperatures to test their resiliency to change. My results confirm and expand on
existing work examining the effect o f temperature on larval development in salmonids.
For bull trout, even apparently small differences in incubation temperature can have
substantial impacts on size and development from emergence to hatch. I have
demonstrated that time to hatch, emergence, size, yolk utilization and potentially muscle
fibre counts are all affected significantly by relatively small changes to the incubation
regime. As shown by previous researchers (see Rombough 1997), early-developmental
stages in fish are particularly sensitive to temperature changes, with the acceleration of
development in salmonids being especially dramatic. Alterations in the stream
environment related to industrial development or changes in global climate can cause
stream warming or changes in groundwater flow patterns. The resulting changes may
cause precocial exhaustion of yolk reserves and early emergence. Thus, changes to the
incubation environment of bull trout, a cold water specialist may disrupt the timing of
development, duration of incubation, size at emergence and ultimately long term
reproductive success.
In summary, based on the results of these studies, I recommend the following to
reso u rces m an ag ers seeking advice o n pro tectin g b u ll tro u t spaw ning habitat. F irstly, bull

trout appear to home to and select specific reaches within stream with high fidelity.
Features such as coarse woody debris, high intergravel permeability, and specific thermal

120

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

regimes appear to result in habitat selection at a fine scale. To maintain high quality
spawning reaches, the structure, function and physical stream processes that govern the
availability of course woody debris and channel features associated with bull trout redds
should be protected and examined through further study.
Secondly, groundwater upwelling at the reach scale and hyporheic exchange
within reaches both appear to be an important component of bull trout spawning habitat
in northern B.C. Based on my work and the work of others, however, there appears to be
considerable variation in the groundwater supply and function between watersheds. Thus
I recommend, in the face o f industrial developments that could alter fluvial and
groundwater hydrology, steps should be taken on a watershed by watershed basis to
better understand how groundwater influences and moderates spawning habitat at both
the watershed and reach scale in significant bull trout spawning streams. These studies
should be undertaken prior to any development and measures for mitigation should be
incorporated into operational planning.
Finally, bull trout appear to be adapted to relatively narrow ranges of temperature
for spawning with apparent differences between northern and southern populations.
Small, changes to temperatures as the result of industrial developments or global
warming have a strong potential to disrupt bull trout spawning and incubation success,
therefore, it is incumbent that the proponents of such works undertake the necessary
studies to ensure that their activities do not degrade the thermal habitat of bull trout
spawning areas. Such work should include efforts to understand local adaptation within
and between watersheds.

121

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

REFERENCES
Anderson, W.G., McKinley, R.S., and Colavecchia, M. 1997. The use of clove oil as an
anaesthetic for rainbow trout and its effect on swimming performance. N. Am. J.
Fish. Manage. 17: 301-307.
Bahr, M.A. 2002. Movement patterns, timing of migration and genetic population
structure of bull trout (Salvelinus confluentus) in the Morice River watershed,
British Columbia. Masters Thesis, University o f Northern British Columbia,
Prince George, B.C.
Bahr, M.A., and Shrimpton, J.M. 2004. Spatial and quantitative differences in
movement o f large bull trout {Salvelinus confluentus) from a watershed in
northwestern British Columbia, Canada, are due to habitat selection and not
differences in life history. Ecol. Freshwater Fish 13: 294-304.
Barton, B.A., and Taylor, B.R. 1996. Oxygen requirements of fishes in northern Alberta
rivers with a general review of the adverse effects of low dissolved oxygen. Wat.
Qual. Res. J. Can. 31: 361-409.
Baxter, C.V., and Hauer, F.R. 2000. Geomorphology, hyporheic exchange, and selection
of spawning habitat by bull trout {Salvelinus confluentus). Can. J. Fish. Aquat.
Sci. 57: 1470-1481.
Baxter, J.S., and McPhail, J.D. 1999. The influence of redd site selection, groundwater
upwelling, and over-winter incubation temperature on survival of bull trout
{Salvelinus confluentus) from egg to alevin Can. J. Zool. 77: 1233-1239.
Beacham, T.D., and Murray, C.B. 1989. Variation in developmental biology of sockeye
salmon {Oncorhynchus nerka) and chinook salmon (O. tshawytscha) in British
Columbia. Can. J. Zool. 67: 2634-2648.
Bjomn, T.C., and Reiser, D.W. 1991. Habitat requirements of salmonids in streams. In
Influences of forest and rangeland management on salmonid fishes and their
habitats. Edited by W.R. Meehan. Am. Fish. Soc. Spec. Publ. No. 19. pp. 112133.
Blanchfield, P. J., and Ridgeway,M. S. 1997. Reproductive timing and use of redd sites
by lake-spawning brook trout {Salvelinus fontinalis). Can. J. Fish. Aquat. Sci. 54:
747-756.
Boag, T.D., and Hvenegaard, P.J. 1997. Spawning movements and habitat selection of
bull trout in a small Alberta foothills stream. Friends of the bull trout conference
proceedings. Bull Trout Task Force (Alberta), Trout Unlimited Canada, Calgary.
Ab. pp. 317-323

