Assessing non-point source pollution in agricultural regions of the upper
St. John River basin using the slimy sculpin (Cottus cognatus)
by
Michelle Anya Gray
BSc (Honours), Trent University, 1996
MSc, Trent University, 1998
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
Doctor of Philosophy
In the Graduate Academic Unit of Biology
Supervisor:
Co-supervisor:
Kelly Munkittrick, PhD, Dept of Biology, UNBSJ
Rick Cunjak, PhD, Dept of Biology, UNBF
Examining Board: Deb MacLatchy, PhD, Dept of Biology, UNBSJ
Katy Haralampides, PhD, Dept of Civil Engineering, UNBF
Charles Bourque, PhD, Faculty of Forestry and Environmental
Management, UNBF
External Examiner: Doug Holdway, PhD, School of Science, University of Ontario
Institute of Technology
This thesis is accepted
_________________________
Dean of Graduate Studies
THE UNIVERSITY OF NEW BRUNSWICK
December, 2003
© Michelle A. Gray, 2003
ABSTRACT
The overall objective of this research project was to assess whether fish populations
in areas of potato cultivation responded to changes in environmental conditions. An
effects-based assessment was conducted in the ‘potato belt’ of northwestern New
Brunswick in the Little River catchment. From 1999-2001, the health and performance
of slimy sculpin (Cottus cognatus) was monitored in agricultural and forested sections
of the river. In the fall of 1999 and 2000, agricultural sites had fewer young-of-the-year
(YOY) sculpin than the forested region. Adult sculpin were larger in the agricultural
region, but had significantly smaller gonads, and female sculpin had smaller livers, and
fewer and smaller eggs than the forested region. By the fall of 2001, only female gonad
size showed a difference from the forested region. These results were used to design
a follow-up study designed to investigate the relative importance of environmental
factors influencing sculpin responses.
The second study investigated the relative influence of temperature and sediment
deposition on slimy sculpin populations across 20 sites on 19 streams in forested and
agricultural catchments in northwestern New Brunswick. YOY sculpin were present at
all forested sites, but only at 2 of 11 agricultural sites. There were no relationships
between body size or density and sediment deposition in either the agricultural or
forested regions, but sculpin density decreased and median YOY size increased with
increasing temperatures. The variability in density of YOY sculpin at agricultural sites
suggested that additional factors beyond temperature might be contributing to
responses.
A secondary overall objective was to evaluate the slimy sculpin as a sentinel and
indicator of site-specific conditions. Stable isotopes of muscle tissues showed little
variability in isotopic signatures, and significant differences between adjacent sites.
Passive integrated transponder (PIT) tags implanted in 112 adult sculpin showed that
ii
75% of sculpin captured over 10 months moved less than 30m. Both isotopes and PIT
tags suggested high spatial and temporal residency of slimy sculpin.
This PhD project showed biological impacts on sculpin populations residing in
streams influenced by non-point source agricultural stressors, and provided support for
the ability of the slimy sculpin to reflect local environmental conditions.
iii
ACKNOWLEDGMENTS
The first and foremost person I need to thank for making this happen is Dr. Kelly
Munkittrick. It has been a long road but he was with me all of the way; whether it was
talking about data over a beer or two, taking me to the emergency room to get stitches
(“don’t cut towards yourself, Michelle!”), zapping me with the electrofisher, being like a
Dad and giving me advice on any topic under the sun, being patient with me as I lost
my ‘oomph’ a couple of time, revising endless pages of abstracts, papers, and
chapters, ruining many song titles (and sneezing) forevermore, and sharing too many
giggles to count. He picked me up whenever I fell down and battled with my lack of
faith in what I could accomplish. I cannot even describe the respect that I have for you,
and how lucky I feel that I was able to work with you for the past 4 years. I know that I
can never convey to you how much I appreciate all that you have done for me.
Dr. Rick Cunjak, my co-supervisor, got me by default and probably got more than
he bargained for! Thanks for answering a million questions and helping me out after I
barged into your office on a semi-daily basis. I am happy that I didn’t fulfill your
prediction that this would take me 10-15 years – I just couldn’t let you be right!
Dr. Allen Curry turned out to be more than just a committee member. He pushed
me hard sometimes to make sure that I kept my questions in focus although I spent a
good few hours cursing him for it. Even still I have a great amount of professional
respect for him and cannot hold a grudge. ☺
My family has been supporting me with encouragement throughout my extended
academic life. My mother has diligently cut out every newspaper article about fish and
science for years, and has always been there to remind me that I could get this thesis
done. My sister has also carefully asked me time and time again, “What EXACTLY is it
that you do again?” so she could tell her accountant friends in Bermuda all about it.
iv
And of course there is Steve…the man of my dreams and my heart who very
fortunately came into my life near the end of my thesis. His patience and ability to
relate to the “thesis writing woes” was exactly what I needed. He listened to me whine,
let me procrastinate when I needed (or wanted) to, and silently helped push me in the
right direction. I am so glad that he came into my life and can’t wait to get on with our
future together now that this hurdle has been passed.
Many thanks to Jean-Louis Daigle and Gordon Fairchild at the Eastern Soil and
Water Conservation Centre in Saint Andre, for providing lab space early in the project
and also for answering many quick questions at the end of the project. Herb Rees at
Agriculture and Agri-Food Canada also came through with lots of helpful information
when I was writing my thesis.
I certainly cannot write an acknowledgements section without thanking those in the
Biology office. Without the help of Marg, Margaret, Marni, and Linda I would still be
running around in circles. Thanks so much for helping me whenever and however I
needed it!
A grant awarded to Kelly Munkittrick by the Toxic Substances Research Initiative
(TSRI) funded the bulk of the thesis research, with additional funding from Crop Life
Canada. A personal post-graduate scholarship from the Natural Science and
Engineering Research Council (NSERC) saved Kelly some money for at least the first
two years.
It is nearly impossible to list, let alone remember, all of the people that came out
and helped in the field or lab throughout my PhD thesis research. In respecting the
hard work that they put in while helping to get me yet some more data, I will attempt a
list but will ask for leniency if I have forgotten a name or two (or more) after 4 years.
Allen Curry
Mark Gautreau
Nicole Duke
Rick Cunjak
Kirk Roach
Chris Cronin
Kelly Munkittrick
Chad Doherty
Megan Findlay
v
Steve Currie
Sandra Brasfield
Tim Rees
Kyle Vodjani
Sullivan Power
Patricia Edwards
Anne MacGeachy
Brendan Galloway
Dan Cartwright
Gordon Yamazaki
Robyn O’Keefe
Coral Cargill
Gerald Tetrault
Karen Gosse
John O’Keefe
Ryan Delong
Lisa Peters
Christine Paton
I finished my MSc thesis acknowledgements with a quote from a Great Big Sea
song to sum up my thesis experience, so I figured I should do the same here:
…
Fog lifts to reveal potential
for generations prophesized
Our growth to be exponential
Our promise finally realized
I can see the earth below me
and I can feel it turn
Feel it turn
Across the sky
The world it learns
So must I
…
- Feel it turn, Great Big Sea 2002
vi
TABLE OF CONTENTS
ABSTRACT..................................................................................................................... ii
ACKNOWLEDGMENTS ................................................................................................ iv
TABLE OF CONTENTS................................................................................................ vii
LIST OF TABLES ........................................................................................................... x
LIST OF FIGURES........................................................................................................xiii
CHAPTER 1
1
GENERAL INTRODUCTION...........................................................................................1
1.1
Cumulative effects assessment .......................................................................1
1.2
Environmental concerns related to agricultural operations..............................3
1.2.1
1.3
Previous research on fish in agricultural regions .............................................5
1.3.1
1.4
Potato productions ...................................................................................5
Background to the present study .............................................................7
Performance measures for effects-based fish assessment.............................8
1.4.1
Choice of monitoring species...................................................................9
1.5
Objectives and outline of thesis .....................................................................11
1.6
References ....................................................................................................18
CHAPTER 2
21
An effects-based assessment of slimy sculpin (Cottus cognatus) populations in potato
agriculture regions of northwestern New Brunswick......................................................21
2.1
Abstract..........................................................................................................21
2.2
Introduction ....................................................................................................22
2.3
Methods .........................................................................................................25
vii
2.4
Results...........................................................................................................30
2.5
Discussion .....................................................................................................36
2.5.1
Conclusions ...........................................................................................43
2.6
Acknowledgements........................................................................................44
2.7
References ....................................................................................................70
CHAPTER 3
73
Investigating the impacts of sediment and temperature on slimy sculpin (Cottus
cognatus) populations in agricultural catchments..........................................................73
3.1
Abstract..........................................................................................................73
3.2
Introduction ....................................................................................................74
3.3
Methods .........................................................................................................76
3.4
Results...........................................................................................................79
3.5
Discussion .....................................................................................................84
3.5.1
Conclusions ...........................................................................................89
3.6
Acknowledgments..........................................................................................90
3.7
References ..................................................................................................104
CHAPTER 4
106
The use of stable isotope analysis to assess the site fidelity of slimy sculpin (Cottus
cognatus).....................................................................................................................106
4.1
Abstract........................................................................................................106
4.2
Introduction ..................................................................................................107
4.3
Methods .......................................................................................................109
4.4
Results.........................................................................................................111
4.5
Discussion ...................................................................................................112
viii
4.5.1
Conclusion ...........................................................................................116
4.6
Acknowledgements......................................................................................116
4.7
References ..................................................................................................121
CHAPTER 5
123
Measuring small-scale movements of the slimy sculpin (Cottus cognatus) to assess site
fidelity in a small river ..................................................................................................123
5.1
Abstract........................................................................................................123
5.2
Introduction ..................................................................................................124
5. 3
Methods .......................................................................................................125
5.4
Results.........................................................................................................130
5.5
Discussion ...................................................................................................132
5.5.1
Conclusions .........................................................................................136
5.6
Acknowledgements......................................................................................137
5.7
References ..................................................................................................145
CHAPTER 6
147
GENERAL DISCUSSION ............................................................................................147
6.1
Effects-based assessment...........................................................................148
6.2
Suitability of the slimy sculpin ......................................................................150
6.3
Important agricultural inputs ........................................................................152
6.3.1
Recent legislation related to potato farming.........................................154
6.4
Conclusions .................................................................................................155
6.5
Future considerations ..................................................................................156
6.6
References ..................................................................................................162
ix
LIST OF TABLES
Table 1.1. A generalized potato farming operation for one season of potato cultivation
in Eastern Canada. Chemical applications include fertilizers (Fe),
insecticides (I), fungicides (Fu), and herbicides (H)...................................14
Table 1.2. Occurrence of fish kill events recorded and total annual rainfall (mm)1 on
Prince Edward Island (1990-2002)2. Blank cell indicates data not available.
...................................................................................................................15
Table 2.1. Little River site characteristics (mean ± SE) measured in August 2000. The
number of measurements at each site is indicated in brackets. ................45
Table 2.2. Sculpin sampling dates for the Little River (sites 1-10) and the Little Forks
Branch stream (LF1-4) from 1999 to 2001. In general, each collection
consisted of 100 sculpin measured and weighed with lethal sampling of
mature fish occurring in the fall and spring. Adjacent land-use was
classified as forested (F), agricultural (A), transitional between forested and
agricultural (T), and urban (U)....................................................................46
Table 2.3. Precipitation and air temperature for the study region was summarized from
data collected at the St. Leonard airport (Environment Canada 2001b),
located approximately 13km northwest from the mouth of the Little River,
and 48km southeast from the mouth of the Little Forks Branch. The noncropping period is from 1 November through 30 April, and the cropping
period is from 1 May through 31 October. .................................................47
Table 2.4. Degree-days (sum of mean daily temperature) for water temperatures
between 27 July and 18 October for the Little River and Little Forks Branch.
Site numbers increase in the downstream direction. .................................48
Table 2.5. Proportions and median sizes of young-of-the-year (YOY) sculpin of sample
populations collected in the Little River, August and November 1999. The
change in the median size of YOY sculpin at each site was used as a
surrogate for growth...................................................................................49
Table 2.6. Mean (± SE) length, carcass weight, and condition factor for mature male
and female slimy sculpin collected in the Little River in November 1999.
Parameter sample sizes different from the overall sample size (N) are
indicated in brackets. Different letters indicate significant differences
between sites for each variable (ANOVA; p<0.05). Condition was assessed
by ANCOVA of carcass weight, with length as the covariate.....................50
Table 2.7. Proportions and median sizes of young-of-the-year (YOY) sculpin of sample
populations collected in the Little River, August and November 2000 and
2001. The change in the median size of YOY sculpin at each site was
used a surrogate for growth. ......................................................................51
Table 2.8. Proportions and median sizes of young-of the year (YOY) sculpin of sample
populations collected along the Little Forks Branch stream in August and
November 2001. The change in the median size of YOY sculpin at each
site was used a surrogate for growth. ........................................................52
Table 2.9. Mean (± SE) length, carcass weight, and condition factor for mature male
slimy sculpin collected in the Little River and Little Forks stream in 2000
and 2001. Within each river and collection period, different letters indicate
significant differences between sites for each variable (p<0.05). Condition
was assessed by ANCOVA of weight, with length as the covariate. .........53
x
Table 2.10. Mean (± SE) length, carcass weight, and condition factor for mature female
slimy sculpin collected in the Little River and Little Forks stream in 2000
and 2001. Parameter sample size differences are indicated in brackets.
Within each river and collection period, different letters indicate significant
differences between sites for each variable (p<0.05). Condition was
assessed by ANCOVA of weight, with length as the covariate..................54
Table 3.1. Location and site identification numbers for the streams used in the study.
Streams are numbered in ascending order from smallest to largest stream
width in forested (F), and agricultural (A) regions. Asterisks indicate sites
common to Welch et al. (1977). The mean width and length of sites
sampled for fishing survey are also listed. .................................................91
Table 3.2. Summary of temperature statistics for streams in forested and agricultural
catchments of northwestern New Brunswick. The values represent
temperatures measured between 5 July and 8 October, 2001. .................92
Table 3.3. Presence/absence of fish species in forested (n=9) and agricultural sites
streams (n=11) in northwestern New Brunswick (October 2001). Values
are the number of sites with the particular fish species present. The values
in brackets represent the adjusted number of sites where Atlantic salmon
could be present (restricted in some cases due to hydroelectric facilities).93
Table 3.4. Summary of slimy sculpin abundance, density, and size of YOY sculpin
collected in forested and agricultural streams of northwestern New
Brunswick (October 2001). ........................................................................94
Table 3.5. Water quality and sediment analysis in forested and agriculture sites.
Values represent mean (±SE), sample sizes are shown in brackets (n), and
asterisks indicate a significant difference between land-use regions (MannWhitney U test; p<0.05). ............................................................................95
Table 3.6. Ranked relative abundance, from lowest to highest, of brook trout and slimy
sculpin collected in 1974 and 2001. Blank cells indicate that fish were
absent from that site during the collection period. The shift in relative
abundance from 1974 to 2001 is indicated as an increase (+), a decrease
(-), or no change (0). ..................................................................................96
Table 5.1. Median length (min-max) for sculpin collected and PIT tagged in the
Kennebecasis River (Group I A-E, Group II Buffer 1-3). Number of sculpin
recaptured from each section and mean ±SE distances [number of
movements] are given for upstream and downstream movements. ........139
Table 5.2. Full model set was assessed using Akaike’s Information Criterion (AIC).
Survival, φ, and recapture probability, p, were kept constant (.), varied by
group (g), time (t), or group and time (g+t). Lowest QAICc values are given
to the most parsimonious model, ∆QAICc = model QAICc - best model
QAICc, and QAICc weight indicates the degree of support given to the
model. ......................................................................................................140
Table 5.3. Maximum likelihood estimates, with standard error (SE) and 95%
confidence limits, for apparent survival (φ) and recapture probabilities (p)
obtained using the program MARK for the best model (see Table 5.2). TITV represent the tracking events. ............................................................141
Table 6.1. Average monthly discharge, and sediment and chemical loading for the
Black Brook Catchment (1992-93) (modified from Chow et al. 1995). The
non-cropping period is 1 November-30 April, and the cropping period is 1
May- 31 October. .....................................................................................158
xi
Table 6.2. Toxicity classification, expected toxicity to fish, chemical properties, and
persistence1 of the top five pesticides used in the agricultural region of the
Black Brook catchment. Chemical ranking was based on the number of
applications multiplied by the area the chemical was applied (H Rees,
Agriculture and Agri-Food Canada, pers. comm.). ..................................159
xii
LIST OF FIGURES
Figure 1.1. Framework for conducting an effects-based cumulative effects assessment
(Munkttrick et al. 2000). .............................................................................16
Figure 1.2. Biological levels of organization where responses may be measured. The
difficulty of measurement, response time for changes, and ecological
significance increase with increasing biological levels...............................17
Figure 2.1. Map of the Little River, New Brunswick indicating site numbers and
locations where sculpin were collected from 1999-2001. ..........................55
Figure 2.2. Map of the forested Little Forks Branch of the Green River indicating site
numbers and locations where sculpin were collected from 2000-2001. ....56
Figure 2.3. Length frequency distributions for sculpin collected in the Little River in
August (grey line) and November (black line) of 1999. Frequencies are
given as % of sample for comparison purposes (minimum N=96).............57
Figure 2.4. Length distributions of sculpin at sites with both an August (top) and
November (bottom) 1999 collection. The ends of the box represent the 25th
and 75th percentile length, whiskers represent the 5th and 95th percentile
lengths, and the horizontal black line represents the median length.
Difference letters above boxes indicate significant differences in length
distributions within each month (Kolmogorov-Smirnov tests; p<0.05). ......59
Figure 2.5. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature male sculpin collected in the Little River in November
1999. Sites 1-5 are located in the forested region (black bars), site 6 in the
transitional region (dark grey bar), and sites 7-9 in the agricultural region
(light grey bars), and site 10 in the urban region (white bar). Different
letters represent statistical differences from ANCOVA on organ weight, with
carcass weight as the covariate (p<0.05). .................................................60
Figure 2.6. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature female sculpin collected in the Little River in November
1999. Sites 1-5 are located in the forested region (black bars), site 6 in the
transitional region (dark grey bar), and sites 7-9 in the agricultural region
(light grey bars), and site 10 in the urban region (white bar). Different
letters represent statistical differences from ANCOVA on organ weight, with
carcass weight as the covariate (p<0.05). .................................................61
Figure 2.7. Length frequency distributions for sculpin collected in the Little Forks
stream in August (grey line) and November (black line) of 2001.
