The Salmon Disturbance Regime: Effects on Biofilm, Sediment and Water

by

Sam J. Albers
BSc. University of Victoria, 2006

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE (ENVIRONMENTAL SCIENCE)
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA

December 2010

©SamAlbers, 2010

1*1

Library and Archives
Canada

Bibliotheque et
Archives Canada

Published Heritage
Branch

Direction du
Patrimoine de I'edition

395 Wellington Street
Ottawa ON K1A 0N4
Canada

395, rue Wellington
Ottawa ON K1A 0N4
Canada
Your file Votre refSrence
ISBN: 978-0-494-75137-4
Our file Notre reference
ISBN: 978-0-494-75137-4

NOTICE:

AVIS:

The author has granted a nonexclusive 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 noncommercial purposes, in microform,
paper, electronic and/or any other
formats.

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 prophete 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.

1+1

Canada

Abstract

n

Abstract

Recent work in salmon spawning streams has shown that sediment resuspended during nest construction aggregates with salmon organic matter to form suspended particles
called floes. These nutrient-rich floes interact with streambed biofilms suggesting a potential floe trapping mechanism that drives biofilm growth. Using the Horsefly spawning
channel, the role of biofilms in trapping fine sediment was evaluated as a mechanism of
salmon-derived nutrient processing. In the active spawn period, biofilm was reduced
in abundance while the streambed sediment infiltration was at its highest level. During
salmon die-off, downstream biofilm abundance recovered to pre-spawn values indicating a nutrient pulse over a small scale. With the re-established biofilm layer, sediment
was increasingly trapped at the streambed surface by biofilms. This increase in biofilm
abundance will likely influence the nutrient dynamics at all levels of the stream foodweb.
Biofilms transfer increases in productivity to higher trophic levels. This transfer has a
positive effect on the next generation of juvenile salmon growth and survivorship.

Contents

in

Contents

Abstract

ii

Contents

iii

List of Tables

vi

List of Figures

viii

Acronyms

xi

Preface

xii

Acknowledgements

xiii

1

Salmon and Biofilm: A Literature Review
1.1 Introduction
1.2 Benthic Response
1.2.1 Biofilms
1.2.1.1 Active-Spawn
1.2.1.2 Post-Spawn
1.2.1.3 MDN contribution
1.2.1.4 Biofilm Composition
1.2.2 Particle Aggregation
1.2.3 Invertebrates
1.2.4 Resident & Juvenile Anadromous Fishes
1.2.5 Water Chemistry

1
1
4
5
7
8
8
9
10
11
12
13

2

Sediment and Biofilm Interactions in a Salmon Spawning Stream
2.1 Introduction
2.1.1 Theme One
2.1.2 Theme Two
2.2 Methods
2.2.1 Study Site
2.2.1.1 Horsefly Spawning Channel
2.2.1.2 Measurement of Site Characteristics
2.2.2 Study Design
2.2.2.1 Spatial and Temporal Controls
2.2.3 Salmon
2.2.4 Biofilms

15
15
16
16
17
17
17
18
19
19
19
21

Contents

2.3 Results
2.3.1 HFC Characteristics
2.3.2 Salmon Numbers
2.3.3 Biofilms
2.3.3.1 Surface Biofilms
2.3.3.2 Sub-surface Biofilms
2.3.3.3 Stable Isotopes
2.3.4 Infiltration Bags
2.3.4.1 Stable Isotopes
2.3.4.2 Particle Size
2.3.5 Suspended Sediment
2.3.6 Intergravel Oxygen
2.3.7 Correlations
2.4 Discussion
2.4.1 Streambed Benthic Response
2.4.2 Stable Isotope Analysis
2.4.2.1 Biofilms
2.4.2.2 Infiltration Bags
2.4.3 Infiltration and Trapping
2.4.4 Implications

21
21
22
23
23
24
25
25
26
26
27
29
31
31
34
34
38
38
39
39
41
43
44
45
48
49
50
52
57

Benthic Biofilm Composition in a Salmon Spawning Stream
3.1 Introduction
3.1.1 Confocal Laser Scanning Microscopy
3.1.2 Objective
3.1.3 Research Questions
3.1.3.1 Question One
3.1.3.2 Question Two
3.1.3.3 Question Three
3.1.4 Secondary Objectives
3.2 Methods
3.2.1 Site Characteristics
3.2.2 Biofilm Growth
3.2.2.1 Confocal Specifications
3.2.2.2 Stains and Fluorescent Markers
3.2.2.3 Image Analysis

59
59
60
61
62
62
62
62
63
63
63
63
64
65
65

2.2.5

2.2.6
2.2.7
2.2.8

3

2.2.4.1 Biofilm Collection
2.2.4.2 Biofilm Characterization
Infiltration Bags
2.2.5.1 Sediment Characterization
2.2.5.2 Particle Size
Suspended Sediment
Piezometers
Statistical Analysis
2.2.8.1 Correlations

IV

3.2.3
3.2.4
3.2.5

3.3

3.4

3.5
4

Contents

v

Nutrient Delivery Estimates from Salmon Carcass Decay
3.2.3.1 Carcass Removal
Assumptions
Data Analysis
3.2.5.1 Manipulations
3.2.5.2 Statistical Analysis

66
67
68
69
69
69
70
71
71
71
74
74
76
79
79
79
81
83
84

Results
3.3.1 Site Characteristics
3.3.2 Microscope Use and Image Analysis
3.3.3 Biofilm Component Patterns
3.3.4 Salmon Nutrient Composition
3.3.4.1 Nutrient Influence
3.3.5 Method Comparison
3.3.6 UNBC Confocal
Discussion
3.4.1 Biofilm Component Patterns
3.4.2 Nutrient Influence
3.4.3 Method Comparison
Conclusions

Conclusions and Management Implications
4.1 Conclusions
4.2 Management Implications

85
85
88

Bibliography

90

A Confocal Laser Scanning Microscopy (CLSM) software scripts
A.l Image J conversion of OIB files to stacked TIFF
A.2 Scion Image macros for the analysis of stacked TIFF files
A.3 R script for data processing

100
100
103
107

List of Tables

VI

List of Tables

1.1

1.2

1.3

Estimated salmon abundance by region. Values are percentages of the total
number of populations in each category. After Quinn (2005); Scheuerell
et al. (2005)
Summary of major biofilm studies in relation to salmon spawning. Unless otherwise indicated responses are either from a salmon presence/absence study or a before, during and after assessment of relevant variables
orboth. a
Summary of water chemistry responses to salmon spawning. NH^" and
soluble reactive phosphorus (SRP) generally increase in response to salmon
run. NO^ values generally increase in a salmon spawning stream but the
increase is not correlated with the arrival of salmon. All nutrient measures
are presented as comparison of concentrations (e.g. mg/1). Note that most
studies on biofilm and salmon have been conducted in coastal areas rather
than interior ecosystem.6

2.1 HFC Site Characteristics. Values represent mean values over the sampling
period. Unless otherwise indicated, values in parentheses are the standard
deviation of that parameter. Grain size measurements are from the B axis
of gravels collected in each experimental section
2.2 Salmon densities for historical usage of the Horsefly River spawning channel
(HFC) and the densities loaded into the experimental sections for this study.
Densities used for the HFC study were slightly higher than historical usage but there are examples of studies using similar densities (Table 1.2;
Chaloner et al. (2002))
2.3 Results from a two-way analysis of variance (ANOVA) of spatial (section)
and temporal (period) salmon treatments on surface chlorophyll a, ash-free
dry mass (AFDM) or inorganic sediment. Interaction contrasts are separated by a |. Contrasts are labelled by the first letter of the corresponding
section and the spawning period (U:Upstream; M:Middle; D:Downstream).
2.4 Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on surface <515N , S13C and the carbon:nitrogen molar
ratio from benthic biofilms. The minimum adequate model (MAM) only
included an interaction term for O N ratio. This interaction term was nonsignificant so only the main effects were tested for all models

4

6

14

27

31

32

38

List of Tables
2.5

vn

Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on 515N and 513C from infiltrated sediment. The minimum adequate model (MAM) included only the residual term for C:N
ratio indicating that neither spawning nor die-off had any effect on intergravel C:N ratios. Therefore, the C:N model summary is not included here.
Dashes (-) indicate a dropped parameter in the MAM
38
2.6 Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on suspended sediment variables. All models were determined using a minimum adequate model (MAM) (Crawley, 2007). OMR
refers to the organic matter ratio of suspended sediment. See section 2.2.6
for details
41
2.7 ANOVA table of spatial and temporal responses of intergravel dissolved
oxygen (DO) from piezometers
43
2.8 Means and marginal means of intergravel DO. Bold font indicates the
grand mean of the model (Table 2.7). Common letters indicate non-significance
in pairwise t-test's with Holm's correction. All values are mg l^1
44
2.9 Summary of literature stable isotope values and C:N of salmon flesh and
values from this study. All values are from adult salmon flesh
48

List of Figures

vm

List of Figures

1.1

1.2

2.1
2.2

2.3

2.4
2.5

2.6

Trophic linkages in salmon spawning system and the general life cycle of
anadromous salmon. The salmon disturbance regime encompasses both
the redd construction and the post-spawn die-off phases of the salmon life
cycle. Energy and nutrient transfers within the river are indicated by solid
lines. Dashed lines refer to the marine component of the salmon life cycle.
Modified from Wipfli et al. 1998
Diagram of a salmon spawning stream illustrating the relationship between
redd construction, biofilm resuspension and flocculation (Rex and Petticrew, 2008)
Location of the Horsefly River spawning channel (HFC) and position of the
experimental reach
Division of the HFC into experimental sections. Downstream spawning
salmon not part of the experiment were excluded from the experimental
section at the lower portion of the downstream reach by an additional steel
fence
Characterization of stream conditions at the HFC. (a) Precipitation; (b) Discharge measurements. Vertical solid lines divide sampling periods defined
by salmon activity.
Live and dead salmon counts by section. Vertical solid lines indicate divisions of the salmon period
Chlorophyll a from gravels sampled in the HFC over the course of a salmon
spawning event. Bar heights are mean values with error bars representing ± 1 standard error of the mean (SEM). Gravels were sampled at three
depths which are indicated by the panel heading. An • indicates a significant difference in the contrast test. A x symbol indicates a non-significant
contrast. The • symbol is an indicator of a non-significant differences in the
starting conditions. All surface means were contrasted according to Table
2.3
AFDM values from gravels sampled in the HFC over the course of a salmon
spawning event. AFDM was determined by ashing glass fiber filter (GFF)
filters. Bar heights are mean values with error bars representing ± 1 SEM.
Gravels were sampled at three depths which are indicated by the panel
heading. An • indicates a significant difference in the contrast test. A x
symbol indicates a non-significant contrast. The • symbol is an indicator of
a non-significant differences in the starting conditions. All surface means
were contrasted according to Table 2.3

2

11
18

20

28
30

33

35

List of Figures
Inorganic sediment values from gravels sampled in the HFC over the course
of a salmon spawning event. Inorganic was determined by ashing GFF filters. Bar heights are mean values with error bars representing ± 1 SEM.
Gravels were sampled at three depths which are indicated by the panel
heading. An • indicates a significant difference in the contrast test. A x
symbol indicates a non-significant contrast. The • symbol is an indicator of
a non-significant differences in the starting conditions. All surface means
were contrasted according to Table 2.3
2.8 Isotopic ratio of benthic biofilms and infiltrated sediment. Ratios moving
towards salmon flesh values indicate a marine nutrient source although
5l3C values can be confounded by photosynthetic processes (Staal et al.,
2007). Isotopic ratios were calculated as per equation 2.1. The far right
panel shows the carbon to nitrogen ratio of both biofilms and sediment.
Salmon flesh values were sampled from fresh salmon carcasses (n=4). Points
are mean values ± 1 SEM
2.9 Range of particle sizes deposited on the stream bed into infiltration bags
as measured by laser in-situ scattering and transmissometry (LISST). Top
panel (a) data points are the mean values ± 1 SEM. Lower panel (b) is the
cumulative distribution of particle size. Both figures are different visual
representations of the same data
2.10 Suspended sediment loads and sediment ratios. Suspended sediment was
sampled by an automatic ISCO water sampler placed in the rear portion
of each section. Loads were calculated by multiplying the concentration of
sediment by the discharge to get a total load. Mean values ± SEM
2.11 Particle size distributions of suspended sediment in the HFC. Background
suspended sediment particle sizes were sampled on July 14*\ 2010 while
the active-spawn particle size was taken during the HFC study on September 24 th , 2009
2.12 Visual difference in the biofilm abundance present on gravels in the upstream and middle sections. Gravels in the middle section are visibly reduced in biofilms and sediment while gravels in the upstream section,
grown in the absence of salmon, are noticeably thicker with biofilm and
sediment. Both images were taken on the same date during the activespawn period
2.13 Relationship between inorganic sediment and biofilms indicators, (a) Relationship between biofilm growth (chlorophyll a) and inorganic sediment
trapped by the biofilm from gravels sampled in the downstream surface
section, (b) Strong correlation between two measures of biofilm growth
suggest an in-stream nutrient source (Hunt and Perry, 1999). (c) Chlorophyll a versus intergravel inorganic sediment collected from infiltration
bags that has deposited into the streambed. Decreased surface biofilm
abundance results in larger masses of fine sediment infiltrating into the
gravelbed (Bag depth=0.30-m). All p-values <0.05

IX

2.7

36

37

40

42

43

46

47

List of Figures
2.14 Dual plots of biofilm (a) and infiltration bag (b) isotopic ratios. Section is
represented using different symbols while the sampling period representative by different error bars line styles. Means ± 1 SEM
2.15 Schematic of a potential mechanism of marine derived nutrient (MDN)
enrichment of the downstream section via the flocculation feedback loop
(Rex, 2009)
3.1

3.2
3.3

3.4

3.5
3.6

3.7

3.8

4.1

Schematic of the slide mounting system used in the HFC. Polycarbonate strips have been previously identified as suitable growth substrates for
confocal laser scanning microscopy (CLSM) analysis of biofilms (Lawrence
et al., 1998)
Site characteristics for the HFC over the CLSM sampling period. See section 2.2 for description of rainfall and discharge collection methods
A single channel example of the image analysis process of thresholding, dilation, and eroding to determine white pixel counts. Eroding and dilation
are important steps to remove noise of the CLSM image
Percent coverage of biofilm components on polycarbonate strips as measured by CLSM. Components are arranged into panels according to the
channel wavelength fluorescence. Error bars indicate SEM which is based
on replicate sample strips from separate tiles in each section. Each strip
was sub-sampled with the CLSM five times (i.e. Five fields of view) to account for spatial variability with the biofilm. These five fields of view were
averaged to give a single value for each tile. Strips from five tiles were
sampled for each sampling date
The number of salmon present, both live and dead, in the HFC during the
confocal study period
Salmon decay products as calculated from equations 3.1 - 3.5 and used for
comparison in Figure 3.7. Shaded portions of this figure represent the period between CLSM sampling dates (Sampling dates also indicated in Figure 3.4)
Relationship between salmon decay products O N ratio and the percent
coverage of biofilm components. Downstream decay product were calculated as the sum of the middle and downstream values. Annotations refer
to results of a Pearson's Product Moment
The left panel illustrated the relationship between spectrophotometrically
measured chlorophyll a and algal coverage measured by CLSM. The right
panel is the relationship between total percent coverage and AFDM
Historical Horsefly River Escapement. Stock enhancement via the HFC
may have contributed to high stocks in the mid-nineties although other
Department of Fisheries and Oceans (DFO) management practices also take
place within the Quesnel watershed (DFO, 2010)

x

51

54

64
72

73

75
76

77

78

78

89

Acronyms

XI

Acronyms
AFDM ash-free dry mass
MAM m i n i m u m adequate model
ANOVA analysis of variance
M D N marine derived nutrient
CLSM confocal laser scanning microscopy
OMR organic matter ratio
DFO Department of Fisheries and Oceans
SA surface area
DIN dissolved inorganic nitrogen
SEM standard error of the mean
DO dissolved oxygen
SRP soluble reactive phosphorus
DOC dissolved organic carbon
TDP total dissolved phosphorus
D O N dissolved organic nitrogen
TIN total inorganic nitrogen
EPS extracellular polymeric substances
TN total nitrogen
GFF glass fiber filter
TP total phosphorus
HFC Horsefly River spawning channel
TSP total soluble phosphorus
KJN Kjeldahl nitrogen
UNBC University
LISST laser in-situ scattering and transColumbia
missometry

of

Northern

British

Preface

xn

Preface

This thesis is divided into four chapters. Chapter 1 provides an overview on the relationship between biofilms and aquatic environments. Specifically, Chapter 1 reviews the ecological impacts of salmon spawning and die-off as a mechanism which structures biofilm
growth. Results presented in Chapter 2 test the hypothesis that biofilm abundance is
driven by in-stream flocculation and resultant streambed salmon nutrient delivery. This
salmon nutrient delivery is in turn driven by the overlap between salmon spawning activity and salmon die-off. A portion of Chapter 2 first appeared as Petticrew and Albers
(2010). Chapter 3 is the first known attempt to characterize changes biofilm components
driven by salmon die-off using confocal laser scanning microscopy. Chapter 4 provides
several conclusions based on the data presented in Chapters 2 and 3 in addition to placing
these findings in a broader context.

Acknowledgements

xni

Acknowledgements
I have many individuals to thank without whom, this project would have never succeeded. First and foremost, I thank my supervisor Ellen Petticrew for her guidance, mentorship, support and willingness to literally get her feet wet in a stream filled with salmon.
Ellen, your calm during some difficult days of my field season was an example to be followed, and an anchor to prevent complete panic. I have benefitted greatly from your
experience and intelligence and I sincerely thank you for this opportunity.
I would like to thank my committee of Dr. Joselito Arocena, Dr. Ian Droppo, Dr. John
Rex and the external reviewer Dr. Mark Wipfli. Each of you have provided significant
contributions throughout the course of this project. Special thanks goes to John Rex for
patiently passing on technical knowledge and developing a significant portion of the conceptual basis upon which this thesis rests.
This project would have not been possible without all the help I received in the field and
the lab. Special thanks goes to Rob Little for his enormous efforts in the field. Rob, I appreciate that snow and ice couldn't keep you away from rotten fish. Thanks to DFO staff Roy
Argue, Doug Turvey and Jason Hwang for support with the channel. Thanks to Jordan
Holmes for assistance and channel-side chats. Thanks to the community of Horsefly for
a wide range of support and interest. Thanks to Rick Holmes and Bill Best of the QRRO
Thanks to George Swerhone and John Lawrence at Environment Canada for taking the
time to teach me confocal techniques. Thanks to Tom Johnston for help with modelling
salmon decay products. I gratefully acknowledge financial assistance from Peace/Williston Aquatic Research Award.
Many thanks to my family and friends who are a continuous source of inspiration. Thanks
to my parents: John and Leslie Albers, my sister: Jane Albers, my brother's family: Jesse,
Heather and Ella Albers and my Nana: Patricia Taylor. Thanks to old buddies: Mike,
Brooke, Leslie, Morgan, Dave, Ben, Quinn, Simon and John. Thanks to everyone at
UNBC. Thanks to COL buddies: Jack, Katrina, Ty, Eric, Becky, Jordie and Lesley. Our
economics discussions provided stimulation and welcome relief from thesis topics.
Finally, I wish to thank Brooke Wilken. Thank you Brooke for agreeing to begin our life
together in Northern BO Thank you for spending long hours listening about sediment
and biofilm. All this is simply an elaborate attempt to impress you.
This entire document was written using freely available open-source software. I gratefully acknowledge the developers and programmers who have freely given their time to
produce such high quality software such as KTpX, Emacs, R and ImageJ.

1

Chapter 1
Salmon and Biofilm: A Literature Review

1.1

Introduction

Anadromous fishes are important nutrient vectors that can substantially alter their natal habitats through yearly nutrient pulses (Naiman et al., 2002) and substantial physical
disturbance (Rex and Petticrew, 2006; Moore, 2006). Each year millions of anadromous
Pacific salmon (Oncorhynchus spp.) return from the ocean to their natal freshwater streams
to spawn and die (Quinn, 2005). Salmon biomass, accumulated primarily at sea, is transported to upstream freshwater spawning grounds (Naiman et al., 2002). This marine
derived nutrient (MDN) pulse is atypical of most rivers. Globally, nutrient and energy
flow in rivers usually proceeds downstream, eventually ending up in the sea (Vannote
et al., 1980). These types of ecosystem linkages are increasingly being seen as crucial
components of a healthy ecosystem (Lamberti et al., 2010). This is particularly true with
a culturally and commercially valuable species such as Pacific salmon.
Pacific salmon migrate against the typical nutrient and energy river gradient, impacting terrestrial and aquatic ecosystems (Quinn, 2005) that are sometimes considerably inland from the marine environment (e.g. Johnston et al., 2004). In the terrestrial landscape, bears and other large mammals will carry salmon carcasses up to 20-m away from
the river bank, fertilizing the surrounding system with unused portions of salmon flesh
(Cederholm et al., 1989) as well as fecal material (Reimchen, 2000). This nutrient input is
a significant driver of primary and secondary productivity in the terrestrial ecosystems
of the Pacific Northwest (Naiman et al., 2002; Drake et al., 2002; Hocking and Reimchen,
2002).

Chapter

1. Salmon and Biofilm: A Literature

Review

2

Similarly, spawning salmon enrich aquatic systems with MDNs via indirect trophic
pathways and direct consumption pathways (Bilby et al. 1996; Wipfli et al. 1998; Figure
1.1). Nowlin et al. (2008) report that nutrient pulses travel more quickly through aquatic
systems than terrestrial systems. This nutrient pulse also tends to be more persistent in
terrestrial systems suggesting that yearly variability in salmon numbers has a large effect
on salmon spawning river systems. Large gaps, however, remain in understanding the
mechanisms by which MDNs enter these aquatic system (Janetski et al., 2009).

Resident Fishes
A

Migration to Ocean'
•
Biomass
Accumulation via
MDN

Invertebrates
A
Increased grazing
and greater
invertebrate density

Return to Natal
Stream
¥
-,
Redd
Construction

Salmon
J
Disturbance *
Regime
1 Post-Spawning :
Die-off

Higher spawner
densities and
increased growth

Biofilms
A

Nutrient uptake and
retention. Increased
primary productivity.

Bioavailable Nutrients

i

Marine Derived Nutrients
Figure 1.1: Trophic linkages in salmon spawning system and the general life cycle of
anadromous salmon. The salmon disturbance regime encompasses both the
redd construction and the post-spawn die-off phases of the salmon life cycle. Energy and nutrient transfers within the river are indicated by solid lines.
Dashed lines refer to the marine component of the salmon life cycle. Modified
from Wipfli et al. 1998.

Depending on the time period and spatial scale examined, salmon can be either a positive material subsidy from carcass nutrient release or a negative process subsidy via nest
(redd) construction (Winemiller et al., 2010). Characterizing the effects and the interaction
of redd construction and subsequent in-situ salmon die-off is a crucial step towards re-

Chapter 1. Salmon and Biofilm: A Literature Review

3

solving the complexities of the salmon ecological picture (Rex and Petticrew, 2008; Moore
and Schindler, 2008). Taken together, these two processes represents a salmon disturbance regime that structures both the biological and physical characteristics of a salmon
bearing stream (Figure 1.1). The simultaneous disturbance and fertilization from salmon
represents a complex and unique ecological process (Moore and Schindler, 2008). The
end result of this salmon disturbance regime on aquatic spawning systems has not been
completely assessed.
Coastal nutrient limited streams receiving inputs of MDNs generally experience a
boost in primary and secondary productivity (Bilby et al., 1996; Quinn, 2005). This trend
also seems to hold true for interior streams although these systems are much less well
studied. Regardless of habitat, the post-spawn die-off leaves many salmon to rot on the
river bed (Quinn, 2005). This rotting salmon flesh can be metabolized into the aquatic system in several ways. Nutrients are either consumed directly by invertebrates (Minakawa
and Gara, 1999) or fish (Wipfli et al., 1998) or are taken up by benthic biofilms (Schindler
etal.,2003).
Regardless of the pathway, this nutrient pulse is usually felt at all trophic levels as
grazers consuming basal components of the food web (biofilms) transfer increases in productivity to higher trophic levels (Bilby et al., 1996). This transfer has a positive effect
on juvenile salmonid growth and survivorship (Wipfli et al., 1999, 2003). This suggests a
positive feedback loop whereby more salmon returning to their natal streams will result
in an positive benefit to the next generation of juvenile salmon populations (Figure 1.1,
Schindler et al. 2003).
In many cases these systems have evolved over thousands of years with an expectation of a reliable MDN pulse. A more complete ecological picture of salmon streams
will aid restoration, management and maintenance of this habitat and the salmon themselves. Accurate prediction of the response of spawning stream ecosystems to the influx
of MDNs will enable a more realistic ecological approach to the management of salmon as

4

Chapter 1. Salmon and Biofilm: A Literature Review

a resource (Pauly et al., 1998). Underlying this need for a better understanding of salmon
habitat is the knowledge that many salmon stocks are under significant threat (Table 1.1).
Diminished salmon runs mean fewer spawning salmon and a consequently smaller nutrient input to the stream. This reduction may have large consequences for conservation
and management goals; fewer nutrients retained within the river systems may inhibit the
ability of the system itself to sustain future salmon populations (Bilby et al., 1996).
Table 1.1: Estimated salmon abundance by region. Values are percentages of the total
number of populations in each category. After Quinn (2005); Scheuerell et al.
(2005).

Region
Southeast Alaska
British Columbia
Washington
Coastal Oregon

1.2

Healthy

In jeopardy

Extinct

Unknown

Total # of Distinct Genetic Pop-'s

10.0
48.3
37.5
32.6

0.1
9.7
22.2
49.7

<0.1
1.3
16.1
6.4

89.9
40.7
24.2
11.3

9228
9038
248
141

Benthic Response

In river ecosystems, the benthic community is the main processor of organic material
(Roman! et al., 2004). Benthic communities can consist of invertebrates, colonial algae,
fishes and biofilm structures. This community forms the basal portions of the river food
chain and is an important component of river productivity (Wipfli et al., 1998). Biofilms,
a mixed autotrophic and heterotrophic community fixed in a polysaccharide material, are
the primary producers in benthic communities (Costerton et al., 1995). Biofilms1 remove
nutrients from a river primarily through uptake and retention (Sabater et al., 2002).
The ecological response of the benthos in a natural system is likely dependent on the
number of salmon spawners, the location of the run, the species examined and the phase
of the salmon run examined. Moore and Schindler (2008) found that at densities of 0.1
1

Biofilms are known by various names within the salmon ecology literature. In the context of this study,
I interpret epilithon and periphyton to equally refer to the material that I call biofilms. Therefore here,
biofilms refer to the biological material growing on gravels in the benthic habitats of salmon spawning
streams.