122

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Brannon, E.L. 1987. Mechanisms stabilizing salmonid fry emergence timing. In Sockeye
salmon (Oncorhynchus nerka) population biology and future management. Edited
by H.D. Smith, L. Margolis, and C.C. Wood. Can. Spec. Publ. Fish. Aquat. Sci.
No. 96. pp. 120-124.
British Columbia Conservation Data Centre. 2006. B.C. Species and Ecosystems
Explorer. [Online] available from: http://srmapps.gov.bc.ca/apps/eswp/ [accessed
September 2006].
Carson, R.J. 2001. Bull trout spawning movements and homing behaviour back to pre­
spawning locations in the McLeod River, Alberta. Bull Trout II Conference
Proceedings; Ecology and Management of Northwest Salmonids. Trout Unlimited
Canada, Calgary, Ab., pp. 137-141.
Cavender, T.M. 1978. Taxonomy and distribution of the bull trout, Salvelinus confluentus
(Suckley), from the American Northwest. Calif. Fish. Game 64: 139-174.
Chandler, J.A., Fedora M.A., and Walters, T.R. 2001. Pre- and post-spawn movements
and spawning observations of resident bull trout in the Pine Creek watershed,
eastern Oregon. Bull Trout II Conference Proceedings. Trout Unlimited Canada,
Calgary, Ab. pp. 167-172
Chapman, D.W. 1988. Critical review of variables used to define effects of fines in redds
of large salmonids. Trans. Am. Fish. Soc. 117: 1-21.
Combs, B.D. 1965. Effect of temperature on the development of salmon eggs. Prog. FishCult. 27:134-137.
Cope, R.S., and Macdonald, J.S. 1998. Responses of sockeye salmon (Oncorhynchus
nerka) embryos to intragravel incubation environments in selected streams within
the Stuart-Takla watershed. Proceedings of the Forest-Fish Conference, Calgary,
Ab. pp. 283-294.
Cunjak, R.A., Power, G., and Barton, D.R. 1986. Reproductive habitat and behaviour of
anadromous arctic char (Salvelinus alpinus) in the Koroc River, Quebec. Nat.
Can. 113: 383-387.
Curry, R.A., Noakes, D.A., and Morgan, G.E. 1995. Groundwater and the incubation and
emergence of brook trout {Salvelinus fontinalis). Can. J. Fish. Aquat. Sci. 52:
1741-1749.
Dent, C.L., Schade, J.D., Grimm, N.B., and Fisher, S.G. 2000. Subsurface influences on
surface biology. In Streams and ground waters. Edited by J.B. Jones and P J.
Mulholland. Academic Press, San Diego, Ca. pp. 381-402.
De March, B.G.E. 1995. Effects of incubation temperature on the hatching success of
arctic char eggs. Prog. Fish-Cult. 57: 132-136.

123

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

DeVries, P. 1997. Riverine salmonid egg burial depths: review o f published data and
implications for scour studies. Can. J. Fish. Aquat. Sci. 54: 1685-1698.
Dill, L.M. 1967. Studies on the early feeding of sockeye alevins. Can. Fish. Cult. 39: 2334.
Dittman, A.H., and Quinn, T.P. 1996. Homing in Pacific salmon: mechanisms and
ecological basis. J. Exp. Biol. 199: 83-91.
Essington, T.E., Sorensen, P.W., and Paron, D.G. 1998. High rate of redd
superimposition by brook trout {Salvelinus fontinalis) and brown trout (Salmo
trutta) in a Minnesota stream cannot be explained by habitat availability alone.
Can. J. Fish. Aquat. Sci. 55: 2310-2316.
Fleming, I.A., and Reynolds, J.D. 2004. Salmon breeding systems. In A.P. Hendry and
S.C. Steams. Evolution Illuminated, Salmon and Their Relatives. Edited by A.P.
Hendry and S.C. Steams. Oxford University Press, New York, N.Y.
Fraley, J.J., and Shepard, B.B. 1989. Life history, ecology and population status of
migratory bull trout {Salvelinus confluentus) in the Flathead Lake and River
system, Montana. Northwest Sci. 6: 133-143.
Freeze, R.A., and Cherry, J.A. 1979. Groundwater. Prentice-Hall Inc., Englewood Cliffs,
N.J.
Garrett, J.W., Bennett, D.H., Frost, F.O., and Thurow, R.F. 1998. Enhanced incubation
success for kokanee spawning in groundwater upwelling sites in a small Idaho
stream. N. Am. J. Fish. Manag. 18: 925-930.
Geist, D.R. 2000. Hyporheic discharge of river water into fall chinook salmon
{Oncorhynchus tshawytscha) spawning areas in the Hanford Reach, Columbia
River. Can. J. Fish. Aquat. Sci. 57: 1647-1656.
Geist, D.R. 2001. Senior Research Scientist. Battelle, Pacific Northwest Research
Laboratory. Richmond, WA.
Geist, D.R., and Dauble, D.D. 1998. Redd site selection and spawning habitat use by fall
chinook salmon: the importance o f geomorphic features in large rivers. Environ.
Manage. 22: 655-669.
Geist, D.R., Joy, M.C., Lee, D.C., and Gonser, T. 1998. A method for installing
piezometers in large cobble bed rivers. Ground Water Monit. Rem. 18: 78-82.
Goetz, F.A. 1997. Distribution of bull trout in Cascade Mountain streams of Oregon and
Washington. Friends of the Bull Trout Conference Proceedings. Bull Trout Task
Force Calgary, Ab., pp. 237-248.