Frequencies are given as % of sample for comparison purposes. ............62
Figure 2.8. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature male sculpin collected in the Little River from the spring
and fall of 2000 and 2001. Sites 1-3 are located in the forested region
(black bars), sites 7 and 9 in the agricultural region (light grey bars), and
site 10 in the urban region (white bar). Within each collection period,
different letters represent statistical differences from ANCOVA on organ
weights, with carcass weight as the covariate (p<0.05). Significant
interactions within ANCOVA are indicated with an asterisk (*). .................63
Figure 2.9. Mean (± SE), liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature male sculpin collected in the Little Forks stream in
November 2000 and 2001. Sites are numbered 1-4 from upstream to
downstream. Within each collection period, different letters represent
statistical differences from ANCOVA on organ weights, with carcass weight
as the covariate (p<0.05). ..........................................................................64
xiii
Figure 2.10. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature female sculpin collected in the Little River from the
spring and fall of 2000 and 2001. Sites 1-3 are located in the forested
region (black bars), sites 7 and 9 in the agricultural region (light grey bars),
and site 10 in the urban region (white bar). Within each collection period,
different letters represent statistical differences from ANCOVA on organ
weights, with carcass weight as the covariate (p<0.05). Significant
interactions within ANCOVA are indicated with an asterisk (*). .................65
Figure 2.11. Mean (± SE), liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature female sculpin collected in the Little Forks stream in
November 2000 and 2001. Sites are numbered 1-4 from upstream to
downstream. Within each collection period, different letters represent
statistical differences from ANCOVA on organ weights, with carcass weight
as the covariate (p<0.05). ..........................................................................66
Figure 2.12. Mean (± SE) fecundity (total number of eggs) (top), number of eggs per
gram of gonad weight (middle), and number of eggs per gram carcass
weight (bottom) for female sculpin collected in the Little River and the Little
Forks stream in the spring of 2000 and 2001. Sites 1, 3, and LF2 are
located in forested regions (black bars), sites 7 and 9 in the agricultural
region (light grey bars). Different letters represent statistical differences
within each collection period from ANOVA on total fecundity, and ANCOVA
on fecundity with gonad weight and carcass weight, respectively, as the
covariates (p<0.05). Significant interactions within ANCOVA are indicated
with an asterisk (*). ....................................................................................67
Figure 2.13. Number of eggs per nest found at site 1 (forested; 10 nests) and site 7
(agricultural; 11 nests) in the Little River in May 2001. The ends of the box
represent the 25th and 75th percentile values, whiskers represent the 5th
and 95th percentile values, and the horizontal black line represents the
median value..............................................................................................69
Figure 3.1. Map of study sites located throughout western and northwestern New
Brunswick. Streams were located in areas dominated by forested (F1-F9)
and agricultural (A1-A11) land-use. Circles identify climate stations used
for precipitation data. .................................................................................97
Figure 3.2. Mean (±SE) daily temperatures for agricultural (black line, n=11), and
forested (grey line, n=9) streams of northwestern New Brunswick, 5 July to
8 October 2001. .........................................................................................98
Figure 3.3. Daily rainfall (mm) from 1 July to 1 October, 2001 at weather stations in St.
Leonard (1), Bon Accord (2), and Woodstock (3) (Environment Canada
2001). The numbers (1-3) represent the location of the weather station on
Figure 3.1...................................................................................................99
Figure 3.4. Mean (±SE) dry weight of total sediment (g) deposited in samplers placed
in forested (grey bars) and agricultural (black bars) streams. Data from
three samplers collected at 1, 2, and 3 months were combined. Line at
200g indicates conservative estimate of natural sediment deposition in the
region based on the forested sites...........................................................100
Figure 3.5. Grain size separation of sediment deposited in solid-walled containers with
>200g dry sediment accumulation in forested and agricultural streams. Silt
and very fine to medium fine sand: 0-500µm; coarse to very coarse sand:
500µm-2mm; and gravel: ≥2mm (Cummins 1962). .................................101
xiv
Figure 3.6. Median size of young-of-the-year (YOY) sculpin (A) and sculpin density (B)
versus maximum mean daily water temperatures, and median size of YOY
sculpin versus sculpin density (C). Triangles represent forested (white)
and agricultural (black) sites. ...................................................................102
Figure 4.1. Schematic map of the Little River, New Brunswick catchment indicating
sites where fish were collected for stable isotope analysis. Sites 1-5 were
located within the forested region, site 6 was within a transitional region
influenced by both forest and agricultural inputs, and sites 7-9 were within
the agricultural region, and site 10 was in an urban area. .......................117
Figure 4.2. Mean (±95%CI) δ15N (top) and δ13C (bottom) isotope values (‰) in sculpin
muscle tissues. Site numbers progress downstream from the forested
region (1-5), the transition region (6), into the agricultural region (7-9), and
at the urban site (10). Statistical differences between sites are denoted
with different letters (Tukey’s HSD test, p<0.05)......................................118
Figure 4.3. Mean (±95%CI) δ15N and δ13C plotted together. Site numbers progress
downstream from the forested region (1-5), the transition region (6), into
the agricultural region (7-9), and at the urban site (10). Site-specific
isotopic signatures are more evident and overlap generally only occurs
between adjacent sites. ...........................................................................119
Figure 4.4. Temporal variation in mean (±95%CI) δ15N and δ13C isotope composition of
sculpin collected at particular sites (1, 3, 7, and 9) on the Little River;
November 1999 – black triangles, March 2000 - white diamonds,
November 2000 – black circles, and April 2001 – white squares. ...........120
Figure 5.1. Schematic of the study area on the Kennebecasis River. Fish were tagged
and released in treatment sections A-E. Upon recapture, fish location was
determined within 0.5m longitudinally and 1-2m laterally using flags placed
along the riverbank and by dividing the river width into 5 equal strata.
Generalized habitat types (run versus riffle habitats) are indicated. ........142
Figure 5.2. Frequency of recapture locations throughout the entire study (by section) of
Group I (black bars) and Group II recaptures (grey bars) in the study area
(refer to Figure 1 for study layout)............................................................143
Figure 5.3. Box and whisker plots of movement (m) upstream and downstream for
recaptured PIT tagged sculpin in Group I (11-306d), and Group II (35212d). Vertical lines represent the maximum and minimum movement, and
the upper and lower limits of the boxes represent the 75th and 25th
percentiles, respectively...........................................................................144
Figure 6.1. Progression of a fish population response pattern showing metabolic
disruption with possible recovery pattern after the disruption is removed.
Response is summarized based on age distribution, energy expenditure,
and energy storage, [x x x], respectively showing no change [0], an
increase [+], or decrease [-]. Modified from Gibbons and Munkittrick
(1994).......................................................................................................160
Figure 6.2. Slimy sculpin egg nest on the underneath of a rock found in the Little River.
The colour difference in the egg masses is indicative of multiple females
laying their eggs in nests maintained by male sculpin. ............................161
xv
CHAPTER 1
GENERAL INTRODUCTION
1.1
Cumulative effects assessment
The Canadian Environmental Assessment Act was promulgated in 1995, and
contained a requirement for future environmental impact assessments (EIA) to include
a cumulative effects assessment (CEA). This was in part based on the understanding
that long-term environmental changes can occur as a result of the combined effects of
numerous environmental stressors, and not only as a result of single stressors
(Hegmann et al. 1999). In the past, EIAs were conducted without explicit identification
of the cumulative effects by focusing on stressor-based methods. This method
approaches the predictive assessment on a stressor-by-stressor basis, for example a
single chemical within a discharge or the discharge of single effluent (Munkittrick et al.
2000). Incorporation of the stressor-based approach into the CEA has commonly
involved the desktop identification of existing and potential stressors, valued ecosystem
components (VECs), and the prediction of impacts based on assumed interactions
between the two (Munkittrick et al. 2000). Although this approach has been successful
at identifying and ameliorating environmental impacts, the process contains little sitespecific information and often does not involve actual monitoring to validate the
predictions (Munkittrick et al. 2000).
Although CEA has been a requirement for almost a decade, there remain concerns
about what the CEA is expected to include. Questions may include how to avoid
assessing everything, how to identify what is important to assess, what other factors
external to the stressors should be considered, over what temporal scale should effects
be assessed, and how to determine the significance of the cumulative effects
(Hegmann et al. 1999). Munkittrick et al. (2000) proposed an effects-driven, or effects1
based, assessment framework as an alternate methodological approach to overcome
many of these issues (Figure 1.1). The effects-based approach relies heavily on
system- or site-specific information to design the study and identify what factors may be
important for assessing the cumulative impact of environmental stressors. This
approach allows for the assessment of biological responses without the necessity of
identifying individual stressors by measuring the "accumulated environmental state" of
the system (Munkittrick et al. 2000). The performance of resident fish populations can
be used to identify where existing conditions are compromising performance and help
to understand the level of stress on a river reach. Therefore, it is the biological effects
of resident fish that effectively drives the study design and sampling by identifying
where performance is affected. Then in a general sense, if fish are able to grow,
reproduce and survive at similar rates to those in reference conditions, you may
conclude that there are no measurable limitations to performance. Alternately, if
resident fish are limited in any of the selected performance measures, then you may
conclude that there are environmental factors contributing to the reduced performance
and design a detailed monitoring program to identify and study the relative contribution
of possible stressors. This removes a certain level of researcher bias from dictating
where the effects would be assumed to be occurring, and may possibly identify limiting
or enhancing factors that may have been missed or ignored by the stressor-based
approach. The main differences between the two approaches is the type of data
required for the assessment, and the development of information from baseline
monitoring, adaptive management, and post-analysis monitoring that is not a
component of the stressor-based approach (Munkittrick et al. 2000).
Most cumulative effects assessments or environmental impact assessments,
regardless of the approach, have been applied to point sources of pollution or
environmental impact. Point source pollution can generally be traced back to a single
2
discharge or definitive source such as an industrial or municipal effluent. Non-point
source pollution results from rainfall or snowmelt carrying natural or synthetic pollutants
from diffuse sources in the air or on the ground and depositing them into aquatic
receiving environments such as lakes, streams and rivers, wetlands, coastal water, or
groundwater (US EPA 1996). Non-point source pollution includes agricultural and
urban runoff, salt from irrigation practices, acid drainage from abandoned mines,
atmospheric deposition, hydrologic modification, and bacteria and pathogens from
livestock, pet wastes, or faulty septic systems (US EPA 1996). Much of the research
related to non-point source pollution has been related to water quality monitoring of
parameters such as water chemistry to define whether there is an alteration to the
system due to pollution. Very little comprehensive monitoring has been done to
evaluate the potential long-term biological effects of various non-point sources of
pollution, perhaps due to the inherent difficulties associated with identifying and
defining the inputs. The application of an effects-based cumulative effects assessment
of fish to non-point source pollution seems particularly relevant since (a) the stressors
do not need to be identified a priori, and (b) it is the effects observed in the fish that
identify where the system is compromised, where the sampling effort should be
focused, and therefore what stressors need to be identified.
1.2
Environmental concerns related to agricultural operations
Within the context of ecological risk assessment, the term stressor may be defined
as “any chemical, physical, or biological entity that can induce adverse effects on
ecological components, that is, individuals, populations, communities, or ecosystems”
(Norton et al. 1992). A recent review of the possible threats to Canadian water
resources identified agriculture as the largest non-point source of nutrient loading to the
environment causing impacts such as eutrophication, loss of habitat, changes in
3
biodiversity, fish kills from ammonia toxicity, declines in amphibian populations, and
degradation of drinking water (Chambers et al. 2001a). Introduction of soil or sediment
from agricultural runoff may cause increased turbidity leading to decreased
photosynthesis, interference with animal behaviours related to sight, impedance of
feeding and respiration, and degradation of habitats required for feeding and spawning
(Chambers et al. 2001b). Agricultural runoff can also introduce significant amounts of
pesticides that can cause fish kills and have potential sublethal effects including
alterations to respiration, reproduction, growth and development, and photosynthesis
(Chambers et al. 2001b). Beyond these direct agricultural stressors there may also be
a multitude of associated indirect stressors including increased water temperatures,
disruption of aquatic food webs, and alteration of local hydrology and habitats (Karr and
Schlosser 1978; Waters 1995). In the recent US Environmental Protection Agency’s
National Water Quality Inventory, the leading source of impairment for rivers and
streams was agricultural activity (US EPA 2002).
In terms of the chemical inputs, there is worldwide concern that environmental
contamination by industrial and agricultural chemicals and other human activities can
disrupt the endocrine systems of wild animals. Agriculture has been identified as a key
research priority area to investigate potential sources of endocrine disruption of wild
animals (Environment Canada 1997; Servos et al. 2000). Although there is yet to be an
official list compiled, there are about 70 pesticides, or components of pesticide
formulations, that have been identified as potential endocrine disrupting compounds
(Kaminuma 2001). In Canada, there are currently 550 pesticides registered under the
Pest Control Products Act, of which 80% are registered for agricultural use (Maguire et
al. 2001).
4
1.2.1
Potato productions
In terms of agricultural activities, potato farming is amongst the most
environmentally degrading, both terrestrially and aquatically (Cairns 2002). Potato
production involves removing large quantities of organic matter from the soil, and high
soil erosion risk as the soil is left bare for long periods, and attracts a wide variety of
pests, generally controlled by relatively intensive chemical applications. A typical
cultivation year of a potato lasts about 7 months (Table 1.1). The majority of the
fertilizer applications occur from the beginning of field preparation and into July.
Insecticides are applied from June forward as insects emerge or arrive at the potato
plants, though most insecticide application is in response to crop scouting reports.
Fungicides, mainly for late blight, are applied multiple times from June or July through
to harvest. Herbicides are applied early in field preparation, as weed control, and then
as a top killer in preparation for harvesting. The top killer is used to slowly kill the leafy
part of the potato plant while allowing the tubers to gain size by converting sugars to
starches and promoting the formation of an outer periderm (Industry Canada 1999).
Depending on conditions during the growing season, a farmer may apply up to 19
pesticide applications (Gray et al. 2000). The amount of tillage during the cultivation
period is an important factor for soil condition, with increased compaction resulting in
reduced soil productivity, decreased rooting depth, and increased potential for erosion
(PEI DAF 1998).
1.3
Previous research on fish in agricultural regions
Despite the fact that agriculture has been identified as the primary activity
responsible for the loss of fish species in the midwestern US (Karr et al. 1985), there
are notably few studies on the sublethal biological impact of exposure to multiple
agricultural stressors. Most of the research on the biological impact of agricultural
5
activities has focused on lab exposures to pesticides or the effect of sediment
deposition on salmonid spawning success (see Waters 1995). From a monitoring
perspective, studies conducted in agricultural regions have focused on fish and
invertebrate density and abundance as performance measures. A 1974 survey of
western New Brunswick streams found that farmed catchments had an average of 52%
fewer brook trout (Salvelinus fontinalis), 92% fewer slimy sculpin (Cottus cognatus),
and 64% less benthic invertebrates than control regions (Welch et al. 1977). A
catchment monitoring study in Iowa found that the stream most influenced by
agriculture had reduced fish populations, but looked only at density and community
assemblage as indications of change (Langel et al. 2001). In fact, the US EPA
recommends a biological assessment in impaired streams and small rivers but
considers only fish community, diversity, and abundance as measures of the health of
the system (Barbour et al. 1999). Fitzgerald et al. (1999) is one of the only published
studies to examine individual fish performance parameters of fish residing in Southern
Ontario streams receiving urban and agricultural inputs and observed impacts on
young-of-the-year and mature female creek chub (Semotilus atromaculatus).
In 1998, a laboratory and field research study was initiated on Prince Edward Island
to determine the potential for pesticides to affect the endocrine systems of fish
downstream of potato fields (Gray et al. 2000). Prince Edward Island was chosen as
an ideal study location due to the intensity of potato production, with about 42 000-45
000 hectares of land under potato cultivation in the past 5 years (Statistics Canada
2003). In order to ensure exposure to runoff from potato fields, the field component
consisted of caging hatchery-reared rainbow trout (Oncorhynchus mykiss) in streams
downstream of potato fields. The results from the caging study showed that there were
no consistent changes to fish condition factor (weight/length3*105) related to
surrounding land use, and vitellogenin induction (an indication of exposure of male fish
6
to estrogenic compounds) and hormonal analyses were inconclusive (Gray et al. 2000).
In the lab study, sediments collected from waterways draining the agricultural regions
showed effects on egg survival, development, and time-to-hatch of Japanese medaka
(Oryzias latipes) (Gray et al. 2000). Although the results did not provide evidence that
the effects were being mediated by the endocrine system, the impacts of early-life
stage exposure of fish to sediments from agricultural regions prompted interest in
applying a more appropriate field design to evaluate the impacts of potato production
on fish populations.
1.3.1
Background to the present study
Prince Edward Island did not provide the opportunity to conduct an effects-based
assessment because there are relatively few wild fish species present, and in many
cases the number of fish is insufficient to sustain a long-term lethal sampling
component. Instead New Brunswick was chosen because most catchments support a
much more diverse and abundant fish community, and because there is also a
significant potato production industry. Until recently, New Brunswick was the second
largest producer of potatoes in Canada behind PEI, with an annual production of about
22 000-23 000 hectares (NB DAFA 2002). The total value of the New Brunswick potato
industry is $400 million annually, including value added (NB DAFA 2002). The
dominant region of potato production is centred in the Saint John River valley from
Woodstock to Grand Falls. Together the western and northwestern regions account for
about 90% of the entire province’s potato production (NB DAFA 2002).