Chapter 1. Salmon and Biofilm: A Literature Review

5

salmon m~~2, active salmon spawning reduces both biofilm and invertebrate abundance
by 75-85%. This density dependent response has implications for interspecific variation
as coho salmon (O. kisutch) often spawn at lower densities than sockeye (O. nerka) or
chum (O. keta) (Bilby et al., 1996). Later in the spawning cycle, during the die-off period,
there is often an increase in biofilm abundance sometimes equalling or exceeding prespawn levels (cf. Johnston et al., 2004). This response may also be density dependent as
smaller runs equate to small MDN additions to spawning grounds. Nevertheless, given
that most salmon spawning streams are oligotrophic (Naiman et al., 2002) even small
MDN additions may constitute important nutrient sources (Bilby et al., 1996).
Table 1.2 summarizes the major findings of several previous biofilm-salmon studies.
Although there is a wide range in biofilm response to MDNs, there is an equally large
range of parameters measured in an attempt to characterize biofilm activity. Of particular importance to this study are the prior attempts to measure the response of benthic
communities to an influx of MDNs. Biofilm growth is usually measured by chlorophyll a
and ash-free dry mass (AFDM) while MDN uptake is measured by 8nC and <515N isotopic
signatures.

1.2.1

Biofilms

The biofilm growth response to the salmon disturbance regime is spatially and temporally
variable and MDN uptake in particular 513C tends to be variable as well (Claeson et al.,
2006). This variable response can be attributed to hydrodynamic conditions (Stoodley
et al., 1999), light, nutrients (Moore and Schindler, 2008), and redd construction (Mitchell
and Lamberti, 2005; Hassan et al., 2008) as well as inadequate characterization of biofilms
(Sabater et al., 2002). Part of this variable response is a temporal issue and can best be assessed by examining the benthic response during both the active-spawn and post-spawn
period (Janetski et al., 2009). This approach to characterize gross effects allows for assessing stream resistance, the ability of the stream to oppose the effects of redd construction,

Table 1.2: S u m m a r y of major biofilm studies in relation to salmon spawning Unless otherwise indicated responses are either
from a salmon presence/absence study or a before, during and after assessment of relevant variables or both a
Study Characteristics

Biofilm Characteristics

Study

System/Location

Time of Year

Salmon
Salmon
D e n s i t y / m 2 Species

AFDM
(mg/cm2)

Chi
a <515N (%)
(ug/cm2)

d13C

Wipfli e t a l (1998)
Peterson a n d Foote (2000)b
Johnston e t a l (2004)
Mitchell a n d Lamberti (2005)
Moore a n d Schindler (2008)6
T i e g s e t a l (2008)
C a k e t a l (2008)
Wipfli e t a l (1998)
C h a l o n e r e t a l (2002)
Mitchell a n d Lamberti (2005)
C l a e s o n e t a l (2006)
Kline e t a l (1990)
Bilby e t a l (1996)
C h a l o n e r e t a l (2002)
McConnachie a n d Petticrew (2006)
Kline e t a l (1993)
C h a l o n e r e t a l (2007)

River/Alaska
River/Wash
River /Interior BC
River/Alaska
River/ Alaska
River/Alaska
Estuary/Alaska
Mesocosm/Alaska
Mesocosm/Alaska
Mesocosm/Alaska
River/Wash
River/ Alaska
River/Wash
River/Alaska
River/Interior BC
Lake/Alaska
River/Alaska

Aug-Oct
July-Aug
July-Oct
July-Sept
Multi-year
July-Oct
July-Sept
Aug-Oct
Aug-Oct
July-Sept
July-Oct
July-April
Whole Year
Aug-Oct
June-Aug
Multi-year
6 years

80 k g / L
08
0 07
0 54
0 57
3 6-4 7
4
15
08
25
0 36
07
2 1 kg/m2

t

-

NSD
NSD

t
t

-

pink
sockeye
sockeye
mix
sockeye
mix
mix
pink
mix
mix
chinook
sockeye
coho
mix
sockeye
sockeye
pink,
chum

t

NSD
-

t

1

1

t
t
-

NSD
-

T
NSD
-

t

t

"NSD no significant difference, - not measured, t increased after salmon run, 4. decrease after salmon run
^Samples were collected during the active spawn period

NSD
-

t
t

t

NSD
NSD

T
t n o test

t

t

t
t

T
—

tno

—

Chapter 1. Salmon and Biofilm: A Literature Review

7

and stream resilience, the ability of the system to process the post-spawn MDN pulse
(Biggs et al, 1999).
1.2.1.1 Active-Spawn
Recent work in spawning salmon streams has identified the role salmon play in enriching
and disturbing the quality of their natal habitats (Naiman et al., 2002; Hassan et al., 2008;
Rex and Petticrew, 2008). During active-spawning salmon can excavate redds that are 1035 cm deep, considerably disturbing streambed biofilms (Schindler et al., 2003). During
this process clay, silt and sand is resuspended into the water column. The ecological
importance of redd construction is highlighted by reduced fine sediment in the gravelbed
(Hassan et al., 2008), subsequent nutrient delivery to the gravelbed (Rex and Petticrew,
2008), and increased egg survival due to scour resistance (Montgomery et al., 1996).
Many studies have reported significant decreases in benthic biofilm abundance during active spawning (Table 1.2). In one example, Moore and Schindler (2008) found no
benefit of MDN to the stream ecosystem as redd construction reduced biofilm abundance
below critical levels suggesting that the system may be unable to recover after the redd
construction disturbance. This result, however, is likely dependent on context and on
many factors including spawners densities, light availability, hydrologic variables, bed
characteristics, background nutrient levels and the timing and species of the run (Moore
and Schindler, 2008). Table 1.2 highlights both the varied biofilm response to a salmon
run and the fact that the full ecological ramifications of this habitat modification are still
unclear. For example, during redd excavation, biofilms are resuspended into the water
column along with sediment. The fate of this biofilm remains unclear. It could either be
transported downstream to receiving lake systems or retained in the stream. The reduction in biofilm abundance due to redd construction, however, may reduce the ability of
the system to process the subsequent nutrient pulse from the post-spawn dead salmon
(Roman! et al., 2004).

Chapter 1. Salmon and Biofilm: A Literature Review
1.2.1.2

8

Post-Spawn

Material from the post-spawn carcass decay is the most substantial nutrient subsidy to
salmon bearing streams particularly in oligotrophic streams (Naiman et al., 2002). Many
studies have reported increases in biofilm density after redd construction sometimes
equalling or exceeding pre-spawn densities (Table 1.2). For example, after the activespawn decrease in biofilm abundance, Moore and Schindler (2008) reported a post-spawn
increase in biofilm abundance that exceeded pre-disturbance levels but only in some
years. The post-spawn MDN pulse may have a longer-lasting effect than active-spawn
disturbance as the in-stream carcass decay can last well after all fish have died (See Johnston et al., 2004). This post-spawn increase, however, is not always found suggesting that
MDNs are not always incorporated into the ecosystem via biofilms.

1.2.1.3 MDN contribution
MDNs have a higher isotopic signature (e.g. £15N and (513C ) that allows for accurate tracing of nutrient pathways (Bilby et al., 1996). Because salmon gain most of their biomass
at sea, they tend have marine carbon and nitrogen isotopic ratios. Convergence to this
marine isotopic signature by freshwater organisms implies utilization of MDNs (Kline
et al., 1993). In particular 515N is a good indicator of marine nutrient origin (Kline et al.,
1990). For example, Bilby et al. (1996) found that all trophic levels in streams were enriched in both (515N and <513C after a salmon run. However, many geographic, analytical,
spatial, and temporal issues still plague the interpretation of MDN uptake and have not
been resolved (Staal et al., 2007). Hence, other studies on the effect of MDNs have yielded
varying results (Table 1.2).
The overall response of biofilms to MDNs has not been conclusively shown to be unidirectionally positive. Claeson et al. (2006) found no increase in 513C values from sampled biofilms after salmon die-off over a period of eight weeks. In contrast, <515N levels
were significantly higher over the same period. Chaloner et al. (2002) found that biofilms

Chapter 1. Salmon and Biofilm: A Literature Review

9

grown in a natural salmon-bearing creek exhibited lower MDN incorporation of carbon
than nitrogen. In another part of the same study, biofilms in a mesocosm channel exhibited lower <515N values than SUC than a non-salmon carcass control (Chaloner et al.,
2002).
Potentially, increased algal growth rates due to MDNs, a factor unaccounted for in
that study, may have been altered via reduced 5UC discrimination during photosynthesis (Chaloner et al., 2002). Using only isotopes and biomass estimates to characterize
biofilm function is a useful but limited approach to identify MDN uptake as the algalbacterial and carbon-nitrogen ratio are also critical measures of biofilm activity (Roman!
and Sabater, 2000). However, to the detriment of a clearer ecological picture, many studies have ignored the more general biofilm literature.
1.2.1.4 Biofilm Composition
Sabater et al. (2002) state that the function of biofilms is intimately tied both to its components and structure. Biofilm composition and structure are in turn determined by a combination of physical, chemical and biological factors. Biologically, biofilms are comprised
of autotrophic and heterotrophic organisms. The ratio of these two biotic components
determines the ability of a biofilm to process organic matter. Predominantly autotrophic
biofilms tend to absorb fewer nutrients because of high internal nutrient cycling (Roman!
et al., 2004). Conversely, more heterotrophic biofilms tend to absorb more nutrients due to
diminished nutrient release by algae available for internal cycling (Roman! et al., 2004).
This biofilm behaviour has important consequences for MDN uptake within a salmon
spawning river. Most studies characterize some combination of these factors but rarely is
biological composition taken into account. Often, this is a practical consideration as the
analytical and microscopic tools needed to properly characterize biofilms are unavailable
to ecologists either due to a lack of specific knowledge or field sampling constraints.
Yoder et al. (2006), one of only a few salmon-biofilm studies that include a measure

Chapter 1. Salmon and Biofilm: A Literature Review

10

of biofilm composition, suggested a nutrient incorporation model whereby heterotrophic
bacteria are the initial streambed colonizers; autotrophic algae lag in response until they
are able to establish populations over microbial mats. This suggests a differential response
of autotrophs to heterotrophs to a salmon nutrient pulse and may explain variable patterns of (513C and <515N uptake by biofilms (Table 1.2). As stated above, to characterize this
process, the full salmon disturbance regime needs to be sampled. It can be expected that
an increase in algae abundance, as seen by Yoder et al. (2006), will result in a second bacterial population spike as heterotrophic bacteria use nutritive products exuded by algae
(Roman! and Sabater, 2000). This suggests a composition-based model to explain how
salmon nutrients are processed within a stream.

1.2.2

Particle Aggregation

The flocculation feedback loop proposed by Rex (2009) elucidated that the temporal overlap between redd construction with increased suspended sediment levels and salmon
die-off with increased salmon decay products in the water column provided ideal conditions for flocculation. The formation of floes within salmon-bearing streams due to
salmon redd construction is well documented (Petticrew and Arocena, 2003; Petticrew
and Rex, 2006). Rex and Petticrew (2008) demonstrated the trapping and deposition into
the gravelbed of suspended microbial MDN floes. These suspended microbial floes or
aggregates consist of organic and inorganic components and can be considered as suspended biofilms (Droppo, 2001). Floes differ from their constituent particle in size, shape
and most importantly in a MDN delivery context, settling velocity (Droppo et al., 1997).
Observed flocculated particles laden with MDNs may deliver nutrients over short riverine distances by increased floe settling rates (Figure 1.2; Rex and Petticrew 2008).
The structure of floes may interact with the structure of biofilms providing a potential
mechanism for biofilm MDN mineralization. External structures called extracellular polymeric substances (EPS) on both biofilms and floe provide a chemical 'stickiness' that may

Chapter

1. Salmon and Biofilm: A Literature

Review

11

Figure 1.2: Diagram of a salmon spawning stream illustrating the relationship between
redd construction, biofilm resuspension and flocculation (Rex and Petticrew,
2008)
aid in particle trapping (Sutherland, 2001). EPS is a comprehensive term for the organic
matrix that house biofilms and floes composed of polysaccharides, proteins and nucleic
acids (Wingender et al., 1999). The interactions between these two EPS sources may play
an important role in particle trapping and subsequent M D N absorption (Droppo, 2001)
particularly in the context of increased biofilm growth (Battin et al., 2003b) after salmon
die-off (Johnston et al., 2004).

1.2.3

Invertebrates

The effect of salmon spawning and die-off on benthic invertebrate communities is an area
of active research and ongoing debate as to whether salmon positively or negatively impact these communities. Some studies have reported reduced invertebrate densities and
altered community structure in streams with actively spawning salmon (Minakawa and
Gara, 2003). Peterson and Foote (2000) found that redd construction caused a significant
increase in invertebrate drift in the water column and reduced invertebrate densities. Ad-

Chapter 1. Salmon and Biofilm: A Literature Review

12

ditionally, Moore and Schindler (2008) found a significant decrease in invertebrate densities above a threshold salmon density of 0.1 salmon per m 2 .
In contrast, salmon carcasses experimentally added to an artificial flume increased
invertebrate density by 8 to 25 times (Wipfli et al., 1998). Over a 50-m reach Claeson
et al. (2006) measured an increase in invertebrate density due to MDNs. Both Claeson
et al. (2006) and Bilby et al. (1996) reported <515N and 513C enrichment of invertebrates
indicating uptake of MDNs in secondary production in response to salmon die-off.
Moore et al. (2004) found that different taxa of invertebrates responded differently to
MDNs. The ability of invertebrates to transfer energy and nutrients gained from biofilms
to higher trophic levels is a crucial link in understanding the effect of MDNs on salmon
productivity. Studies that have found an increase in invertebrate densities tend to explicitly account for spatial variability within streams, especially downstream linkages of
invertebrates to an MDN pulse. Although the exact nutrient pathways have not been determined, the increase in productivity seen in biofilm does seem to transfer upwards to
resident fishes (Wipfli et al, 2003).

1.2.4

Resident & Juvenile Anadromous Fishes

In contrast to the disparate response of invertebrates, resident fishes seem to exhibit a
clear trend of increased growth in the presence of MDNs. This response manifests itself
in several ways. Resident and juvenile anadromous fishes become enriched with d15N
and <513C (Wipfli et al, 2003) indicating incorporation of MDNs (Bilby et al., 1996; Claeson
et al., 2006). Moreover, increased growth rates of cutthroat trout (O. clarki), Dolly Varden
(Salvelinus malma) and anadromous salmon (Oncorhynchus spp.) suggest a trophic transfer
of energy and nutrients from decaying salmon to subsequent fish populations (Wipfli
et al., 2003). Finally, salmon flesh and eggs provide a rich food source for resident fishes
(Schindler et al, 2003).

Chapter 1. Salmon and Biofilm: A Literature Review

1.2.5

13

Water Chemistry-

Table 1.3 summarizes the response of common aquatic nutrient measures to salmon spawning. Consistently, NH^ and soluble reactive phosphorus (SRP) levels are significantly elevated in the active-spawn period due to waste products (Mclntyre et al., 2008) and during
the post-spawn period in-situ carcass decay (Hood et al., 2007). The consistent pattern of
NH^ and SRP levels may be due to their prominence in the chemical makeup of fish excretion (Moore and Schindler, 2008) and decay (Cak et al., 2008) products. Modelling of
nutrient dynamics, however, by Johnston et al. (2004) suggests that the decay products
are primarily responsible for this increase in N H | and SRP. Given the ephemeral nature
of NH^" as a nitrogen form, large fluctuations in NH^ levels would indicate that fast microbial metabolic pathways are the main processors of organic material (Johnston et al.,
2004). The increase in flocculated material in the water column reported by Rex and Petticrew (2008) and the concomitant increase in suspended bacterial matter (in the form of
floes) (Droppo, 2001) suggests one potential mechanism for this processing. Most studies,
however, have only been reproduced in coastal areas which contrast greatly with interior
systems in terms of nutrient availability and demand.
Additionally, Table 1.3 highlights the range of measures used to characterize the water
column during salmon spawning. Often, NO^T and dissolved organic carbon (DOC) are
observed to increase. These NO^ and DOC fluxes, however, are often not correlated to
any point of the salmon spawning cycle (Cak et al., 2008).

Table 1.3: Summary of water chemistry responses to salmon spawning NHJ~ and SRP generally increase m response to salmon
run NO^~ values generally increase in a salmon s p a w n m g stream but the increase is not correlated with the arrival
of salmon All nutrient measures are presented as comparison of concentrations (e g mg/1) Note that most studies
on biofilm and salmon have been conducted m coastal areas rather than interior ecosystem b
Study

Type of Control

Study Location

Bilby etal (1996)
Chaloneretal (2004)
Johnston etal (2004)
Claeson etal (2006)
Mitchell and Lamberti (2005)
Hood etal (2007)
Cak etal (2008)
Tiegsetal (2008)
Tiegsetal (2009)
Chaloneretal (2007)
Wipfli etal (2010)

Experimental
Spatial
None
Spatial
Spatial
Spatial
Spatial
None
Cage Exclusion
Spatial
Experimental

Wash
Alaska
Interior BC
Wash
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska
Alaska

b

NH+

NO^

DOC

SRP

t

t
t

nsd
-

nsd
nsd
T(nc)
T(nc)
T(nc)
nsd
-

f4 x
f40 x
nsd
t4 x
tlO x
t6x

t

nsd
nsd
t(nc)
nsd
t(nc)
T(nc)
T(nc)
nsd

T40 x

t
t

T10 x
tl00 x
t40 x
T

t

TN

nc

NO^

TDN

TSP

KJN

t

nsd
used for TN

nsd

nsd

t
t
t
T

nsd, no significant difference, -, not measured, j , increased in the spawning stream, 4-, decrease in the salmon spawnmg stream, nc, not correlated
with the salmon run

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

15

Chapter 2
Sediment and Biofilm Interactions in a Salmon Spawning Stream:
Aspects of Marine-Derived Nutrient Infiltration and Trapping

2.1

Introduction

Biofilms remain an understudied component of the nutrient cycling pathways in rivers
(Battin, 2000). Similarly, interior salmon spawning streams remain understudied habitats
compared to coastal systems (See Tables 1.2 & 1.3). Interior systems differ considerably
from coastal systems. The longer distances from the ocean (Quinn, 2005) to the interior
habitats diminish the marine connectivity as the downstream nutrient sinks are often rearing lakes rather than the ocean. This connectivity between rearing habitat and spawning
grounds in interior systems has considerable implications for the spatial component of
nutrient retention. The life history of interior salmon stocks reflects this different habitat
as juvenile fish often spend a year in the rearing lake prior to proceeding out to sea.
Considerable information has been gained on salmon spawning ecology (Schindler
et al., 2003; Janetski et al., 2009) from using both artificial stream-based studies (Claeson
et al., 2006; Rex and Petticrew, 2008) as well as field observations (Moore and Schindler,
2008). The Horsefly River spawning channel (HFC) represents a unique environment
that spans the manipulability of an artificial stream with the realism of a natural habitat.
For example, hydrodynamic conditions can be kept constant while salmon activity still
closely mimics that of a typical stream. This type of ecological realism has been previously
highlighted as vitally important for making consistent observations (Janetski et al., 2009).
The objective of this study is to examine the significance of biofilm uptake and retention of MDNs as related to salmon spawning and die-off in an interior British Columbia

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

16

river. The interaction between sediment and biofilms is examined in the context of a
salmon spawning event (i.e. the salmon disturbance regime). This research explored two
broad themes as they relate to salmon spawning and biofilm ecology.

2.1.1 Theme One
How does biofilm biomass change in response to the salmon disturbance regime? Changes
in biofilm biomass and isotopic signature have been observed to reflect nutrient uptake
by biofilms in other freshwater systems (Bilby et al., 1996; Sabater et al., 2002; Yoder et al.,
2006). What effect does redd construction have on biofilm abundance? Does a reduction
of biofilm during the active-spawn period inhibit growth of biofilms during the postspawn phase? Is there a spatial component to this growth? Do downstream biofilms
become enriched with MDNs from upstream salmon carcass decay? Chlorophyll a and
AFDM will be used as indicators of change in biofilm biomass and isotopic signatures
will be used to infer MDN addition.

2.1.2

Theme Two

What effect does salmon redd construction have on biofilm and sediment resuspension
and subsequent nutrient delivery to downstream streambeds? Salmon resuspend biofilm
and sediment when cleaning gravels during redd construction (Figure 1.2). Resuspension
of these materials into the water column via redd construction has an unknown impact on
the overall nutrient retention capacity of streambeds both spatially and temporally. The
fate of these reworked biofilms may represent an important in-stream nutrient vector, if
they are deposited within the channel. The fate of resuspended sediment may be a key
indicator of mechanisms transferring MDNs to the streambed (Rex and Petticrew, 2008).

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

2.2

17

Methods

2.2.1 Study Site
The Horsefly watershed (52° 19'N/121° l'W ) is located within the Cariboo region of
British Columbia. The Horsefly River is the largest tributary of Quesnel Lake. Historically, the Horsefly River supports a large sockeye escapement although recent years have
seen steep declines. The Horsefly river escapement is often more than the combined total
of all other Quesnel River tributaries (Lawrence, 2004).
2.2.1.1 Horsefly Spawning Channel
The Horsefly River spawning channel (HFC) is an artificial salmon stock enhancement
stream constructed by the Department of Fisheries and Oceans (DFO) (Figure 2.1). The
channel is 1600-m in length and approximately 10-m in width with a slope of 1%. Water
flow into the channel is controlled by a large siphon supplying water from a settling pond
which is directly connected to the Horsefly River. The mouth of the channel flows into the
Horsefly River (Figure 2.1). Sockeye salmon (Oncorhynchus nerka) enter the HFC via the
Horsefly River. Upstream access to the Horsefly River for the salmon is restricted within
the HFC by a permanent gate at the upper portion of the channel. Confined salmon then
spawn inside the channel.
In summer and fall of 2009, in addition to regular DFO enhancement activities, a portion of the HFC was converted into an experimental reach (Figure 2.1). Sampling began
on August 28 th , 2009 and lasted until October 26"\ 2009. Prior to the start of the sampling
period (August 26*'l/ 2009), the channel bed was cleaned of sediment and biofilms using a
rake with 30-cm teeth mounted on a bulldozer. The rake resuspended material from the
streambed, which was then flushed out of the channel to a downstream settling pond via
artificially-generated high discharge.

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

18

Horsefly

Horsefly
River

Figure 2.1: Location of the Horsefly River spawning channel (HFC) and position of the
experimental reach.
2.2.1.2

Measurement of Site Characteristics

Discharge in the HFC was monitored through a combination of staff gauge readings and
a pressure transducer (Unidata 8007 WPD) and applied to a calibrated rating curve. The
rating curve was estimated by measuring flow velocity at 0.6 of the depth at 1-m intervals
across the channel and at a range of representative stage heights (n=4). Stage height w a s
an excellent predictor of discharge (r 2 =0.997). Precipitation was measured using three
anchored buckets located in three open fields near the study site. Precipitation was calculated by averaging the volume of water in each bucket and normalizing it per unit area.
Water temperature was continuously recorded in each section using calibrated Tidbit Temp Loggers (Onset Corp.). Tidbits were calibrated by placing them in a bucket of
water at a known temperature and establishing a correction term for each Tidbit based
on the difference between the measured and known temperatures. Conductivity and p H
were recorded using a Combo Multiparameter Meter with two-point calibration (Geo Scientific, Ltd.). Average grain size was determined by measuring the three major axes of

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

19

gravels (n=100) randomly selected from each section. Other resident fishes observed in
the channel during the experimental period were a small number of Rainbow trout (O.
mykiss), Kokanee (O. nerka) and Chinook Salmon (O. tshawytscha). Light levels were similar across the entire experimental reach of the HFC. The west side of the channel was
devoid of tree cover. The east side of the channel had a 3-m strip of deciduous trees that
provided uniform shade over the length of the experimental reach.

2.2.2

Study Design

2.2.2.1 Spatial and Temporal Controls
The study site was located in the upper portion of the spawning channel (Figure 2.1).
A 60-m section of the channel was divided into three sections using steel fences which
limited salmon entry into a particular section. Channel sections will be henceforth referred to as 'Upstream' for the spatial control, 'Middle' for the middle section with the
most active salmon spawning area and 'Downstream' as the downstream deposition area
(Figure 2.2). A 1-m buffer around each fence was excluded from sampling and used for
movement through the channel.
Sampling was divided into three spawning periods. All samples collected prior to
salmon arrival were defined as 'Pre-Spawn'. Samples collected during the most active
spawning period were termed 'Active-Spawn' and the die-off period was termed 'PostSpawn'. Collectively these temporal divisions of the salmon spawning cycle are referred
to as 'Period'.

2.2.3

Salmon

Salmon densities were enumerated visually and with a digital camera. The salmon were
visually counted by two individuals. In instances where the counts differed greatly (>10
salmon), the salmon were recounted until a similar count was reached. Where live salmon

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

20

Direction of Water Flow
^

20 m

i

1
I
1
1
D o w n s t r e a m Reach

'
l
i
i

i
20 m
—JU.
hi
i
i
Middle Reach
I Upstream Reach
i
i
i
20 m

m |

<U 1

,°> 1
c I
o'

LL
t=|
O

>.

u'

l/l 1

31

4i
UJ

Figure 2.2: Division of the HFC into experimental sections Downstream spawning
salmon not part of the experiment were excluded from the experimental section at the lower portion of the downstream reach by an additional steel fence.
densities were too active to be counted visually (September Ylth-2Sth), a digital photograph was taken of the reach and salmon were counted at a later date. Freshly dead
salmon1 flesh (n=4) was sampled and analyzed for d15N , d13C and %C and %N.
A range of salmon numbers were loaded into each section. The downstream and
middle sections were used as areas to assess the effect of active salmon spawning and
die-off. The upstream section was intended to remain free of fish as the spatial control on
salmon redd construction and die-off. A small number of salmon, however, escaped into
the upstream section diminishing the spatial control. Live salmon were removed from
the upstream section when possible, minimizing spawning activity. Dead salmon were
also immediately taken out of the upstream section removing any potential die-off effects
from the upstream control.
Salmon were sampled within an lh of dying In each case, the fish was observed to cease swimming
and sampled once it had completely stopped movmg

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

2.2.4

Biofilms

2.2.4.1

Biofilm Collection

21

Streambed biofilms were collected in each section from randomly sampled gravels during the pre-spawn, active-spawn and post-spawn periods. On every sample date, five
rocks were randomly collected at each of three streambed depths from the three section
sections. Five samples have been previously identified as an appropriate sample size to
characterize the spatial variability of biofilm within a reach of similar size (see Chaloner
et al., 2004). Measurements taken from these five rocks were averaged to generate a mean
data point for each sampling day. The three sampling depths were streambed surface,
5-cm below the streambed (^ d50) and 10-cm below the streambed (« 2xd50). Because of
a small number of escapees into the upstream section, rocks were collected in this section
where there had clearly been no redd building activity. Rock surface area (SA) was determined using the method of Graham et al. (1988). A second surface biofilm sample was
collected on each sampling date in triplicate from each section for stable isotope analysis.

