Embed Size (px)
Citation preview
Abiotic Conditions in Contrasting Environments: an
Examination of Precambrian Shield Lotic Communities
by
Margaret Rose Neff
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Ecology and Evolutionary Biology University of Toronto
© Copyright by Margaret R. Neff 2011
ii
Abiotic Conditions in Contrasting Environments: an Examination
of Precambrian Shield Lotic Communities
Margaret R. Neff
Doctor of Philosophy
Ecology and Evolutionary Biology University of Toronto
2011
Abstract
The inherent complexity of the natural world has long been a central theme in ecological
research, as the patterns and processes that govern ecosystems can operate at multiple spatial and
temporal scales. It is clear that to develop general ecological frameworks, we must consider
many different factors at different scales, and incorporate ideas from other disciplines. This
thesis touches on several of these ideas, first through an analysis of literature, and then with field
research examining the role of broad-scale abiotic factors on lotic systems. To determine how
integrated aquatic science is currently understood among different researchers, I provide an
analysis on communication and exchange of ideas among various subfields in aquatic science. I
show that there are clear divisions within the aquatic science literature, suggesting that there is
progress to be made on the integration of methods and ideas. Next, I examine the impact of a
large-scale geological feature, the Canadian Precambrian Shield, on abiotic conditions in lotic
systems, and how these conditions in turn influence the species assemblages of aquatic
organisms. This is addressed with both historical survey data, as well as contemporary data, and
as a whole, incorporates ideas concerning the relative influence of regional versus local factors,
the importance of historical factors on species distributions, and the relationship between the
abiotic environment and biological communities. These analyses show that there are distinct fish
iii
and macroinvertebrate communities in Shield lotic systems compared to those found in nearby
off-Shield sites, indicating that the Shield is an important broad-scale factor influencing local
biological communities. This finding, in conjunction with previous knowledge on the influence
of historical factors, provides further insight on the structuring of lotic fish and
macroinvertebrate communities in Ontario.
iv
Acknowledgments
I would like to thank my PhD committee, Drs. Doug Currie, Nigel Lester, Nick Mandrak and
Keith Somers for their insight, criticism and ideas throughout this whole process. Your
knowledge and accomplishments will continue to be a goal to strive for, throughout the rest of
my career. In particular I would like to acknowledge Don Jackson, my supervisor – I will be
forever grateful that you took a chance on me. I have learned so much from you, about ecology,
statistics, science, natural history, politics, and history. Your advice has always been thoughtful
and sound, and it has been a pleasure getting to know you. Thank you for your time,
encouragement and genuine interest in my life.
Thank you to all the Jackson lab members, past and present: Karen Alofs, Andrew Drake, Brie
Edwards, Monica Granados, Jaewoo Kim, Mark Poos, Sapna Sharma, Les Stanfield, Meg St.
John, Angela Strecker, Steve Walker, and Lifei Wang. In particular I would like to thank Steve,
for his patience and willingness to answer my numerous questions; for Mark and Andrew, for
offering their expert advice on field sampling; for Brie, for being an awesome conference
companion and a great listener; and for Monica, for always trying to help me whenever I found
myself completely confused or overwhelmed. In addition, I would like to acknowledge all of the
“honorary” lab members and/or other graduate students and post-docs who were always
available to chat: Maria Bennell, Christine Ngai, Anna Price, Caren Scott, Anna Simonsen,
Bronwyn Rayfield, Crystal Vincent and Cameron Weadick.
I would also like to thank all of my field and lab assistants – it is no exaggeration to say that
without their help, none of this would have been possible: Jesse Elders, Julie Ellis, Vernon
Lewis, Chris Lusczeck, Ariel Nesbitt, and Tommy Tam. Particular thanks go to Alex Manning,
who braved long hours of field sampling in one of the most beautiful places on earth.
Besides learning more than I even hoped to learn about ecology and science, I have come away
with so much more in the way of understanding myself, others and life in general. I would be
remiss if I did not acknowledge those people who were instrumental in this particular journey,
because without them, I’m not sure where I would be. First, thank you to the T.R.E.E., for
coming into my life at the exact right moment, and for being my family away from home. Thank
v
you to the University College Waterdragons – joining this team was one of those key, life-
changing moments, and I can’t even begin to describe the amount of inspiration that I gathered
from being a part of this group.
Thank you to my amazing, amazing family. Every single day I thank the universe for this
incredible support system. We are a diverse group of people, but we all share a deep-rooted
optimism and ability to find the humor (or “yumor”) in any situation. Here’s to living each day
like it’s “the best yet!” In addition, my parents, John and Debbie Neff, deserve special mention.
My greatest achievements will always be those where I make you both proud.
Last, but in no way least, thank you to the series of decisions or twists of fate that set me on the
path to my continuing association with Algonquin Park. Spending those summers at the family
cabin on Smoke Lake undoubtedly cemented my love of the natural world, and when I finally
realized that I wanted to be an ecologist, it all made sense. My experiences there have shaped
nearly all aspects of the person I am today. Thank you, thank you, thank you.
vi
Table of Contents
Acknowledgments.......................................................................................................................... iv
Table of Contents........................................................................................................................... vi
List of Tables .................................................................................................................................. x
Chapter II ................................................................................................................................... x
Chapter III .................................................................................................................................. x
Chapter IV.................................................................................................................................. x
Chapter V .................................................................................................................................. xi
Appendix A ............................................................................................................................... xi
Appendix B ............................................................................................................................... xi
Appendix C .............................................................................................................................. xii
List of Figures .............................................................................................................................. xiv
Chapter II ................................................................................................................................ xiv
Chapter III ................................................................................................................................ xv
Chapter IV............................................................................................................................... xvi
Chapter V ............................................................................................................................... xvii
Appendix A ........................................................................................................................... xviii
Appendix C ........................................................................................................................... xviii
List of Appendices ........................................................................................................................ xx
ChapterI: General Introduction ........................................................................................................1
Research Objectives ....................................................................................................................5
References ...................................................................................................................................5
ChapterI: Communication and cohesion in aquatic science literature...........................................14
Abstract .....................................................................................................................................14
Introduction ...............................................................................................................................14
vii
Methods.....................................................................................................................................16
Interaction among subfields of aquatic science .................................................................16
Patterns within aquatic science subfields...........................................................................17
Statistical Analysis ....................................................................................................................20
Results .......................................................................................................................................20
Interaction among subfields of aquatic ecology ................................................................20
Patterns within aquatic ecology subfields..........................................................................23
Discussion .................................................................................................................................32
Interaction among subfields of aquatic science .................................................................32
Patterns within aquatic science subfields...........................................................................34
Aquatic science today ........................................................................................................36
References .................................................................................................................................37
Chapter III: Effects of broad scale geological changes on patterns in macroinvertebrate assemblages...............................................................................................................................39
Abstract .....................................................................................................................................39
Introduction ...............................................................................................................................39
Methods.....................................................................................................................................42
Statistical Analysis ....................................................................................................................47
Results .......................................................................................................................................47
Data set A: regional community composition ...................................................................47
Data set B: community composition and abiotic environment..........................................48
Discussion .................................................................................................................................54
References .................................................................................................................................59
Chapter IV: Effects of a geological transitoin on stream fish communities: the influence of the Canadian Precambrian Shield .............................................................................................65
Abstract .....................................................................................................................................65
Introduction ...............................................................................................................................65
viii
Methods.....................................................................................................................................69
Statistical Analysis ....................................................................................................................76
Results .......................................................................................................................................77
Patterns in fish community composition............................................................................77
Patterns in abiotic conditions .............................................................................................85
Patterns in abiotic and biotic composition between Shield and off-Shield systems..........85
Discussion .................................................................................................................................87
References .................................................................................................................................92
Chapter V: Regional-scale patterns in community concordanc: the influence of the Precambrian Shield on the structure of lotic assemblages ........................................................98
Abstract .....................................................................................................................................98
Introduction ...............................................................................................................................98
Methods...................................................................................................................................103
Abiotic sampling..............................................................................................................105
Fish sampling ...................................................................................................................108
Macroinvertebrate sampling ............................................................................................108
Statistical Analysis ..................................................................................................................108
Results .....................................................................................................................................110
Patterns in abiotic conditions across Shield and off-Shield lotic systems .......................110
Patterns in fish communities............................................................................................112
Patterns in macroinvertebrate communities.....................................................................114
Abiotic and community associations ...............................................................................116
Community concordance and abiotic associations ..........................................................119
Discussion ...............................................................................................................................122
References ...............................................................................................................................128
Chapter VI: General Discussion ..................................................................................................135
ix
References ...............................................................................................................................138
Appendices...................................................................................................................................143
Appendix A .............................................................................................................................145
Appendix B .............................................................................................................................149
Appendix C .............................................................................................................................159
Copyright Acknowledgements.....................................................................................................175
x
List of Tables
Chapter II
Table 2-1. Summary of topics included in analysis by category, including a) research questions,
b) experimental design, c) level of study, d) study system, e) taxonomic group and f) methods of
analysis…………………………………………………………………………………………...19
Chapter III
Table 3-1. Compositional abiotic variables analyzed using correspondence analysis (CA) and
their interpretations based on ordination scores. The first 2 CA axes were used for substrate
composition (summarizing 39.1% of the variation), the first 4 were used for landscape
composition (72.4%), and the 1st axis was used for each of canopy cover (59.4%), gradient
(62.2%), and bank stability (40.5%)……………………………………………………………..45
Table 3-2. Results from linear discriminant analysis of data set B environmental data from
Shield and off-Shield sites. Variables with strong negative scores were associated with off-Shield
sites, whereas variables with strong positive scores were associated with Shield sites. Linear
discriminant analysis correctly classified 95.8% of Shield sites and 94% of off-Shield sites based
on their environmental conditions. See Table 3-1 for abbreviations of compositional abiotic
variables (S1, S2, L1, L2, L3, L4)……………………………………………………………...51
Chapter IV
Table 4-1. Results from Correspondence Analysis on each of the % composition variables from
the raw data: substrate composition, landscape composition, canopy cover and gradient. CA
scores for each site for each group of variables were used in further analyses to represent the
variation within these compositional variables. This table describes how values for each new
variable (S1, S2, L1, L2, Can and G) relate to the original data…………………………………73
Table 4-2. Five functional attributes and their respective categories used in this analysis.
Individual fish species were classified into a single category for each functional attribute.
xi
Numbers in brackets denote the number of species in the dataset that fell within each specified
category………………………………………………………………………………………….74
Table 4-3. Indicator Values (%) for each species from Indicator Species Analysis. Species are
listed alphabetically. Indicator values shown in bold are those which are >25% for either the off-
Shield or Shield group. Note that Creek Chub and Northern Redbelly Dace have indicator values
>25% for both the off-Shield and Shield group, and that most Shield indicator species also have
high values for off-Shield sites…………………………………………………………………..80
Chapter V
Table 5-1. Interpretation of compositional abiotic variables. Raw substrate and land-use
composition data at each site (see Table C-4, Appendix C) were summarized using
correspondence analysis into axes scores representing the dominant trends for each group of
variables. This table describes how values for each new variable (S1, S2, S3 and ALIS1, ALIS2)
relate to the original data……………………………………………………………………….107
Appendix A
Table A-1. Macroinvertebrate taxa of the subset of OMNR Aquatic Habitat Inventory data used
in Chapter 3. Taxa in brackets indicate historical nomenclature used in original data.……….141
Table A-2. Invertebrate taxa and associated abbreviations used in Chapter 3 ordination figures…………………………………………………………………….…………………….145
Appendix B
Table B-1. Summary of all fish species collected, the identification code (ID) used elsewhere in
the chapter and their % occurrence in both Shield and off-Shield systems…………………….149
Table B-2. Summary of all abiotic variables, with mean, median values, and standard deviations
of sites sampled from Shield and off-Shield systems………………………………………….151
Table B-3. Summary of fish functional traits (Goldstein and Meador 2003, Eakins 2010),
including a) temperature preferences, b) fish trophic status, c) fish geomorphology preferences,
d) substrate preferences, and e) reproductive behaviors………………………………………152
xii
Table B-4. Summary of ALIS1 and ALIS 2 summary variables, and how they are interpreted.
To assess the potential effect of land use on the patterns observed in fish communities in Shield
and off-Shield sites, the original landscape data, which was assessed up to 50 m from the
sampling reach at each site, were replaced with catchment land-use data using the Aquatic
Landscape Inventory Software (ALIS). Land use was calculated for each sampling site, where
the entire catchment area was divided among 28 land-use categories, so that a percentage of the
total catchment area was determined for each category. These data were then summarized into
two land-use variables using CA, in the same manner as substrate composition, gradient, and
canopy cover (Appendix A). This resulted in two land-use variables, ALIS1 and ALIS2, which
collectively summarize 40.3% of the variation in land use across all sites. Replacing the original
landscape variables (L1 and L2) with ALIS1 and ALIS2, we used CCA to examine patterns in
fish community composition with the abiotic variables. Using partial CCA, we were able to
determine the variation explained solely by ALIS1 and ALIS2 in comparison to the variation
explained by the other abiotic variables. ……………………………………………………….158
Table B-5. Summary of the variation explained by land use (ALIS1 and ALIS2), the remaining
abiotic variables, shared variation between land use and abiotic variables, and variation
unexplained by any of the included variables…………………………………………………..158
Appendix C
Table C-1. 2007-2008 sampling sites and associated bedrock geology. Sites in grey were
removed from subsequent analyses due to equipment malfunctions resulting in loss of water
chemistry data (sites 5-9) or absence of fish (sites 44 and 60)…………………………………159
Table C-2. Summary of all fish species collected, with % occurrence in both Shield and off-
Shield systems…………………………………………………………………………………..161
Table C-3. Summary of all macroinvertebrate families collected, with % occurrence in both
Shield and off-Shield systems…………………………………………………………………..162
Table C-4. Summary of all abiotic variables, with mean, standard deviation and median values
for both Shield and off-Shield systems…………………………………………………………164
xiii
Table C-5. Summary of fish functional traits (Goldstein and Meador 2003, Eakins 2010),
including a) temperature preferences, b) trophic status, c) geomorphology preferences, d)
substrate preferences, and e) reproductive behaviors…………………………………………..166
Table C-6. List of macroinvertebrate taxa and abbreviations used in ordination plots……….171
xiv
List of Figures
Chapter II
Fig. 2-1. Correspondence analysis for (a) 1982 (N=483) and (b) 2005 (N=768) citations for the
four journals, showing the association between articles and the source journal of their associated
citations. Articles located close together are more likely to include references to the same
journal(s). The journal coordinates form a triangular shape on the plot, representing the three-
dimensional space (ternary space) of the CA results. Articles found near a particular journal cite
predominately from that journal, whereas articles found in the center of the triangle cite papers
from all four journals…………………………………………………………………………….22
Fig. 2-2. Graphical representation of the degree of change between 1982 and 2005 in the
composition of citations for each journal. The change in position of a journal between years
provides a measure of the relative change in its composition of citations over that time period.
For example, the 2005 position for Freshwater Biology is positioned slightly below its 1982
indicating a reduction in the relative frequency of citations to itself and inclusion of more papers
from the other three journals……………………………………………………………………..26
Fig. 2-3. A correspondence analysis of all data categories (see Table 2-1) was conducted for all
papers published from the 5 journals in 2005. This figure and figs. 2-4 through 2-6 use the same
coordinate system to provide details regarding various research issues; therefore relative
positions of variables can be readily compared among the various figures. This figure shows the
coordinates for research topics and the journal centroids. The proximity of the research topics to
one another or to a journal centroid provides a summary of their association. Smaller distances
represent closer association and angles relative to the origin provide a measure of whether the
variables (categories) are positively or negatively associated. For example, studies of behavior
(BEH) and species composition (SC) are strongly, but negatively associated in their co-
occurrence in these journals……………………………………………………………………...27
Fig. 2-4. Subset of the 2005 CA showing coordinates for experimental design and level of study
with average coordinate scores for all articles in each journal. The scale matches that used in
xv
Figures 2-3 through 2-6 so associations between points in each figure can be made based on their
proximity to one another…………………………………………………………………………28
Fig. 2-5. Subset of the 2005 CA showing coordinates for taxonomic group(s) and study system
with average coordinate scores for all articles in each journal. The scale matches that used in
Figures 2-3 though 2-6 so associations between points in each figure can be made based on their
proximity to one another…………………………………………………………………………29
Fig. 2-6. Subset of the 2005 CA showing coordinates for method of analysis with average
coordinate scores for all articles in each journal. The scale matches that used in Figures 2-3
through 2-5 so associations between points in each figure can be made based on their proximity
to one another…………………………………………………………………………………….30
Fig. 2-7. Bar graphs indicating the overall percentages of regional (a) author location and (b)
study site location for the five journals. Study site location includes two additional regions: the
Atlantic, Pacific or other ocean(s), and the category “none” for studies which did not clearly state
the location……………………………………………………………………………………….31
Chapter III
Fig. 3-1. Map depicting bedrock geology and sampling sites in south-central Ontario. The heavy
dark lines depict the Precambrian Shield boundary in this area of the province………………46
Fig. 3-2. The ordination resulting from a Correspondence Analysis (CA) of a) 121
macroinvertebrate taxa and b) 125 study sites located either on or off the Precambrian Shield.
The first 2 CA axes explained 11% of the variability of invertebrate community composition.
See Table A-1 for taxon abbreviations. Labels are centered over the point for each taxon……..52
Fig. 3-3. a) Ordination from a Canonical Correspondence Analysis (CCA) of 109 invertebrate
taxa, 21 environmental variables, and 58 sites. Vector variables are: dissolved O2 (DO), velocity
(V), alkalinity (Alk), conductivity (Cond), turbidity (Turb), width, depth, discharge, organic
debris (OD), woody debris (WD), rocks, undercut banks (UCBank), landscape variables (L1, L2,
L3, and L4), substrate variables (S1, S2), gradient (G), canopy cover (Cov), and bank stability
(BS). Variability explained by CCA axis 1 = 7% and CCA axis 2 = 5%. b) The ordination from
xvi
a partial CCA of invertebrate, environmental, and site data shown in Fig. 3-3a, with spatial
information in the form of Shield/off-Shield site location partialled out of the data……………53
Chapter IV
Fig. 4-1. MNR Aquatic Habitat Inventory sites overlying a map of surficial geology in south-
central Ontario. Circles denote sites with complete abiotic data (dataset B), and triangles denote
additional sites with fish community data (dataset A). Dashed line indicates the geological
boundary between Shield and off-Shield sites…………………………………………………...75
Fig. 4-2. Correspondence analysis results for dataset A, where a) shows patterns among
sampling sites and b) shows associations between fish species. Green circles indicate Shield
sites and blue circles indicate off-Shield sites. Species labels are positioned at the center of each
point, with minor adjustments made for ease of reading (label key can be found in Table 4-3).
The first two axes for dataset A summarize 16.4% of the variation……………………………..81
Fig. 4-3. Correspondence analysis results for dataset B, where a) shows patterns among
sampling sites and b) shows associations between fish species. Green circles indicate Shield
sites and blue circles indicate off-Shield sites. Species labels are positioned at the center of each
point, with minor adjustments made for ease of reading. The first two axes for dataset A
summarize 18.8% of the total variation………………………………………………………….82
Fig. 4-4. Correspondence analyses for the five functional attribute analyses: a) temperature, b)
geomorphology, c) substrate, d) trophic status, and e) reproductive behavior. Shield sites are
represented as green circles, and off-Shield sites are blue circles……………………………….83
Fig. 4-5. Correspondence analysis for a) dataset A and b) dataset B, with sites coded according
to their location north of the Kirkfield Outlet or south of the Kirkfield Outlet/in areas covered by
the maximum extent of Glacial Lake Algonquin (according to Hinch et al. 1991, Fig. 6)……...84
Fig. 4-6. Canonical correspondence analysis (CCA) results for dataset B. Variables include
conductivity (Cond), alkalinity, turbidity (Turb), pH, undercut banks (UCBank), gradient (G),
canopy cover (Cov), water temperature (T), dissolved oxygen (DO), woody debris (WD), water
velocity (V), discharge, width, depth, instream rock cover (Rocks), marl/muck/rubble substrate
xvii
(S1), rock/rubble/boulder substrate (S2), meadow/cultivated landscape (L1), and shrub marsh
landscape (L2)…………................................................................................................................86
Chapter V
Fig. 5-1. 2007-2008 map of 57 sampling sites in south-central Ontario. Closed circles indicate
sites located on the Canadian Shield, open circles indicate sites located off-Shield, and grey
circles indicate “transition” sites. The heavy dark line denotes the Shield boundary in this area
of the province………………………………………………………………………………….104
Fig. 5-2. Principal component analysis of abiotic variables standardized to zero mean and unit
variance. Closed circles indicate Shield sites, gray circles indicate transition sites and open
circles indicate off-Shield sites…………………………………………………………………111
Fig. 5-3. Correspondence analysis of fish community data, showing positions of a) sites and b)
species. Species labels are positioned at the center of each point, with minor adjustments made
for ease of reading. A guide to species labels can be found in Table C-6 (Appendix C)……...113
Fig. 5-4. Correspondence analysis of macroinvertebrate community data, showing positions of
a) sites and b) taxa. Taxa labels are positioned at the center of each point, with minor
adjustments made for ease of reading. A guide to taxa labels can be found in Table C-6
(Appendix C)…………………………………………………………………………………...115
Fig 5-5. Full (a) and partial (b-d) canonical correspondence analyses (CCA) of fish community
data, using abiotic variables standardized to zero mean and unit variance. Abiotic variables
include pH, conductivity (Cond), percent gradient (G), velocity (V), dissolved oxygen (DO),
water temperature (T), moss (M), woody debris (WD), algae (AL), water vegetation (WV),
depth, substrate (S1, S2, S3) and land use (ALIS1, ALIS2). Full and partial CCA analyses were
used to determine which variables were important in distinguishing Shield and off-Shield fish
communities, where b) shows a partial CCA with variation attributed to land use variables
removed, c) partial CCA with water chemistry variables removed, and d) partial CCA with
remaining physical habitat variables removed………………………………………………….117
Fig. 5-6. Full (a) and partial (b-d) canonical correspondence analyses (CCA) of
macroinvertebrate community data, using abiotic variables standardized to zero mean and unit
xviii
variance. Abiotic variables include pH, conductivity (Cond), percent gradient (G), velocity (V),
dissolved oxygen (DO), water temperature (T), moss (M), woody debris (WD), algae (AL),
water vegetation (WV), depth, substrate (S1, S2, S3) and land use (ALIS1, ALIS2). Full and
partial CCA analyses were used to determine which variables were important in distinguishing
Shield and off-Shield macroinvertebrate communities, where b) shows a partial CCA with
variation attributed to land use variables removed, c) partial CCA with water chemistry variables
removed, and d) partial CCA with remaining physical habitat variables removed…………….118
Fig. 5-7. Graph indicating magnitude of Procrustes residuals for each site, grouped by geology.
Numbers at the top of each column indicate site number (see Table C-1, Appendix C)……….121
Appendix A
Fig. A-1. Ordination from a CA of 58 study sites and 109 macroinvertebrate taxa (data set B).
The first 2 CA axes explain 13.8% of the total variation……………………………………….148
Appendix C
Fig. C-1. Relationship between site area (m2) and a) fish and b) macroinvertebrate taxa richness
for each site. Shield sites are displayed as green circles, transition sites as pink circles, and off-
Shield sites as blue circles. No apparent relationships exist between these two variables for any
type of site, indicating that differences in total sampled area were not related to differences in
taxa richness between sites.……………………………………………………………...……..172
Fig. C-2. Survey of electrofishing efficiency of a one-pass protocol compared to a three-pass
protocol for three Shield sites (Plastic, Stoneleigh and Jerry). Each site was blocked upstream
and downstream of sampling area with block nets, sampled three times using the backpack
electrofisher. Fishes caught in each pass were kept separate from each other to determine how
many additional individuals were caught with each successive pass. Since Shield streams are
lower in conductivity than off-Shield streams and may pose a problem for effective sampling by
electrofisher, three Shield streams with varying conductivity were chosen for this analysis to
assess any differences in efficiency of the one-pass protocol in low conductivity waters. a) Plot
showing number of individual fish caught in each pass. Stoneleigh, with the highest conductivity
at 83µS, showed the least amount of decline in individuals caught between passes, suggesting
xix
that low conductivity did not reduce the efficiency of the first pass compared to the number of
individuals caught with three passes. b) Plot showing the number of unique fish species caught
in each pass. For example, three fish species were caught in the first pass at Stoneleigh, and one
additional species (not captured in the first pass) was captured in the second pass. No additional
new species were caught in the second or third passes at Jerry or
Plastic.…………………………………………………………………………………………..173
Fig. C-3. Macroinvertebrate taxa depletion curves for seven sites: Livingstone, South Nelson,
Parkside, Young, Rutherford, Waiman and Mariposa. For these sites, each pool and riffle
sample was divided into ten equal subsamples and then sorted and identified. Depletion curves
were calculated to ascertain whether a certain percentage of the overall sample was sufficient to
obtain all taxa found in the full sample…………………………………………………………174
xx
List of Appendices
Appendix A. Supplementary information for Chapter III, including ordination plot for Data Set
B (Fig. A-1) and a list of macroinvertebrate families in this analysis (Table A-1) and the
abbreviations used in all plots (Table A-2).
Appendix B. Supplementary information for Chapter IV, including a summary of the data used
in this analysis (Table B-1), a summary of all abiotic variables (Table B-2), a summary of fish
functional traits by species (Table B-3), further explanation on land-use variables (Table B-4),
and information on the amount of variation explained in the partial CCA analysis with land-use
variables (Table B-5).
Appendix C. Supplementary information for Chapter V, including a list of all sampling sites
and their associated bedrock geology (Table C-1), a summary of all fish (Table C-2),
macroinvertebrate (Table C-3) and abiotic (Table C-4) data used in this analysis, a summary of
fish functional traits by species (Table C-5), a list of macroinvertebrate families and the
abbreviations used for all plots (Table C-6), relationships between site area and taxa richness
(Fig. C-1), further explanation on electrofishing depletion surveys (Fig. C-2) and further
explanation on macroinvertebrate subsampling procedures (Fig. C-3).
1
Chapter I General Introduction
Since its inception, the field of ecology has strived to explain patterns and processes observed in
the natural world, whether on the scale of individual species, populations, communities or entire
ecosystems. Due to its inherent complexity, it is unlikely that one single concept will be able to
explain these patterns, especially at the community and ecosystem scale. Instead, ecologists
have over time developed a collection of hypotheses to help describe biotic patterns. For
example, Southwood (1977, 1988) spearheaded the idea of a habitat template, examining the
relationship between a species’ habitat and its traits. In community ecology, several hypotheses
about the rules governing biotic assemblages have propelled our understanding of why certain
species are found together and others are not. For example, Diamond (1975) proposed that
competitive interactions between species lead to nonrandom co-occurrence patterns, while others
have proposed that similar patterns could also arise by random colonization (Connor and
Simberloff 1979). Other theories have focused more on the influence of abiotic drivers on
patterns in community composition. Smith and Powell (1971) hypothesized that local
communities could be the result of a series of hierarchical abiotic and biotic “filters”, whereby
the total regional species pool is successively whittled down to the local community by factors
operating at different spatial scales. This theory has since gained some traction within ecology,
with numerous studies either expanding on this hypothesis (e.g. Tonn 1990, Poff 1997) or using
it as a framework for further study in a number of different types of communities, including
terrestrial plants (e.g., Eskelinen and Virtanen 2000), birds (e.g. Robinson et al. 2000), mammals
(e.g., Hortal et al. 2008), aquatic macroinvertebrates (e.g., Heino et al. 2003) and fishes (e.g.,
Wang et al. 2003). In recent years, there has been increased interest in the role of species
functional traits in regards to abiotic and biotic filters acting at regional or local scales, and/or
how they match with the traditional taxonomic characterization of biotic communities (e.g.,
Angermeier and Winston 1998, Layman and Winemiller 2004, Heino et al. 2007, Hoeinghaus et
al. 2007, Buisson and Grenouillet 2009, Montana and Winemiller 2010).
Within lotic community ecology, examination of the processes that ultimately determine
community composition, structure and function has long been framed within the idea that
streams and rivers are unlike other ecological systems in several key areas – namely, they have
2
strong, unidirectional flow, are dynamic in space and time, and are dominated by movements of
energy (Thompson and Lake 2010). These key ideas have spurred numerous developments
concerning the theoretical function of lotic systems, from the River Continuum Concept
(Vannote et al. 1980) to the Patch Dynamics Concept (Pringle et al. 1988, Townsend 1989), to
viewing the river as a continuous landscape, or riverscape (e.g., Schlosser 1995, Fausch et al.
2002, Allan 2004). Most recently, researchers have begun to acknowledge the links between
lotic, lentic, and riparian systems, as well as the need to further expand lotic theory to
incorporate these types of systems (Jones 2010).
These paradigms have provided a theoretical framework for studies more focused on the
influence of particular abiotic factors on lotic communities. Jackson et al. (2001) provided a
review of the key factors important in the composition, structure and function of lotic fish
communities. Water temperature can limit the range of species at different spatial scales and, in
particular areas, low temperatures may limit distributions or have impacts on stream ice cover
and the disturbance caused by ice scour. Other factors, like dissolved oxygen availability and
pH, can also limit species distribution and abundance, or be tied to behavioral or physiological
adaptations necessary to survive in unfavorable conditions. Stream morphology and other
physical aspects of the lotic habitat can also influence fish communities, where factors such as
depth and pool-riffle patterns can influence fish species richness and composition. More recent
studies have further incorporated ideas concerning spatial and temporal context, and the relative
roles of regional versus local factors. Montana and Winemiller (2010) investigated the
relationship between habitat complexity and both taxonomic and functional diversity at both a
local and macrohabitat scale, whereas other studies found additional evidence for the importance
of ecoregion or landscape-scale variables on predicting fish assemblages (e.g. Johnson et al.
2007, Stewart-Koster et al. 2007, Pinto et al 2009). However, while the importance of regional-
scale factors continues to be considered, the roles of local variables and species interactions are
also major foci of investigation. Buisson et al. (2008) found that the spatial distribution of fishes
within a river system was primarily determined by mean temperature when compared to other
local physical factors, although other studies have further examined the role of species
interactions – such as predation and competition – on the structure of lotic assemblages (e.g.,
Young 2001, Miyasaka et al. 2003, Layman and Winemiller 2004). In addition, there has been
continued attention on the role of disturbance on lotic fish communities, whether naturally
3
occurring events such as drought or flooding (e.g., Oberdorff et al. 2001, Magalhaes et al. 2007)
or through human-mediated activities such as land-use changes due to agriculture and/or
urbanization (e.g., Sutherland et al. 2002, Van Sickle et al. 2004, Wipfli and Musslewhite 2004,
Stammler et al. 2005, Orego et al. 2009), or climate change (e.g., Buisson and Grenouillet 2009).
Studies concerning macroinvertebrate community ecology often focus on lotic systems, with
particular attention paid to the role of substrate, water temperature, disturbance and water
chemistry. Studies have examined the importance of substrate size (e.g., Bourassa and Morin
1995, Williams and Smith 1996, Richards et al. 1997), roughness (e.g., Erman and Erman 1984),
and heterogeneity (e.g., Williams 1980) on macroinvertebrate taxa richness, size distribution,
productivity and colonization patterns. Townsend et al. (1983) found temperature to be an
important variable in structuring both fish and invertebrate communities, while the effects of
disturbance and flow has been investigated in relation to theories concerning community
dynamics (Resh et al. 1988, Lake 2000), as well as its impacts on macroinvertebrate assemblages
(e.g., Lepori and Malmqvist 2007, Nhiwatiwa et al. 2009, Arkle et al. 2010). Investigations into
the role played by water chemistry on macroinvertebrate communities has been extensive,
particularly the effects of pH, alkalinity and calcium availability. Several studies have linked
higher macroinvertebrate production with alkalinity (Egglishaw 1968, Dillon and Benfield 1982,
Kruger and Waters 1983, Eggert and Burton 1984). Other studies have examined more specific
effects of differences in water hardness and/or calcium on species distributions (McKillop and
Harrison 1972, Rooke and Mackie 1984) or, more broadly, the impacts of high specific
conductance, due to factors such as salinity (e.g., Kay et al. 2001) or nutrient enrichment (e.g.,
Slavik et al. 2004). Acidity also plays an important role in macroinvertebrate community
structure and function (Townsend et al. 1983), and has received attention in light of the
prevalence of acid precipitation and other anthropogenically induced changes to pH in aquatic
systems (e.g., Hall and Ide 1987, Dangles et al. 2004). Examinations of macroinvertebrate
patterns at broader spatial scales have shown that invertebrate communities can be structured by
chemical characteristics (lake invertebrates; Jackson and Harvey 1993), and in particular pH
(Longergan and Rasmussen 1996). A number of more recent studies have begun to examine
how all of the aforementioned factors operate and influence macroinvertebrate assemblages at
different spatial scales (e.g., Li et al. 2001, Sponseller et al. 2001, Townsend et al. 2003, Heino et
al. 2007, Nhiwatiwa et al. 2009). In particular, the influence of land-use patterns on
4
macroinvertebrate communities has gained significant attention (Sponseller et al. 2001, Roy et
al. 2003, Van Sickle et al. 2004), as well as specific land-use changes due to anthropogenic
activities such as agricultural development, urbanization, impoundments and forestry practices
(e.g., Stanley et al. 2002, Roy et al. 2003, Van Sickle et al. 2004, Wipfli and Musslewhite 2004,
Moore and Palmer 2005.)
All of the abiotic factors described thus far as important drivers of aspects of both fish and
macroinvertebrate lotic communities are, to some degree, influenced by other factors acting at
large spatial scales. For example, the surficial geology in a landscape can constrain and
determine abiotic factors in aquatic systems, such as drainage patterns, sediment supply, channel
morphology and water chemistry (Esselman et al. 2006 and references therein). In addition, this
feature can further influence patterns in anthropogenic land use that, in turn, can have major
effects on lotic environments (Allan et al. 1997, Allan 2004). As in many lotic studies that
broaden into a landscape perspective of the stream environment and its associated biota (e.g.,
Fausch et al. 2002), ecologists have begun to incorporate aspects of geology as a large-scale
variable potentially influencing other local factors and communities (e.g., Richards et al. 1996,
1997, Wiley et al. 1997, Wang et al. 2003, Sandin and Johnson 2004, Esselman et al. 2006,
Kratzer et al. 2006).
In North America, the Precambrian Shield is a large geological region comprised of seven
geological provinces with distinctive characteristics that are differentiated from the surrounding
areas, which collectively occupy almost half of Canada’s surface (Natural Resources Canada
2009). The Shield is characterized by ancient, Precambrian rocks underlying approximately two-
thirds of North America, the northern portions exposed during the glacial retreat roughly 10,000
years ago, and now lay at the surface or are covered with a thin layer of nutrient-poor soil.
Given our understanding of the influence of geology on other abiotic factors important to fish
and macroinvertebrate lotic communities, my doctoral thesis will examine the role of the
Precambrian Shield on the abiotic and biotic conditions of lotic systems. This issue will be
addressed within the context of identifying the major abiotic drivers of fish and
macroinvertebrate lotic communities, and how a regional-scale factor, like the Precambrian
Shield, can influence various aspects of the local lotic environment, including water chemistry,
physical habitat, and land-use patterns within a stream catchment.
5
Research Objectives
To begin, this thesis conducts an assessment of aquatic science literature across a range of
subdisciplines, with the aim of examining the degree communication and integration within
aquatic science, both currently and historically. This chapter provides a broad overview to the
field of aquatic science, and contextualizes the studies in chapters II-IV. In Chapters II and III, I
use historical data to identify the broad geographical patterns in macroinvertebrate and fish
community composition in lotic systems of south-central Ontario. I specifically examine the
influence of the Canadian Precambrian Shield on these patterns relative to adjacent systems
located off-Shield, in addition to land-use patterns and historical effects such as postglacial
dispersal. In Chapter IV, I further examine the influence of the Shield vs. off-Shield on fish and
macroinvertebrate communities at a more limited spatial scale, and tease apart the relative
influence of water chemistry, physical habitat and land-use patterns on both groups of taxa. By
collecting fishes and macroinvertebrates concurrently at all sampled sites, I am also able to
examine the degree of community concordance between fishes and macroinvertebrates,
providing further insight into the key abiotic variables driving the differences in community
composition between Shield and off-Shield sites.
Chapter II has been previously published in the Canadian Journal of Fisheries and Aquatic
Sciences, and is included in this thesis with permission from the publisher:
Neff, M.R. and D.A. Jackson. 2009. Communication and cohesion in aquatic science literature.
Canadian Journal of Fisheries and Aquatic Sciences 66: 701-712.
Chapter III has been previously published the Journal of the North American Benthological
Society, and is included in this thesis with permission from the publisher:
Neff, M.R. and D.A. Jackson. 2011. Effects of broad-scale geological changes on patterns in
macroinvertebrate assemblages. Journal of the North American Benthological Society 30: 459-
473.
References
6
Allan, J.D. 2004. Landscapes and riverscapes: the influence of land use on stream ecosystems.
Annual Review of Ecology, Evolution, and Systematics 35: 357-384.
Allan, J.D., D.L. Erickson, and J. Fay. 1997. The influence of catchment land use on stream
integrity across multiple spatial scales. Freshwater Biology 37: 149-161.
Angermeier, P.L. and M.R. Winston. 1998. Local vs. regional influences on local diversity in
stream fish communities of Virginia. Ecology 79: 911-927.