Our laboratory had begun a cumulative effects assessment of the St. John River
system in New Brunswick during the summer of 1999 to assess the effects on the
resident fish of industrial and municipal inputs, hydroelectric dams, and forestry and
agricultural land-use practices (Munkittrick et al. 2002). In order to separate out the
7
potential effects of agriculture from the other point and non-point stressors, we
designed a study to focus on tributaries of the St. John River and ask the question, are
there effects on fish populations of agricultural regions within the St. John River basin?
Coincidently, during the summer we initiated the research study in New Brunswick,
the impacts of potato production became a much more public issue due to a significant
increase in the frequency and magnitude of fish kills occurring in PEI streams (Table
1.2). Although there were difficulties finding direct evidence that pesticides were
responsible for the fish kills, the occurrence in cultivated catchments receiving pesticide
applications and immediately following rainfall events, suggests that the runoff from
potato fields was responsible (Mutch et al. 2002). The absence of a fish kill event
during the summer of 2001, a very dry year, also suggests a link between the fish kills,
precipitation, and runoff.
1.4
Performance measures for effects-based fish assessment
The methods developed by Munkittrick et al. (2000) for an effects-based cumulative
effects assessment of fish in the Moose River basin in northeastern Ontario were used
as a framework for applying an effects-based assessment of fish in the agricultural
region of the St. John River basin. The level of organization was chosen a priori to
focus research efforts at the fish level within the food web and conduct a “top-down”
monitoring approach. While monitoring studies may choose to conduct a “bottom-up”
approach and start with a lower trophic level, we felt that because the biological
impacts of agricultural stressors were not well understood it was important to establish
first whether there were impacts at a higher level of organization. The level of
biological organization within which to measure responses can be a ‘philosophical’
decision, but also one based on the purpose of the assessment (Munkittrick et al.
2000). Changes in the performance of an organism can be evident at the individual,
8
population and community levels of organization (Figure 1.2). As you progress from
the introduction of a pollutant to the ecosystem level, there is an increase in the
ecological relevance and importance of responses, but also an increase in the time lag
for detection of response and an increasing difficulty to link responses to particular
pollutants or stressors. Collection of information at the individual level offers the
capacity to measure responses that are relevant for the population level, but which are
also directly linked to physiological indicators that may be used to establish causeeffect relationships (Munkittrick et al. 2000).
Individual level performance measures were chosen to reflect the growth,
reproduction, and population structure of sample fish populations. Relative fish size,
and size of fish at the end of the growing season can be used as surrogates for growth,
liver size and fish condition as indicators of energy storage, gonad size and fecundity
as indicators of reproductive investment, and size distributions as an indication of the
survival or performance of different age classes.
1.4.1
Choice of monitoring species
The resident fish community limited the research in agricultural regions of PEI. The
need to cage hatchery-reared rainbow trout in order to ensure exposure and sufficient
samples sizes was problematic. First, there were concerns about accidental
introduction of rainbow trout in systems where they had not previously been introduced.
Second, caging fish that had been raised under hatchery conditions was likely stressful
for the fish as they were placed in a very foreign environment and were not receiving
food multiple times per day. Under these circumstances, any responses would need to
be carefully regarded within the context of the methodology.
In order to choose an appropriate study species for the New Brunswick studies, a
preliminary fish community survey was conducted in the Little River catchment, located
9
just north of Grand Falls (Figure 2.1). The catchment is located in the northwestern
region of the major potato production area. Fish collected in the preliminary survey
included brook trout, slimy sculpin, blacknose dace (Rhinichthys atratulus), three-spine
stickleback (Gasterosteus aculeatus), and creek chub. Although brook trout were
collected throughout the system, their abundance was low at some sites in the
agricultural region. The only species collected throughout the system and in sufficient
numbers at all sites was the slimy sculpin.
The slimy sculpin displays many characteristics that make it a particularly suitable
sentinel species. In the past, the choice of large fish species for environmental
monitoring was due more to their commercial importance rather than their suitability as
a sentinel species (Gibbons et al. 1998a). Gibbons et al. (1998a,b) recently proposed
the use of small-bodied, non-migratory fish species as alternative sentinel species to
overcome issues related to residency and exposure, abundance, response time to
environmental changes, and lack of commercial or sport fishing pressures. The slimy
sculpin is a cool-water fish species with a very broad North American geographical
distribution ranging from Virginia to the Arctic, and from coast to coast (Scott and
Crossman 1998). The sculpin is reported to have small home ranges due in part to a
lack of swim bladder and to territorial behaviour (Van Vliet 1964). Various markrecapture studies of freshwater sculpins have found that the majority of recaptured fish
moved <50m (McLeave 1964; Brown and Downhower 1982; Hill and Grossman 1987;
Morgan and Ringler 1992). It is intuitive that the greater the probability that a fish
resides in a particular area, the greater its value as a sentinel species for monitoring
whether environmental conditions in that area are causing biological impacts on fish.
The average life span of the sculpin is 4-5 years, reaching sexual maturation relatively
quickly in 1-2 years (Anderson 1985; Scott and Crossman 1998). In general, a shorter
lifespan means that they may respond earlier to environmental changes, possibly
10
showing alterations in reproduction and growth faster than longer-lived species
(Munkittrick et al. 2000).
1.5
Objectives and outline of thesis
The main objective of the thesis research was to determine whether there were
sublethal impacts of agricultural stressors on fish residing in streams receiving inputs
from agricultural activities. We focused the fish assessment in the Little River
catchment due to its location within New Brunswick’s intense potato producing region.
The uppermost region of the catchment is predominantly forested while agricultural
fields dominate the landscape in the lower region. In terms of study site design, the
upper, forested region was considered the reference area because there were no
upstream influences from agriculture. The collection of information from the forested
region provided important basic information about the life history of the species in order
to be able to appropriately interpret the presence or absence of responses.
The primary objective of the second chapter of the thesis was to conduct an effectsbased fish assessment and determine whether there were differences in the health and
performance of slimy sculpin residing in forested and agricultural regions of the Little
River. Secondary objectives included determining whether the responses were
observable over time. The results from the research conducted on sculpin in the Little
River are presented in Chapter 2, “An effects-based assessment of slimy sculpin
(Cottus cognatus) populations in agricultural regions of northwestern New Brunswick”.
After identifying a variety of responses in sculpin in the Little River system, the next
question was whether some of the same responses could be confirmed in other
streams and systems experiencing inputs from agricultural activities. The focus of the
second study was on the observations that proportions of young-of-the-year (YOY)
sculpin were reduced and that sculpin tended to be larger in the agricultural region in
11
the Little River. We predicted that there would be fewer but larger sculpin present in
streams in agricultural regions compared with forested regions. As a basic template,
we used a study conducted by Fisheries and Oceans in 1974, that looked at sediment
deposition and the density of fishes and invertebrates in control streams compared with
those influenced by clearcut forestry or agricultural in northwestern New Brunswick
(Welch et al. 1977). We measured sediment deposition, temperature, and fish size and
density at 20 sites on 19 different streams; 9 in forested and 11 in agricultural regions.
The results from this work are presented in Chapter 3, “Investigating the impacts of
sediment and temperature on slimy sculpin (Cottus cognatus) populations in
agricultural catchments”.
Another objective of the overall thesis research was to evaluate the slimy sculpin as
a sentinel environmental monitoring species. The use of sentinel species is rooted in
the principle that the status of a sentinel will reflect the overall condition of the aquatic
environment in which the fish reside (Munkittrick 1992). As previously stated, an
important assumption of sentinel monitoring is that fish collected at specific sites will
exhibit responses that reflect their local environment (Gibbons 1998a). We used stable
carbon and nitrogen isotopes to determine whether sculpin demonstrated site-specific
isotope signatures related to their residency within certain parts of the river. If they
were moving substantially within the system, and thus feeding over a broad spatial
scale, their isotopic signatures would overlap. Alternatively, if they remained within a
more restricted area, they could display site-specific isotopic signatures as a
consequence of the local food sources. The results from that study are presented in
Chapter 4, “The use of stable isotope analysis to assess the site fidelity of slimy sculpin
(Cottus cognatus)”.
An ideal fish species would also then be one that does not move freely throughout
the system, and remains within a relatively small spatial scale over a relatively long
12
temporal scale. The cryptic and benthic behaviour of the sculpin also suggests that is
unlikely to actively move great distances. To further address the issue of whether the
sculpin is likely to reflect local conditions due to limited mobility, we inserted passive
integrated transponder (PIT) tags into slimy sculpin and monitored their movements
over a period of 10 months. The results from that work are contained in Chapter 5,
“Measuring small-scale movements of the slimy sculpin (Cottus cognatus) to assess
movements in a small river”.
The final chapter, Chapter 6, provides a discussion and summary of the response of
slimy sculpin in agricultural regions of New Brunswick and a discussion of the suitability
of the slimy sculpin as a sentinel monitoring species for non-point source pollution. The
issue of intense potato production in the region is considered with respect to the timing
of input of potentially critical stressors. Suggestions for future research needs and
conclusions from the research are also provided.
13
Table 1.1. A generalized potato farming operation for one season of potato cultivation in Eastern Canada. Chemical
applications include fertilizers (Fe), insecticides (I), fungicides (Fu), and herbicides (H).
Month
Activity
Chemical applications
April
field preparation
Fe
late April: plant early maturing varieties
May
mid to late May: plant late maturing varieties
Fe, I ,H
June
spray program begins for blight
Fe, Fu. I
emergence of Colorado Potato Beetle
July
crop scouting for insects
Fe, Fu, I
spray program for blight continued
mid July: “new potatoes” harvested
August
crop scouting for insects and virus and diseased plants removed
Fu, I, H
spray program for blight continued
top kill for seed potato crops
September
spray program for blight continued
Fu, I, H
crop scouting scaling down
mid Sept: top kill processing potato crop
mid Sept: harvest seed potato crop
October
2nd application of top kill for processing crop
H
harvest processing crop
plough fields for next season
Sources: Industry Canada (1999) and H Rees, Agriculture and Agri-Food Canada (pers. comm.)
14
Table 1.2. Occurrence of fish kill events recorded and total annual rainfall (mm)1 on
Prince Edward Island (1990-2002)2. Blank cell indicates data not available.
Year
Number of fish kills
Number of dead fish
Total rainfall
reported
reported
(mm)
1990
1
200-300
1251.7
1991
0
>2000
813.7
1992
0
664.6
1993
0
591.5
1994
2
719.6
1995
2
1996
1
1012.1
1997
0
633.0
1998
1
1055.6
1999
8
2000
5
976.0
2001
0
494.8
2002
9
>35 0003
>5000
773.3
918.6
~20 000
1 – Annual rainfall data from meteorological station at Long River, PEI (46º30N, 63º33W),
Environment Canada (2001).
2 – Modified from Mutch et al. (2002) and B Ernst (pers. comm.).
3 – All salmon parr in semi-natural rearing pond, plus all fish in a 4km stretch of river.
15
Effects-based Assessment
Define the system
Develop key indicators
Develop performance assessment
Data analysis
Identify impaired aspects
Identify limiting factors and potential
interactions
Identify critical stressors
Confirm diagnosis
Risk management
and remediation
Risk assessment:
develop predictive model
Figure 1.1. Framework for conducting an effects-based cumulative effects assessment
(Munkttrick et al. 2000).
16
Increasing importance and ecological relevance
Ecosystem
Community
composition
Population
changes
Whole organism
responses
Physiological
changes
Biochemical
changes
Pollutant
Increasing response time
Increasing difficulty of linkage to specific chemicals
Figure 1.2. Biological levels of organization where responses may be measured. The
difficulty of measurement, response time for changes, and ecological significance
increase with increasing biological levels.
17
1.6
References
Anderson CS. 1985. The structure of sculpin populations along a stream size gradient.
Environ. Biol. Fishes. 13: 93-102.
Barbour MT, Gerritsen J, Snyder BD, and Stribling JB. 1999. Rapid Bioassessment
Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic
Macroinvertebrates and Fish, 2nd edition. EPA 841-B-99-002. US Environmental
Protection Agency, Office of Water, Washington, DC, USA.
Brown L, and Downhower JF. 1982. Summer movements of mottled sculpin, Cottus
bairdi (Pisces: Cottidae). Copeia. 1982:450-453.
Cairns DK. 2002. Land use and aquatic resources of Prince Edward Island streams and
estuaries: an introduction. In Effects of land use practices on fish, shellfish, and their
habitats on Prince Edward Island. Cairns DK (ed). Can. Tech. Rep. Fish. Aquat. Sci.
no. 2048 pp. 1-13.
Chambers PA, Guy M, Roberts E, Charlton MN, Kent R, Gagnon C, Grove G, Foster
N, DeKimpe C, and Giddings M. 2001a. Nutrients. Section 6. Threats to Sources of
Drinking Water and Aquatic Ecosystem Health in Canada. NWRI Scientific
Assessment Report Series No.1. National Water Research Institute, Burlington, ON,
Canada. pp 23-36.
Chambers P, DeKimpe C, Foster N, Goss N, Miller J, and Prepas E. 2001b.
Agricultural and forestry land use impacts. Section 13. Threats to Sources of
Drinking Water and Aquatic Ecosystem Health in Canada. NWRI Scientific
Assessment Report Series No.1. National Water Research Institute, Burlington, ON,
Canada. pp 57-63.
Environment Canada. 1997. Identifying Research Needs and Priorities. Proceedings of
the Environment Canada Workshop on Endocrine Disruptor Compounds. Servos M,
Luce S (eds). October 23-24. Niagara Falls, ON, Canada.
Environment Canada. 2001. Canadian daily climate data. CD-ROM.
Fitzgerald DG, Lanno RP, and Dixon DG. 1999. A comparison of a sentinel species
evaluation using creek chub (Semotilus atromaculatus) to a fish community
evaluation for the initial identification of environmental stressors in small streams.
Ecotoxicol. 8:33-48.
Gibbons WN, Munkittrick KR, McMaster ME, and Taylor WD. 1998a. Monitoring aquatic
environments receiving industrial effluents using small fish species 1: response of
spoonhead sculpin (Cottus ricei) downstream of a bleached-kraft pulp mill. Environ.
Toxicol. Chem. 17:2227-2237.
18
Gibbons WN, Munkittrick KR, McMaster ME, and Taylor WD. 1998b. Monitoring aquatic
environments receiving industrial effluents using small fish species 2: comparison
between responses of trout perch (Percopsis omiscomaycus) and white sucker
(Catostomus commersoni) down of a pulp mill. Environ. Toxicol. Chem. 17:22382245.
Gray MA, Teather KL, Sherry J, McMaster M, Hewitt M, and Mroz R. 2000. Endocrine
disrupting potential in freshwater ecosystems near agricultural areas on Prince
Edward Island. EPS-5-AR-99-6. Environment Canada Surveillance Report.
Environment Canada, Dartmouth, NS, Canada.
Hegmann G, Cocklin C, Creasey R, Dupuis S, Kennedy A, Kingsley L, Ross W, Spaling
H, and Stalker D. 1999. Cumulative Effects Assessment Practitioners Guide.
Prepared by AXYS Environmental Consulting Ltd. and the CEA Working Group for
the Canadian Environmental Assessment Agency, Hull, PQ, Canada.
Hill J, and Grossman GD. 1987. Home range estimates for three North American
stream fishes. Copeia. 1987:376-380.
Industry Canada. 1999. The potato: then & now.
http//:collections.ic.gc.ca/potato/index.asp. Downloaded information 15 July 2003.
Karr JR, and Schlosser IJ. 1978. Water resources and the land-water interface.
Science. 201:229-234.
Karr, JR, Toth LA, and Dudley DR. 1985. Fish communities of midwestern rivers: a
history of degredation. Bioscience. 35:90-95.
Kaminuma T. http://www.nihs.go.jp/hse/endocrine-e/paradigm/pesticide/pest-list-e.html.
Updated 17 November 2001, downloaded 2 August 2003.
Langel RJ, Seigley LS, Wilton TF, May JE, Schueller MD, Birmingham MW, Wunder G,
Polton V, Tisl J, and Palas E. 2001. Sny Magill Nonpoint Source Pollution
Monitoring Project, Clayton County, Iowa: Water Years 1995 – 1998. Iowa
Department of Natural Resources, Iowa City, IA, USA.
Maguire RJ, Sibley PK, Solomon KR, and Delorme P. 2001. Pesticides. Section 3.
Threats to Sources of Drinking Water and Aquatic Ecosystem Health in Canada.
NWRI Scientific Assessment Report Series No.1. National Water Research Institute,
Burlington, ON, Canada. pp 9-12.
McCleave JD. 1964. Movement and population of the mottled sculpin (Cottus bairdi
Girard) in a small Montana stream. Copeia. 1964:506-512.
Morgan CR, and Ringler NH. 1992. Experimental manipulation of sculpin (Cottus
cognatus) populations in a small stream. J. Freshw. Ecol. 7:227-232.
Munkittrick KR. 1992. A review and evaluation of study design considerations for sitespecifially assessing the health of fish populations. J. Aquat. Ecosys. Health. 1:283293.
19
Munkittrick KR, McMaster ME, Van Der Kraak G, Portt C, Gibbons WN, Farwell A, and
Gray M. 2000. Development of methods for effects-driven cumulative effects
assessment using fish populations: Moose River project. Society of Environmental
Toxicology and Chemistry. Pensacola, FL, USA.
Munkittrick KR, Hewitt M, Curry A, Cunjak R, and Van Der Kraak G. 2002.
Development of a cumulative effects assessment strategy for the Saint John River.
Final Report for Toxic Substances Research Initiative (TSRI) Project No. 205.
Mutch JP, Savard MA, Julien GRJ, MacLean B, Raymond B, and Doull J. 2002.
Pesticide monitoring and fish kill investigations on Prince Edward Island, 19941999. In Effects of land use practices on fish, shellfish, and their habitats on Prince
Edward Island. Cairns DK (ed).Can. Tech. Rep. Fish. Aquat. Sci. no. 2048 pp. 94115.
NB DAFA. 2002. Annual report 2000-2001. New Brunswick Department of Agriculture,
Fisheries, and Aquaculture. Fredericton, NB, Canada.