124

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Goetz, F. 1989. Biology of the bull trout Salvelinus confluentus a literature review.
Willamette National Forest. Eugene, Or.
Gould, W.R. 1987. Features in the early development of bull trout (Salvelinus
confluentus) Northwest Sci. 61: 264-268.
Groot, C. and Margolis, L. 1991. Pacific Salmon Life Histories. University of British
Columbia Press, Vancouver, B.C.
Haas, G.R., and McPhail, J.D. 1991. The systematics, zoogeography and evolution of
Dolly Varden {Salvelinus malma) and bull trout {Salvelinus confluentus) in North
America. Can. J. Fish. Aquat. Sci. 48: 2191-2211.
Hamor, T., and Garside, E.T. 1976. Developmental rates of embryos of Atlantic salmon,
Salmo salar L., in response to various levels of temperature, dissolved oxygen
and water exchange. Can. J. Zool. 54: 1912-1917.
Hansen, E.A. 1975. Some effects of groundwater on brown trout redds. Trans. Am. Fish.
Soc. 1: 100-110.
Heath, D.D., Fox, C.W., and Heath, J.W. 1999. Maternal effects on offspring size:
variation through early development o f chinook salmon. Evolution. 53: 16051611.
Heming, T.A. 1982. Effects of temperature on utilization o f yolk by chinook salmon
{Oncorhynchus tshawytscha) eggs and alevins. Can. J. Fish. Aquat. Sci. 39: 184190.
Heming, T.A., Mclnemey, J.E., and Alderdice, D.F. 1982. Effect o f temperature on initial
feeding in alevins of chinook salmon {Oncorhynchus tshawytscha ). Can. J. Fish.
Aquat. Sci. 39: 1554-1562.
Hendry, A.P., Hensleigh, J.E., and Reisenbichler, R.R. 1998. Incubation temperature,
developmental biology, and the divergence of sockeye salmon {Oncorhynchus
nerka) within Lake Washington. Can. J. Fish. Aquat. Sci. 55: 1387-1394.
Hengeveld, H.G. 1990. Global climate change: implications for air temperature and water
supply in Canada. Trans. Am. Fish. Soc. 119: 176-182.
Holtby, L.B. 1988. Effects of logging on stream temperature in Carnation Creek, British
C olum bia, an d asso ciated im pacts o n coho salm on { O n c o r h y n c h u s k is u tc h ) C an.

J. Fish. Aquat. Sci. 45: 502-515.
Humpesch, U.H. 1985. Inter- and intra-specific variation in hatching success and
embryonic development of five species of salmonids and Thymallus thymallus.
Arch. Hydrobiol. 104: 129-144.

125

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Hutchings, J.A. 1991. Fitness consequences of variation in egg size and food abundance
in brook trout Salvelinus fontinalis Evolution. 45: 1162-1168.
Huuskonen, H., Penttinen, O., and Piironen, J. 2003. Effects of temperature and parental
background on the embryonic survival and metabolic rate of newly hatched Arctic
charr. Proceedings of the 26th Annual Larval Fish Conference. 2003. Institute of
Marine Research, Bergen, No.
Ihssen, P. 1978. Inheritance of bluesac disease for hatchery charr of the genus Salvelinus
Environm. Biol. Fishes 3: 317-320.
Jaworski, A., and Kamler, E. 2002. Development of a bioenergetics model for fish
embryos and larvae during the yolk feeding period. J. Fish. Biol. 60: 785-809.
Johnston, I. A. 2001. Genetic and environmental determinants of muscle growth patterns.
In Muscle development and growth, Fish Physiology Volume 18. Edited by I.A.
Johnston. Academic Press, New York, N.Y. pp. 141-186.
Johnston , I.A., Alderson, D., Sandham, C., Dingwall, A., Mitchell, D., Selkirk, C.,
Nikell, B., Baker, R., Roberston, B., Whyte, A., and Springate, J. 2000a. Patterns
of muscle growth in early and late maturing populations of Atlantic salmon
{Salmo salar L.). Aquaculture. 189: 307-333.
Johnston , I.A., Alderson, D., Sandham, C., Dingwall, A., Mitchell, D., Selkirk, C.,
Nikell, B., Baker, R., Roberston, B., Whyte, A., and Springate, J. 2000b. Muscle
fibre density in relation to the colour and texture of smoked Atlantic salmon
{Salmo salar L.). Aquaculture 189: 335-349.
Kieman, J.A. 1999. Histological and histochemical methods: theory and practice. 3rd
ed. Butterworth Heinemann, Boston, Ms.
Kinnison, M., Unwin, M., Boustead, N., and Quinn, T.P. 1998. Population-specific
variation in body dimensions o f adult chinook salmon {Oncorhynchus
tshawytscha) from New Zealand and their source population, 90 years after
introduction. Can. J. Fish. Aquat. Sci. 55: 554-563.
Kitano, S., Maekawa, K., Nakano, S., and Fausch, K.D. 1994. Spawning behaviour of
bull trout in the upper Flathead Drainage, Montana, with special reference to
hybridization with bull trout. Trans. Am. Fish. Soc. 123: 988-992.
L ee, D .R ., an d C herry, J.A . 1978. A field exercise on g ro undw ater flow u sin g seepage