Arkle, R.S., D.S. Pilliod, and K. Strickler. 2010. Fire, flow and dynamic equilibrium in stream
macroinvertebrate communities. Freshwater Biology 55: 299-314.
Bourassa, N. and A. Morin. 1995. Relationships between size structure of invertebrate
assemblages and trophy and substrate composition in streams. Journal of the North
American Benthological Society 14: 393-403.
Buisson, L., L. Blanc and G. Grenouillet. 2008. Modeling stream fish species distribution in a
river network: the relative effects of temperature versus physical factors. Ecology of
Freshwater Fish 17: 244-257.
Buisson, L. and G. Grenouillet. 2009. Contrasted impacts of climate change on stream fish
assemblages along an environmental gradient. Diversity and Distributions 15: 613-626.
Connor, E.F. and D. Simberloff. 1979. The assembly of species communities: chance or
competition? Ecology 60: 1132-1140.
Dangles, O., B. Malmqvist and H. Laudon. 2004. Naturally acid freshwater ecosystems are
diverse and functional: evidence from boreal streams. Oikos 104: 149-155.
Diamond, J.M. 1975. Assembly of species communities. In Ecology and Evolution of
Communities, eds. M.L. Cody and J.M. Diamond. Harvard University Press, Cambridge.
Dillon Jr., R.T., and E.F. Benfield. 1982. Distribution of pulmonate snails in the New River of
Virginia and North Carolina, U.S.A.: interaction between alkalinity and stream drainage
area. Freshwater Biology 12: 179-186.
7
Eggert, S.L. and T.M. Burton. 1984. A comparison of Acroneuria lycorias (Plecoptera)
production and growth in northern Michigan hard- and soft-water streams. Freshwater
Biology 32: 21-31.
Egglishaw, H.J. 1968. The quantitative relationship between bottom fauna and plant detritus in
streams of different calcium concentrations. Journal of Applied Ecology 5: 731-740.
Erman, D.C. and N.A. Erman. 1984. The response of stream macroinvertebrates to substrate size
and heterogeneity. Hydrobiologia 108: 75-82.
Eskelinen, A. and R. Virtanen. 2005. Local and regional processes in low-productive mountain
plant communities: the roles of seed and microsite limitation in relation to grazing.
Oikos 110: 360-368.
Esselman, P.C., M.C. Freeman, and C.M. Pringle. 2006. Fish-assemblage variation between
geologically defined regions and across a longitudinal gradient in the Monkey River
Basin, Belize. Journal of the North American Benthological Society 25: 142-156.
Fausch, K.D., C.E. Torgersen, C.V. Baxter and H.W. Li. 2002. Landscapes to riverscapes:
bridging the gap between research and conservation of stream fishes. Bioscience 52:
483-498.
Hall, R.J. and F.P. Ide. 1987. Evidence of acidification effects on stream insect communities in
central Ontario between 1937 and 1985. Canadian Journal of Fisheries and Aquatic
Sciences 44: 1652-1657.
Heino, J., T. Muotka, and R. Paavola. 2003. Determinants of macroinvertebrate diversity in
headwater streams: regional and local influences. Journal of Animal Ecology 72: 425-
434.
Heino, J., H. Mykra, J. Kotanen and T. Muotka. 2007. Ecological filters and variability in
stream macroinvertebrate communities: do taxonomic and functional structure follow the
same path? Ecography 30: 217-230.
8
Hoeinghaus, D.J., K.O. Winemiller and J.S. Birnbaum. 2007. Local and regional determinants
of stream fish assemblage structure: inferences based on taxonomic vs. functional groups.
Journal of Biogeography 34: 324-338.
Hortal, J., J. Rodriguez, M. Nieto-Diaz and J.M. Lobo. 2008. Regional and environmental
effects on the species richness of mammal assemblages. Journal of Biogeography 35:
1202-1214.
Jackson, D.A. and H.H. Harvey. 1993. Fish and benthic invertebrates: community concordance
and community-environment relationships. Canadian Journal of Fisheries and Aquatic
Sciences 50: 2641-2650.
Jackson, D.A., P.R. Peres-Neto and J.D. Olden. 2001. What controls who is where in freshwater
fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of
Fisheries and Aquatic Sciences 58: 157-170.
Johnson, R.K., M.T. Furse, D. Hering, and L. Sandin. 2007. Ecological relationships between
stream communities and spatial scale: implications for designing catchment-level
monitoring programmes. Freshwater Biology 52: 939-958.
Jones, N.E. 2010. Incorporating lakes within the river discontinuum: longitudinal changes in
ecological characteristics in stream-lake networks. Canadian Journal of Fisheries and
Aquatic Sciences 67: 1350-1362.
Kay, W.R., S.A. Halse, M.D. Scanlon and M.J. Smith. 2001. Distribution and environmental
tolerances of aquatic macroinvertebrate families in the agricultural zone of southwestern
Australia. Journal of the North American Benthological Society 20: 182-199.
Kratzer, E.B., J.K. Jackson, D.B. Arscott, A.K. Aufdenkampe, C.L. Dow, L.A. Kaplan, J.D.
Newbold and B.W. Sweeney. Macroinvertebrate distribution in relation to land use and
water chemistry in New York City drinking-water-supply watersheds. Journal of the
North American Benthological Society 25: 954-976.
Kruger, C.C. and T.F. Waters. 1983. Annual production of macroinvertebrates in three streams
of different water quality. Ecology 64: 840-850.
9
Lake, P.S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North
American Benthological Society 19: 573-592.
Layman, C.A. and K.O. Winemiller. 2004. Size-based responses of prey to piscivore exclusion
in a species-rich neotropical river. Ecology 85: 1311-1320.
Lepori, F. and B. Malmqvist. 2007. Predictable changes in trophic community structure along a
spatial disturbance gradient in streams. Freshwater Biology 52: 2184-2195.
Li, J., A. Herlihy, W. Gerth, P. Kaufmann, S. Gregory, S. Urquhart and D.P. Larsen. 2001.
Variability in stream macroinvertebrates at multiple spatial scales. Freshwater Biology
46: 87-97.
Lonergan, S.P. and J.B. Rasmussen. 1996. A multi-taxonomic indicator of acidification:
isolating the effects of pH from other water-chemistry variables. Canadian Journal of
Fisheries and Aquatic Sciences 53: 1778-1787.
McKillop, W.B. and A.D. Harrison. 1972. Distribution of aquatic gastropods across an interface
between the Canadian Shield and limestone formations. Canadian Journal of Zoology
50: 1433-1445.
Magalhaes, F.M., P. Beja, I.J. Schlosser and M.J. Collares-Pereira. 2007. Effects of multi-year
droughts on fish assemblages of seasonally drying Mediterranean streams. Freshwater
Biology 52: 1494-1510.
Miyasaka, H., M. Genkai-Kato, N. Kuhara and S. Nakano. 2003. Predatory fish impact on
competition between stream insect grazers: a consideration of behaviorally- and density-
mediated effects on an apparent coexistence pattern. Oikos 101: 511-520.
Montana, C.G. and K.O. Winemiller. 2010. Local-scale habitat influences morphological
diversity of species assemblages of cichlid fishes in a tropical floodplain river. Ecology
of Freshwater Fish 19: 216-227.
Moore, A.A. and M.A. Palmer. 2005. Invertebrate biodiversity in agricultural and urban
headwater streams: implications for conservation and management. Ecological
Applications 15: 1169-1177.
10
Natural Resources Canada 2009. The atlas of Canada. Online at
http://atlas.nrcan.gc.ca/site/english/maps/environment/geology/geologicalprovinces.
Nhiwatiwa, T., T. De Bie, B. Vervaeke, M. Barson, M. Stevens, M.P.M. Vanhove, L.
Brendonck. 2009. Invertebrate communities in dry-season pools of a large subtropical
river: patterns and processes. Hydrobiologia 630: 169-186.
Oberdorff, T., B. Hugueny, and T. Vigneron. 2001. Is assemblage variability related to
environmental variability? An answer for riverine fish. Oikos 93: 419-428.
Orrego, R., S.M. Adams, R. Barra, G. Chiang and J.F. Gavilan. 2009. Patterns of fish
community composition along a river affected by agricultural and urban disturbance in
south-central Chile. Hydrobiologia 620: 35-46.
Pinto, B.C.T., F.G. Araujo, V.D. Rodrigues and R.M. Hughes. 2009. Local and ecoregion
affects on fish assemblage structure in tributaries of the Rio Paraiba do Sul, Brazil.
Freshwater Biology 54: 2600-2615.
Poff, N.L. 1997. Landscape filters and species traits: towards mechanistic understanding and
prediction in stream ecology. Journal of the North American Benthological Society 16:
391-409.
Pringle, C.M., R.J. Naiman, G. Bretschko, J.R. Karr, M.W. Oswood, J.R. Webster, R.L.
Welcomme and M.J. Winterbourn. 1988. Patch dynamics in lotic systems: the stream as
a mosaic. Journal of the North American Benthological Society 7: 503-524.
Resh, V.H., A.V. Brown, A.P. Covich, M.E. Gurtz, H.W. Li, G.W. Minshall, S.R. Reice, A.L.
Sheldon, B. Wallace and R.C. Wissmar. 1988. The role of disturbance in stream
ecology. Journal of the North American Benthological Society 7: 433-455.
Richards, C., L.B. Johnson, and G.E. Host. 1996. Landscape-scale influences on stream habitats
and biota. Canadian Journal of Fisheries and Aquatic Sciences 53: 295-311.
Richards, C., R.J. Haro, L.B. Johnson, and G.E. Host. 1997. Catchment and reach-scale
properties as indicators of macroinvertebrate species traits. Freshwater Biology 37: 219-
230.
11
Robinson, W.D., J.D. Brawn and S.K. Robinson. 2000. Forest bird community structure in
central Panama: influence of spatial scale and biogeography. Ecological Monographs 70:
209-235.
Rooke, J.B. and G.L. Mackie. 1984. Mollusca of six low-alkalinity lakes in Ontario. Canadian
Journal of Fisheries and Aquatic Sciences 41: 777-782.
Roy, A.H., A.D. Rosemund, M.J. Paul, D.S. Leigh and J.B. Wallace. 2003. Stream
macroinvertebrate response to catchment urbanization (Georgia, U.S.A.). Freshwater
Biology 48: 329-346.
Sandin, L. and R.K. Johnson. 2004. Local, landscape and regional factors structuring benthic
macroinvertebrate assemblages in Swedish streams. Landscape Ecology 19: 501-514.
Schlosser, I.J. 1995. Critical landscape attributes that influence fish population dynamics in
headwater streams. Hydrobiologia 303: 71-81.
Slavik, K., B.J. Peterson, L.A. Deegan, W.B. Bowden, A.E. Hershey and J.E. Hobbie. 2004.
Long-tern responses of the Kuparuk River ecosystem to phosphorous fertilization.
Ecology 85: 939-954.
Smith, C.L. and C.R. Powell. 1971. The summer fish communities of Brier Creek, Marshall
County, Oklahoma. American Museum Novitiates: 2458.
Southwood, T.R.E. 1977. Habitat, the templet for ecological strategies? Journal of Animal
Ecology 46: 336-365.
Southwood, T.R.E. 1988. Tactics, strategies and templets. Oikos 52: 3-18.
Sponseller, R.A., E.F. Benfield and H.M. Valett. 2001. Relationships between land use, spatial
scale and stream macroinvertebrate communities. Freshwater Biology 46: 1409-1424.
Stammler, K.L., R.L. McLaughlin and N.E. Mandrak. 2005. Streams modified for drainage
provide fish habitat in agricultural areas. Canadian Journal of Fisheries and Aquatic
Sciences 65: 509-522.
12
Stanley, E.H., M.A. Luebke, M.W. Doyle, D.W. Marshall. 2002. Short-term changes in channel
form and macroinvertebrate communities following low-head dam removal. Journal of
the North American Benthological Society 21: 172-187.
Stewart-Koster, B., M.J. Kennard, B.D. Harch, F. Sheldon, A.H. Arthington and B.J. Pusey.
2007. Partitioning the variation in stream fish assemblages within a spatio-temporal
hierarchy. Marine and Freshwater Research 58: 675-686.
Sutherland, A.B., J.L. Meyer and E.P. Gardiner. 2002. Effects of land cover on sediment regime
and fish assemblage structure in four southern Appalachian streams. Freshwater Biology
47: 1791-1805.
Thompson, R.M. and P.S. Lake. 2010. Reconciling theory and practise: the role of stream
ecology. River Research and Applications 26: 5-14.
Tonn, W.M. 1990. Climate change and fish communities: a conceptual framework.
Transactions of the American Fisheries Society 119: 337-352.
Townsend, C.R., A.G. Hildrew and J. Francis. 1983. Community structure in some southern
English streams: the influence of physiochemical factors. Freshwater Biology 13: 521-
544.
Townsend, C.R. 1989. The patch dynamics concept of stream community ecology. Journal of
the North American Benthological Society 8: 36-50.
Townsend, C.R., S. Doledec, R. Norris, K. Peacock and C. Arbuckle. 2003. The influence of
scale and geography on relationships between stream community composition and
landscape variables: description and prediction. Freshwater Biology 48: 768-785.
Vannote, R.L., G.W. Minshall, K.W. Cummins, J.R. Sedell and C.E. Cushing. 1980. The river
continuum concept. Canadian Journal of Fisheries and Aquatic Sciences 37: 130-137.
Van Sickle, J., J. Baker, A. Herlihy, P. Bayley, S. Gregory, P. Haggarty, L. Ashkenas, J. Li.
2004. Projecting the biological condition of streams under alternative scenarios of
human land use. Ecological Applications 14: 368-380
13
Wang, L., J. Lyons, P. Rasmussen, P. Seelbach, T. Simon, M. Wiley, P. Kanehl, E. Baker, S.
Niemela, and P.M. Stewart. 2003. Watershed, reach and riparian influences on stream
fish assemblages in the Northern Lakes and Forest Ecoregion, U.S.A. Canadian Journal
of Fisheries and Aquatic Sciences 60: 491-505.
Wiley, M.J., S.L. Kohler and P.W. Seelbach. 1997. Reconciling landscape and local views of
aquatic communities: lessons from Michigan trout streams. Freshwater Biology 37: 133-
148.
Williams, D.D. and M.R. Smith. 1996. Colonization dynamics of river benthos in response to
local changes in bed characteristics. Freshwater Biology 36: 237-248.
Williams, D.D. 1980. Some relationships between stream benthos and substrate heterogeneity.
Limnology and Oceanography 25: 166-172.
Wipfli, M.S. and J. Musslewhite. 2004. Density of red alder (Alnus rubra) in headwaters
influences invertebrate and detritus subsidies to downstream fish habitats in Alaska.
Hydrobiologia 520: 153-163.
Young, K. 2001. Habitat diversity and species diversity: testing the competition hypothesis
with juvenile salmonids. Oikos 95: 87-93.
14
Chapter II
Communication and cohesion in aquatic science literature
Abstract
In 1982, Frank Rigler challenged limnologists and fisheries biologists to address gaps in theory,
experimental research and management practices that have limited the advancement of both
fields. We followed up on Rigler’s concerns using a literature study to determine the objectives
and methodologies of studies across a range of subdisciplines within aquatic science. We
surveyed both recent and historical literature from five leading journals that range in emphasis to
include a broad array of subjects in aquatic science. Literature from 1982 was compared to
recent publications to determine how communication and integration within aquatic science has
changed. We found limited changes in the breadth of coverage provided by any journal. We
further analyzed contemporary literature according to subject matter, methods of analysis,
location of the research, and scale of study. We used correspondence analysis to identify the
differences and associations across these fields, and to uncover those particular research areas
that have more clearly bridged some of these gaps previously identified. Our findings indicate
that there are still clear divisions within modern aquatic science literature and that the journals
considered typically show specific emphasis in the types of questions posed, methods of analysis
and the geographic representation of authors.
Introduction
In 1982, Frank Rigler published a paper in Transactions of the American Fisheries Society
entitled “The relationship between fisheries management and limnology,” in which he addressed
the lack of interaction between limnologists and fisheries biologists, and described both possible
causes and potential solutions. As one of these solutions, Rigler advocated the further
development and use of empirical ecology, a method which concentrates on making ecological
predictions, as a way for scientists in the two fields to effectively collaborate and integrate.
Some subsequent research in limnology and fisheries science addressed some of the empirical
issues (e.g. Carpenter et al. 1985, Kerr and Ryder 1988, Poff 1997); however, it cannot be said
15
whether any significant gains were achieved in closing this perceived gap between two principal
branches of aquatic science. The purpose of this chapter, therefore, is to address Rigler’s
discussion about the lack of communication and collaboration among fields in aquatic science:
namely, what is the state of communication, collaboration and integration within the aquatic
sciences? This broad area of research includes many different subjects and research questions,
but we would argue that the overall aims and goals of these many fields are not only related, but
interconnected. For example, limnologists and oceanographers are broadly concerned with
abiotic and biotic properties such as biomass, productivity, nutrient fluxes and cycling, and to
some degree the role of biotic populations and communities in aquatic ecosystems (Lehman
1986, Peters 1986). Aquatic biologists and fisheries ecologists in general study the interactions
of individuals, species, populations and communities with their environment, whether biotic or
abiotic (for recent, well-cited examples, see Mackay 1992, Jackson et. al. 2001, and Vander
Zanden and Rasmussen 2001). Even while embracing different aspects of the umbrella term
aquatic science, these subfields have goals which are not mutually exclusive—all together, they
seek to understand the properties and dynamics of aquatic ecosystems, whether freshwater or
marine in nature. Therefore, it is reasonable to conclude that researchers in these subfields are,
or should be, interested in those areas immediately outside their own.
However, beyond general differences in subject matter, there may be other differences that tend
to limit integration and collaboration. For example, one field may have a strong emphasis on
modeling, whereas another may focus on hypothesis testing and manipulative experiments. In
addition, some scientists may focus their research on a particular species rather than a
community or ecosystem, thereby addressing questions that are more specific to that organism,
rather than to general theory.
Based on the reasoning that researchers in these subfields benefit from the advances made in
other closely related aspects of aquatic science, we endeavored to answer the following
questions. (1) To what extent do researchers in a subfield of aquatic science address and
incorporate scientific advances, i.e. literature, from other subfields? (2) Within aquatic science
literature, what are the differences and similarities among subfields in terms of the scientific
approach (research objectives, experimental design, and methods of analysis)? and (3) What
effect do these differences have on interaction among subfields?
16
To address these questions, we took a quantitative, comparative approach to the analysis of
trends in aquatic science literature. Quantitative approaches to literature analysis are often very
useful. For example, Brown (1999) surveyed the titles, abstracts and keywords in the literature to
show the trend of publications in the field of macroecology. This method effectively indicated
the rising interest in this particular ecological research philosophy since its inception some ten
years earlier.
Methods
In order to examine how certain areas of aquatic science relate to one another, we singled out
five leading journals in aquatic science to provide a cross-section of various research areas. To
date, literature analyses of this nature typically only examine a single journal in this area (e.g.,
Hildrew and Townsend 2007). Such focused reviews can offer insight into trends within a given
journal over time, but they do not address interactions with other areas of aquatic ecology.
There are a number of journals specializing in aquatic science, many with a particular focus
related to one or more of the different subfields—for example, topics in fish biology, limnology,
oceanography, or benthic invertebrate ecology. We selected five leading journals using overall
impact factor for 2005 and their relevance to aquatic science and primary subject matter in order
to cover a range of topics in aquatic science. The resulting journals include: Limnology and
Oceanography (L&O), Freshwater Biology (FB), Transactions of the American Fisheries
Society (TAFS), Journal of the North American Benthological Society (JNABS), and the
Canadian Journal of Fisheries and Aquatic Sciences (CJFAS). We do not see this set as
providing an exhaustive coverage, but rather a representation of the broader literature.
Following Brown (1999), all original research articles published in 2005 from these five journals
were assessed in several different ways in order to address our questions. Review and opinion
articles were excluded from the analysis.
Interaction among subfields of aquatic science
To assess how often authors in one subfield incorporated research from other subfields into their
own work, we quantified the number of references cited in each article from the five journals.
For example, in looking at an article published in TAFS, we tallied the number of papers from
17
Freshwater Biology, Limnology & Oceanography, Canadian Journal of Fisheries and Aquatic
Sciences, the Journal of the North American Benthological Society, and Transactions of the
American Fisheries Society that were cited within that article. The purpose of this summary was
to see the extent to which authors chose references published in journals inside and outside of
their subfield, thereby providing a measure of the broader integration of the literature. To
determine whether and how these relationships have changed since the time of Rigler’s
publication, the same analysis was performed on the 1982 volumes of these journals in order to
compare it to the 2005 information. As JNABS did not exist in 1982, it was not included in this
temporal analysis.
Patterns within aquatic science subfields
Scientific approach. The second aspect of this study was aimed at providing a more in-depth
analysis of the research published in these five journals in 2005. A summary of the topics
included in the analysis is found in Table 2-1. To assess the scientific approach used in each
study, each article was first categorized according to the research questions, ideas and/or issues
addressed within the paper. Based on our initial survey of the papers involved, we developed a
list of subject categories (Table 2-1a) that covered the range of topics often encountered in
aquatic science. It was not intended to be exhaustive in its coverage of finely resolved
categories, but rather to make certain each category included multiple papers. Articles could be
associated with more than one subject category, and subjects were chosen for each article based
on the scientific questions asked, the study objectives and/or applications.
Next, the experimental design and analysis (Table 2-1b) was assessed by describing the technical
aspects of each study, such as whether it included laboratory or field experiments, or
incorporated multiple spatial scales. This assessment was then further expanded to include
information regarding the organismal level of study (Table 2-1c), the study system (Table 2-1d)
and the taxonomic groups used in the experiment (Table 2-1e). Lastly, we examined the
methods of quantitative analysis employed in each study (Table 2-1f), such as the type of
statistical analysis or mathematical modeling methods used.
Geographical location of authors and experiment sites. For each article, the geographical
location of each author was recorded to determine whether there were regional differences in the
18
types of studies and approaches used. Those locations were then grouped into one (per author)
of six regions: North America, South America, Europe, Asia, South Pacific (including Australia
and New Zealand), and the Middle East/Africa. In addition, the geographical locations where
experiments or field sampling were carried out were recorded and entered into eight categories:
North America, South America, Europe, Asia, Atlantic Ocean, Pacific Ocean and Antarctica, the
South Pacific, and Unknown. Sampling site locations included two additional categories: one
for studies that involved the world’s major oceans (“Atlantic Ocean, Pacific Ocean and
Antarctica”), and the “Unknown/None” category, which included all studies where the sampling
site was not explicitly reported or where the study took place in a laboratory of unspecified
location.
19
20
Statistical Analysis
Given the large number of research categories and hundreds of papers included, a means of
resolving general patterns was needed. Correspondence analysis (CA) was used to highlight
general relationships among variables and journals and to summarize the major patterns in a few
dimensions. All CA analyses were run using the Biplot macro for Microsoft Excel
(http://filebox.vt.edu/artsci/stats/vining/keying/biplot_final.zip). Four different analyses were
used for this study: one analysis for the 1982 citation data, one for the 2005 citation data, one
combining common journals from 1982 and 2005, and one analysis for the 2005 scientific
approach and geographical location data. The resulting datasets were presence-absence matrices,
which were not standardized or transformed prior to analysis.
Correspondence analysis is a useful technique to show how different points – in this case,
articles—are related to one another. Points found near each other on a CA plot are more similar
than points that are positioned far apart, thereby providing a graphical representation of patterns
among points (i.e. in this case the association between individual papers and the various types of
categories) in a dataset.
Results
Interaction among subfields of aquatic ecology
The citation counts in the 1982 papers (N=483) were summarized using a correspondence
analysis for the four journals (Fig. 2-1a). The first two axes collectively explain 76% of the
variation. Note the triangular pattern, with Freshwater Biology and Limnology & Oceanography
each contributing one corner, and Canadian Journal of Fisheries and Aquatic Sciences and
Transactions of the American Fisheries Society forming the third corner collectively. Articles
that cited heavily from a particular journal were found closer to where the journal is positioned.
For example, articles clustered around L&O predominately featured citations to that same
journal. There was a strong association of TAFS articles with the TAFS journal coordinate,
showing that many TAFS articles cited papers from TAFS. In contrast, CJFAS articles were
slightly more diffuse in terms of their spatial pattern, with some points positioned closer to the
center of the array and others extending to the L&O area of the plot. Articles along a
21
“perimeter” line of the “triangle” cite articles from the two journals at the points connecting the
line, but not from the journal at the remaining apex. For example, the line of L&O points
extending from the L&O journal coordinate to the FB journal coordinate would cite both L&O
articles as well as from FB, but none from CJFAS or TAFS.
The correspondence analysis of citation counts for 2005 (N=798; Fig. 2-1b) showed a similar
triangular pattern as the 1982 data, with each journal also in a similar position in coordinate
space. There were more articles published in each journal during 2005 compared to 1982, and
so the coverage of the plot is more complete than the 1982 plot. The first two axes collectively
explain 81% of the variation. The clustering of TAFS articles around the TAFS journal centroid
remained evident, although this time there were a few points positioned further away, mostly
along the perimeter line extending up to Freshwater Biology. CJFAS, FB and L&O articles all
were concentrated around their respective journals, but also had many points which diffused out
towards the center of the array showing greater breadth in the journals cited. The points found in
the center indicated contributions that cited all the journals presented in the study with equal
frequency. Departures from equal frequency shift the location towards more commonly cited
journals. For example, the CJFAS article indicated by “A” has two citations from L&O, four
from FB, and two from CJFAS.
To further investigate how each journal shifted in the predominance of one or more journals in
the citations, or alternatively the breadth of its citations, we used correspondence analysis on the
combined 1982 and 2005 data (Fig. 2-2). Here, journal positions for 1982 and 2005 are
displayed as multivariate centroids and provided a visual representation of how citation
preferences in each journal have shifted between 1982 and 2005. As the resulting plot
demonstrated, the breadth of the citations used within papers has changed little in all journals
that were surveyed. The plot contrasts TAFS, L&O, and FB as being quite different regardless of
which year was considered. CJFAS falls towards the TAFS region (Fig. 2-2), but represented a
more intermediate location as the papers cited within CJFAS encompassed a greater range of the
total space defined (Fig. 2-1b).
22
Fig. 2-1. Correspondence analysis for (a) 1982 (N=483) and (b) 2005 (N=768) citations for the four journals, showing the association between articles and the source journal of their associated citations. Articles located close together are more likely to include references to the same journal(s). The journal coordinates form a triangular shape on the plot, representing the three-dimensional space (ternary space) of the CA results. Articles found near a particular journal cite predominately from that journal, whereas articles found in the center of the triangle cite papers from all four journals.
23
Patterns within aquatic ecology subfields
The 2005 data set was used for the more detailed examination of relationships within and among
the aquatic ecology subfields. As the full analysis included 100 categories, we present different
sets of the results in a series of plots based on a single analysis. In order to provide reference
points and aid in the interpretation, the journal centroids are presented on each plot.
Scientific approach. The relationship between research question variables and journals shows a
pattern where papers focused on species composition (SC), landscape morphology (LM),
productivity (P), and biogeochemical cycling (BC) are associated more strongly with FB, JNABS
and L&O publications. Management issues (MI), statistical and analytical methods (AES),
sampling methodologies (SM), genetics (GEN), behavior (BEH), and physiology or morphology
(PM) are most closely associated with TAFS and CJFAS publications (Fig. 2-3).
Experimental design, study scale and level are summarized for each article (Fig. 2-4). There is a
close association between mensurative or observational studies, field experiments, and
experiments including multiple spatial scales on large time scales. On the other hand,
manipulative experiments, laboratory experiments, experiments on a single spatial scale and
shorter time units (minutes, hours, days and weeks), and studies with multiple experiments are
all closely associated. In general, JNABS studies tend to be more field-oriented, encompassing
greater spatial and temporal scales than those published in L&O. TAFS and CJFAS show no
particular association to either manipulative experimental studies or observational studies, i.e.
these journals publish studies covering a more complete range of these scales and types of
studies.
For the biological level of study, there is a general trend ranging from taxonomically specific to
the broad, with individual and population-level experiments in the upper right quadrant, and
community and ecosystem-level studies in the lower left (Fig. 2-4). TAFS and CJFAS are more
associated with population- and individual-level studies, whereas community- and ecosystem-
level studies are more common in JNABS, L&O, and FB. Species- and molecular/cellular-level
studies are relatively rare in the dataset, therefore positioned near the periphery of the plot, and
tend to be more associated with CJFAS, TAFS, and L&O and ecosystem-level studies not be well
represented in TAFS and CJFAS.
24
There are several strong associations between the study system(s), the taxonomy of the study
organism(s) and the journal in which the paper was published (Fig. 2-5). For example, fish,
macrocrustaceans (e.g. decapods), mollusks, amphibians and reptiles are all primarily associated
with TAFS and CJFAS, whereas coral, zooplankton, phytoplankton and cyanobacteria are
associated with L&O. Zoobenthos and birds are closely associated with JNABS and FB. Studies
of lakes, ponds, reefs, wetlands, reservoirs, and terrestrial environments are more specific to
L&O, FB and JNABS, but less important in general within CJFAS and TAFS. In contrast, these
latter journals have a greater emphasis on open-ocean studies and about an equal split between
them and L&O for marine coastal studies. Estuarine and river/stream studies appear to be
equally frequent across the journals, with possibly less coverage in L&O and more in FB. Lab-
based studies are generally not common in FB or JNABS, but are more commonly found in TAFS
and L&O.
The last aspect to the scientific approach examined the methods of quantitative analysis used in
each study (Fig. 2-6). The coordinates for “univariate statistical analyses” and “hypothesis
testing” are located in the center of the CA biplot approximately midway from all journals,
indicating that these techniques were generally associated with all journals; multivariate analyses
and descriptive studies fell a bit further out in the lower left quadrant and relates to their greater
prevalence within JNABS and FB and less so in TAFS and CJFAS. Articles with an emphasis on
theory, statistical analysis of genetic relationships, mathematical modeling, testing and
comparing statistical models, or the use of no statistics at all fell along the positive end of CA
axis 2, closest to TAFS and CJFAS, suggesting strong theoretical and modeling focus in these
journals, as well as those being the predominant outlets for papers having no quantitative
elements.
Geographical location of authors and experiment sites. Regional geographical information for
each author and field or lab research location was examined separately to see if there were any
patterns in the geographical extent of authorship in any journal. CJFAS (79.6%), JNABS (68.7%)
and especially TAFS (99.2%) showed strong author representation in North America (NA),
whereas a more equal split between North America and Europe was seen in L&O (55.4% and
42.6%, respectively) and an emphasis towards European authors was seen in FB (32.1% and
57.2%, NA and Europe respectively). FB also had contributions from the South Pacific, with
13.2% of the articles having one or more authors from that region. TAFS had almost no
25
representation outside of North America, with 1.5% coming from Europe, 0.8% from Asia, and
0% from the remaining regions of the Middle East/Africa, South America, and the South Pacific.
Study site location had a strong association with the location of the authors, and showed the
same trends within each journal (Fig. 2-7). Again, the majority of study sites published in TAFS
(81.8%), JNABS (62.7%), and CJFAS (56%) were located in North America, whereas FB (24.5%
and 44%) and L&O (35.4% and 27.2%) were more evenly represented by North America and
Europe. In addition, studies conducted in the Atlantic/Antarctic/Pacific were more likely to
involve authors from North America and the South Pacific than any of the other author locations.
In order to ascertain the amount of cross-regional collaboration in each journal, the number of
articles with authors from more than one region were averaged across the total number of articles
published in that journal in 2005. L&O (15.4%) and FB (12.6%) had the highest percentage of
articles with authors from more than one region, whereas TAFS had the lowest percentage
(1.5%). L&O (4 articles) and CJFAS (3 articles) also had papers with authorship representing
three or more regions.
26
Fig. 2-2. Graphical representation of the degree of change between 1982 and 2005 in the composition of
citations for each journal. The change in position of a journal between years provides a measure of the relative
change in its composition of citations over that time period. For example, the 2005 position for Freshwater
Biology is positioned slightly below its 1982 indicating a reduction in the relative frequency of citations to itself
and inclusion of more papers from the other three journals.
27
Fig. 2-3. A correspondence analysis of all data categories (see Table 2-1) was conducted for all papers
published from the 5 journals in 2005. This figure and figs. 2-4 through 2-6 use the same coordinate system to
provide details regarding various research issues; therefore relative positions of variables can be readily
compared among the various figures. This figure shows the coordinates for research topics and the journal
centroids. The proximity of the research topics to one another or to a journal centroid provides a summary of
their association. Smaller distances represent closer association and angles relative to the origin provide a
measure of whether the variables (categories) are positively or negatively associated. For example, studies of
behavior (BEH) and species composition (SC) are strongly, but negatively associated in their co-occurrence in
these journals.
28
Fig. 2-4. Subset of the 2005 CA showing coordinates for experimental design and level of study with average
coordinate scores for all articles in each journal. The scale matches that used in Figures 2-3 through 2-6 so
associations between points in each figure can be made based on their proximity to one another.
29
Fig. 2-5. Subset of the 2005 CA showing coordinates for taxonomic group(s) and study system with average
coordinate scores for all articles in each journal. The scale matches that used in Figures 2-3 through 2-6 so
associations between points in each figure can be made based on their proximity to one another.
30
Fig. 2-6. Subset of the 2005 CA showing coordinates for method of analysis with average coordinate scores for
all articles in each journal. The scale matches that used in Figures 2-3 through 2-5 so associations between points
in each figure can be made based on their proximity to one another.
31
Fig. 2-7. Bar graphs indicating the overall percentages of regional (a) author location and (b) study site location
for the five journals. Study site location includes two additional regions: the Atlantic, Pacific or other ocean(s),
and the category “none” for studies which did not clearly state the location.
32
Discussion
The intent of this study was to examine the broad field of aquatic science with three objectives in
mind: firstly, to assess the extent of integration between subfields of aquatic ecology in the
current literature; secondly, to determine whether the journals and subfields have shown
increased integration over the near quarter century since Frank Rigler challenged us to do so; and
thirdly, to provide a descriptive analysis of the research objectives, geographical patterns and the
experimental and analytical methods within each subfield. Although this study does not provide
an in-depth examination of the amount of collaboration and breadth of each study across the
entire field, the analysis does allow some general statements about the interconnections between
various subfields in aquatic science.
Interaction among subfields of aquatic science
In general, this analysis indicates that there are still substantial differences in citation patterns
among the included subfields of aquatic science, represented by the four journals that were
selected for comparison between 1982 and 2005. Although this analysis provides only a visual
comparison, it is evident that the locations of all four journals show little change in coordinate
space across the years. The same triangular pattern in the CA plots of 1982 and 2005 and the
overall clustering of articles near the journal in which they are published indicates that
researchers still tend to cite studies within the same subfield and journal. As the journals tend to
have more specialized research areas, such a finding is not greatly surprising, although we did
anticipate an increase in the amount of overlap and use of literature from the other journals over
time. It appears that generally speaking, Rigler’s challenge has not been met and the various
areas of aquatic sciences tend not to capitalize on the findings of other fields to further their own
research agendas. It is discouraging that after 23 years, there are still very few examples of
studies that broadly integrate the literature from across these areas.
Our analysis also indicated a pattern between citation diversity and the research focus of a
publication. In general, we found that papers that examined several research questions on a
community or ecosystem scale were likely to cite papers from two or more of the journals that
were included in this analysis. In contrast, studies that addressed a single research question or a
single taxon were more likely to cite more papers that were published in the same journal, and
33
tended to include fewer citations from any of the remaining journals included in this analysis. It
should be noted that low citation diversity should not necessarily be an indication that a
particular study is narrow in scope. However, this analysis shows that studies which embrace a
broader view of aquatic ecology and involve several research topics are more likely to include
contributions from multiple subfields in formulating ideas and supporting outcomes. In turn, it is
likely that these particular studies will be of greater interest not only to researchers who are
looking at similar ideas, but also researchers from other areas of aquatic science.
Delving further into this analysis, there are particular subfields which tend to be more insular
than others. Looking at the journals that show the least amount of dispersion among the different
publications, it is apparent that TAFS papers are tightly bound to TAFS itself and CJFAS, as
TAFS articles are in general strongly clustered around the TAFS and CJFAS journal coordinates.
In addition, there is a striking absence of TAFS articles which cite papers from all four other
journals used in this analysis or relatively few including L&O, FB or JNABS. L&O also had few
publications which cited papers from TAFS, likely due to the non-vertebrate emphasis within the
L&O publications, whereas CJFAS and FB show the most variation across the journals. In the
case of TAFS as an example, this insularity could indicate that the focus of TAFS is more
specifically related to fish and less about broader areas of aquatic science. However, those
interested in fish communities, habitat requirements, or food web studies would seemingly
benefit from literature published outside of this subfield.
One issue to consider when analyzing how and which references are cited in an article is the
advent of electronic database searches through which searches are conducted. With these tools,
it is possible to find articles on a particular topic regardless of the journals in which they were
published. Therefore with the increasing use of these electronic resources there is an expectation
that the breadth of literature easily located and accessed should increase, thereby facilitating a
broader use of the literature. However, the analysis of temporal patterns does not offer much
support for that hypothesis.