Norton SB, Rodier DJ, Gentile JH, Van Der Schalie WH, Wood WP, and Slimak MW.
1992. A framework for ecological risk assessment at the EPA. Environ. Toxicol.
Chem. 11:1663-1672.
PEI DAF. 1998. Best management practices: Soil conservation for potato production.
PEI Department of Agriculture and Forestry. Charlottetown, PE, Canada.
Scott WB, and Crossman EJ. 1998. Freshwater fishes of Canada. Galt House
Publications Ltd., Oakville, ON, Canada.
Servos M, Delorme P, Fox G, Sutcliffe R, and Wade M. 2000. Proceedings of the 5NR Workshop: Establishing a national agenda for the scientific assessment of
endocrine disrupting substances.
Statistics Canada. 2003. Canadian potato production: Statistical bulletin. Vol. 1. No. 1.
22-008-XIE. Statistics Canada, Ottawa, ON, Canada.
US EPA. 1996. Nonpoint source pollution: the nation’s largest water quality problem.
EPA-841-96-004A. Pointer No.1. US Environmental Protection Agency.
Washington, DC, USA.
US EPA. 2002. National Water Quality Inventory 2000 report. EPA-841-R-02-001. US
Environmental Protection Agency. Washington, DC, USA.
Van Vliet WH. 1964. An ecological study of Cottus cognatus (Richardson) in northern
Saskatchewan. MA Thesis, University of Saskatchewan, Saskatoon, SK, Canada.
Waters TF. 1995. Sediment in streams: Sources, biological effects and control.
American Fisheries Society Monograph 7. Bethesda, MD, USA.
Welch HE, Symons PEK, and Narver DW. 1977. Some effects of potato farming and
forest clearcutting on small New Brunswick streams. Technical Report No. 745.
Fisheries and Marine Service, St. Andrews, NB, Canada.
20
CHAPTER 2
An effects-based assessment of slimy sculpin (Cottus cognatus) populations in
potato agriculture regions of northwestern New Brunswick1
2.1
Abstract
Recently in Atlantic Canada, there has been increased concern associated with
potato farming as a result of an increase in the frequency and magnitude of fish kills
downstream of cultivation activities following major storm events. Over a period of
three years, we monitored the population structure and physiological performance of
slimy sculpin (Cottus cognatus) in an intensive potato cultivation region of northwestern
New Brunswick. The slimy sculpin, a small-bodied benthic fish, was considered
suitable for monitoring due to its natural abundance throughout the system, limited
mobility, lack of fishing pressures, and easily measured life history characteristics.
Rather than focus on particular agricultural stressors, an effects-based assessment of
the fish in the system was conducted to determine whether there were observable and
consistent responses of sculpin in the agricultural region. We found that the local
population structure at agricultural sites consisted of fewer young-of-the-year fish in two
of three years. In the spring, the number of sculpin eggs per nest was reduced in an
agricultural site versus a forested site. In comparison with forested reaches, adult
sculpin were larger, but with smaller gonads, and females had smaller livers, and fewer
and smaller eggs. These biological responses were reduced in the fall of 2001
following slightly drier conditions than previous two years. The effects-based approach
successfully demonstrated biological impacts on sculpin temporally and spatially and
1
This chapter is currently in preparation for submission to the journal Water Quality
Research Journal of Canada under joint authorship with Kelly R. Munkittrick.
21
therefore the species’ potential for studying non-point source impacts in environmental
assessments.
2.2
Introduction
An increase in the frequency and magnitude of fish kills downstream from potato
fields in Prince Edward Island (Mutch et al. 2002) is one reason the issue of non-point
source pollution has recently become a focus of concern in eastern Canada. Between
1990 and 1998, PEI experienced a total of 7 fish kills, while there have been 22 fish
kills between 1999 and 2002 (Mutch et al. 2002; B Ernst, Environment Canada, pers.
comm.). Although pesticides are implicated in most lethal events, aquatic stressors
from row crop cultivation may include other direct inputs such as nutrients and soil, and
indirect effects such as increased temperatures (e.g. associated with clearing of bankside vegetation), habitat alteration, impacts on food sources, changes in local
hydrology, and reduced dissolved oxygen (Karr and Schlosser 1978; Waters 1995).
Surprisingly, there are few studies that have attempted to investigate the in situ
biological effects on fish residing in the streams or rivers in agricultural regions. Most
field research and monitoring of aquatic systems in agricultural regions have looked for
changes in diversity and abundance at the community level (e.g. Barbour et al. 1999),
while laboratory studies have focused largely on relatively short-term exposure to
individual pesticides (e.g. Teather et al. 2001). The sublethal biological impacts of nonpoint source agricultural stressors over the long-term was recently identified as a
significant research deficiency by Environment Canada (Environment Canada 2001a).
The traditional ecotoxicological approach is to study the environmental
contamination and the potential responses to exposure. This stressor-based approach
depends on correlation or extrapolation to evaluate the potential ecological
22
consequences of exposure. The consequence of focusing on the stressors is that
many assessments focus heavily on documenting the potential impacts of known
stressors and known impact pathways, and ignore the potential interaction of the
variety of potential stresses and unknown local environmental factors. The value of the
predicted effects is often limited by the lack of site-specific information, the correlative
nature of predictions, and reliance upon known or established relationships between
stressors and organisms (Munkittrick et al. 2000). Farm-specific differences in
chemical use patterns, the timing of applications, and local differences in rainfall
patterns and water extractions for irrigation all affect the relevance of predictive work for
assessing potential agricultural impacts. Interpretation of the impacts of agricultural
activity is further complicated by the potential unknown interactions of chemical
exposures with a wide variety of other potential stressors mentioned above.
Munkittrick et al. (2000) proposed an effects-driven, or effects-based, biological
assessment as an alternate approach to overcome situations where limiting or
enhancing factors may be missed or ignored by the stressor-based approach. The
effects-based approach relies heavily on system- or site-specific information to design
the study and identify what factors may be important for assessing the cumulative
impact of multiple environmental stressors. This approach allows for the assessment of
biological responses without the necessity of identifying individual stressors by
measuring the "accumulated environmental state" of the system, and is particularly well
suited for studying the impacts of non-point source pollution like agriculture. The
performance of resident fish populations is used to identify where existing conditions
are compromising performance and help to understand the level of stress on a river
reach. The investigation into what, in particular, may be causing the observed
responses is restricted to follow-up studies once the existence of ecologically relevant
responses has been documented (Munkittrick et al. 2000).
23
With annual potato cultivation of approximately 22 to 23 000 hectares, the potato
industry in New Brunswick is valued at about $400 million annually (NB DAFA 2002).
The northwestern region of New Brunswick is responsible for about 35% of the potato
production in the province (NB DAFA 2000). One particularly intensive potato
cultivation area within the region is the Little River catchment near Grand Falls, NB.
Although it is estimated that only 15% of the catchment is cultivated (H Rees,
Agriculture and Agri-Food Canada, pers. comm.), all of the agricultural activity is
concentrated in the lower catchment, where agricultural development has reached its
capacity. Availability of land for agricultural expansion continues to be an issue in this
area (NB DAFA 2000). A preliminary fish community survey in the Little River
catchment identified the slimy sculpin (Cottus cognatus) as the only fish species
present in sufficient numbers for assessing effects throughout the catchment (Gray et
al. 2002).
The objectives of this study were to conduct an effects-based assessment of the
potential impacts of non-point source agriculture on the slimy sculpin. Over a threeyear period, we compared the health and reproductive performance of sculpin residing
in agricultural and forested regions to determine whether possible chronic exposure to
non-point source agricultural stressors had observable and consistent impacts at the
whole-organism level. The effects-based assessment also documented baseline
conditions of sculpin performance, yearly and seasonally, and trends in responses over
time in the forested region. We used various performance parameters to evaluate
whether fish were able to survive, grow, and reproduce similar to fish at reference
locations (Munkittrick and McMaster 2000).
24
2.3
Methods
Study sites
The Little River (47º03’08.4”N, 67º44’20.8”W) drains a total area of about 340km2,
and is located northeast of Grand Falls, New Brunswick (Figure 2.1). The mainstem
Little River is a 4thorder stream that originates in a forested landscape and drains
predominantly agricultural lands in its lower reaches. The Little River terminates at the
St. John River immediately upstream of Grand Falls. Sculpin sampling sites were
classified a priori based on the dominant surrounding land-cover that was forested, a
transition area between forestry and agriculture, an area where agriculture was the
predominant land-use, and the farthest downstream site which was predominantly
urban (Grand Falls has a population of just under 6000). Average width, depth,
velocity, and cobble size (i.e. largest rocks located along random transects) was
measured in August 2000 at some of the main sites used in the study (Table 2.1).
From the farthest upstream to the farthest downstream site, the Little River increases in
width and velocity, with the exception of a narrower and deeper section at site 9. Mean
cobble size varied over a relatively small range of about 9 cm between sites 1 to 10
(corresponding stream distance of 30km). The mean annual discharge was estimated
at approximately 3.2 x 107 m3.
In the fall of 2000 and 2001, and the spring of 2001 sculpin were also collected
along the Little Forks Branch (47º30’25.9”N, 68º11’57.4”W) of the Green River to
provide reference information of sculpin populations along an upstream-downstream
gradient in a river uninfluenced by agriculture. The Little Forks stream is a 4th order
stream, which drains a completely forested catchment area of about 150km2, and is
located about 62km northwest of Grand Falls (Figure 2.1). The Little Forks stream
increases in size similar to the Little River, with a width of 10m and 19m at the
25
uppermost (LF1) and lowermost (LF4) sites, respectively (corresponding stream
distance of 12km). The two streams had relatively comparable depth, velocity, and
cobble size at the sites chosen for sculpin collections.
Fish collections
The main objective of the first fish collection in August 1999 was to characterize the
fish community in the Little River catchment in order to choose fish species to use for
monitoring. Sites were enclosed with barrier nets and fish were collected using a
backpack electrofishing unit (Smith-Root C-15), with at least two persons with dipnets
(mesh size 4mm stretched). All fish were identified, measured for length (±1mm) and
weight (±0.1g), and released back into the site when fishing was completed. Fish
species collected included brook trout (Salvelinus fontinalis), three-spine stickleback
(Gasterosteus aculeatus), creek chub (Semotilus atromaculatus), blacknose dace
(Rhinichthys atratulus), and slimy sculpin (Cottus cognatus). Brook trout and slimy
sculpin were the only species present at all sites, three-spine stickleback were present
only at some forested sites, and blacknose dace and creek chub were present only at
some agricultural sites. Although brook trout were present throughout the basin, they
were present only in very small numbers at some agricultural sites (<5 per site),
precluding its use as a monitoring species for the entire system. The slimy sculpin was
the only fish species present at all sites with sufficient densities to be considered for
monitoring.
Slimy sculpin were sampled at various sites along the Little River from 1999 to 2001
and in the Little Forks Branch of the Green River from 2000 to 2001 (Table 2.2). When
and where possible, sampling consisted of the collection of 100 sculpin, regardless of
fishing effort or stream area covered. All captured fish were measured for length
26
(±1mm) and weight (±0.01g). As slimy sculpin do not have reliable secondary sex
characteristics, the 30 largest sculpin were retained for lethal sampling in order to get
sufficient sample size of both male and female sculpin. In the fall of 2000, in an attempt
to reduce the amount of lethal sampling and to reduce variability due to size, only 20
fish were kept, with 60-70mm as the target size range. Unequal sex ratios resulted in
low sample sizes for some sexes and subsequent lethal sampling required a minimum
of 30 mature fish again. The remaining fish not kept for lethal sampling were returned
to the site of collection after measurements were taken.
For the purposes of lethal sampling, the sculpin were sacrificed by severing the
spinal cord. In general, the liver and gonad were removed and weighed (±0.01g). With
the exceptions of November 1999 and May 2000, the stomach was also removed and
weighed in order to calculate carcass (i.e. eviscerated) weights. The fecundity (total
number of eggs) of female sculpin sampled in the spring collections was also
determined. Fecundity was not determined in the autumn collections as the female
gonads were in early recrudescence. Carcasses were frozen for stable isotope
analysis (Chapter 4).
Slimy sculpin were collected on three separate occasions in the spring of 2000
(early and late March, and early May) to observe the changes in fish condition and
gonad development during pre-spawning conditions. Mature males and females were
sampled in early and late March, and only females were sampled in early May. The
sex of mature sculpin was easily determined during these collections due to the greatly
enlarged abdomen of the female sculpin.
In the autumn of 2000, the objective was to collect sculpin of similar sizes at three
sites along each of the Little River and the Little Forks Branch of the Green River to
compare fish from a forest to agricultural continuum of land-use and a solely forested
river. A similar size range of fish was chosen for sampling in order to compare fish
27
across different rivers. Collections were not successful for adults at the lowermost Little
River site (Site 10); consequently this site was excluded from the analysis. Due to high
water levels following a rainstorm, site 7 was not sampled until two weeks after site 1 in
November 2000.
In May 2001, we conducted a sculpin nest survey at site 1 and site 7. Nests were
identified by overturning rocks that appeared to have had a recent excavation beneath
the downstream side, and also by collecting guarding males by electrofishing and
flipping rocks in the vicinity of their capture. Guarding males are easily distinguished
due to dark pigmentation around the head and pectoral fins (Van Vliet 1964). Eggs
were removed from the nest and counted.
Air and water temperature
Climate data compiled by Environment Canada at the St. Leonard airport (Figure
2.1), located approximately 13km northwest from the mouth of the Little River and
48km southeast from the mouth of the Little Forks Branch stream, were used to
examine air temperature and rainfall amounts in the region between 1998 and 2001
(Table 2.3). Data from 1998 were included to indicate what the conditions were like
previous to our first sculpin collections. There was a decrease in annual rainfall
amounts over the four years. The wettest spring (Apr-Jun) was in 2000, and the
wettest summer (Jul-Sep) and fall (Oct-Dec) were both in 1999. Mean monthly air
temperatures indicate that 1999 experienced a slightly warmer spring and summer,
2001 had the warmest fall, and overall 2000 was the coolest year (Table 2.3). Mean
summer air temperatures (July-September) were 16.0, 17.1, 14.8, and 16.5ºC, for
1998-2001 respectively.
28
Thermologgers (VEMCO Ltd., Shad Bay, NS) were installed at sites 1, 3, 5, 7, 9,
and 10 on the Little River that recorded water temperature each hour. Most began
recording on July 27, 2000, with the exceptions of Site 7, Site 10, and one placed about
2km below Site 5, which began recording in December 1999. Thermologgers were
installed at the sites on the Little Forks Branch stream in the fall of 2000. For the
purposes of statistical comparison, temperature profiles were summarized between
July 27 and October 18 for 2000 and 2001 due to the absence of gaps in any dataset
(Table 2.4).
Data analyses
Population distributions
Young-of-the-year (YOY) sculpin hatch in mid to late June at a size of 10mm and
are generally too small to be caught in dipnets until they are closer to 20mm (Gray et al.
2002). By the following spring, the rapid growth of the YOY class causes an overlap
with the 1+ year-class in some years, making the resolution of age classes difficult
(Gray et al. 2002). For this reason, length frequency distributions were constructed for
only the late summer and fall collections, where possible, using size class increments
of 2mm, and calculating the frequency as a percentage of the total sample size in order
to compare sites without distortions due to sample size. Young-of-the-year sculpin
were identified based on the first mode of length frequency distributions. Length
distributions were compared between sites using the two-sample Kolmogorov-Smirnov
(K-S) test. The K-S test is a non-parametric, distribution free test that compares two
datasets by assessing the similarity of the two cumulative distribution functions: H0:
F(X) = F(Y); HA: F(X) ≠ F(Y) (Sokal and Rohlf 1995).
29
Adult fish data
Normality and homoscedasticity were assessed by visual examination of normal
probability plots and residual plots, respectively. All lethal data were log-transformed
before statistical analysis to improve conformation with the assumptions of parametric
analysis. Length and weight data, and number of eggs in a nest were analyzed by
analysis of variance (ANOVA). Condition factor (k) was calculated using k =
weight/length3*10000. Analysis of covariance (ANCOVA) was used to compare the
condition of the fish across sites using carcass weight as the dependent variable and
length as the covariate. For comparisons of organ size (liver and gonad) and fecundity
(# eggs), carcass weight was used as the covariate. Carcass weight was used to avoid
erroneous analysis of parameters related to the weight of the fish caused by potential
differences in organ sizes. ANCOVA is considered relatively robust to small differences
in slopes (Hamilton et al. 1993), so slopes were considered different when p<0.03.
Following significant ANOVA and ANCOVA tests, Tukey’s post-hoc test was used to
determine site differences (p<0.05).
2.4
Results
Population structure – 1999 Little River
In 1999, the young-of-the-year (YOY) mode of the length frequency distribution was
very prominent in the late summer and fall collections, but was markedly reduced in the
agricultural sites (sites 7-9; Figure 2.3). In August, YOY sculpin comprised 42-46% of
the populations sampled in the forested and transition region, and only 17-23% in the
agricultural region (Figure 2.3 grey line and Table 2.5). The median size of YOY
sculpin (Table 2.5) and the median length within each site’s population increased in the
downstream direction (Figure 2.4). Overall, the length distributions at all forested sites
30
and the transition site were significantly different from those at the agricultural sites (all
p<0.001; Figure 2.4 top).
Within each site, the shapes of the length frequency distributions were similar
between August and November 1999 (Figure 2.3). By November 1999, the relative
YOY proportions increased slightly to 48-55% in the forested region, and to 28-35% in
the agricultural region (Table 2.5). The loss of older, larger sculpin over the late
summer presumably results in an inflation of the relative frequency of younger smaller
sculpin (Figure 2.3 black line and Table 2.5). The same relationship did not hold for the
transition site where the relative proportion of YOY sculpin dropped from 46% to 29%
over the three-month period. The size difference gradient was still evident across the
regions, but was not as distinct as August (Figure 2.4 bottom). Within the forested
region itself, the length distributions remained stable over the three-month period with a
shift upwards reflecting growth. There was a slight decrease in the median length likely
as a result of mortality in the oldest age classes. Conversely, it appears that the
median length at the transition site increased in response to a decrease in the younger
age classes (Table 2.5 and Figure 2.4). Using the change in median size as a
surrogate for growth over the three-month period, YOY sculpin grew more at the
agricultural sites, attaining larger sizes nearing the end of the first growing season
(Table 2.5).