meters and mini-piezometers. J. Geol. Educ. 27: 6-10.
Leman, V.N. 1993. Spawning sites of chum salmon, Oncorhynchus keta:
microhydrological regime and viability o f progeny in redds (Kamchatca River
Basin). J. Ichthyol. 33: 104-117.

126

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Markevich, N.B., and Bilenskaya, N.I., 1992. Viability and weight growth of sockeye
salmon young (Oncorhynchus nerka) completing the early stages of
embryogenesis at various temperatures. J. Ichthyol. 32: 105-118.
Markle, D.F. 1992. Evidence of bull trout x brook trout hybrids in Oregon. Proceedings
of the Gearhart Mountain bull trout workshop. American Fisheries Society,
Oregon Chapter, Corvallis, Or. pp. 58-67.
McCart, P. 1997. Bull trout in Alberta, a review. Friends of the Bull Trout Conference
Proceedings. Bull Trout Task Force, Calgary, Ab. pp. 191-207.
McPhail, J.D., and Baxter, J.S. 1996. A review of bull trout (Salvelinus confluentus) life
history and habitat use in relation to compensation and improvement
opportunities. Fisheries Management Rep. No. 104, Province of British Columbia,
Victoria, B.C.
McPhail, J.D., and Murray, C.B. 1979. The early life history and ecology of Dolly
Varden {Salvelinus malma) in the upper Arrow Lakes. Report o f the Institute of
Animal Resource Ecology and Department of Zoology, University of British
Columbia, to B.C. Hydro and Ministry of Environment, Victoria, B.C.
Merz, J.E., Setka, J.D., Pasternack, G.B., Wheaton, J.M. 2004. Predicting benefits of
spawning-habitat rehabilitation to salmonid {Oncorhynchus spp.) fry production
in a regulated California river. Can. J. Fish. Aquat. Sci. 61: 1433-1446.
Murray, C.B., and J.D. McPhail. 1988. Effects of incubation temperature on development
of five species of Pacific salmon {Oncorhynchus) embryos and alevins. Can. J.
Zool. 66: 266-273.
Nathanailides, C., Lopez-Albors, O., Stickland, N.C., 1995. Influence on pre-hatch
temperature on the development o f muscle cellularity in post-hatch Atlantic
salmon Salmo salar. Can. J. Fish. Aquat. Sci. 52: 675-680.
O'Brien, D.S., and Zimmerman, J.T. 2000. Davis River bull trout study 1999 radio
telemetry progress report. Ministry of Environment Land and Parks, Fisheries
Branch, Prince George, B.C.
Ojanguren, A.F., Reyes-Gavilan, F.G., and Munozl, R.R. 1999. Effects of temperature
on growth and efficiency of yolk utilisation in eggs and pre-feeding larval stages
of Atlantic Salmon. Aquacult. Int. 7: 81-87
Peterson, R.H., Spinney, H.C.E., and Sreedharan., A. 1977. Development of Atlantic
salmon {Salmo salar) eggs and alevins under varied temperature regimes. J. Fish.
Res. Board Can. 34: 31-43.