Another issue to consider is that we only included how often the four journals that were used for
this overlap study were cited. This means that potentially, many authors may have cited papers
from a range other journals (journals with a broad focus on ecology, genetics, etc), information
which was not included in this analysis due to our more restricted focus. In addition, there were
34
twenty articles that did not cite any of our four study journals, and were not included as they
could not be used in the multivariate approach. Therefore, this aspect of our study is a summary
of publications that cited at least one article in at least one of the journals we considered.
Patterns within aquatic science subfields
The descriptive analysis of the type of research published in each of the five journals aims at
elucidating similarities and differences across aquatic science subfields in terms of scientific
approach (research objectives, experimental design and methods of analysis) and geographical
associations. Ultimately, this can be used to comment on how these differences and similarities
might limit or enhance interaction among subfields. Overall, the results indicate particular
patterns in each journal regarding the types of studies that are typically published. One striking
pattern is the separation of CJFAS and TAFS, two journals with a heavy emphasis on fish, from
the remaining three. Besides the more obvious differences, such as predominant study organism
– fish and mollusks for CJFAS and TAFS, and zoobenthos for JNABS and FB and organisms
such as phyto- and zooplankton for L&O – the journals differ in other ways as well. Both CJFAS
and TAFS have a greater tendency to publish studies at the population or individual level, often
including the development and implementation of mathematical models, and/or with an emphasis
on the more applied aspects of aquatic science, such as management and development of
sampling and statistical methods. Conversely, JNABS, FB and L&O place more emphasis on
community- and ecosystem-level studies, with research questions to reflect that emphasis, such
as landscape factors, productivity and biogeochemical cycling. In addition, although univariate
statistics and hypothesis testing were common analytical features across all five journals,
multivariate statistics were represented more predominantly in JNABS and FB.
There were some points of similarity among all five journals, as well as some patterns which do
not parcel out strongly according to journal. For example, research topics such as anthropogenic
effects, abiotic/environmental effects, and non-indigenous species were well-represented across
the board. These particular subjects are both broad and high-profile issues in current ecological
literature, and it is likely that these topics would be of interest to readers of all journals. The
results regarding experimental design indicate two general types of designs: long-running
mensurative field studies over multiple spatial scales, and manipulative laboratory experiments
on shorter time scales, often including multiple experiments. For example, mensurative studies
35
with long sampling intervals and large spatial scales were more associated with JNABS papers,
whereas L&O had a greater emphasis on manipulative studies of shorter temporal and spatial
scales. However, both types show up in all five journals.
One of the most striking patterns in this analysis is the association between region of publication,
author location, and collaboration across geographical regions. To varying degrees, North
American journals (TAFS, CJFAS, and JNABS) attract authors predominantly from North
America, with L&O being the exception. Freshwater Biology, based out of England and
Australia, and L&O both attract authors from a wider area, primarily Europe and Asia. In
addition, FB and L&O published more studies with authors that have collaborated across
geographical regions – for example, one paper might include a researcher from North America,
another from Europe and another from Asia. In the case of L&O, this trend may be due to in part
to the natural tendency of oceanographic studies to focus on larger spatial scales, allowing for
collaboration across the world. In general, it is not surprising that a journal attracts authors from
the same region in which it is published, given many have connections to professional societies
having similar geographic associations. Similarly, many researchers conduct their field or lab
studies in areas close to their institutional home due to many reasons, e.g. ease of logistics,
reduced cost, common language. However, it can be argued that journals that attract papers from
all over the world provide a more representative and broader snapshot of their field given their
greater likelihood of highlighting research that may otherwise go unnoticed if published in
journals having a more regional specialization.
The trends in aquatic science literature highlighted by this analysis offer some insight into the
trends seen in the citation analysis mentioned previously. As noted, TAFS in particular, is very
likely to cite papers predominantly from TAFS or from CJFAS rather than the remaining three
journals. For example, analytical methods such as model development and implementation,
particularly in regards to application in management protocols, were more prevalent TAFS and
CJFAS, which also are very likely to include studies which often site either TAFS or CJFAS and
not the other three journals included here. These citation patterns, therefore, could be reflecting
the idea that there are particular tools or ideas which are more dominant in one subfield over
another. This could further support why integration among subfields is sometimes difficult, i.e.
some of the basic approaches to studying the organisms differ amongst these fields.
36
Although our analysis indicates that there are patterns associated with each journal, it is
important to note that some researchers may attempt to achieve diversification in their publishing
record by choosing to publish in a wide array of journals rather than submitting a broad paper to
a narrowly focused journal. This may be particularly relevant today, as there are many more
journals available now than in 1982, many of which specialize in a particular area of aquatic
science. Rather than helping to integrate and cross-fertilize the literature, this abundance of
publishing outlets could potentially lead to even greater specialization and insular literatures.
This interpretation of the publication decisions and connections represents only some of the
possible factors determining the outcomes, as this analysis is limited to describing patterns and
not able to determine the underlying causes of those patterns.
Aquatic science today
Rigler’s (1982) comments on the relationship between limnology and fisheries science sparked
several response papers, some of which have been cited considerably, all focused around the
need for predictive methods to be the main objective in aquatic science research (e.g., Peters
1986, Pace 2001). In this analysis, we aimed to take a step further back from that debate about
prediction in fisheries science and limnology and look at more general trends in aquatic science.
We do see some minor steps towards a more integrative discipline when we compare one
measure of the level of interaction between subdisciplines from 1982 to 2005. However, on the
whole the degree of cross-citation (or alternatively, cross-fertilization of ideas) has remained
relatively unchanged. In addition, our overall assessment of patterns within aquatic science
indicates that there are some areas where subdisciplines remain largely separate from one
another.
A more standardized method of approaching science — for example, through the use of
predominately predictive techniques — is possibly one way to strengthen ties between
subdisciplines, but ultimately may be impractical or take many years for it to be integrated into
common practice. Instead, we suggest that there should more effort put into events and
organizations which promote inter-disciplinary collaboration. Examples of such organizations
are the US National Center for Ecological Analysis and Synthesis (NCEAS) or the Canadian
Institute for Ecology and Evolution. Such forums are broad in their overall mandate rather than
just aquatic science, but the overall aim is to advance knowledge through synthesis, collaboration
37
and data sharing between institutions and research groups, which can lead to improvements in
accessibility and applicability for resource managers and policy makers.
Another way to promote interdisciplinary collaboration is through integration at the level of
scientific meetings and conferences. These venues already offer important opportunities for
scientists to meet, discuss ideas and gain exposure to current research in their field. More efforts
to hold multiple-disciplinary meetings could further the interactions between aquatic ecologists.
For example, the Society of Canadian Limnologists (SCL) and the Canadian Conference for
Fisheries Research (CCFFR) have been holding their annual conferences jointly for many years,
often with increasingly integrated sessions and symposia. More meetings, sessions, and
organizations that incorporate several subdisciplines of aquatic science could go a long way in
both promoting advances made in each field, as well as encouraging interdisciplinary
collaboration.
Change that are to be made within the sciences are likely to be gradual; however, it is our hope
that with time, the ties between the many facets of aquatic science will be strengthened with the
ultimate goal of better understanding and management of our aquatic ecosystems. In the light of
the physical, climatic and biological changes this planet is currently facing and will continue to
face, it is in our best interest.
References
Brown, James H. 1999. Macroecology: Progress and prospect. Oikos 87: 3-14.
Carpenter, S.R., JF Kitchell, and JR Hodgson. 1985. Cascading trophic interactions and lake
productivity. Bioscience 35(10): 634-639
Hildrew, A.G. and C.R. Townsend. 2007. Freshwater Biology – looking back, looking forward.
Freshw. Biol. 52(10): 1863-1867.
Jackson, D.A., P.R. Peres-Neto, and J.D. Olden. 2001. What controls who is where in
freshwater fish communities – the roles of biotic, abiotic, and spatial factors. Can. J.
Fish. Aquat. Sci. 58(1): 157-170.
38
Kerr, S.R., and R.A. Ryder. 1988. The applicability of fish yield indexes in freshwater and
marine ecosystems. Limnol. Oceanogr. 33(4): 973-981.
Lehman, J.T. 1986. The goal of understanding in limnology. Limnol. Oceanogr. 31(5): 1160-
1166.
Mackay, R.J. 1992. Colonization by lotic macroinvertebrates – a review of processes and
patterns. Can. J. Fish. Aquat. Sci. 49(3): 617-628.
Pace, M.L. 2001. Prediction and the aquatic sciences. Can. J. Fish. Aquat. Sci. 58(1): 63-72.
Peters, R.H. 1986. The role of prediction in limnology. Limnol. Oceanogr. 31 (5): 1143-1156.
Poff, N.L. 1997. Landscape filters and species traits: Towards mechanistic understanding and
prediction in stream ecology. J. North Am. Benthol. Soc. 16(2): 391-409.
Rigler, F.H. 1982. The relation between fisheries management and limnology. Trans. Am.
Fish. Soc. 111(2): 121-132.
Vander Zanden, M.J. and J.B. Ramussen. 2001. Variation in delta N-15 and delta C-13 trophic
fractionation: Implications for aquatic food webs studies. Limnol. Oceanogr. 46(8):
2061-2066
39
Chapter III
Effects of broad-scale geological changes on patterns in macroinvertebrate assemblages
Abstract
Understanding the broad-scale factors that influence biological communities has long been a goal
of community ecology. We used benthic macroinvertebrate data to identify broad geographical
patterns in macroinvertebrate community composition and specifically to examine the influence
of the Precambrian Shield on stream abiotic and biotic conditions. The Precambrian Shield is a
geological feature that encompasses most of northern North America. Geology differs between
Shield and off-Shield areas, creating distinctly different physical and chemical conditions in
aquatic systems. We focused this regional-scale study on south-central Ontario, where both
Shield and off-Shield conditions are found in adjacent areas. We used constrained and
unconstrained multivariate analyses to examine patterns in biotic, abiotic, and spatial variables.
Our results showed that, in low-order lotic systems, macroinvertebrate communities differ
between Shield and off-Shield streams. Shield streams have higher dissolved O2, velocity, and
discharge, larger amounts of woody debris, and greater canopy cover than off-Shield streams. In
contrast, off-Shield streams have higher conductivity, alkalinity, pH, turbidity, and water
temperature, and frequently are surrounded by meadow, cultivated, or pastured land. In general,
macroinvertebrate communities at off-Shield sites had a greater proportion of taxa preferring
pool or depositional habitats, whereas macroinvertebrate communities at Shield sites contained
taxa typically associated with riffles or erosional habitats. Analysis of spatial location indicated
that the Shield/off-Shield distinction probably is the result of a combination of intertwined
abiotic and spatial factors.
Introduction
Community ecologists have highlighted a need to better understand better the relative roles of
regional and local processes in determining species composition. Ricklefs (1987) suggested that
to broaden concepts of the regulation of community structure, we must unite the effects of local
and regional processes in community theory. Many studies (Jackson and Harvey 1989, Tonn et
40
al. 1990, Wiley et al. 1997, Angermeier and Winston 1998, Vinson and Hawkins 1998, Wang et
al. 2003, Allan 2004, Hoeinghaus et al. 2007) have incorporated these ideas into various
examinations of aquatic communities in an attempt to disentangle regional and local effects,
particularly with regard to managerial decisions.
Much of northern North America is a vast network of lake and river systems as a consequence of
the last glaciation. Several investigators have examined broad-scale patterns to understand better
the regional factors structuring lake fish communities. Jackson and Harvey (1989) found that
regional patterns in Ontario fish communities reflect broad-scale factors, such as postglacial
colonization and extrinsically regulated environmental conditions (e.g., lake thermal regimes),
whereas variation within regions is explained by local-scale geomorphic and chemical
characteristics of individual lakes. Mandrak (1995) further examined these patterns and found
that postglacial dispersal and climatic processes structure regional patterns of fish species
richness in Ontario lakes.
Fish community structure seems to be highly dependent on historical pathways of dispersal in the
years following the glacial retreat, but there is good reason to think that different mechanisms
may be responsible for the distribution and community composition of macroinvertebrates. In
particular, dispersal mechanisms of most macroinvertebrates (i.e., an aerial dispersal stage) differ
from those of fishes and, therefore, macroinvertebrates may be less constrained by waterways for
post-glacial dispersal. As suggested by Hynes (1970) and later reviewed by Vinson and Hawkins
(1998), many factors can affect the distribution and community composition of aquatic
macroinvertebrates. The most notable factors include dispersal ability (Mackay 1992, Bohonak
and Jenkins 2003), water chemistry (McKillop and Harrison 1972, Minshall and Minshall 1978,
Huryn et al. 1995), physical habitat features (Allan 1975, Erman and Erman 1984, Bourassa and
Morin 1995), disturbance (Wallace 1990, Huryn et al. 1995) and, at a local scale, competition
and predation (McAuliffe 1984, Holomuzki and Short 1988, Burdon and Harding 2008).
Geology is one broad-scale feature of a landscape that can profoundly affect the physical and
chemical attributes of aquatic systems (Johnson et al. 1997, Dow et al. 2006), which can, in turn,
affect the biological communities of those systems (Esselman et al. 2006, Kratzer et al. 2006).
The Precambrian Canadian Shield (referred to herein as the Shield) is a large-scale and well-
known attribute of the North American landscape. The Shield encompasses a broad area
41
covering much of Canada, with extensions into the northern US in the Midwest, New York, and
New England. In this expansive area, ancient Precambrian metamorphic and igneous bedrock
lies at, or close to, the surface and is covered by only a thin layer of nutrient-poor soil. In south-
central Ontario, Shield geology is in sharp contrast to the younger, primarily sedimentary
bedrock of southern Ontario and most of remaining North America (Chapman and Putnam
1984). The weathering characteristics of these 2 geological types differ sharply. Shield rock
weathers because of freeze/thaw fracturing, whereas off-Shield sedimentary rock often weathers
via chemical dissolution by water (Chapman and Putnam 1984). Thus, the physical and chemical
attributes of aquatic systems on the Shield are likely quite different from off-Shield systems. For
example, extremely low conductivities and a propensity for acidification resulting from poor
buffering capacity against the threat of acid precipitation have been widely reported in Shield
systems (Ricker 1934, Beamish and Harvey 1972, Schindler et al. 1980, Hall and Ide 1987,
Kelso et al. 1990, Bowman et al. 2006). Recently, researchers have noted changes to
zooplankton and crayfish communities because of declining Ca levels in Shield systems caused
by recovery from acidification (Jeziorski et al. 2008, Edwards et al. 2009). In addition to
differences in water chemistry on- and off-Shield, the more immediate proximity of Shield
bedrock to the surface may affect the physical structure of aquatic systems, particularly during
development of classical pool–riffle morphology of streams (Wohl and Legleiter 2003), but this
effect less well studied. The prevalence of bedrock outcroppings on the Shield suggests that the
substrate of Shield streams will be dominated by coarser materials, e.g., bedrock and boulders,
and that systems with areas of exposed bedrock will have poorly developed classical pool–riffle
relationships simply because the rock resists eroding into pools. The slow erosional rates of
Shield bedrock will translate chemically to waters with low conductivity, low Ca levels, and poor
buffering capacity against acid inputs.
Aquatic systems of the Canadian Shield have been the subject of some research (Giberson and
Mackay 1991, Jackson and Harvey 1995, Yan et al. 1996, Jeziorski et al. 2008, Edwards et al.
2009), but many of these studies were focused exclusively on lentic systems or involved only
fish communities. Broad-scale patterns of lotic macroinvertebrate assemblages have been the
focus of some recent studies, particularly with regard to differences attributed to ecoregions
(Hawkins et al. 2000 and references therein, Waite et al. 2000, Van Sickle and Hughes 2000,
Kratzer et al. 2006), but these patterns have yet to be examined in relation to the Canadian
42
Shield. We examined abiotic and biotic patterns across southern and central Ontario to address
two main objectives. First, we identified broad geographical patterns in macroinvertebrate
community composition. We predicted that abiotic factors associated with the Shield vs off-
Shield contrast would affect biotic composition in lotic systems in south-central Ontario. We
identified the abiotic variables that characterized Shield and off-Shield systems and asked
whether differences between macroinvertebrate communities were related to differences in those
factors. We expected to observe significant differences in physical habitat and water chemistry
between Shield and off-Shield streams. We also expected to find significant changes in
macroinvertebrate community composition associated with abiotic changes. For example, we
expected to find O2-sensitive or acid-tolerant species prevalent in Shield systems because of
steeper physical gradients (i.e., slope) and more erosional habitats, whereas we expected Ca-
sensitive species to be less common because of the poor buffering capacity of the Shield
geology.
Methods
In the 1970s, the Ontario Ministry of Natural Resources (OMNR) began a sampling program
called the Aquatic Habitat Inventory Survey (AHIS). The purpose of this survey was to provide a
record of fish and macroinvertebrate communities in lotic and lentic systems across the province.
Field crews from local and regional OMNR offices sampled streams and rivers following
protocols outlined in the AHIS manual (AHIS 1979). The geographical extent of this data set
allows for broad-scale analysis of historical macroinvertebrate communities in a large number of
lotic systems in Ontario.
The AHIS method involved extensive qualitative sampling designed to collect most of the
species present at a site. Sampling occurred in as many different habitats as possible and was
done with a variety of equipment, including dip nets, Surber samplers, bottom dredges, and drift
nets. The sampling and processing protocol was designed to retain invertebrates exceeding a
standard sieve size of 59 µm (AHIS 1979). The AHIS did not call for a rigidly standardized
sampling procedure but emphasized exhaustive sampling that yielded a similar sampling effort at
all sites. All invertebrate specimens collected at each site were preserved for identification,
unless a large number of crayfish or large clams (>20) were collected, in which case a subsample
of �10 individuals was retained (Aquatic Habitat Inventory Surveys 1979).
43
We used a subset of all AHIS records to examine patterns in macroinvertebrate community
composition. We retained all 1st- to 3rd-order stream sites (determined from 1:50,000 topographic
maps) from 6 OMNR districts covering a wide area of south-central Ontario to yield 125 sites
(Fig.3-1). We restricted the analysis to small systems to ensure that any differences in
community composition were not caused by effects of stream order and that sampled stream size
did not differ systematically among districts. In addition, we restricted this analysis to south-
central Ontario to minimize differences in climate patterns among sites, particularly between
Shield and off-Shield locations. Sites from both data sets were mapped onto a geographical
information system (GIS) layer of bedrock geology to determine whether they were on the
Canadian Shield (Fig. 3-1) before classifying them.
After collection, all macroinvertebrate samples were identified to the lowest possible taxonomic
level, which, in many cases, was to genus or species. Identifications were completed by qualified
personnel in the Ontario Ministry of Natural Resources. For this analysis, we scaled all
identifications to family to account for possible misidentifications at lower taxonomic levels and
to account for groups identified to only levels above species or genus. We converted original
records from abundance to presence/absence to ameliorate the potential disparity in sampling
effort among sites and districts. The final data set consisted of a 125 site × 121 invertebrate taxa
(families) presence/absence matrix.
Records for 58 of the 125 sites included in this analysis contained extensive abiotic data
collected at the same time as the macroinvertebrate sampling. This information included
physical–chemical data, such as water temperature (T), dissolved O2 (DO), pH, turbidity (Turb),
alkalinity (Alk), specific conductance (Cond), amount of in-stream cover represented as organic
debris (OD; included accumulations of leaves, twigs, and algae), woody debris (WD), in-stream
rock cover, undercut banks (UCBank), width, depth, velocity, and discharge. In addition, the
original data set also included % composition or categorical variables, including substrate size
(categories: rock, boulder, rubble, gravel, sand, silt, clay, muck, marl, and detritus), % canopy
cover of stream, degree of gradient (i.e., channel slope; categorized as low, medium, or high),
bank stability (a measure of bank erosion, bank slumping, and amount of trees and other
vegetation), and the surrounding landscape type assessed laterally to 50 m from the sampling
reach (categories: cultivated, firm pasture, meadow, upland hardwood, upland conifer, swamp
hardwood, swamp conifer, shrub marsh, and open marsh). Variables within a set of categories or
44
their percentages are not independent (Jackson 1997), so we used correspondence analysis (CA)
to summarize the variation within each group of variables (i.e., substrate, canopy cover, channel
slope, bank stability, and landscape type) into �1 correspondence analysis axes and used their
resulting site score as a new variable. These summary variables included 2 substrate axes (S1,
S2), 1 canopy-cover axis (Cov), 1 channel-slope axis (G), 1 bank-stability axis (BS), and 4 axes
summarizing landscape composition (L1, L2, L3, L4) (Table 3-1).
For these analyses, the 1st data set (A) included 125 sites with macroinvertebrate
presence/absence of 121 taxa collected between 1972–1977, and the 2nd data set (B) included the
58-site subset with macroinvertebrate presence/absence of 109 taxa plus corresponding
environmental information represented in 23 abiotic variables. Data set A was used to assess
broad-scale patterns in macroinvertebrate community composition because this data set included
more sites and covered a greater spatial extent. Data set B was used to examine the relationship
between macroinvertebrate community composition and the abiotic environment.
45
46
Fig. 3-1. Map depicting bedrock geology and sampling sites in south-central Ontario. The heavy dark lines
depict the Precambrian Shield boundary in this area of the province.
47
Statistical Analysis
CA was used to examine the general patterns in regional macroinvertebrate community
composition in data set A. CA is a multivariate indirect gradient analysis well suited to analyzing
presence/absence data (Jackson and Harvey 1989). Discriminant analysis was used to elucidate
patterns in environmental variation between Shield and off-Shield systems. This analysis
summarized covariation among the abiotic variables in data set B to test whether environmental
characteristics differed between Shield and off-Shield sites. Both CA and Canonical
Correspondence Analysis (CCA) with data set B were used to determine the relationship
between biotic and abiotic variables and to elucidate which environmental variables best
explained macroinvertebrate community patterns across south-central Ontario. CCA with a
single predictor categorical variable (Shield/off-Shield) was used to test whether differences in
macroinvertebrate community composition existed between Shield and off-Shield streams. A
permutation test (n = 9999) was then used to determine if the resulting constrained axis was
significant, i.e., were communities from the two regions significantly different?
Last, partial CCA was used to examine the relative influence of spatial position in the landscape
on observed biotic and abiotic patterns (Borcard et al. 1992). We used Shield/off-Shield as a
conditional variable to remove the variation attributable to the large-scale spatial location from
the biological data set before constraining it by the abiotic variables. This technique effectively
controlled for differences between Shield or off-Shield in the interpretation of the patterns in
invertebrate communities and environmental variables.
All statistical analyses were done with the vegan package (Oksansen et al. 2009) in R (version
2.10.1; R Development Core Team, Vienna, Austria).
Results
Data set A: regional community composition
CA axis 1 for data set A summarized a gradient of taxa that tend to prefer slow-moving waters
(positive values) over fast-flowing waters (negative values) (Fig. 3-2a; see Table A-1 (Appendix
A) for taxon codes). Nearly all families of Trichoptera, Plecoptera, and Ephemeroptera had
negative values, whereas nearly all families of Hemiptera, Gastropoda, Amphipoda, Hirudinea,
48
and Isopoda had positive values. Generalist taxa (e.g., Chironomidae, Baetidae) tended to be
weakly negative or positive, a result indicating that they were well represented across most sites
or that they lacked affinity for particular conditions. The composition of gastropods underwent a
transition along the axis from pulmonate snails (e.g., Planoboridae, Physidae, Lymnaeidae),
which had strong positive values on axis 1, to prosobranch snails (e.g., Bithyniidae, Viviparidae),
which had weakly positive to negative values. Hirudinidae (Hirudinea) and all families of
Amphipoda and Isopoda had strongly positive values on axis 1. Prosobranch snails and most taxa
of Odonata, Hirudinea, Hydrachnida, Oligochaeta, Nematomorpha, Tricladida, and Unionidae
were strongly associated with the far negative end of CA axis 2. When the first 2 axes were
examined simultaneously, 3 families of Odonata (Macromiidae, Gomphidae, and
Calopterygidae) and 2 families of Diptera (Ceratopogonidae, and Empididae) were strongly
associated with each other.
Each site was coded according to the bedrock geology to examine patterns of community
composition in relation to the Canadian Shield (Fig. 3-2b). The result was a strong pattern along
the first 2 CA axes. Shield sites occurred predominantly on the negative ends of both axes, and
off-Shield sites occurred primarily on the positive ends of both axes. This pattern indicated that
macroinvertebrate composition at a site differed strongly between Shield and off-Shield sites.
Data set B: community composition and abiotic environment
Linear discriminant analysis revealed that specific conductance and alkalinity were most
important for distinguishing between Shield and off-Shield sites (Table 3-2). A jackknife (i.e.,
cross-validated; Olden and Jackson 2001) approach was used to classify individual sites as
Shield or off-Shield based on their environmental characteristics. Twenty-three Shield sites
(95.8%) and 32 off-Shield sites (94%) were correctly classified in this analysis. In general, off-
Shield sites had with higher conductivity and alkalinity values, whereas Shield sites had higher
water velocity, more dense canopy cover, unstable banks, and greater stream depths.
Relationships found in the unconstrained CA on data set B were consistent with those for data set
A, a result indicating that sites in data set B were a representative subset of the broader set (Fig.
A-1). When the abiotic variables were used to constrain the invertebrate community data in a
CCA of data set B, 47.3% of the overall variation was explained by the resulting CCA axes (Fig.
49
3-3a). On CCA axis 1, sites with negative scores were associated with higher DO concentrations,
higher velocity, greater amounts of woody debris, a denser canopy, and higher % sand and %
clay substrate. Taxa associated with the negative side of CCA axis 1 were most families of
Trichoptera, Plecoptera, Ephemeroptera, Lepidoptera, Viviparidae (a prosobranch gastropod),
Unionidae, and Torrenticolidae (a hydrachnid). Sites at the positive end of the axis were more
alkaline, with higher specific conductance, turbidity, and warmer thermal conditions. In addition,
these sites had increased amounts of meadow, cultivated or pasture riparian areas, and larger
substrates. Associated taxa included most families of Zygoptera, slow-water Hemiptera, most
gastropods, amphipods, Hirudinea, the remaining Hydrachnida families, and Isopoda. CCA axis
2 also showed similar patterns, with higher values of pH, specific conductance, alkalinity,
meadow, and cultivated or pastured landscapes associated with sites and species positioned
toward the negative end of the axis. Associated taxa included predominantly shredder and
scraper families of Trichoptera, Plecoptera, and Coleoptera; skater families of Hemiptera;
pulmonate gastropods; Amphipoda; and Isopoda. Sites toward the positive end of CCA axis 2
were associated with greater stream depth, width, discharge, unstable banks, and more in-stream
woody debris. Taxa associated with the positive end of CCA axis 2 were most odonate families,
Ephemeroptera preferring depositional habitats, prosobranch gastropods, and most families of
Hirudinea, Bivalvia, Lepidoptera, Hydrachnida, and Oligochaeta.
The single CCA axis produced by constraining the macroinvertebrate community data location
(Shield/off-Shield) was highly significant (p < 0.0001). This result indicated that sites could be
grouped by geological location with confidence based on their macroinvertebrate community
composition.
Partial CCA was used to assess the relative influence of spatial location and environment on the
patterns observed among individual sites. Environmental information alone explained 44.8%,
location (Shield/off-Shield; i.e., regional differences in space) alone explained 2.5%, and the
covariation between space and environment explained 2.6% of the total variation. When the
effect of Shield/off-Shield location was partialled out of these data, the ability of abiotic
variables to separate Shield/off-shield sites was reduced. Some Shield sites were grouped in the
ordination plot, but scores of Shield and off-Shield overlapped strongly (Fig. 3-3b). Abiotic
variables with negative scores on CCA axis 1 were DO, pH, % sand and clay substrates, and
velocity. Among the abiotic variables with positive scores, the importance of conductivity and
50
alkalinity decreased and depth and water temperature increased. Patterns on the spatially
conditioned CCA axis 2 were similar to those on the unconditioned CCA axis 2. pH and %
pasture or cultivated land use had strong negative scores associated mainly with off-Shield sites,
whereas bank instability and discharge were strongly associated mainly with Shield sites.
However, width, depth, and woody debris had less influence on the spatially conditioned CCA
axis 2, and both channel slope and % sand and clay substrates were more influential on CCA axis
1.
When spatially-structured abiotic information was partialled out of the data set, the
macroinvertebrate communities at Shield/off-Shield sites were less distinctly separated based on
abiotic variables. This result suggested that some key abiotic differences that distinguish Shield
and off-Shield sites were removed with the conditional Shield/off-Shield categorical variable.
51
52
Fig. 3-2. The ordination resulting from a Correspondence Analysis (CA) of a) 121 macroinvertebrate taxa and b)
125 study sites located either on or off the Precambrian Shield. The first 2 CA axes explained 11% of the
variability of invertebrate community composition. See Table A-1 for taxon abbreviations. Labels are centered
over the point for each taxon.
53
Fig. 3-3. a) Ordination from a Canonical Correspondence Analysis (CCA) of 109 invertebrate taxa, 21
environmental variables, and 58 sites. Vector variables are: dissolved O2 (DO), velocity (V), alkalinity (Alk),
conductivity (Cond), turbidity (Turb), width, depth, discharge, organic debris (OD), woody debris (WD), rocks,
undercut banks (UCBank), landscape variables (L1, L2, L3, and L4), substrate variables (S1, S2), gradient (G),
canopy cover (Cov), and bank stability (BS). Variability explained by CCA axis 1 = 7% and CCA axis 2 = 5%.
b) The ordination from a partial CCA of invertebrate, environmental, and site data shown in Fig. 3-3a, with
spatial information in the form of Shield/off-Shield site location partialled out of the data.
54
Discussion
We showed that the Canadian Precambrian Shield, a large-scale, abrupt geological transition,
influences both abiotic conditions and regional patterns in macroinvertebrate community
composition in lotic systems of south-central Ontario. These analyses accurately distinguished
sites as Shield or off-Shield based on environmental characteristics alone or based on
macroinvertebrate community composition data alone. Other investigators have shown that
patterns in macroinvertebrate community composition can be caused by differences in physical
habitat (Allan 1975, Erman and Erman 1984, Bourassa and Morin 1995) and water chemistry
(McKillop and Harrison 1972, Minshall and Minshall 1978, Huryn et al. 1995), both of which
differ between Shield and off-Shield sites. Previous studies in which the effect of a geological
transition was examined explicitly tended to be focused on only a few taxa, such as communities
or populations of gastropods (McKillop and Harrison 1972, Huryn et al. 1995). Results of other
studies incorporating changes in geology between ecoregions have been mixed. For example,
Waite et al. (2000) compared lotic macroinvertebrate assemblages in ecoregions in the Mid-
Atlantic Highlands (USA) where geological differences between some ecoregions are similar to
the differences between Shield and off-Shield locations. Waite et al. (2000) found no clear site
clusters by ecoregion in a CA ordination of macroinvertebrate assemblages, but they did find that
sites could be arranged consistently according to stream slope, size, and water chemistry, which
are important variables distinguishing Shield from off-Shield sites. In contrast, Kratzer et al.
(2006) showed that geology might be an important determinant of macroinvertebrate
communities in ecoregions in the New York City drinking-water-supply watersheds, but they
pointed out that the relative importance of geological variables and patterns of land use were
difficult to disentangle.
Two driving variables in the distinction between Shield and off-Shield lotic systems appeared to
be conductivity and alkalinity, both of which typically were much higher at off-Shield than at
Shield sites. This result was expected because of the differences in weathering rates between
Shield (predominantly granites and gneiss) and off-Shield (limestone) bedrock in Ontario
(Chapman and Putnam 1984). These differences in weathering rates produce differences in
chemical composition (e.g., increased levels of many elements, including P, and higher pH) that
ultimately influence primary productivity (Dillon and Rigler 1974). Shield sites tended to have
55
higher water velocity, steeper channel slopes, denser canopy cover, less stable banks, and greater
channel depths, whereas off-Shield sites were more turbid, with more stable banks and open
canopies, and were more frequently surrounded by meadow, cultivated, or pastured land. These
physical differences all fit within expectations for Shield and off-Shield systems, except that
unstable banks were more prevalent in Shield systems. However, the measure of bank stability
was based on both slope and lack of vegetation, which could apply to systems with steep, bare,
boulder or bedrock banks (i.e., steep slopes are regarded as an indicator of unstable banks, but
may also be the result of granitic bedrock that resists erosion), conditions would be expected at
Shield sites.
Other variables were important in distinguishing macroinvertebrate communities at Shield and
off-Shield sites. For example, Shield sites were more northerly than off-Shield sites because of
the geographical location of the Canadian Shield. However, restricting the geographic scope of
the study area to south-central Ontario effectively reduced the influence of broad-scale climatic
differences that could have confounded the analyses. Moreover, much of the spatial variation in
invertebrate communities was associated with an east–west gradient that probably was caused by
lake-effect climate variability and proximity of the study area to Lake Huron.
The interplay between geology and land use in determining the ecological condition of lotic
systems has been well documented (e.g., Allan 2004, Dow et al. 2006, Kratzer et al. 2006), with
the general conclusion that separating the effects of geology and land use on macroinvertebrate
communities is tricky at best. Geology often plays a large role in determining land use, and this
is the case in Ontario, where the nature of the Shield limits agricultural development in Shield
areas. Land-use data relevant to the time period covered by our data set were not available for the
geographical extent of the study area, so we were unable to include a more detailed comparison
of the effects of land use on macroinvertebrate communities in Shield and off-Shield systems. A
cursory analysis of current land-use data (Aquatic Landscape Inventory Software [ALIS], 2001–
2003), indicated that usage differs between Shield and off-Shield areas. Off-Shield areas tend to
have more cropland, pasture, and other types of less forested land uses than Shield areas. These
land-use data were compiled ~30 y after the historical surveys used in our study, but the
predominant uses have not changed during this time interval for this region. We used a partial
CCA to attempt to control for differences in land use. The results indicated that land-use
accounted for only a very small portion of the total variation in macroinvertebrate communities.
56
When land-use information was partialled out of the data set, the distinction between Shield and
off-Shield sites remained. Previous studies comparing Shield and off-Shield lentic systems also
revealed patterns similar to those found in this study (Jackson and Harvey 1989). That is, Shield
and off-Shield sites were distinct, regardless of the fact that the off-Shield sites encompassed a
range of different habitat types and conditions.
Physical habitat often is an important variable when distinguishing macroinvertebrate
communities. Substrate size was not an influential abiotic variable in the distinction between
Shield and off-Shield sites, but some variability in macroinvertebrate communities could be
attributed to preferences for erosional or depositional habitats. In general, more taxa that tend to
prefer depositional habitats, such as Odonata, Hemiptera, Coleoptera, Hirudinea, Gastropoda,
Diptera, Hydrachnida, and Isopoda, were found at off-Shield than at Shield sites. These taxa are
found predominantly in slower waters and have adaptations, such as sprawling, climbing,
burrowing, and diving habits, that enable them to exploit food resources primarily found in
depositional habitats. In contrast, more taxa that tend to prefer erosional habitats, such as
Trichoptera, Plecoptera, and Ephemeroptera, were found at Shield than at off-Shield sites. Many
of these taxa are clingers, with adaptations to hold on to substrates in fast water and exploit food
resources found in swifter currents. Variation exists in the types of habitat exploited by species
within each invertebrate family, but this general distinction in the types of taxa associated with
Shield and off-Shield systems is consistent with the physical characteristics of each system.
We addressed the effect of spatial organization of sites by using a partial CCA to tease apart the
effects of location (Shield/off-Shield) and environmental conditions on the arrangement of sites
according to their species composition. Spatial information accounted for only a minor
percentage of the variation explained by the analysis, but it had a significant influence on the
separation of Shield and off-Shield sites in the resulting biplot of the first 2 constrained axes.
This difference in the community ordination suggests that some key abiotic differences that
distinguish Shield and off-Shield sites were removed once the effect of spatial location was
removed. Given the strong differences in conductivity and pH and the relatively minor
differences in most other variables between regions, we suggest that water chemistry was the
most influential determinant of differences between communities in the two regions.
Some taxa may be constrained to off-Shield sites by physiological limitations related to the low
57
ionic composition or low calcium concentration in Shield waters. Calcium is an essential element
for all organisms for a variety of physiological and structural processes, and its concentration in
aquatic systems is often dependent on substrate rock type (Webster and Patten 1979). Our
analyses showed that most gastropods and other crustaceans, such as Amphipoda and
Cambaridae, were strongly associated with off-Shield sites. For example, only one gastropod
family (Viviparidae, a prosobranch snail that requires higher O2 levels and permanent water) was
strongly associated with Shield streams. McKillop and Harrison (1972) found a similar pattern in
which densities of pulmonate snails were higher in hard water (325–885 µS/cm) and of
prosobranch snails were higher in medium to soft water (55–86 µS/cm). They suggested that
pulmonate snails may be physiologically limited from occurring in low-conductivity water, but
the exact mechanism for their limitation was not specified. Nearly 45% of all freshwater
gastropods are restricted to waters with calcium concentrations >25 mg/L (Thorp and Covich
1991). Ontario Ministry of the Environment data (Ministry of Environment, unpublished data)
indicate that calcium levels in streams in the south-central Ontario region of the Canadian Shield
range from 1.0 to 15.3 mg/L (higher values often are related to local deposits of glacial till
containing carbonate rocks) and conductivity values range from 14.4 to 161 µS/cm (n = 126). In
our study, conductivity ranged from 26 to 181 µS/cm at Shield sites and from 312 to 600 µS/cm
at off-Shield sites. The low conductivity of Shield water could interact with biological factors to
lead to indirect biotic interactions. For example, if pulmonate snails are less well adapted than
prosobranch snails to low-calcium waters, then their sparse abundance in Shield waters might be
the result of physiological stress or of other factors, such as competitive interactions with
prosobranch snails. In addition, many fish species that occur in Shield waters are able to crack
the shells of gastropods and prey upon them effectively (Scott and Crossman 1998). Pulmonate
shells tend to be thinner than prosobranch shells (Dillon 2000), so pulmonate snails may be at
more risk than prosobranch snails in environments with snail-predating fishes. A lack of
available calcium in the water may make snails with thin shells even more vulnerable to
predation. However, Dillon (2000) noted that large interspecific variation exists in calcium
tolerance in freshwater gastropods, and the breadth of tolerance of the taxa in the study region is
not known. The patterns of association match well with predictions based on the literature, but
further support is required from experimental work to determine the role of calcium in regulating
the composition of gastropods and, potentially, other taxa. For example, similar distribution
patterns were observed for Amphipoda and Isopoda families, but too little information is
58
available in the literature concerning the physical and chemical habitat preferences of these
organisms to draw any further conclusions. Minshall and Minshall (1978) noted that some
indirect evidence exists to suggest that Gammarus pulex (Amphipoda) may be affected by
calcium levels. Glazier et al. (1992) found that some Amphipod species were absent from
systems with conductivities <25 µS/cm At present, little direct evidence exists to support a
hypothesis of calcium limitation on the distribution of benthic invertebrates, although attention to
this question is increasing because of declining calcium concentrations in areas on the Canadian
Shield (Jeziorski et al. 2008, Edwards et al. 2009).