2.2.4.2

Biofilm Characterization

Immediately after collection, gravels were scrapped with a toothbrush and rinsed with
distilled water to remove all biofilm and inorganic sediment. The resultant slurry was
filtered onto a pre-ashed, pre-weighed GFF, protected from light and frozen at -20°C until further analysis (Mitchell and Lamberti, 2005). Chlorophyll a was extracted from the
slurry in 25 ml of 90% buffered acetone for 24h at 4°C and the extract was centrifuged at
3100 rpm for 20 min. Extracted chlorophyll a was analyzed spectrophotometrically correcting for phaeophytins by acidification with HC1 (APHA, 1995; Steinman and Lamberti,
1996; Mitchell and Lamberti, 2005). Any material left on the GFF after extraction and material centrifuged into a plug in the above centrifugation step was combined, dried at
60°C for 12h, weighed, ashed at 550°C for 2h and weighed again. The mass lost during

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

22

the ashing step was defined as ash-free dry mass (AFDM) and the material left on the GFF
was the amount of inorganic sediment trapped by the biofilm.
Biofilms for stable isotopes analysis were scraped from each gravel in the same manner as described above except the biofilm slurry was frozen in a microcentrifuge tube.
Upon returning to the lab, samples were freeze-dried and analyzed for 8l3C , S15N and
total C and N (Pacific Centre for Isotopic and Geochemical Research, University of British
Columbia). Isotope enrichment was determined as follows (Kline et al., 1993):
%o S13C or 515N = Rsampl* ~ Rstandard

x

1000

(2.1)

^standard

where R is the ratio of the heavy isotope to the light isotope. The standard for C is Peedee
Belemnite and for N is air (Bilby et al., 1996).

2.2.5

Infiltration Bags

Sediment and nutrient infiltration in the streambed was assessed using modified infiltration bags which allow for vertical and horizontal sediment delivery to a sample column
of gravel (Rex and Petticrew, 2006). Prior to sampling, three 0.35-m holes were dug in
each section. Plastic bucket frames covered with galvanized steel mesh (aperture 0.025m) were placed in each hole. The plastic frames prevented outside gravels from filling
the hole, while the steel mesh allowed for normal water flow through the gravels. Additionally, the plastic frames allowed for periodic removal of the sample column and replacement with clean gravel in the same unaltered position within the streambed. In each
experimental section, infiltration bags were placed at the base of three buckets and covered with gravel cleaned of sediment <2-mm. Bags were removed weekly, and replaced
with new bags and cleaned gravel. For each weekly sampling date, gravels were rinsed
through a 2-mm sieve to remove all <2-mm sediment into a volumetrically calibrated
bucket.

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream
2.2.5.1

23

Sediment Characterization

Fine sediment was sampled by resuspending all the material collected from infiltration
bags in a sample bucket, waiting 10-s and sub-sampling the top portion of water. This
method allows for larger particles to settle out and ensures that only fine sediment (<70pm) is sampled (Rex and Petticrew, 2006; Petticrew and Albers, 2010). Samples were
taken for particle size analysis and for AFDM and inorganic sediment characterization.
Fine sediment from the infiltration bags was filtered onto GFF, dried at 60°C for 12h,
weighed, ashed at 550°C for 2h and weighed again. Response variables derived from this
process were AFDM and inorganic sediment. Sediment <2-mm subsamples were also
analyzed for <515N , 513C and total N and C in the same manner as biofilms as described
above (Section 2.2.4.2).

2.2.5.2

Particle Size

Particle size distributions were determined using laser in-situ scattering and transmissometry
(LISST) (Sequoia Scientific, Inc.). A LISST probe measures the degree of diffraction when
a laser is passed through a 60-ml sample to determine a distribution of 32 size classes
of particles ranging from 2pm to 460pm (Agrawal and Pottsmith, 2000). The LISST protocol was adapted from Rex (2009, Appendix 1). Samples (60-ml) were analyzed with
the LISST within 5 days of sampling. Samples were gently mixed and poured into the
LISST sample chamber to prevent the formation of air bubbles. Rex (2009) identified that
bulk samples of sediments are more easily collected than water column aggregates and
can reflect particle sizes distribution shifts in the water column. Three subsamples from
each infiltration bag were collected and averaged to account for any potential variability
associated with LISST sampling (Williams et al., 2007). LISST data were processed using
a semi-automated macro with MS Excel (Microsoft, Inc.) that calculated cumulative distributions, measures of central tendency as well as diagnostic parameters. Standard error
bar plots of die, d50 and d84 values as per Kondolf (2000) and cumulative distribution plots

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

24

were used to qualitatively compare changing particle size during the salmon spawning
cycle.

2.2.6

Suspended Sediment

ISCO automatic water samplers (Teledyne ISCO, Inc.) were placed streamside with the
sampling tube located in the thalweg of the HFC near the rear of each section to sample
for suspended sediment. Water was sampled every 3 hours 8 times per day to form one
daily composite sample. ISCO water samples were also filtered with GFF, dried at 60°C
for 12h, weighed, ashed at 550°C for 2h and weighed again. The ratio of the material
lost during the ashing step to the material that remained was used as a response variable
for the ISCO water samples. This variable was defined as the organic matter ratio (OMR)
and used to assess the nutrient quality of suspended sediment in the HFC as it reflects the
organic and inorganic components of the sediment. The total mass of sediment (inorganic
or organic fractions) in the water column was determined by multiplying the sediment
concentration by the discharge to estimate the total daily suspended sediment load.
Particle size of suspended sediment was examined using the LISST on September 24"1,
2009 and July 1414, 2010. A particle size measurement with the LISST was taken on July
14th, 2010 in the HFC to determine background or pre-spawn particle size distribution.
Measurements taken on July 1414, 2010 were conducted under similar channel conditions
and it is assumed that measurements taken on this day are representative of pre-spawn
conditions in 2009. On both sample dates, the LISST was placed in the rear portion of the
middle section to measure the size of suspended particles passing through the instrument
aperture. The LISST was programmed to sample particles every 15-s. The LISST was
left in the channel to sample for approximately 15-min. Data was processed in a similar
manner to section 2.2.5.2. Post-processing included averaging every three observations
to calculate a single point. A Kolmorov-Smirnov test (Siegel, 1957) was used to compare
mean background and active-spawn distributions and cumulative distribution plots were

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

25

used to examining grain coarsening patterns.

2.2.7

Piezometers

To monitor intergravel DO, three piezometers constructed of plastic tubing were buried in
each section at a depth of 28-cm in the bed prior to beginning the experiment. Piezometers
were sampled daily, and analyzed for DO following an evacuation of the tubing to ensure
residual water was not sampled. The oxygen meter was calibrated for each sampling date
by determining the saturation point of oxygen corrected for temperature and elevation.

2.2.8

Statistical Analysis

All biofilm, piezometer and infiltration bag response variables were analyzed with a
two-way ANOVA using period and section as fixed effects. Reaches were considered
to be adequately independent to merit an ANOVA approach although I acknowledge
that some temporal dependence may exist (Cak et al., 2008). The minimum adequate
model (MAM) for each parameter was determined by comparing the F-ratio of a full and
reduced ANOVA model (Whittingham et al, 2006; Crawley, 2007). Replicate biofilm,
infiltration bag and piezometers measurements for each sampling day were averaged.
These single data points from each sampling day served as replicates for the period of
salmon activity factor. A significant period x section interaction indicated that a particular reach demonstrated a different temporal trend as the salmon run progressed. Linear
contrasts of means were used to test specific hypotheses if a significant interaction was
determined (sensu Mills and Bever, 1998). Each contrast was compared both to its temporal and spatial controls. Contrasts were chosen to determine the effect of the salmon
disturbance regime on biofilm abundance. To confirm similar starting conditions in each
section, the contrasts tested for differences in the means at the outset of the experiment.
Contrasts were coded according to the conditions for linear contrasts set out by Fox (1997).

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

26

Models without a significant interaction term had to be interpreted solely on their
main effects. In this case an effect of salmon was still inferred. However, the causality of
this effect is less clear although significant trends may still be assessed. In the absence of
a significant interaction term, pairwise multiple t-test's with Holm's p-value correction
was used to compare mean differences for the main effects (Fox, 1997). Other mean comparisons test all possible combination of factors leading to overly conservative estimates
(Quinn and Keough, 2002). Response variables were log and square-root transformed as
necessary to meet the assumptions of parametric tests. Null hypotheses were rejected at
an a level of 0.05. All statistical analysis were conducted using R 2.11.1 (2010). All graphics were created using R 2.11.1 (2010) with the memisc (Elff, 2010), ggplot2, (Wickham,
2009) and lattice (Sarkar, 2008) packages.

2.2.8.1

Correlations

All correlations were analyzed using Pearson's product moment correlation (Quinn and
Keough, 2002). The relationship between AFDM and chlorophyll a of biofilms was analyzed using individual surface rock values. The relationship between inorganic sediment
and chlorophyll a was limited to post-spawn downstream surface biofilms to examine
deposition characteristics. The correlation between infiltration bag sediment and surface
chlorophyll a from biofilms was a comparison between the weekly values of both parameters from all sections.

2.3

Results

Any study that attempts to demonstrate the net impact of salmon on their natal stream
should ideally sample to characterize all stages of salmon activity, while maintaining concurrent spatial controls to assess how a system responds in the absence of salmon (Janetski
et al., 2009). This type of study design would provide a means to test the interaction be-

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

27

tween spatial and temporal controls and assess the resistance and resiliency of the stream
ecosystem in its totality (Biggs et al., 1999). In this way the entire salmon disturbance
regime can be assessed and the stability of the system can be accurately measured (Biggs
et al, 1999).
In this study, many of the parameters measured did not yield a significant interaction
term but rather significant main effects. These models are no less significant or important
than those with a significant interaction term. However, interpreting these main effects
is more difficult as the hypothesis tested only allows for the comparisons of the marginal
means 2 . Therefore, statistical evidence of differences in cell means is only present in cases
where there is significant interaction term. This statistical framework provides the context
in which to view the patterns seen in the response variables.
Table 2.1: HFC Site Characteristics. Values represent mean values over the sampling period Unless otherwise indicated, values in parentheses are the standard deviation of that parameter. Grain size measurements are from the B axis of gravels
collected in each experimental section
Section
Upstream
Middle
Downstream

Water Temperature (°C, max/mm)

Conductivity (uS)

pH

Gram Size (cm)

12 33 (-0 04, 21 67)
12 41 (0 05, 21 86)
Equipment Failure

126 (44)
133 (45)
120 (39)

7 56 (0 43)
7 51(0 46)
7 77 (0 54)

4 11(119)
4 22(1 11)
4 24 (117)

2.3.1 HFC Characteristics
Site characteristics of the HFC remained relatively constant across stream sections and
varied in a similar manner over the course of the study (Table 2.1). Sections experienced
similar maxima and minima in the physical parameters tested. These differences across
sections were relatively minor compared to differences in biofilm, sediment and intergravel DO.
2

Margmal means are defined as the mean of one factor averaged over all the levels of the other factor
(Quinn and Keough, 2002) Thus, in the context of this study, an example of a marginal mean is the value
of a parameter in the post-spawn period averaged over all sections

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

Date

Figure 2.3: Characterization of stream conditions at the HFC. (a) Precipitation; (b) Discharge measurements. Vertical solid lines divide sampling periods defined by
salmon activity.

28

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

29

In the summer and fall of 2009, the Horsefly region experienced relatively little precipitation (Figure 2.3a). Excluding one observation (October 7th; snow), all precipitation fell
as rain. The frequency of rain events increased during the active-spawn and post-spawn
periods although these were still relatively small storms that were experienced equally
by all three sections. Aside from one period, discharge remained stable over the course
of the sampling period (Figure 2.3b). For a brief period, discharge increased by ~50% («
September 19t/l-28t/l). The high discharge reading was likely caused by the daily build-up
of salmon carcasses at a separation fence downstream of the study site rather than an a
50% increase in discharge. The build-up caused an increase in water depth at the staff
gauge as water backed-up which may have resulted in overestimated discharge values.

2.3.2

Salmon Numbers

Discharge, the number of salmon loaded into the channel, and fish densities reflected
natural spawning conditions and historical usage of the HFC. Salmon arrived in small
numbers inside the experimental area on September 2nd (Figure 2.4). However these fish
were not actively spawning and present in such small numbers that I still considered this
period as pre-spawn. Salmon numbers peaked on September 12 ih in both the middle and
downstream sections. Peak die-off in the middle section occurred on October 7th and this
date was defined as the beginning of the post-spawn period.
On September 25th dead salmon were removed from the middle and downstream
sections. Removal of dead fish from the HFC by black bears (Ursus americanus) occurred
on several occasions and is accounted for by daily counts of dead salmon numbers. There
was no indication that black bears were removing any live fish from the channel. As
a result of escaped salmon from the enhancement section into the experimental section
of the HFC, salmon numbers exceeded target densities although they were still within
a natural range (See Table 1.2 & 2.2). Sockeye also typically spawn at extremely high
densities for salmon (Bilby et al., 1996).

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

Live salmon
Pre-Spawn

Active-Spawn

Post-Spawn

Dead salmon
Downstream
Middle
Upstream

Date

Figure 2.4: Live and dead salmon counts by section. Vertical solid lines indicate divisions
of the salmon period.

30

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

31

Table 2.2: Salmon densities for historical usage of the HFC and the densities loaded into
the experimental sections for this study. Densities used for the HFC study were
slightly higher than historical usage but there are examples of studies using
similar densities (Table 1.2; Chaloner et al. (2002))

2.3.3

Biofilms

2.3.3.1

Surface Biofilms

Year/Section

Density ( F i s h / m 2 )

Middle
Downstream
1989
1990
1991
1992
1993
1994
1995
1996
1998
1999
2000
2003
2006
2007
2009

2.5
1.5
1.7
2.1
1.3
0.3
1.3
1.4
1.2
0.8
1.8
0.9
0.2
1.6
1.4
0.5
0.6

The section x period interaction was significant for surface chlorophyll a, AFDM and
inorganic sediment (Table 2.3). Pre-spawn conditions in each section were not different
enough to reject the null hypothesis that there was no difference in all biofilm parameters
at the outset of the experiment (Figure 2.5, 2.6 & 2.7). Mean chlorophyll a values during
the active-spawn period were significantly different from spatial and temporal controls.
During the active spawning period mean chlorophyll a values in the middle section were
significantly reduced by 4.2 x from the upstream control over the same time period and by
1.5 x from the pre-spawn values from the same section (Figure 2.5-Surface). Downstream
surface biofilms were 2.7x higher in chlorophyll a than the active-spawn period in the
same reach and 1.4x than the upstream control during the same period.

Post-spawn

chlorophyll a in the middle section was not significantly different from the active-spawn

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

32

period in the same reach and the upstream section for the same period.
Table 2.3: Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on surface chlorophyll a, AFDM or inorganic sediment. Interaction contrasts are separated by a |. Contrasts are labelled by the first letter
of the corresponding section and the spawning period (U:Upstream; M:Middle;
D:Downstream).
Chlorophyll a

AFDM

Inorganic Sediment

Source of Variation

Df

Sum Sq

Pr(>F)

SumSq

Pr(>F)

Sum Sq

Pr(>F)

Section
Period
Section x Period
M:Active|U:Active & M:Pre
D:Post|D:Active & U:Post
M:Post|M:Active & U:Post
Starting Conditions
Residuals

2
2
4
1
1
1
1
18

2.288
3.309
3.297
2.424
0.868
0.000
0.005
3.342

0.009
0.002
0.011
0.002
0.044
0.965
0.878

1.779
4.060
2.009
1.337
0.663
0.008
0.002
2.483

0.008
0.000
0.024
0.006
0.042
0.812
0.916

0.338
0.866
0.281
0.212
0.001
0.057
0.011
0.181

<0.000
<0.000
0.001
<0.000
0.819
0.029
0.300

Surface mean ash-free dry mass (AFDM) values demonstrated a similar pattern as the
chlorophyll a values described above. AFDM was significantly reduced during the active
spawning period by 3.2 x compared to the upstream section during the same period and
1.4 x times compared to pre-spawn values in the same reach. Post-spawn values in the
downstream reach were significantly higher (3.0 x) than the active-spawn period in the
same reach. Similar to the chlorophyll a values, no significant difference was found in the
middle section in the post-spawn period contrasted to the appropriate controls (Figure
2.6-Surface).
Inorganic sediment found in the biofilm samples followed a slightly different pattern than the two parameters described above. Like chlorophyll a and AFDM, inorganic
sediment trapped by biofilms in the middle reach during active spawning was reduced
compared to the pre-spawn values in the same reach (3.0 x) and the upstream section
during the same period (3.4x; Figure 2.7-Surface). In contrast to measurements on the
other two biofilm parameters, inorganic sediment was not significantly different in the
downstream section in the post-spawn period but was significant in the middle section
during the post-spawn period. Inorganic sediment was higher when compared to active

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

33

Surface

v^IIJ it:

D50

Section
^ H Down
. Mid
Up

2xD50

te
Pre-Spawn

Active-Spawn

Post-Spawn

Sampling Period
Figure 2.5: Chlorophyll a from gravels sampled in the HFC over the course of a salmon
spawning event Bar heights are mean values with error bars representing
± 1 SEM Gravels were sampled at three depths which are indicated by the
panel heading An • indicates a significant difference in the contrast test A x
symbol indicates a non-significant contrast The • symbol is an indicator of a
non-significant differences in the starting conditions All surface means were
contrasted according to Table 2 3

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

34

spawning period in the middle section (3.7 x) but lower when compared to the upstream
control during the same period (1.6x). Both values were contrasted in the same manner
as above (See Table 2.3).
2.3.3.2

Sub-surface Biofilms

The mean chlorophyll a, AFDM and inorganic sediment from biofilms sampled below the
surface at 5 and 10 cm demonstrated no statistically significant difference with respect to
spawning period and experimental section (all p>0.05). The lower two panels of Figures
2.5, 2.6 & 2.7 demonstrate the muted response seen at depth over the course of the salmon
disturbance regime. The pattern of reduced chlorophyll a and AFDM in the middle section during the active spawn period appeared to only extend to the 5 cm depth although
this result is not significant.
2.3.3.3

Stable Isotopes

Across all three parameters tested, there was a significant effect of sample period (Table
2.4). Additionally, section was a significant factor for <)15N values. All post-hoc tests were
pairwise comparisons used t-tests with Holm's correction. Levels of 515N were significantly higher during the active-spawn (p<0.000) and post-spawn (p<0.000) periods than
the pre-spawn temporal control. Post-hoc comparisons revealed no significant differences
in the mean amount of <*)15N across sections. Levels of <513C were significantly greater during the active spawn period than the post-spawn period (p=0.019). However, there was
no significant difference in the remaining period comparisons for r513C (See Figure 2.8).
All pairwise comparisons of period means of C:N ratios were significantly different from
each other (all p<0.05).

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

35

Surface

D50

( Section
^ H Down

' 1

i

Mid

i

Up

1
2xD50

Pre-Spawn

Active-Spawn

r '1

Post-Spawn

Sampling Period
Figure 2.6: AFDM values from gravels sampled in the HFC over the course of a salmon
spawning event. AFDM was determined by ashing GFF filters. Bar heights are
mean values with error bars representing ± 1 SEM. Gravels were sampled at
three depths which are indicated by the panel heading. An A indicates a significant difference in the contrast test. A x symbol indicates a non-significant
contrast. The • symbol is an indicator of a non-significant differences in the
starting conditions. All surface means were contrasted according to Table 2.3.

Chapter 2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

36

Surface

•

i

D50

Section
^ H Down
Mid
Up

2xD50

Pre-Spawn

Active-Spawn

Post-Spawn

Sampling Period
Figure 2.7: Inorganic sediment values from gravels sampled in the HFC over the course
of a salmon spawning event. Inorganic was determined by ashing GFF filters.
Bar heights are mean values with error bars representing ± 1 SEM. Gravels
were sampled at three depths which are indicated by the panel heading. An
• indicates a significant difference in the contrast test. A x symbol indicates
a non-significant contrast. The • symbol is an indicator of a non-significant
differences in the starting conditions. All surface means were contrasted according to Table 2.3.

Middle

Downstream °

Upstream v

513C
24 0 -

14 0 -

Salmon Flesh 6 1 5 N = 10 78
40

9

C N

*r

Salmon Flesh C . N = 3 84 .
u

24 2 J

M

24 4 -

35

13 5 -

\

3
!

24 6 -

30 -

Salmon Flesh 513 C = - 2 0 61J

25
55

13 0 -

"

'

\

V

24 8 -

c

CL

i

12 5 -

a

12 -

o

w

-24 -

50-

T

i'

45

*'

40

T

:

10-

n

^-^

9-

§"

K^_

-25 -

11 -

Q.
^—1.

8 -

-26 -

35

Si

"
7 i

.**<*

^

At

^

At

^

At
•^c

^

At
<^

^

^

«^°

^

At

n&

I

ctf*

^

.^
At

At
«&

^

I

At <**

r§

<1<P'

Period

Figure 2.8: Isotopic ratio of benthic biofilms and infiltrated sediment. Ratios moving towards salmon flesh values indicate a
marine nutrient source although <513C values can be confounded by photosynthetic processes (Staal et al., 2007).
Isotopic ratios were calculated as per equation 2.1. The far right panel shows the carbon to nitrogen ratio of both
biofilms and sediment. Salmon flesh values were sampled from fresh salmon carcasses (n=4). Points are mean
values ± 1 SEM.

I
5"
&

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

38

Table 2.4: Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on surface <515N , <513C and the carbon:nitrogen molar ratio
from benthic biofilms. The minimum adequate model (MAM) only included
an interaction term for C:N ratio. This interaction term was non-significant so
only the main effects were tested for all models.
615

813 ; C

'N

C:N

Source of Variation

Df

Sum Sq

Pr(>F)

Sum Sq

Pr(>F)

Df

Sum Sq

Pr(>F)

Period
Section
Section x Period
Residuals

2
2
22

0.351
0.066
0.196

0.001
0.041
-

8.769
4.618
17.897

0.012
0.080
-

2
2
4
14

3.677
21.807
5.837
7.570

<0.000
0.063
0.074

2.3.4

Infiltration Bags

2.3.4.1

Stable Isotopes

Table 2.5: Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on r515N and <513C from infiltrated sediment. The minimum
adequate model (MAM) included only the residual term for C:N ratio indicating that neither spawning nor die-off had any effect on intergravel C:N ratios.
Therefore, the C:N model summary is not included here. Dashes (-) indicate a
dropped parameter in the MAM.
t515N

<5 13 C

Source of Variation

Df

Sum Sq

Pr(>F)

Df

Sum Sq

Pr(>F)

Period
Section
Residuals

2

0.166

0.128

0.887

1.405
0.213
1.127

<0.000
0.148

24

2
2
22

Only (513C of intergravel sediment was affected by the presence of salmon (Table 2.5).
A significant period term in the ANOVA model and significant post-hoc comparisons
indicate a large spike in 513C levels during the active-spawn period and a subsequent
decrease in the post-spawn period (All p<0.01) although this trend is consistent across
across all sections and is small in magnitude (Figure 2.8).

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream
2.3.4.2

39

Particle Size

The lower end of the particle size distribution (<i16) remained relatively stable in each
section. Similarly, the pre-spawn and active-spawn d8i remain stable across all sections
(Figure 2.9a). The most notable change in particle size is seen in the post-spawn period
in the downstream section. Post-spawn values in this section were higher than similar
particle sizes in the middle and upstream section. There was, however, a general trend of
increasing particle size in the post-spawn period but this result is not significant (p>0.20).
Figure 2.9b supports a general coarsening trend of material being deposited in infiltration
bags corresponding to salmon activity. This result, however, was not statistically tested
due to a lack of rigourous multi-sample distribution tests. Furthermore, distribution comparisons, like the Kolmogorov-Smirnov test, only test for difference and not location of
difference reducing their usefulness in this particular context (Quinn and Keough 2002;
But see section 2.3.5). A qualitative assessment of particle size via Figure 2.9 indicated that
during the post-spawn period, downstream particle size was bigger with approximately
35% of particles greater than 200 pm.

2.3.5

Suspended Sediment

Table 2.6 summarizes the results of an ANOVA on three suspended sediment variables.
The MAM with the organic response variable only included a period term, which was
significant. Pairwise comparisons, however, were not significantly different for any section making interpretation of the organic variable ambiguous. Both period and section
had a significant effect on inorganic suspended sediment. Pre-spawn inorganic sediment was significantly lower than both the active-spawn (p=0.0076) and the post-spawn
periods (p=0.0183). Upstream inorganic suspended sediment was significantly higher
than sediment collected from the middle section (p=0.0026) and the downstream section
(p=0.0012).

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Downstream
i

i

1

Middle

l

1

Active-Spawn

I
o

I

I

150 -

£
D

a

V

V

~~

i

a

50 7
1

(a)

I

2

I

s:

I

0

100 -

i

Post-Spawn

d84 o
d50 a
dl6 v
250 -

0

40

Upstream

I

Pre-Spawn

200 -

Stream

__I

3>

2

M

»

i

1

1

Upstieam

Upstream

Experimental Section

1

I

1

1

1

i

i

i

i

i

i

i

Active-Spawn

Pre-Spawn

i

i

i

Post-Spawn

(b)

„

7>* t

ft

r/

r

fa
Q
N

40

Downstream
Middle
Upstream
i

i
100

1
200

300

1

1

I

1

1

1

o
a
v
1

400

Particle Size (ll m)

Figure 2.9: Range of particle sizes deposited on the stream bed into infiltration bags as
measured by laser in-situ scattering and transmissometry (LISST). Top panel
(a) data points are the mean values ± 1 SEM. Lower panel (b) is the cumulative
distribution of particle size. Both figures are different visual representations
of the same data.