127

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Pillipow, R. and Williamson, C. 2004. Goat River bull trout {Salvelinus confluentus)
biotelemetry and spawning assessments 2002-03. B.C. J. Ecosys. Manage. 4: 2937.
Power, G., Brown, R.S., and Louie, D. 1999. Groundwater and fish - insights from
northern North America. Hydrol. Proc. 13: 401-422.
Prince, A., and Powell, C. 2000. Clove oil as an anaesthetic for invasive field procedures
on adult rainbow trout. North Am. J. Fish. Manage. 20: 1029-1032.
Quinn, T.P. 1984. Homing and straying in Pacific salmon. In Mechanisms of migration in
fishes. Edited by J. D. McCleave, G. P. Arnold, J. J. Dodson, and W. H. Neill, pp.
357-362. Plenum Press, New York., N.Y.
Quinn T.P., Hendry, A.P., Wetzel., L.A., De Roos, A.M., Sabelis, M.W. 1995. The
influence of life history trade-offs and the size of incubation gravels on egg size
variation in sockeye salmon {Oncorhynchus nerka). Oikos. 74: 347-424.
Quinn, T.P., Wetzel, L., Bishop, S., Overberg, K., and Rogers, D.E. 2001. Influence of
breeding habitat on bear predation and age at maturity and sexual dimorphism of
sockeye salmon populations. Can. J. Zool. 79: 1782-1793.
Rasband, W.S., 1997-2006. ImageJ, U. S. National Institutes of Health, Bethesda,
Maryland, USA, [online] Available from http://rsb.info.nih.gov/ij/. [accessed
January 2005]
Ridgeway, M.S., and Blanchfield, P J. 1998. Brook trout spawning areas in lakes. Ecol.
Freshw. Fish. 7: 140-145.
Rieman, B.E., and McIntyre, J.D. 1995. Occurrence of bull trout in naturally fragmented
habitat patches of varied size. Trans. Am. Fish. Soc. 124: 285-296.
Ripley, T.D., Boyce, M., and Scrimgeour, G J . 2004. Effects of industrial activity on bull
trout populations in Alberta’s boreal forest: an evaluation of current and future
impacts. Proceedings Forest-Land-Fish Conference II. Edmonton, Ab. pp. 41-42
Ripley, T.D., Scrimgeour, G.J., and Boyce, M. 2005. Bull trout {Salvelinus confluentus)
occurrence and abundance influenced by cumulative industrial developments in a
Canadian boreal forest watershed. Can. J. Fish. Aquat. Sci. 62: 2431-2442.
Rombough, P.J. 1997. The effects of temperature on embryonic and larval development.
In Global Warming: Implications for freshwater and marine fish. Edited by C.M.
Wood and D.G. McDonald, Cambridge University Press, Cambridge, U.K. pp
177-223.
Salvanes, A.G.V., Aksnes, D.L., and Giske, J. 1995. A surface-dependent gastric
evacuation model for fish. J. Fish Biol. 47: 679.

128

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Scholz, A., McClellan, H., Geist, D.R., Brown, R.S. 2005. Investigations of migratory
bull trout {Salvelinus confluentus) in relation to fish passage at Albeni Falls Dam.
Battelle Pacific Northwest Division, Richland, WA. Final report to Department
of the Army, Seattle District Corps of Engineers, Report No. PNWD-3521.
Seattle, Wa.
Shrimpton, J.M., Bourgeois, J.F., and Blouw, D.M. 1998. Forest harvesting increases
stream temperature: impacts on juvenile chinook salmon. In Stress in Fish. Edited
by T.P. Barry, B.A. Barton and D.D. MacKinlay. American Fisheries Society.
Bethesda, Md. pp. 77-82.
Spruell, P., Hemmingsen, A.R., Howell, P.J., Kanda, N. and Allendorf, F.W. 2003.
Conservation genetics of bull trout: geographic distribution o f variation at micro
satellite loci. Cons. Genetics. 4: 17-29.
Sokal, R.R., and Rohlf, F.J. 1995. Biometry: The Principles and Practice of Statistics in
Biological Research. W.H. Freeman and Co., New York, N.Y.
Steen, R.P., and Quinn, T.P. 1999. Egg burial depth by sockeye salmon {Oncorhynchus
nerka): implications for survival of embryos and natural selection on female body
size. Can. J. Zool. 77: 836-841.
Swanberg, T.R. 1997. Movements of bull trout {Salvelinus confluentus) in the Clark Fork
River system after transport upstream of Milltown Dam. Northwest Sci. 71: 313317.
Tabachnick, B.G., and Fidell, L.S. 2001. Using Multivariate Statistics. Needham Heights,
Allyn and Bacon. Boston, Ma.
Tang, J., Bryant, M.D., and Brannon, E.L. 1987. Effect o f temperature extremes on the
mortality and development rates of coho salmon embryos and alevins. Prog. FishCult. 49: 167-174.
Thomas, J.E., Groft, D., Chew, L., and Clayton, T. 2001. Genetic variation in bull trout
{Salvelinus confluentus) from Alberta and northern Montana. Bull Trout II
Conference Proceedings. Trout Unlimited Canada, Calgary, Ab. pp. 117-118.
Thurow, R.F., Lee, D.C., and Rieman, B.E. 1997. Distribution and status of seven native
salmonids in the interior Columbia River Basin and portions of the Klamath River
and Great Basins. North Am. J. Fish. Manage. 17: 1094-1110.
Watson, W., and Hillman, T.W. 1997. Factors affecting the distribution and abundance of
bull trout: an investigation at hierarchical scales. North Am. J. Fish. Manage. 17:
237-252.