Another potential limiting variable constraining taxa to Shield or off-Shield sites is pH. In some
ways, pH is related to water-chemistry variables like calcium, but pH is of particular interest in
Shield systems because of the effects of acidic precipitation in the area. In a study of Canadian
Shield lakes, Lonergan and Rasmussen (1996) noted that the abundances of several
macroinvertebrate taxa were significantly associated with pH after other water-chemistry
variables were partialled out of the analysis. For instance, Hyallela azteca (Amphipoda) was an
important indicator of pH variation among lakes and was found predominantly in waters with
higher pH (e.g., Glazier et al. 1992 and references therein). This strong association with pH also
may help explain the association of Amphipoda taxa with off-Shield sites in this analysis because
Shield sites were generally more acidic than off-Shield sites.
Four Shield sites—North, Stoddard, Esson, Beaver, and McCue—consistently grouped with off-
Shield sites in both abiotic and biotic analyses. A possible explanation for this outcome is that all
of these sites are near the southern edge of the Canadian Shield, where the geology is a more
complex mix of metamorphic and igneous rock than the gneiss-dominated deposits further north.
For example, McCue Creek is the only Shield site on carbonate metasedimentary bedrock, and
this distinction could explain why both the environmental and biological characteristics of this
stream had scores in the area of overlap between Shield and off-Shield points in both the CA and
CCA. Therefore, sites close to the transition area between Shield and off-Shield geology may
exhibit abiotic and biotic characteristics of both areas.
The influence of the Canadian Precambrian Shield on both the physiochemical habitat and
macroinvertebrate communities of lotic systems suggests that this feature is an important
regional-scale landscape determinant of biological patterns in south-central Ontario. Such broad-
59
scale environmental discontinuities probably are important determinants of biodiversity, and
could enhance regional biodiversity by providing strong contrasts in environmental conditions. It
is not clear whether such differences in community composition are common in response to such
natural discontinuities, and we acknowledge that we were unable to disentangle fully the
potential effects of differences in land-use patterns between Shield and off-Shield sites that may
influence macroinvertebrate communities. Regardless, macroinvertebrate community
composition differed significantly between Shield and off-Shield lotic systems in this analysis.
These differences may be important when considering the potential for species ranges to change
in response to factors, such as climate change. Some conditions, such as thermal environments,
may change, but underlying physiological conditions (caused by water chemistry in this
instance) may limit the potential for changes in distributions of taxa.
References
AHIS (Aquatic Habitat Inventory Survey). 1979 and 1987. Manual of instructions. Ontario
Ministry of Natural Resources – Fisheries Branch. Queen’s Printer for Ontario, Ontario,
Canada.
Allan, J. D. 1975. The distributional ecology and diversity of benthic insects in Cement Creek,
Colorado. Ecology 56:1040–1053.
Allan, J. D. 2004. Landscapes and riverscapes: the influence of land use on stream ecosystems.
Annual Review of Ecology, Evolution, and Systematics 35:357–384.
Angermeier, P. L., and M. R. Winston. 1998. Local vs. regional influences on local diversity in
stream fish communities of Virginia. Ecology 79:911–927.
Beamish, R. J., and H. H. Harvey. 1972. Acidification of the La Cloche mountain lakes, Ontario,
and resulting fish mortalities. Journal of the Fisheries Research Board of Canada
29:1131–1143.
Bohonak, A. J., and D. G. Jenkins. 2003. Ecological and evolutionary significance of dispersal
by freshwater invertebrates. Ecology Letters 6:783–796.
Borcard, D., P. Legendre, and P. Drapeau. 1992. Partialling out the spatial component of
60
ecological variation. Ecology 73:1045–1055.
Bourassa, N., and A. Morin. 1995. Relationships between size structure of invertebrate
assemblages and trophy and substrate composition in streams. Journal of the North
American Benthological Society 14:393–403.
Bowman, M. F., K. M. Somers, R. A. Reid, and L. D. Scott. 2006. Temporal response of stream
benthic macroinvertebrate communities to the synergistic effects of anthropogenic
acidification and natural drought events. Freshwater Biology 51:768–782.
Burdon, F. J., and J. S. Harding. 2008. The linkage between riparian predators and aquatic
insects across a stream–resource spectrum. Freshwater Biology 53:330–346.
Chapman, L. J., and D. F. Putnam. 1984. The physiography of Southern Ontario. Special
Volume 2. Ontario Geological Survey, Sudbury, Ontario.
Dillon, P. J., and F. H. Rigler. 1974. The phosphorus–chlorophyll relationship in lakes.
Limnology and Oceanography 19:767–773.
Dillon, R. T. 2000. The ecology of freshwater Molluscs. Cambridge University Press,
Cambridge, UK.
Dow, C.L., D. B. Arscott, and J. D. Newbold. 2006. Relating major ions and nutrients to
watershed conditions across a mixed-use, water-supply watershed. Journal of the North
American Benthological Society 25:887–911.
Edwards, B. A., D. A. Jackson, and K. M. Somers. 2009. Multispecies crayfish declines in lakes:
implications for species distributions and richness. Journal of the North American
Benthological Society 28:719–732.
Erman, D. C., and N. A. Erman. 1984. The response of stream macroinvertebrates to substrate
size and heterogeneity. Hydrobiologia 108:75–82.
Esselman, P. C., M. C. Freeman, and C. M. Pringle. 2006. Fish-assemblage variation between
geologically defined regions and across a longitudinal gradient in the Monkey River
Basin, Belize. 2006. Journal of the North American Benthological Society 25:142–156.
61
Giberson, D. J., and R. J. Mackay. 1991. Life history and distribution of mayflies
(Ephemeroptera) in some acid streams in south central Ontario, Canada. Canadian
Journal of Zoology 69:899–910.
Glazier, D. S., M. T. Horne, and M. E. Lehman. 1992. Abundance, body composition and
reproductive output of Gammarus minus (Crustacea: Amphipoda) in ten cold springs
differing in pH and ionic content. Freshwater Biology 28:149–163.
Hall, R. J., and F. P. Ide. 1987. Evidence of acidification effects on stream insect communities in
central Ontario between 1937 and 1985. Canadian Journal of Fisheries and Aquatic
Sciences 44:1652–1657.
Hawkins, C. P., R. H. Norris, J. Gerritsen, R. M. Hughes, S. K. Jackson, R. K. Johnson, and R. J.
Stevenson. 2000. Evaluation of the use of landscape classifications for the prediction of
freshwater biota: synthesis and recommendations. Journal of the North American
Benthological Society 19:541–556.
Hoeinghaus, D. J., K. O. Winemiller, and J. S. Birnbaum. 2007. Local and regional determinants
of stream fish assemblage structure: inferences based on taxonomic vs. functional groups.
Journal of Biogeography 34:324–338.
Holomuzki, J. R., and T. M. Short. 1988. Habitat use and fish avoidance behaviors by the
stream-dwelling isopod Lirceus fontinalis. Oikos 52:79–86.
Huryn, A. D., A. C. Benke, and G. M. Ward. 1995. Direct and indirect effects of geology on the
distribution, biomass and production of the freshwater snail Elimia. Journal of the North
American Benthological Society 14:519–534.
Hynes, H. B. N. 1970. The ecology of stream insects. Annual Review of Entomology 15:24–42.
Jackson, D. A. 1997. Compositional data in community ecology: the paradigm or peril of
proportions? Ecology 78:929–940
Jackson, D. A., and H. H. Harvey. 1989. Biogeographic associations in fish assemblages: local
vs. regional processes. Ecology 70:1472–1484.
62
Jackson, D. A., and H. H. Harvey. 1995. Gradual reduction and extinction of fish populations in
acid lakes. Water, Air and Soil Pollution 85:389–394.
Jeziorski, A.., N. D. Yan, A. M. Paterson, A. M. DeSellas, M. A. Turner, D. S. Jeffries, B.
Keller, R. C. Weeber, D. K. McNicol, M. E. Palmer, K. McIver, K. Arsenaeau, B. K.
Ginn, B. F. Cumming, and J. P. Smol. 2008. The widespread threat of calcium decline in
fresh waters. Science 322:1374–1377.
Johnson, L. B., C. Richards, G. E. Host, and J. W. Arthur. 1997. Landscape influences on water
chemistry in Midwestern stream ecosystems. Freshwater Biology 37:193–208.
Kelso, J. R. M., M. A. Shaw, C. K. Minns, and K. H. Mills. 1990. An evaluation of the effects of
atmospheric acidic deposition on fish and the fishery resource of Canada. Canadian
Journal of Fisheries and Aquatic Sciences 47:644–655.
Kratzer, E. B., J. K. Jackson, D. B. Arscott, A. K. Aufdenkampe, C. L. Dow, L. A. Kaplan, J. D.
Newbold, and B. W. Sweeney. 2006. Macroinvertebrate distribution in relation to land
use and water chemistry in New York City drinking-water-supply watersheds. Journal of
the North American Benthological Society 25:954–976.
Lonergan, S. P., and J. B. Rasmussen. 1996. A multi-taxonomic indicator of acidification:
isolating the effects of pH from other water-chemistry variables. Canadian Journal of
Fisheries and Aquatic Sciences 53:1778–1787.
Mackay, R. J. 1992. Colonization by lotic macroinvertebrates: a review of processes and
patterns. Canadian Journal of Fisheries and Aquatic Sciences 49:617–628.
Mandrak, N. E. 1995. Biogeographic patterns of fish species richness in Ontario lakes in relation
to historical and environmental factors. Canadian Journal of Fisheries and Aquatic
Sciences 52:1462–1474.
McAuliffe, J. R. 1984. Competition for space, disturbance and the structure of a benthic stream
community. Ecology 65:894–908.
McKillop, W. B., and A. D. Harrison. 1972. Distribution of aquatic gastropods across an
interface between the Canadian Shield and limestone formations. Canadian Journal of
63
Zoology 50:1433–1445.
Minshall, G. W., and J. N. Minshall. 1978. Further evidence on the role of chemical factors in
determining the distribution of benthic invertebrates in the River Duddon. Archiv für
Hydrobiologie. 53:324–355.
Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, R. G. O’Hara, G. L. Simpson, P. Solymos,
M. Henry, H. Stevens, and H. Wagner. 2009. Vegan: community ecology package. R
library. R Project for Statistical Computing, Vienna, Austria.
Olden, J. D., and D. A. Jackson. 2001. Torturing data for the sake of generality: how valid are
our regression models? Ecoscience 7:501–510.
Ricker, W. E. 1934. An ecological classification of certain Ontario streams. University of
Toronto Studies, Biological Series Number 37. Publications of the Ontario Fisheries
Research Laboratory 49:1–114.
Ricklefs, R. E. 1987. Community diversity: relative roles of local and regional processes.
Science 235:167–171.
Schindler, D. W., R. H. Hesslein, and R. Wagemann. 1980. Effects of acidification on
mobilization of heavy–metals and radionuclides from the sediments of a freshwater lake.
Canadian Journal of Fisheries and Aquatic Sciences 37:373–377.
Scott, W. B., and E. J. Crossman. 1998. Freshwater fishes of Canada. Galt House Publications
Limited, Oakville, Ontario.
Thorp, J. H., and A. P. Covich (EDITORS). 1991. Ecology and classification of North American
freshwater invertebrates. Academic Press, San Diego, California.
Tonn, W. M., J. J. Magnuson, M. Rask, and J. Toivonen. 1990. Intercontinental comparison of
small-lake fish assemblages: the balance between local and regional processes. American
Naturalist 136:345–375.
Van Sickle, J. and R.M. Hughes. 2000. Classification strengths of ecoregions, catchments, and
geographic clusters for aquatic vertebrates in Oregon. Journal of the North American
64
Benthological Society 19: 370-384.
Vinson, M. R., and C. P. Hawkins. 1998. Biodiversity of stream insects: Variation at local, basin
and regional scales. Annual Review of Entomology 43:271–293.
Waite, I.R., A.T. Herlihy, D.P. Larsen and D.J. Klemm. 2000. Comparing strengths of
geographic and nongeographic classifications of stream benthic macroinvertebrates in the
Mid-Atlantic Highlands, USA. Journal of the North American Benthological Society 19:
429-441.
Wallace, J. B. 1990. Recovery of lotic macroinvertebrate communities from disturbance.
Environmental Management 14:605–620.
Wang, L., J. Lyons, P. Rasmussen, P. Seelbach, T. Simon, M. Wiley, P. Kenehl, E. Baker, S.
Niemela, and P. M. Stewart. 2003. Watershed, reach and riparian influences on stream
fish assemblages in the Northern Lakes and Forest Ecoregion, U.S.A. Canadian Journal
of Fisheries and Aquatic Sciences 60:491–505.
Webster, J. R., and B. C. Patten. 1979. Effects of watershed perturbation on stream potassium
and calcium dynamics. Ecological Monographs 49:51–72.
Wiley, M. J., S. L. Kohler, and P. W. Seelbach. 1997. Reconciling landscape and local views of
aquatic communities: lessons from Michigan trout streams. Freshwater Biology 37:133–
148.
Wohl, E., and C. J. Legleiter. 2003. Controls on pool characteristics along a resistant-boundary
channel. Journal of Geology 111:103–114.
Yan, N. D., W. Keller, N. M. Scully, D. R. S. Lean, and P. J. Dillon. 1996. Increased UV-B
penetration in a lake owing to drought-induced acidification. Nature 381:141–143.
65
Chapter IV
Effects of a geological transition on stream fish communities: the influence of the Canadian Precambrian Shield
Abstract
Biological communities are influenced by a variety of regional and local factors, and
understanding the relative importance of these factors has long been a goal of community
ecology. In this study, I use fish data from low-order lotic systems to identify broad
geographical patterns in community composition, and to specifically examine the influence of
the Precambrian Shield on both stream abiotic and biotic conditions. The Precambrian Shield is
a geological feature that encompasses most of northern North America, and contrasts to areas
off-Shield, creating distinct physical and chemical conditions in aquatic systems. This regional-
scale study focuses on south-central Ontario, where both Shield and off-Shield conditions are
found adjacent to each other. Both constrained and unconstrained multivariate analyses were
used to quantify relationships in biotic and abiotic variables, in addition to Indicator Species
Analysis and a fish species’ functional trait analysis. The results show that in low-order lotic
systems, Shield streams have fish communities distinct from off-Shield streams. Lotic fish
communities on the Shield are associated with higher velocity, increased amounts of instream
woody debris, dissolved oxygen and shrub marsh landscapes. In contrast, off-Shield
communities are characterized by higher conductivity, alkalinity, water temperature, and
turbidity, and are frequently surrounded by meadow, cultivated or pastured land. In general, off-
Shield sites were more species rich, and were associated with species preferring a wide variety of
habitats, whereas Shield sites were associated with a subset of regional species preferring more
specific habitats. Overall, my results indicate distinct abiotic and biotic conditions between
Shield and off-Shield low-order lotic systems within a small geographic range.
Introduction
Community ecologists have long sought to ascertain the key factors important in determining
species richness, composition, abundance and other metrics describing biological communities
(e.g., McAuliffe 1984, Marsh-Matthews and Matthews 2000, Condit et al. 2002). Studies in
community ecology range from examining the effects of biotic interactions (e.g. McAuliffe
66
1984), the applicability of various predictive models of community organization (e.g. Condit et
al. 2002), to the effects of multiple factors on multiple spatial and temporal scales (e.g. Marsh-
Matthews and Matthews 2000). In many cases, the importance of one factor versus another
depends on the system and the communities in question. For fish communities, a variety of
factors concerning both the abiotic and biotic environment have been identified as being
particularly relevant. These factors can include predation, competition via resource partitioning,
climate, dissolved oxygen levels, acidity, and physical attributes of the habitat (Jackson et al.
2001). For lotic fishes, these attributes might also include water velocity, stream morphology
and structural complexity of the environment.
Determining which of these factors is most important in structuring a particular community is no
easy task but, in general, the literature suggests that species composition and richness are thought
to be the product of a hierarchy of distribution patterns over a range of spatial scales (Smith and
Powell 1971, Tonn 1990, Jackson et al. 2001). For example, regional patterns of fish species
distribution may be attributed to factors such as climate, dispersal barriers and historical
biogeography (e.g. Ricklefs 1987, Jackson and Harvey 1989, Minns 1989, Mandrak 1995),
whereas local patterns may reflect habitat diversity, competition and predation (e.g., Layman and
Winemiller 2004, Buisson et al. 2008). This hierarchical organization of processes is often
described as a series of filters that the total species pool passes through to reach a local
assemblage (Smith and Powell 1971, Tonn 1990). Ricklefs (1987) stressed the importance of
examining multiple scales, or considering both regional and local factors.
Many studies since Ricklefs (1987) have demonstrated that regional factors are important in
determining various aspects of fish community structure, including diversity, composition and
abundance of species. Angermeier and Winston (1998) found that regional fish diversity
consistently predicted local diversity, whereas Tonn et al. (1990) found that for two systems with
distinct regional and historical influences but similar local characteristics, the two systems shared
similar species richness but different assemblage patterns. More recent studies have included
aspects of landscape and watershed level factors, as well as analysis of fish functional traits, to
better describe aspects of fish assemblages. Wang et al. (2003) found that reach-scale
characteristics are influenced by watershed characteristics that, in turn, affect distribution,
abundance and other community attributes of fishes. Hoeninghaus et al. (2007) determined that
taxonomic-based community patterns were best explained by species’ geographic distributions,
67
whereas functional patterns in fish communities were explained by an equal influence of both
local and regional factors.
Several studies have attempted to explain large-scale patterns in fish communities in the
province of Ontario, Canada, with the overarching conclusion that fish species distributions are
largely shaped by historical and environmental processes. Jackson and Harvey (1989) found that
regional patterns among lentic fish assemblages reflect large-scale factors such as historical post-
glacial colonization routes, whereas variation within regions is likely influenced by small-scale
morphological and chemical characteristics of lakes. Mandrak (1995) expanded on their study,
finding provincial gradients in lentic fish species richness, concluding that patterns in species
richness are directly related to post-glacial colonization and climate, once spatial autocorrelation
is removed from the analysis. Hinch et al. (1991) highlighted certain species’ distributions that
are likely influenced by the location and maximum extent of glacial Lake Algonquin, a water
body composed of glacial melt waters that historically covered the area of present-day Lake
Huron and surrounding lands, and the Kirkfield Outlet, which connected Lake Algonquin to
Lake Ontario and provided an important dispersal route for cool- and cold-water fishes into the
new melt waters left behind by the retreating glaciers in the uplands of central Ontario.
While these studies provide valuable insight into regional-scale patterns in fish species
distributions, most studies have focused primarily on lentic systems and few have examined
patterns in relation to an important landscape feature at the same regional scale – the Canadian
Precambrian Shield. The Shield, as it is commonly known, is a broad area characterized by
Precambrian metamorphic and igneous bedrock lying at or close to the surface, covered by a thin
layer of nutrient-poor soil (Chapman and Putnam 1984). This geological feature covers much of
Canada, with extensions into the northern United States in the Midwest, New York, and New
England. In south-central Ontario, the Shield geology stands in sharp contrast to the younger,
primarily sedimentary bedrock of southern Ontario. The inherent differences in weathering
characteristics of Shield geology compared to the limestone-dominated bedrock in “off-Shield”
areas of Ontario has a particularly strong influence on aquatic systems. Shield streams and lakes
are known to have extremely low conductivity and a propensity for acidification due to the poor
buffering capabilities of Shield bedrock (Jackson and Harvey 1989), and lotic systems may
further show different physical development, particularly with a general lack of classical pool-
riffle morphology. By focusing on a smaller geographic area containing both Shield and off-
68
Shield regions, the effect of this strong environmental discontinuity on fish communities can be
effectively examined.
Although the influence of the Canadian Shield on fish community structure is largely unknown,
there is also a lack of insight into the functioning of lotic ecosystems in this region. Historically,
lotic studies have been primarily confined to off-Shield southern and southwestern areas of the
province, where streams and rivers are more likely to conform to classical ideas concerning
geomorphology and water chemistry (e.g. Bowlby and Roff 1986, Kilgour and Barton 1999).
Many of the variables thought to be important in fish community structure, as outlined in
Jackson et al. (2001), may be directly influenced by Shield geology. For example, Shield
systems may be unable to support many acid-sensitive species, due to these systems’ inability to
buffer against anthropogenically generated acid precipitation inputs, which are of particular
concern in this area (e.g. Hall and Ide 1987, Kelso et al. 1990, Bowman et al. 2006).
In this study, I address two main objectives. First, I identify the broad geographical patterns in
low-order lotic fish community composition, with particular attention to the differences
attributed to the Canadian Shield. While previous work has shown that historical colonization
routes and climate are important factors in determining patterns in fish species richness and
assemblage in Ontario (Jackson and Harvey 1989, Hinch et al. 1991, Mandrak 1995), I propose
that the conditions imposed by Canadian Shield act as a regional filter. I predict that the abiotic
environment created by the Shield geology will directly influence fish community composition
in lotic systems, whereby nearby Shield and off-Shield systems will have distinct fish
communities from each other. By restricting the geographic extent of the study area to south-
central Ontario, I can reduce the influence of other important factors, such as climate, that would
otherwise be a factor at a larger geographic scale (Jackson and Harvey 1989, Mandrak 1995). As
noted previously, the location of glacial refugial origins and routes for dispersal following the
last ice age is also an important factor in explaining present-day fish species distributions. By
incorporating this knowledge with an examination of how the Shield might further influence
species distributions, we will have a clearer understanding of how fish community composition
is determined at the regional scale in Ontario.
The second objective is to identify the particular physical and chemical features that characterize
Shield and off-Shield systems in south-central Ontario, and relate these factors to characteristics
69
of fish community composition. Aspects of water chemistry, such as conductivity and acidity,
should be important in distinguishing these systems, and I expect fish assemblages will show
similar responses, with acid-sensitive species being less prominent in Shield systems. In
addition, Shield streams are expected to be marked by low productivity, high gradient, greater
water velocities and dominance of larger substrates, as a product of the resistant underlying
bedrock. I predict fish assemblages to reflect these physical attributes, with a predominance of
faster water/riffle/run-preferring species in Shield streams. Overall, I expect to find lower fish
diversity on the Shield due to differences in overall productivity, as well as potentially less
complex habitat (e.g. pools and riffles).
Methods
This study was completed using data available from a province-wide sampling program
conducted by the Ontario Ministry of Natural Resources (OMNR) in the 1970s and 1980s. The
Aquatic Habitat Inventory Survey (AHI) was implemented to provide a record of fish and
macroinvertebrate communities in both lotic and lentic systems across the province. Field crews
from OMNR district offices sampled streams and rivers following protocols outlined in the AHI
manual (OMNR 1979). There is no specific information available describing how sampling sites
were specifically chosen, but protocols indicate that it was up to the discretion of each district’s
sampling crews, with the overall aim of obtaining a general survey of lotic fish communities.
The geographical extent of this historical dataset allows for a large-scale analysis of fish
communities in a large number of lotic systems.
The sampling protocols for the AHI program were geared towards a qualitative assessment of all
fish species present at a site. Various sampling gears were recommended for use and the
backpack electrofisher was the predominant method of capture (Gareth Goodchild, personal
communication, formerly of OMNR). Although the AHI did not rigidly standardize the
sampling procedure, there was an emphasis on exhaustive sampling, such that all sites included
in this survey were sampled extensively but without providing standardized catch per unit effort
measures. Fish species were identified in the field, with voucher specimens collected and
verified by experts at the Royal Ontario Museum in Toronto, Ontario.
70
For this analysis, I chose a subset of all survey records to examine patterns of fish community
composition in regards to the Canadian Shield. As surveys were conducted through individual
OMNR districts, data from six districts (Algonquin, Bancroft, Bracebridge, Minden, Huron and
Lindsay) were included to cover lotic systems from a wide area of south-central Ontario (Fig. 4-
1) to reduce differences in climatic patterns among sites, particularly between Shield and off-
Shield sites. From the full set of sampled sites from the six districts, the dataset was then
restricted to small systems (1st – 3rd order, determined from 1:50,000 topographic maps) to
ensure that any differences in community composition were not due to the effects of stream
order, and that there was no systematic difference in stream size sampled in different districts
(Chapter III). All sites within these six districts that met these criteria were included in the
analysis.
Original records were converted from abundance to presence-absence, to minimize any potential
disparity in sampling effort between sites and districts. While abundance data may provide
additional valuable insights into patterns in fish community composition and structure, a lack of
information on specific sampling methods and effort, particularly among OMNR districts leaves
us unable to describe or quantify any sampling biases. In addition, any record that was recorded
at a taxonomic level only above species (e.g. family, genus) was removed from the data, so that
only species-level identifications were included. The final dataset (dataset A) included 104 sites
and 50 fish species, with 53 sites identified as being on-Shield and 51 sites off-Shield based on a
GIS layer of bedrock geology (Ontario Geological Survey 2003). To examine these sites relative
to historical colonization processes in this region following deglaciation, these sites were also
assessed for their location in relation to the maximum extent of glacial Lake Algonquin and the
Kirkfield Outlet (Hinch et al. 1991, Fig. 6), which connected Lake Algonquin with Lake Ontario
and provided an important dispersal route following the retreat of the glaciers. Sixty-five sites
were located north of the Kirkfield Outlet, 19 sites to the south, and 20 sites were in areas
covered by the maximum extent of Lake Algonquin.
A subset of sites in dataset A (dataset B: 17 Shield and 36 off-Shield sites) also included
complete data for a variety of abiotic variables. These variables include water temperature (T),
dissolved oxygen (DO), pH, turbidity (Turb), alkalinity (Alk), conductivity (Cond), amount of
instream cover represented as organic debris (OD), which includes accumulations of leaves,
twigs and algae, large woody debris (WD), instream rock cover, undercut banks (UCBank),
71
width, depth, velocity and discharge. In addition, the original data also included a number of
variables recorded as either percent composition or category, including substrate composition,
canopy cover, gradient, and surrounding landscape type. As percent composition variables are
not independent of each other and most standard statistical approaches are unsuitable as a result
(Jackson 1997), correspondence analysis was used to summarize the variation within each group
of variables into one or more correspondence analysis axes summarizing the dominant trends
(Table 4-1). The corresponding site score was used as a new variable summarizing the dominant
trends. These summary variables included two substrate variables (S1, S2), two landscape
variables (L1, L2), one canopy cover variable (Cov) and one gradient variable (G). A summary
of the biotic and abiotic data used in this analysis is found in Tables B-1 and B-2 (Appendix B).
Other studies examining the effects of regional factors on biological communities often highlight
the importance of land use in regards to the ecological conditions of lotic systems (e.g. Allan
2004, Dow et al. 2006, Kratzer et al. 2006). Geology and land use are often closely associated,
and this is certainly the case when considering the Canadian Shield, as the physical nature of this
geological feature leaves Shield areas generally unsuitable for agricultural development. As high
quality land-use data for the entire catchments were not available for the geographical and
temporal extent of the study area, I was not able to determine whether differences in land use
between Shield and off-Shield was the driving variable in structuring fish community
composition in these systems, rather than location relative to the Shield. In an attempt to
examine this issue, I performed additional analyses where the original landscape data collected at
the time of the sampling (i.e., variables L1 and L2) were replaced with 2001-2003 land-use data
from the Aquatic Landscape Inventory Software (ALIS), which provides percentages of different
land-use types for the entire catchment of each site. As much of the land use in this region has
not changed dramatically between the two time periods, these recent data provided an
opportunity to examine this issue. For additional information on this latter analysis, see Table B-
4 (Appendix B).
In addition to assessing differences in community structure based on taxonomic composition, I
also considered potential differences in the functional traits of communities. For example,
Shield systems may be comprised of predominantly riffle- or run-preferring species as opposed
to those preferring slower habitats. For this analysis, a combination of resources was used to
classify the fish species in both datasets A and B according to five categories of fish functional
72
traits or attributes. These five categories include temperature preferences, trophic status, stream
geomorphology preferences, substrate preferences, and reproduction behaviors (Table 4-2).
Information on species functional traits as described in Goldstein & Meador (2003) were used to
categorize species, and cross-checked these classifications with the Ontario Online Fish Database
(Eakins 2010), which primarily synthesizes information from Coker et al. (2001). Where data in
the Ontario Fish Database differed from traits described in Goldstein & Meador (2003), I chose
traits as described in the database as that information is more likely to represent characteristics of
Ontario populations for these species. This exercise resulted in five frequency matrices, which
were then transformed to relative frequency prior to statistical analysis. For further explanation
on the classification of fish species’ functional traits, refer to Table B-3 (Appendix B).
73
74
75
Fig. 4-1. MNR Aquatic Habitat Inventory sites overlying a map of surficial geology in south-central Ontario.
Circles denote sites with complete abiotic data (dataset B), and triangles denote additional sites with fish
community data (dataset A). Dashed line indicates the geological boundary between Shield and off-Shield sites.
76
Statistical Analysis
To analyze patterns in fish community composition across south-central Ontario, I used
Correspondence Analysis on the fish community data in datasets A and B. I then used Canonical
Correspondence Analysis (CCA) to test if Shield and off-Shield sites differed statistically based
on their community composition. This analysis constrains the CA using a single predictor
variable indicating a site’s location on- or off-Shield, and then performs a permutation test
(n=1000) to determine if the resulting axis is significant – in other words, whether the
communities from the two regions are significantly different from each other, or can be viewed
as a homogenous set.
To further examine patterns in species composition between Shield and off-Shield sites, I used
Indicator Value Analysis (Dufrene and Legendre 1997) to identify the indicator species
corresponding to two types of sites. This method identifies the most characteristic species of
each group by highlighting those species that are found mostly in either Shield or off-Shield sites
and present in the majority of sites belonging to that group. I then examined patterns in fish
functional traits between Shield and off-Shield sites using Correspondence Analysis (CA) on
each of the five frequency matrices for temperature preferences, trophic position,
geomorphology preferences, substrate preferences, and reproductive behaviors. CA is a
multivariate, indirect gradient analysis that allows the simultaneous assessment of both sites and
variables – in this case, functional traits. Each matrix was then constrained by a single predictor
variable indicating a site’s location on- or off-Shield using Canonical Correspondence Analysis
(CCA), and then permuted (n=1000) to determine if the resulting constrained axis is significant.
Fish community composition patterns were then assessed in the context of the glacial history of
the region. In the same CA result, sites were recoded according to their location north of the
Kirkfield Outlet or south of the Kirkfield Outlet/within the maximum extent of glacial Lake
Algonquin. CA was also used on data for fish thermal preferences, as described above, and CCA
was used with a single predictor variable (location north or south of the Kirkfield Outlet) and
permuted (n=1000) to test for significant associations between thermal preferences and site
location in relation to the glacial waters.
77
To examine patterns in abiotic conditions between Shield and off-Shield sites, a jackknifed
Linear Discriminant Analysis (LDA) was used to summarize the covariation between the abiotic
variables in dataset B, and to statistically test whether the two types of sites differ in their
environmental characteristics. Using CCA, the community composition data were then
constrained by the abiotic variables to determine the most important variables in explaining fish
community composition patterns across south-central Ontario.
All analyses were performed using R (R Development Core Team 2010).
Results
Patterns in fish community composition
A CA plot of the first two axes for dataset A shows groups of sites based on their Shield or off-
Shield location, but with considerable overlap (Fig. 4-2a). Burbot (Lota lota) and Muskellunge
(Esox masquinongy) are two rare (i.e., present at <5% of sites)) species found at a small number
of off-Shield sites, and have highly positive values on CA axis 1 (Fig. 4-2b). Rosyface Shiner
(Notropis rubellus), Longnose Sucker (Catostomus catostomus), Channel Catfish (Ictalurus
punctatus) and Mimic Shiner (Notropis volucellus) are other less common species with positive
values on CA axis 1, but more associated with a small number of Shield sites. Slimy Sculpin
(Cottus cognatus), Rainbow Trout (Oncorhynchus mykiss) and Splake (Lake Trout x Brook
Trout hybrid) appear on the far end of the second axis, associated with a small number of off-
Shield sites. Both Rainbow Trout and Splake are stocked species, and represent introductions
via human vectors. Removal of these species from the analysis did not change the results, and so
were retained to portray an accurate representation of the fish community in this time period.
When constrained by a single predictor variable indicating a site’s location on- or off-Shield, the
constrained axis was significant (n=1000, p<0.001). A CA plot of the first two axes for dataset
B, which is a subset of the sites in dataset A, shows a similar pattern, and the constrained axis is
also significant (n=1000, p<0.001) (Fig. 4-3).
Off-Shield sites were significantly more species-rich (Kruskal-Wallis chi-squared = 11.8711,
df=1, p<0.001), with a median of seven species for Shield sites and 10 species for off-Shield
sites. Indicator Value Analysis supported this finding, with 10 species found to have an indicator
78
value >25% for the off-Shield group (Table 4-3). The 25% cutoff indicated species that are
present in >50% of the sites in that group (Chaves et al. 2008). Off-Shield indicator species
included, in descending order, Blacknose Dace (Rhinichthys astralutus), White Sucker
(Catostomus commersonii), Mottled Sculpin (Cottus bairdii), Brook Stickleback (Culaea
inconstans), Central Mudminnow (Umbra limi), Fathead Minnow (Pimephales promelas), Creek
Chub (Semotilus atromaculatus), Northern Redbelly Dace (Chrosomus eos), Bluntnose Minnow
(Pimephales notatus), Rock Bass (Ambloplites rupestris), Brassy Minnow (Hybognathus
hankinsoni), and introduced species Rainbow Trout (Oncorhynchus mykiss). For Shield sites,
only four species had an indicator value >25% – Creek Chub, Common Shiner (Luxilus
cornutus), Brook Trout (Salvelinus fontinalis) and Northern Redbelly Dace. Two species,
Northern Redbelly Dace and Creek Chub, had an indicator value >25% for both Shield and off-
Shield sites. This indicated that these species are not particularly associated with either Shield or
off-Shield systems, but are well-represented in both regions and, subsequently, cannot be
considered an “indicator” of either.
Patterns in fish species composition were analyzed according to species’ preferences for water
temperature, trophic group, geomorphology, substrate and reproduction (Fig. 4-4). There was no
clear visual distinction between Shield and off-Shield sites for temperature preferences, and the
constrained CCA axis was not significant (p=0.125, Fig. 4-4a). Analysis of geomorphology
preferences shows that Shield sites tend to be associated with strictly riffle or strictly pool
species, such as Northern Redbelly Dace and Common Shiner, whereas off-Shield sites are
additionally associated with species that use a variety of habitat types, such as Central
Mudminnow and Bluntnose Minnow (p<0.001, Fig. 4-4b). Shield sites were also associated with
species preferring smaller substrates, such as Northern Redbelly Dace, Creek Chub and Common
Shiner, while off-Shield sites were associated with species preferring a broader range of
substrates (p<0.001, Fig. 4-4c). For trophic group variables, there was not a clear distinction
between groups of sites on the first two CA axes, but were some Shield/off-Shield groupings on
axes 3 and 4, and the constrained CCA axis was significant (p<0.001, Fig. 4-4d). Shield sites
were associated with herbivore-invertivore species and planktivore-invertivore species, while
off-Shield sites were more associated with strict invertivores and detritivores. Lastly, Shield
sites were generally associated with species that do not guard their eggs or young, such as Lake
Chub (Couesius plumbeus), Blackchin Shiner (Notropis heterodon) and Golden Shiner
79
(Notemigonus crysoleucas), whereas off-Shield sites encompass species with a broader variety of
reproductive behaviors (p<0.001, Fig. 4-4e).
Analysis of fish community composition incorporating the glacial history of the region showed
similar patterns as when examining community composition according to location on- or off-
Shield for both dataset A (Fig. 4-5a) and B (Fig. 4-5b). Analysis of fish thermal preferences
showed a weak association of cool- and cold-water species with sites north of the Kirkfield
Outlet, but the constrained axis was non-significant (p=0.084).