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

41

Table 2.6: Results from a two-way ANOVA of spatial (section) and temporal (period)
salmon treatments on suspended sediment variables. All models were determined using a minimum adequate model (MAM) (Crawley, 2007). OMR refers
to the organic matter ratio of suspended sediment. See section 2.2.6 for details.
Organic

OMR

Inorganic

Source of Variation

Df

Sum Sq

Pr(>F)

Df

Sum Sq

Pr(>F)

Sum Sq

Pr(>F)

Period
Section
Residuals

2

0.919

0.044

0.601
2.347
27.205

0.229
0.004

19.698

2.791
4.037
33.055

0.004
<0.000

137

2
2
135

There was a significant effect of section on the suspended sediment quality (OMR;
Table 2.6). Pairwise comparisons indicated that upstream OMR was significantly lower
than both the downstream (p=0.0083) and middle (p=0.0080) sections. All sections experienced a decrease in the ratio during the active-spawn period although this drop was
most pronounced in the middle and downstream sections (Figure 2.10).
Suspended sediment particle size comparisons between July 14t/l, 2010 (proxy background/ pre-spawn conditions) and September 24th were significantly different (K-S test:
D=5312, p-value<0.000). The K-S test was conducted on the mean distributions of each
sampling date (Figure 2.11). The background proxy sample exhibited greater variability
in particle size and a greater percentage of smaller particles in the system. In contrast, the
particle size characterization taken during the active-spawn period exhibited less variation in particle size and a greater percentage of larger particles.

2.3.6

Intergravel Oxygen

Intergravel DO measured using piezometers was significantly affected by the presence
and arrival of salmon into the HFC (Table 2.7). Post-hoc comparisons are summarized
in Table 2.8. Pre-spawn and active-spawn intergravel DO measurements did not significantly differ. However all other pairwise comparisons for period were significantly
different (p<0.05). The mean post-spawn oxygen level was significantly higher than the

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Downstream

o

•

Middle

Stream

Upstream V

i

Orgamclnorgamc Ratio

1 6 -

14 -

,0

1 2 -

^^^\
10 -

T
^"^^^

i

1

-''

'*

,.'

—^

I

08 -

"
-L

Organic
0 007 'to
-a
o
^
c

'r

yS

0 006 -

1

T3

/

u 0 005 c
u
a.

f'

3
C/3

/

//
f'

/

-5 1

—
—____^^"~
- ' "

- -J?
"

i

2r " "

y'

,'' _

,'

Ii '

0 004 -

-1In organic
0012 -

'7

0010 -

-1
0 008 -

-r-

^r
0 006 -

^*^**^ - -"1
^^- -''
-

s^''
0 004 -

Pre-Spawn

Active-Sp iwn

Post-Spawn

Sampling Penod

Figure 2.10: Suspended sediment loads and sediment ratios. Suspended sediment was
sampled by an automatic ISCO water sampler placed in the rear portion of
each section. Loads were calculated by multiplying the concentration of sediment by the discharge to get a total load. Mean values ± SEM.

42

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

Stream

43

Sample
— Background Proxy
— Active-Spawn

CD

2'c

"o

»•;

Vf,

Particle Size (|i m)

Figure 2.11: Particle size distributions of suspended sediment in the HFC. Background
suspended sediment particle sizes were sampled on July 14t/l, 2010 while the
active-spawn particle size was taken during the HFC study on September
24*\ 2009.
other two periods (Table 2.8). All spatial section pairwise comparisons were significantly
different. The downstream oxygen values were the highest while the middle section had
significantly lower DO levels.
Table 2.7: ANOVA table of spatial and temporal responses of intergravel dissolved
oxygen (DO) from piezometers.

2.3.7

Source of Variation

Df

Sum Sq

Pr(>F)

Section
Period
Residuals

2
2
166

47.768
83.000
187.664

<0.000
<0.000

Correlations

Downstream surface biofilm chlorophyll a was significantly and highly correlated to downstream surface biofilm inorganic sediment (p-value<0.000, r=0.815) and ash-free dry mass

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

44

Table 2.8: Means and marginal means of intergravel DO. Bold font indicates the grand
mean of the model (Table 2.7). Common letters indicate non-significance in
pairwise t-test's with Holm's correction. All values are mg 1_1.
Downstream

Middle

Upstream

Period

Pre-Spawn
Active-Spawn
Post-Spawn

8.3
9.0
10.4

8.0
7.4
9.4

8.2
8.4
9.4

8.2°
83ab

Section

9.4*

8.F

8.7Z

8.8

Variable

9.7C

(AFDM) (p-value<0.000, r=0.926). Intergravel sediment (the mass of inorganic material
collected from the infiltration bags) and chlorophyll a were significantly negatively correlated (p-value<0.006, r=-0.509)

2.4

Discussion

The results presented here correspond well with other studies that have reported decreases in biofilm abundance during active salmon spawning followed by a post-spawn
increase in biofilm abundance (See Table 1.2; e.g. Moore and Schindler, 2008). Factors
that influence biofilm abundance include the supply of light and nutrients in addition to
hydrologic and physical disturbances (Biggs et al., 1999; Sabater et al., 2006). All biofilms
in each section received approximately the same level of light because of similar tree
cover and were subject to the same experimental flow conditions (Table 2.1). Thus, temporal and spatial biofilm abundance patterns were primarily driven by disturbance from
salmon redd construction and nutrients from salmon carcass decay. Taken together, these
two processes can be called the salmon disturbance regime. Patterns seen in chlorophyll
a, AFDM and inorganic sediment provide strong evidence that the system is unable to
resist the disturbance of redd creation while the system is ultimately resilient evidenced
by its ability to process nutrients from the salmon die-off.
Similar to other studies, the salmon disturbance regime was characterized by two
main periods (active- and post-spawn). The middle section during the active-spawn pe-

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

45

riod was meant to simulate a natural spawning ground. The downstream section during the post-spawn period was intended to simulate a nearfield habitat downstream of
salmon carcass decay. The middle section, during active-spawning was evidenced by
lower biofilm abundance, increased sediment infiltration into the bed, decreased levels of
intergravel DO and a lower OMR. The post-spawn period in the downstream section was
characterized by higher biofilm abundance, lower sediment infiltration, increased levels
of intergravel DO and a higher OMR.

2.4.1

Streambed Benthic Response

Initial low pre-spawn biofilm abundance can be attributed to the preparatory channel
cleaning and suggests young immature biofilms. Chlorophyll a, AFDM and inorganic
sediment patterns in upstream biofilms, therefore, reflect natural biofilm growth in the
absence of salmon. During the active-spawn period, salmon disturbed the streambed
by creating redds to the extent that areas of disturbance were visibly reduced in biofilm
abundance. In contrast, biofilms growing in the absence of salmon (upstream section)
were noticeably thicker and uniformly spread out throughout the reach (Pers. Obs.; Figure 2.12). This pattern is supported by chlorophyll a and AFDM measurements in the
upstream section. The standing stock of surface chlorophyll a and AFDM was significantly reduced in the middle section during the active-spawning period (Figure 2.5 &
2.6-Surface). This decrease in biofilm biomass during active spawning has been previously reported and is usually attributed to the physical reworking of gravels by salmon
(e.g. Moore and Schindler, 2008). Other major biofilm disturbance forces like light and
hydrologic conditions (Peterson, 1996) were kept constant across sections, indicating that
the reduction in biofilm abundance is primarily due to salmon redd construction.
Post-spawn downstream chlorophyll a and AFDM increases suggest a salmon carcass decay influence on biofilm growth (Figure 2.5 & 2.6-Surface). Over the same time
period, biofilms in the middle section did not experience the same level of growth. Com-

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawnmg Stream

46

Figure 2.12: Visual difference in the biofilm abundance present on gravels in the upstream
and middle sections. Gravels in the middle section are visibly reduced in
biofilms and sediment while gravels in the upstream section, grown in the
absence of salmon, are noticeably thicker with biofilm and sediment. Both
images were taken on the same date during the active-spawn period.
parisons with spatial and temporal controls suggest that the increase is not simply due
to natural biofilm succession and growth (Vannote et al., 1980; Peterson, 1996, Figure
2.5-Surface). Rather differences in post-spawn middle and downstream biofilm growth
can be explained by the position of decaying salmon. During the post-spawn period, in
contrast to the middle section, the downstream section had a large upstream source of
decaying salmon (Figure 2.4). Biofilms in the middle section, recovered from the salmon
disturbance primarily via natural succession while the downstream section had the added
nutrient pulse of decaying salmon. Hunt and Perry (1999) found that a strong correlation
between AFDM and chlorophyll a levels suggests an in-stream nutrient source. Thus,
Figure 2.13b provides further evidence that decaying upstream salmon are the source of
the biofilm abundance increase.
An alternative explanation may be that downstream biofilms experienced an "immigration effect" where resuspended biofilm from the middle section acted as a nutrient

Chapter

2. Sediment

and Biofilm Interactions in a Salmon Spawning

25 - (a)

•

•

s
s
e

Stream

04

r = 0.815

gP 20 -

1 15"

••

•

OJ

o
rt

o
c
<u
4*

•

i n .

•

••

•

•

s 02

•

&
Q

••
•
0 5

I

•

-

3
CO

01

00 -

00
i

00

i

i

i

05

10

15

00

05

Surface Chlorophyll a (|4 g/cm 2 )

10

Surface Chlorophyll a (u\ g/cm 2 )

10

05 -

02

04

06

Surface Chlorophyll a (\i g/cm")

Figure 2.13: Relationship between inorganic sediment and biofilms indicators, (a) Relationship between biofilm growth (chlorophyll a) and inorganic sediment
trapped by the biofilm from gravels sampled in the downstream surface section, (b) Strong correlation between two measures of biofilm growth suggest an in-stream nutrient source (Hunt and Perry, 1999). (c) Chlorophyll
a versvis intergravel inorganic sediment collected from infiltration bags that
has deposited into the streambed. Decreased surface biofilm abundance results in larger masses of fine sediment infiltrating into the gravelbed (Bag
depth=0.30-m). All p-values <0.05.

47

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

48

source for the downstream biofilm growth (Peterson, 1996). This explanation, however,
is not supported by either the timing of salmon redd building activity (Figure 2.4) or isotope data (See Section 2.4.2). Moreover, the ultimate source of the nutrients would still
be salmon. Instead the difference in post-spawn biofilms sampled from the downstream
and middle sections is attributed to MDNs.
It should be noted that chlorophyll a values are very similar to values reported in
the literature. In particular, the timing and magnitude of chlorophyll a increases and
decreases reported by Moore and Schindler (2008) is very similar to Figure 2.5-Surface.
Post-spawn increases in chlorophyll a are also very comparable to values reported by Cak
et al. (2008). Both of these studies were conducted in coastal environments suggesting a
common response across a range of habitats.

2.4.2

Stable Isotope Analysis

Table 2.9: Summary of literature stable isotope values and C.N of salmon flesh and values
from this study All values are from adult salmon flesh
Study

Type of Study

Species <513C (%o) <515N (%o) C N Ratios

HFC Study
McConnachie and Petticrew (2006)
Johnston e t a l (1997)
Claeson e t a l (2006)
Bilby e t a l (1996)
C h a l o n e r e t a l (2002)
Kline e t a l (1993)
Hilderbrand et al (1996)

Interior
Interior
Interior
Flume
Coastal
Coastal
Coastal
Coastal

sockeye
sockeye
sockeye
chinook
coho
pink
sockeye
chinook

-20.6
-212
-20 6
-171
-17 9
-20 6
-19 6
-201

10.8
10.9
10 6
15 9
14 2
12 7
12 3
112

3.9
3 41
-

Salmon carcass tissue sampled from the HFC was within the range of other reported
values both for sockeye and other species of salmon (Table 2.9). Salmon S15N was in the
lower end of the range of the examined values although comparable to studies conducted
at similar latitudes and distances from the ocean (Johnston et al., 1997; McConnachie and
Petticrew, 2006). Convergence to the salmon isotopic signature can usually be attributed
to utilization of that nutrient source for growth and development and suggests processing

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

49

of MDNs (Kline et al., 1990) although this is sometimes problematic with rJ13C (See below).
2.4.2.1

Biofilms

Middle and downstream biofilm <515N isotope ratios increased compared to the upstream
control over the course of the salmon spawning event (Figure 2.8). Low source S15N, as
shown in Table 2.9, may account for observed low biofilm <515N values (3.1-5.6%o) as reported values are generally higher (e.g. Bilby et al., 1996, <515N = 7.1%o). Isotopic values
from biofilms collected from the HFC were highest in the post-spawn period suggesting progressively greater MDN enrichment (Kline et al. 1990; Figure 2.8). These results
are typical of other studies that have found increases in <515N values and indicate salmon
nutrient sequestration by biofilms. The increase in <515N also corresponds to increases in
biofilm abundance (Figure 2.5 & 2.6-Surface) suggesting that the abundance increase is
due to a MDN source (Bilby et al., 1996). The increase of 515N is statistically significant indicating that the rapid growth of biofilm following the active-spawn decrease is primarily
driven by marine-derived nitrogen.
The patterns of biofilm <513C isotopic enrichment are not typical of previous reports
that use 513C as a tracer for MDNs (e.g. Kline et al, 1990; Bilby et a l , 1996). The least
negative (i.e. higher) <513C value in this study (-24.0%o) corresponded to the active spawn
period in the upstream section where there were few salmon present (Figure 2.4). Interpretation of biofilm 513C values, therefore, is confounded by several processes that
prevents establishing a clear relationship between MDNs and <513C ratios.
Firstly, variable discrimination of isotopes by biofilms at different stages of development tends to confound isotopic ratios (Kline et al., 1990; Peterson and Fry, 1987). Older,
thicker, more mature biofilms, for example, tend to have higher (less negative) S13C isotopic ratios because of well developed internal carbon cycling processes (Staal et al., 2007).
Carbon isotopic ratios were higher in the upstream section both in the active and postspawn period. Biofilms growing in that area remained undisturbed by salmon and as a

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

50

result were visibly thicker and more mature than disturbed biofilms growing in the middle and downstream sections. Furthermore, elevated 813C levels in mature biofilms may
have been exacerbated by relatively low 813C values of salmon flesh (Table 2.9) which
may have been masked by internal cycling processes.
Second, typically in studies that use the prevalence of marine isotopes to infer a salmon
nutrient source, the degree of r513C fractionation is usually assumed to be small. This assumption, however, may not be warranted. Some fractionation usually occurs in the
uptake of C by algal-dominated biofilms, resulting in elevated 513C biofilm values (Peterson and Fry, 1987; Bilby et al., 1996; Schindler et al., 1997). Surface chlorophyll a values
(Figure 2.5) suggest algal-mediated isotopic fractionation may have occurred (Peterson
and Fry, 1987) and may further explain variation in 813C that does not correspond to the
presence of salmon.
Thirdly, there is considerable natural variability associated with <513C values (France,
1995; Johnston et al., 1997). Three pseudoreplicates per day and replications within each
sample period may have not been sufficient to characterized patterns of <513C enrichment
through natural variability. This variability is in contrast to 515N values (Figure 2.14) and
potentially limits the usefulness of r513C as a nutrient tracer in this type of study design.
Lastly, the patterns of 513C may suggest alternative sources of carbon for biofilms.
In particular, nutrient leaching from upstream macrophytes may have contributed to
an isotopic signature that did not directly represent salmon influence. Isotopic carbon
decreased during the post-spawn period. The decrease coupled with the rapid biofilm
growth during the same period suggest that biofilms were not limited by a carbon source
and may have been using other sources of carbon.

2.4.2.2

Infiltration Bags

Isotopic ratios of infiltrated sediment follow similar patterns in terms of 515N during the
active spawn period. During the post-spawn period, however, there was no difference

(a) Biofilms

Pre-Spawn
Active-Spawn
Post-Spawn

(b) Infiltration Bags

A Downstream
x Middle
N7 Upstream

H
n-

h H |-Jr_-_-_-_-_-_-J4-_-_-_-_-_-^—|
-j-

_l_

r ~"

j

-I

111
ryi

^
^~+--?

rliT
HH-1-1"!-"1

•-:
j

X

V

— I

-27

1

1

1

1

1

1

1

1

-26

-25

-24

-23

-27

-26

-25

-24

513 C %0

1

—

-23

513 C %o

2.14: Dual plots of biofilm (a) and infiltration bag (b) isotopic ratios. Section is represented using different symbols while
the sampling period representative by different error bars line styles. Means ± 1 SEM.

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

52

when comparing downstream and middle section 515N values with upstream <515N values. This pattern corresponds to the re-establishment of the biofilm layer and suggests
that surface biofilms are sequestering most of the MDNs being released into the system
by decaying salmon. Infiltrated values of <513C , while statistically significant, may not
be biologically significant as the range of values are not sufficient to imply incorporation
of salmon organic matter (Figure 2.8). Active-spawn isotope values may be from salmon
waste products (Mclntyre et al., 2008) although these products should have been absent
from the upstream section.

2.4.3 Infiltration and Trapping
Sediment is often ignored in discussions of biofilm and salmon. This omission is usually
attributed to the fact that the relevant studies use a fisheries rather than a geomorphological approach (Kondolf, 2000; But see McConnachie and Petticrew, 2006). The results of
this experiment indicate that sediment plays a crucial role in two ecological processes outlined below, highlighting the importance of abiotic factors on benthic stability and on the
salmon disturbance regime. Furthermore, the inorganic sediment component of biofilm
analysis is rarely reported alongside measures like chlorophyll a and AFDM. These findings highlight important insights that can be gained from using this information.
A persistently high level of suspended inorganic sediment in the upstream section
(Figure 2.10) can be attributed to the structure of the HFC. Most spawning channels in
British Columbia are built with a 2% grade. The HFC, however, was built with a 1%
grade and is annually cleaned to prevent sediment build-up (See section 2.2.1.1). A similar pattern of high background inorganic sediment was also found in the HFC in 2007
(Hulsman and Wubben, 2008) supporting the assertion that background sediment levels
are indicative of regular channel conditions.
Biofilm inorganic sediment levels generally followed the same temporal and spatial
patterns as biofilm abundance (Figure 2.7). During the active-spawn period, surface grav-

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

53

els were significantly cleaned of sediment in the middle and downstream section. Similar
to the decrease in biofilm abundance, this decrease is attributed to physical reworking of
gravel by salmon and implies salmon-mediated addition of inorganic sediment to the water column. One potential concern is that the biofilm communities may have been diatomdominated causing an overestimation in the inorganic sediment values (Conover, 1966).
Microscopic visualization of biofilms, however, suggest that this is not the case although
this is only a qualitative assessment (See Chapter 3).
Salmon-mediated inorganic sediment resuspension to the water column is supported
by suspended sediment values (Figure 2.10). Inorganic sediment load increased in the
water column during the active-spawn period in both the middle and downstream section presumably via redd construction. Surface gravel reaccumulation of sediment then
occurred during the post-spawn period when salmon activity decreased evidenced by
inorganic sediment values in biofilms (Figure 2.7-Surface). Increased sedimentation has
been demonstrated to reduce biofilm growth by shading the algal component (Power,
1990) and inhibit stream restoration (Moerke and Lamberti, 2003). Sediment shading,
however, appears to have not influenced biofilms growing in the HFC. In most cases,
greater sediment deposition accompanied increases in AFDM and chlorophyll a (Figure
2.13a-b). Instead of shading, the increase in inorganic sediment trapped by biofilms suggests a mechanism by which MDNs are delivered to biofilms.
Redd construction resuspends a broad range of particle sizes of sediment some of
which deposits onto the streambed (Rex and Petticrew, 2006). This resuspended sediment, which includes gravel bed aggregated particles, has been demonstrated to form
in-stream flocculated particles when exposed to MDNs (Rex, 2009). Coarsening of suspended sediment suggest the presence of either flocculated particles or aggregates in
the water column (Figure 2.11). Increases in downstream surface inorganic sediment
are closely related to increases in biofilm abundance (Figure 2.13a) which suggests two
mechanisms operating in concert.

Chapter

2. Sediment

and Biofilm Interactions

Biofilm and
Sediment
Resuspension

Salmon
Nutrients

Flocculation

in a Salmon Spawning

k+,

Modified
Settling

Stream

54

Water Flow

«g^_

Redd
Contruction

Biofilm
Retention/
— T ~ ^ — T T V -\: • .-./' \;~N ' >,—^i—

.•v\dw, • j rV'•"-.••':• •*:. • v

Downstream •

X

/

<• • iv,- 1 -: \ A ' r 7 — r y r — T r r ; ' 1 - . • * * ••* '

,«J-V^

r,-.r

;.-.;.--V-y .-„ < )X:-.

w^-..^ •

.•-;. V A . A J'v .->:

U p s t r e a m Control

Middle

/

/ •

/

Figure 2.15: Schematic of a potential mechanism of MDN enrichment of the downstream
section via the flocculation feedback loop (Rex, 2009).
Firstly, it seem clear that floes are forming in the middle section and settling over a
small spatial scale in the downstream section (Figure 2.15). Secondly, increased biofilm
growth may be facilitating particle trapping via EPS and fine sediment trapping interactions (Romani and Sabater, 2000). The biofilm abundance increase is likely being driven
by MDNs (Section 2.4.2) which suggest a positive feedback loop whereby biofilm biomass
increase allows for greater subsequent nutrient enrichment. Rather than shading biofilms,
floes and particle aggregates appear to the source of inorganic sediment and salmon nutrients, aiding in streambed nutrient delivery. Biofilm abundance increases and 5l5N values
suggest these are then rapidly processed by downstream biofilms.
The presence of floes during the post-spawn period is also supported by increases in
the OMR of suspended sediment (Figure 2.10). Since the ISCOs were placed in the rear
of each section, each value is an indication of the suspended sediment being delivered
to the next section. For example, a high middle section value during the post-spawn

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

55

period is the material being delivered to the downstream section in that same period. This
increase provides evidence that floes are forming as floes tend to have a higher organic
matter content. The timing of this increase when salmon are actively decaying in the
water column suggests that the organic matter enrichment is due to salmon. More organic
material present in the water column provide ideal conditions for flocculation (Droppo
et al., 1997).
This floe settling mechanism is also supported by an increase in the size of particles
deposited on the streambed in the post-spawn downstream section (Figure 2.9) as well as
the size of particles present in the water column during active-spawn (Figure 2.11). Increased settling rates of MDN particles from the middle section as a result of flocculation
would explain the nearfield biofilm response (Droppo et al., 1997, Figure 2.15). Additionally, this result provides further evidence in support of the flocculation feedback loop
proposed by Rex and Petticrew (2008).
This work also identifies a previously unreported effect of salmon redd construction.
Rapid growth of biofilms after redd construction appears to aid surface sequestration of
MDNs. A significant negative relationship between intergravel sediment as measured
by infiltration bags and biofilm growth indicates that greater biofilm abundance in the
post-spawn period decreases infiltration of sediment into the streambed (Figure 2.13c).
Similarly, low biofilm abundance during redd construction (Figure 2.5 & 2.6-Surface) is
accompanied by higher sediment infiltration (Figure 2.13c). This suggests intergravel
storage of MDNs during the active-spawn periods when biofilm abundance is low and
streambed surface storage when biofilm abundance is high. A growing surface biofilm
layer may facilitate rapid MDN uptake as photosynthetic activity is diminished deeper
in the streambed (Gibert and Deharveng, 2002; Also see sub-surface panels in Figures 2.5,
2.6 & 2.7). This idea corresponds well with a subsurface increase in bacterial abundance as
seen by Rex and Petticrew (2008) in an artificial flume in response to MDNs and suggests
that bacteria are processing sediment stored during active-spawn (Petticrew and Albers,

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

56

2010).
Surface processing of MDNs is indirectly supported by post-spawn intergravel DO
patterns (Table 2.8). Processing of MDNs by heterotrophically-dominated microbial communities (Yoder et al., 2006) would likely deplete oxygen levels (Horn and Hempel, 1997)
and account for post-spawn decreases in intergravel DO. Reduced light penetration into
the intergravel would have favoured heterotrophic microbes leading to the depletion of
the DO pool (Bastviken et al., 2004). Conversely, during the post-spawn period in the
downstream section, an increase in intergravel DO suggests a decrease in intergravel biological activity by decreased sediment infiltration into the bed (Figure 2.13c) and increased surface biofilm abundance (Figure 2.5). This drop in DO may have also been facilitated by nest digging of salmon. Salmon can construct redds 10-35 cm deep (Schindler
et al., 2003) in the streambed which is approximately the piezometer depth that intergravel DO was sample from. The digging may have facilitated organic matter delivery
further depleting DO levels.
Interpretation of the intergravel DO is complicated by patterns seen during the activespawn period. Sections experiencing approximately the same level of salmon activity
(Middle and Downstream) exhibit opposite patterns of intergravel DO (Table 2.8). From
a management perspective, these differences are not biologically significant as they are
still within the acceptable DO range for proper egg development (Cope and Macdonald,
1996). The difference, however, is persistence across replicates and is statistically significant (Table 2.7). For a brief period, during the active-spawn period in the middle section,
the apparatus used to sample piezometer water was also sampling a small portion of surface water. This discrepancy, however, likely lead to higher levels of DO, not lower. This
difference during the active-spawn period remains unexplained.
As mentioned above, high biofilm abundance during the post-spawn period was accompanied by low sediment infiltration rates (Figure 2.13c). During the post-spawn period in the downstream section biofilms likely trapped MDN laden floes, preventing them

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

57

from penetrating further into the bed. The EPS of both biofilms and floes are known to interact suggesting a possible mechanism for this trapping ability of biofilms (Sutherland,
2001). The biofilm layer at the surface thus can utilize nutrients and photosynthesize
more readily allowing for quick uptake of MDNs. Furthermore, this biofilm layer acts
as a barrier to subsurface storage. These results have implications for salmon as habitat
regulators as the temporal and spatial conditions of spawning and die-off may dictate the
degree in which MDNs are incorporated.