129

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Weaver, T.M., and White, R.G. 1985. Coal Creek fisheries monitoring study number III.
Contract No. 53-0385-3-2685. U.S. Forest Service Flathead National Forest,
Kalispell, Mt.
Webb, J.H., and McLay, H.A. 1996. Variations in the time of spawning of Atlantic
salmon {Salmo salar) and its relationship to temperature in the Aberdeenshire
Dee, Scotland. Can. J. Fish. Aquat. Sci. 53: 2739-2744.
Webster, D.A., and Eiriksdottir, G. 1976. Upwelling water as a factor influencing choice
of spawning sites by brook trout {Salvelinus fontinalis). Trans. Am. Fish. Soc.
105:416-421.
Withler, R.E. 1987. Genetic variation in survival of chinook salmon {Oncorhynchus
tshawytscha) alevins exposed to an unidentified agent of mortality. Genome. 29:
839-845
Wood, C.M., and McDonald, D.M. Editors. 1997. Global Warming - Implications for
Freshwater and Marine Fish. Cambridge Seminar Series No. 61.

130

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

APPENDIX A: ANOVA TABLES FOR ANALYSIS OF SURVIVAL.
Table 10. ANOVA table for year one analysis of survival, which includes the factor
family in the test.
Source
Date
Reach
Family
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
2
3
5
13
99

Adj SS
0 .94979
0.18221
1.47025
0.37464
0.46330
7 .45723

Adj MS
0 .94979
0.09110
0.49008
0.07493
0.03564
0.07533

F
12 .61
1.21
6.51
0.99
0.47

P
0.001
0.303
0.000
0.425
0. 935

Table 11. ANOVA table for year two analysis of survival, which includes the factor
family in the test.
Source
Date
Reach
Family(Reach)
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
5
11
12
20
148

Adj SS
0.96516
1.57559
2.46987
0. 63453
0.44578
3.42766

Adj MS
0.96516
0.31512
0.22453
0.05288
0.02229
0.02316

F
41..67
13..61
9..69
2..28
0..96

P
0,.000
0..000
0,.000
0 ..011
0..511

Table 12. ANOVA table for year two analysis of survival, which includes the factors
male and female in the test.
DF
1
2
1
1
5
13
101

Source
Date
Reach
Male
Female
Redd(Reach)
Capsule(Reach Redd)
Error

Adj SS
1.02249
0.24930
0.25196
0.09818
0.40529
0.34446
0.09150

Adj MS
1.02249
0.12465
0.25196
0.09818
0.08106
0.02650

F
11.17
1.36
2.75
1 .07
0.89
0.29

P
0.001
0.261
0.100
0.303
0.494
0. 992

Table 13. ANOVA table for year two analysis of survival, which includes the factor
male in the test.
Source
Date
Reach
Male(Reach)
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
5
5
12
20
154

Adj SS
0.98004
1.50021
1.97723
0.56639
0.43543
3.92030

Adj MS
0.98004
0.30004
0.39545
0.04720
0.02177
0.02546

F
38.50
11.79
15.53
1.85
0.86

P
0.000
0.000
0.000
0.044
0.643

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

APPENDIX A: CONTINUED
Table 14. ANOVA table for year two analysis of survival including the factor female
the test.
Source
Date
Reach
Female(Reach)
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
5
5
12
20
154

Adj SS
0.94675
1.50506
0.36987
0.58192
0.40301
5.52766

Adj MS
0.94675
0.30101
0.07397
0.04849
0.02015
0.03589

F
26.38
8.39
2.06
1.35
0.56

P
0.000
0.000
0.073
0.196
0. 933

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

APPENDIX B: ANOVA TABLES FOR ANALYSIS OF GROUNDWATER
CHEMISTRY.
Table 15. ANOVA tables for examination of variance in specific discharge within three
collection periods (September, March and April, respectively).
Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
22

Adj SS
0.0014488
0.0024253
0.0067749
0.0050423

Adj MS
0.0014488
0.0006063
0.0011291
0.0002292

F
6.32
2.65
4.93

P
0.020
0.061
0.002

Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
22

Adj SS
0.04259
0.41181
1.75580
0.99818

Adj MS
0.04259
0.10295
0.29263
0.04537

F
0.94
2 .27
6.45

P
0.343
0.094
0.000

Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
18

Adj SS
0.015153
0.039049
0.137054
0.096133

Adj MS
0.015153
0.009762
0.022842
0.005341

F
2.84
1.83
4.28

P
0.109
0.167
0.008

Table 16. ANOVA tables for examination of variance in specific conductivity within
three collections periods (September, March and April, respectively).
Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
23

Adj SS
0.49779
1.09618
0.06682
0.16928

Adj MS
0.49779
0 .27405
0.01114
0.00736

F
67 .64
37.23
1.51

P
0.000
0.000
0.218

Source
Use
Reach (Use)
Piezometer(Use Reach)
Error

DF
1
4
6
23

Adj SS
1.68652
2 .32312
0.06717
0 .60278

Adj MS
1.68652
0.58078
0.01120
0.02621

F
64.35
22.16
0.43

P
0.000
0.000
0.853

Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
18

Adj SS
0.1669
47.4355
0.8186
3.6125

Adj MS
0 .1669
11.8589
0. 1364
0.2007

F
0.83
59.09
0. 68

P
0.374
0.000
0. 668

133

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 17. ANOVA tables for examination o f variance in oxygen concentration within
three collections periods (September, March and April, respectively).
Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
24