80
81
Fig. 4-2. Correspondence analysis results for dataset A, where a) shows patterns among sampling sites and b)
shows associations between fish species. Green circles indicate Shield sites and blue circles indicate off-Shield
sites. Species labels are positioned at the center of each point, with minor adjustments made for ease of reading
(label key can be found in Table 4-3). The first two axes for dataset A summarize 16.4% of the variation.
82
Fig. 4-3. Correspondence analysis results for dataset B, where a) shows patterns among sampling sites and b)
shows associations between fish species. Green circles indicate Shield sites and blue circles indicate off-Shield
sites. Species labels are positioned at the center of each point, with minor adjustments made for ease of reading.
The first two axes for dataset A summarize 18.8% of the total variation.
83
Fig. 4-4. Correspondence analyses for the five functional attribute analyses: a) temperature, b) geomorphology,
c) substrate, d) trophic status, and e) reproductive behavior. Shield sites are represented as green circles, and off-
Shield sites are blue circles.
84
Fig. 4-5. Correspondence analysis for a) dataset A and b) dataset B, with sites coded according to their location
north of the Kirkfield Outlet or south of the Kirkfield Outlet/in areas covered by the maximum extent of glacial
Lake Algonquin (according to Hinch et al. 1991, Fig. 6).
85
Patterns in abiotic conditions
Linear discriminant analysis correctly classified 94.3% of sites into their respective Shield or off-
Shield group using the abiotic variables. By region, 94.1% (16/17) of Shield sites and 94.4%
(34/36) of off-Shield sites were correctly classified in a cross-validated analysis. This result
suggests these streams can be accurately classified into their geological group by their physical
and chemical traits.
Patterns in abiotic and biotic composition between Shield and off-Shield
systems
Using CCA to find the association between fish species composition and abiotic conditions, I
found that the first two CCA axes separated sites into their respective Shield and off-Shield
groups (Fig. 4-6). High conductivity, alkalinity, temperature, turbidity, meadow and cultivated
landscapes (L2), and pH were the dominant abiotic variables associated with off-Shield sites,
whereas higher velocity, woody debris, shrub marsh (L1) and, to a lesser degree, dissolved
oxygen (DO). Notably, substrate variables were not important abiotic drivers in separating
Shield and off-Shield sites.
A second CCA where the original landscape variables L1 and L2 were replaced with current
land-use data was also performed. By partialling out the variation attributed to current land use,
I found that those land-use variables only accounted for a small portion of the total variation in
fish communities (5.2%) compared to the rest of the abiotic variables (35.2%) (Table B-5,
Appendix B). This result suggests that while differences in land use are likely important in
structuring these fish communities, there are other important contributing abiotic variables.
86
Fig. 4-6. Canonical correspondence analysis (CCA) results for dataset B. Variables include conductivity
(Cond), alkalinity, turbidity (Turb), pH, undercut banks (UCBank), gradient (G), canopy cover (Cov), water
temperature (T), dissolved oxygen (DO), woody debris (WD), water velocity (V), discharge, width, depth,
instream rock cover (Rocks), marl/muck/rubble substrate (S1), rock/rubble/boulder substrate (S2),
meadow/cultivated landscape (L1), and shrub marsh landscape (L2).
87
Discussion
The characteristic geology of the Canadian Shield creates unique abiotic and biotic conditions in
lotic systems compared to off-Shield areas in south-central Ontario. My results show that Shield
and off-Shield sites have distinct physical and chemical conditions, as well as distinct fish
community composition patterns. Systems located on the Canadian Shield are characterized by
having higher current velocity, dissolved oxygen, woody debris, and discharge and lower
conductivity, alkalinity, water temperature, and pH. Analysis of fish functional traits shows an
association of fast-water species at Shield sites, preferring smaller substrates and less likely to be
nest guarders. Indicator value analysis highlights other species associations. Brook Trout, a
species intolerant of low dissolved oxygen levels and high temperatures, is strongly associated
with Shield systems, as is Common Shiner, a species often found in pools in clear, cool streams.
Other species, like Central Mudminnow and Brook Stickleback, tolerant of low dissolved
oxygen, were strongly associated with off-Shield sites. Off-Shield sites are characterized by
higher conductivity, alkalinity, temperature, pH, and meadow or cultivated landscapes. These
sites are associated with warm temperature-tolerant species such as Fathead Minnow, Central
Mudminnow and Brassy Minnow, and species like Brook Stickleback with wide pH tolerances.
Overall, the general patterns suggest that off-Shield sites are associated with broadly tolerant
species (e.g. Fathead Minnow, Central Mudminnow, Creek Chub and Northern Redbelly Dace),
which are indicator species for both types of streams and have habitat preferences that could
describe either Shield or off-Shield streams. Creek Chub is known as a broadly tolerant species,
able to live in systems with lower dissolved oxygen, but prefers small, clear and cold streams
(Scott and Crossman 1998). Northern Redbelly Dace, also widespread throughout southern and
central Ontario, but is often found in dystrophic waters on the Shield (Holm et al. 2009).
I predicted that Shield sites would be dominated by larger substrates, due to the surface
proximity of bedrock on the Shield, and this condition would potentially influence fish species
composition. However, neither of the two substrate variables was an important abiotic driver in
separating Shield and off-Shield sites, perhaps reflecting the comparable size of the systems
considered and given that substrate size is related generally to stream gradient and system size.
Analysis of fish functional traits indicates that Shield sites are more associated with species
preferring smaller substrates and, while the most important substrate variable associated with
88
Shield systems (S1) represents marl and muck substrates, this variable was not an important
characteristic of Shield systems, either in the abiotic analysis or when the environment and fish
communities were analyzed together. Overall, these results suggest that substrate composition is
not one of the most important physical differences between Shield and off-Shield streams, and
instead may be more important structuring community composition within Shield or off-Shield
sites.
Although Shield and off-Shield sites can be distinguished based on their fish assemblages,
patterns in species richness, indicator analysis and analysis of functional traits, all results suggest
that Shield species may be a subset of the total species pool whereas the off-Shield sites contain
a broader portion of the species pool. Analysis of species by their functional traits or preferences
indicates that Shield sites have significantly different fish functional attributes from off-Shield
sites, but associations with those functional attributes show that Shield systems seem to be
comprised of species covering only one or two functional traits, whereas off-Shield sites are
home to species representing a broader suite of traits. As species richness was significantly
lower in Shield sites, this smaller degree of functional diversity may in part be a function of
lower species richness. However, if there was no association with functional traits and
Shield/off-Shield habitat, then we might expect a wide diversity of traits even within the small
number of species found on the Shield, which was not the case here. In general, Shield sites are
most associated with strictly riffle or pool species that prefer smaller substrates and have non-
nest-guarding reproductive behaviors. In contrast, off-Shield communities tend to exhibit a wider
variety of preferences in every category. In addition, in most cases, there are many species
found off-Shield that are either absent or rare in Shield sites. According to the indicator species
analysis, there were no species that were unequivocally “Shield” – even those species with a
>25% indicator value still had high values for off-Shield sites. The lack of strong indicators
specific to Shield systems provides further support to the idea that Shield systems may therefore
represent a lower diversity of habitats than found off-Shield, and only the species that are best
adapted to Shield-like environments are able to successfully inhabit those systems.
These results suggest that some aspect of the abiotic or biotic environment in Shield systems has
“filtered” the regional species pool, such that only a specific subset of fish species is found in
Shield systems. One potential filter could be temperature, as it is an abiotic factor known to be
important in structuring fish communities and, in this analysis, was important variable in
89
abiotically distinguishing Shield and off-Shield systems. The post-glacial dispersal of fishes in
this region following the last ice age is thought to be tightly linked to species’ thermal tolerances.
The location of glacial Lake Algonquin, as noted in Hinch et al. (1991), is an important indicator
of fish species distributions based on thermal tolerances. During this time, the Kirkfield Outlet
flowed southeast from glacial Lake Algonquin to Lake Ontario, and connected waterways that
drained northern uplands in a north-to-south direction (Hinch et al. 1991, fig. 6). Recolonization
of waters in the north above the outlet after the recession of the glaciers was only possible for a
short period of time, as isostatic rebound altered the connection of these waterways. As the
glacial meltwaters would have been very cold, it is likely that fishes adapted to cool or cold
waters would have recolonized first. However, by the time the meltwaters warmed enough for
colonization by warm water species, the Kirkfield Outlet no longer existed, effectively restricting
colonization to northern areas. Therefore, we might expect more cool- and cold-water fishes to
be found in waters north of the Kirkfield Outlet, and warm-water species restricted to southern
Ontario. There was a weak association between prevalence of cool- and cold-water fish species
and sites north of the outlet. However, this association was not significant, suggesting either that
this relationship is not very strong, or that there have been sufficient colonizations by, or
introductions of, warm-water species into this area by the time these systems were surveyed. As
present-day drainage patterns of lotic systems in Ontario are tied to historical connections and
dispersal pathways, I also investigated the relationship between watershed and fish community
patterns. The effect of secondary watershed – which for this study, included Central Ottawa,
Eastern Georgian Bay, Lake Ontario and Niagara River, or Northern Lake Erie – revealed no
apparent influence on the observed patterns in fish communities, as there was fair representation
of the Eastern Georgian Bay watershed in both Shield and off-Shield sites.
There was no indication of an association between certain temperature preferences and Shield
and off-Shield sites, indicating that that there are warm, cool, and cold water species in both
types of sites. It should be noted that the location of the Kirkfield Outlet more or less follows the
Shield boundary, making it difficult to determine whether patterns in fish composition are due to
a site’s location in relation to the glacial lake and outlet, or to the Shield boundary. However,
there are a number of off-Shield “above outlet” sites among the “below outlet” and “Lake
Algonquin” sites in the analysis, suggesting that the patterns in fish community composition are
better explained by location on- or off-Shield. There were some associations of typically cool-
90
and cold-water species, such as Brook Trout, with Shield sites, which suggests that water
temperature may yet be an important abiotic factor in structuring Shield and off-Shield fish
communities in south-central Ontario, at least in supporting populations of particular species.
The role of temperature in structuring fish community composition in lotic systems has been
demonstrated in numerous studies, whether across broad spatial scales (e.g., Lyons 1996, Marsh-
Matthews and Matthews 2000), as a local-scale factor (e.g. Wang et al. 2003), or within a system
moving from headwaters to higher-order streams (e.g. Townsend et al. 1983). The role of
temperature in this study in regards to fish communities of the Shield is still unclear, and to
further explore this factor, it will be necessary to design an analysis specifically designed to test
this hypothesis.
Chapter III showed that across Shield and off-Shield lotic systems, many gastropod taxa were
rare in Shield systems, and lower conductivity and alkalinity levels may have contributed to
these differences. The two strongest abiotic factors separating Shield and off-Shield sites are
chemical variables, conductivity and alkalinity. This is not surprising given the geological
differences of the two areas, especially as many off-Shield areas in south-central Ontario are
almost entirely limestone. Lower conductivity levels may contribute to increased physiological
stress and energy expenditures to maintain suitable body requirements. For example, one study
found that Fathead Minnow raised in low conductivity water (105 µS/cm) had lower survival but
greater biomass than those reared in high conductivity water (502 µS/cm) (Blanksma et al.
2009). These conductivities are similar to the difference between Shield and off-Shield systems,
with Shield systems generally lower than those used in the experimental conditions. It has also
been demonstrated that acclimation to soft-water can cause a significant reduction in the
swimming ability of juvenile Rainbow Trout, where a greater cardiac output is necessary to
achieve the same amount of exercise as fish acclimated in hard-water (Dussault et al 2008).
Many factors can influence conductivity, including levels of cations such as calcium and
magnesium. Calcium, along with phosphorus, is important in the development and maintenance
of the skeletal system, where the majority of the required calcium is absorbed from the water at
the gills (Lall and Lewis-McCrea 2007). The effects of calcium have also been prominently
studied in relation to the effect of low pH and heavy metal toxicity to fishes, where it has been
shown that water calcium can ameliorate the negative effects of low pH and aluminum on the
ionoregulatory mechanisms in fishes (e.g. Wood et al. 1990, and references therein). The effects
91
of low pH on fishes is also relatively well known, and include reproductive failure, harm to gill
mucus and membranes, loss of salts and lowered capacity of hemoglobin (Matthews 1998, and
references therein). Therefore, low conductivity and pH of streams on the Shield may be key
chemical drivers, or filters, either excluding various fish species due to low tolerance to these
conditions, or to creating stressful environments that further exacerbate the effects of
competition, predation or other unfavorable interactions.
I acknowledge that differences in land-use patterns between Shield and off-Shield areas may
influence patterns in fish community composition. The nature of the Shield, with thin, nutrient-
poor soils and large expanses of exposed bedrock, effectively limits the agricultural development
in the Shield region. It is possible that the differences in fish communities between Shield and
off-Shield streams are a result of not the Shield itself, but due to differences in the amount of
agricultural, urban or other land uses within each catchment. A lack of data for catchment-scale
land-use patterns for the sampling time period prevented any in-depth analysis on this subject,
but analysis using current land-use data for these sampling sites shows that the variation
attributed solely to land-use is a very small portion of the overall variation in fish community
composition. I suggest that differences in land-use between Shield and off-Shield areas be
viewed not as a confounding factor, but instead as another abiotic factor influenced by the
geology of the Shield, like conductivity and pH, that, in turn, ultimately affects fish community
composition.
Overall, the results indicate distinct abiotic and biotic conditions between Shield and off-Shield
low-order lotic systems. Shield systems have reduced conductivity, alkalinity, pH, higher
dissolved oxygen and instream and canopy cover compared to off-Shield streams, which are
distinguished by higher temperature, turbidity and corresponding increases in all three water
chemistry variables. Shield fish communities are composed of species less tolerant to low
dissolved oxygen and high turbidity, or those preferring cool, clear waters. This suggests that
Shield systems may exhibit a more limited range, or a different range, of environmental
conditions, thereby restricting the fish species pool to a subset of what is found off-Shield in
south-central Ontario suggesting that the Canadian Shield may also be a regional environmental
filter in south-central Ontario in the hierarchy of limiting factors on local fish assemblages
(Smith and Powell 1971, Tonn 1990; Jackson et al. 2001). These findings differ from a study in
South America that found geology affected various physical and chemical parameters in
92
headwater streams, but fish communities did not differ (Esselman et al. 2006). Although barriers
(e.g. falls, rapids) may have limited historical and contemporary dispersal, thereby influencing
current distributions of fish species, Chapter III also showed distinct differences in invertebrate
communities encompassing these same two regions. As the invertebrates are not limited in their
dispersal in the same manner as fishes (e.g. many invertebrates have aerial adult forms),
dispersal limitations cannot sufficiently explain the differences in invertebrate communities
between the two regions, but rather environmental differences must account for such differences.
Similarly, although the fish species composition of some individual sites may have been
influenced by dispersal limitations, it is likely that environmental conditions account for these
broad-scale differences in fish communities between Shield and off-Shield regions. While some
species may be excluded from Shield sites for abiotic reasons, exclusions could also be due to
biotic interactions, particularly those that become more pronounced in a more stressful abiotic
environment (i.e., low conductivity and/or pH). It is likely that a combination of abiotic and
biotic factors ultimately define the distinction between Shield and off-Shield fish communities,
but it is clear that the Canadian Shield itself has a distinct effect on patterns in fish community
composition in south-central Ontario.
References
Allan, J.D. 2004. Landscapes and riverscapes: the influence of land use on stream ecosystems.
Annual Review of Ecology, Evolution, and Systematics 35: 357-384.
Angermeier, P.L. and M.R. Winston. 1998. Local vs. regional influences on local diversity in
stream fish communities of Virginia. Ecology 79: 911-927.
Ontario Ministry of Natural Resources (OMNR). 1979. Aquatic Habitat Inventory Surveys:
Manual of Instructions. Ontario Ministry of Natural Resources – Fisheries Branch.
Queen’s Printer for Ontario, Ontario, Canada. Updated and reprinted in 1987.
Aquatic Landscape Information System (ALIS) dataset. 2008. Land Information Ontario,
managed by Ontario Ministry of Natural Resources, Peterborough, Ontario, Canada.
93
Blanksma, C., B. Eguia, K. Lott, J.M. Lazorchak, M.E. Smith, M. Wratschko, T.D. Dawson, C.
Elonen, M. Kahl, and H.L. Schoenfuss. 2009. Effects of water hardness on skeletal
development and growth in juvenile fathead minnows. Aquaculture 286: 226-232.
Bowlby, J.N. and J.C. Roff. 1986. Trophic structure in southern Ontario streams. Ecology 67:
1670-1679.
Bowman, M.F., K.M. Somers, R.A. Reid and L.D. Scott. 2006. Temporal response of stream
benthic macroinvertebrate communities to the synergistic effects of anthropogenic
acidification and natural drought events. Freshwater Biology 51: 768-782.
Buisson, L., L. Blanc and G. Grenouillet. 2008. Modeling stream fish species distribution in a
river network: the relative effects of temperature versus physical factors. Ecology of
Freshwater Fish 17: 244-257.
Chapman, L. J., and D. F. Putnam. 1984. The physiography of Southern Ontario. Special
Volume 2. Ontario Geological Survey, Sudbury, Ontario.
Chaves, M.L., M. Rieradevall, P. Chainho, J.L. Costa, M.J. Costa, and N. Prat. 2008.
Macroinvertebrate communities of non-glacial high altitude intermittent streams.
Freshwater Biology 53: 55-76.
Coker, G.A., C.B. Portt, and C.K. Minns. 2001. Morphological and ecological characteristics of
Canadian freshwater fishes. Canadian Manuscript Report of Fisheries and Aquatic
Sciences 2554: iv – 89.
Condit, R. N. Pitman, E.G. Leigh Jr., J. Chave, J. Terborgh, R.B. Foster, P. Nunez V., S. Aguilar,
R. Valencia, G. Villa, H.C. Muller-Landau, E. Losos, and S.P. Hubbell. 2002. Beta-
diversity in tropical forest trees. Science 295: 666-669.
Dow, C.L., D.B. Arscott and J.D. Newbold. 2006. Relating major ions and nutrients to
watershed conditions across a mixed-use, water-supply watershed. Journal of the North
American Benthological Society 25: 887-911.
Dufrene, M. and P. Legendre. 1997. Species assemblages and indicator species: the need for a
flexible asymmetrical approach. Ecological Monographs 67: 345-366.
94
Dussault, E.B., R.C. Playle, D.G. Dixon, R.S. McKinley. 2008. Effects of soft-water
acclimation on the physiology, swimming performance, and cardiac parameters of the
rainbow trout, Oncorhynchus mykiss. Fish Physiology and Biochemistry 34: 313-322.
Eakins, R. J. (2010) Ontario Freshwater Fishes Life History Database. Version 3.88. On-line
database. (http://www.fishdb.ca), accessed 21 October 2010.
Esselman, P.C., M.C. Freeman, and C.M. Pringle. 2006. Fish-assemblage variation between
geologically defined regions and across a longitudinal gradient in the Monkey River
Basin, Belize. Journal of the North American Benthological Society 25: 142-156.
Goldstein, R.M. and M.R. Meador. 2004. Comparisons of fish species traits from small streams
to large rivers. Transactions of the American Fisheries Society 133: 971-983.
Hall, R.J. and F.P. Ide. 1987. Evidence of acidification effects on stream insect communities in
central Ontario between 1937 and 1985. Canadian Journal of Fisheries and Aquatic
Sciences 44: 1652-1657.
Hinch, S.G., N.C. Collins, and H.H. Harvey. 1991. Relative abundance of littoral zone fishes:
biotic interactions, abiotic factors, and postglacial colonization. Ecology 72: 1314-1324.
Hoeinghaus, D.J., K.O. Winemiller, and J.S. Birnbaum. 2007. Local and regional determinants
of stream fish assemblage structure: inferences based on taxonomic vs. functional groups.
Journal of Biogeography 34: 324-338.
Jackson, D.A. and H.H. Harvey. 1989. Biogeographic associations in fish assemblages: local
vs. regional processes. Ecology 70: 1472-1484.
Jackson, D.A., K.M. Somers, and H.H. Harvey. 1989. Similarity coefficients: Measures of co-
occurrence and association or simply measures of occurrence? The American Naturalist
133: 436-453.
Jackson, DA. 1997. Compositional data in community ecology: The paradigm or peril of
proportions? Ecology 78: 929-940
95
Jackson, D.A., P.R. Peres-Neto and J.D. Olden. 2001. What controls who is where in freshwater
fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of
Fisheries and Aquatic Sciences 58: 157-170.
Kelso, J.R.M., M.A. Shaw, C.K. Minns and K.H. Mills. 1990. An evaluation of the effects of
atmospheric acidic deposition on fish and the fishery resource of Canada. Canadian
Journal of Fisheries and Aquatic Sciences 47: 644-655.
Kilgour, B.W. and D.R. Barton. 1999. Associations between stream fish and benthos across
environmental gradients across southern Ontario, Canada. Freshwater Biology 41: 553-
566.
Kratzer, E.B., J.K. Jackson, D.B. Arscott, A.K. Aufdenkampe, C.L. Dow, L.A. Kaplan, J.D.
Newbold and B.W. Sweeney. 2006. Macroinvertebrate distribution in relation to land
use and water chemistry in New York City drinking-water-supply watersheds. Journal of
the North American Benthological Society 25: 954-976.
Lall, S.P. and L.M. Lewis-McCrea. 2007. Role of nutrients in skeletal metabolism and
pathology in fish – An overview. Aquaculture 267: 3-19.
Layman, C.A. and K.O. Winemiller. 2004. Size-based responses of prey to piscivore exclusion
in a species-rich neotropical river. Ecology 85: 1311-1320.
Lyons, J. 1996. Patterns in the species composition of fish assemblages among Wisconsin
streams. Environmental Biology of Fishes 45: 329-341.
McAuliffe, J.R. 1984. Competition for space, disturbance and the structure of a benthic stream
community. Ecology 65: 894-908.
Mandrak, N.E. 1995. Biogeographic patterns of fish species richness in Ontario lakes in relation
to historical and environmental factors. Canadian Journal of Fisheries and Aquatic
Sciences 52: 1462-1474.
Marsh-Matthews, E. and W.J. Matthews. 2000. Geographic, terrestrial and aquatic factors:
which most influence the structure of stream fish assemblages in the midwestern United
States? Ecology of Freshwater Fish 9: 9-21.
96
Matthews, W.J. 1998. Patterns in Freshwater Fish Ecology. Chapman & Hall, New York,
U.S.A.
Minns, C.K. 1989. Factors affecting fish species richness in Ontario lakes. Transactions of the
American Fisheries Society 118: 533-545.
Ontario Geological Survey. 2003. Surficial Geology of Southern Ontario – (including Toronto);
Ontario Geological Survey, Miscellaneous Release Data 128.
R Development Core Team. 2010. R 2.10.1. A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria.
Ricklefs, R.E. 1987. Community diversity: relative roles of local and regional processes.
Science 235: 167-171.
Smith, C.L. and C.R. Powell. 1971. The summer fish communities of Brier Creek, Marshall
County, Oklahoma. American Museum Novitiates: 2458.
Tonn, W.M. 1990. Climate change and fish communities: a conceptual framework.
Transactions of the American Fisheries Society 119: 337-352.
Tonn, W.M., J.J. Magnuson, M. Rask and J. Toivonen. 1990. Intercontinental comparison of
small-lake fish assemblages: the balance between local and regional processes. The
American Naturalist 136: 345-375.
Townsend, C.R., A.G. Hildrew and J. Francis. 1983. Community structure in some southern
English streams: the influence of physiochemical factors. Freshwater Biology 13: 521-
544.
Wang, L., J. Lyons, P. Rasmussen, P. Seelbach, T. Simon, M. Wiley, P. Kanehl, E. Baker, S.
Niemela, and P.M. Stewart. 2003. Watershed, reach and riparian influences on stream
fish assemblages in the Northern Lakes and Forest Ecoregion, U.S.A. Canadian Journal
of Fisheries and Aquatic Sciences 60: 491-505.
Wood, C.M., D.G. McDonald, C.G. Ingersoll, D.R. Mount, O.E. Johannsson, S. Landsberger,
and H.L. Bergman. 1990. Effects of water acidity, calcium and aluminum on whole
97
body ions of brook trout (Salvelinus fontinalis) continuously exposed from fertilization to
swim-up: a study by instrumental neutron activation analysis. Canadian Journal of
Fisheries and Aquatic Sciences 47: 1593-1603.
98
Chapter V
Regional-scale patterns in community concordance: the influence of the Precambrian Shield on the structure of lotic
assemblages
Abstract
The determinants of a local-scale biological community can include both abiotic and biotic
factors acting at a variety of spatial and temporal scales. For example, factors such as historical
processes, regional climate, physical habitat and competition may all act simultaneously to
produce an assemblage of species capable of coexisting at a particular location. Previous studies
have showed that a broad-scale geological feature, the Precambrian Shield, leads to distinct
abiotic conditions in lotic systems as compared to similar systems off-Shield. In addition, Shield
systems are characterized by distinct fish and macroinvertebrate assemblages. In this study, I use
contemporary fish and macroinvertebrate community data from south-central Ontario to examine
the degree of community concordance between fishes and macroinvertebrates, and to determine
the main abiotic drivers for this biological distinction between Shield and off-Shield areas. This
study clearly shows that low-order lotic systems on the Shield differ both abiotically and
biotically from adjacent off-Shield systems, and that similar abiotic drivers are responsible for
structuring both groups of taxa. However, there is some indication that aspects of water
chemistry and land-use are most important in structuring macroinvertebrate communities, while
physical habitat plays a larger role for fish communities.
Introduction
Local species composition of biological communities can depend on a variety of factors,
including the abiotic environment, interspecific interactions, historical colonization patterns,
spatial proximity of habitats and the size of the species pool (Chase 2003). The relative
importance of such factors in determining species composition and richness in a local habitat
depends on the scale at which those factors operate. A popular paradigm in ecology depicts such
factors as a series of filters through which the total regional species pool passes in order to obtain
99
a local assemblage (Smith and Powell 1971, Tonn 1990, Poff 1997, Jackson et al. 2001). While
many early investigations in community ecology focused on local-scale factors, in more recent
years studies on the relative effects of local and regional processes on the overall local
assemblage have been conducted in a variety of systems, including terrestrial plant communities
(e.g., Eskelinen and Virtanen 2005 ), marine communities (e.g., Hughes et al. 1999), bird
communities (e.g., Robinson et al. 2000), mammal communities (e.g., Williams et al. 2002,
Hortal et al. 2008), fish communities (e.g., Angermeier and Winston 1998, Marsh-Matthews and
Matthews 2000, Wang et al. 2003) and aquatic invertebrate communities (e.g., Heino et al. 2003,
Sandin and Johnson 2004, Mykra et al. 2007, Chapter III).
In aquatic systems, it has been shown that often regional factors play a large role in determining
local species assemblage patterns. Regional factors are important in determining diversity,
composition and abundance of fish species, where regional fish diversity can predict local
diversity (Angermeier and Winston 1998), and where two systems with distinct regional and
historical influences, but similar local characteristics, can have similar species richness but
different assemblage patterns (Tonn et al. 1990). Accordingly, it has been shown that reach-
scale characteristics in lotic systems are influenced by broader watershed characteristics that, in
turn, affect distribution and abundance of fish species (Wang et al. 2003). Similar results have
been found for macroinvertebrate communities (Poff 1997, Malmqvist 2002), where the
importance of local-scale variables decreased with increasing spatial scale of a study (Mykra et
al. 2007), and where regional invertebrate richness is the most important predictor of local
richness in some systems (Heino et al. 2003). Beche and Statzner (2009) concluded that
environmental filtering of invertebrate functional traits was likely responsible for the redundancy
in stream invertebrate communities across the USA, and found that there was no distinguishable
difference in the importance of local vs. watershed-scale variables – instead, a combination of
both was responsible for structuring the community. While both regional and local factors are
ultimately responsible for structuring communities, regional variables do indeed have a strong
influence on the local assemblages of both fish and macroinvertebrates.
The geological composition of a region can greatly influence the physical and chemical attributes
of lakes, streams and rivers, and thus have an important impact on local species assemblages
(e.g. Chapter III). In North America, the Precambrian Shield is one such geological feature,
encompassing much of Canada and parts of northern United States in the Midwest and New
100
England. The Shield, as it is familiarly known, is characterized by ancient metamorphic and
igneous bedrock lying at or close to the surface, having been exposed during the glacial retreat at
the end of the last ice age in North America (Chapman and Putnam 1984). In south-central
Ontario, the Shield landscape is covered by only a thin layer of nutrient-poor soils and stands in
sharp contrast to the younger, primarily sedimentary bedrock of southern Ontario and most of
remaining North America. Differences in the weathering characteristics of these two geological
types create distinct physical and chemical attributes in aquatic systems – Shield rock weathers
primarily by freeze-thaw fracturing, whereas off-Shield limestone rock weathers via chemical
dissolution in water (Chapman and Putnam 1984). Slow weathering rates of Shield bedrock
result in poor buffering capacity, and Shield waters tend to be characterized by low
conductivities, a propensity for acidification, and calcium depletion (see Edwards et al. 2008,
Jeziorski et al. 2008). Proximity of the bedrock to the surface may also affect the physical
structure of aquatic systems, particularly streams and rivers, where erosional resistance of the
underlying bedrock may result in poor pool development (Wohl and Legleiter 2003).
Various abiotic variables can influence fish community composition and structure, including
dissolved oxygen levels, acidity, water velocity, stream morphology, structural complexity of the
habitat (Jackson et al. 2001) and patterns in land use (Allan et al. 1997). Similar variables have
been found to influence lotic macroinvertebrate communities, including physical habitat
structure (e.g. Allan 1975, Erman and Erman 1984, Bourassa and Morin 1995) and water
chemistry (e.g. McKillop and Harrison 1972, Minshall and Minshall 1978, Huryn et al. 1995).
In Chapters III and IV, I showed that, at a large scale, lotic systems on the Precambrian Shield do
not only exhibit distinct physical and chemical characteristics, but also have distinct fish and
macroinvertebrate communities compared to streams of similar size found in adjacent off-Shield
sites in south-central Ontario. Shield systems are generally characterized by low conductivity
and alkalinity levels, and are faster flowing systems with higher dissolved oxygen and denser
canopy cover than off-Shield counterparts. Macroinvertebrate communities are characterized by
taxa preferring erosional habitats, while depauperate in taxa such as Gastropoda (particularly
Pulmonata), Isopoda and Amphipoda (Chapter III). Fish communities of the Shield appear to be
a subset of species found off-Shield, where Shield systems are characterized by species that
prefer either strictly pools or riffles and exhibiting non-nest guarding reproductive behaviors. In
contrast, off-Shield communities tend to exhibit a wider variety of preferences (Chapter IV).
101
Examining community patterns within two very different groups of taxa allows us to broaden the
scope of interpretation and generality of ecological understanding. By comparing fish and
invertebrate community relationships, we can better understand the effects of the Canadian
Shield on aquatic taxa. For example, is the change in conductivity from Shield to off-Shield
systems the major abiotic driver determining aquatic species composition, or is it a physical
factor such as substratum characteristics or stream geomorphology? As well, do fishes and
macroinvertebrates respond to the same abiotic drivers, and in the same ways? Evidence from
other studies suggests that there is community concordance between different groups of aquatic
taxa, although the results vary depending on spatial and temporal scale, as well as the abiotic
factors in question (Jackson and Harvey 1993, Kilgour and Barton 1999, Heino 2002, Paavola et
al. 2006, Bowman et al. 2008). In some cases, there was concordance between fish and
invertebrate communities, but the different groups responded to different environmental drivers
(Jackson and Harvey 1993). In other cases, distributions of fish and macroinvertebrate species
were correlated to similar environmental variables, such as Kilgour and Barton (1999)
demonstrated when examining streams across natural environmental gradients in south-central
Ontario. Other studies have highlighted the importance of spatial and temporal scale in the
degree of concordance (Paavola et al 2006, Bowman et al 2008), where Paavola et al. (2006)
suggested that the community composition of both fishes and invertebrates is controlled by
similar regional-scale environmental factors, but there is weaker concordance when examining
communities at a local scale.
One of the major ecological differences between fishes and macroinvertebrates lies in their
dispersal abilities. Many macroinvertebrate taxa have an aerial life-history stage, whereby adults
emerge from aquatic systems with the ability to fly overland to new locations (Bilton et al. 2001,
Bohonak and Jenkins 2003). In contrast, fish species are limited to aquatic pathways for their
dispersal, and can face barriers such as waterfalls and dams. Historical limits to dispersal can
also play a role in present-day species distributions. Previous studies have shown that regional
patterns in fish species distributions in Ontario have been determined, in part, by the location of
post-glacial colonization routes into the province (Jackson and Harvey 1989, Mandrak 1995),
and it has been suggested that species’ distributions are likely influenced by dispersal barriers
during the glacial retreat following the last glaciation, when only cold water species were able to
colonize waters at the edge of the melting ice sheets (Hinch et al. 1991). By the time these
102
waters in the region warmed to a suitable degree for warm water fishes, aquatic pathways to
upland areas of the province had been disconnected due to isostatic rebound of the landscape and
subsequent formation of modern-day watersheds. Thus, both historical and current dispersal
limitations could play a large role in determining which fish species are found where throughout
the province, and observations concerning the influence of the Precambrian Shield on fish
communities could be confounded with these dispersal limitations. However, if
macroinvertebrate communities -- which should not show the same dispersal limitations that
require aquatic connections – also show distinct assemblages in Shield and off-Shield systems, it
supports the hypothesis that the abiotic conditions imposed in Shield lotic systems are important
factors affecting the community composition of both fish and macroinvertebrate communities,
and that patterns seen in fish communities are not merely an artifact of dispersal limitation.
In Chapters III and IV, historical data from a broad region in south-central Ontario were used to
examine regional patterns in abiotic conditions, fish communities and macroinvertebrate
communities, specifically in relation to the Precambrian Shield. In this study, I examine Shield
and off-Shield systems in a smaller geographic range to better understand the effects of the
Precambrian Shield on the physical, chemical and biological composition of low-order lotic
systems. I have three main objectives: 1) Identify the distinctions in physical and chemical
properties between Shield and off-Shield lotic systems; 2) Identify the differences in fish and
macroinvertebrate community composition between the two types of systems using taxonomic
and functional traits, and identify the major abiotic drivers affecting community composition in
these systems; and, 3) Examine the degree of concordance between fish and macroinvertebrates,
and identify if these two groups of taxa respond to the same abiotic drivers.
I predict similar abiotic differences in Shield and off-Shield lotic systems as reported in Chapters
III and IV, where Shield systems will be fast-water streams lower in conductivity and pH
compared to off-Shield systems. For sites located at or near the Shield/off-Shield boundary, I
expect intermediate conditions, reflecting the interface of two types of geological composition. I
expect to find fewer fish species in Shield systems, due to a more limited range in available
habitats, and these species will prefer faster waters and larger substrates, and these systems to be
comprised of species tolerant of low pH. I expect to see similar patterns in macroinvertebrate
taxa, with fewer taxa found in Shield systems, and generally depauperate in acid-intolerant
and/or low-calcium intolerant taxa. I also expect dominance of taxa preferring faster (erosional)
103
habitats and larger substrates, while expecting a wider variety of tolerances, preferences and
habits in off-Shield systems. Lastly, previous studies (e.g., Chapters III and IV) indicate that fish
and invertebrate communities of Shield and off-Shield systems are influenced by similar abiotic
drivers; therefore, these communities may show a strong degree of concordance with each other.
However, I predict that chemical drivers may be more important in structuring macroinvertebrate
communities. For example, calcium levels (correlated with conductivity data) may be more
important in structuring invertebrate taxa, as organisms such as mollusks and bivalves may be
the most sensitive to this variable. In contrast, the effects of low conductivity may not directly
limit certain fish species, but may be an indirect factor by inducing physiological stress that
exacerbates other interactions such as competition or predation.
Methods
In the summer of 2007 and 2008, 1st-3rd order lotic systems located on- and off-Shield were
sampled to assess fish and macroinvertebrate community composition as well as physical and
chemical attributes. Sampling was conducted within a limited geographic range extending from
~44.3 to 45.9°N and ~ -78.5 to –79.5°W in south-central Ontario (Fig. 5-1). In this area, the
bedrock geology is either Shield or off-Shield, with a strong delineation between the two regions.
The latitudinal and longitudinal restriction aimed to reduce the effects of climatic differences
between sites. To select sites, systems within this area that met the following criteria were
considered for sampling: 1) between 1st and 3rd stream order inclusively; 2) relatively
unimpacted by anthropogenic influences; and, 3) on land locally accessible from roads or
pathways. Sites were classified as “Shield” or “off-Shield” using a GIS layer of bedrock geology
(Ontario Geological Survey 2003). Some sites classified as “Shield” using the GIS data had
unusually high conductance and/or were very close to off-Shield geology, where there was
potential for the true site location to be off-Shield, or on a combination of Shield and off-Shield
geology. These sites were tentatively deemed to be “transition” sites, and were examined more
closely in the subsequent analyses. In July-August 2007, 35 sites were sampled (34 Shield, one
transition), and in June-August 2008, 16 off-Shield sites, four transition sites and an additional
seven Shield sites were sampled, resulting in a total of 64 sites. Due to water chemistry meter
equipment malfunctions and absence of fishes, seven sites were removed from the data set to
result in 36 Shield, five transition and 16 off-Shield sites. Table C-1 (Appendix C) lists all sites
and their associated bedrock geology).