2.4.4

Implications

This pattern of biofilm abundance shifts in response to salmon spawning and die-off is
indicative of the high benthic resiliency to the salmon disturbance regime as biofilm abundance rebounded from low active-spawn levels. Additionally this highlights the important role that biofilms play in trapping sediment. As a basal portion of benthic foodwebs,
biofilms often structure benthic resilience and resistance ultimately determining overall
system stability (DeAngelis et al., 1990; Stoodley et al., 2002). The interaction and thus
the importance, however, between sediment and biofilms with respect to intergravel and
streambed MDN storage patterns is also highlighted by these results.
Well oxygenated intergravel DO is crucial for proper egg development and is one of
the primary goals of spawning channels (Toews and Brownlee, 1981). The establishment
of the biofilm layer and the corresponding increase in DO suggest that the downstream
areas in the post-spawn period are crucial both for nutrient retention and viable egg development. Active-spawn DO levels are well below saturated levels and sediment levels
are also likely too high for proper egg development. In an example, however, of the resiliency of the total system, DO levels rebounded and sediment infiltration decreased,
re-establishing saturated DO which are ideal conditions for egg development. Primary
remineralization of MDNs has also been demonstrated to occur mostly via autochthonous
production supporting the link between adult to juvenile sockeye via increased basal

Chapter 2. Sediment and Biofilm Interactions in a Salmon Spawning Stream

58

component abundance (Kline et al., 1993). Thus nutrient retention at the surface provides both a spatially stable nutrient source for optimal incubation conditions and future
salmon.

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

59

Chapter 3
Benthic Biofilm Composition in a Salmon Spawning Stream: A
Microscopic Approach to Biofilm Characterization

3.1

Introduction

Improved experimental methods have reformed our perception of microorganisms beyond the single culture models into more complex biofilm frameworks. That is, there is
now a recognition that most microorganisms reside within some biofilm matrix (Sutherland, 2001). With a greater knowledge of biofilm structure and function, it is now realized
that biofilms represent complex microbial communities that display the same characteristics and properties as any other ecosystem (Battin et al., 2007).
Biofilms can be broadly defined as the accumulation of single-celled organisms (both
prokaryotic and eukaryotic) and extracellular polymeric substances (EPS) (Costerton et al.,
1995). In the context of aquatic systems and this study, biofilms are specifically defined as
the autotrophic and heterotrophic microbial communities embedded in EPS that grow in
the benthic zone. The morphology and amount of EPS associated with a biofilm growth
is important for the nutrient trapping ability of biofilms (Neu and Lawrence, 1997). Thus
changes in EPS structure and growth due to a MDN pulse may be important for the assimilation of suspended nutrients into biofilms. Inorganic sediment contributes to biofilm
structure as a stabilizer. EPS binds sediment and cellular material together and this interaction between EPS and sediment confers a stability to the biofilm (Wolfaardt et al., 1999)
enabling greater development.
The function of a biofilm matrix is intimately tied to its structure. Stoodley et al.
(1999) found that biofilms grown under nutrient rich conditions developed significantly

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

60

different morphologies than biofilms grown under nutrient poor conditions. Specifically,
biofilms in nutrient rich conditions respond with greater biomass, increased biofilm thickness, increased algal/bacterial ratio and greater SA coverage allowing for greater nutrient uptake (Stoodley et al., 1999; Sabater et al., 2002). This pattern, however, has not been
tested in an MDN context. The presence of MDNs, particularly during the post-spawn
period, would suggest elevated nutrient concentrations (e.g Tiegs et al., 2009) and a similar structure change as demonstrated by Stoodley et al. (1999).
In addition, biofilm morphologies adapted to changing nutrient conditions by developing 'streamers' and 'cell clusters'. Battin et al. (2003b) demonstrated the functional
significance of this morphological change. Battin et al. (2003b) showed that streamers
grown due to enriched nutrient conditions altered the hydrodynamic environment near
the streambed interface. This change in microhydrodynamic conditions resulted in increased settling velocity of suspended organic particles. This suggests a positive feedback
loop in biofilms whereby nutrient rich water set the conditions necessary for the optimal
utilization of nutrients by biofilms.

3.1.1

Confocal Laser Scanning Microscopy

Since the first usage of confocal laser scanning microscopy (CLSM) for the analysis of
river biofilms (Lawrence et al., 1991), it has become both a powerful and standard tool in
biofilm research. The use of CLSM has significantly expanded our knowledge of biofilm
structure (Neu et al., 2005; Battin et al., 2003a), composition (Lawrence et al., 1998; Neu
et al., 2005), functioning (Neu et al., 2001) and development (Neu and Lawrence, 1997;
Mohamed et al., 1998). Use of CLSM in biofilm research has become so pervasive that in
many respects it now the de facto tool with regards to process-based biofilm research.
Conducting biofilm research using CLSM is advantageous over other microscopic
methods (scanning electron or environmental-scanning electron microscopes) when the
research questions being asked are biologically process-based. Numerous unique com-

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

61

pounds exist within the biofilm matrix, each contributing to the process and functioning
of biofilms (Wingender et al., 1999). Traditional analytical methods require separation,
purification and identification of unique biofilm components to infer function and address process-based questions (Neu et al., 2005). With CLSM, paired with fluorescent
markers, biofilm components can be viewed in a fully hydrated and natural state. Samples can be non-destructively sampled by growing biofilms on a portable and removable
medium (e.g. polycarbonate slides). Samples can be taken directly from a field site with
little to no preservation. Other methods of visualizing biofilms do not always preserve
the natural composition of the sample as sample preservation often disturbs these parameters (Neu and Lawrence, 2004). The need for sample preparation is also greatly reduced
as compared to other microscopic methods (Lawrence and Neu, 1999).
The need for higher resolution biofilm research in all areas of river aquatic ecology
is strong. The ecological importance of the biofilm layer is recognized (Battin et al.,
2007) while information on attached bacterial and algal communities (biofilms) is lacking
(Manz, 1999; Neu and Lawrence, 2004). Phenomenal progress has been made in terms of
our understanding of biofilm and nutrient dynamics and much of this progress can be attributed to sophisticated CLSM techniques (See Battin et al., 2003a; Pauling and WagnerDob ler, 2006). This progress, however, has not translated across all related fields and
CLSM remains primarily a specialized microbiological tool rather than an broad ecological one.

3.1.2

Objective

Table 1.2 lists several important bio film-based studies of salmon habitat and spawning.
Although, ash-free dry mass (AFDM) and chlorophyll a remain important coarse measures of biofilm mass and composition, finer details of biofilm structure and function require higher resolution tools. CLSM characterization has been highlighted as a promising
technique to aid research in nutrient dynamics within aquatic systems (Cross et al., 2005).

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

62

The Horsefly River spawning channel (HFC) and the controlled loading of salmon into
the HFC served as an experimental environment to a) examine changes in biofilm composition in response to salmon carcass decay and b) test the feasibility of CLSM analysis
in a remote field setting.
The goal of this research was to image two functional components of biofilms (bacteria
and algae) growing in the HFC after a salmon spawning event. This was done via multiple parameter imaging to determine changes in abundance patterns of these components
in response to salmon carcass decay which is hypothesized to affect biofilm community
composition. Three research questions were addressed to accomplish this goal.

3.1.3

Research Questions

3.1.3.1 Question One
How do the bacterial and algal components of biofilms change in response to varying
nutrient concentrations from marine derived nutrients (MDNs)? Rotting salmon release
nutrients that are utilized by the benthic community and in particular biofilms (Johnston
et al, 2004).
3.1.3.2

Question Two

What is the impact of these compositional changes in algae and bacteria on temporal and
spatial patterns of MDN storage by biofilms?
3.1.3.3

Question Three

How do common destructive measurements of biofilm composition and biomass (chlorophyll a and AFDM) compare to CLSM measurements of these same parameters? Previous
salmon-biofilm studies have not made use of CLSM to characterize the biofilm response
to MDNs (Table 1.2). This study will relate these two types of measurements and explore

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

63

sources of variation.

3.1.4

Secondary Objectives

A secondary objective of this work was to establish a protocol for future users of the
CLSM at University of Northern British Columbia (UNBC) to adapt to their own uses
especially in a remote field setting.

3.2

Methods

3.2.1

Site Characteristics

Biofilms were sampled from the HFC in Horsefly, British Columbia. A more detailed description of site characteristics can be found in section 2.2. Briefly, experimental sections
of the HFC (n=3) were characterized by two different densities of salmon spawners and
a spatial control with no salmon (Figure 2.2). The upstream section served as a spatial
control with little salmon influence while the middle and downstream sections received
inputs of salmon organic matter.

3.2.2

Biofilm Growth

For the analysis of river biofilms, removable slides were attached to ceramic tiles that
were placed directly on the streambed. The slides were made of polycarbonate strips,
a suitable growth substrate for biofilms (Lawrence and Neu, 1999). For each sampling
date, five slides (one stain control & four samples) were collected from each experimental
section, immediately immersed in river water and placed in a sealed container. Samples
were transported for three hours in a refrigeration unit and stained and viewed on the
same day.
Five ceramic tiles were placed in each section of the HFC on September 25 t/l , 2009 and

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

64

Figure 3.1: Schematic of the slide mounting system used in the HFC. Polycarbonate strips
have been previously identified as suitable growth substrates for CLSM analysis of biofilms (Lawrence et al., 1998).
sampled weekly until October 27th, 2009. Tiles were placed in the thalweg of the HFC.
Because of reduced salmon activity, the slides were not physically disturbed during this
period and any growth on the polycarbonate strips can be viewed can cumulative biofilm
growth over the post-spawn period.
3.2.2.1 Confocal Specifications
An Olympus Fluoview 1000 with a multiline argon gas laser (458,488 & 515 nm) and two
independent Helium-Neon gas lasers (543 & 633 nm) mounted with an inverted Olympus
microscope was used to image biofilms. Observations of biofilms grown on polycarbonate strips were made on a 60x water immersible lens, 1.2 numerical aperture. Scanning
was done sequentially to minimize photobleaching (Pawley, 1995). Optimum settings
for each sample period were determined so that all microscope parameters were kept
constant with a minimum number of saturated pixels (i.e. white value of 255) to ensure
intercomparability between samples. Five fields of view from each sample were taken to
account for biofilm spatial variability (Neu et al., 2005). These five fields of view were
averaged to form a composite sample.

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream
3.2.2.2

65

Stains and Fluorescent Markers

All staining protocol was derived chiefly from Lawrence et al. (1998), Strathmann et al.
(2002) and Neu and Lawrence (2004). Bacteria nucleic acids were stained with the nucleic
acid stain, SYT09 (ex=488 nm, em=522/32 nm; Molecular Probes, Inc.). For each staining
period 1 pi of SYTO 9 and 1 ml of distilled water were mixed to form a stock staining solution. Two to three drops of this stock solution were added to a fresh biofilm sample under
subdued light and incubated at room temperature for 15 minutes (Neu and Lawrence,
2004). A washing step is not necessary with SYT09 (Lawrence et al., 1998).
Autofluorescence in the far red channel was used as a measure of algal abundance.
Lawrence et al. (1998) found that algal cells fluorescence brightly at a excitation wavelength of 647 nm and a detection of emission of 680/32 nm. The Olympus Fluoview technical specification required a slight modification to the laser parameters used to detect
algal autofluorescence which was deemed acceptable (ex=633, em=647nm; John Lawrence
pers. comm.). Algal abundance was corrected for cyanobacteria autofluorescent interference by subtracting any autofluorescence in the red channel (ex=543nm, em=578nm) from
algal autofluorescence in the far red channel (ex=633, em=647nm).

3.2.2.3

Image Analysis

Image analysis was conducted by the same person using the same computer monitor using the same brightness and contrast settings. This consistent approach to image analysis
ensures that any potential biases or errors are kept constant throughout the experiment
(Neu, 2000). Image J (Rasband and ImageJ, 2009) was used to convert the Olympus file
formats to 24-bit RGB stacked TIFF files with a semi-automated macros. These files were
sufficiently high in contrast to justify any loss in dynamic range by the conversion to a
24-bit TIFF file. Images collected from each field of view were 512 pixels by 512 pixels.
The TIFF files were then loaded into Scion Image for subsequent image analysis. Images from each CLSM channel were thresholded to define the boundary of the objects in

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

66

the image. A semi-automated macros was then used to make the image binary, dilate and
erode the image and then count the number of white pixels in that particular channel.
Dilation and erosion help better define the borders of the object while eliminating noise
from the signal for a more accurate white pixel count. This white pixel count was used
as the percent coverage for either the algal and bacterial components of the biofilm for a
field of view.

3.2.3

Nutrient Delivery Estimates from Salmon Carcass Decay

Salmon decay products were modelled using estimated loss rates from the Takla River
(Johnston et al., 2004) and daily fish counts from the Horsefly Channel in 2009. The
amount of nutrients released into the water column and available for biofilm sequestration was estimated using the same equations outlined by Johnston et al. (2004). The
model outlined below, therefore, generates an estimated value for the mass of nutrients
lost from salmon carcasses and delivered to the water column on a daily basis.
The total new nutrient contribution of all salmon dying on a given day, t, is estimated
by:
Nut(t) = D(t) x % Nutrient (Carbon or Nitrogen)

(3.1)

where D(t) is the number of dead salmon that died on day t and % Nutrient is the average
percent composition of salmon flesh for either carbon and nitrogen. The total amount of
salmon nutrients present in the stream on day t is given by:

Ini(t) = Nut(t) + Rem(t)

(3.2)

where Nut(t) is the new arrival of nutrients on day t (equation 3.1) and the Remit) is the
total amount of nutrients remaining in the system on day t from all previous days. Rem(t)

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

67

accounts for the mass and nutrient loss of carcasses from previous days and is given by:

Remit) = Im(t) - Loss{t)

(3.3)

where Ini(t) is given by equation 3.2. Thus the daily mass loss of salmon (for both carbon
and nitrogen; Loss(t)) to the water column on day t is given by:
Im(t) x (1 - ek)

(3.4)

where Ini(t) is calculated according to equations 3.1-3.3 and k is the constant decay rate
of sockeye salmon given by Johnston et al. (2004) from Takla River. The decay rate from
Johnston et al. (2004) for carbon was -0.0360 kg/day and -0.0460 kg/day for nitrogen. The
downstream section of the HFC received nutrient contributions from salmon decaying in
both the middle and downstream sections (See Figure 2.2). Therefore, the nutrient load
received by the downstream section from decaying salmon was the sum of the middle
and downstream values generated by equations 3.1-3.5.
3.2.3.1 Carcass Removal
Over the course of the experiment, some salmon were removed from the HFC system
either via dead pitching or black bear (Ursus americanus) consumption. Dead pitching was
done to reduce the number of decaying salmon down to a more natural representation of
a spawning stream. If there was a net loss of fish on a given day (i.e. D(t) — D(t — 1)
was <0), the salmon nutrients removed on that day from the stream was estimated by
using the average nutrient content of salmon from the previous day and multiplying that
number by the total number of fish removed from the system. Total nutrient removal is
therefore estimated by:

flm(t-l)\

Total nutrient removal = I —
— * Fish Removed(t)
V D(t - 1) )

(3.5)

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

68

Dead salmon (D(t)) were enumerated visually by two individuals on a daily basis.
In instances where the counts differed greatly (>10 salmon), the salmon were recounted
until a similar count was reached. The mean nutrient content of salmon was estimated by
randomly sampling four freshly dead salmon from the study reach on October 5th, 2009.
A small portion of somatic tissue was removed from the fish and immediately frozen.
Samples were then freeze dried and shipped to an external laboratory (Pacific Centre for
Isotopic and Geochemical Research, University of British Columbia). Using equations
3.1-3.5, a carbon to nitrogen molar ratio was estimated for the sampling period.

3.2.4

Assumptions

Several assumptions are made in the accounting of dead pitched and bear removed salmon
carcasses. The first is that carcasses were removed randomly, such that all levels of decay have an equal chance of removal. Evidence of bear removal was seen throughout the
channel not just the banks. Moreover, bears removed fish from the HFC with evidently
little preference for any particular decay stage of fish. Salmon in an advanced state of decay were consumed by bears just as readily as fresh carcasses. Dead pitching was done on
one day (September 26"1, 2009) and all dead fish were removed from the system. This indicates that fish were randomly removed from the channel and this method of accounting
for salmon nutrient content is appropriate.
A second key assumption is that the calculated loss rate for each nutrient does not vary
between the Takla River (Johnston et al., 2004) and the Horsefly River. Water temperature
during the period of salmon decay would likely be the biggest difference between sample
areas. To assess the differences, average temperature during the decay period from both
streams were compared. The mean temperature in the middle section of the HFC during
the post spawn period was 9.63°C (SEM=0.116). The mean temperature for Takla River
was taken from Figure 3 in Johnston et al. (2004) and was calculated using Engauge Digitizer (Mitchell, 2010). The mean value of all streams for every year during the post-spawn

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

69

period was calculated to give a single value. This mean temperature was 9.88°C (SEM=
0.121). These are comparable river temperatures and this small difference indicates that
using the Takla River loss rate is acceptable to use in the Horsefly system.
A final assumption is that biofilms sampled for CLSM analysis are utilizing MDNs.
Biofilms used for microscopic analysis were not specific analyzed for r515N and <513C isotopes. Results presented in Chapter 2, however, indicate that biofilm growth in the HFC
was driven by MDNs. Thus the assumption for this chapter is that biofilms growing on
polycarbonate slides follow a similar pattern of MDN incorporation.

3.2.5

Data Analysis

3.2.5.1

Manipulations

Biofilms were sampled on four occasions while nutrient release by salmon was estimated
on a daily basis. To examine the relationship between salmon carcass decay and biofilm
composition, the average daily nutrient release over the biofilm sampling interval was
used as a comparison against biofilm measures. For example, if biofilms were sampled
on October 5th and October 15 th , the average daily nutrient release from salmon carcasses
in between those dates was used as a measure of salmon nutrient contribution to the
October 15*'1 biofilm sampling date. For the downstream comparison, the decay products
from both the downstream and middle sections were used as biofilms in that section had
nutrient contributions from both sections.

3.2.5.2

Statistical Analysis

Mean differences in biofilm composition between each section was compared statistically
using a two-way ANOVA. Three separate ANOVAs were conducted using the percent
coverage of bacteria, algae and the algakbacterial ratio as response variables. Sampling
week and experimental section were used as factors. Pearson's correlation test was used

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

70

to relate estimated salmon decay products to biofilm composition. Given that salmon
were only present in the middle and downstream sections, upstream composition values
were excluded from this particular nutrient analysis.
An additional objective of this research was to compare CLSM to the method outlined
in section 2.2.4.2. Results of that spectrophotometric method were compared to algae
percent coverage measured with CLSM using Pearson's correlation test. A maximum
of two days separated spectrophotometric sampling and CLSM sampling. Sampling occurred approximately every week and it is these samples which are compared with Pearson's correlation test. In addition to spectrophotometrically measured algae, ash-free dry
mass (AFDM) was measured to determine the total biological material present on gravels. AFDM was compared to the total percent coverage as an additional test. All statistical analysis were conducted using R 2.11.1 (2010). All graphics were created using
R 2.11.1 (2010) with the memisc (Elff, 2010), ggplotl, (Wickham, 2009) and lattice (Sarkar,
2008) packages.

3.3

Results

It was the initial goal of this biofilm research to collect samples for CLSM from the entire
salmon disturbance regime 1 . Salmon activity, however, prevented sampling of biofilms
throughout the active spawn period. Salmon redd construction moved sufficient gravel
to completely bury most of the clay tiles. Therefore, the clay tiles were redeployed after
salmon activity had subsided and samples were only collected in the post-spawn period.
All subsequent analysis of biofilm composition change applies only to the post-spawn period of the salmon disturbance regime. This limits the inference of this particular portion
of the study to the response of biofilm to in-stream carcass decay of salmon.
Designing studies that incorporate the entire salmon disturbance regime is a more accurate representation of the ecological role of salmon (See Chapter 2; Janetski et al., 2009).
^ e e section 1.1 for a complete description of the salmon disturbance regime

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

71

Four sampling dates constitutes relatively weak statistical power and some care should
be taken in over-extrapolating these results. Nevertheless, many biofilm studies are conducted in the absence of any statistical and quantitative analyses (But see Lawrence et al.,
1998).

3.3.1

Site Characteristics

Site characteristics are identical to those presented in section 2.2.1.2. Figure 3.2 has been
truncated from Figure 2.4 to reflect the CLSM sampling period. Conditions in the HFC
remained relatively stable over the course of the sampling period. The HFC experienced
a slight increase in discharge at the beginning of the sampling period. This increase was
due to a build-up of salmon carcasses at a downstream fence causing water to back-up,
subsequently raising the water level at the staff gauge. The Horsefly region experienced
relatively few high intensity storms or rain events during the sampling period. The small
events shown in Figure 3.2 were not likely important in terms of biofilm composition.

3.3.2

Microscope Use and Image Analysis

For each sampling period, the CLSM was run under optimal conditions using freshly
stained samples with a single operator. The microscope lens was cleaned in between
samples using the manufacturer's lens cleaner and 90% ethanol. Figure 3.3 is a screenshot
of the image analysis process used to quantitatively assess CLSM images.

3.3.3

Biofilm Component Patterns

Biofilm components were assessed based on the levels of coverage demonstrated by a particular parameter. The total number of pixels occupied by all objects in the corresponding
channel (algae or bacteria) was divided by the total number of pixels in the field of view
(512x512) to determine a percent coverage. Figure 3.4 shows the patterns of composition

Chapter

3. Benthic Biofilm Composition

in a Salmon Spawning

Stream

72

(a)

(b)

.*
Date

Figure 3.2: Site characteristics for the HFC over the CLSM sampling period. See section
2.2 for description of rainfall and discharge collection methods.

Chapter

3. Benthic Biofilm Composition

in a Salmon Spawnmg

El

Stream

|;j§.

t, 'X.
•P'.-

i«„

•fe

-1,'V

A^SSGIS.*-

. s- <Bt t»i

P

V ^VJ '_KJ

JN«

Dlate

df
Fig

.3: A single channel example of the image analysis process of thresholding, dilation, and eroding to determine white pixel counts. Eroding and dilation are
important steps to remove noise of the CLSM image.

73

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

74

change over four weeks of post-spawn salmon decay.
Biofilm components were assessed statistically using a two-way ANOVA. No significant differences between sections and sampling week (with one exception) for any
response variable was demonstrated over the post-spawn period (all p<0.05) although
statistical power was weak due to low degrees of freedom and ultimately a small sample
size. The lone exception in this analysis was a significant effect of time on the percent
coverage of algae (Fi6=0.945, p=0.034) although the effect was not significant across section. A qualitative assessment of Figure 3.4 suggests a general increase in total percent
coverage of biofilm. Slides collected on October 21 i/l had a greater proportion of algae in
the biofilm from samples collected in the middle section. On the subsequent sample day
(October 27th), there is a noticeable drop in algae coverage in the middle section but an
increase in downstream bacterial coverage. This result, however, is purely qualitative.

3.3.4

Salmon Nutrient Composition

The number of salmon present (live and dead) in the HFC during the sampling period
is presented in Figure 3.5. The salmon counts in Figure 3.5 were values used for D(t)
from equation 3.1. A subsample of fish were analyzed for elemental carbon and nitrogen
(See section 2.2.3). The average carbon composition of sockeye salmon in the channel was
47.81% (SD=1.06; n=4) of the total mass. The average nitrogen composition of salmon
in the channel was 14.51% (SD=0.40; n=4) of the total mass. These values were used,
respectively for nitrogen and carbon, as % Nutrient in equation 3.1.

3.3.4.1 Nutrient Influence
Figure 3.6 represents daily salmon decay products as calculated by equations 3.1-3.5.
Salmon were primarily in a die-off phase when CLSM sampling begun (Figure 3.5). There
was only one significant correlation between salmon decay products and biofilm composition. This significant relationship was in the middle section between the C:N ratio and

Chapter

3. Benthic Biofilm Composition

in a Salmon Spawning

Stream

Section
H | | Downstream
1 5-

j||

Middle
Upstream

1 0-

05-

0)

>

o
O

c
a>
o 1 0-

aj

06040200-

I

2009-10-05

2009-10-15

2009-10-21

2009-10-27

Date

Figure 3.4: Percent coverage of biofilm components on polycarbonate strips as measured
by CLSM Components are arranged into panels according to the channel
wavelength fluorescence Error bars indicate SEM which is based on replicate
sample strips from separate tiles in each section Each strip was sub-sampled
with the CLSM five times (l e Five fields of view) to account for spatial variability with the biofilm These five fields of view were averaged to give a single
value for each tile Strips from five tiles were sampled for each sampling date

75

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

1

i

i

i

i

1

i

1

1

1

1

I

1

l

1

r-v

*

Middle
Upstream

©

m ©
©

1

Live salmon

om <N
o
o<N -

B
O

1

76

' -^

-

©

6
VI

©

-

Dead salmon

o

a
3

©
©

-

\

©

z ©© -

----

/

©

o i

l
^

T!P

i

^

^

i

^

^

i

i

<S<^
cf> 0 <>

1

d*

1

d*

1
1
1
-?
O
•$•
O*
O*
d*

I

1

l

/P
cf

l

l

A>
S?
o°
d*

1

A>
0°

Date

Figure 3.5: The number of salmon present, both live and dead, in the HFC during the
confocal study period.
the bacterial coverage. In addition, this was a strong positive correlation. In contrast,
the algal component of the growing biofilms were not significantly related to either the
downstream or middle O N ratio (Algae; Figure 3.7) nor were biofilms in the downstream
section related to salmon decay products.

3.3.5

Method Comparison

In addition to biofilm composition patterns in the post-spawn period, this research also
assessed differences in a) spectrophotometrically measured algae abundance and CLSM
measured algae via autofluorescence and b) AFDM and total biofilm coverage as measured by CLSM. The method for spectrophotometrically measured algae abundance and
AFDM is outlined in section 2.2.4.2. Neither coarse measure of biofilm (Chlorophyll a or
AFDM) was significantly correlated to either algal or total percent coverage (Figure 3.8).

Chapter

3. Benthic Biofilm Composition

in a Salmon Spawning

Stream

77

Downstream

<$ ¥

T>' ¥

¥ / « " ¥ # & & # $ <§> ¥ •& NN O V * • Nfc <N • ^ <£ i> T> # TK T? ^
Date

T> $>

Figure 3.6: Salmon decay products as calculated from equations 3.1 - 3.5 and used for
comparison in Figure 3 7. Shaded portions of this figure represent the period
between CLSM sampling dates (Sampling dates also indicated in Figure 3 4)

Chapter 3. Benthic Biofilm Composition

in a Salmon Spawning Stream

Downstream

Middle

5 ^ - p-value = 0.395

78

p-value = 0.195

O
O

os--

p - value = 0.488

p - value = 0.009
r= 0.991
A

j&

Salmon Decay Products C N Ratio

Figure 3.7: Relationship between salmon decay products C:N ratio and the percent coverage of biofilm components. Downstream decay product were calculated as
the sum of the middle and downstream values. Annotations refer to results of
a Pearson's Product Moment.