Energy storage and expenditure -1999
In November 1999, both adult males and females were larger in the downstream
reaches than upstream forested reaches (Table 2.6). Mean fish length and size
increased from the forested region to the transition site, decreased slightly into the
agricultural region, then increased at the urban site. The condition factor for sculpin
31
ranged between 0.88 and 1.04, and did not follow the same progression as size, with
fluctuating values showing no apparent relation to site location. In November, male
liver size (as a % of carcass weight) ranged between 1.2 and 1.8%, with a significant
increase seen at the urban site relative to some upstream sites (Figure 2.5 top). Male
gonadosomatic indices (GSI) ranged from 1.4 to 2.2%, with the lowest values observed
in the agricultural and urban sites (Figure 2.5 bottom). The only significant difference
was site 7 compared to the farthest upstream forested site. Female LSI values were
higher than the male values, and ranged from 1.9 to 2.9% (Figure 2.6 top). As with the
males, there was a significant increase in female liver size at the urban site. Female
liver and gonad sizes were both significantly smaller at the agricultural sites compared
to female sculpin collected at the forested sites (Figure 2.6).
The 1999 sculpin collections in the Little River showed that the proportions of YOY
sculpin were reduced in the agricultural and urban region relative to the forested region,
and that the site-specific population distributions were very similar over a three-month
period. The size of the fish tended to increase in the downstream direction, from the
forested to the transition region, and from the agricultural to the urban region. Fish
condition was variable and did not appear to reflect any particular differences related to
the site where the fish were collected. Adult male liver size, and female liver and gonad
sizes relative to the size of the fish were all smaller in the agricultural region. Finally,
adult sculpin had larger livers at the urban site relative to some sites in the forested and
agricultural region. After the first field season, the number of sites was collapsed to
enable a shift in the focus to confirming responses over time.
32
Population structure - 2000-2001
Proportions of YOY sculpin were reduced in the agricultural region in 2000, but
were similar in 2001 compared with the forested region (Table 2.7). The growth of YOY
sculpin was similar in 2000 and 2001, despite warmer water temperatures in 2001. The
differences in proportions of YOY in the agricultural region was not likely a longitudinal
(upstream to downstream) phenomenon, as there were no differences observed across
sites in a comparable, forested river system (the Little Forks stream; Figure 2.7). The
proportion of YOY sculpin at the Little Forks Branch sites ranged from 46-55%, with a
small increase in YOY size in the downstream direction (Table 2.8). As in the Little
River, the shapes of the length frequency distributions were also particularly similar
within a site from August to November (Figure 2.7).
Energy storage and expenditure - 2000-2001
In November 2000, fish were selected for a common size range in order to compare
fish from different rivers. There were no significant differences in length or weight for
females (p>0.05). Male sculpin weight increased downstream despite similar mean
lengths, and condition factor was elevated for both males and female sculpin at the
agricultural site (Tables 2.9 and 2.10). In many of the collections, the fish tended to be
larger in the agricultural and urban region, though most differences were not
significantly different. As in 1999, fish condition factor fluctuated among sites, with no
apparent trends related to the region where the fish were collected in 2000 and 2001
(Tables 2.9 and 2.10). There was no specific gradient of sizes in the downstream
direction within the Little Forks Branch comparison stream. There was a decrease in
female size at the lowermost site in the fall of 2000, and increasing sizes of sculpin
33
towards LF3 in the fall of 2001, but followed by a decrease in fish size at the lowermost
site (Tables 2.9 and 2.10).
Male sculpin in the Little River had LSI values ranging between 1.3 and 2.3% for all
fall sampling periods (Figure 2.8 top). There was little seasonal change, with a range of
LSI between 2 and 3% in the spring collections. There were no significant differences
until November 2001 when male sculpin at site 7 had significantly smaller livers relative
to their body sizes than sculpin in the forested region. Male sculpin GSI values ranged
between 2 and 3% in the fall with a reduction to between 1 and 2% in the spring (Figure
2.8 bottom). Male gonads tended to be smaller in the agricultural region in the spring of
2000, but were not significantly different from males in the forested sites in all other
collections. There was no gradient of liver size or gonad size associated with stream
location observed in the male sculpin collected in the Little Forks stream in November
2000 or 2001 (Figure 2.9).
Adult female sculpin showed much more seasonal variation in their somatic indices.
In the Little River, female LSI values ranged from about 3-4.4 % in the fall to 5-6.5% in
the earliest spring collection (Figure 2.10 top). Female liver size then decreased to
between 3 and 4% of their carcass weight as the fish approached spawning in May. In
the spring of 2000, female liver sizes were smaller in the agricultural region than the
forested region until immediately before spawning. There was no difference in liver
size in the fall of 2000, though the two-week lag in sampling between the two sites may
have been sufficient time for the fish at the agricultural site to catch up to the forested
site during this period of growth for female livers. Female GSI values increased from
between 3 and 5% in the fall to 10-12% in the earliest spring collection (Figure 2.10
bottom). Female gonads tended to be smaller in the agricultural region but were only
significantly smaller in the agricultural region in late March 2000 and again in November
2001 (Figure 2.10). Female sculpin collected in the Little Forks stream did not show
34
any differences in liver or gonad sizes associated with location along the stream
continuum (Figure 2.11). Female gonad size increased between LF1 to LF3, and then
decreased at the lowermost site LF4.
Total fecundity of female sculpin was lower in the agricultural region compared with
the forested region in all spring collections except for early March 2000 (Figure 2.12
top). The number of eggs per gram of gonad weight, effectively a surrogate for the size
of the eggs, was lower at the agricultural sites in late March; ANCOVA was not possible
for May 2000 and April 2001 due to a difference in the slopes of the regression lines
(Figure 2.12 middle). The number of eggs divided by the gonad weight was not
significantly different in May 2000 (p= 0.41), with the two forested sites (site 1 and 3)
having significantly larger eggs than site 7 in April 2001 (p<0.001). When the weight of
the fish was accounted for, the numbers of eggs was significantly lower in the
agricultural sites during all spring collections (Figure 2.12 bottom). The number of eggs
per gram carcass weight was 48-59 in the forested region, and 30-39 in the agricultural
region of the Little River. The total fecundity of females in the Little Forks stream was
lower than for females at site 1 in the Little River, but the number of eggs per gram
gonad weight and carcass weight, were within the range seen for female sculpin in the
forested section of the Little River. The number of eggs per nest found in the spring of
2001 was not significantly different at the forested and agricultural site (ANOVA;
p=0.21), though the variability was quite high. The median number of eggs was lower
at the site in the agricultural region compared to the forested site (170 versus 281,
respectively) (Figure 2.13).
35
Water temperature
Within the Little River, the late summer and fall of 2001 was warmer than 2000, with
a difference at most sites of about 100 degree-days. Over the 84d period, this amounts
to an average increase in water temperatures of about 1.2ºC per day. For both the
Little River and the Little Forks Branch, water temperatures increased in the
downstream direction. Within the forested sites, over a comparable distance, the water
temperatures increased about 130 degree-days on the Little River and 80 degree-days
on the Little Forks Branch.
2.5
Discussion
The objective of our research was to determine whether sculpin populations in
agricultural areas near the Little River, NB, showed changes in health or performance
relative to the upstream forested region. We conducted an effects-based assessment
by examining population structures and whole organism characteristics, comparing fish
residing in the agricultural region with fish living in the forested region. Results from
three years of monitoring showed differences in population structure in the late summer
and fall of 1999 and fall of 2000. Despite adult sculpin tending to be larger in the
downstream reaches, we observed reductions in organ size for both male and female
sculpin. Various indicators of reproductive investment were also reduced for female
sculpin in the agricultural region. Comparisons with the Little Forks Branch, a
completely forested stream, suggests that the organ responses observed in the Little
River did not follow an upstream-downstream gradient but were more likely associated
with land-use inputs from adjacent agricultural activities.
Studying the impacts of non-point source pollution such as agriculture is
confounded by many factors. Being a non-point source of pollution, it is difficult to trace
36
or identify the source of pollution back to a single location to quantify the inputs.
Secondly, the input of direct stressors generally occurs via runoff during rainfall events.
Thus, the frequency and magnitude of exposure of direct stressors to biota is
intermittent and unpredictable and the relative contribution of each runoff event to
indirect stressors (e.g. increased sedimentation causing habitat degradation) is also
unpredictable and very difficult to measure. Monitoring over a three-year period
allowed for the study of effects to be conducted over a range of environmental
conditions. The period from 1 May to 31 October can be considered the inclusive
cropping period (Chow et al. 2000), from when potatoes are planted to the latest
harvest. The 1999 cropping period was the wettest during the study, but was
comparable to 2000 in terms of the number of substantial rainfall events (one day
rainfall >20mm). Relative to 1999 and 2000, total rainfall in the same period on 2001
was reduced by between 10 and 20%, respectively, and there were approximately half
the number of substantial rainfall events. Air temperatures and water temperatures
indicated that 2000 was the coolest year during the study. With a difference of about
100 degree days within each site during the same period of 2000 and 2001, water
temperatures were about 1ºC cooler in 2000. This difference in water temperature may
account for the greater median sizes attained by YOY sculpin in 2001 compared to
2000. The within river difference in the Little River between the uppermost and
lowermost sites, approximately 120-130 degree days, may also be largely responsible
for the difference in the size of YOY and older sculpin observed between regions.
Sculpin collected in the Little Forks forested stream also showed increased size in the
downstream direction, though not always statistically significant, associated with an
increase in degree-days.
The proportion of young-of-the-year (YOY) sculpin in the late summer can be used
as a rough measure of reproductive success, representing the combined outcome of
37
reproductive output, egg viability, and early-life stage survival within a particular sample
population. In the forested sites on the Little River and the Little Forks Branch, which
represent the reference situation for sculpin, the proportion of YOY in sample
populations was generally 40% or more. In the late summer of 1999 and 2000, the
proportions of YOY sculpin in sample populations in the agricultural region were
reduced to almost half that of the reference situation. Sampling again in November
confirmed that the responses persisted over time within particular sites. In contrast, the
proportions of YOY in 2001 at all sites within the Little River and the Little Forks Branch
were similar and quite stable suggesting no influence of location along the stream
continuum or adjacent land-use. The relatively drier conditions in 2001 could be a
factor in the improved survival of YOY sculpin in the agricultural region of the Little
River, as the influence of adjacent agricultural land-use would be reduced with possibly
lower runoff amounts. Fitzgerald et al. (1999) looked at creek chub populations in
sections of a stream receiving agricultural and urban inputs and also found that the 0+
age class was limited or absent from many of the impacted sites. The authors
suggested that habitat degradation might have been a factor because of the
reproductive strategy of the creek chub. Creek chub have a similar lithophilous
reproductive behaviour as the sculpin that requires they have a clean, stony substrate
for egg laying and development (Balon 1975; Scott and Crossman 1998). Male sculpin
prepare and maintain nests beneath rocks where females lay their eggs, and the males
defend and guard the nests for over a month until the sculpin larvae have dispersed
(Van Vliet 1964). Thus, degradation of suitable spawning and rearing habitats may
have significant impacts on egg or early life-stage survival.
Other conditions or stressors that may influence the performance of the YOY ageclass could include reduced egg production or egg viability, as well as increased
predation or water temperature, or reductions in food availability (Fitzgerald et al.
38
1999). Predation by other fish is not likely a significant factor in the Little River system,
as brook trout densities were lowest in the agricultural region. Food availability was not
believed to be limiting in this region, although there were some indications of a
decrease in benthic invertebrate community diversity and key clean water taxa (RA
Curry, UNB, pers. comm.). Increased temperatures can be a factor influencing egg
viability and early survival (Rand et al. 1995). Relative to the forested region, summer
water temperatures were consistently 2-3ºC warmer in the agricultural region with mean
daily temperatures reaching 18ºC during the summer of 2000, and 20ºC during the
summer of 2001. Upper lethal limits for older slimy sculpin have been reported around
23-25ºC (Symons et al. 1976; Otto & Rice 1977). Van Vliet (1964) found that sculpin
fry were able to sustain higher temperatures than adult sculpin. In addition, because
the YOY proportions were comparable in both forested and agricultural regions in the
warmer summer (2001), it is unlikely that temperature was responsible for reducing the
numbers of YOY sculpin. It is uncertain whether egg viability was reduced in the
agricultural region as this was not assessed in our study.
The relative increase in size and growth of YOY sculpin in the downstream direction
may be related to the warmer water temperatures. An increase in median size of about
5-6mm was generally seen for YOY at the uppermost forested and the lowermost
agricultural site on the Little River. If the distance between sites is accounted for, the
increase in size of 3mm from uppermost to lowermost site on the Little Forks Branch is
probably similar to the YOY growth in the Little River. Downstream water temperatures
were also warmer in the Little Forks Branch with a consistent 1-2ºC difference, though
overall the Little Forks Branch was a cooler river with mean daily temperatures
exceeding 16ºC only once at LF4 during the summer of 2001. Based on the general
ecology of running waters it is accepted that streams and rivers, without significant
groundwater inputs, will tend to be warmer in downstream sections because they are
39
wider and more open, thus receiving less shading from streamside vegetation (Vannote
and Sweeney 1980; Allan 1995).
Condition factor and liversomatic index are both generally accepted as indications
of nutritional status and energy storage (Munkittrick et al. 2000; Barton et al. 2002),
though increased liver size has been found in response to exposure to contaminants in
pulp mill effluent (Munkittrick et al. 2000). We found no consistent differences in the
condition of mature sculpin in the different land-use regions of the LittleRiver, nor
among the sites of the Little Forks stream. In terms of liver size, there were conflicting
responses observed in the agricultural region. There were no differences in liver size
for male sculpin, whereas female sculpin in the lower reaches consistently showed
reductions until the fall of 2001. This reduction, however, was not associated with
decreased condition; in fact in the fall of 2000, it was associated with increased
condition. It is uncertain why there were contradictory responses with the two
parameters, and why there may be a sex-specific response. Spoonhead sculpin
(Cottus ricei) and slimy sculpin collected downstream of pulp mill effluents displayed
increased condition and liver size and with no differences between sexes (Gibbons et
al. 1998; Galloway et al. 2003).
Over the three years of monitoring mature male and female sculpin both had
smaller gonads in the agricultural region in 1999 and 2000, with no consistent
differences observed in sculpin gonad size in the spring or fall of 2001. Reduced
fecundity was observed in female sculpin from the agricultural region in both 2000 and
2001 spring collections. There was one exception in early March 2000, where
fecundity was higher at the agricultural site, however the relationship was opposite
when standardized for body size. For a given body weight, female sculpin consistently
exhibited reduced fecundity in the agricultural region for all spring collections. There
was also evidence that the size of their eggs was reduced in 2000 and 2001. Egg size
40
often has a covariant relationship with fecundity, and can also be influenced by
environmental conditions (Greeley 2002). Reduced fecundity may be connected to the
observation of fewer eggs in nests in the agricultural region, and smaller egg sizes may
have implications for egg viability or early-life stage survival (e.g. McKim 1977; Greeley
2002). Higher temperatures generally lead to an increase in fecundity and a decrease
in egg size. The fact that we see decreased fecundity and egg size in the agricultural
region may indicate that temperature is not a factor in the response.
Fecundity has been established as a sensitive indicator of exposure to xenobiotics,
and has been used extensively as a reproductive metric due to its high ecological value
and proven relationship to anthropogenic stress (Greeley 2002). However, Greeley
(2002) cautions that because indicators of reproductive performance such as fecundity
and egg size can be influenced by environmental conditions (e.g. temperature, salinity,
and nutrition), they should be accounted for within a suite of reproductive indicators as
possible covariates. In this study we were able to demonstrate smaller gonad size,
reduced fecundity, smaller egg size, possibly fewer eggs per nest, and reduced
proportions of YOY sculpin at sites in the agricultural region. All of these parameters
are either directly or indirectly interrelated and point to some form of reproductive
impairment. In their study of creek chub in streams influenced by urban and
agricultural inputs, Fitzgerald et al. (1999) found no change in GSI, but did detect a
reduction in fecundity when fish length was accounted for.
Gibbons and Munkittrick (1994) suggested generalized stress response patterns for
white sucker subjected to natural and anthropogenic environmental impacts. The
framework uses various biological criteria including age structure, condition, and
fecundity to determine whether the fish population is experiencing one of eight
proposed response patterns: no response, exploitation, recruitment failure, chronic
recruitment failure, food limitation, metabolic redistribution, niche shift, and multiple
41
stressors (Gibbons and Munkittrick 1994). Using their framework, the general response
of sculpin in the agricultural region of the Little River fits many of the biological criteria
for metabolic redistribution, though the shift of the population due to the reduced YOY
proportions also points to exposure to multiple stressors. The conflicting responses
seen with increased size but decreased liver size and gonad size could be a result of a
metabolic disruption problem with differential energy utilization and allocation to
somatic and reproductive tissue growth.
One major limitation with using a small-bodied fish for environmental monitoring is
that information on the basic life-history characteristics is generally lacking (Munkittrick
et al. 2000). Interpretation of biological responses requires an awareness of the natural
variability within the species both temporally and spatially. Sampling sculpin at multiple
reference sites in forested regions and at different times of the year provided essential
baseline information about the changes in sculpin population structure and wholeorganism parameters. Sampling in the fall and spring over three years provided
information on the progression of organ development and energy storage as the sculpin
enter into gonadal recrudescence and prepare for spawning. In the forested region of
the Little River, male liver size increased about 50% between the fall and spring
collections, while gonad size remained relatively stable with a small decrease from fall
to spring as the testis undergoes reorganization in preparation for spawning. Because
female fish require more energy during gonadal recrudescence to account for the
greater reproductive output, there is a larger change in female sculpin liver and gonad
size between the fall and spring. At the reference sites, female liver size more than
doubled from the fall to the early spring, then steadily declined as spawning time
approached. This almost certainly coincides with vitellogenesis in female sculpin when
the protein vitellogenin, produced in the liver, is transferred and incorporated into the
developing oocytes. Concomitantly, female GSI increased about 400% from the fall to
42
the early spring and then effectively tripled again immediately before spawning in May.