104
Fig. 5-1. The 57 sites sampled in 2007-2008 in south-central Ontario. Green circles indicate sites located on the
Canadian Shield, blue circles indicate sites located off-Shield, and pink circles indicate “transition” sites. The
heavy dark line denotes the Shield boundary in this area of the province.
105
Abiotic sampling
Each site was assessed for a set of physical and chemical parameters. Site length was first
determined using a standard method for both Shield and off-Shield systems, which was 10 times
the average wetted width of the stream. As it was hypothesized that Shield systems may not
adhere to the same pattern in meander wavelengths as off-Shield systems, this method allowed
standardization across sites while still accounting for minor differences in stream size. Thus, this
method incorporates the idea that as stream size increases, the effort required to adequately
sample that system should increase proportionally, and that the use of a fixed site length could
result in under-representation of certain habitat features in larger systems (Flotemersch et al.
2011). The relationships between site area (m2) and species/taxa richness are displayed in Fig.
C-1 (Appendix C), and did not differ from random for either fish (linear regression, p=0.5704) or
macroinvertebrates (linear regression, p=0.8297).
Dissolved oxygen, pH, conductivity and water temperature were measured three times within the
sampling reach and then averaged. Physical variables measured included percent canopy cover
(visual estimate, confirmed by two individuals), stream gradient, mean water velocity (measured
three times in both pool and riffle segment), substrate composition and depth. Each site was
divided into a series of transects, following the methodology outlined in the Ontario Stream
Assessment Protocol (OSAP 2005). Along each transect, substrate composition was assessed by
both taking a random sample as well as an overall visual assessment at three, five or six points,
depending on the length of the transect. The presence of moss, woody debris, algae and aquatic
macrophytes was also assessed at each point along the transect. Presence or absence of these
features at each point along the transect was then used to calculate a proportional measure of
abundance – for example, if moss was found at five point measurements along a transect with six
points, it received a value of 0.83 for that transect.
The resulting abiotic data set thus contained the following variables: pH, conductivity (Cond),
percent stream gradient (G), current velocity (V), dissolved oxygen (DO), water temperature (T),
moss (M), woody debris (WD), algae (AL), aquatic macrophytes (WV), and depth. As substrate
composition record at each site represents percent composition variables that are not independent
of each other (Jackson 1997), Correspondence Analysis (CA) was used to summarize the
variation across the different substrate categories into axes summarizing the dominant trends.
106
The resulting site scores for the first three axes were used as new variables, resulting in three
substrate variables, S1, S2 and S3. Lastly, land-use data for the upstream catchment area for
each site were collected using the Aquatic Landscape Inventory Software (ALIS). Percent
upstream catchment area of 28 land-use categories was assessed for each sampling site, and then
CA was used to summarize the variation across the difference categories in the same manner as
the substrate variables, resulting in two summary land-use variables, ALIS1 and ALIS2. A
summary on the interpretation of substrate and land use variables is provided in Table 5-1.
Variation in climatic attributes can impact biotic assemblages in broad regional studies such as
this one. While this aspect of the abiotic environment was not explicitly assessed, care was taken
to minimize climatic variation within the study region, as the off-Shield to Shield transition lies
along a latitudinal gradient. A cursory examination of average yearly temperatures revealed
minimal latitudinal variation (i.e., ~1 °C), in addition to longitudinal variation, likely attributed
to the influence of Lake Huron (Environment Canada 2011).
107
108
Fish sampling
Fish community composition was sampled at each site using a single pass with a backpack
electrofisher. Block nets were set up at both ends of the site to prevent escapement during
sampling. Fishes were identified on site and returned to the stream after processing; reference
samples were taken to confirm identifications in the lab using taxonomic keys in Hubbs and
Lagler 2004. As the conductivity of Shield systems is typically quite low and may influence the
effectiveness of electrofishing as a method of capture, three-pass depletion surveys were
conducted on representative Shield sites to determine the efficiency of the single-pass protocol in
collecting species susceptible to electrofishing (Fig. C-2, Appendix C).
Macroinvertebrate sampling
Macroinvertebrate community composition was sampled at each site with two kick-net samples
from both a riffle and pool habitat. If there was no distinct riffle or pool, one sample was taken
from a faster section of the site and the other from a slower section. Sampling was done along a
transect starting from the water’s edge and out into the center of the stream for one minute.
Samples were preserved in the field in 75% ethanol, and brought to the lab for sorting and
identification to family. Pool and riffle samples were pooled for each site prior to analysis. The
entire sample was sorted and identified for each site, as species depletion analyses indicated that
a representative sub-sample of all species present in the sample could not be obtained using a
portion of the total sample (Fig. C-3, Appendix C). Therefore, all organisms in each sample
were identified and counted. Identifications to family were conducted in-lab using various
taxonomic keys, including Clarke (1981), Pecarksky et. al (1990), Thorp and Covich (1991) and
Merrit and Cummins (1998).
A summary of the biotic and abiotic data for this study is found in Tables C-2, C-3 and C-4
(Appendix C).
Statistical Analysis
To examine abiotic patterns in Shield and off-Shield sites, I used Principal Components Analysis
(PCA) to summarize the variation in physical and chemical variables. I used cross-validated
Linear Discriminant Analysis (LDA) to classify sites into two groups based on their abiotic
109
environment, and then predicted group membership for the five transition sites. Abiotic data
were standardized to zero mean and unit variance prior to all analyses.
To examine general patterns in fish and macroinvertebrate community composition for Shield
and off-Shield sites, I used CA on log-transformed abundance data. I then used Canonical
Correspondence Analysis (CCA) to test if Shield and off-Shield sites differ significantly based
on their community composition. This analysis constrains the CA result using a single predictor
variable indicating a site’s location on- or off-Shield, and then performs a permutation test
(n=1000) to determine if the group membership in resulting constrained axis differs from
random. Differences in fish and macroinvertebrate taxa richness between Shield and off-Shield
sites was assessed using a Kruskal-Wallis test. Lastly, fish and macroinvertebrate community
data were constrained by abiotic variables using CCA to determine the most important factors in
explaining fish community patterns in these sites. I used partial CCA to examine the relative
influence of land-use, water chemistry and physical habitat variables on the overall explained
variation.
I examined patterns in fish functional traits using a combination of resources to classify fish
species according to five categories: temperature preferences, trophic status, stream
geomorphology preferences, substrate preferences, and reproduction methods/behaviors. I used
descriptions of species’ functional traits as described in Goldstein and Meador (2003), and
classifications listed in the Ontario Online Fish Database (Eakins 2010, http://www.fishdb.ca)
(for further explanation on the classification of fish species’ functional traits, refer to Table C-5,
Appendix C). This exercise resulted in five frequency matrices that were analyzed using CA to
examine patterns in these traits between Shield and off-Shield sites. Each matrix was then
constrained by a single predictor variable indicating a site’s location on- or off-Shield analyzed
using CCA, and permuted (n=1000) to determine if the resulting constrained axis is significant.
Due to the greater number of macroinvertebrate taxa collected in this analysis and corresponding
functional categories, a similar analysis of macroinvertebrate functional traits is not statistically
practical. However, I applied knowledge of functional traits and preference from Thorp and
Covich (1991) and Merritt and Cummins (1996) to make connections between aquatic habitats
and the representative macroinvertebrate communities.
110
To assess community concordance, I used Procrustes analysis (Jackson 1995; Peres-Neto and
Jackson 2001) on the fish and macroinvertebrate CA results, which rotates an ordination result to
produce the maximum similarity with another ordination result. Significance was then assessed
using the permutation-based PROTEST function (n=1000).
All analyses were performed using R (R Development Core Team 2010).
Results
Patterns in abiotic conditions across Shield and off-Shield lotic systems
The first PCA axis summarized 22.7% of the total variation and shows a clear separation of
Shield and off-Shield sites (Fig. 5-2). Off-Shield sites were characterized by high conductivity,
pH, depth, clay and detrital substrates (S1), dissolved oxygen and aquatic macrophytes. Shield
sites were characterized by an abundance of moss, high gradient, denser canopy cover and
bedrock and detritus substrates (S3). The second PCA axis increased the variation summarized
to 37.2% and further characterized many Shield sites as having abundant woody debris.
There was 100% correct classification of Shield and off-Shield sites using cross-validated Linear
Discriminant Analysis, and this result was used to predict group membership (Shield or off-
Shield) for the five transition sites. Four sites were classified as “Shield” and one site, Shadow
(site 41), was classified as “off-Shield.” This result was consistent with PCA, where Shadow did
fall among the off-Shield sites in the PCA plot of the first two axes. It should be noted, however,
that the other four transition sites, while being classified as Shield with LDA, did fall at the
interface of Shield and off-Shield points in the PCA plot.
111
Fig. 5-2. Principal component analysis of abiotic variables standardized to zero mean and unit variance. Green
circles indicate Shield sites, pink circles indicate transition sites and blue circles indicate off-Shield sites.
112
Patterns in fish communities
The CA on log-abundance fish data showed a degree of separation of Shield and off-Shield sites
primarily along the first two axes (Fig. 5-3a). Constraint by a single predictor variable of a site’s
location on- or off-Shield indicated that sites could be grouped according to their geological
location based on their fish community composition with confidence (p=0.005). The analysis
showed species such as Creek Chub (Semotilus atromaculatus), Common Shiner (Luxilus
cornutus), and White Sucker (Catostomus commersonii), were associated predominately with
Shield sites, along with less common species like Brook Trout (Salvelinus fontinalis), Golden
Shiner (Notemigonus crysoleucas), Largemouth Bass (Micropterus salmoides), Blacknose Dace
(Rhinichthys obtusus), Brown Bullhead (Ameiurus nebulosus) and Smallmouth Bass (M.
dolomieu) (Fig. 5-3b, see Table C-6, Appendix C for species labels). Off-Shield sites had higher
abundances and/or occurrences of Central Mudminnow (Umbra limi), Brook Stickleback
(Culaea inconstans), Blackchin Shiner (Notropis heterodon) and Iowa Darter (Etheostoma exile),
as well as rare species such as Rainbow Darter (Etheostoma caeruleum) and the predominately
lake-dwelling Spottail Shiner (Notropis hudsonius).
Analysis of fish functional traits showed significant results for trophic position (pseudo-F =
0.3423, p = 0.0007), substrate preferences (pseudo-F = 0.5604, p = 0.0013) and reproductive
behaviors (pseudo-F = 0.1534, p = 0.011), suggesting that certain fish functional traits are
significantly associated with Shield or off-Shield sites. Off-Shield sites had species with lower
trophic level, including planktivore/invertivore/detritivore combinations, species preferring
smaller substrates and vegetated habitats, and species that exhibit complex nesting behaviors.
Shield sites were conversely associated with more invertivore-carnivores, species preferring
sand, gravel, cobble and boulder substrates without vegetation, and species exhibiting non-nest
guarding behaviors (either brood hiders or open substrate nesters). Analysis of geomorphology
preferences and water temperature preferences were not significant (pseudo-F = 0.102486, p =
0.08 and pseudo-F = 0.060848, p = 0.1207, respectively).
113
Fig. 5-3. Correspondence analysis of fish community data, showing positions of a) sites and b) species. Species
labels are positioned at the center of each point, with minor adjustments made for ease of reading. A guide to
species labels can be found in Table C-6 (Appendix C).
114
Patterns in macroinvertebrate communities
The CA of macroinvertebrate log-transformed abundance data showed a clear separation of
Shield and off-Shield sites using the first two axes (Fig. 5-4a). Constraint by a single predictor
variable of a site’s location on- or off-Shield indicated that sites could be grouped according to
their geological location based on their macroinvertebrate community composition with
confidence (p<0.001). The analysis showed that there were a number of taxa that were located
near the origin of the plot, such as Chironomidae, Simuliidae and Tipulidae (Diptera),
Hydropsychidae (Trichoptera), Elmidae (Coleoptera), Oligeochaeta and Bivalvia, indicating that
these taxa are widely distributed and common in both groups of sites (Fig. 5-4b, see Table C-6,
Appendix C for taxa labels). Shield sites were associated with taxa such as Aeshnidae,
Cordulegastridae, and Gomphidae (Odonata), Baetiscidae, Ephemerellidae, Heptegeniidae and
Leptophlebidae (Ephemeroptera), Glossosomatidae, Leptoceridae, Philopotamidae,
Odontoceridae, Polycentropidae and Rhyacophilidae (Trichoptera), Sialidae and Corydalidae
(Megaloptera) and Leuctridae, Nemouridae and Perlidae (Plecoptera). Off-Shield sites were
associated with Amphipoda, Asellidae (Isopoda), Baetidae and Caenidae (Ephemeroptera),
Corixidae (Hemiptera), Haliplidae, Gyrinidae, Hydrophilidae and Dytiscidae (Coleoptera),
Libellulidae and Lestidae (Odonata), Sciomyzidae and Tabanidae (Diptera), Hirudinea, and
Lymnaeidae, Physidae, Planoboridae and Valvatidae (Gastropoda).
115
Fig. 5-4. Correspondence analysis of macroinvertebrate community data, showing positions of a) sites and b)
taxa. Taxa labels are positioned at the center of each point, with minor adjustments made for ease of reading. A
guide to taxa labels can be found in Table C-6 (Appendix C).
116
Abiotic and community associations
Shield and off-Shield fish communities were generally separated using the first two CCA axes
(Fig. 5-5a), where off-Shield communities tended to be associated with clay and detritus
substrates (S1), high conductivity, water vegetation, pH, depth and, to a lesser degree, wetlands
and/or alvar landscapes (ALIS2). Shield fish communities were primarily associated increased
density of canopy cover, treed bogs, unknown, settlement, water and/or recent cutover land use
(ALIS1), steeper gradient, higher dissolved oxygen, in-stream moss, and instream woody debris.
Sites were more clearly separated in a plot of the first two CCA axes of macroinvertebrate
community composition constrained by abiotic variables (Fig. 5-6a). Off-Shield
macroinvertebrate communities were primarily associated with high conductivity, pH, small
substrates (clay, detritus, sand-silt), bedrock and clay substrates, algae, wetlands and/or alvars
(ALIS2), and depth. Shield communities were associated with land-use variable ALIS1 (greater
treed bogs, settlements, water bodies and/or recent cutovers), woody debris, higher gradient, in-
stream moss, water velocity and temperature.
Partial CCA was used to assess the relative contribution of land-use, water chemistry and
physical habitat variables to the overall explained variation in the constrained analyses. CCA
summarized 40.3% of the variation in fish communities, with 4% attributed to water chemistry
variables (pH and conductivity), 3.2% to land-use variables (ALIS1 and ALIS2), 29.9% to the
remaining physical habitat variables, and 3.2% shared variation. For macroinvertebrates, CCA
summarized 43.3% of the total variation in macroinvertebrate communities, with 3.5% attributed
to water chemistry variables, 2.6% to land-use variables, 29.7% to the remaining physical habitat
variables, and 7.5% shared variation. Plots of the various partial CCAs (i.e., water chemistry
removed, land use removed, and physical habitat variation removed) revealed that the Shield/off-
Shield separation of sites in the CCA plot could be attributed to variation in all three types of
abiotic factors for fish communities (Figs. 5-5b and 5-5d). However, for macroinvertebrate
communities, the distinction between Shield and off-Shield points was only maintained when
physical variables are partialled out (Fig. 5-6d), as the Shield/off-Shield distinction broke down
when either land use or water chemistry was removed (Figs. 5-6b and 5-6c, respectively).
117
Fig 5-5. Full (a) and partial (b-d) canonical correspondence analyses (CCA) of fish community data, using
abiotic variables standardized to zero mean and unit variance. Abiotic variables include pH, conductivity
(Cond), percent gradient (G), velocity (V), dissolved oxygen (DO), water temperature (T), moss (M), woody
debris (WD), algae (AL), water vegetation (WV), depth, substrate (S1, S2, S3) and land use (ALIS1, ALIS2).
Full and partial CCA analyses were used to determine which variables were important in distinguishing Shield
and off-Shield fish communities, where b) shows a partial CCA with variation attributed to land use variables
removed, c) partial CCA with water chemistry variables removed, and d) partial CCA with remaining physical
habitat variables removed.
118
Fig. 5-6. Full (a) and partial (b-d) canonical correspondence analyses (CCA) of macroinvertebrate community
data, using abiotic variables standardized to zero mean and unit variance. Abiotic variables include pH,
conductivity (Cond), percent gradient (G), velocity (V), dissolved oxygen (DO), water temperature (T), moss
(M), woody debris (WD), algae (AL), water vegetation (WV), depth, substrate (S1, S2, S3) and land use (ALIS1,
ALIS2). Full and partial CCA analyses were used to determine which variables were important in distinguishing
Shield and off-Shield macroinvertebrate communities, where b) shows a partial CCA with variation attributed to
land use variables removed, c) partial CCA with water chemistry variables removed, and d) partial CCA with
remaining physical habitat variables removed.
119
Community concordance and abiotic associations
Species richness significantly differed between Shield and off-Shield sites for macroinvertebrate
communities (Shield family richness mean = 15.8 and median = 17; off-Shield family richness
mean = 19.7 and median = 20, Kruskal-Wallis chi-square = 8.7007, p = 0.003), but did not differ
significantly for fish communities (Shield species richness mean = 3.6 and median = 3; off-
Shield species richness mean = 4.1 and median = 4.5; Kruskal-Wallis chi-square = 1.7653, p =
0.184). Procrustes analysis between CA ordinations of fish and macroinvertebrate abundances
showed that the two ordinations are significantly associated (m2=0.771, p<0.001). Analysis of
residuals showed that sites with unusual fish communities had the largest residuals -- i.e., the
greatest difference in site position between fish and macroinvertebrate CA results (Fig. 5-7). The
size of residuals for Shield and off-Shield sites were significantly different, with off-Shield sites
having more discordant fish and macroinvertebrate communities (Mann-Whitney test, p=0.022).
This pattern appears to be driven by four sites – Ten Mile (Shield, site 1), Kingfisher (off-Shield,
site 37), Waiman (off-Shield, site 37) and Beaverton (off-Shield, site 48) – which had very large
residuals compared to all other sites. Procrustes residuals were positively correlated with pH
(r2=0.1533), conductivity (r2=0.07), and algae (r2=0.1291), while negatively correlated with
ALIS1 (r2=0.1645), canopy cover (r2=0.1423), gradient (r2=0.0828), and woody debris
(r2=0.1256).
While not showing strong patterns in regards to Procrustes residuals, there were several sites that
exhibited patterns in either abiotic or biotic conditions that blurred the distinction between a
Shield and off-Shield site. Queen’s Line (Shield, site 34) exhibited abiotic conditions typical of
other Shield sites and had a representative fish community, but had more off-Shield-like
macroinvertebrate composition. Compared to other Shield sites, this site was less associated
with high dissolved oxygen and had a greater proportion of smaller substrates. Hawkers (site 57),
Rutherford (site 45) and Silver (site 49) are off-Shield sites that exhibited typical abiotic
characteristics and macroinvertebrate communities, but had a more Shield-like fish community.
These sites had higher canopy cover and were dominated by larger substrates compared to
average values for other off-Shield sites. These sites were more similar to off-Shield due to their
higher conductivity and pH compared to Shield sites, as these were two of the most important
120
variables in the abiotic analysis. However, other variables like canopy cover, substrate size and
woody debris may lead to more Shield-like fish communities at these sites.
Abiotically, all transition sites except for Shadow (site 41) were classified as Shield. Shadow
had higher conductivity and higher abundance of clay and detrital substrates than other Shield
sites. The other four transition sites were Shield-like in all variables except for having higher
conductivity values than other Shield sites. Transition sites Baker-Beech (site 35) and South
Kashe (site 62), classified as Shield using abiotic data, exhibited Shield fish and
macroinvertebrate communities, while Shadow (site 41), classified as off-Shield using abiotic
data, exhibited communities at the interface of Shield and off-Shield compositional patterns.
South Beaver (site 51) and Burnt 2 (site 56), also abiotically classified as Shield, were less
biotically defined. South Beaver exhibited a Shield-like fish community, due to high abundances
of species such as Blacknose Dace, Creek Chub and Golden Shiner, but was associated with a
more off-Shield macroinvertebrate community. This site was characterized by many off-Shield
abiotic attributes, including higher pH and conductivity compared to other Shield sites.
Concordantly, the invertebrate community reflected this chemical environment, being comprised
of nearly all off-Shield taxa including Amphipoda, Hirudinea, Oligochaeta, Planoboridae and
Viviparidae. However, it also contained low abundances of Shield-like taxa, including
Coenagrionidae, Empididae, Gomphidae and Leuctridae.
Lastly, three sites (Burnt 1, 2 and 3) are located on different tributaries of the same river system.
The most southern of these sites, Burnt 3 (site 63), is the closest one to the Shield boundary and
is abiotically classified as a Shield site. However, it was associated with more off-Shield biota,
such as low abundances of Isopoda, Amphipoda, Hydracharina and Gastropoda, as well as Brook
Stickleback and Central Mudminnow. Compared to the average values for other Shield streams,
it had a lower gradient and dissolved oxygen, but much higher values for substrate variables
which included detritus (i.e., S1 and S3). Burnt 2 (site 56), which is nearby, had a similar fish
and invertebrate community, and was classified as a Shield site but had higher conductivity than
on average. Burnt 1 (site 55), which is the furthest from the Shield/off-Shield interface,
exhibited Shield abiotic and biotic characteristics.
121
Fig. 5-7. Magnitude of Procrustes residuals for each site, grouped by geology. Numbers at the top of each
column indicate site number (see Table C-1, Appendix C).
122
Discussion
It is clear that low-order lotic systems on the Canadian Shield differ abiotically and biotically
from adjacent off-Shield systems in south-central Ontario. The major differences in abiotic
conditions coincide with differences in the community composition of both fish and
macroinvertebrates, such that there is community concordance between the two groups of taxa.
It appears that overall, similar abiotic drivers are responsible for structuring both groups of taxa.
Shield fish and macroinvertebrate communities were both associated with dense canopy cover,
treed bog landscapes, higher gradient and dissolved oxygen concentrations, and greater amounts
of instream moss and woody debris compared to off-Shield systems. These systems were also
lower in conductivity, pH, and depth, were less associated with aquatic macrophytes, and were
dominated by larger substrates (boulder, cobble, and gravel). The particular taxa found at Shield
sites exhibited the expected suite of functional traits. For example, fish species associated with
Shield systems tended to prefer gravel, cobble and boulder substrates, whereas off-Shield
systems were associated with species preferring smaller substrates and water vegetation. I
predicted that Shield fish communities would represent a subset of the off-Shield community,
including only those species that are able to live in the presumably more limited physical and
chemical habitat of Shield systems. However, analysis of patterns in fish species distributions
found no difference in species richness between Shield and off-Shield sites, and there were a
number of species that were exclusively found in Shield streams. This finding is in contrast with
results from Chapter IV, where it was suggested that Shield fish communities could be
considered a subset of the total species pool. In this analysis, assessment of this pattern may be
confounded by the appearance of these Shield-exclusive species. In most cases, such as with
Grass Pickerel, these species occurrences may represent isolated populations that otherwise are
rare in south-central Ontario, whether on- or off-Shield (Mandrak and Crossman 1992a). In
addition, other species such as Brown Bullhead, Emerald Shiner, Largemouth Bass, Rock Bass,
Smallmouth Bass and Yellow Perch are not only also known to occur off-Shield locations
(Mandrak and Crossman 1992a), but also occurred in very low abundance in this data set.
However, overall there is no evidence in this analysis to support the hypothesis that Shield fish
communities represent a subset of the total regional species pool, and that Shield fish
communities tend to exhibit a narrower range of functional characteristics.
123
For invertebrates, off-Shield systems were significantly more taxa-rich than Shield systems.
Community associations show that off-Shield systems were generally characterized by a greater
abundance of taxa with potentially greater cation requirements, such as Isopoda, Gastropoda, and
the Diptera Sciomyzidae (which have calcium carbonate crystals in their integument), or those
sensitive to low pH such as Amphipoda (Glazier et al. 1991 and references therein). In addition,
these systems were associated with swimming crawling predators (e.g., Coleoptera and
Hemiptera) and other taxa preferring depositional and vegetated habitats. Shield systems were
much lower in abundance of Amphipoda, Isopoda and Gastropoda, and were instead
characterized by crawling or non-swimming predators such as two families of Polycentropidae
and Rhyacophilidae (Trichoptera), Perlidae (Plecoptera) and both families of Megaloptera.
In this study, fishes and macroinvertebrates were sampled from the same locations at the same
time, allowing an examination of the degree of community concordance between these two
groups of taxa. This approach has three main advantages for the purposes of examining the role
of the Canadian Shield in structuring lotic communities in south-central Ontario. First, an
assessment can be made as to whether fish and macroinvertebrate composition respond in the
same way to the abiotic factors influenced by the Shield – in other words, for two sites with
similar fish communities, do they also have similar macroinvertebrate communities? Secondly, if
there is concordance between the two groups of taxa, this suggests that any potential limitation in
dispersal of fishes is not likely the main driver in distinctions between Shield and off-Shield
systems. Lastly, this analysis can also highlight those sites where there might be a mismatch
between fish and macroinvertebrate communities. This provides an opportunity to relate abiotic
conditions at that site back to the different communities, providing further insight into the
important abiotic drivers for the two groups of taxa.
Evidence from other studies suggests that there is community concordance between different
groups of aquatic taxa, although the results vary depending on spatial and temporal scale, as well
as the abiotic factors in question (Jackson and Harvey 1993, Kilgour and Barton 1999, Heino
2002, Paavola et al. 2006, and Bowman et al. 2008). In this study, fish and macroinvertebrate
communities were significantly concordant with each other, such that sites with similar fish
communities had similar macroinvertebrate communities. This result is similar to what was
found in Kilgour and Barton (1999) and, overall, these results suggest that fish and
macroinvertebrate communities are responding to similar abiotic drivers. These results further
124
support the hypothesis that the Shield, via its influence on various physical and chemical
components of lotic systems, is an important large-scale feature that affects lotic community
composition. The patterns seen here are not consistent with the hypothesis of potential dispersal
limitations or historical factors such as post-glacial recolonization routes (Mandrak 1995) being
the dominant factors structuring these communities. Although I was not setting out to explicitly
test the relative influence of regional and local factors, this work provides additional support to
earlier studies showing the importance of regional factors in structuring biological communities
(e.g., Tonn et al. 1990, Angermier and Winston 1998, Heino et al. 2003, Mykra et al. 2007).
While there was significant concordance between fish and macroinvertebrate communities,
examination of the Procrustes residuals shows that as the abiotic environment moves to more off-
Shield-like conditions, there is a greater mismatch between fish and macroinvertebrate
communities. However, this pattern appears to be driven by four sites – Ten Mile (Shield, site
1), Kingfisher (off-Shield, site 37), Waiman (off-Shield, site 37) and Beaverton (off-Shield, site
48) – which had very large residuals compared to all other sites. There is not one variable or
attribute in common with these four sites, except that they all have unusual fish communities
compared to other sites that have similar macroinvertebrate communities. For example, sites
Ten Mile, Beaverton and Waiman were generally more species-rich than other sites, and tended
to contain a greater number of less common species, such as Slimy Sculpin (Cottus cognatus)
and Spottail Shiner. As Ten Mile is in close proximity to a large lake, it is possible that the
presence of rare species is the result predominately lake-dwelling species moving into lotic areas.
However, there was no evidence of a relationship between site proximity to lakes or large
wetlands and residual size. In contrast, Kingfisher was relatively species-poor, lacking common
species such as Common Shiner and Creek Chub, but had another rare species, Iowa Darter.
Sites with abiotic or biotic conditions that blur the distinction between Shield and off-Shield sites
offer further insight into the driving abiotic factors for fish and macroinvertebrate taxa in these
systems. While fish and macroinvertebrate communities are, for the most part, responding to the
same abiotic drivers in distinguishing Shield and off-Shield systems, it appears that the
macroinvertebrate taxa most responsible for the Shield/off-Shield distinction are those that
appear limited to certain chemical conditions. In contrast, fish communities seem to be
responding more to differences in physical habitat rather than to chemical conditions. For
example, sites that are chemically off-Shield-like but physically Shield-like (e.g., Hawkers,
125
Rutherford, South Beaver) have macroinvertebrate communities similar to other off-Shield sites,
due mainly to an increase in Isopoda, Amphipoda and Gastropoda families, but have fish
communities more similar to Shield sites. This result is consistent with the results reported in
Jackson and Harvey (1993), where significant concordance was observed between fish and
invertebrate communities in lakes in central Ontario, but where the fish communities were
structured by morphological characteristics of lakes, and the invertebrate communities by
chemical characteristics.
Aspects of water chemistry can limit distributions of some taxa, such Gastropoda (McKillop and
Harrison 1972, Chapter III) and Amphipoda (Glazier et al. 1992, Chapter III). As evidenced by
some transition and off-Shield sites described previously, some off-Shield sites (e.g., higher
conductivity and pH) with physical attributes similar to Shield sites (e.g., larger substrates) have
macroinvertebrate communities that contain taxa characteristic of both off-Shield and Shield
sites. This pattern is complemented by results from the patterns in Shield and off-Shield
macroinvertebrate communities when variation attributed to water chemistry, land use and
physical habitat is examined separately. Removing variation attributed to land use or water
chemistry resulted in no distinction between Shield and off-Shield sites in their respective CCA
plot, suggesting that these two groups of variables are important drivers in distinguishing the
macroinvertebrate communities of these two geological regions. When variation attributed to
physical habitat is removed and only water chemistry and land-use variation are included, the
Shield/off-Shield distinction remains, thereby indicating that these latter variables are the more
important types of structuring factors. However, when only water chemistry or land-use
variables are removed, the distinction again breaks down, further highlighting the importance of
both sets of variables and their covariation, rather than just land-use or water chemistry alone.
However, differences in Shield and off-Shield fish communities appear to be more tightly linked
to physical habitat preferences, as there is a shift in community composition between Shield and
off-Shield sites, rather than a combination of Shield and off-Shield species at sites with off-
Shield chemistry and Shield physical habitat. For example, sites with a greater proportion of
larger substrates – even when also high in conductivity and pH –were associated with species
preferring larger substrates. When examining the role of water chemistry, physical habitat and
land-use simultaneously, there is little change in the distinction of Shield and off-Shield sites
when variation attributed to any group of variables is removed, suggesting that unlike
126
macroinvertebrate communities, all three types of factors play a role in distinguishing Shield and
off-Shield fish communities. It has been argued that while parameters such as conductivity and
pH can impact stream fishes (e.g., Wood et al. 1990, Matthews 1998, Dussault et al. 2008), the
effects may be indirect also, where low conductivity and pH can negatively impact fish behavior
or interactions with other species, but may not directly lead to absences of species except for in
extreme conditions (Chapter IV). For example, species of darters were very rare in Shield
systems, both in this study and historically (Chapter IV), and distribution maps indicate that
several species are mostly limited to off-Shield areas in Ontario (Mandrak and Crossman 1992a).
This could be due to historical factors, such as pathways of post-glacial recolonization (Mandrak
and Crossman 1992b), but there may also be other limiting factors such as unfavorable water
chemistry which exacerbates any effects of predation or competition with other Shield species.
However, further study is necessary to address this hypothesis.
In this study, I consider the geological composition of south-central Ontario to be an “umbrella”
abiotic factor directly influencing water chemistry, land-use patterns and physical habitat of lotic
systems. Other studies have sought to separate the relative effects of geology and land-use
patterns on aquatic communities (e.g., Kratzer et al. 2006, Dow et al. 2006), and generally
conclude that it is difficult at best. Geology often plays a large role in determining land use
(Allan et al. 1997). For example, the prominent bedrock outcroppings and nutrient-poor soil of
the Shield effectively limits agricultural development, but encourages other practices such as
mining and logging. Land-use was an important aspect in the distinction of Shield and off-
Shield systems, and was an important set of variables in determining both fish and
macroinvertebrate communities. The goal of this study, however, was not to separate the effects
of land use from that of geology on aquatic communities, but to examine the Shield overall as a
potential regional driver of the total species pool. In other words, the geological composition of
the Shield influences water chemistry, physical habitat composition in aquatic systems and land-
use patterns in the surrounding catchments that, in turn, create either favorable or unfavorable
conditions for various species. For macroinvertebrates, I found that patterns in water chemistry
and land use were most important in distinguishing Shield and off-Shield communities, whereas
fish communities were distinguished by combinations of all three. While overall the two groups
of taxa were generally concordant with one another, similar to results presented by Kilgour and
Barton (1999), there was evidence that different abiotic drivers are potentially more influential to
127
one group over the other, similar to a previous studies on lake systems (Jackson and Harvey
1989, 1993). However, concordance between these two groups of taxa supports the idea that the
distinctions between Shield and off-Shield fish communities are not likely not due to dispersal
limitation. If this had been the case, I would have expected differences in the respective patterns
of fish and invertebrate communities given their differing dispersal potential. However, the
results are consistent with the hypothesis of differences in abiotic factors imposed by the
geological distinction between Shield and off-Shield. While previous studies have shown that
historical processes can play a role in determining fish species distributions in the province of
Ontario (Jackson and Harvey 1989, Hinch et al. 1991, Mandrak 1995), these results suggest that,
as a whole, the Precambrian Shield is another important broad-scale influence on the physical
and chemical habitat of lotic systems and their biological communities. While historical factors
no doubt influence the present-day community composition, their influence appears less
dominant than that imposed by the environmental conditions distinguishing Shield from off-
Shield sites.
Consideration of the relative influence of regional and local factors in structuring biological
communities continues to be a relevant point of interest among community ecologists. A
number of papers within the aquatic literature have stressed the importance of regional-level
factors in structuring local fish and macroinvertebrate communities (e.g., Jackson and Harvey
1989, Angermeier and Winston 1998, Wang et al. 2003, Heino et al. 2003). This study lends
further support to this idea, where a broad-scale geological feature leads to distinct physical,
chemical and landscape characteristics in aquatic systems. However, it is important to note that
while the abiotic conditions created by the Shield lead to distinct biological communities across
the landscape of south-central Ontario, there is still potential for variation in abiotic conditions
within these two groups of systems. This variation in local conditions can also have important
implications for the types of biota found at specific locations, and the range of conditions within
a region can have important implications for regional species richness. For example, if a region
is comprised of a set of nearly uniform habitats, that region is likely to have a smaller species
pool than where there is a range of abiotic conditions, all within the context of the broader-scale
filters represented by the region. Anecdotal information regarding the types of systems found on
the Canadian Shield suggests that there is some variation in physical and chemical conditions
among different lotic systems – for example, some Shield streams are slow and boggy, with dark
128
tea-colored water, while others are fast, clear and run over exposed bedrock. Further study will
be required to examine the local abiotic variation in Shield lotic systems in relation to biotic
communities. However, it has been shown here that, at a broad scale, the Shield influences
regional abiotic characteristics of lotic systems such that we see distinct fish and
macroinvertebrate communities from adjacent systems off-Shield, further solidifying the
importance of large-scale factors on biological communities.
References
Allan, J.D. 1975. The distributional ecology and diversity of benthic insects in Cement Creek,
Colorado. Ecology 56: 1040-1053.
Allan, J.D., D.L. Erickson and J. Fay. 1997. The influence of catchment land use on stream
integrity across multiple spatial scales. Freshwater Biology 37: 149-161.
Angermeier, P.L. and M.R. Winston. 1998. Local vs. regional influences on local diversity in
stream fish communities of Virginia. Ecology 79: 911-927.
Aquatic Landscape Information System (ALIS) dataset. 2008. Land Information Ontario,
managed by Ontario Ministry of Natural Resources, Peterborough, Ontario, Canada.
Beche, L.A. and B. Statzner. 2009. Richness gradients of stream invertebrates across the USA:
taxonomy- and trait-based approaches. Biodiversity and Conservation 18: 3909-3930.
Bilton, D.T., J.R. Freeland and B. Okamura. 2001. Dispersal in freshwater invertebrates.
Annual Review of Ecology, Evolution, and Systematics 32: 159-181
Bohonak, A.J. and D.G. Jenkins. 2003. Ecological and evolutionary significance of dispersal by
freshwater invertebrates. Ecology Letters 6: 783-796.
Bourassa, N. and A. Morin. 1995. Relationships between size structure of invertebrate
assemblages and trophy and substrate composition in streams. Journal of the North
American Benthological Society 14: 393-403.
Bowman, M.F., R. Ingram, R.A. Reid, K.M. Somers, N.D. Yan, A.M. Paterson, G.E. Morgan
and J.M. Gunn. 2008. Temporal and spatial concordance in community composition of
129
phytoplankton, zooplankton, macroinvertebrate, crayfish, and fish on the Precambrian
Shield. Canadian Journal of Fisheries and Aquatic Sciences 65: 919-932.
Chase, J.M. 2003. Community assembly: when should history matter? Oecologia 136: 489-
498.
Chaves, M.L., M. Rieradevall, P. Chainho, J.L. Costa, M.J. Costa, and N. Prat. 2008.
Macroinvertebrate communities of non-glacial high altitude intermittent streams.
Freshwater Biology 53: 55-76.
Clarke, A.H. 1981. The Freshwater Molluscs of Canada. National Museum of Natural
Sciences, National Museum of Canada, Ottawa, Canada.
Dow, C.L., D.B. Arscott and J.D. Newbold. 2006. Relating major ions and nutrients to
watershed conditions across a mixed-use, water-supply watershed. Journal of the North
American Benthological Society 25: 887-911.