Algae

Total (Bacteria + Algae)
A

p-value = 0 14

p - value = 0 23

cu
en

Section

(0
O

O

o

o

Downstream

A

Middle

o

Upstream

o
1_

(O
Q.

o
s
1 i

i
'

.

r

o,

Chlorophyll a (u. g/cm2)

< H

i'

)

o 1.
AFDM (mg/cm2)

1/:

Figure 3.8: The left panel illustrated the relationship between spectrophotometrically
measured chlorophyll a and algal coverage measured by CLSM. The right
panel is the relationship between total percent coverage and AFDM

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

3.3.6

79

UNBC Confocal

The results presented here were also used to establish a protocol that will assist future
users at UNBC in integrating ecologically based studies with CLSM methods. The primary difficulty encountered during this study was the distance from the sampling site to
the location of the CLSM (« 275-km). Additionally, the software required to undertake
the CLSM image analysis had to be developed from scratch. Software routines and semiautomated macros were established to assist the analysis of CLSM aiding future users
(See Appendix A). Future users will benefit from these software routines in addition to
the protocol developed.

3.4

Discussion

Biofilms thrive in any system that has sufficient nutrient resources (Costerton et al., 1995).
Elevated nutrient levels are a characteristic in spawning streams during the post-spawning
period of the salmon spawning cycle (Naiman et al., 2002). Thus, biofilm growth occurs in
salmon streams in the post-spawn after the fish carcasses begin to decay and release nutrients (e.g. Chaloner et al., 2007; Chapter 2). Chapter 2 demonstrated an increase in biofilm
abundance after spawning and <515N values indicated that this response is likely due to
salmon. The results presented here suggest that the community shift from an autotrophic
(algae) dominated biofilm to a heterotrophic (bacterial) one has a functional significance
for stream ecosystems opposite to pattern seen by Droppo et al. (2007). This shift would
help explain the long-term biological storage of MDNs.

3.4.1

Biofilm Component Patterns

A study by Yoder et al. (2006) is the only other known published attempt to characterize microbial biofilm communities at the functional component level. Yoder et al. (2006)
found increases in algae followed by increases in bacteria of biofilms during the post-

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

80

spawn period. The elevated microbial response in the presence of salmon nutrients as
identified by Yoder et al. (2006) is in contrast to the results presented in Figure 3.4. Yoder
et al. (2006) proposed that an initial increase in bacterial usage of MDNs. This initial colonization by bacteria provides subsequent algal colonization sites leading to a predominately autotrophic biofilm (Azam et al., 1983). The initial bacterial population bloom
facilitate algal colonization by taking advantage of carbon, nitrogen and organic nutrient
release from decaying carcasses (Yoder et al., 2006).
Results presented in Figure 3.4 do not demonstrate the same initial bacteria bloom.
This difference may be accounted for by several reasons. The pattern seen in Figure 3.4
may simply be natural variation. An alternative explanation is that because Yoder et al.
(2006) sampled on a monthly basis, in contrast to weekly sampling basis of this study,
Figure 3.4 may represent a finer resolution picture of biofilm component changes.
In the context of this study, however, observed changes in biofilm composition may
also reflect an important MDN sequestration pathway. Romani and Sabater (2000) report
that benthic bacterial populations use excreted algal decay product as a nutritive source.
As an algal bloom begins to decline, algal exudates become a nutrient within biofilms
for proximate (within the biofilm) bacteria growth (Kaplan and Bott, 1989). This pattern
manifests itself as a bacterial increase (October 27th) after a decrease in algae abundance
(After October 21*^). Figure 3.4 suggests this type of pattern whereby algal exudates,
themselves a product of MDN enrichment, provide microbial heterotrophic MDN storage
potential. This storage potential is supported by findings that autotrophically dominated
biofilm communities in aquatic systems are subject to high herbivory and decomposition
and thus accumulate less carbon (Cebrian and McClelland, 1998).
Should this pattern by replicated in future studies, it may have implications for winter biofilm storage of MDNs as microbial heterotrophically dominated communities are
less limited by photosynthetic activity and thus may retain MDNs over a smaller spatial
scale as this study saw a downstream bacterial increases over 20-m. This result, however,

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

81

would need to be replicated with a greater sample size and further into the post-spawn
period to confirm this hypothesis.

3.4.2

Nutrient Influence

The observed changes in biofilm composition presented in the above section (3.4.1) are
likely being driven by changing nutrient concentrations (Lawrence et al., 1998). Mohamed et al. (1998) found variable nutrient conditions, driven by pulp mill effluent, caused
changes in biofilm community composition. Neu et al. (2005) state that nutrients directly
affect "the nutritive value of the biofilm for grazers and play a role in water quality", linking changes in absorbed MDNs in biofilms to foodweb level impacts. For example, Wipfli
et al. (1999) found that this increase in biofilm quality resulted in higher salmonid densities. Thus, the ecosystem level impacts of biofilm enrichment appear to transfer directly
back up to salmon (See Figure 1.1). Mechanisms for this enrichment, however, are still not
well understood. I related the C:N ratio of salmon decay products to biofilm components
to determine how that ratio affects biofilm component coverage.
Biofilm growth followed the same general development pattern in response to nutrient levels outlined by Neu and Lawrence (1997), Romani and Sabater (2000) & Neu et al.
(2005). Neu et al. (2005) found that a nitrogen addition to biofilms sampled from the South
Saskatchewan River resulted in a decrease in bacterial numbers. Bacterial coverage from
slides sampled from the HFC (middle section) responded to higher O N ratio with increased coverage (Figure 3.7) suggesting a similar response as a lower C:N ratios results
in less bacterial coverage. This is consistent with the general aquatic science paradigm
that organic carbon limits bacterial productivity (Mohamed et al., 1998).
Relating biofilms sampled from the downstream section is complicated by the source
of nutrients. Biofilms in the downstream section were subjected to upstream nutrient
loads from decaying salmon in the middle section as well as decaying salmon from the
downstream section (Figure 3.5). This spatial variation may explain why the relation-

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

82

ship breaks down in the downstream section. A lack of a relationship in the downstream
section may indicate that advective losses were occurring in the middle section due to
biofilm growth. In addition, because there were no salmon present in the downstream
section late in post-spawn period, estimated nutrient contributions from decaying carcasses in the downstream section were zero, skewing already established biofilm cultures
against nutrient values that would predict low biofilm coverage.
Algal coverage was not related to the presence of salmon in either section suggesting
other influences on algal productivity (Figure 3.7). Neu et al. (2005) suggest that algal
abundance is more limited than bacteria by specific nutrient ratios. This, however, was
not a factor explicitly tested. It was not practical or desirable to isolate a single nutrient
from the MDN source. An alternative explanation is that algal abundance is limited by
solar radiation and that during the late fall sampling period, this was the dominating
factor driving algal productivity. This lack of increasing algal abundance is in contrast
to the results that are presented in Chapter 2 and are likely a consequence of the biofilm
developmental stage being examined (France, 1995).
Future studies should consider measuring the nutrient ratios present in the water and
of the biofilm being viewed under CLSM. If nutrients dictate biofilm composition, as
Neu et al. (2005) indicate, nutrient ratios may provide useful measurements of the nutrient processing capabilities of biofilms. Identification of biofilms, beyond the level of
algae and bacteria, to specific taxa (sensu Peterson and Grimm, 1992) will also likely be an
important future research focus. Peterson and Grimm (1992) found that different taxa of
algae responded differently to variable nutrient and grazer regimes. Moreover, the successional patterns of algal species after the redd construction disturbance will likely be
an important determinant of nutrient processing. This assessment of community composition at the level of taxa will be crucial for future microbiological studies of biofilms in
salmon streams.

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

3.4.3

83

Method Comparison

A stated goal of utilizing the CLSM was to compare conventional methods of biofilm
characterization to more sophisticated CLSM methods. Figure 3.7 is the result of this comparison across all experimental sections. Figure 3.8 illustrates the comparison between a)
both microscopically and spectrophotometrically measured algae and b) AFDM, a measure of total biomass, and total biological material, the sum of algae and bacterial percent coverage. Spectrophotometrically measured chlorophyll a and AFDM are commonly
used measures of biofilm biomass in freshwater systems (Steinman and Lamberti, 1996).
Neither comparison was statistically significant. Because of salmon activity, tiles had to
be redeployed in the post-spawn period. Thus, biofilms measured spectrophotometrically were at a different developmental stage than biofilm examined microscopically. The
difference in developmental stage explains a lack of a clear relationship between these
parameters (Figure 3.8). Furthermore, this result provide evidence that different developmental stages causes different biofilm composition which in turn causes different isotope
values (Staal et al., 2007) as seen in section 2.4.2. By using live/dead fluorescent probes,
Neu and Lawrence (1997) demonstrated that biofilm composition changes with development, although the exact nature of the change is in itself variable.
In terms of CLSM protocol, this result highlights the temporally sensitive nature of
sampling comparisons. The timing of sampling has an impact on the composition of samples. There is a clear need to have deployable CLSM slides that can withstand the rigours
of the entire salmon disturbance regime. Salmon spawning disturbs the composition of
streambed biofilms in an unknown way. Sampling these disturbed biofilms for CLSM
would improve our understanding of the recovery of biofilms after the active-spawn period, when the potential for MDN storage is high.

Chapter 3. Benthic Biofilm Composition in a Salmon Spawning Stream

3.5

84

Conclusions

Battin et al. (2007) noted the absence of a biofilm theoretical framework. Microbiology
often still views microorganisms as single-cultures studied in isolation. Battin et al. (2007)
proposed that biofilms should exist under the umbrella of landscape ecology. One of the
key components of landscape ecology is the need to identify the role of disturbance in
ecosystems (Urban et al., 1987). The work presented in this chapter addresses this need.
The findings presented here demonstrate biofilm community changes in response to an
ecological disturbance (salmon spawning) situated within an ecological framework that
explicitly acknowledges the role of sediment and nutrients on biofilm storage of MDNs.
Internal biofilm cycling, as evidenced by the increases in bacteria seen after a reduction
algal percent coverage (Figure 3.4), identifies a potential mechanism by which MDNs are
retained within a biofilm and in the stream rather than flushed immediately downstream.
This mechanism has been previously recognized in the biofilm literature (Neu et al., 2005).
This is, however, is the first known attempt to apply the same methodological approach
to aquatic systems that receive MDN pulses. These results highlight that biofilms have a
disproportionate impact on ecosystems in relation to their size and mass.

Chapter 4. Conclusions and Management

Implications

85

Chapter 4
Conclusions and Management Implications

4.1

Conclusions

This thesis demonstrates the importance of the salmon disturbance regime on benthic
biofilm communities. More generally, this thesis contributes to the broader literature on
the ecosystem effects of salmon spawning and die-off. The results presented in Chapter 2 and 3 suggest a biofilm trapping mechanism of nutrients and sediment driven by
MDNs. Chapter 3 provided an initial explanation of benthic biofilm composition and
storage patterns in response to MDNs. These mechanisms of MDN movement through
a lotic ecosystem provide an additional perspective in which to view the watershed level
benefits of salmon.
The findings in Chapter 2 support a growing body of literature that MDNs can be
delivered and retained over a small downstream scale and that the driving mechanism
behind this delivery and retention is flocculation. These results add a biofilm trapping
component to this floe delivery process. Moore et al. (2004) demonstrated that the impact
of salmon is vastly different depending on their status as a live spawner or decaying
carcass. The results presented in Chapter 2 suggest that the significant temporal overlap
between these two processes may be the most important determinant of the impact of
salmon.
These results also suggest an absence of this biofilm trapping ability during the activespawn period via a reduced biofilm layer. Increased sediment infiltration into the streambed
during the active-spawn disturbance and a subsequent decrease in infiltration during the
post-spawn fertilization period supports this trapping ability of biofilms (Chapter 2). This

Chapter 4. Conclusions and Management

Implications

86

relationship between biofilm growth and streambed sediment infiltration has not been
previously reported.
The significant post-spawn increase of biofilm abundance and isotopic signature 20m downstream of rotting carcasses further indicates that MDN flushing to downstream
rearing lakes is limited by in-stream flocculation processes. This abundance increase is an
example of the connection between salmon, a primarily marine organism, and freshwater
biofilms (Lamberti et al., 2010). Salmon and biofilm form the linkage point that connects
the marine and freshwater ecosystems. This is particularly true for interior Fraser River
stocks of salmon and has significant implications for rearing lake productivity.
Fraser River sockeye salmon exhibit four year cyclical patterns involving one high
return dominant year followed by a lower return sub-dominant year and two very low
return sub-dominant years. The mechanisms behind this cyclical pattern remain unclear
(Hume et al., 1996). In sub-dominant years (i.e. low salmon densities), the number of
returning adult salmon directly confers a size and survival benefit to the next generation of juvenile salmon. In a dominant year, however, when there is a large number of
spawners, the relationship between the number of adults and the next generation of juvenile salmon is not linear (Hyatt et al., 2004). There appears to be a threshold number
of returning spawners, above which little benefit is transfered to the next generation of
juvenile salmon (Hyatt et al., 2004). This is often vaguely attributed to habitat limitations
and carrying capacity.
Results presented in Chapters 2 and 3, however, has identified a link between salmon
spawning activity and streambed nutrient storage. At higher densities of salmon spawners, streambed nutrient storage increases the potential for the river to act as a nutrient
sink; therefore, fewer nutrients being transported downstream may limit salmon-driven
productivity increases in rearing lakes due to high spawning activity. This retardation of
nutrient transfer, via mechanisms outlined in this thesis, may result in a reduced MDN
input to the rearing lake, diminishing the capacity for a lake productivity boost from

Chapter 4. Conclusions and Management

Implications

87

salmon. With the river acting as a nutrient sink at high spawner densities, fewer nutrients would be transfered downstream, limiting the next generation of juvenile salmon
productivity. The timing of this nutrient transfer may be limited by river flow conditions
and suggests that the spring melt may be important for mobilizing stored nutrients. Thus,
MDN enrichment may become particularly important at low escapements as the rearing
habitat may still be operating below its carrying capacity. In this situation, MDNs would
then have a significant impact.
A recent attempt to characterize the importance of various nutrient sources at the watershed level by Wipfli and Baxter (2010) highlights the magnitude of the contribution that
MDNs make to the overall watershed nutrient budget. During the active-spawning period, Wipfli and Baxter (2010) estimate that approximately 50% of a stream's nutritional
content is contributed by salmon. Wipfli and Baxter (2010) identify the need to focus
"on the specific trophic processes and pathways that limit the productivity of riverine
food webs that sustain production of salmon during their freshwater phase". The biofilm
abundance increase presented in Chapter 2 presents the basal portion of one these key
trophic pathways. The salmon disturbance regime structures the key trophic pathways.
Wipfli et al. (2003) demonstrated this MDN mediated trophic pathway translated into
faster growth of resident fish. This finding highlights the feedback nature of salmon systems presented in Chapter 1 and helps assess the magnitude of the impact presented in
Chapter 2.
This substantial MDN contribution, however, may be an underestimate as Wipfli and
Baxter (2010) do not explicitly account for intergravel and sediment storage mechanisms
of MDNs. Intergravel and sediment storage of MDNs (via flocculation) are an important
part of the nutrient cycle within salmon streams (Chapter 2; McConnachie and Petticrew,
2006; Rex and Petticrew, 2008; Petticrew and Albers, 2010). MDN storage in stream sediments (as flocculated particles) are less likely to be flushed downstream as floes settle
out of the water column faster (Droppo, 2001; Rex and Petticrew, 2008). This storage po-

Chapter 4. Conclusions and Management

Implications

88

tential allows for the longer term release of nutrients temporally extending the impact of
salmon beyond that proposed by Wipfli and Baxter (2010). Whichever way that MDNs
are processed by watersheds, the ecological value of salmon to watersheds is significant
(Wipfli and Baxter, 2010).

4.2

Management Implications

Since the early eighties there has been a steady increase in the amount of sockeye returning to the Horsefly River (Figure 4.1). This increase peaked in 2001 with an escapement
in excess of 1.5 million (Figure 4.1). At that time, many fisherman throughout the Fraser
Basin raised concerns that spawning grounds had become saturated with salmon to the
point where it was having a negative impact on overall stock health. The phenomenon
was termed over-spawning 1 and was used to advocate for higher recruitment. This view
can be summarized in this way (From the House of Commons Standing Committee on
Fisheries and Oceans cited in Walters et al., 2004):
"There's a high correlation between over-escapement and poor return, particularly for sockeye. Every major over-escapement event since 1956 has resulted in a near-collapse in the Skeena, in Rivers Inlet, and in the Fraser River.
But our managers go on dumping more and more fish on the spawning grounds."
In 2010, after several years extremely low escapements, the Fraser River experienced the
largest sockeye return in nearly a century. Soon after the total size of the Fraser River sockeye run was realized, many media outlets reported on the fear of over-spawning. Several
groups, particularly those advocating for higher recruitment, proposed that fish escaping
fishery pressure would over crowd spawning streams, experience high pre-spawn mortality and generally be wasted if there were not caught in a fisherman's net. After the high
escapement in 2001 a report was commissioned by DFO (Walters et al., 2004) to examine
1

Also referred to as over-escapement

Chapter 4. Conclusions and Management

Implications

89

TO " „ ( ) E

,Tt
Year

Figure 4.1: Historical Horsefly River Escapement. Stock enhancement via the HFC may
have contributed to high stocks in the mid-nineties although other DFO management practices also take place within the Quesnel watershed (DFO, 2010).
if over-spawning causes stock collapses. This report found that "there is no evidence that
high spawning runs place stocks at risk of collapse" (pp. 5 Walters et al., 2004).
Given the important cultural and economic legacy of salmon as a resource it is obviously crucial to strike a balance between fishery and conservation values. Traditional as
well as commercial fisheries are important components of the British Columbian landscape. The concept of over-spawning, however, attempts to frame the argument in the
absence of ecological values and rather portrays decaying salmon as a "waste". The results presented in this thesis demonstrate an increase in the basal portion of the foodweb
that is directly due to salmon die-off. This material is recycled through the ecosystem
and provides considerable ecological value (Wipfli et al., 2010). Categorizing this ecologically valuable material as "waste" overcompensates in the direction of the fishery and is
indicative of an outdated single species management regime.

Bibliography

90

Bibliography
AGRAWAL, Y. C , AND H. C. POTTSMITH. 2000. Instruments for particle size and settling
velocity observations in sediment transport. Marine Geology 168: 89-114.

APHA. 1995. Standard methods for the examination of water and wastewater, 18th
edition. American Public Health Association.
A Z A M , F., T. FENCHEL, J. FIELD, J. GRAY, L. MEYER-REIL, AND F. THINGSTAD. 1983.

The ecological role of water-column microbes in the sea. Marine Ecology Progress Series.
10: 257-263.
BASTVIKEN, D., L. PERSSON, G. O D H A M , AND L. TRANVIK. 2004. Degradation of dis-

solved organic matter in oxic and anoxic lake water. Limnology and Oceanography 49:
109-116.
BATTIN, T. J. 2000. Hydrodynamics is a major determinant of streambed biofilm activity:
From the sediment to the reach scale. Limnology and Oceanography 45: 1308-1319.
BATTIN, T. J., L. A. K A P L A N , J. D. NEWBOLD, X. H. C H E N G , AND C. M. H A N S E N .

2003a. Effects of current velocity on the nascent architecture of stream microbial
biofilms. Applied and Environmental Microbiology 69: 5443-5452, doi:10.1128/AEM.
69.9.5443-5452.2003.
BATTIN, T. J., L. A. K A P L A N , J. D. NEWBOLD, AND C. M. H A N S E N . 2003b. Contributions of microbial biofilms to ecosystem processes in stream mesocosms. Nature 426:
439-442.
BATTIN, T. J., W. T. S L O A N , S. KJELLEBERG, H. DAIMS, I. M. H E A D , T. P. CURTIS, AND
L. EBERL. 2007. Microbial landscapes: New paths to biofilm research. Nature Reviews

Microbiology 5: 76-81.
BIGGS, B. J., N. C. T U C H M A N ,

R. L. LOWE, AND R. J. STEVENSON. 1999. Resource
stress alters hydrological disturbance effects in a stream periphyton community. Oikos
85: 95-108.
BILBY, R. E., B. R. FRANSEN, AND P. A. BiSSON. 1996. Incorporation of nitrogen and

carbon from spawning coho salmon into the trophic system of small streams: Evidence
from stable isotopes. Canadian Journal of Fisheries and Aquatic Sciences 53: 164-173.
CAK, A. D., D. T. CHALONER, AND G. A. LAMBERTI. 2008. Effects of spawning salmon

on dissolved nutrients and epilithon in coupled stream-estuary systems of southeastern
Alaska. Aquatic Sciences 70: 169-178, doi:10.1007/s00027-008-8090-5.

Bibliography

91

CEBRIAN, W., AND V. M C C L E L L A N D . 1998. The dependence of heterotrophic consump-

tion and C accumulation on autotrophic nutrient content in ecosystems. Ecology Letters
1: 165-170.
CEDERHOLM, C. J., D. B. H O U S T O N , D. L. COLE, AND W. J. SCARLETT. 1989. Fate of
coho salmon (Oncorhynchus kisutch) carcasses in spawning streams. Canadian Journal of
Fisheries and Aquatic Sciences 46: 1347-1355.
CHALONER, D. T., K. M. MARTIN, M. S. WIPFLI, P. H. OSTROM, AND G. A. LAMBERTI.

2002. Marine carbon and nitrogen in southeastern Alaska stream food webs: Evidence
from artificial and natural streams. Canadian Journal of Fisheries and Aquatic Science
59: 1257-1265.
CHALONER, D. T., G. A. LAMBERTI, R. W. MERRITT, N. L. MITCHELL, P. H. OSTROM,
AND M. S. WIPFLI. 2004. Variation in responses to spawning Pacific salmon among three

south-eastern Alaska streams. Freshwater Biology 49: 587-599.
CHALONER, D. T., G. A. LAMBERTI, A. D. CAK, N. L. BLAIR, AND R. T. EDWARDS.

2007. Inter-annual variation in responses of water chemistry and epilithon to Pacific
salmon spawners in an Alaskan stream. Freshwater Biology 52: 478-490.
CLAESON, S. M., J. L. Li, J. E. C O M P T O N , AND P. A. BISSON. 2006.

Response of nutrients, biofilm and benthic insects to salmon carcass addition. Canadian Journal of
Fisheries and Aquatic Sciences 63: 1230-1241.
CONOVER, R. J. 1966. Assimilation of organic matter by zooplankton.

Limnology and

Oceanography 11: 338-345.
C O P E , R. S., AND J. S. M A C D O N A L D . 1996. Responses of sockeye salmon (Oncorhynchus

nerka) embryos to intragravel incubation environments in selected streams within the
Stuart-Takla watershed, pp. 1-4. In Conf. Proc, Land Management Practices Affecting
Aquatic Ecosystems. M.K. Brewin and D.N.A. Monita (technical co-ordinators).
COSTERTON, J. W., Z. LEWANDOWSKI, D. E. CALDWELL, D. R. KORBER, AND H. M.
LAPPINSCOTT. 1995. Microbial biofilms. Annual Review of Microbiology 49: 711-745.
CRAWLEY, M. J. 2007. The R book. John Wiley & Sons Inc.
CROSS, W. F., J. P. BENSTEAD, P. C. FROST, AND S. A. THOMAS. 2005. Ecological stoichiometry in freshwater benthic systems: Recent progress and perspectives. Freshwater
Biology 50: 1895-1912.
DEANGELIS, D. L., P. J. M U L H O L L A N D , J. W. ELWOOD, A. V. PALUMBO, AND A. D.
STEINMAN. 1990. Biogeochemical cycling constraints on stream ecosystem recovery.

Environmental Management 14: 685-697.
DRAKE, D. C , R. J. N A I M A N , AND J. M. HELFIELD. 2002. Reconstructing salmon abundance in rivers: An initial dendrochronological evaluation. Ecology 83: 2971-2977.

Bibliography

92

DROPPO, I. G. 2001. Rethinking what constitutes suspended sediment. Hydrological
Processes 15: 1551-1564.
D R O P P O , I. G., G. G. LEPPARD, D. T. FLANNIGAN, AND S. N. LISS. 1997. The freshwa-

ter floe: A functional relationship of water and organic and inorganic floe constituents
affecting suspended sediment properties. Water, Air, & Soil Pollution 99: 43-53.
D R O P P O , I. G., N. Ross, M. SKAFEL, AND S. N. LISS. 2007. Biostabilization of cohesive

sediment beds in a freshwater wave-dominated environment. Limnology and Oceanography 52: 577-589.
ELFF, M. 2010. memisc: Tools for Management of Survey Data, Graphics, Programming, Statistics, and Simulation, h t t p : / / C R A N . R - p r o j e c t . o r g / p a c k a g e = m e m i s c .
R package version 0.95-30.

Fox, J. 1997. Applied regression analysis, linear models, and related methods. Sage
Publications, Inc.
FRANCE, R. 1995. Differentiation between littoral and pelagic food webs in lakes using

stable carbon isotopes. Limnology and Oceanography 40: 1310-1313.
GlBERT, J., AND L. DEHARVENG. 2002. Subterranean ecosystems: A truncated functional
biodiversity. BioScience 52: 473-481.
G R A H A M , A. A., D. J. M C C A U G H A N , AND F. S. M C K E E . 1988. Measurement of surface
area of stones. Hydrobiologia 157: 85-87.
H A S S A N , M.,

A.

GOTTESFELD, D.

MONTGOMERY, J. TUNNICLIFFE, G.