Adj SS
4.0778
7.3322
2.3950
3.5890

Adj MS
4.0778
1.8331
0.3992
0.1495

F
27.27
12.26
2.67

Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
23

Adj SS
0.88602
1.05606
0.27555
0.91292

Adj MS
0.88602
0.26402
0.04592
0.03969

Source
Use
Reach(Use)
Piezometer(Use Reach)
Error

DF
1
4
6
18

Adj SS
0.005629
0.157701
0.009778
0.034778

Adj MS
0.005629
0.039425
0.001630
0.001932

P
0. 000
0. 000
0. 040

F
22.32
6.65
1.16

F
2.91
20.41
0.84

P
0.000
0.001
0.363

P
0.105
0.000
0.553

Table 18. ANOVA tables for analysis of variance in specific discharge, specific
conductivity, and oxygen concentration respectively, between two collections periods
(September April).
Specific Discharge
Source
Use
Reach(Use)
Depth(Use Reach)
Error

Adj SS
2.8228
17.9182
4.7837
20.4166

Adj MS
2.8228
4.4796
0.7973
0.9280

F
3.04
4.83
0.86

DF
1
4
11
21

Adj SS
2.31614
1.27799
1.41522
1.83416

Adj MS
2.31614
0.31950
0.12866
0.08734

F
26.52
3.66
1.47

P
0.000
0.021
0.214

DF
1
4
12
22

Adj SS
0.01615
1. 56149
0.33695
0.60113

Adj MS
0.01615
0.39037
0.02808
0.02732

F
0.59
14 .29
1.03

P
0.450
0.000
0.459

DF
1
4
6
22

P
0,.095
0,.006
0,.540

Specific Conductivity
Source
Use
Reach(Use)
Depth(Use Reach)
Error
Oxygen
Source
Use
Reach(Use)
Depth(Use Reach)
Error

134

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

APPENDIX C: ANOVA TABLES FOR ANALYSIS OF EMBRYO SIZE AND
YOLK UTILIZATION.
Table 19. ANOVA table for year one analyses of length with factors for male and female
parent included in the test.
Source
Date
Female
Male
Reach
Female*Male
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
1
1
2
1
5
13
1293

Adj SS
17928.3
5.8
7.8
2666.8
3.5
78.7
56.2
1827.2

Adj MS
17928.3
5.8
7.8
1333.4
3.5
15.7
4.3
1.4

F
12686.93
4.08
5.51
943.58
2 .47
11.13
3.06

P
0 ..000
0 .,044
0 .,019
0.,000
0 .,116
0 .,000
0 .,000

Table 20. ANOVA table for year one analyses of total dry weight with
and female parent included in the test.
Source
Date
Female
Male
Female*Male
Reach
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
1
1
1
2
5
13
1246

Adj SS
0.0037224
0.0000156
0.0000007
0.0039215
0.0008054
0.0001078
0.0001322
0.0045192

Adj MS
0.0037224
0.0000156
0.0000007
0.0039215
0.0004027
0.0000216
0.0000102
0.0000036

F
1026.,31
4 .,31
0 .,20
1081.,20
111.,03
5.,95
2 .,80

P
0.000
0 .038
0 .657
0.000
0.000
0.000
0.001

Table 21. ANOVA tables for year one analyses of dry-body weight with factors for male
and female parent included in the test.
Source
Date
Female
Male
Female*Male
Reach
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
1
1
1
2
5
13
1265

Adj SS
0.434128
0.000644
0.004301
0.000253
0.093867
0.002351
0.006025
0.092059

Adj MS
0.434128
0.000644
0.004301
0.000253
0.046933
0.000470
0.000463
0.000073

F
5965.43
8.85
59.10
3.47
644.92
6.46
6.37

P
0.000
0.003
0.000
0.063
0.000
0.000
0.000

Table 22. ANOVA tables for year one analyses of dry yolk weight with factors for male
and female parent included in the test.
Source
Date
Female
Male
Female*Male
Reach
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
1
1
1
2
5
13
1278

Adj SS
0.0199154
0.0000014
0.0000873
0.0039291
0.0049296
0.0001863
0.0002479
0.0068176

Adj MS
0 .0199154
0.0000014
0.0000873
0.0039291
0.0024648
0.0000373
0.0000191
0.0000053

F
3733.28
0.26
16.36
736.53
462.04
6.99
3.57

P
0.000
0. 608
0.000
0.000
0.000
0.000
0 .000

135

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 23. ANOVA tables for year two analyses of length with the factor family included
in the test.
Source
Date
Reach
Family(Reach)
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
5
11
12
20
702

Adj SS
2833..42
1682..13
78..61
66..88
63..10
559..57

Adj MS
2833..42
336,.43
7 .15
5..57
3,.16
0,.80

F
3554..63
422.,06
8..97
6..99
3..96

P
0,.000
0,.000
0,.000
0 ,.000
0,.000

Table 24. ANOVA tables for year two analyses of total dry weight with the factor family
included in the test.
Source
Date
Reach
Family
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
5
4
12
20
701