Dussault, E.B., R.C. Playle, D.G. Dixon, R.S. McKinley. 2008. Effects of soft-water acclimation
on the physiology, swimming performance, and cardiac parameters of the rainbow trout,
Oncorhynchus mykiss. Fish Physiology and Biochemistry 34: 313-322.
Eakins, R.J. 2010. Ontario Freshwater Fishes Life History Database. Version 3.91. On-line
database (http://www.fishdb.ca).
Edwards, B.A., D.A. Jackson and K.M. Somers. 2009. Multispecies crayfish declines in lakes:
implications for species distributions and richness. Journal of the North American
Benthological Society 28: 719-732.
Environment Canada. 2011. Canadian Climate Normals or Averages, 1971-2000. National
Climate Data and Information Archive. Online at
http://www.climate.weatheroffice.gc.ca.
Erman, D.C. and N.A. Erman. 1984. The response of stream macroinvertebrates to substrate size
and heterogeneity. Hydrobiologia 108: 75-82.
130
Eskelinen, A. and R. Virtanen. 2005. Local and regional processes in low-productive mountain
plant communities: the roles of seed and microsite limitation in relation to grazing.
Oikos 110: 360-368.
Glazier, D.S., M.T. Horne, and M.E. Lehman. 1992. Abundance, body composition and
reproductive output of Gammarus minus (Crustacea: Amphipoda) in ten cold springs
differing in pH and ionic content. Freshwater Biology 28: 149-163.
Goldstein, R.M. and M.R. Meador. 2004. Comparisons of fish species traits from small streams
to large rivers. Transactions of the American Fisheries Society 133: 971-983.
Heino, J. 2002. Concordance of species richness patterns among multiple freshwater taxa: a
regional perspective. Biodiversity and Conservation 11: 137-147.
Heino, J., T. Muotka, and R. Paavola. 2003. Determinants of macroinvertebrate diversity in
headwater streams: regional and local influences. Journal of Animal Ecology 72: 425-
434.
Hinch, S.G., N.C. Collins and H.H. Harvey. 1991. Relative abundance of littoral zone fishes:
biotic interactions, abiotic factors, and postglacial colonization. Ecology 72: 1314-1324.
Hortal, J., J. Rodriguez, M. Nieto-Diaz and J.M. Lobo. 2008. Regional and environmental
effects on the species richness of mammal assemblages. Journal of Biogeography 35:
1202-1214.
Hughes, T.P., A.H. Baird, E.A. Dinsdale, N.A. Moltschaniwskyj, M.S. Pratchett, J.E. Tanner and
B.L. Willis. 1999. Patterns of recruitment and abundance of corals along the Great
Barrier Reef. Nature 397: 59-63.
Huryn, A.D., A.C. Benke and G.M. Ward. 1995. Direct and indirect effects of geology on the
distribution, biomass and production of the freshwater snail Elimia. Journal of the North
American Benthological Society 14: 519-534.
Jackson, D.A. 1995. PROTEST: A Procrustean randomization test of community environment
concordance. Ecoscience 2:297-303.
131
Jackson, D.A. and H.H. Harvey. 1989. Biogeographic associations in fish assemblages: local
vs. regional processes. Ecology 70: 1472-1484.
Jackson, D.A. and H.H. Harvey. 1993. Fish and benthic invertebrates: community concordance
and community-environment relationships. Canadian Journal of Fisheries and Aquatic
Sciences 50: 2641-2650.
Jackson, D.A., P.R. Peres-Neto and J.D. Olden. 2001. What controls who is where in freshwater
fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of
Fisheries and Aquatic Sciences 58: 157-170.
Jeziorski, A., N.D. Yan, A.M. Paterson, A.M. DeSellas, M.A. Turner, D.S. Jeffries, B. Keller,
R.C. Weeber, D.K. McNicol, M.E. Palmer, K. McIver, K. Arsenaeau, B.K. Ginn, B.F.
Cumming and J.P. Smol. 2008. The widespread threat of calcium decline in fresh waters.
Science 322: 1374-1377.
Kilgour, B.W. and D.R. Barton. 1999. Associations between stream fish and benthos across
environmental gradients across southern Ontario, Canada. Freshwater Biology 41: 553-
566.
Kratzer, E.B., J.K. Jackson, D.B. Arscott, A.K. Aufdenkampe, C.L. Dow, L.A. Kaplan, J.D.
Newbold and B.W. Sweeney. Macroinvertebrate distribution in relation to land use and
water chemistry in New York City drinking-water-supply watersheds. Journal of the
North American Benthological Society 25: 954-976.
Malmqvist, B. 2002. Aquatic invertebrates in riverine landscapes. Freshwater Biology 47: 679-
694.
Mandrak, N.E. and E.J. Crossman. 1992a. A checklist of Ontario freshwater fishes. Royal
Ontario Museum, Toronto, Canada.
Mandrak, N.E. and E.J. Crossman. 1992b. Postglacial dispersal of freshwater fishes into
Ontario. Canadian Journal of Zoology 70: 2247-2259.
132
Mandrak, N.E. 1995. Biogeographic patterns of fish species richness in Ontario lakes in relation
to historical and environmental factors. Canadian Journal of Fisheries and Aquatic
Sciences 52: 1462-1474.
Marsh-Matthews, E. and W.J. Matthews. 2000. Geographic, terrestrial and aquatic factors:
which most influence the structure of stream fish assemblages in the midwestern United
States? Ecology of Freshwater Fish 9: 9-21.
Matthews, W.J. 1998. Patterns in Freshwater Fish Ecology. Chapman & Hall, New York,
U.S.A.
McKillop, W.B. and A.D. Harrison. 1972. Distribution of aquatic gastropods across an interface
between the Canadian Shield and limestone formations. Canadian Journal of Zoology
50: 1433-1445.
Merritt, R.W. and K.W. Cummins, eds. 1996. An introduction to the aquatic insects of North
America, 3rd edition. Kendall Hunt Publishing Company, Iowa, U.S.A.
Minshall, G.W. and J.N. Minshall. 1978. Further evidence on the role of chemical factors in
determining the distribution of benthic invertebrates in the River Duddon. Arch.
Hydriobiol. 53: 324-355.
Mykra, H., J. Heino and T. Muotka. 2007. Scale-related patterns in the spatial and
environmental components of stream macroinvertebrate assemblage variation. Global
Ecology and Biogeography 16: 149-159.
Ontario Geological Survey 2003. Surficial Geology of Southern Ontario – (including Toronto);
Ontario Geological Survey, Miscellaneous Release Data 128.
Paavola, R., T. Muotka, R. Virtanen, J. Heino, D. Jackson, A. Maki-Petays. 2006. Spatial scale
affects community concordance among fishes, benthic macroinvertebrates, and
bryophytes in streams. Ecological Applications 16: 368-379.
133
Peckarsky, B.L., P. Fraissinet, M.A. Pention, and D.J. Conklin, Jr. 1990. Freshwater
Macroinvertebrates of Northeastern North America. Cornell University Press, Ithaca,
New York.
Peres-Neto, P.R. and D.A. Jackson. 2001. How well do multivariate data sets match? The
advantages of a Procrustean superimposition approach over the Mantel test. Oecologia
129:169-178.
Poff, N.L. 1997. Landscape filters and species traits: Towards mechanistic understanding and
prediction in stream ecology. Journal of the North American Benthological Society 16:
391-409.
R Development Core Team. 2010. R 2.10.1. A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria.
Robinson, W.D., J.D. Brawn and S.K. Robinson. 2000. Forest bird community structure in
central Panama: influence of spatial scale and biogeography. Ecological Monographs 70:
209-235.
Sandin, L. and R.K. Johnson. 2004. Local, landscape and regional factors structuring benthic
macroinvertebrate assemblages in Swedish streams. Landscape Ecology 19: 501-514.
Smith, C.L. and C.R. Powell. 1971. The summer fish communities of Brier Creek, Marshall
County, Oklahoma. American Museum Novitates: 2458.
Stanfield, L., ed. 2005. Ontario Stream Assessment Protocol (OSAP), v.7.
Thorp, J.H. and A.P. Covich, eds. 1991. Ecology and classification of North American
freshwater invertebrates. Academic Press Inc. San Diego, California.
Tonn, W.M. 1990. Climate change and fish communities: a conceptual framework.
Transactions of the American Fisheries Society 119: 337-352.
Tonn, W.M., J.J. Magnuson, M. Rask and J. Toivonen. 1990. Intercontinental comparison of
small-lake fish assemblages: the balance between local and regional processes. The
American Naturalist 136: 345-375.
134
Wang, L., J. Lyons, P. Rasmussen, P. Seelbach, T. Simon, M. Wiley, P. Kanehl, E. Baker, S.
Niemela, and P.M. Stewart. 2003. Watershed, reach and riparian influences on stream
fish assemblages in the Northern Lakes and Forest Ecoregion, U.S.A. Canadian Journal
of Fisheries and Aquatic Sciences 60: 491-505.
Williams, S.E., H. Marsh and J. Winter. 2002. Spatial scale, species diversity, and habitat
structure: small mammals in Australian tropical rain forest. Ecology 83: 1317-1329.
Wohl, E. and C.J. Legleiter. 2003. Controls on pool characteristics along a resistant-boundary
channel. Journal of Geology 111: 103-114.
Wood, C.M., D.G. McDonald, C.G. Ingersoll, D.R. Mount, O.E. Johannsson, S. Landsberger,
and H.L. Bergman. 1990. Effects of water acidity, calcium and aluminum on whole
body ions of brook trout (Salvelinus fontinalis) continuously exposed from fertilization to
swim-up: a study by instrumental neutron activation analysis. Canadian Journal of
Fisheries and Aquatic Sciences 47: 1593-1603.
135
Chapter VI
General Discussion
In this thesis, I first examined patterns of integration among the various subfields of aquatic
science by examining study subject and citation patterns within five prominent aquatic journals.
This chapter showed that while various subfields in aquatic ecology tend not to utilize the
findings of other subfields to further their research, those studies that do refer to fields outside
their specific area of research tend to address research questions having a broader view of
aquatic ecology. In the other chapters of my thesis, I have attempted to integrate analyses of fish
communities, macroinvertebrate communities, and elements of habitat in order to examine the
influence of a broad-scale geological feature on lotic systems. Using historical data, I was able
to first address this goal at a broader geographical scale (Chapters III and IV), and then follow
this with a similar examination at a finer scale (Chapter IV). Results from these chapters show
that there are distinct differences in both abiotic and biotic conditions of systems located on the
Canadian Shield compared to those found off-Shield. Furthermore, Chapter IV suggests that,
overall, fish and macroinvertebrate communities are responding to similar abiotic drivers.
Ecology is a broad field encompassing many different subjects and research questions, but with
the overall aim being to understand the properties and dynamics of ecosystems, whether at the
population, community, or ecosystem level. In community ecology, there has been a noticeable
shift towards viewing local communities with a wider lens, where aspects of that community are
examined at various spatial and temporal scales in order to best describe biotic patterns. As
more and more is learned about the driving abiotic and biotic factors that structure local
communities, it becomes apparent that nearly every research question must be first put into the
context of scale.
Community ecology research throughout the past 40 years has increasingly stressed the relative
effects of local and regional processes on the structure and composition of local biological
communities (e.g., Smith and Powell 1971, Ricklefs 1987, Tonn 1990, Jackson et al. 2001).
Particularly in lotic systems, past studies have focused primarily on the influence of local factors
136
– for example, many studies on aspects of aquatic macroinvertebrate communities have focused
on factors such as substrate (e.g., Williams 1980, Erman and Erman 1984, Bourassa and Morin
1995), predation (e.g., Holomuzki and Short 1988) and competition (e.g., McAuliffe 1984).
Following key papers, such as Ricklefs (1987), that urged ecologists to consider the context of
spatial and temporal scale in ecology, research began to highlight the importance of regional and
historical factors on aquatic communities (e.g., Jackson and Harvey 1989, Mandrak 1995,
Johnson et al. 2007, Heino et al. 2007, Montana and Winemiller 2010). For lotic systems, it has
been suggested that it is particularly important to recognize the influence of these broad-scale
factors, particularly when interpreting how these systems react to disturbances (Resh et al. 1988).
This thesis further contributes to our knowledge on the role of broad-scale factors in structuring
aquatic communities by examining the influence of the Precambrian Shield on lotic systems in
south-central Ontario. Analysis of this broad question, applied to both historical and
contemporary fish and macroinvertebrate communities, shows that the Shield is an important
regional-scale geological feature that creates abiotic conditions distinct from lotic systems found
off-Shield. While acknowledging that other broad-scale processes, such as historical factors,
may influence patterns in species distributions (e.g., Hinch et al. 1991, Mandrak 1995), my thesis
shows that aquatic systems on the Shield differ in water chemistry, physical habitat, and land-use
characteristics, such that the Shield may act as a regional-scale factor leading to particular abiotic
conditions and, thereby, biotic communities suited to those conditions. Different aspects of the
abiotic environment may be more or less important for the different groups of taxa investigated,
but, overall, the significant degree of concordance in macroinvertebrate and fish communities
indicates that these two groups are responding similarly to the abiotic differences between the
two regions.
The ability to examine patterns in two groups of taxa sampled at the same locations and at the
same time in Chapter V further expands on our knowledge of the relative influence of historical
and contemporary barriers to dispersal in this area. In Chapter IV, it was proposed that patterns
in fish species distributions may be influenced by historical barriers to dispersal following glacial
retreat in this region. The results of that study indicated that, while these historical barriers have
likely influenced fish species distributions in some aspects as was shown by Hinch et al. (1991)
and Mandrak (1995), it appears that the influence of the Shield on abiotic conditions in these
systems also plays a part in the observed patterns. Chapter V further supports this hypothesis, as
137
there was significant concordance between fish and macroinvertebrate communities across all
sites. As macroinvertebrates are not likely to be influenced by the same barriers to dispersal as
fishes due to aerial adult life stages in many taxa, this outcome suggests that any potential
limitation in the dispersal of fishes is not likely to be the main driver in distinctions between
Shield and off-Shield systems.
To incorporate broad-scale factors in studies examining local lotic communities, researchers
have begun to adopt a landscape perspective on the stream environment (e.g., Fausch et al.
2002). This perspective has in many cases lead to the incorporation of geology as a broad-scale
variable potentially influencing local factors (e.g., Richards et al. 1996, Richards et al. 1997,
Wiley et al. 1997, Wang et al. 2003, Sandin and Johnson 2004, Esselman et al. 2006, Kratzer et
al. 2006). While there have been varying results regarding the relative influence of geology on
local abiotic and biotic factors, it is clear that geology can be an important master variable to
consider in the hierarchy of the biotic “filters” proposed by Smith and Powell (1971). For south-
central Ontario, the sharp distinction between two geological regions influences aquatic abiotic
conditions strongly enough to lead to distinct biological communities. This distinction could be
useful in order to develop management and/or conservation practices that are optimally
specialized for the type of system. For the most part, lotic research in the province of Ontario
and North America in general, has been focused on temperate streams with more classical and
predictable pool-riffle morphology (e.g., Bowlby and Roff 1986, Kilgour and Barton 1999). As
this study has shown that lotic systems on the Canadian Shield are abiotically distinct from other
such systems, there is potential for specific protocols or objectives to ensure proper management.
My study examined differences in species composition between Shield and off-Shield lotic
systems, but there are other community attributes that may be affected by the abiotic differences
between the two regions. It has been shown that considering biotic communities by their
functional traits, rather than taxonomic classification, can provide additional insight into the
structuring forces of aquatic communities (e.g., Angermier and Winston 1998, Peres-Neto 2004,
Bhat 2005, Heino et al. 2007). Functional attributes of Shield and off-Shield macroinvertebrate
communities were coarsely examined in this study using known information concerning habit
and habitat preferences by various taxonomic families to make generalizations about the two
types of communities, and how they match to the abiotic environment. Functional aspects of fish
communities were more specifically examined by placing species into functional categories and
138
re-analyzing the community matrix to look for differences between Shield and off-Shield
systems. For both groups of taxa, there was an association between certain functional traits and
the abiotic environment in which they were found. However, one aspect that was not explored in
this thesis is the possibility of intraspecific variation in traits such as body morphology or
behavior. For example, there may be species found frequently in both Shield and off-Shield
systems, but with Shield populations that exhibit differences in behavior or morphology in
response to the abiotic environment in which they live. In some fish species, certain
morphometric features, such as the location of the pectoral fin in relation to other anatomical
markers, may vary within a species in relation to differences in its abiotic or biotic habitat (e.g.,
Ehlinger and Wilson 1988, Chipps et al. 2004). Given the distinct and rather abrupt change in
geology—and, as was shown in this thesis, in the physical and chemical habitat available to
fishes and macroinvertebrates—at the Shield/off-Shield contact zone, this may provide an
opportunity to explore this feature of various fish species more thoroughly.
In conclusion, this thesis has further defined our understanding of the importance of broad -scale
features – in this case, geology – on abiotic and biotic factors of local ecosystems. As ecologists
continue to search for laws or rules that govern the organization and functioning of organisms
within the environment, we must continue to emphasize the necessity of spatial and temporal
context. This, in addition to the continued integration of ideas across various subfields – for
example, landscape ecology with aquatic ecology – will allow us to continue to unravel the
complexity of the world around us, and thus provide sound knowledge to those tasked with
managing and protecting the natural world.
References
Angermeier, P.L. and M.R. Winston. 1998. Local vs. regional influences on local diversity in
stream fish communities of Virginia. Ecology 79: 911-927.
Bhat, A. 2005. Ecomorphological correlates in tropical stream fishes in southern India.
Environmental Biology of Fishes 73: 211-225.
Bourassa, N., and A. Morin. 1995. Relationships between size structure of invertebrate
assemblages and trophy and substrate composition in streams. Journal of the North
American Benthological Society 14:393–403.
139
Bowlby, J.N. and J.C. Roff. 1986. Trophic structure in southern Ontario streams. Ecology 67:
1670-1679.
Chapman, L. J., and D. F. Putnam. 1984. The physiography of Southern Ontario. Special
Volume 2. Ontario Geological Survey, Sudbury, Ontario.
Chipps, S.R., J.A. Dunbar and D.H. Wahl. 2004. Phenotypic variation and vulnerability to
predation in juvenile bluegill sunfish (Lepomis macrochirus). Oecologia 138: 32-38.
Ehlinger, T.J. and D.S. Wilson. 1988. Complex foraging polymorphism in bluegill sunfish.
Proceedings of the National Academy of Sciences 85: 1878-1882.
Erman, D. C., and N. A. Erman. 1984. The response of stream macroinvertebrates to substrate
size and heterogeneity. Hydrobiologia 108:75–82.
Esselman, P. C., M. C. Freeman, and C. M. Pringle. 2006. Fish-assemblage variation between
geologically defined regions and across a longitudinal gradient in the Monkey River
Basin, Belize. 2006. Journal of the North American Benthological Society 25:142–156.
Fausch, K.D., Torgersen, C.E., Baxter, C.V., and H.W. Li. 2002. Landscapes to riverscapes:
Bridging the gap between research and conservation of stream fishes. Bioscience 52 (6):
483-498.
Flotemersch, J.E., J.B. Stribling, R.M. Hughes, L. Reynolds, M.J. Paul and C. Wolter. 2011.
Site length for biological assessment of boatable rivers. River Research and Applications
27: 520-535.
Heino, J., H. Mykra, J. Kotanen and T. Muotka. 2007. Ecological filters and variability in
stream macroinvertebrate communities: do taxonomic and functional structure follow the
same path? Ecography 30: 217-230.
Hinch, S.G., N.C. Collins, and H.H. Harvey. 1991. Relative abundance of littoral zone fishes:
biotic interactions, abiotic factors, and postglacial colonization. Ecology 72: 1314-1324.
Holm, E., N.E. Mandrak and M.E. Burridge. 2009. The Royal Ontario Museum field guide to
freshwater fishes of Ontario. Royal Ontario Museum, Toronto, Ontario.
140
Holomuzki, J. R., and T. M. Short. 1988. Habitat use and fish avoidance behaviors by the
stream-dwelling isopod Lirceus fontinalis. Oikos 52:79–86.
Hubbs, C.L. and K.F. Lagler; revised by G.R. Smith. 2004. Fishes of the Great Lakes Region.
University of Michigan Press, Michigan, USA.
Jackson, D. A., and H. H. Harvey. 1989. Biogeographic associations in fish assemblages: local
vs. regional processes. Ecology 70:1472–1484.
Jackson, D.A., P.R. Peres-Neto and J.D. Olden. 2001. What controls who is where in freshwater
fish communities – the roles of biotic, abiotic, and spatial factors. Canadian Journal of
Fisheries and Aquatic Sciences 58: 157-170.
Johnson, R.K., M.T. Furse, D. Hering, and L. Sandin. 2007. Ecological relationships between
stream communities and spatial scale: implications for designing catchment-level
monitoring programmes. Freshwater Biology 52: 939-958.
Kilgour, B.W. and D.R. Barton. 1999. Associations between stream fish and benthos across
environmental gradients across southern Ontario, Canada. Freshwater Biology 41: 553-
566.
Kratzer, E. B., J. K. Jackson, D. B. Arscott, A. K. Aufdenkampe, C. L. Dow, L. A. Kaplan, J. D.
Newbold, and B. W. Sweeney. 2006. Macroinvertebrate distribution in relation to land
use and water chemistry in New York City drinking-water-supply watersheds. Journal of
the North American Benthological Society 25:954–976.
Mandrak, N. E. 1995. Biogeographic patterns of fish species richness in Ontario lakes in relation
to historical and environmental factors. Canadian Journal of Fisheries and Aquatic
Sciences 52:1462–1474.
McAuliffe, J. R. 1984. Competition for space, disturbance and the structure of a benthic stream
community. Ecology 65:894–908.
Montana, C.G. and K.O. Winemiller. 2010. Local-scale habitat influences morphological
diversity of species assemblages of cichlid fishes in a tropical floodplain river. Ecology
of Freshwater Fish 19: 216-227.
141
Peres-Neto, Pedro R. 2004. Patterns in the co-occurrence of fish species in streams: the role of
site suitability, morphology and phylogeny versus species interactions. Oecologia 140:
352-360.
Resh, V.H., A.V. Brown, A.P. Covich, M.E. Gurtz, H.W. Li, G.W. Minshall, S.R. Reice, A.L.
Sheldon, B. Wallace and R.C. Wissmar. 1988. The role of disturbance in stream
ecology. Journal of the North American Benthological Society 7: 433-455.
Richards, C., L.B. Johnson, and G.E. Host. 1996. Landscape-scale influences on stream habitats
and biota. Canadian Journal of Fisheries and Aquatic Sciences 53: 295-311.
Richards, C., R.J. Haro, L.B. Johnson, and G.E. Host. 1997. Catchment and reach-scale
properties as indicators of macroinvertebrate species traits. Freshwater Biology 37: 219-
230.
Ricklefs, R. E. 1987. Community diversity: relative roles of local and regional processes.
Science 235:167–171.
Sandin, L. and R.K. Johnson. 2004. Local, landscape and regional factors structuring benthic
macroinvertebrate assemblages in Swedish streams. Landscape Ecology 19: 501-514.
Scott, W.B. and E.J. Crossman. 1998. Freshwater fishes of Canada. Galt House Publications,
Ltd., Oakville, Ontario, Canada.
Smith, C.L. and C.R. Powell. 1971. The summer fish communities of Brier Creek, Marshall
County, Oklahoma. American Museum Novitiates: 2458.
Tonn, W.M. 1990. Climate change and fish communities: a conceptual framework.
Transactions of the American Fisheries Society 119: 337-352.
Wang, L., J. Lyons, P. Rasmussen, P. Seelbach, T. Simon, M. Wiley, P. Kenehl, E. Baker, S.
Niemela, and P. M. Stewart. 2003. Watershed, reach and riparian influences on stream
fish assemblages in the Northern Lakes and Forest Ecoregion, U.S.A. Canadian Journal
of Fisheries and Aquatic Sciences 60:491–505.
Wiley, M. J., S. L. Kohler, and P. W. Seelbach. 1997. Reconciling landscape and local views of
142
aquatic communities: lessons from Michigan trout streams. Freshwater Biology 37:133–
148.
Williams, D.D. 1980. Some relationships between stream benthos and substrate heterogeneity.
Limnology and Oceanography 25: 166-172.
143
Appendices
Appendix A
Table A-1. Macroinvertebrate taxa of the subset of OMNR Aquatic Habitat Inventory data used in Chapter 3. Taxa in brackets indicate historical nomenclature used in original data.
Phylum Subphylum Class Subclass Order Family Branchiobdellae Branchiobdellida Branchiobdellidae
Erpobdellidae Arhynchobdellida Hirudinidae Glossiphoniidae
Hirudinea
Rhynchobdellida Piscicolidae
Haplotaxida
Annelida Clitellata
Oligochaeta Lumbriculida Lumbriculidae
Eylaidae Hydrachnidae Hygrobatidae Hydrodromidae Hydryphantidae Pionidae
Chelicerata Arachnida Acari Trombidiformes
Torrenticolidae Crangonyctidae1 Gammaridae
Amphipoda
Hyalellidae (Talitridae) Decapoda Cambaridae
Crustacea Malacostraca Eumalocostraca
Isopoda Asellidae Chrysomelidae Curculionidae Dryopidae Dytiscidae Elmidae Gyrinidae Haliplidae Heteroceridae Hydraenidae Hydrophilidae Microsporidae (Sphaeriidae) Psephenidae
Coleoptera
Staphylinidae Athericidae Blephariceridae
Arthropoda
Hexapoda Insecta
Diptera
Ceratopogonidae
1 Family of cave-dwelling amphipod; may be possible misidentification.
144
Chironomidae Culicidae Dixidae Empididae Ephydridae Muscidae Psychodidae Ptychopteridae Sciomyzidae Simulidae Stratiomyidae Syrphidae Tabanidae Tipulidae Baetidae Baetiscidae Caenidae Ephemeridae Ephemerellidae Heptageniidae Isonychiidae Leptophlebiidae Siphlonuridae
Ephemeroptera
Tricorythidae Belostomatidae Corixidae Gerridae Mesovelidae2 Nepidae Notonectidae Pleidae2 Reduviidae3
Hemiptera
Veliidae Crambidae Noctuidae Pyralidae
Lepidoptera
Tortricidae4 Sialidae Megaloptera Corydalidae Aeshnidae Calopterygidae Coenagrionidae
Odonata
Cordulegastridae
2 Generally considered to be a lentic taxa.
3 Non-aquatic.
4 Non-aquatic; associated with emergent and floating vegetation.
145
Corduliidae Gomphidae Lestidae Libellulidae Macromiidae Capniidae Chloroperlidae Leuctridae Nemouridae Perlidae Perlodidae Peltoperlidae Pteronarcidae
Plecoptera
Taeniopterygeridae Beraeidae Brachycentridae Glossosomatidae Helicopsychidae Hydropsychidae Hydroptilidae Lepidostomatidae Leptoceridae Limnephilidae Molannidae Philopotamidae Phryganeidae Polycentropodidae Psychomyiidae
Trichoptera
Rhyacophilidae Heterodonta Veneroida Pisidiidae Bivalvia Palaeoheterodonta Unionoida Unionidae
Architaenioglossa Viviparidae Ancylidae Lymnaeidae Physidae
Basommatophora
Planorbidae Bithyniidae Hydrobiidae
Sorbeoconcha
Pleuroceridae
Mollusca
Gastropoda Orthogastropoda
Triganglionata Valvatidae Chordodea Chordodidae Nematomorpha Gordioida Gordea Gordiidae
Platyhelminthes Turbellaria Seriata Tricladida Planariidae
Table A-2. Invertebrate taxa and associated abbreviations used in Chapter 3 ordination figures.
Family Abbreviation Family Abbreviation Family Abbreviation Aeshnidae AESH Glossiphoniidae GLO Notonectidae NOT
146
Ancylidae ANC Glossosomatidae GLOS Peltoperlidae PEL Asellidae ASE Gomphidae GOM Perlidae PER Atheceridae ATH Gordiidae GOR Perlodidae PERL Baetidae BAE Gyrinidae GRY Philopotamidae PHI Belostomatidae BEL Haliplidae HAL Phryganeidae PHR Beraeidae BER Haplotaxida HAP Physidae PHY Bithyniidae BIT Helicopsychidae HEL Pionidae PIO Blephiceridae BLE Heptageniidae HEP Piscicolidae PIS Brachycentridae BRA Heteroceridae HET Pisidiidae PISI Branchiobdellida BRAN Hirudinidae HIR Planariidae PLA Caenidae CAE Hyalellidae HYA Planorbidae PLAN Calopterygidae CAL Hydrachnidae HYAC Pleidae PLE Cambaridae CAM Hydraenidae HYAE Pleuroceridae PLEU Capniidae CAP Hydrobiidae HYBI Polycentropodidae POL Ceratopogonidae CER Hydrodromidae HYDRO Psephenidae PSE Chironomidae CHI Hydrophilidae HYPH Psychodidae PSY Chloroperlidae CHL Hydropsychidae HYPS Psychomyiidae PSYC Chordodidae CHO Hydroptilidae HYPT Pteronarcidae PTE Chrysomelidae CHR Hydryphantidae HYDR Ptychopteridae PTY Coenagrionidae COE Hygrobatidae HYG Pyralidae PYR Cordulegastridae COR Isonychiidae ISO Rhyacophilidae RHY Corduliidae CORD Lebertiidae LEB Riduviidae RID Corixidae CORI Lepidostomatidae LEPI Sciomyzidae SCI Corydalidae CORY Leptoceridae LEP Sialidae SIA Crambidae CRA Leptophlebiidae LEPT Simulidae SIM Crangonyctidae CRAG Lestidae LES Siphlonuridae SIP Culicidae CUL Leuctridae LEU Staphylinidae STA Curculionidae CUR Libellulidae LIB Stratiomyidae STR Dixidae DIX Limnephilidae LIM Syrphidae SYR Dryopidae DRY Lumbriculida LUM Tabanidae TAB Dytiscidae DYT Lymnaeidae LYM Taeniopterygeridae TAE Elmidae ELM Macromiidae MAC Tipulidae TIP Empididae EMP Mesovelidae MES Torrenticolidae TOR Ephemerellidae EPHL Microsporidae MIC Tortricidae TORT Ephemeridae EPH Molannidae MOL Tricorythidae TRI Ephydridae EPHY Muscidae MUS Unionidae UNI Erpobdellidae ERP Nemouridae NEM Valvatidae VAL Eylaidae EYL Nepidae NEP Veliidae VEL Gammaridae GAM Noctuidae NOC Viviparidae VIV Gerridae GER
Table A-2 (revised). Invertebrate taxa and associated abbreviations used in Chapter 3 ordination figures, revised
from original publication to address errors and updates to taxonomic nomenclature.
Family Abbreviation Family Abbreviation Family Abbreviation Aeshnidae AESH Glossiphoniidae GLO Notonectidae NOT Ancylidae ANC Glossosomatidae GLOS Peltoperlidae PEL Asellidae ASE Gomphidae GOM Perlidae PER Athericidae ATH Gordiidae GOR Perlodidae PERL Baetidae BAE Gyrinidae GRY Philopotamidae PHI Belostomatidae BEL Haliplidae HAL Phryganeidae PHR
147
Beraeidae BER Haplotaxida HAP Physidae PHY Bithyniidae BIT Helicopsychidae HEL Pionidae PIO Blephariceridae BLE Heptageniidae HEP Piscicolidae PIS Brachycentridae BRA Heteroceridae HET Pisidiidae PISI Branchiobdellida BRAN Hirudinidae HIR Planariidae PLA Caenidae CAE Hyalellidae HYA Planorbidae PLAN Calopterygidae CAL Hydrachnidae HYAC Pleidae PLE Cambaridae CAM Hydraenidae HYAE Pleuroceridae PLEU Capniidae CAP Hydrobiidae HYBI Polycentropodidae POL Ceratopogonidae CER Hydrodromidae HYDRO Psephenidae PSE Chironomidae CHI Hydrophilidae HYPH Psychodidae PSY Chloroperlidae CHL Hydropsychidae HYPS Psychomyiidae PSYC Chordodidae CHO Hydroptilidae HYPT Pteronarcidae PTE Chrysomelidae CHR Hydryphantidae HYDR Ptychopteridae PTY Coenagrionidae COE Hygrobatidae HYG Pyralidae PYR Cordulegastridae COR Isonychiidae ISO Rhyacophilidae RHY Corduliidae CORD Lebertiidae LEB Reduviidae RID Corixidae CORI Lepidostomatidae LEPI Sciomyzidae SCI Corydalidae CORY Leptoceridae LEP Sialidae SIA Crambidae CRA Leptophlebiidae LEPT Simulidae SIM Crangonyctidae CRAG Lestidae LES Siphlonuridae SIP Culicidae CUL Leuctridae LEU Staphylinidae STA Curculionidae CUR Libellulidae LIB Stratiomyidae STR Dixidae DIX Limnephilidae LIM Syrphidae SYR Dryopidae DRY Lumbriculidae LUM Tabanidae TAB Dytiscidae DYT Lymnaeidae LYM Taeniopterygeridae TAE Elmidae ELM Macromiidae MAC Tipulidae TIP Empididae EMP Mesovelidae MES Torrenticolidae TOR Ephemerellidae EPHL Microsporidae MIC Tortricidae TORT Ephemeridae EPH Molannidae MOL Tricorythidae TRI Ephydridae EPHY Muscidae MUS Unionidae UNI Erpobdellidae ERP Nemouridae NEM Valvatidae VAL Eylaidae EYL Nepidae NEP Veliidae VEL Gammaridae GAM Noctuidae NOC Viviparidae VIV Gerridae GER
148
Fig. A-1. Ordination from a CA of 58 study sites and 109 macroinvertebrate taxa (data set B). The first 2 CA
axes explain 13.8% of the total variation.
149
Appendix B
Table B-1. Summary of all fish species collected, the identification code (ID) used elsewhere in the chapter and
their % occurrence in both Shield and off-Shield systems.
Family Common Name ID Scientific Name
% Shield Sites
Present
% Off-Shield Sites
Present Petromyzontidae Silver Lamprey SIL Ichthyomyzon unicuspis 0 2 Sea Lamprey SEL Petromyzon marinus 0 2 Cyprinidae Northern Redbelly Dace NRD Chrosomus eos 58.5 64.7 Finescale Dace FD Chrosomus neogaeus 17 15.7 Lake Chub LC Couesius plumbeus 11.3 0 Brassy Minnow BM Hybognathus hankinsoni 5.7 29.4 Hornyhead Chub HC Nocomis biguttatus 0 5.9 Golden Shiner GS Notemigonus crysoleucas 18.9 2 Common Shiner CS Luxilus cornutus 67.9 45.1 Blackchin Shiner BCS Notropis heterodon 30.2 7.8 Blacknose Shiner BNS Notropis heterolepis 0 3.9 Rosyface Shiner RS Notropis rubellus 3.8 0 Spotfin Shiner SS Cyprinella spiloptera 0 2 Mimic Shiner MS Notropis volucellus 1.9 0 Bluntnose Minnow BLM Pimephales notatus 24.5 43.1 Fathead Minnow FM Pimephales promelas 20.8 49 Blacknose Dace BD Rhinichthys atratulus 9.4 56.9 Longnose Dace LD Rhinichthys cataractae 22.6 35.3 Creek Chub CC Semotilus atromaculatus 88.7 74.5 Fallfish FF Semotilus corporalis 3.8 3.9 Pearl Dace PD Margariscus margarita 39.6 39.2 Catostomidae Longnose Sucker LS Catostomus catostomus 3.8 5.9 White Sucker WS Catostomus commersonii 50.9 74.5 Ictaluridae Brown Bullhead BB Ameiurus nebulosus 13.2 11.8 Channel Catfish CHC Ictalurus punctatus 1.9 0 Esocidae Northern Pike NP Esox lucius 0 2 Muskellunge MUS Esox masquinongy 1.9 3.9 Umbridae Central Mudminnow CM Umbra limi 1.9 39.2 Salmonidae Rainbow Trout RT Oncorhynchus mykiss 3.8 33.3 Brown Trout BT Salmo trutta 0 11.8 Brook Trout BKT Salvelinus fontinalis 52.8 37.3 Splake SPL S. namaycush x S. fontinalis 0 2 Percopsidae Trout-Perch TP Percopsis omiscomaycus 9.4 0 Gadidae Burbot BUR Lota lota 1.9 2 Gasterosteidae Brook Stickleback BS Culaea inconstans 24.5 56.9 Cottidae Mottled Sculpin MSC Cottus bairdii 3.8 45.1 Slimy Sculpin SLS Cottus cognatus 1.9 2 Centrarchidae Rock Bass RB Ambloplites rupestris 7.5 31.4 Pumpkinseed PS Lepomis gibbosus 28.3 35.3 Bluegill BG Lepomis macrochirus 0 2
150
Smallmouth Bass SB Micropterus dolomieu 15.1 19.6 Largemouth Bass LB Micropterus salmoides 1.9 7.8 Percidae Yellow Perch YP Perca flavescens 30.2 23.5 Walleye WLE Sander vitreus 1.9 3.9 Rainbow Darter RD Etheostoma caeruleum 0 5.9 Iowa Darter ID Etheostoma exile 5.7 11.8 Fantail Darter FTD Etheostoma flabellare 0 3.9 Least Darter LED Etheostoma microperca 0 2 Johnny Darter JD Etheostoma nigrum 5.7 23.5 Logperch LP Percina caprodes 1.9 5.9
151
Table B-2. Summary of all abiotic variables, with mean, median values, and standard deviations of sites sampled
from Shield and off-Shield systems.