During the same period, the number of eggs in the gonad drops by about 50%, as the
eggs increase in volume. Although there was a drop in the condition of females in the
spring of 2000 relative to the fall, which would not be surprising during such a period of
high energy requirements, there was not a similar decrease in condition observed in the
spring of 2001.
2.5.1
Conclusions
Using an effect-based approach we monitored sculpin populations in the Little River
over a three-year period allowing us to observe and confirm potential effects from
exposure to non-point source pollution from agricultural activities both spatially and
temporally. Sculpin populations in the agricultural region of the Little River had fewer
YOY sculpin in the fall of 1999 and 2000 than the forested regions. Also in the
agricultural region, mature male sculpin had smaller gonads, and female sculpin had
smaller livers and gonads relative to sculpin of the same body size in the reference
forested region. These effects were determined to be unrelated to stream continuum
by observing sculpin along a reference forested river, the Little Forks Branch. By the
fall of 2001, sculpin population structures appeared stable and there were fewer
significant differences in organ sizes in adult sculpin in the agricultural region. The
cropping season of 2001 experienced between 10 and 20% less rainfall than the
preceding two years, and about half of the major rainfall events, which may have
played a role in reducing the potential effects of the agricultural stressors. Although
natural stressors such as temperature could influence many of the variables that we
measured, an increase in temperature is generally associated with a concomitant
increase in somatic and reproductive tissue growth. In the agricultural region we
43
observed some evidence of warmer temperatures, increased size, but associated with
a consistent observation of decreased gonad size for both males and females. Further
research is needed to confirm the responses observed in the Little River in other
systems receiving agricultural inputs, and follow-up studies need to be designed to
attempt to tease out the relative influence of temperature, sedimentation, nutrient, and
pesticides on the responses observed in the local fish populations.
2.6
Acknowledgements
Funding for this study was provided primarily by the Toxic Substances Research
Initiative (TSRI), with additional support from Crop Life Canada. MAG was funded by a
post-graduate scholarship from the Natural Sciences and Engineering Research
Council (NSERC), and KRM is supported as a Canadian Research Chair. The field
and laboratory work was conducted with the invaluable help of technicians and
graduate students from the New Brunswick Cooperative Fish and Wildlife Research
Unit and the Canadian Rivers Institute. The authors would also like to acknowledge the
assistance given by Agriculture and Agri-Food Canada and the Eastern Canadian Soil
and Water Conservation Centre.
44
Table 2.1. Little River site characteristics (mean ± SE) measured in August 2000. The
number of measurements at each site is indicated in brackets.
Site
Width (m)
Depth (m)
Velocity (cm.s-1)
Discharge1
Cobble size
(m3.s-1)
(cm)
1
7.9 ± 0.4 (3)
0.30 ± 0.03 (10)
6.5 ± 0.7 (9)
0.15
8.5 ± 0.8 (30)
3
10.8 ± 1.3 (3)
0.23 ± 0.02 (15)
9.0 ± 1.2 (13)
0.22
16.0 ± 2.1 (45)
6
21.8 ± 0.3 (3)
0.25 ± 0.03 (15)
11.9 ± 1.3 (15)
0.64
17.3 ± 0.9 (45)
7
22.9 ± 0.6 (3)
0.24 ± 0.01 (15)
13.3 ± 0.7 (15)
0.72
10.3 ± 0.4 (45)
9
16.4 ± 0.4 (3)
0.44 ± 0.04 (15)
13.3 ± 1.5 (15)
0.95
12.9 ± 0.7 (45)
10
25.0 ± 0.3 (3)
0.25 ± 0.01 (15)
16.3 ± 1.0 (15)
1.01
16.0 ± 0.7 (45)
1 - Discharge was calculated by multiplying mean width, depth, and velocity.
45
Table 2.2. Sculpin sampling dates for the Little River (sites 1-10) and the Little Forks Branch stream (LF1-4) from 1999 to
2001. In general, each collection consisted of 100 sculpin measured and weighed with lethal sampling of mature fish
occurring in the fall and spring. Adjacent land-use was classified as forested (F), agricultural (A), transitional between
forested and agricultural (T), and urban (U).
Site
Land-use
1999
Aug
1
2
3
4
5
6
7
8
9
10
LF1
LF2
LF3
LF4
F
F
F
F
F
T
A
A
A
A
F
F
F
F
•
•
•
•
•
•
•
2000
2001
Nov
Mar
May
Aug
Nov
Apr
•
•
•
•
•
•
•
•
•
•
•
1
•
•
•
•
•
2
3
•
•2
•1
•3
•
•
•
•
•
•5
1 - fish were collected in both early March (8 March) and late March (26-27 March)
2 - fish only collected in late March
3 - only mature female sculpin were collected and sampled
4 - no adults were collected
5 - no other Little Forks Branch sites were accessible at this time
46
Nov
•
•
•
•
4
•
•
•
Aug
•
•
•
•
•
•
•
•
•
•
•
•
Table 2.3. Precipitation and air temperature for the study region was summarized from data collected at the St. Leonard
airport (Environment Canada 2001b), located approximately 13km northwest from the mouth of the Little River, and 48km
southeast from the mouth of the Little Forks Branch. The non-cropping period is from 1 November through 30 April, and the
cropping period is from 1 May through 31 October.
Total rainfall (mm)
Month
Mean Air T (ºC)
1998
1999
2000
2001
1998
1999
2000
2001
January
1.0
35.4
8.4
0
-6.69
-12.62
-12.27
-12.17
February
52.6
4.2
9.0
9.6
-3.37
-8.75
-10.11
-11.49
March
89.6
21.0
13.2
0
-3.09
-1.38
-1.46
-5.17
April
32.6
17.4
105.8
13.4
3.77
3.22
2.04
2.38
May
83.6
43.2
111.8
63.6
12.99
13.48
8.82
12.70
June
107.6
89.2
140.4
104.4
15.69
17.43
13.89
15.78
July
178.0
109.4
123.8
121.6
18.54
18.58
16.77
17.82
August
88.4
114.2
88.2
53.4
16.86
16.45
16.35
18.12
September
97.8
180.8
43.2
106.2
12.40
16.16
11.16
13.40
October
52.8
75.8
43.2
46.4
5.77
4.10
5.08
3.29
November
33.0
71.6
41.4
34.0
-1.70
0.87
1.23
1.28
December
17.6
46.6
34.6
18.4
-7.34
-6.25
-9.96
-3.83
Non-cropping
227.4
196.2
212.4
75.4
Cropping
625.6
612.6
550.6
495.6
Total
47
Table 2.4. Degree-days (sum of mean daily temperature) for water temperatures
between 27 July and 18 October for the Little River and Little Forks Branch. Site
numbers increase in the downstream direction.
Little River
2000
2001
Little Forks
2001
Site 1
941.5
1045.4
LF1
928.4
3
903.6
1004.4
LF2
980.0
5
930.3
LF3
978.8
6
977.2
LF4
1003.1
7
1014.7
1136.3
9
1024.8
1124.9
10
1064.3
1197.5
48
Table 2.5. Proportions and median sizes of young-of-the-year (YOY) sculpin of sample
populations collected in the Little River, August and November 1999. The change in
the median size of YOY sculpin at each site was used as a surrogate for growth.
August 1999
Site
November 1999
Proportion of
Median size
Proportion of
Median size
Growth
population
(mm)
population
(mm)
(mm)
0.47
34
0.54
32
0.54
28
1
2
0.46
26
3
6
4
0.43
26
0.48
30
4
5
0.42
28
0.55
32
4
6
0.46
28
0.29
32
4
7
0.23
29
0.30
37
8
8
0.17
30
0.35
37
7
9
0.21
31
0.28
38
7
0.16
1
10
46
1– gap in lengths up to 39mm may cause an overestimation of this value
49
Table 2.6. Mean (± SE) length, carcass weight, and condition factor for mature male
and female slimy sculpin collected in the Little River in November 1999. Parameter
sample sizes different from the overall sample size (N) are indicated in brackets.
Different letters indicate significant differences between sites for each variable
(ANOVA; p<0.05). Condition was assessed by ANCOVA of carcass weight, with length
as the covariate.
Carcass weight (mm) 1
Condition factor (k)
A
2.29 ± 0.12
A
0.98 ± 0.02
AC
65.55 ± 1.65
AB
2.72 ± 0.24
A
0.94 ± 0.02
ABC
6
66.00 ± 2.25
AB
2.85 ± 0.32
A
0.97 ± 0.03
ABC
Site 4
9
67.89 ± 2.63
AB
2.92 ± 0.32
A
0.90 ± 0.02
AB
Site 5
10
68.80 ± 2.67
AB
3.39 ± 0.37
A
1.00 ± 0.02
AC
Site 6
14
70.36 ± 2.09
BC
3.35 ± 0.33(13)
A
0.91 ± 0.02(13)
AB
Site 7
12
66.42 ± 1.99
AC
2.64 ± 0.24
A
0.88 ± 0.03
B
Site 8
12
70.83 ± 2.42
BC
3.44 ± 0.39
A
0.93 ± 0.02
ABC
Site 9
7
69.57 ± 3.63
AC
3.39 ± 0.59
A
0.94 ± 0.03
ABC
Site 10
20
79.8 ± 1.30
D
5.43 ± 0.36
B
1.04 ± 0.02
C
Site 1
12
56.00 ± 1.15
A
1.62 ± 0.08
A
0.92 ± 0.02
AB
Site 2
12
58.17 ± 1.31
A
1.75 ± 0.13
A
0.87 ± 0.02
A
Site 3
19
57.74 ± 1.36
A
1.80 ± 0.14
A
0.90 ± 0.02
AB
Site 4
13
56.38 ± 1.47
A
1.63 ± 0.13(12)
A
0.91 ± 0.02(12)
AB
Site 5
14
56.14 ± 1.47
A
1.63 ± 0.16
A
0.89 ± 0.02
AB
Site 6
16
68.00 ± 1.96
B
2.85 ± 0.27
B
0.87 ± 0.01
A
Site 7
15
60.20 ± 1.88
AC
2.05 ± 0.24
AC
0.89 ± 0.01
AB
Site 8
14
62.14 ± 0.89
AB
2.15 ± 0.12
AB
0.89 ± 0.02
AB
Site 9
22
65.68 ± 1.18
BC
2.49 ± 0.17(21)
BC
0.86 ± 0.02(21)
A
Site 10
10
77.00 ± 1.39
D
4.53 ± 0.23
D
0.99 ± 0.03
B
N
Length (mm)
Site 1
15
61.33 ± 0.89
Site 2
11
Site 3
Males
Females
50
Table 2.7. Proportions and median sizes of young-of-the-year (YOY) sculpin of sample populations collected in the Little
River, August and November 2000 and 2001. The change in the median size of YOY sculpin at each site was used a
surrogate for growth.
August 2000
Site
1
November 2000
August 2001
November 2001
Proportion of
Median
Proportion of
Median
Growth
Proportion of
Median
Proportion of
Median
Growth
population
size (mm)
population
size (mm)
(mm)
population
size (mm)
population
size (mm)
(mm)
0.71
25
0.66
36
11.0
0.57
27
0.56
35.5
8.5
0.62
32
0.57
44
12.0
0.57
36
0.71
47
11.0
2
6
0.23
24
7
0.26
28
9
0.51
33
10
0.86
31
0.48
0.41
341
42.5
6.01
11.5
1 – gap in lengths from 31-41mm may cause an underestimation of these values.
51
Table 2.8. Proportions and median sizes of young-of the year (YOY) sculpin of sample
populations collected along the Little Forks Branch stream in August and November
2001. The change in the median size of YOY sculpin at each site was used a surrogate
for growth.
August 2001
Site
November 2001
Proportion of
Median size
Proportion of
Median
Growth
population
(mm)
population
size (mm)
(mm)
LF1
0.50
21
0.55
29
8
LF2
0.46
22
0.41
31
9
LF4
0.52
24
0.62
32
8
52
Table 2.9. Mean (± SE) length, carcass weight, and condition factor for mature male
slimy sculpin collected in the Little River and Little Forks stream in 2000 and 2001.
Within each river and collection period, different letters indicate significant differences
between sites for each variable (p<0.05). Condition was assessed by ANCOVA of
weight, with length as the covariate.
Little River sites
Early March 2000
Site 1
Site 9
Late March 2000
Site 1
Site 3
Site 7
Site 9
November 2000
Site 1
Site 7
April 2001
Site 1
Site 3
1
Site 7
Site 9
November 2001
Site 2
Site 3
Site 7
Site 9
Site 10
Little Forks sites
November 2000
LF2
LF3
LF4
April 2001
2
LF2
November 2001
LF1
LF2
LF3
LF4
N
Length (mm)
Carcass weight (g)
Condition factor (k)
10
3
63.2 ± 1.9
78.7 ± 5.2
A
B
2.45 ± 0.29
4.85 ± 1.12
A
B
0.93 ± 0.04
0.96 ± 0.03
A
A
11
9
6
4
70.6 ± 2.4
79.3 ± 3.1
68.8 ± 3.9
79.8 ± 5.2
A
A
A
A
3.64 ± 0.37
5.22 ± 0.76
3.22 ± 0.52
5.51 ± 1.15
A
A
A
A
1.01 ± 0.02
0.99 ± 0.04
0.95 ± 0.04
1.04 ± 0.03
A
A
A
A
8
5
65.1 ± 1.2
68.0 ± 0.5
A
A
2.90 ± 0.19
3.74 ± 0.13
A
B
1.04 ± 0.03
1.19 ± 0.03
A
B
7
6
1
7
66.7 ± 2.7
69.2 ± 1.9
80
75.1 ± 4.6
A
A
A
A
A
1.17 ± 0.05
1.12 ± 0.02
1.15
1.08 ± 0.05
A
AB
A
3.63 ± 0.48
3.74 ± 0.31
5.90
5.11 ± 1.26
9
8
18
8
14
70.2 ± 2.6
75.3 ± 1.9
75.8 ± 1.2
73.0 ± 2.5
79.6 ± 0.9
A
AB
AB
AB
B
3.24 ± 0.37
4.19 ± 0.39
4.16 ± 0.19
4.10 ± 0.41
4.96 ± 0.20
A
AB
AB
AB
B
0.90 ± 0.02
0.96 ± 0.03
0.94 ± 0.01
1.03 ± 0.02
0.97 ± 0.02
A
AB
A
BC
AC
N
Length (mm)
4
5
5
66.0 ± 1.7
70.2 ± 1.6
70.0 ± 1.5
6
73.0 ± 3.5
7
9
20
13
64.4 ± 2.7
68.3 ± 3.6
81.9 ± 1.2
73.2 ± 2.3
Carcass weight (g)
A
A
A
3.02 ± 0.27
3.48 ± 0.20
3.42 ± 0.26
A
A
A
4.26 ± 0.60
A
A
B
A
2.49 ± 0.31
3.28 ± 0.50
5.14 ± 0.26
3.54 ± 0.34
Condition factor (k)
1.04 ± 0.03
1.01 ± 0.05
0.99 ± 0.02
A
A
A
1.06 ± 0.02
A
A
B
A
0.91 ± 0.03
0.96 ± 0.02
0.92 ± 0.01
0.87 ± 0.02
1- only one mature male was collected
2- this was the only Little Forks Branch site accessible due to snow depth
53
B
AB
A
AB
B
Table 2.10. Mean (± SE) length, carcass weight, and condition factor for mature female
slimy sculpin collected in the Little River and Little Forks stream in 2000 and 2001.
Parameter sample size differences are indicated in brackets. Within each river and
collection period, different letters indicate significant differences between sites for each
variable (p<0.05). Condition was assessed by ANCOVA of weight, with length as the
covariate.
Little River sites
Early March 2000
Site 1
Site 9
Late March 2000
Site 1
Site 3
Site 7
Site 9
May 2000
Site 1
Site 7
November 2000
Site 1
Site 7
April 2001
Site 1
Site 3
Site 7
Site 9
November 2001
Site 2
Site 3
Site 7
Site 9
Site 10
N
Length (mm)
Carcass weight (g)
Condition factor (k)
16
16
54.2 ± 1.4
66.8 ± 1.9
A
B
1.30 ± 0.11
2.61 ± 0.20
A
B
0.79 ± 0.02
0.85 ± 0.02
A
B
10
14
20
18
60.5 ± 1.2
67.0 ± 2.5
60.0 ± 1.3
65.6 ± 1.3
AB
A
B
A
1.99 ± 0.12
2.75 ± 0.38
1.89 ± 0.14
2.65 ± 0.18(17)
AB
A
B
A
0.89 ± 0.02
0.85 ± 0.02
0.86 ± 0.01
0.91 ± 0.02(17)
A
A
A
A
8
25
57.9 ± 1.9
57.0 ± 1.4
A
A
1.56 ± 0.17
1.61 ± 0.13
A
A
0.78 ± 0.03
0.84 ± 0.02
A
A
11
10
63.6 ± 1.1
64.1 ± 1.2
A
A
2.57 ± 0.15
2.98 ± 0.22
A
A
0.99 ± 0.02
1.11 ± 0.03
A
B
9
6
12
3
63.1 ± 2.7
61.7 ± 2.3
62.9 ± 1.0
67.0 ± 2.7
A
A
A
A
2.76 ± 0.37
2.40 ± 0.31
2.36 ± 0.10
2.69 ± 0.23
A
A
A
A
1.05 ± 0.02
1.00 ± 0.03
0.94 ± 0.01
0.89 ± 0.03
A
AB
B
B
16
18
9
18
15
66.4 ± 2.0
66.2 ± 2.1
70.4 ± 1.4
65.9 ± 1.5
71.3 ± 1.5
A
A
A
A
A
2.66 ± 0.23
2.92 ± 0.34
3.22 ± 0.15
2.85 ± 0.24
3.55 ± 0.24
A
A
A
A
A
0.87 ± 0.02
0.95 ± 0.02
0.92 ± 0.02
0.97 ± 0.02
0.96 ± 0.01
A
B
AB
B
B
Little Forks sites
N
Length (mm)
Carcass weight (g)
Condition factor (k)
November 2000
LF2
10
67.3 ± 1.4
A
3.11 ± 0.22
A
1.01 ± 0.02
A
LF3
10
67.2 ± 1.0
A
3.11 ± 0.14
A
1.02 ± 0.02
A
LF4
10
62.9 ± 1.1
B
2.59 ± 0.14
A
1.03 ± 0.02
A
April 2001
LF21
12
62.4 ± 2.3
2.24 ± 0.27
0.88 ± 0.01
November 2001
LF1
18
54.2 ± 1.3
A
1.38 ± 0.12
A
0.84 ± 0.01
A
LF2
17
62.7 ± 1.8
B
2.25 ± 0.18
B
0.89 ± 0.02
A
LF3
11
74.6 ± 1.5
C
3.78 ± 0.27
C
0.90 ± 0.02
A
LF4
11
59.4 ± 2.5
AB
1.94 ± 0.25
AB
0.88 ± 0.03
A
1- this was the only Little Forks Branch site accessible due to snow depth
54
Little Forks
St. Leonard airport
Little River
Grand Falls
Flow
NEW BRUNSWICK
Little River
St. John River
Fredericton
Saint John
1
2
3
4
5
6
7
8
0km
9
5km
10
St. John River
Figure 2.1. Map of the Little River, New Brunswick indicating site numbers and
locations where sculpin were collected from 1999-2001.