CLARKE,

G. W Y N N , H. JONES-COX, R. POIRIER, E. M A C I S A A C , H. HERUNTER, ET AL. 2008.
Salmon-driven bed load transport and bed morphology in mountain streams. Geophysical Research Letters 35: L04405.
HlLDERBRAND, G., S. FARLEY, C. ROBBINS, T. HANLEY, K. TlTUS, AND C. SERVHEEN.
1996. Use of stable isotopes to determine diets of living and extinct bears. Canadian
Journal of Zoology 74: 2080-2088.
H O C K I N G , M. D., AND T. E. REIMCHEN. 2002. Salmon-derived nitrogen in terrestrial

invertebrates from coniferous forests of the Pacific Northwest. BMC Ecology 2:4. h t t p :
//www.biomedcentral.com/1472-67 85/2/4.
HOOD, E., J. FELLMAN, AND R. T. EDWARDS. 2007. Salmon influences on dissolved organic matter in a coastal temperate brown-water stream: An application of fluorescence
spectroscopy. Limnology and Oceanography 52: 1580-1587.
HORN, H., AND D. C. HEMPEL. 1997. Growth and decay in an auto-/heterotrophic
biofilm. Water Research 31: 2243-2252.
HULSMAN, H., AND L. WUBBEN. 2008. Salmon spawning and sediment dynamics in
the Horsefly River in British Columbia. Master's thesis, Universiteit Utrecht.

Bibliography

93

H U M E , J., K. SHORTREED, AND K. MORTON. 1996. Juvenile sockeye rearing capacity

of three lakes in the Fraser River system. Canadian Journal of Fisheries and Aquatic
Sciences 53: 719-733.
H U N T , A. P., AND J. D. PERRY. 1999. Algal-bacterial relationships in streams with varying DOC concentration and source, pp. 34-42. In C. W. Keevil, A. Godfree, D. Holt, and
C. Dow [eds.], Biofilms in the Aquatic Environment. The Royal Society of Chemistry.
HYATT, K., D. M C Q U E E N ,

K. SHORTREED, AND D. RANKIN. 2004. Sockeye salmon
(Oncorhynchus nerka) nursery lake fertilization: Review and summary of results. Environmental Reviews 12: 133-162.
JANETSKI, D. J., D. T. CHALONER, S. D. TIEGS, AND G. A. LAMBERTI. 2009.

Pacific
salmon effects on stream ecosystems: A quantitative synthesis. Oecologia 159: 583-595.

N., J. M A C D O N A L D , K. H A L L , AND P. J. TSCHAPLINSKI. 1997. A preliminary study of the role of sockeye salmon (Oncorhynchus nerka) carcasses as carbon
and nitrogen sources for benthic insects and fishes in the "Early Stuart" stock spawning
streams, 1050 km from the ocean. Fisheries Project Report No. RD55, British Columbia
Ministry of Environment, Lands and Parks, Victoria, BC h t t p : / /www. f or . gov. be .
ca/hfd/library/fisheries/FPRRD55.pdf.
JOHNSTON,

JOHNSTON, N. T., E. A. M A C I S A A C , P. J. TSCHAPLINSKI, AND K. J. H A L L . 2004. Effects

of the abundance of spawning sockeye salmon (Oncorhynchus nerka) on nutrients and
algal biomass in forested streams. Canadian Journal of Fisheries and Aquatic Sciences
61: 384-403.
KAPLAN, L., AND T. BOTT. 1989. Diel fluctuations in bacterial activity on streambed
substrata during vernal algal blooms: Effects of temperature, water chemistry, and habitat. Limnology and Oceanography 34: 718-733.
KLINE, T. C , J. J. GOERING, O. A. MATHISEN, P. H. P O E , AND P. L. PARKER. 1990.

Re-

cycling of elements transported upstream by runs of Pacific salmon: I. (515N and <513C evidence in Sashin Creek, southeastern Alaska. Canadian Journal of Fisheries and Aquatic
Sciences 47: 136-144.
KLINE, T. C , J. J. GOERING, O. A. MATHIESEN, P. H. P O E , P. L. PARKER, AND R. S.

SCALAN. 1993. Recycling of elements transported upstream by runs of Pacific salmon:
II. 515N and 513C evidence in the Kvichak River watershed, Bristol Bay, Southwestern
Alaska. Canadian Journal of Fisheries and Aquatic Sciences 50: 2350-2365.
KONDOLF, G. M. 2000. Assessing salmonid spawning gravel quality. Transactions of
the American Fisheries Society 129: 262-281.
G. A., D. T. CHALONER, AND A. E. HERSHEY. 2010. Linkages among
aquatic ecosystems. Journal of the North American Benthological Society 29: 245-263.
LAMBERTI,

LAWRENCE, D. F. 2004. The feasibility of conducting a watershed-based fish sustainability plan on the Horsefly River watershed. Master's thesis, Royal Roads University.

Bibliography

94

LAWRENCE, J., AND T. N E U . 1999. Confocal laser scanning microscopy for analysis of

microbial biofilms. Methods in Enzymology 310: 131-144.
LAWRENCE, J., D. KORBER, B. HOYLE, J. COSTERTON, AND D. CALDWELL. 1991. Opti-

cal sectioning of microbial biofilms. Journal of Bacteriology 173: 6558-6567.
LAWRENCE, J., T. N E U , AND G. SWERHONE. 1998. Application of multiple parameter
imaging for the quantification of algal, bacterial and exopolymer components of microbial biofilms. Journal of Microbiological Methods 32: 253-261.
M A N Z , W. 1999. Phylogenetic composition, spatial structure, and dynamics of lotic
bacterial biofilms investigated by fluorescent in situ hybridization and confocal laser
scanning microscopy. Microbial Ecology 37: 225-237.
M C C O N N A C H I E , J. L., AND E. L. PETTICREW. 2006. Tracing organic matter sources in

riverine suspended sediment: Implications for fine sediment transfers. Geomorphology
79: 13-26.
MCINTYRE, P., A. FLECKER, M. VANNI, J. H O O D , B. TAYLOR, AND S. THOMAS.

2008. Fish distributions and nutrient cycling in streams: Can fish create biogeochemical hotspots? Ecology 89: 2335-2346.
MILLS, K., AND J. BEVER. 1998. Maintenance of diversity within plant communities:
Soil pathogens as agents of negative feedback. Ecology 79: 1595-1601.

MlNAKAWA, N., AND R. I. GARA. 1999. Ecological effects of a chum salmon (Oncorhynchus keta) spawning run in a small stream of the Pacific Northwest. Journal of
Freshwater Ecology 14: 327-335.
MlNAKAWA, N., AND R. I. GARA. 2003. Effects of chum salmon redd excavation on
benthic communities in a stream in the Pacific Northwest. Transactions of the American
Fisheries Society 132: 598-604.
MITCHELL, M. 2010. Engauge Digitizer - Digitizing software, h t t p : / / d i g i t i z e r .
sourceforge.net/.
MITCHELL, N. L., AND G. A. LAMBERTI. 2005. Responses in dissolved nutrients and

epilithon abundance to spawning salmon in southeast Alaska streams. Limnology and
Oceanography 50: 217-227.
MOERKE, A. H., AND G. A. LAMBERTI. 2003. Responses in fish community structure to
restoration of two Indiana streams. North American Journal of Fisheries Management
23: 748-759.
M O H A M E D , M., J. LAWRENCE, AND R. ROBARTS. 1998. Phosphorus limitation of heterotrophic biofilms from the Fraser River, British Columbia, and the effect of pulp mill
effluent. Microbial Ecology 36: 121-130.

Bibliography

95

MONTGOMERY, D. R., J. M. BUFFINGTON, N. P. PETERSON, D. SCHUETT-HAMES, AND

T. P. QuiNN. 1996. Stream-bed scour, egg burial depths, and the influence of salmonid
spawning on bed surface mobility and embryo survival. Canadian Journal of Fisheries
and Aquatic Sciences 53: 1061-1070.
MOORE, J. W. 2006. Animal ecosystem engineers in streams. BioScience 56: 237-246.
M O O R E , J. W., AND D. E. SCHINDLER. 2008. Biotic disturbance and benthic community
dynamics in salmon-bearing streams. Journal of Animal Ecology 77: 275-284.
M O O R E , J. W., D. E. SCHINDLER, AND M. D. SCHEUERELL. 2004. Disturbance of fresh-

water habitats by anadromous salmon in Alaska. Oecologia 139: 298-308.
R. J., R. E. BILBY, D. E. SCHINDLER, AND J. M. HELFIELD. 2002. Pacific
salmon, nutrients, and the dynamics of freshwater and riparian ecosystems. Ecosystems
5: 399-417.
NAIMAN,

N E U , T., AND J. LAWRENCE. 2004. One-photon versus two-photon laser scanning microscopy and digital image analysis of microbial biofilms. Methods in Microbiology 34:
89-136.
N E U , T., G. SWERHONE, AND J. LAWRENCE. 2001. Assessment of lectin-binding analysis

for in situ detection of glycoconjugates in biofilm systems. Microbiology 147: 299-313.
N E U , T., G. SWERHONE, U. BOCKELMANN, AND J. LAWRENCE. 2005. Effect of CNP on
composition and structure of lotic biofilms as detected with lectin-specific glycoconjugates. Aquatic Microbial Ecology 38: 283-294.

NEU, T. R. 2000. Confocal laser scanning microscopy (CLSM) of biofilms, chap. 13,
pp. 211-219. In H. Flemming, U. Szewzyk, and T. Griebe [eds.], Biofilms: Investigative
Methods & Applications. CRC Press.
N E U , T. R., AND J. R. LAWRENCE. 1997. Development and structure of microbial
biofilms in river water studied by confocal laser scanning microscopy. FEMS Microbiology Ecology 24: 11-25.

NOWLIN, W. H., M. J. VANNI, AND L. H. YANG. 2008. Comparing resource pulses in
aquatic and terrestrial ecosystems. Ecology 89: 647-659.
PAULING, B., AND I. WAGNER-DOBLER. 2006. Stream microcosm for investigating GEM
impact on the indigenous bacterial community in river water and sediment. Process
Biochemistry 41: 2129-2137.
PAULY, D., V. CHRISTENSEN, J. DALSGAARD, R. FROESE, AND F. TORRES JR. 1998. Fishing down marine food webs. Science 279: 860-863.

PAWLEY, J. [ed.]. 1995. Handbook of Biological Confocal Microscopy, 2nd ed. Springer.
PETERSON, B. J., AND B. FRY. 1987. Stable isotopes in ecosystem studies. Annual Review

of Ecology and Systematics 18: 293-320.

Bibliography

96

PETERSON, C. G. 1996.

Algal ecology: Freshwater benthic ecosystems, chap. 13: Response of benthic algal communities to natural physical disturbance, pp. 375-402. Academic Press.
PETERSON, C. G., AND N. B. GRIMM. 1992. Temporal variation in enrichment effects
during periphyton succession in a nitrogen-limited desert stream ecosystem. Journal of
the North American Benthological Society 11: 20-36.

PETERSON, D., AND C. FOOTE. 2000. Disturbance of small-stream habitat by spawning
sockeye salmon in Alaska. Transactions of the American Fisheries Society 129: 924-934.
PETTICREW, E., AND J. REX. 2006. The importance of temporal changes in gravel-stored
fine sediment on habitat conditions in a salmon spawning stream. International Association of Hydrological Sciences Publication 306: 434-441.
PETTICREW, E. L., AND S. J. ALBERS. 2010. Salmon as biogeomorphic agents: Temporal and spatial effects on sediment quantity and quality in a northern British Columbia
spawning channel. International Association of Hydrological Sciences Publication 337:
251-257.
PETTICREW, E. L., AND J. M. AROCENA. 2003. Organic matter composition of gravelstored sediments from salmon-bearing streams. Hydrobiologia 494: 17-24.

POWER, M. 1990. Resource enhancement by indirect effects of grazers: Armored catfish,
algae, and sediment. Ecology 71: 897-904.
G., AND M. J. KEOUGH. 2002. Experimental Design and Data Analysis for
Biologists. Cambridge University Press.
QUINN,

QUINN, T. 2005. The behavior and ecology of Pacific salmon and trout. American Fisheries Society.
R 2.11.1, R. D. C. T. 2010. R: A Language and Environment for Statistical Computing.
R Foundation for Statistical Computing, Vienna, Austria, h t t p : //www. R-pro j e c t .
org. ISBN 3-900051-07-0.
RASBAND, W. S., AND IMAGEJ. 2009. ImageJ, 1.43b. National Institutes of Health,
Bethesda, MD, USA. h t t p : / / r s b w e b . n i h . g o v / i j / .
REIMCHEN, T. 2000. Some ecological and evolutionary aspects of bear-salmon interactions in coastal British Columbia. Canadian Journal of Zoology 78: 448-457.
REX, J., AND E. PETTICREW. 2006. Pacific salmon and sediment flocculation: Nutrient
cycling and intergravel habitat quality. International Association of Hydrological Sciences Publication 306: 1-8.
REX, J., AND E. PETTICREW. 2008. Delivery of marine-derived nutrients to streambeds
by Pacific Salmon. Nature Geoscience 1: 840-843.

Bibliography

97

REX, J. F. 2009. The flocculation feedback loop: Delivery of marine derived nutrients in
Pacific salmon streams. Ph.D. thesis, University of Northern British Columbia.
R O M A N I , A., A N D S. SABATER. 2000. Influence of algal biomass on extracellular enzyme

activity in river biofilms. Microbial Ecology 40: 16-24.
R O M A N I , A., H . G U A S C H , I. M U N O Z , J. R U A N A , E. V I L A L T A , T. S C H W A R T Z , F. E M T I A Z I , A N D S. S A B A T E R . 2004.

Biofilm structure and function and possible implications
for riverine DOC dynamics. Microbial Ecology 47: 316-328.
SABATER, S., H . G U A S C H , A. R O M A N I , A N D I. M U N O Z . 2002.

The effect of biological
factors on the efficiency of river biofilms in improving water quality. Hydrobiologia 469:
149-156.
SABATER, S., H. G U A S C H , I. M U N O Z , A N D A. R O M A N I . 2006. Hydrology, light and the

use of organic and inorganic materials as structuring factors of biological communities
in Mediterranean streams. Limnetica 25: 335-347.
SARKAR, D. 2008. Lattice: Multivariate Data Visualization with R. Springer,
/ / l m d v r . r - f o r g e . r - p r o j e c t . o r g . ISBN 978-0-387-75968-5.

http:

SCHEUERELL, M., P. LEVIN, R. ZABEL, J. WILLIAMS, AND B. SANDERSON. 2005. A
new perspective on the importance of marine-derived nutrients to threatened stocks of
Pacific salmon (Oncorhynchus spp.). Canadian Journal of Fisheries and Aquatic Sciences
62: 961-964.
SCHINDLER, D., S. CARPENTER, J. C O L E , J. KlTCHELL, AND M. PACE. 1997. Influence
of food web structure on carbon exchange between lakes and the atmosphere. Science
277: 248-251.
SCHINDLER, D. E., M. D. SCHEUERELL, J. W. MOORE, S. M. GENDE, T. FRANCIS, AND
W. P A L E N . 2003. Pacific salmon and the ecology of coastal ecosystems. Frontiers in
Ecology and the Environment 1: 31-37.
SlEGEL, S. 1957. Nonparametric statistics. American Statistician 11: 13-19.
STAAL,

M.,

R.

THAR,

M.

KUHL,

M.

C.

M.

VAN

LOOSDRECHT,

G.

WOLF,

J. D E B R O U W E R , A N D J. RIJSTENBIL. 2007. Different carbon isotope fractionation patterns during the development of phototrophic freshwater and marine biofilms. Biogeosciences 4: 613-626.
S T E I N M A N , A. D., A N D G. A. L A M B E R T I . 1996. Biomass a n d pigments of benthic algae,

chap. 17, p p . 295-313. In R. Hauer and G. A. Lamberti [eds.], Methods in Stream Ecology.
Academic Press.
S T O O D L E Y , P., I. D O D D S , J. D. B O Y L E , A N D H. M. L A P P I N - S C O T T .

1999. Influence of
hydrodynamics and nutrients on biofilm structure. Journal of Applied Microbiology 85:
19-28.

Bibliography

98

STOODLEY, P., K. SAUER, D. DAVIES, AND J. COSTERTON. 2002.

Biofilms as complex
differentiated communities. Annual Reviews in Microbiology 56: 187-209.
STRATHMANN, M., J. W I N G E N D E R , AND H. FLEMMING. 2002. Application of fluores-

cently labelled lectins for the visualization and biochemical characterization of polysaccharides in biofilms of Pseudomonas aeruginosa. Journal of Microbiological Methods 50:
237-248.
SUTHERLAND, I. W. 2001. The biofilm matrix-An immobilized but dynamic microbial

environments. Trends in Microbiology 9: 222-227.
TIEGS, S., D. CHALONER, P. LEVI, J. RUEGG, J. TANK, AND G. LAMBERTI. 2008. Timber
harvest transforms ecological roles of salmon in southeast Alaska rain forest streams.
Ecological Applications: A publication of the Ecological Society of America 18: 4-11.
TIEGS, S. D., E. Y. CAMPBELL, P. S. LEVI, J. RUEGG, M. E. BENBOW, D. T. CHALONER,
R. W. MERRITT, J. L. TANK, AND G. A. LAMBERTI. 2009. Separating physical distur-

bance and nutrient enrichment caused by Pacific salmon in stream ecosystems. Freshwater Biology 54: 1864-1875, doi:{10.1111/j.l365-2427.2009.02232.x}.
TOEWS, D. A. A., AND M. J. BROWNLEE. 1981. A handbook for fish habitat protection
on forest lands in British Columbia, Special Publication ed. Department of Fisheries and
Oceans, Government of Canada.
URBAN, D., R. O ' N E I L L , AND H. SHUGART. 1987. Landscape Ecology. BioScience 37:

119-127.
VANNOTE, R. L., G. W. MINSHALL, K. W. C U M M I N S , J. R. SEDELL, AND C. CUSHING.

1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences
37: 130-137.
WALTERS, C. J., P. L E B L O N D , AND B. RIDELL. 2004. Does over-escapement cause
salmon stock collapse? Technical paper. Pacific Fisheries Resource Conservation
Council. h t t p : //www . f i s h .be . c a / f i l e s / O v e r e s c a p e m e n t R e p o r t _ 2 0 0 4_0_
Complete.pdf.
W H I T T I N G H A M , M. J., P. A. STEPHENS, R. B. BRADBURY, AND R. P. FRECKLETON. 2006.
Why do we still use stepwise modelling in ecology and behaviour? Journal of Animal
Ecology 75: 1182-1189.

WlCKHAM, H. 2009. ggplot2: elegant graphics for data analysis. Springer New York.
http://had.co.nz/ggplot2/book.
WILLIAMS, N. D., D. E. WALLING, AND G. J. L. LEEKS. 2007. High temporal resolution

in situ measurement of the effective particle size characteristics of fluvial suspended
sediment. Water Research 41: 1081-1093.

Bibliography

99

WINEMILLER, K., A. FLECKER, AND D. HOEINGHAUS. 2010. Patch dynamics and environmental heterogeneity in lotic ecosystems. Journal of the North American Benthological Society 29: 84-99.
WINGENDER, J., T. N E U , AND H. FLEMMING. 1999. What are bacterial extracellular
polymeric substances?, Microbial Extracellular Polymeric Substances. Springer.
WIPFLI, M., AND C. BAXTER. 2010. Linking ecosystems, food webs, and fish production:

Subsidies in salmonid watersheds. Fisheries 35: 373-387.
WIPFLI, M. S., J. H U D S O N , AND J. CAOUETTE. 1998. Influence of salmon carcasses on

stream productivity: Response of biofilm and benthic macroinvertebrates in southeastern Alaska, USA. Canadian Journal of Fisheries and Aquatic Sciences 55: 1503-1511.
WIPFLI, M. S., J. P. H U D S O N , D. T. CHALONER, AND J. P. CAOUETTE. 1999. Influence of

salmon spawner densities on stream productivity in southeast Alaska. Canadian Journal
of Fisheries and Aquatic Sciences 56: 1600-1611.
WIPFLI, M. S., J. P. H U D S O N , J. P. CAOUETTE, AND D. T. CHALONER. 2003. Marine sub-

sidies in freshwater ecosystems: salmon carcasses increase the growth rates of streamresident salmonids. Transactions of the American Fisheries Society 132: 371-381.
WIPFLI, M. S., J. P. H U D S O N , J. P. CAOUETTE, N. L. MITCHELL, J. L. LESSARD, R. A.
H E I N T Z , AND D. T. CHALONER. 2010. Salmon carcasses increase stream productivity

more than inorganic fertilizer pellets: A test on multiple trophic levels in streamside
experimental channels. Transactions of the American Fisheries Society 139: 824-839.
WOLFAARDT, G., J. LAWRENCE, AND D. KORBER. 1999. Function of EPS, chap. 10, pp.

171-200. In Microbial Extracellular Polymeric Substances: Characterization, Structure
and Function. Springer.
YODER, D. M., A. VIRAMONTES, L. L. KIRK, AND L. F. H A N N E . 2006. Impact of salmon

spawning on microbial communities in a northern California river. Journal of Freshwater
Ecology 21: 147-155.

Appendix

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

100

Appendix A
Confocal Laser Scanning Microscopy (CLSM) software scripts
A.l

Image J conversion of OIB files to stacked TIFF

//ver
3 7e
1/7/2010
/ / W r i t t e n by Glen MacDonald, Core for Communication Research
/ / I m a g i n g and Microscopy Core of
V i r g i n i a M e r r i l l Bloedel Hearing Research Center and the C e l l u l a r Morphology
for Human Development and D i s a b i l i t y
/ / B o x 357923
/ / U n i v e r s i t y of Washington
/ / S e a t t l e , WA 98195-7923
//glenmac@u Washington edu
/ / R e q u i r e s the L O C I . t o o l s p l u g i n
/ / On running the f i r s t macro, s e l e c t a f i l e , the d i a l o g window d i s p l a y s number of c h a n n e l s and b i t depth
/ / S e l e c t merge o p t i o n s of "RGB", "Composite H y p e r s t a c k " or "No Merge"
//
S e l e c t a c h a n n e l (0 —n) for each c o l o r channel
Error t r a p p i n g
If you choose a
/ / c h a n n e l not p r e s e n t , the macro e x i t s with a message t h a t the c h a n n e l is not p r e s e n t
/ / P l a c i n g any c h a n n e l i n t o 2 c o l o r s w i l l c r e a t e an RGB merge
/ / A n y channel may be a s s i g n e d to more than 1 c o l o r
/ / T h i s v e r s i o n w i l l g e n e r a t e a p r o j e c t i o n image based on c h a n n e l s in the i n i t i a l s e l e c t e d f i l e
/ / " S a v e as RGB TIFF" w i l l save s t a c k s and p r o j e c t i o n s ( i f s e l e c t e d , t o o ) as they are c r e a t e d and c l o s e them
/ / Batch mode w i l l check for number of c h a n n e l s and z—steps
If a f i l e has a d i f f e r e n t number of c h a n n e l s from
s e l e c t e d image,
/ / i t s filename is w r i t t e n to the log window, but not opened, s i n c e the merge will be off
/ / I f p r o j e c t i o n s are s e l e c t e d in batchmode, but a f i l e with the same number of c h a n n e l s is e n c o u n t e r e d t h a t is
/ / i t will be opened, b u t i t s name w i l l be w r i t t e n to the log window i n s t e a d of a l l o w i n g ImageJ to choke making

Core of the

the

Center

initially

not a z—series ,
a p r o j e c t i o n of

var name, o r i g f o r m a t , pType , mergelD , dir , ext , c o u n t , path , 1 , rede , blue , grnc , gryc , bsizeC , c r e a t e , c r e a t e t y p e , i c s e x t ,name, prefix , b s i z e Z ,
/ / g l o b a l s with i n i t i a l
var r e d - " 2 " ,
var g r n - " l " ,
var b l u = " 0 " ,
var gry="None",
var merge="RGB",
var mip="None",
var b a t c h = 0 ,
var s d i r - " " ,
var d e s t d i r = ",
var saveMe=false ,
var closeMe = f a l s e ,
var fformat = "RGB",

defaults

macro "Import Channel Order [ 1 ] " {
requires ( 1 42d '),
run{"Bio—Foimats
Macro Extensions
'),
path = F lie openDialog (" Select
a file')
,//
path+filenaine
//
redirectErrorMessagesO,
dir = File get Pai ent (path ) + /"J/path
to
file
naine=File get Name ( path ) , / / get
filename
naniel =
lertgthOf(name),
basename=Flie
nameWithout Extension ,
if (name-=b a sen a me )
exit (" This file
name doeb not have an extension
to indicate
tif ,
jpeg ',
tcs , etc ') ,
basenamel = lengthOf(
basenante),
e\t = $ubstnng
(name , basenamel + 1 .naniel),
Ext getFormat (path ,
format)
origformat=format,
Ext
setld(path),
Ext getSizeC ( sizeC ) , / / deliver s the number of
channels
Ext
getSizeZ(sizeZ),
Ext
getSizeT(sizeT),
chinned)Array {s izeC +1),
for (} =l,j<stzeC + l,]++)
cH[j]=d2s(j
-1,0),
cH[Q] = '*None* ,
Ext getPixelType
(pixelType
),
pclType = pixelType ,
matchPixelTypc
(pType) ,

file

if (MzeZ>l)
zplancb = '\nThis
elbe
zplanes = "
if
(sizeT>l)
tplaneb='\nThis
else

is a Z—series

with

is a timelapse

series

+^tzeZ + ' image

with

planes

'+sizeT + ' time

points

type

\nPleabe

rename

with

the

appropi

tate

extension

E g

it

Appendix

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

tplane* = ' ,
msg=" Reading image file "+name+" \ nAssign
"+sizeC+" channels
to merged colors \nBit
Depth is
Dialog create ("Channel
Merge Options
3 7e '),
Dialog addChoice(" Merge Type ", newArrayi 'RGB" , 'Composite',
'No merge"'),
merge),
Dialog addMesbage (msg+zplanes + t planes ) ,
Dialog addMebsage (" Red merged channel
") ,
Dialog addChoice("Channels
",
chi.red),
Dialog addMessage (" Green merged channel
') ,
Dialog addChotce( ' Channels
',
cH,grn),
Dialog addMessage ( 'Blue merged channel
') ,
Dialog addChotcef
Channels
',
cH.blu),
Dialog addMessage ("Gray merged channel
') ,
Dialog addChoice (' Channels
',
cH,gry),
Dialog addChoice (' Create Projection
Image ',new Array ( 'None , 'Average
Intensity",
Max Intensity
Standard
Deviation ', 'Median")
,mip),
Dialog addCheckbox ( ' Save opened images ', saveMe),
Dialog addChoice ("Saved image format
', newArrayi 'TIFF—RGB ', "ICS—Compostte ')
.fformat),
Dialog addCheckboxf" Process all images in directory
',batch),
Dialog addCheckbox(" Close file
on save " , closeMe),
Dialog show () ,
merge=Dialog
getChotce () ,
red=Dialog
getChotce () ,
grn = Dialog getChotce () ,
bin-Dialog
getCltoice () ,
gry = Dialog getChotce () ,
mtp=Dialog getChotce 0 ,
fformat = Dialog getChotce 0 ,
saveMe-Dialog
getCheckbox () ,
batch = Dialog getCheckbox () ,
c!oseMe= Dialog getCheckbox () ,
if (closeMe==true&&saveMe==fa
Ise )