Adj SS
0.0006681
0.0025632
0.0006643
0.0000914
0.0000708
0.0022147

Adj MS
0.0006681
0.0005126
0.0001661
0.0000076
0 .0000035
0.0000032

F
211.47
162.26
52.57
2.41
1.12

P
0.000
0.000
0.000
0.005
0.322

Table 25. ANOVA tables for year two analyses o f dry-body weight with the factor
family included in the test.
Source

DF

Adj SS

Adj MS

F

P

Date
Reach
Family
Redd(Reach)
Capsule(Reach Redd)
Error

1
5
4
12
20
718

0.0036610
0.0018186
0.0000052
0.0000778
0.0000519
0.0008321

0.0036610
0.0003637
0.0000013
0.0000065
0.0000026
0.0000012

3158.99
313.85
1.13
5.59
2.24

0.000
0.000
0.342
0.000
0.002

Table 26. ANOVA tables for year two analyses of dry yolk weight the factor family
included in the test.
Source
Date
Reach
Family
Redd(Reach)
Capsule(Reach Redd)
Error

DF
1
5
4
12
20
702

Adj SS
0.0072683
0.0078339
0.0007359
0.0001459
0.0001558
0.0026085

Adj MS
0.0072683
0.0015668
0.0001840
0.0000122
0.0000078
0.0000037

F
1956.05
421.65
49.51
3.27
2.10

P
0.000
0.000
0.000
0.000
0.003

136

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

APPENDIX D: ANOVA TABLES FOR ANALYSIS OF LABORATORY
GROWTH.
Table 27. ANOVA table for the laboratory analysis of length.
Source
Date
Temperature
Female
Male
Female*Male
Tray(Temperature)
Error

DF
8
2
1
1
1
9
609

Adj SS
5271.12
2308.54
22 .60
0.29
0 .02
42 .81
348.00

Adj MS
658 .89
1154 .27
22 .60
0 .29
0 .02
4 .76
0 .57

F
1153.07
2020.00
39.56
0.50
0.04
8.32

P
0..000
0..000
0..000
0,.479
0..847
0..000

Table 28. ANOVA table for the laboratory analysis of dry-body w
Source
Date
Temperature
Female
Male
Female*Male
Tray(Temperature)
Error

DF
8
2
1
1
1
9
618

Adj SS
0.0042494
0.0021008
0.0000029
0.0000002
0.0000034
0.0000360
0.0007440

Adj MS

0. 0005312
0. 0010504
0. 0000029
0. 0000002
0. 0000034
0. 0000040
0. 0000012

F
441.19
872.45
2.39
0. 17
2 .82
3.32

P

0.000
0.000
0.123
0. 684
0 .094
0.001

Table 29. ANOVA table for the laboratory analysis of dry-yolk weight.
Source
Date
Temperature
Female
Male
Female*Male
Tray(Temperature)
Error

DF
8
2
1
1
1
9
618

Adj SS
0.0148260
0.0059096
0.0005417
0.0000001
0.0000105
0.0000850
0.0019498

Adj MS
0.0018532
0.0029548
0.0005417
0.0000001
0.0000105
0.0000094
0.0000032

F
587.39
936.52
171.68
0 .03
3.32
2.99

P
0.000
0.000
0.000
0.855
0.069
0.002

Table 30. Four ANOVA tables for the laboratory analysis of muscle fibres for myomeres
1-3 and 4 & 5 combined, respectively.
Myomere 1
Source
Treatment
Male
Female
Stage(Treatment)
Tray(Treatment)
Error
M yom ere

DF
2
1
1
3
9
97

Adj SS
85524
527
1499
10312
49135
596752

Adj MS
42762
527
1499
3437
5459
6152

F
6..95
0..09
0..24
0..56
0..89

P
0..002
0..770
0..623
0..644
0..539

DF
2
1
1
3
9
97

Adj SS
62856
6121
13028
1606
61611
445453

Adj MS
31428
6121
13028
535
6846
4592

F
6.84
1.33
2.84
0.12
1.49

P
0.002
0.251
0.095
0.950
0.162

2

Source
Treatment
Male
Female
Stage(Treatment)
Tray(Treatment)
Error

137

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.

Table 30 continued.
Myomere 3
Source
Treatment
Male
Female
Stage(Treatment)
Tray(Treatment)
Error

DF Adj SS
2 82889
892
1
1 18018
8617
3
9 80194
97 584033

Adj MS
41444
892
18018
2872
8910
6021

F
6.88
0.15
2.99
0.48
1.48

Adj MS
246313
109
46645
28088
41709
12265

F
20.08
0.01
3.80
2.29
3.40

P
0.002
0.701
0.087
0.699
0.166

Myomeres 4 and 5 Combined
Source
Treatment
Male
Female
Stage(Treatment)
Tray(Treatment)
Error

DF Adj SS
2 492627
1
109
1
46645
84264
3
9 375380
97 1189692

0
0
0
0
0

R ep ro d u ced with p erm ission o f th e copyright ow ner. Further reproduction prohibited w ithout perm ission.