55
Little Forks
Little Forks
St. Leonard airport
Little River
Grand Falls
NEW BRUNSWICK
Flow
St. John River
Fredericton
Saint John
LF1
LF2
LF3
0km
5km
LF4
Green River
Figure 2.2. Map of the forested Little Forks Branch of the Green River indicating site
numbers and locations where sculpin were collected from 2000-2001.
56
Figure 2.3. Length frequency distributions for sculpin collected in the Little River in
August (grey line) and November (black line) of 1999. Frequencies are given as % of
sample for comparison purposes (minimum N=96).
57
20
20
15
10
10
5
5
0
20
0
20
15
Site 6
15
Site 2
10
10
5
5
0
20
0
20
15
Site 7
15
Site 3
Site 8
10
10
5
5
0
20
0
20
15
15
Site 4
Site 9
0
20
0
20
15
15
Site 5
58
70
62
54
46
38
30
22
More
94
86
78
70
62
54
46
0
38
0
30
5
22
5
Length (mm)
Site 10
10
14
10
Length (mm)
More
5
94
5
86
10
78
10
14
% of N
15
Site 1
100
AUGUST
A
A
A
A
B
B
B
2
4
5
6
7
8
9
Length (mm)
80
60
40
20
0
100
NOVEMBER
A
A
A
B
CD
BC
D
2
4
5
6
7
8
9
Length (mm)
80
60
40
20
0
Site
Figure 2.4. Length distributions of sculpin at sites with both an August (top) and
November (bottom) 1999 collection. The ends of the box represent the 25th and 75th
percentile length, whiskers represent the 5th and 95th percentile lengths, and the
horizontal black line represents the median length. Difference letters above boxes
indicate significant differences in length distributions within each month (KolmogorovSmirnov tests; p<0.05).
59
2
AB AB
AB
AB
A
A
AB
AB
AB
B
7
8
9
10
AB
AB
9
10
1.8
1.6
LSI (%)
1.4
1.2
1
0.8
0.6
0.4
0.2
0
1
2
3
4
5
6
A
AB
AB
AB
AB
AB
B
AB
1
2
3
4
5
6
7
8
3
GSI (%)
2.5
2
1.5
1
0.5
0
Site
Figure 2.5. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature male sculpin collected in the Little River in November 1999. Sites 15 are located in the forested region (black bars), site 6 in the transitional region (dark
grey bar), and sites 7-9 in the agricultural region (light grey bars), and site 10 in the
urban region (white bar). Different letters represent statistical differences from
ANCOVA on organ weight, with carcass weight as the covariate (p<0.05).
60
3.5
A
AB
1
2
A
A
1
2
ABC ABCD ABCD AB
BCD CD
D
A
9
10
3
LSI (%)
2.5
2
1.5
1
0.5
0
3
3
ABC
4
5
6
7
8
AC ABC A
BD
D
7
8
D
CD
2.5
GSI (%)
2
1.5
1
0.5
0
3
4
5
6
9
10
Site
Figure 2.6. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature female sculpin collected in the Little River in November 1999. Sites
1-5 are located in the forested region (black bars), site 6 in the transitional region (dark
grey bar), and sites 7-9 in the agricultural region (light grey bars), and site 10 in the
urban region (white bar). Different letters represent statistical differences from
ANCOVA on organ weight, with carcass weight as the covariate (p<0.05).
61
20
LF1
15
10
5
0
20
% of N
15
LF2
10
5
0
20
15
LF4
10
5
94
86
More
Length (mm)
78
70
62
54
46
38
30
22
14
0
Figure 2.7. Length frequency distributions for sculpin collected in the Little Forks
stream in August (grey line) and November (black line) of 2001. Frequencies are given
as % of sample for comparison purposes.
62
4
3.5
Early
March
2000
A A
Late March
2000
November
2000
A A A A
April 2001
A
A A
3
LSI (%)
November
2001
AB AB C A BC
2.5
A A
2
1.5
1
0.5
0
1 7
1 3 7 9
3
1 7
1 3 9
A A
*
GSI (%)
2.5
2
A A
A A B AB
1 7
1 3 7 9
2 3 7 9 10
A A A
1.5
1
0.5
0
1 7
1 3 9
2 3 7 9 10
Site
Figure 2.8. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature male sculpin collected in the Little River from the spring and fall of
2000 and 2001. Sites 1-3 are located in the forested region (black bars), sites 7 and 9
in the agricultural region (light grey bars), and site 10 in the urban region (white bar).
Within each collection period, different letters represent statistical differences from
ANCOVA on organ weights, with carcass weight as the covariate (p<0.05). Significant
interactions within ANCOVA are indicated with an asterisk (*).
63
Nov 2000
1.8
Nov 2001
AB
B
A
AB
LF3
LF4
1.6
LSI (%)
1.4
A
A
A
LF2
LF3
LF4
LF1
LF2
A
A
A
A
A
LF2
LF3
LF4
LF1
LF2
1.2
1
0.8
0.6
0.4
0.2
0
3
A
A
GSI (%)
2.5
2
1.5
1
0.5
0
LF3
LF4
Site
Figure 2.9. Mean (± SE), liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature male sculpin collected in the Little Forks stream in November 2000
and 2001. Sites are numbered 1-4 from upstream to downstream. Within each
collection period, different letters represent statistical differences from ANCOVA on
organ weights, with carcass weight as the covariate (p<0.05).
64
Early
March
2000
Late March
2000
May
2000
Nov
2000
April 2001
November
2001
8
7 A B
A B B AB
A A B AB
LSI (%)
6
A A
5
A B
A A A A A
4
3
2
1
0
1 7
1 3 7 9
1 9
1 7
1 3 7 9
2 3 7 9 10
A B
35.9 ± 2.8
30
A A A A
GSI (%)
25
A A B B
20
15
*
10
A A
AB A AB BC C
5
0
1 7
1 3 7 9
1 9
1 7
Site
1 3 7 9
2 3 7 9 10
Figure 2.10. Mean (± SE) liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature female sculpin collected in the Little River from the spring and fall of
2000 and 2001. Sites 1-3 are located in the forested region (black bars), sites 7 and 9
in the agricultural region (light grey bars), and site 10 in the urban region (white bar).
Within each collection period, different letters represent statistical differences from
ANCOVA on organ weights, with carcass weight as the covariate (p<0.05). Significant
interactions within ANCOVA are indicated with an asterisk (*).
65
Nov 2000
3.5
Nov 2001
A
A
A
LF2
LF3
LF4
A
A
A
LF2
LF3
LF4
A
A
A
A
LF1
LF2
LF3
LF4
B
A
LSI (%)
3
2.5
2
1.5
1
0.5
0
5
A
AB
GSI (%)
4
3
2
1
0
LF1
LF2
LF3
LF4
Site
Figure 2.11. Mean (± SE), liversomatic (LSI; top), and gonadosomatic index (GSI;
bottom) for mature female sculpin collected in the Little Forks stream in November
2000 and 2001. Sites are numbered 1-4 from upstream to downstream. Within each
collection period, different letters represent statistical differences from ANCOVA on
organ weights, with carcass weight as the covariate (p<0.05).
66
Figure 2.12. Mean (± SE) fecundity (total number of eggs) (top), number of eggs per
gram of gonad weight (middle), and number of eggs per gram carcass weight (bottom)
for female sculpin collected in the Little River and the Little Forks stream in the spring of
2000 and 2001. Sites 1, 3, and LF2 are located in forested regions (black bars), sites 7
and 9 in the agricultural region (light grey bars). Different letters represent statistical
differences within each collection period from ANOVA on total fecundity, and ANCOVA
on fecundity with gonad weight and carcass weight, respectively, as the covariates
(p<0.05). Significant interactions within ANCOVA are indicated with an asterisk (*).
67
200
Early
March
2000
Late March
2000
May
2000
AB A B B
April 2001
A AB B AB
Fecundity
150
A B
A B
100
50
Eggs/g gonad weight
0
500
A A
400
AC A B C
300
*
200
*
100
0
Eggs/g carcass weight
60
A B
A A B B
1 7
1 3 7 9
A B
A AB C BC
50
40
30
20
10
0
68
1 7
Site
1 3 7 9
LF2
700
600
# eggs per nest
500
400
300
200
100
0
Site 1
Site 7
Figure 2.13. Number of eggs per nest found at site 1 (forested; 10 nests) and site 7
(agricultural; 11 nests) in the Little River in May 2001. The ends of the box represent
the 25th and 75th percentile values, whiskers represent the 5th and 95th percentile
values, and the horizontal black line represents the median value.
69
2.7
References
Allan JD. 1995. Stream ecology: Structure and function of running waters. Chapman &
Hall, London, UK.
Balon EK. 1975. Reproductive guilds of fishes: A proposal and definition. J. Fish. Res.
Board Can. 26:1429-1438.
Barbour MT, Gerritsen J, Snyder BD, and Stribling JB. 1999. Rapid Bioassessment
Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic
Macroinvertebrates and Fish, 2nd edition. EPA 841-B-99-002. US Environmental
Protection Agency, Office of Water, Washington, DC, USA.
Barton BA, Morgan JD, and Vijayan MM. 2002. Physiological and condition-related
indicators of environmental stress in fish. In Biological indicators of aquatic
ecosystem stress. Adams SM (ed). American Fisheries Society, Bethesda, MD,
USA. pp. 111-148.
Chow TL, Rees HW, and Monteith J. 2000. Seasonal distribution of runoff and soil loss
under four tillage treatments in the upper St. John River valley New Brunswick,
Canada. Can. J. Soil Sci. 80:649-660.
Environment Canada. 2001a. Threats to sources of drinking water and aquatic
ecosystem health in Canada. NWRI Scientific Assessment Report Series No.1.
National Water Research Institute, Burlington, ON, Canada.
Environment Canada. 2001b. Canadian daily climate data. CD-ROM.
Fitzgerald DG, Lanno RP, and Dixon DG. 1999. A comparison of a sentinel species
evaluation using creek chub (Semotilus atromaculatus) to a fish community
evaluation for the initial identification of environmental stressors in small streams.
Ecotoxicol. 8:33-48.
Galloway B, Munkittrick KR, Currie S, Gray M, Curry RA, and Wood C. 2003.
Examination of the responses of slimy sculpin (Cottus cognatus) and white sucker
(Catostomus commersoni) collected on the Saint John River downstream of pulp
mill, paper mill, and sewage discharges. Environ. Toxicol. Chem. In press.
Gibbons WN, and Munkittrick KR. 1994. A sentinel monitoring framework for identifying
fish population responses to industrial discharges. J. Aquat. Ecosys. Health. 3:227237.
Gibbons WN, Munkittrick KR, McMaster ME, and Taylor WD. 1998. Monitoring aquatic
environments receiving industrial effluents using small fish species 1: response of
spoonhead sculpin (Cottus ricei) downstream of a bleached-kraft pulp mill. Environ.
Toxicol. Chem. 17:2227-2237.
Gray MA, Curry RA, and Munkttrick KR. 2002. Non-lethal sampling techniques for
assessing fish populations for environmental assessment. Water Qual. Res. J. Can.
37:195-211.
70
Greeley MS. 2002. Reproductive indicators of environmental stress. In Biological
indicators of aquatic ecosystem stress. Adams SM (ed). American Fisheries
Society, Bethesda, MD, USA. pp 312-377.
Hamilton H, Paine MD, Gibson W, and Conley D. 1993. 1992 Operational monitoring
survery of the Lesser Slave River. Vol. VI. Prepared for the Slave Lake Pulp
Corporation by EVS Consultants Ltd., North Vancouver, BC, Canada.
Karr JR, and Schlosser IJ. 1978. Water resources and the land-water interface.
Science. 201:229-234.
McKim JM. 1977. Evaluation of tests with early life stages of fish for predicting longterm toxicity. J. Fish. Res. Board Can. 34:1148-1154.
Munkittrick KR, and McMaster ME. 2000. Effects-driven assessment of multiple
stressors using fish populations. In Multiple stressors in ecological risk and impact
assessment: Approach to risk estimation. Ferenc SA, and Foran JA (eds). Society
of Environmental Toxicology and Chemistry Press. Pensacola, FL. pp. 27-65
Munkittrick KR, McMaster ME, Van Der Kraak G, Portt C, Gibbons WN, Farwell A, and
Gray M. 2000. Development of methods for effects-driven cumulative effects
assessment using fish populations: Moose River project. Society of Environmental
Toxicology and Chemistry. Pensacola, FL, USA.
Mutch JP, Savard MA, Julien GRJ, MacLean B, Raymond B, and Doull J. 2002.
Pesticide monitoring and fish kill investigations on Prince Edward Island, 19941999. In Effects of land use practices on fish, shellfish, and their habitats on Prince
Edward Island. Cairns DK (ed). Can. Tech. Rep. Fish. Aquat. Sci. no. 2048 pp. 94115.
NB DAFA. 2000. Annual report 1999-2000. New Brunswick Department of Agriculture,
Fisheries, and Aquaculture. Fredericton, NB, Canada.
NB DAFA. 2002. Annual report 2000-2001. New Brunswick Department of Agriculture,
Fisheries, and Aquaculture. Fredericton, NB, Canada.
Otto RG, and Rice JO. 1977. Responses of a freshwater sculpin (Cottus cognatus
gracilis) to temperature. Trans. Am. Fish. Soc. 106:89-94.
Rand GM, Wells PG, and McCarty LS. 1995. Introduction to aquatic toxicology. In
Fundementals of aquatic toxicology. Rand GM (ed). Taylor Francis, Washington
DC, USA. pp 3-67.
Scott WB, and Crossman EJ. 1998. Freshwater fishes of Canada. Galt House
Publications Ltd. Oakville, ON, Canada.
Sokal RR, and FJ Rohlf. 1995. Biometry. 3rd ed. WH Freeman and Co. New York, NY,
USA.
71
Symons PEK, Metcalfe JL, and Harding GD. 1976. Upper lethal and preferred
temperatures of the slimy sculpin, Cottus cognatus. J. Fish. Res. Board Can.
33:180-183.
Teather KL, Harris M, Boswell JL, and Gray MA. 2001. Effects of Acrobat-MZ ® and
Tattoo-C® on Japanese medaka (Oryzias latipes) development and adult male
behaviour. Aquat. Toxicol. 51:419-430.
Vannote RL, and Sweeney BW. 1980. Geographic analysis of thermal equilibria: A
conceptual model for evaluating the effect of natural and modified thermal regimes
on aquatic insect communities. Am. Naturalist. 115:667-695.
Van Vliet WH. 1964. An ecological study of Cottus cognatus (Richardson) in northern
Saskatchewan. MA Thesis, University of Saskatchewan. Saskatoon, SK, Canada.
Waters TF. 1995. Sediment in streams: Sources, biological effects and control.
American Fisheries Society Monograph 7. Bethesda, MD, USA.
72
CHAPTER 3
Investigating the impacts of sediment and temperature on slimy sculpin (Cottus
cognatus) populations in agricultural catchments2
3.1
Abstract
A study carried out in 1974 (Welch et al., 1977) found reductions in the number of
brook trout (Salvelinus fontinalis), slimy sculpin (Cottus cognatus), and benthos
associated with increased sedimentation in farmed catchments. During the summer of
2001, we repeated a similar study in the same region of New Brunswick with 20
streams in forested (FOR) or agricultural (AGR) catchments. The streams in
agricultural catchments were warmer than streams in forested catchments (median =
16.0ºC and 13.3ºC, respectively). The elevated temperature was associated with
increased sizes and decreased densities of young-of-the-year (YOY) slimy sculpin, and
all sites where water temperatures reached 25ºC were devoid of YOY sculpin.
Sediment deposition was greatest at the agricultural sites, with increased fine
sediments deposited at agricultural sites and an increase in larger, coarse sands at two
sites with active forestry operations. Sites that were common to both the 1974 and
2001 studies (n=11) were compared, and we found that relative fish abundance did not
change much between 1974 and 2001. Five agricultural sites from 1974 that reported
no sculpin now had at least one adult sculpin, though none of them had YOY sculpin
present. There are indications of slight improvements over the past 27 years; however,
the absence of YOY fish in agricultural regions is a major concern. Although this study
did not find a significant effect of sediment on the fish population, sediment is known to
2
This chapter is currently in preparation for submission to the journal Environmental
Toxicology and Chemistry under joint authorship with R. Allen Curry and Kelly R. Munkittrick.