'+pType + " per

, "Mm

exit ("There
is no point m closing
files
after
opening unless you select
to actually
save them1
(red ' = "*None*'&&gm
'= "*None*"&&bltt '= '*Nonc'&&gry
'= "*None*'&&merge== RGB ')
exit ("RGB cannot merge more than 3 channels
Switch to 'Composite'
to merge 4 channels
"),
'f (8]y '- "*None*' &&merge=="RGB )
exit ("Gray is only a valid option
with
Composite'
meige
) ,
if (mtpi = 'None
&&stzeZ==l)
exit('Tltis
/s not a z—series , projections
cannot be selected
) ,
format = substring
(fformat ,0 , tndexOf (fformat,'
— ")),
tcsext =" tcs
cHs = newArray (4) ,//contains
the vars from choices
1 2 3,4,
'None*
cHs[0l=red,
cHsll]=grn,
cHs[2]=blu,
cHs[31=gry,
cHsc=new Array ( Red', Green ','Blue
' 'day'),
//labels
for chosen
channels
if
(saveMe==l)
destdir
= getDirectory
("Choose Destination
Directory
),
if (merger'RGB')
{
create = ' ',
if (batch ' = 1)
doMerge () ,
else
batchAll () ,
}

pixel

Intensity

101

",

'

'Sum

Slices

) ,

if

if (merge--"Composite") {
create create
,
if (batch' = V
doMerge 0 ,
else
balcltAllO
,
}
if ( m e r g e - - " N o merge") {
if (batch' = 1)
noMergeO ,
else
batchAll () ,

>

}
f u n c t i o n doMerge () {
setBatchMode(true),
cHsm=newAiray(4)
, //airange
image labels
in order of merger
for ( } =0, ]<cHs length /++){ //cHs sets channel
import order for
r,g,b,gy
if (cHslj i'= *None* ){
cHbo=(cHs[ j j) + 2, // convert
channel
numbering for impoi t
opttons='open-[
+-path + 'f
an toscale
specify .range
s pi it .channels
oiew=[ Standai d Image} J stack .order^Default
c.end=cHso
c_btep=l
,
showStattis ( opening file
),
run ( ' Bio—Formatb Importer ', options ) ,
rename (chisel j }) , //name for color LUT
cH^m[} l=cHscl j ],
1/ make a new array for the merging with channel LUT names and *None* in oide
}

else{
cHsml j j="*None*
}

',

>

c r e a t e t y p e = " r e d = ["+cHsm[Q] + "] green = ["+cHsrn[l ] + "] blue =["+cHsm[2] + '] gray = ["+cHsm[3]+ "] " + c r e a t e ,
run("Merge Channels
", createtype),
p r e f i x = s u b s t n n g (name,0 , indexOf (name, ' ") ) ,
rename ( prefix+"_merge ' ) ,
file-getTitle,

c_begin=cHso

Appendix

filelD=get!mageID,
if (mip' = "None")
doProjechon (mip),
if ( s a v e M e - - l ) {
selectlmage
(filelD)
,
if
(forinat=="lCSfl)
run (" Bio—Formats Exporter ", "save=["+destdir+ftle
else
save As (format , destdir+fi
le ) ,
if
(closeMe==tnte)
close () ;
}
setBatchMode(" e x i t and d i s p l a y " ) ,

161
163
165
167
169
171
173
175
177
179
181

183
185
187
189
191
193
195

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

+ icsext

+"]"),

>
function noMerge () {
setBatchMode
(true),
for ( }=0, y<cHs length, j ++){ I / cHs selects
channels
to import
and order for
r,g,b,gy
if (cHsl}
}>='*None*'){
cH$o = (cHs[j }) +1; //convert
channel
numbering for
import
op t ion s-" open =["+ path +" ] a u toscale
specify .range
spl it -channels
view=[ Standard
Image} } stack -order = Default
c.end=cHso
c.step =1 ",
showStattisi"
opening file
"),
run (" Bio—Formats Importer ", opt tons ) ,
prefix = bitbstnng(name,0,
indexOf(name,
" ") ) ,
renamef preftx+"—Ch"+cHs[ j }) , //name for color LUT
file=getTttle
,
filelD=get!magelD,
if (mtp' = "None")
doProjection
(imp),
if (saveMe = = l){
select Image (fit el D ) ,
if
(format=="lCS")
run (" Bio—Formats Exporter",
" save =l"+destdir+file
+
icsext+"]"),
else
save As (format , des td ir+fi le) ,

>

197

if

199

(closeMe—true)
c l o s e () ,

>

201

setBatchMode (" e x i t and d i s p l a y " ) ,
203 }
function b a t c h A l l ( ) {
205
list = getFileList
(dir),
for (t=0,i<ltst
length , i++){
207
if ( File isDirectory(dn+hst[i])
' = !){
if (endbWithdtbt
lil,
ext)){
209
Ext setld (dtr+list
I i]) ,
Ext
getSizeC(bbizeC),
211
Evf
getStzeZ(bsizeZ),
;/
(bt,izeC==stzeC){
213
pat!i = dir+hst
[ i \,
name=list { i } ,
!
215
if (mcrge = "No meige")
doMerge () ,
217
else
noMeige () ,
219
} / / i f bsizec
else
221
p r i n t ( d i r + l i s t [ i ]+ " c o n t a i n s " + b s i z e C + " c h a n n e l s " ) ,
} / / i f endswith
223
else
, / / s k i p f o l d e r s and o t h e r f i l e t y p e s and loop to n e x t f i l e
225
} / / i f f i l e is
}//for
227
}//fxn
229
231
233
235
237
239
241
243
245

102

f u n c t i o n match Pi xelType ( p e l T y p e ) {
if ( matcites ( pixel Type , ' tttntS "))
pType = 'unsigned
8—bits ",
if ( matches ( ptxelType ," itintl6
'))
pType = 'unsigned
16—bits ',
/ / (matches ( pixelType ," uint32 "))
pType - "unsigned
32— bits",
if ( matches ( pi x el Type , " tnt S '))
pType = "signed
8-bitb",
if ( matches ( pixel fypc ," mil 6 "))
pType = "signed
16—bits",
if ( ma tcheb( pixelType , ' int32 "))
pType = "signed
32—bits",
if ( matches (pixelType
."float
"))
pType - "floating
point
32—bits",
if ( matches ( pixelType
."double"))
pType = "64— bits double pi ecibion ',

247 }
249 macro "Composite Hyperstack to RGB [ 2 ] " {
run (" Channels Tool
"),
251 run ("Stack
to RGB"),
}

c.begin=cHso

Appendix

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

macro " B r i g h t e s t P o i n t P r o j e c t i o n [ 3 ] " {
noslices = nSltces ,
if ( nos lices -=1)
exit ("This
is not a stack
"),
run("Z Project
", " project ion =[Max

103

Intensity}"),

>
f u n c t i o n d o P r o j e c t i o n (mip){
if
(bsizeZ==l)
prtnt(dir+file+
" is not a z— stack
"),
else
{
mergeID=getlmageID
() ,
merge=getTttle
,
merge! = lengthOf(merge)
,
end=substring
(imp ,0,3) ,
if (endsWith (merge ," -merge ")= = 1)
merge=substrtng
(file ,0 , name! —6),
if
(end=="Ave")
end = "AVG",
if
(end=="Max")
end = "MAX";
if (end-="Min ")
end = "_MIN",
if
(end=="Sum")
end = " SUM",
if (end=="Sta ")
end = ".STD",
if
(end=="Med")
end='MED",
run("Z Project
" , "projection
-["+mip + "}");
rename
(merge+end),
if (saveMe= = l){
saveAb( ' tiff " , des td ir + merge + end),
if
(c!oseMe==l)
close () ,

>
}
}
macro "Commands Help [ 4 ] " {
Dialog create ("Commands Help "),
Dialog addMessage( '' Import Channel Order [11' — Opens dialog windozo with options
for file
merger \n' Composite
Hyperstack
to RGB
[2]' — Couveits
an active
image with 16— bits / channel
into an RGB file
with 8—bits / channel
Useful to convert
images for
Photobhop or PowerPoint
\n' Brightest
Point Pi ojection
[3}' — Creates a maximum intensity
projection
from the active
window
\n'Close
All Imageb [7J' — Closes al! open image windows (without
saving anything)
\nSee
the Impoit Channel Order Macro
documentation
for complete
information
\ nSelecting
a file
with the wrong filename
extension
or a misspelled
extension
will
result
in a long Java error message at the Log Window
"),
Dialog show () ,

>
macro "Channels Help [ 5 ] " {
Dialog create ("Channel
Help "),
Dialog addMessage (' Merge Type \n'RGB'
meiges stacks
at 8—bits per channel
\n' Composite
merges stacks
at >S bits per channel,
and
letams
more metadata \n'No Merge' opens each channel as a separate
stack \nFV—1000 channel
order is 0 [shortest
luavelength
] to n [longest
wavelength ]
Eg
0=bltte , l=gi een , 2 - red , 3= far i ed \n
Fewer channels
change the numbei nig e g 2
channels
0=green , l = red or 0=blue and l=red
\n Limit which image channels
are opened by 'No Merge' by assigning
to a color
, even though they will be opened in grayscale
\n").
Dialog show () ,
}
macro "Check Box Help [ 6 ] " {
Dialog create (' Checkbox
Help"),
Dialog addMesbage ("These options
are applied
when opening files
through
the Import Channel Oi der macro \n' Ct eate Projection
Image '
— Select
a projection
method
This rcq uireb a Z—s tack '\n' Save opened imageb ' —\n
Images and projections
are saved to
the destination
directory
\u'Saved
image format'
— \n
RGB—TIFF converts
images to 8— bits per channel
(RGB) and saves as
a TIFF \n
1CS—Composite saves Composite
merged images at native
bit depth,
biich as 16—bits per channel
\n' Process
all
images in directory'
—\nApplies
your settings
to all data files
in the directory
that have the same filename
extension
ALL
FILES MUST HAVE THE SAME NUMBER OF CHANNELS' \n 'Close on save'Each image is closed
after
it is saved
"),
Dialog show () ,

>
macro "Close All Images [ 7 ] " {
for (o-D,0< nlmages , o ++)
close () ,

>

A.2

Scion Image macros for the analysis of stacked TIFF files

macro ' l e c t i n

macro [ 1 ] '

{lectin

processing
image}

image
dilated

macro makes

it

so the

average

is

the

avei age

of the

pixels

that

are

being

counted

in an eroded

and

Appendix

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

{a macro to do our typical
image analysis
wlticli is for each slice
in a z series
certain
value and the mean pixel
value of the image above the threshold
{setting
the threshold
value for the stack and starting
the macro}
var
Thrvalue integer,
Addvalue i n t e g e r ,
1 integer,
n,mean,mode,min,max real ,
npx i n t e g e r ,
mn r e a l ,
nsl i n t e g e r ,
sin i n t e g e r ,
begin
{ THRESHOLD VALUE FOR LECTIN IMAGES }
T h r v a l u e =GetNumber( ' T h r e s h o l d value for s e r i e s is ' , 2 5 0 ) ,
Addvalue =255 — T h r v a l u e ,
nsl = n S l i c e s ,
sin = 0 ,
Measure,
r U s e r l [ r C o u n t ] =11111 1 1 1 1 1 ,
r U s e r 2 [ r C o u n t ] =111 1 1 1 1 1 ,
UpdateResults,
begin
if n s h c e s = 0 then begin
P u t M e s s a g e ( ' T h i s window is not a s t a c k ' ) ,
exit,
end,
for I =1 to n S h c e s do begin
S e l e c t S h c e (1 ) ,
{ ( I n s ib where you put in what to do for the
slice}
{ ***************** n e w
s/uff*»*****}
SelectAll ,
Copy,
MakeNewWindow( 'temp ' ) ,
SetNewSize(512,512),
Paste,
Se tTh res hold ( T h r v a l u e ) ,
MakeBinary,
Dilate,
Erode,
Copy,
Dispose ,
ScaleMathf f a l s e ) ,
Paste,
Add,

get the number
value}

of white

thresholded

pixels

104

above a

AddConstant(Addvalue) ,
{Getb mean and puts it in user! }
SetUser2Label( 'mean') ,
Measure,
GetRe s u i t s (n, mean, mode, m m , max) ,
mn -Mean,
{tin e shot ding to 255 and making binary s u b r o u 11 n t }
SetThreshold(255),
MakeBinary,
{ Count White Pixels
subroutine
Counts the number
zvlute pixels
in the
current
selection
and stores
the counts in the Userl columns
}
SetUser1Label( ' W h i t e ' ) ,
r U s e r l [ r C o u n t ] =262144 - h i s t o g r a m [ 2 5 5 ] ,
rUser2 [ rCount ] -mn,
UpdateResults,
s i n - s I n +1,
{cowmenf out the next three lines
if you dont want the file
closed at the end}
if s l n - n s l then begin
Dispose ,
end,
end,
end,
end,
end,
end,
macro ' b a c t e r i a

(green

pixels)[b]'

{a macro to determine
the area of bacteria
colored gieen in our usual
dilates}
{setting
the threshold
value for the stack and starting
the macro}
var
Thrvalue i n t e g e r ,
Addvalue i n t e g e r ,
l integer,
n , mean, mode, mm , max r e a l ,
npx i n t e g e r ,
mn real ,
nsl i n t e g e r ,
sin integer,
begin { THRESHOLD VALUE FOR BACTERIA HERE }
T h r v a l u e -GetNumber( ' T h r e s h o l d value for s e r i e s is ' ,200) ,
Addvalue = 2 5 5 - T h r v a l u e ,
nsl = n S l i c e s ,
sin = 0 ,

image

sequence

stained

with

SYTO 9 this

macro erodes

and

Appendix A. Confocal Laser Scanning Microscopy (CLSM) software scripts

105

Measure,
rUserl [rCount] -11111 11111,
r U s e r 2 [ r C o u n t ] =111 1 1 1 1 1 ,
UpdateResults,
begin
if n s h c e s = 0 then begin
P u t M e s s a g e ( ' T h i s window is not a s t a c k ' ) ,
exit,
end,
for i =1 to n S l i c e s do begin
SelectShce (l),
{ ( I n s is where you put in what to do for the
slice}
{thresholding
to thrvalue
and making binary
subroutine}
SetThreshold (Thrvalue) ,
MakeBinary,
Dilate ,
Erode,
Measure,
{ Count White Pixels
subroutine
Counts the number
zohite pixels
in the
current
selection
and stores
the counts in the Userl columns
}
SetUserlLabel( 'White') ,
r U s e r l [ r C o u n t ] -262144 - h i s t o g r a m [ 2 5 5 ] ,
rUser2[rCount] =0,
UpdateResults,
sin = s l n + l ,
{comment out the next three lines
if you dont want the file
closed at the end}
if sln = nsl then begin
Dispose,
end,
end,
end,
end,
end,
end,
macro ' a l g a e

(far

red p i x e l s ) [a ] '

{a macro to determine
the area of bacteria
colored
green in our usual image sequence
dilates}
{setting
the threshold
value for the stack and starting
the macro}
var
Thrvalue i n t e g e r ,
Addvalue i n t e g e r ,
l integer,
n , mean, mode, mm, max r e a l ,
npx i n t e g e r ,
mn r e a l ,
nsl i n t e g e r ,
sin i n t e g e r ,
begin { THRESHOLD VALUE FOR ALGAE HERE }
T h r v a l u e =GetNumber( ' Threshold value for s e r i e s is ' , 2 1 0 ) ,
Addvalue -255 — T h r v a l u e ,
nsl - n S l i c e s ,
sin = 0 ,
Measure,
r U s e r l [ r C o u n t ] =11111 H i l l ,
r U s e r 2 [ r C o u n t ] =111 1 1 1 1 1 ,
UpdateResults ,
begin
if n s l i c e s = 0 then begin
PutMessage( "I his window is not a s t a c k ' ) ,
exit,
end,
for i =1 to n S l i c e s do begin
SelectSIice(i),
{this
is 'where you put in what to do for the
slice}
{thresholding
to thrvalue
and making binary
subroutine}
SetThreshold(Thrvalue),
MakeBinary,
Dilate,
Erode,
Measure,
{ Count White Pixels
subi online
Counts the number
white pixels
in the
tuirent
selection
and stores
the counts in the Uberl columns
}
SetUserlLabel('White'),
r U s e r l [rCount] =262144 - h i s t o g r a m [ 2 5 5 ] ,
rUser2[rCount] =0,
UpdateResults,
sin = s 1 n + 1 ,
{comment out the next three lines
if you dont want the file
closed at the end}
if sln = nsl then begin
Dispose ,
end,
end,
end,
end,
end,
end,
macro ' cyanos (red p i x e l s ) [c ] '

stained

with

SYTO 9 this

macro erodes

and

Appendix

190
192
194
196
198
200
202
204
206
208
210
212
214
216
218
220
222
224
226
228
230
232
234

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

{a macro to determine
the area of bacteria
colored green m our usual image sequence
dilates}
{setting
the threshold
value for the stack and starting
the macro}
var
Thrvalue integer,
Addvalue i n t e g e r ,
1 integer,
n,mean,mode,mm,max real,
npx i n t e g e r ,
mn real ,
nsl integer ,
sin i n t e g e r ,
begin { THRESHOLD VALUE FOR ALGAE HERE }
T h r v a l u e =GetNumber( ' T h r e s h o l d value for s e r i e s is ' , 2 1 0 ) ,
Addvalue =255 — T h r v a l u e ,
nsl = n S h c e s ,
sin =0,
Measure,
r U s e r l [ r C o u n t ] = 1 1 1 1 1 11111,
r U s e r 2 [ r C o u n t ] =111 1 1 1 1 1 ,
UpdateResults ,
begin
if n s l i c e s = 0 then begin
P u t M e s s a g e ( ' T h i s window is not a s t a c k ' ) ,
exit,
end,
for i =1 to n S l i c e s do b e g i n
S e l e c t S l i c e (1 ) ,
{this
is where you put in what to do for the
slice}
{thresholding
to thrvalue
and making binary
subroutine}
SetThreshold (Thrvalue) ,
MakeBinary,
Dilate,
Erode,
Measure,
{ Count White Pixels
subroutine
Counts the number
white pixels
in the
current
select ion and stores
the counts in the Userl columns
}
Se t U s e r l Label ( ' W h i t e ' ) ,
r U s e r l [ r C o u n t ] =262144 - h i s t o g r a m [ 2 5 5 ] ,
rUser2 [ r C o u n t ] = 0 ,
UpdateResults,
sin - s i n + 1 ,
{comment out the next three lines
if you dont want the file
closed at the end}
if sln = nsl then begin
Dispose,
end,
end,
end,
end,
end,
end,

stained

with

SYTO 9 this

that

being

counted

106

macro erodes

and

236
238 macro 'polymer macro [ p ] '
{lectin

image
dilated

processing
image}

macto

makes

it

so the

average

is

the

average

of

the

{a macro to do our typical
image analysts
which is for each '-lice in a z series
certain
value and the mean pixel
value of the image above the threshold
the threshold
value for the stack and starting
the macro}

242 {setting
var
244
246
248
250
252 begin
254
256
258
260

Thrvalue integer,
Addvalue i n t e g e r ,
I integer,
n, mean, mode, mm, max r e a l ,
npx i n t e g e r ,
mn real ,
nsl i n t e g e r ,
sin i n t e g e r ,
{ FHRESHOLD VALUE FOR POLYMER IMAGES }
T h r v a l u e =GetNumber( ' Threshold value for s e r i e s
Addvalue = 2 5 5 - T h r v a l u e ,
nsl = n S h c e s ,
sin = 0 ,
Measure,
r U s e r l [ r C o u n t ] =11111 1 1 1 1 1 ,
r U s e r 2 [ r C o u n t ] =111 1 1 1 1 1 ,
UpdateResults,

is

begin
262
264

if n s l i c e s - 0 then begin
Put Mess age ( ' T h i s window is not a s t a c k ' ) ,
exi t ,

end,
for i =1 to n S l i c e s do begin
SelectShce(i),
268
{this
is whete you put in what
{ **» **************„ t . r y stuff *******}
270 S e l e c t A H ,
Copy,
272 MakeNewWindow( 'temp ' ) ,
SetNewSize(512,512),

266

to do for

the

slice}

',245),

pixels

get the
value}

are

number

of white

in an eroded

thresholded

and

pixels

above a

Appendix

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

107

274

Paste,
SetThres h o l d ( T h r v a l u e ) ,
276 MakeBinary,
Dilate,
278 Erode,
Copy,
280 Dispose
ScaleMath( f a l s e ) ,
282 P a s t e ,
Add,
284 { , . . , . , ****»********,**.***\
AddConstantf Addvalue) ,
286 {Gets mean and puts it in user! }
SetUser2Label( ' m e a n ' ) ,
288
Measure,
Get R e s u l t s ( n , mean, mode, nun, m a x ) ,
mn =Mean,
290
292 {threshc
294
296
298
300
302
304
306
308
310

Iding
to 255 and making binary
subroutine}
SetThreshold(255),
MakeBinary,
{ Count White Pixels
subroutine
Counts the number
white pixels
in the
current
selectIO n and stores
the counts
in the Userl columns
}
SetUserlLabel('White'),
r U s e r l [ r C o u n t ] =262144 - h i s t o g r a m [ 2 5 5 ] ,
rUser2 [ rCount ] =mn.
UpdateResults ,
sin = s l n + 1 ,
{commen
out the next three lines
if you dont want the file
closed at the end}
if s t n = n s l then b e g i n
Dispose,
end,
end,
end,
end,
end,
end,

A.3

R script for data processing

# S c r i p t to pi ocebS raw confocal
image files
that have been run using
14, 2010
#CSV files
can be exported
from any spreadsheet
program
# csv inputted
needed to have the following
headei
format
# Sample Frame
rep Total Count User
itOtherwibe
one just
needs to change the names
pro <—read csv ( " f i l e n a m e and f i l e p a t h csv' , h e a d e r = l"RUE, s e p - ' , " )
#TIus is i e moving the sepai at or that the macros adiled
p r o < - p r o [ ! (pro$Count = = l 1111 11) ,]

the

Ceoige

Swerhone

macros

Written

by Sam Albers

on March

#These

aren t the actual
channel names but it makes it easier
to sort if we Hunk of the channel
name as what we aie trying
out of the channel
Order matters
here so if your data ib organized
differently
move the labels
around
accordingly
p r o $ c h a n n e l = c ( " c y a n o " / ' b a c t e r i a " ," a l g a e " )
pro$macro=c( " b a c t e r i a " , ' b a c t e r i a " , " b a c t e r i a " , ' a l g a e " , " a l g a e " , "a I g a e " , " c y a n o " , " c y a n o " ," c y a n o " )

to

^Organizes
the data so that the correct
channel
is
each image so we initially
btart
with 3 tunes
p r o <— p r o [ p r o $ c h a n n e l - = p r o $ m a c r o , ]

on

being selected
agaiitbt
the correct
macio
the amount of data
Thib command removes

Each macios
this
pail

was run

three

times

get

#Collapse
data by the fiame
establishes
mean white count for
frames
pro mean <— w i t h ( p r o , a g g r e g a t e (pro , b y = T i s t ( r e p , Macro=macro) , mean))
# Conditional
statement
that adds
pro m e a n $ s t a i n < — w i t h ( p r o mean,
sytolec ', " s y t o l e c " ) ) ) ) )

the correct
stain
to the dataframe
May need to modified
if the staining
regime wab d iffe rent
i f el se ( rep== ' 1 " , " n o n e " , i f el se ( r e p = = " 2 " , " s y t o " , 1 f el se ( r e p = = " 3 " , " sy t o l e c " , i f el se ( r e p = = " 4 " , "

############################################################################
#Thts

ib a bit of a i oundabout
process
to subtiact
the
the algae coverage
which only shows up in the far
p r o mean <— pro mean[ o r d e r ( p r o mean$rep) , ]

cyanobacteria
red

cooerage

pro mean$algaecyano <— u n l i s t ( l a p p l y ( s p l it ( p r o mean, p r o m e a n $ r e p ) ,
f u n c t i o n ( x ) x $ c a h b 2 <— x$Count— x[x$Macro == " c y a n o " ,
pro mean <— pro mean[ o r d e r ( p r o mean$Macro) , ]

which

'Count"])

shows

up in

the

red and fat

red channel

from

)

################################################################
#Syto

Control
Subtract
the control
from the othei
values
The assumption
here is that
representative
of the fluorescence
in the entire
sample
pro mean$sy t o c o n t r o l <— u n l i s t ( l a p p l y ( s p l i t ( pro mean, pro mean$Mai_ro) ,
f u n c t i o n ( x ) x $ c a l i b <— x$Count— x [ x $ s t a i n == " n o n e " ,
"Count"]) )

the fl uroescence

in

the

control

is

# This gathers
the data we need from the above subtraction
and combines
it into a column called
percentcov
p r o meanSpercentcov <— ( c ( p r o mean[pro mean$Macro==" a l g a e " , " a l g a e c y a n o " ] , pro mean[pro mean$Macro==" b a c t e r i a " , " s y t o c o n t r o l " ] ,
p r o mean[pro mean$Macro=="cyano" , "Count" ] ) / p r o m e a n $ T o t a l ) * 1 00

Appendix

A. Confocal Laser Scanning Microscopy (CLSM) software scripts

############################################################
#Just a trim here
fust
to keep only what we need
pro mean <— s u b s e t ( p r o mean, s e l e c t = c(Macro, r e p , p e r c e n t c o v ,
^Comment out section
labels
as needed
pro mean$section="down"
#pro mean$sectton=
'inid
#pro
mean$section="up'
####File can then be exported
write c s v ( p r o mean, " f i l e n a m e

to a
csv")

csv

file

which

stain))

can be opened

#################################################################
pro mean
#################################################################

by most

spreadsheet

software

packages

108