Shield sites Off-Shield sites Variables Mean SD Median Mean SD Median
Water Temperature (°C) 14.7505 4.1402 14 19.7936 2.7616 20.1667 Dissolved Oxygen (ppm) 9.8565 1.4265 10.1 8.0712 2.1542 8.45 pH 6.8391 0.5423 6.66 7.7594 0.8609 8.1875 Turbidity (F.T.U.) 6.9029 11.0959 1.4 24.6308 30.2057 15.1548 Alk (mg/l) 24.782 40.5389 10 206.0739 52.7898 197.0833 Water
chemistry Conductivity (µohms) 77.8353 90.3647 42 416.2153 78.5247 401.6667 Undercut bank 0.0484 0.0588 0.05 0.0907 0.115 0.05 Rocks 0.0956 0.1163 0.07 0.0724 0.0689 0.05 Woody debris 0.1098 0.1033 0.1 0.0614 0.0566 0.05 Instream
cover Organic debris 0.0592 0.0589 0.04 0.0641 0.078 0.0396 % Rock 0.4118 1.2776 0 5.4444 9.7993 0 % Boulder 11.3529 13.8967 5 9.3681 9.1997 6.625 % Rubble 9.7059 11.1286 5 19.7778 18.2868 13.75 % Gravel 9.2353 7.9885 10 15.6389 12.051 12.125 % Sand 19.7059 16.0459 20 18.5764 21.4357 10 % Silt 16.1765 18.1495 10 14.3542 13.2816 10 % Clay 5.7059 13.6783 0 4.2986 8.2091 0 % Muck 18.0588 21.7125 10 8.7083 12.4748 2.75 % Marl 0.2941 1.2127 0 1.2222 5.4569 0 Substrate
Composition % Detritus 9.3529 9.7784 5 2.6111 5.619 0 Width (m) 6.2825 4.5751 4.245 4.8118 2.9581 3.9625 Depth (m) 0.32 0.2311 0.328 0.2448 0.1364 0.1937 Velocity (m/s) 0.4655 0.2955 0.3333 0.251 0.1498 0.2 Stream
morphology Discharge (m3/s) 1.3 2.9667 0.34 0.2866 0.4225 0.1242 % Cultivated 0 0 0 6.2537 10.2427 0 % Pasture 0.2179 0.8983 0 10.0844 11.7679 4.955 % Meadow 1.9826 7.2754 0 20.5233 22.8417 17.1888 % Upland Hardwood 20.9501 29.7062 0 24.9189 17.7079 22.6852 % Upland Conifer 17.7537 26.8514 3.7037 13.2973 14.2228 12.5 % Swamp Hardwood 16.8325 32.5201 0 11.1317 14.9577 6.3186 % Swamp Conifer 5.2711 9.8244 0 4.2986 7.9905 0 % Shrub Marsh 28.925 41.9631 7.1429 6.0353 9.5168 0 Landscape
Composition % Open Marsh 8.0672 24.2959 0 3.4567 8.2811 0 % Dense 13.25 21.8682 0 13.9722 19.4918 6.125 % Partly Open 36.1176 35.6715 27.5 27.3819 20.4634 25 Canopy
Cover % Open 50.6324 38.4525 45 58.6458 25.7437 58 % Low (<10°) 93.2353 24.299 100 81 30.4471 100
Gradient % Medium (10°-60°) 6.7647 24.299 0 19 30.4471 0
152
Table B-3. Summary of fish functional traits (Goldstein and Meador 2003, Eakins 2010), including a) temperature
preferences, b) fish trophic status, c) fish geomorphology preferences, d) substrate preferences, and e) reproductive
behaviors.
a) Temperature Cold Cool Warm
Species (<19 °°°°C) (19-25 °°°°C) (>25 °°°°C) Silver Lamprey X Sea Lamprey X Rainbow Trout X Brown Trout X Brook Trout X Northern Pike X Muskellunge X Central Mudminnow X Longnose Sucker X White Sucker X Northern Redbelly Dace X Finescale Dace X Lake Chub X Brassy Minnow X Hornyhead Chub X Golden Shiner X Common Shiner X Blackchin Shiner X Blacknose Shiner X Rosyface Shiner X Spotfin Shiner X Mimic Shiner X Bluntnose Minnow X Fathead Minnow X Blacknose Dace X Longnose Dace X Creek Chub X Fallfish X Pearl Dace X Brown Bullhead X Channel Catfish X Burbot X Brook Stickleback X Trout-Perch X Rock Bass X Pumpkinseed X Bluegill X Smallmouth Bass X Largemouth Bass X Yellow Perch X Walleye X Rainbow Darter X Iowa Darter X Fantail Darter X
153
Least Darter X Johnny Darter X Logperch X Mottled Sculpin X Slimy Sculpin X
154
b) Trophic Status
Species PLANK-
DET PLANK-
INV DET DET-INV HER-INV INV
INV-CAR
VAR-OMNI
Silver Lamprey X Sea Lamprey X Rainbow Trout X Brown Trout X Brook Trout X Northern Pike X Muskellunge X Central Mudminnow X Longnose Sucker X White Sucker X Northern Redbelly Dace X Finescale Dace X Lake Chub X Brassy Minnow X Hornyhead Chub X Golden Shiner X Common Shiner X Blackchin Shiner X Blacknose Shiner X Rosyface Shiner X Spotfin Shiner X Mimic Shiner X Bluntnose Minnow X Fathead Minnow X Blacknose Dace X Longnose Dace X Creek Chub X Fallfish X Pearl Dace X Brown Bullhead X Channel Catfish X Burbot X Brook Stickleback X Trout-Perch X Rock Bass X Pumpkinseed X Bluegill X Smallmouth Bass X Largemouth Bass X Yellow Perch X Walleye X Rainbow Darter X Iowa Darter X Fantail Darter X Least Darter X Johnny Darter X Logperch X Mottled Sculpin X Slimy Sculpin X
155
c) Geomorphology Species Riffle Riffle/Run Riffle/Pool Run/Pool Pool All Slow Variable
Silver Lamprey X Sea Lamprey X Rainbow Trout X Brown Trout X Brook Trout X Northern Pike X Muskellunge X Central Mudminnow X Longnose Sucker X White Sucker X Northern Redbelly Dace X Finescale Dace X Lake Chub X Brassy Minnow X Hornyhead Chub X Golden Shiner X Common Shiner X Blackchin Shiner X Blacknose Shiner X Rosyface Shiner X Spotfin Shiner X Mimic Shiner X Bluntnose Minnow X Fathead Minnow X Blacknose Dace X Longnose Dace X Creek Chub X Fallfish X Pearl Dace X Brown Bullhead X Channel Catfish X Burbot X Brook Stickleback X Trout-Perch X Rock Bass X Pumpkinseed X Bluegill X Smallmouth Bass X Largemouth Bass X Yellow Perch X Walleye X Rainbow Darter X Iowa Darter X Fantail Darter X Least Darter X Johnny Darter X Logperch X Mottled Sculpin X Slimy Sculpin X
156
d) Substrate Species SM VEG-SM SM-MED MED VEG-MED MED-LG VAR VEG-VAR
Silver Lamprey X Sea Lamprey X Rainbow Trout X Brown Trout X Brook Trout X Northern Pike X Muskellunge X Central Mudminnow X Longnose Sucker X White Sucker X Northern Redbelly Dace X Finescale Dace X Lake Chub X Brassy Minnow X Hornyhead Chub X Golden Shiner X Common Shiner X Blackchin Shiner X Blacknose Shiner X Rosyface Shiner X Spotfin Shiner X Mimic Shiner X Bluntnose Minnow X Fathead Minnow X Blacknose Dace X Longnose Dace X Creek Chub X Fallfish X Pearl Dace X Brown Bullhead X Channel Catfish X Burbot X Brook Stickleback X Trout-Perch X Rock Bass X Pumpkinseed X Bluegill X Smallmouth Bass X Largemouth Bass X Yellow Perch X Walleye X Rainbow Darter X Iowa Darter X Fantail Darter X Least Darter X Johnny Darter X Logperch X Mottled Sculpin X Slimy Sculpin X
157
e) Reproduction Species Open substrate, nonguarder Brood hider, nonguarder Complex nester
Silver Lamprey X Sea Lamprey X Rainbow Trout X Brown Trout X Brook Trout X Northern Pike X Muskellunge X Central Mudminnow X Longnose Sucker X White Sucker X Northern Redbelly Dace X Finescale Dace X Lake Chub X Brassy Minnow X Hornyhead Chub X Golden Shiner X Common Shiner X Blackchin Shiner X Blacknose Shiner X Rosyface Shiner X Spotfin Shiner X Mimic Shiner X Bluntnose Minnow X Fathead Minnow X Blacknose Dace X Longnose Dace X Creek Chub X Fallfish X Pearl Dace X Brown Bullhead X Channel Catfish X Burbot X Brook Stickleback X Trout-Perch X Rock Bass X Pumpkinseed X Bluegill X Smallmouth Bass X Largemouth Bass X Yellow Perch X Walleye X Rainbow Darter X Iowa Darter X Fantail Darter X Least Darter X Johnny Darter X Logperch X Mottled Sculpin X Slimy Sculpin X
158
Table B-4. Summary of ALIS1 and ALIS 2 summary variables, and how they are interpreted. To assess the
potential effect of land use on the patterns observed in fish communities in Shield and off-Shield sites, the original
landscape data, which was assessed up to 50 m from the sampling reach at each site, were replaced with catchment
land-use data using the Aquatic Landscape Inventory Software (ALIS). Land use was calculated for each sampling
site, where the entire catchment area was divided among 28 land-use categories, so that a percentage of the total
catchment area was determined for each category. These data were then summarized into two land-use variables
using CA, in the same manner as substrate composition, gradient, and canopy cover (Appendix A). This resulted in
two land-use variables, ALIS1 and ALIS2, which collectively summarize 40.3% of the variation in land use across
all sites. Replacing the original landscape variables (L1 and L2) with ALIS1 and ALIS2, we used CCA to examine
patterns in fish community composition with the abiotic variables. Using partial CCA, we were able to determine
the variation explained solely by ALIS1 and ALIS2 in comparison to the variation explained by the other abiotic
variables.
Abbreviation CA
score Description Positive Water, sparse coniferous, mixed deciduous
ALIS1 Negative Cropland, deciduous swamp, pasture, conifer swamp Positive Sparse coniferous, alvars or mine tailings, unclassified, treed bog
ALIS2 Negative Mixed deciduous, dense deciduous
Table B-5. Summary of the variation explained by land use (ALIS1 and ALIS2), the remaining abiotic variables,
shared variation between land use and abiotic variables, and variation unexplained by any of the included variables.
Variables % Variation
explained Unknown 58.35 Land use only 5.19 Other abiotic factors 35.22 Shared land use and abiotic 1.21
159
Appendix C
Table C-1. 2007-2008 sampling sites and associated bedrock geology. Sites in grey were removed from subsequent analyses due to equipment malfunctions resulting in loss of water chemistry data (sites 5-9) or absence of fish (sites 44 and 60).
ID Name Group Latitude Longitude Day Year Geology 1 Ten Mile Shield 45.27158 78.96019 185 2007 Migmatitic rocks and gneisses 2 Plastic Shield 45.17302 78.82317 186 2007 Felsic igneous rocks 3 Cinder Shield 45.05996 78.97592 187 2007 Migmatitic rocks and gneisses 4 Tramway Shield 45.22254 78.92044 188 2007 Migmatitic rocks and gneisses 5 Shoelace Shield 45.19188 78.77663 189 2007 Tectonite 6 Cashman Shield 45.6013 79.0894 191 2007 Migmatitic rocks and gneisses 7 Tonawanda Shield 45.59262 79.11647 191 2007 Migmatitic rocks and gneisses 8 Thumb Shield 45.37595 78.90462 192 2007 Migmatitic rocks and gneisses 9 Jerry Shield 45.37464 79.12198 193 2007 Migmatitic rocks and gneisses
10 Sixteen Mile Shield 45.33965 79.0093 200 2007 Migmatitic rocks and gneisses 11 Heeny Shield 45.12748 79.09679 201 2007 Migmatitic rocks and gneisses 12 Livingstone Shield 45.41147 78.72363 202 2007 Migmatitic rocks and gneisses 13 Fletcher Shield 45.3406 78.82065 202 2007 Migmatitic rocks and gneisses 14 South Nelson Shield 45.46183 78.9541 203 2007 Migmatitic rocks and gneisses 15 Troutspawn Shield 45.40947 78.73228 204 2007 Migmatitic rocks and gneisses 16 Parkside Shield 45.44115 78.72317 205 2007 Migmatitic rocks and gneisses 17 Charcoal Shield 45.26222 78.90252 205 2007 Felsic igneous rocks 18 Pine Springs Shield 45.11983 78.86145 206 2007 Felsic igneous rocks 19 Hardy Shield 45.4764 78.9132 208 2007 Felsic igneous rocks 20 Lynx Shield 45.23483 79.23875 221 2007 Migmatitic rocks and gneisses 21 Stoneleigh Shield 45.1118 79.23485 221 2007 Migmatitic rocks and gneisses 22 Hinterland Shield 45.39525 78.67182 222 2007 Migmatitic rocks and gneisses 23 Niger Shield 45.41273 78.84092 224 2007 Migmatitic rocks and gneisses 24 Fletcher Bay Shield 45.30648 78.83372 224 2007 Migmatitic rocks and gneisses 25 Clear Shield 45.05842 78.98987 225 2007 Migmatitic rocks and gneisses 26 Ridout Shield 45.10715 78.97965 225 2007 Migmatitic rocks and gneisses 27 Jessop Shield 45.4209 79.25743 227 2007 Migmatitic rocks and gneisses 28 South Waseosa Shield 45.36733 79.28572 227 2007 Felsic igneous rocks
160
29 Siding Shield 45.2734 79.3178 228 2007 Felsic igneous rocks 30 McKay Shield 45.04988 79.18865 229 2007 Felsic igneous rocks 31 Devine Shield 45.18525 79.25912 233 2007 Migmatitic rocks and gneisses 32 Avery Shield 45.19512 78.80612 233 2007 Felsic igneous rocks 33 Harvey Shield 45.29883 78.84782 235 2007 Felsic igneous rocks 34 Queen's Line Shield 45.01433 78.75588 235 2007 Tectonite 35 Baker-Beech Transition 45.07555 78.68612 237 2007 Carbonate metasedimentary rocks 36 White off-Shield 44.44004 79.03178 162 2008 Limestone, dolostone, shale, arkose, sandstone 37 Kingfisher off-Shield 44.56253 78.91637 169 2008 Limestone, dolostone, shale, arkose, sandstone 38 Goose off-Shield 44.42411 78.88838 169 2008 Limestone, dolostone, shale, arkose, sandstone 39 Grass off-Shield 44.73296 79.39757 176 2008 Limestone, dolostone, shale, arkose, sandstone 40 Perch off-Shield 44.7178 78.88615 178 2008 Limestone, dolostone, shale, arkose, sandstone 41 Shadow Transition 44.70583 78.79969 178 2008 Mafic to felsic metavolcanic rocks 42 Mud off-Shield 44.64486 79.26009 183 2008 Limestone, dolostone, shale, arkose, sandstone 43 Young off-Shield 44.7064 79.16663 183 2008 Limestone, dolostone, shale, arkose, sandstone 44 Gull Shield 44.73371 78.81676 190 2008 Early felsic plutonic rocks 45 Rutherford off-Shield 44.51722 78.70612 191 2008 Limestone, dolostone, shale, arkose, sandstone 46 Waiman off-Shield 44.6421 79.21719 197 2008 Limestone, dolostone, shale, arkose, sandstone 47 Emily off-Shield 44.41308 78.65809 198 2008 Limestone, dolostone, shale, arkose, sandstone 48 Beaverton off-Shield 44.39409 79.08286 198 2008 Limestone, dolostone, shale, arkose, sandstone 49 Silver off-Shield 44.6523 79.43703 204 2008 Limestone, dolostone, shale, arkose, sandstone 50 Head off-Shield 44.73063 79.06881 205 2008 Felsic igneous rocks 51 S. Beaver Transition 44.76712 78.77389 211 2008 Mafic to felsic metavolcanic rocks 52 Drag Shield 44.89582 78.64134 212 2008 Early felsic plutonic rocks 53 Kendrick Shield 44.86602 78.67896 212 2008 Early felsic plutonic rocks 54 Crego Shield 44.76714 78.69348 213 2008 Early felsic plutonic rocks 55 Burnt 1 Shield 44.86057 78.62112 213 2008 Early felsic plutonic rocks 56 Burnt 2 Transition 44.75504 78.68589 214 2008 Early felsic plutonic rocks 57 Hawkers off-Shield 44.5588 78.61983 218 2008 Limestone, dolostone, shale, arkose, sandstone 58 Dalrymple off-Shield 44.63468 79.17615 219 2008 Limestone, dolostone, shale, arkose, sandstone 59 Martin off-Shield 44.56486 78.58577 225 2008 Limestone, dolostone, shale, arkose, sandstone 60 Deverells Transition 44.69873 79.11577 226 2008 Felsic igneous rocks 61 Black Shield 44.7881 79.23194 239 2008 Migmatitic rocks and gneisses 62 South Kashe Transition 44.81431 79.33841 239 2008 Felsic igneous rocks 63 Burnt 3 Shield 44.73222 78.67768 240 2008 Early felsic plutonic rocks 64 Mariposa off-Shield 44.36409 78.93342 240 2008 Limestone, dolostone, shale, arkose, sandstone
161
Table C-2. Summary of all fish species collected, with % occurrence in both Shield and off-Shield systems.
% Shield Sites
% Transition
Sites
% Off-Shield Sites
Family Common Name Scientific Name Present Present Present Cyprinidae Northern Redbelly Dace Chrosomus eos 27.8 0 31.3 Finescale Dace Chrosomus neogaeus 5.6 0 12.5 Lake Chub Couesius plumbeus 5.6 0 0 Brassy Minnow Hybognathus hankinsoni 2.8 0 12.5 Pearl Dace Margariscus nachtriebi 5.6 20 6.3 Hornyhead Chub Nocomis biguttatus 5.6 0 6.3 Golden Shiner Notemigonus crysoleucas 0 20 0 Emerald Shiner Notropis atherinoides 2.8 0 0 Common Shiner Luxilus cornutus 27.8 60 25 Blackchin Shiner Notropis heterodon 0 0 6.3 Spottail Shiner Notropis hudsonius 2.8 0 12.5 Bluntnose Minnow Pimephales notatus 5.6 20 0 Fathead Minnow Pimephales promelas 0 0 12.5 Longnose Dace Rhinichthys cataractae 11.1 0 6.3 Blacknose Dace Rhinichthys obtusus 36.1 40 31.3 Creek Chub Semotilus atromaculatus 97.2 80 68.8 Catostomidae White Sucker Catostomus commersonii 16.7 60 12.5 Ictaluridae Brown Bullhead Ameiurus nebulosus 5.6 0 0 Esocidae Grass Pickerel Esox americanus vermiculatus 0 20 0 Umbridae Central Mudminnow Umbra limi 16.7 40 37.5 Salmonidae Brook Trout Salvelinus fontinalis 19.4 0 6.3 Gadidae Burbot Lota lota 0 0 6.3 Gasterosteidae Brook Stickleback Culaea inconstans 25 60 87.5 Cottidae Slimy Sculpin Cottus cognatus 2.8 0 0 Centrarchidae Rock Bass Ambloplites rupestris 2.8 0 0 Pumpkinseed Lepomis gibbosus 11.1 0 6.3 Smallmouth Bass Micropterus dolomieu 2.8 0 0 Largemouth Bass Micropterus salmoides 8.3 0 0 Percidae Rainbow Darter Etheostoma caeruleum 0 0 6.3 Iowa Darter Etheostoma exile 0 0 12.5 Johnny Darter Etheostoma nigrum 0 20 6.3 Yellow Perch Perca flavescens 2.8 0 0
162
Table C-3. Summary of all macroinvertebrate taxa collected, with % occurrence in both Shield and off-Shield
systems.
Taxa
Class Subclass/Clade Order Family % Shield sites
present % Transition sites present
%Off-Shield sites present
Arachnida Hydracarina 30.56 40 43.75 Bivalvia 86.11 100 100 Clitellata Hirudinea 25 40 68.75 Oligochaeta 75 100 93.75 Entognatha Collembola 16.67 0 25 Gastropoda Heterostropha Valvatidae 0 0 25
Neotaenioglossa Hydrobiidae 2.78 20 18.75 Pulmonata Lymnaeidae 0 0 75 Physidae 2.78 20 75 Planoboridae 16.67 40 87.5 Prosobranchia Architaenioglossa Viviparidae 5.56 20 6.25 Insecta Coleoptera Chyrsomelidae 0 20 0 Dysticidae 2.78 20 25 Elmidae 58.33 80 75 Gyrinidae 0 0 6.25 Haliplidae 0 0 37.5 Heteroceridae 0 0 6.25 Hydrophilidae 2.78 20 25 Psephenidae 13.89 20 12.5 Staphylinidae 2.78 0 0 Diptera Ceratopogonidae 94.44 100 87.5 Chaoboridae1 2.78 0 0 Chironomidae 100 100 100 Dixidae 2.78 0 6.25 Empididae 33.33 60 18.75 Psychodidae 2.78 0 0 Ptychopteridae2 0 20 0 Sciomyzidae 0 0 6.25 Simuliidae 72.22 60 81.25 Stratiomyidae 0 0 18.75 Tabanidae 22.22 0 31.25 Tipulidae 83.33 100 50 Ephemeroptera Baetidae 41.67 40 50 Baetiscidae 2.78 0 0 Caenidae 8.33 20 43.75 Ephemerellidae 33.33 0 12.5 Heptageniidae 44.44 0 12.5 Isonychiidae 2.78 0 6.25 Leptophlebiidae 63.89 40 37.5
1 Not considered a lotic taxa.
2 Not considered a lotic taxa.
163
Tricorythidae 5.56 0 0 Hemiptera Belostomatidae 0 0 6.25 Corixidae 5.56 0 31.25 Gerridae 5.56 0 6.25 Hebridae3 0 0 6.25 Macroveliidae 0 0 6.25 Veliidae 2.78 0 0 Lepidoptera Noctuidae4 8.33 0 0 Pyralidae 2.78 0 12.5 Megaloptera Corydalidae 33.33 40 31.25 Sialidae 16.67 0 6.25 Odonata Aeshnidae 47.22 0 31.25 Calopterygidae 30.56 20 18.75 Coenagrionidae 2.78 20 6.25 Cordulegastridae 50 20 25 Cordulidae 8.33 0 25 Gomphidae 33.33 40 6.25 Lestidae 0 0 6.25 Libellulidae 2.78 0 6.25 Plecoptera Chloroperlidae 11.11 20 6.25 Leuctridae 19.44 40 6.25 Nemouridae 5.56 0 6.25 Perlidae 30.56 20 6.25 Trichoptera Glossosomatidae 0 20 0 Hydropsychidae 94.44 80 68.75 Hydroptilidae 25 0 0 Lepidostomatidae 5.56 0 18.75 Leptoceridae 11.11 0 6.25 Limnephilidae 38.89 40 37.5 Molannidae 5.56 0 0 Odontoceridae 5.56 0 0 Philopotamidae 36.11 40 18.75 Phryganeidae 11.11 20 25 Polycentropidae 38.89 20 18.75 Psychomyiidae 16.67 0 18.75 Rhyacophilidae 8.33 0 0 Malacostraca Amphipoda 11.11 40 75 Isopoda Asellidae 2.78 20 56.25
3 Semi-aquatic.
4 Associated with emergent and floating vascular hyrophytes.
164
Table C-4. Summary of all abiotic variables, with mean, standard deviation and median values for both Shield and off-Shield systems.
Shield Transition Off-Shield Variables Mean SD Median Mean SD Median Mean SD Median pH 6.6705 0.5208 6.6283 7.524 0.7268 7.3233 8.0021 0.4422 8.1133 Water
chemistry Conductivity (Cond) 40.1195 32.5685 26.8333 304.2807 118.6775 283.6667 441.3958 125.8081 443 Canopy (Can) 0.5861 0.3537 0.775 0.33 0.3599 0.3 0.3094 0.372 0.05 Gradient (G) 1.7673 2.0547 1.1741 0.878 0.3984 0.92 0.7621 0.7613 0.6016 Velocity (V) 0.1936 0.1567 0.1471 0.2802 0.0615 0.2737 0.1868 0.1176 0.2086 Dissolved oxygen (DO) 6.3231 1.2956 6.4367 6.974 2.269 7.46 7.3758 2.2839 7.9233 Temperature (T) 19.0407 2.7195 19.65 18.26 2.1335 17.3 18.5875 2.7541 18.3167 Moss (M) 0.0821 0.1133 0.0333 0.0167 0.0289 0 0.0097 0.0259 0 Woody debris (WD) 0.2833 0.1754 0.2472 0.21 0.074 0.2333 0.176 0.144 0.1667 Algae (AL) 0.0596 0.0863 0.0278 0.0033 0.0075 0 0.0656 0.1593 0 Water vegetation (WV) 0.0955 0.1597 0.025 0.1133 0.125 0.0833 0.2045 0.2788 0.05 Depth 0.1432 0.0929 0.1122 0.313 0.3218 0.1699 0.1993 0.1119 0.1775 % Detritus 9.3739 10.118 7.4537 10.7864 12.0769 6.9524 11.7613 15.8263 7.0539 % Clay 1.0643 3.9813 0 0 0 0 10.1917 16.1579 1.4583 % Sand-silt 36.7653 23.8921 36.5 54.361 9.841 56.482 42.0555 23.2732 41.2103 % Gravel 26.1515 18.159 25.8333 26.3657 15.215 33.064 21.9029 19.9118 19.2094 % Cobble 15.8606 18.2999 6.5 7.5869 13.7459 1.1364 4.1745 6.5808 1.9156 % Boulder 5.5042 7.8982 1.5 0.9 2.0125 0 1.2968 2.0304 0
Physical habitat
% Bedrock 5.2802 11.8955 0 0 0 0 8.6173 18.5813 0 % Water 9.9031 6.1398 9.913 8.4124 6.2508 9.4576 1.6462 5.5887 0 % Wetlands 0 0 0 0.0043 0.0095 0 1.5928 6.2708 0 % Deciduous Swamp 0.0448 0.2679 0 0.9235 1.2663 0.423 4.0867 7.0818 0.3976 % Conifer Swamp 0 0 0 0.9492 1.9179 0 3.7717 8.0853 0.3769 % Open Fen 0 0 0 0.0501 0.0933 0 0 0 0 % Open Bog 0.08 0.3856 0 0.0052 0.0117 0 0 0 0 % Treed Bog 1.6775 1.3825 1.3524 0.4983 0.7028 0 0 0 0 % Dense Deciduous 29.1374 22.2761 26.8019 17.7959 7.8109 16.7842 16.0209 22.3627 7.2589 % Dense Coniferous 3.9556 2.593 3.5768 9.1234 10.2432 5.2734 10.7177 15.5104 2.6748
Land Use
% Coniferous 0.0001 0.0008 0 0.0009 0.002 0 0 0 0
165
Plantation % Mixed - Deciduous 22.6385 12.2002 22.8302 14.3034 4.6482 14.2145 2.367 4.5067 0.8413 % Mixed - Coniferous 20.8742 16.9645 14.5443 27.5579 6.5969 25.48 3.7118 4.5416 2.6737 % Sparse Coniferous 1.6477 4.044 0.0504 5.9081 10.6952 2.0137 0 0 0 % Sparse Deciduous 5.8929 7.8723 3.2288 9.0387 7.3003 11.6904 2.5887 3.5818 1.2965 % Recent Cutovers 0.089 0.2555 0 0.3456 0.4587 0.0688 0 0 0 % Settlement 0.2797 0.6232 0 0.124 0.2773 0 0 0 0 % Pasture 0.9561 2.5592 0.0344 2.4 2.3904 1.5872 12.9088 13.0489 11.7264 % Cropland 2.7486 15.5333 0 2.3872 3.3901 1.3874 37.8791 30.7679 25.9031 % Alvar or Mine Tailings 0.0726 0.2677 0 0.1701 0.3804 0 2.7087 7.406 0 % Unclassified 0.0021 0.0123 0 0.0021 0.0046 0 0 0 0
166
Table C-5. Summary of fish functional traits (Goldstein and Meador 2003, Eakins 2010), including a) temperature
preferences, b) trophic status, c) geomorphology preferences, d) substrate preferences, and e) reproductive
behaviors.
a) Temperature Cold Cool Warm
Species (<19°°°°
C) (19-25
°°°°C) (>25 °°°°C)
Blackchin Shiner X Blacknose Dace X Bluntnose Minnow X Brassy Minnow X Brook Stickleback X Brook Trout X Brown Bullhead X Burbot X Central Mudminnow X Common Shiner X Creek Chub X Emerald Shiner X Fathead Minnow X Finescale Dace X Golden Shiner X Grass Pickerel X Hornyhead Chub X Iowa Darter X Johnny Darter X Lake Chub X Largemouth Bass X Longnose Dace X Northern Redbelly Dace X Pearl Dace X Pumpkinseed X Rainbow Darter X Rock Bass X Slimy Sculpin X Smallmouth Bass X Spottail Shiner X White Sucker X Yellow Perch X
167
b) Trophic
Species PLANK PLANK-
DET PLANK-
INV DET DET-INV
HER-INV INV
INV-CARN CARN VAR
Blackchin Shiner X Blacknose Dace X Bluntnose Minnow X Brassy Minnow X Brook Stickleback X Brook Trout X Brown Bullhead X Burbot X Central Mudminnow X Common Shiner X Creek Chub X Emerald Shiner X Fathead Minnow X Finescale Dace X Golden Shiner X Grass Pickerel X Hornyhead Chub X Iowa Darter X Johnny Darter X Lake Chub X Largemouth Bass X Longnose Dace X Northern Redbelly Dace X Pearl Dace X Pumpkinseed X Rainbow Darter X Rock Bass X Slimy Sculpin X Smallmouth Bass X Spottail Shiner X White Sucker X Yellow Perch X
168
c) Geomorphology
Species RF RF/RUN RF/P RUN/P P All
slow VAR Blackchin Shiner X Blacknose Dace X Bluntnose Minnow X Brassy Minnow X Brook Stickleback X Brook Trout X Brown Bullhead X Burbot X Central Mudminnow X Common Shiner X Creek Chub X Emerald Shiner X Fathead Minnow X Finescale Dace X Golden Shiner X Grass Pickerel X Hornyhead Chub X Iowa Darter X Johnny Darter X Lake Chub X Largemouth Bass X Longnose Dace X Northern Redbelly Dace X Pearl Dace X Pumpkinseed X Rainbow Darter X Rock Bass X Slimy Sculpin X Smallmouth Bass X Spottail Shiner X White Sucker X Yellow Perch X
169
d) Substrate
Species VEG-MUD
VEG-SN
VEG-G-C
VEG-VAR
MUD-VEG-SN
MUD-SN
MUD-SN-G
SN-G
SN-G-C G
G-C G-C-B VAR
Blackchin Shiner X Blacknose Dace X Bluntnose Minnow X Brassy Minnow X Brook Stickleback X Brook Trout X Brown Bullhead X Burbot X Central Mudminnow X Common Shiner X Creek Chub X Emerald Shiner X Fathead Minnow X Finescale Dace X Golden Shiner X Grass Pickerel X Hornyhead Chub X Iowa Darter X Johnny Darter X Lake Chub X Largemouth Bass X Longnose Dace X Northern Redbelly Dace X Pearl Dace X Pumpkinseed X Rainbow Darter X Rock Bass X Slimy Sculpin X Smallmouth Bass X Spottail Shiner X White Sucker X Yellow Perch X
170
e) Reproduction
Species Open,
nonguarder
Brood hider,
nonguarder Complex
nester Blackchin Shiner X Blacknose Dace X Bluntnose Minnow X Brassy Minnow X Brook Stickleback X Brook Trout X Brown Bullhead X Burbot X Central Mudminnow X Common Shiner X Creek Chub X Emerald Shiner X Fathead Minnow X Finescale Dace X Golden Shiner X Grass Pickerel X Hornyhead Chub X Iowa Darter X Johnny Darter X Lake Chub X Largemouth Bass X Longnose Dace X Northern Redbelly Dace X Pearl Dace X Pumpkinseed X Rainbow Darter X Rock Bass X Slimy Sculpin X Smallmouth Bass X Spottail Shiner X White Sucker X Yellow Perch X
171
Table C-6. List of macroinvertebrate taxa and abbreviations used in ordination plots.
Macroinvertebrates Fish AESH Aeshnidae LEPI Lepidostomatidae BCS Blackchin Shiner AMP Amphipoda LEP Leptoceridae BD Blacknose Dace ASE Asellidae LEPT Leptophlebiidae BM Bluntnose Minnow BAE Baetidae LES Lestidae BRM Brassy Minnow
BAET Baetiscidae LEU Leuctridae BS Brook Stickleback BEL Belostomatidae LIB Libellulidae BT Brook Trout BIV Bivalvia LIM Limnephilidae BB Brown Bullhead CAE Caenidae LYM Lymnaeidae BUR Burbot CAL Calopterygidae MAC Macroveliidae CM Central Mudminnow CER Ceratopogonidae MOL Molannidae CS Common Shiner CHA Chaoboridae NEM Nemouridae CC Creek Chub CHI Chironomidae NOC Noctuidae ES Emerald Shiner CHL Chloroperlidae ODO Odontoceridae FM Fathead Minnow CHR Chyrsomelidae OLI Oligochaeta FD Finescale Dace COE Coenagrionidae PER Perlidae GS Golden Shiner COL Collembola PHI Philopotamidae GP Grass Pickerel COR Cordulegastridae PHR Phryganeidae HC Hornyhead Chub
CORD Cordulidae PHY Physidae ID Iowa Darter CORI Corixidae PLAN Planoboridae JD Johnny Darter CORY Corydalidae POL Polycentropidae LC Lake Chub
DIX Dixidae PSE Psephenidae LB Largemouth Bass DYT Dysticidae PSY Psychodidae LD Longnose Dace ELM Elmidae PSYC Psychomyiidae NRD Northern Redbelly Dace EMP Empididae PTY Ptychopteridae PD Pearl dace EPHL Ephemerellidae PRY Pyralidae PS Pumpkinseed GER Gerridae RHY Rhyachophilidae RD Rainbow Darter
GLOS Glossomatidae SCI Sciomyzidae RB Rock Bass GOM Gomphidae SIA Sialidae SS Slimy Sculpin GRY Gyrinidae SIM Simuliidae SB Smallmouth Bass HAL Haliplidae STA Staphylinidae SPS Spottail Shiner HEB Hebridae STR Stratiomyidae WS White Sucker HEP Heptageniidae TAB Tabanidae YP Yellow Perch HET Heteroceridae TIP Tipulidae HIR Hirudinea TRI Tricorythidae
HYAC Hydracarina VAL Valvatidae HYBI Hydrobiidae VEL Veliidae HYPH Hydrophilidae VIV Viviparidae HYPS Hydropsychidae HYPT Hydroptilidae
ISO Isonychiidae
172
Fig. C-1. Relationship between site area (m2) and a) fish and b) macroinvertebrate taxa richness for each site.
Shield sites are displayed as green circles, transition sites as pink circles, and off-Shield sites as blue circles.
Linear regression for fish species richness and site area showed no significant relationships (adjusted R2 = -
0.0122, p=0.5704), as well as no significant relationships for macroinvertebrate taxa richness and site area
(adjusted R2=-0.0173, p=0.8297). This indicate that differences in total sampled area were not related to
differences in taxa richness between sites.
173
Fig. C-2. Survey of electrofishing efficiency of a one-pass protocol compared to a three-pass protocol for three
Shield sites (Plastic, Stoneleigh and Jerry). Each site was blocked upstream and downstream of sampling area
with block nets, sampled three times using the backpack electrofisher. Fishes caught in each pass were kept
separate from each other to determine how many additional individuals were caught with each successive pass.
Since Shield streams are lower in conductivity than off-Shield streams and may pose a problem for effective
sampling by electrofisher, three Shield streams with varying conductivity were chosen for this analysis to assess
any differences in efficiency of the one-pass protocol in low conductivity waters. a) Plot showing number of
individual fish caught in each pass. Stoneleigh, with the highest conductivity at 83µS, showed the least amount
of decline in individuals caught between passes, suggesting that low conductivity did not reduce the efficiency of
the first pass compared to the number of individuals caught with three passes. b) Plot showing the number of
unique fish species caught in each pass. For example, three fish species were caught in the first pass at
Stoneleigh, and one additional species (not captured in the first pass) was captured in the second pass. No
additional new species were caught in the second or third passes at Jerry or Plastic.
174
Fig. C-3. Macroinvertebrate taxa depletion curves for seven sites: Livingstone, South Nelson, Parkside, Young,
Rutherford, Waiman and Mariposa. For these sites, each pool and riffle sample was divided into ten equal
subsamples and then sorted and identified. Depletion curves were calculated to ascertain whether a certain
percentage of the overall sample was sufficient to obtain all taxa found in the full sample.
175
Copyright Acknowledgements
Chapter II has been previously published in the Canadian Journal for Fisheries and Aquatic
Sciences and is used in this thesis with permission.
Chapter III has been accepted for publication in the Journal of the North American Benthological
Society and is used in this thesis with permission.