73
be associated with smothering, habitat damage, transport of chemicals off the fields
and reductions in food availability.
3.2
Introduction
Intensive potato farming places high levels of stress on streams and rivers
associated with the chemicals, nutrients, and soil that enter water systems during runoff events. In recent years, there has been an increase in the frequency and magnitude
of fish kills downstream from potato farms following major storm events in Atlantic
Canada (Mutch et al. 2002). Chemical agents are suspected as a major factor in these
episodic events, and agrochemicals have been detected in winter groundwater
discharges through redds (nests) of brook trout (Salvelinus fontinalis) in potato-growing
regions (Curry and MacNeill 2004). However, assessing the sublethal impacts of
agricultural run-off is complicated by the unknown relative impact of direct and indirect
stressors, and acute and chronic exposures. Possible indirect effects include habitat
alteration and degradation, impacts on invertebrate food sources, changes to the local
hydrology, increased temperatures, and reduced dissolved oxygen (Waters 1995).
Within the province of New Brunswick, 90% of all potato production is located within
the western (55%) and northwestern (35%) regions (NB DAFA 2000). In 1974, Welch
et al. (1977) surveyed northern New Brunswick streams for impacts of agriculture and
clear-cutting on fish, benthos, and physical factors. In streams draining agricultural
lands, they found 54% fewer trout, 92% fewer sculpin, and 64% less benthos when
compared with reference streams. The authors cited chemical contamination and
sedimentation as the most likely contributing factors. We have been investigating the
impacts of potato cultivation on slimy sculpin (Cottus cognatus) populations in Little
River in the same area of northwestern New Brunswick (Gray et al. 2002). We
74
observed that young-of-the-year (YOY) sculpin were less abundant and were larger in
agricultural compared to forested streams in the same system. Factors influencing the
growth of the YOY sculpin in the agricultural region may include food availability,
eutrophication, reduced competition, temperature, or impacts from pollution.
The slimy sculpin is a small, cool-water benthic fish species that inhabits both lotic
and lacustrine environments. They are classified as a lithophilous fish species, which
require clean, stony substrates for reproduction (Balon 1975). Sculpin spawn on the
underside of rocks with males tending the nest of eggs until the larvae emerge (Van
Vliet 1964). Sculpin feed on insect larvae that inhabit gravel and cobble substrates.
Berkman and Rabeni (1987) found that increasing fine sediment in streams has the
largest negative effect on benthic insectivores, herbivores and lithophilous fish. Given
these life history characteristics, the slimy sculpin has the potential to be seriously
impacted by sediment added to streams.
We initiated the present survey to determine if our initial observations of impacts on
YOY sculpin in an agricultural region of a single river persisted across a larger spatial
scale. In addition, we wanted to investigate the relationship between sculpin size and
distribution with sediment deposition and stream temperature. We quantified sediment
deposition and temperature profiles for 9 forested and 11 agricultural sites located in
northwestern New Brunswick between July and October 2001. The resident fish
population was assessed to investigate correlations with abiotic measurements. With
11 sites in common with the Welch et al. (1977) survey, a tertiary objective was to
compare temporal scale changes in fish abundance after almost three decades.
75
3.3
Methods
Site selection
Attempts were made to locate the original sites of Welch et al. (1977). They were
categorized as control, clear-cut, and farmed sites. The present study had greater
resolution of land-use based on available GIS information, and sites were more
accurately classified as either forested (considered as the reference condition) or
agricultural streams. Eleven sites from the older study were sampled (4 forested, 7
agricultural), with an additional 5 forested and 4 agricultural sites added for this study
(Figure 3.1). Within each land-use category, sites were numbered in ascending order
based on the width of the site (Table 3.1). Reasons for not using an original 1974 site
included absence of current access roads, dried-up or diverted streambeds, inaccurate
coordinates, or a change in the surrounding land-use.
Sediment traps
Aluminium cans (17.5cm x 15.5cm; height x diameter) with solid walls were installed
at each site to quantify sediment deposition. Cans were filled with crushed commercial
gravel (>1.9cm) that was sieved and washed over 0.6cm screen. Three cans were
buried with the brim flush with the stream substrate across (>1m separation) the head
of the same riffle where possible (n=12; 5 forested, 7 agricultural sites), and at the head
of successive riffles if the stream was too narrow (n=8; 4 forested, 4 agricultural sites).
Cans were installed 1-2 July, and one can from each site was retrieved 1, 2, and 3
months post-installation. The retrieval process involved placing a plastic lid over the
top of the can and lifting it out of the substrate. The water within the can was decanted
into a separate container, and both were placed in coolers for transport back to the
laboratory. At the laboratory, the contents of individual cans and the decanted water
76
were mixed in steel trays. Samples were dried at approximately 50°C for a minimum of
24h. The dried sample was separated into three subsamples that were passed through
a series of sieves (37, 150, 250, and 500µm, and 1 and 2mm) and shaken for 5min.
The total amount of dry material per grain size was weighed to the nearest 1g.
Fish Assessment
Fish were surveyed after the removal of the final sediment sample (9-18 October).
Fish were collected using a backpack electrofishing unit (Smith-Root C-15) with a
standardized effort of two persons with dipnets and 1000s of electrofishing output.
Barrier nets were not installed, as we have previously found no significant differences
with sculpin collection in open versus closed sites using one sweep through an area.
All fish were identified and measured for fork length, or total length for sculpins (±1mm),
weight (±0.01g), and were then released back into the site where they were collected.
Site data
Temperature recorders (12-bit, Minilog-TR, Vemco Ltd. Shad Bay, NS, Canada)
were placed at each site immediately downstream of the sediment samplers where they
recorded hourly temperatures. Temperature was recorded beginning in July prior to
sediment sampling until mid-October. The timing encompassed the period of potential
growth for YOY sculpin, from the time of approximate emergence from the nests to the
end of growth for the first growing season (Gray et al. 2002). For purposes of
temperature data analysis, a common time period (5 Jul – 8 Oct) was examined when
recorders were present at all sites.
Water samples were collected at each site in December 2001 using sterile Nalgene
bottles. The samples were kept at approximately 4°C, and transported back to the lab
77
for standard water chemistry and analysis (alkalinity, conductivity, hardness, colour, Ca,
Mg, nitrate, pH, organic C, Ph, NH3, total N, turbidity). Sediments from 16 sites (10
agriculture, 6 forested) were analyzed by dry combustion using a carbon, nitrogen, and
sulphur elemental analyzer (LECO CNS-2000).
Comparison of fish communities (1974 and 2001)
The methods used for the fish survey differ between 1974 and 2001, complicating
direct comparisons of the data. In 1974, multiple sweeps (4-5) of the same section
were made and the densities of fish were expressed as # of fish per 100m (Welch et
al., 1977). Due to variations in habitat quality and complexity, we standardized fishing
effort by conducting one sweep through the site for a standard time of 1000s of
electrofishing output. Although this standardized the time spent collecting fish, it did not
standardize the amount and type of habitats that were covered among sites.
Expressing our fish abundance by 100m of stream length was not comparable to the
1974 study and would significantly overestimate fish abundance in wide streams
relative to narrow streams. To compare the data between years, we ranked the sites
by the relative abundance of trout or sculpin. The change in rankings was then used to
compare relative fish abundance within a land-use region.
Data treatment
Normality and homoscedasticity were assessed by visual examination of normal
probability plots and residual plots, respectively. Temperature and sediment
accumulation data were analyzed for differences among land-use regions using nonparametric Mann-Whitney U tests. The relationships between mean daily temperature
fluctuation and stream size, and YOY sculpin size and density were assessed using
78
linear regression. A two-factor ANOVA was used to assess the difference in sediment
accumulation in the two land-use regions over time (fixed factors: month (1-3) and landuse (FOR vs. AGR)). Relationships between temperature and sediment deposition
(independent variables), and fish size and density (dependent variables assessed
separately) were assessed using a backwards step-wise linear regression analysis (F
to remove set at 3.9). A two-factor ANOVA was used to assess the composition of the
sediment with respect to land-use (FOR vs. AGR; fixed factor) and sediment
accumulation (low vs. high; fixed factor). The percentage of carbon, nitrogen, and
sulphur bound to sediments was transformed (arcsine) and analyzed using MannWhitney U-tests with accumulation (low vs. high) as the grouping factor. Statistical
analysis was completed using Systat© (v10, SPSS Inc., Chicago), with the exception of
the step-wise linear regression which was completed using SigmaStat© (v2.03, SPSS
Inc., Chicago).
3.4
Results
Temperature
Although patterns of the temperature profiles followed predicted seasonal trends for
the two land-use regions, the average daily temperatures were higher in the agricultural
sites compared to the forested sites (Figure 3.2; U=7003, df=1, p<0.001). The median
temperatures were 16.0 and 13.3°C for the agricultural sites and forested sites,
respectively. The greatest difference between the regions occurred in late July during
the warmest period of the study, with the temperature profiles converging by early
October.
Cumulative degree-days (sum of mean daily temperatures) during the study period
were significantly higher in the agricultural sites than the forested sites (U=84, df=1,
79
p=0.009; Table 3.2). Degree-days above 10ºC were calculated as a surrogate for
growing time (see Munkittrick et al. 2000) and were likewise elevated in the agricultural
regions (U=81, df=1, p=0.017). In general, the temperature dropped below 10ºC
sooner in the forested than the agricultural regions. The mean daily difference between
the daily maximum and minimum temperatures was calculated to provide an estimate
of the daily temperature fluctuation experienced at each site. There was no significant
difference between mean daily differences in the forested and agricultural regions
(U=51, df=1, p=0.91). There was a significant positive relationship between mean daily
fluctuation and the rank of stream width in the forested region (linear regression,
r2=0.90, p<0.001), but not in the agricultural region (r2=0.31; p=0.08). The number of
hours that the temperature was above 23ºC and 25ºC were also calculated as a relative
indication of thermal stress for sculpin (Symons et al. 1976; Otto and Rice 1977) at
each site (Table 3.2). Temperatures exceeded 25ºC at four of the agricultural sites.
Precipitation
Precipitation data (Environment Canada 2001) for the period 1 July to 1 October
showed that the central study sites received more rain than the northern and southern
sites (Figure 3.3). The total rainfall recorded during that period for St. Leonard, Bon
Accord, and Woodstock was 281.2, 356.8, and 273.6mm, respectively.
Sediment deposition
There was not a significant relationship between time (1, 2 or 3 months) and
sediment accumulation in the forested and agricultural sites (ANOVA; F=0.33, df=2,
p=0.72), and so sediment data from the three cans at each site were combined for
subsequent analysis. Within a land-use region, total accumulations of sediment were
80
greater in agricultural streams (median = 334.4g) than forested streams (median =
76.7g; U=600, df=1, p=0.009). Accumulation by site varied within land-use region, with
some sites collecting a minimal amount of material regardless of surrounding land-use
(Figure 3.4). Moderate to heavy siltation typically occurred in agricultural streams, but
appreciable amounts of fine material did not accumulate at all sites. At sites A6 and
A9, it was observed that the gravel substrate at the top of these cans was covered in a
thick layer of periphyton. Although this may have prevented sediments from depositing
in the can since the surrounding substrate was covered in fine sediments, we did not
quantify the periphyton and have no further evidence to support the relationship with
reduced sediment deposition. Based on observed background accumulations of
sediment among the forested sites (7 sites; 45-117g), we chose 200g as a conservative
indication of natural sediment deposition for streams in these regions (Figure 3.4). Two
forested and seven agricultural streams had deposition above this level.
Grain size analysis
For streams with “above natural” deposition, we observed that within the two
forested streams, coarse to very coarse sand dominated the collections (40-70% by
weight; Figure 3.5). These two streams were located in catchments with active forest
operations, and as the study progressed, clear-cutting was observed upstream of site
F8. Silt and fine sand (35-80%) dominated deposited materials in the agricultural
streams (Figure 3.5).
Fish abundance
Slimy sculpin and brook trout dominated the fish communities occurring at 17 of the
20 sites (Table 3.3). Fish that were present in forested sites, but were absent from
81
agricultural sites included Atlantic salmon (Salmo salar), lake chub (Couesius
plumbeus), and burbot (Lota lota). There are some natural restrictions to Atlantic
salmon due to hydroelectric facilities, so the number of sites was adjusted to account
for those where salmon were possibly present.
Slimy sculpin
YOY sculpin were identified by plotting length-frequency distributions for each site.
Slimy sculpin were completely absent from 3 agricultural streams, and only 2 of the
remaining 8 streams had young-of-the-year (YOY) sculpin present (Table 3.4). Where
sculpin (all sizes) occurred, their median densities were 0.90 and 0.22/m2, in the
forested and agricultural catchments, respectively. The two agricultural streams with
YOY sculpin had substantial densities, similar to and greater than most forested sites.
Young-of-the-year sculpin comprised between 10 and 80% of the sample collected at
forested sites, which was not significantly different from the agricultural streams (U=14,
df=1, p=0.24).
At the forested sites, temperature was a significant variable for predicting median
YOY sculpin size (step-wise regression, F=37.59, r2= 0.84, p<0.001; Figure 3.6 A), and
density (all sizes) (F=24.48, r2=0.78, p=0.003; Figure 3.6 B). Sediment deposition was
not a significant variable for predicting sculpin size or density in the forested region
(F=0.01 and 0.38, and p=0.91 and 0.56, respectively). Within the agricultural sites,
neither temperature or sediment deposition was a significant variable for predicting
sculpin density (F=2.88 and 1.59, and p=0.13 and 0.24, respectively).
If agricultural sites are excluded (due to small sample size), the median size of YOY
sculpin was negatively related to sculpin density (linear regression, r2=0.58, df=8,
82
p=0.020; Figure 3.6 C). Although, the relationship is not strong, this indicates some
support for a density-dependent effect on the size of YOY sculpin.
Water quality and sediment
The forested streams had lower conductivity (91 µS/cm) than the agricultural
streams (309 µS/cm). Several parameters that can be linked to agricultural operations
were elevated in the agricultural streams including alkalinity, magnesium, nitrate, and
total nitrogen (Table 3.5). No differences were seen in the colour, total organic carbon,
total phosphorus, and turbidity, which may have been due to the timing of water
sampling (i.e. after the active agriculture season had ended).
There were differences detected among the inorganic compounds (carbon,
nitrogen, and sulphur) bound to the deposited sediments only when accumulation (low
or high; low< 200g and high>200g) was taken into account (ANOVA; all F>74, df=12,
all p<0.001). The proportion of carbon, sulphur, and nitrogen bound to accumulated
sediments were all lower at sites where accumulation was greater than 200g,
regardless of land-use (Tukey’s test; all p<0.05).
Comparison of fish communities (1974 and 2001)
There were no trends in the changes in brook trout abundance (ranked parameter)
between survey periods (Table 3.6). There was a shift from low to higher abundance of
trout in some streams; however, there were an equal number of streams that declined
to a lower relative abundance. In the agricultural regions, there are more sites that
improved in brook trout abundance over the 27-year period. The best agricultural site
(A6) in 1974 continued to maintain the highest relative abundance of trout in 2001.
83
There were no changes in the relative rankings of sculpin abundance among the
forested sites between 1974 and 2001 (Table 3.6). Of the five agricultural sites that
reported no sculpin in 1974 (A1 through A5), four of these sites were found to contain
at least one adult sculpin in 2001, and the remaining site (A2) was still devoid of
sculpin. Of the two sites that reported sculpin in 1974, A10 still had the highest sculpin
abundance among agricultural sites, while A6 was now devoid of sculpin. Densities of
sculpin collected at the comparison sites ranged from 0.68-1.40/m2 in the forested
regions compared to 0.02-0.32/m2 at the agricultural sites.
3.5
Discussion
In forested streams, higher water temperature was significantly associated with
increased size and density of slimy sculpin. Young-of-the-year (YOY) sculpin were
larger at warmer sites, and densities of all sculpin decreased as maximum
temperatures increased. The increased size of YOY sculpin is possibly due in part to
density-dependent effects, though the relationship is not strong. In the agricultural
sites, sculpin density also decreased with increasing water temperatures, but the data
were much more variable and the relationship was not significant. The size of YOY
sculpin at agricultural sites could not be evaluated because they were present at only 2
of the 11 agricultural sites. The absence of YOY sculpin at 9 of the agricultural sites
suggests that there are other factors than temperature influencing density at those
sites.
These observations confirm our previous findings of larger and fewer YOY sculpin
at agricultural sites than forest sites in the Little River (Chapter 2), but over a much
broader spatial scale and range of stream size influenced by varying degrees of
surrounding land-use activities. Although it was predicted a priori that sediment would
84
play an important role in determining the performance of sculpin (abundance, growth,
survival), we were unable to provide evidence to support any such correlation. It is a
challenge to discriminate the effects of sedimentation from other effects of agricultural
activities including increased temperatures, and chemical and nutrient inputs (Waters
1995), and will require controlled experiments to elucidate their individual contributions.
The temperature of running waters is dictated by factors such as elevation, regional
climate, season, and relative groundwater inputs (Allan 1995). In the same region,
larger rivers will tend to be warmer than small rivers and streams because they are
wider and more open, thus receiving less shading from streamside vegetation.
Clearing of lands adjacent to smaller waterways for agriculture can modify the thermal
regime through a reduction in canopy shading and as a consequence of altered
hydrology from increased erosion and decreased inputs of woody debris (Stauffer et al.
2000). Our results support this general observation with temperatures, on average,
warmer in the streams surrounded by agricultural fields and in the wider forested
streams. Daily fluctuations between the minimum and maximum temperatures were
also greatest at the agricultural sites, where there was little or no canopy adjacent to or
over the stream.
As a cool-water stenotherm, the sculpin has a narrow temperature range with an
upper lethal limit of approximately 23-25°C (Symons et al. 1976; Otto and Rice 1977),
and would thus be unable to survive in waters above 25ºC for very long (Kuehne 1962).
Symons et al. (1976) suggested that sustained temperatures over 19ºC would lead to a
decrease or disappearance of brook trout and slimy sculpin. Edwards (2001) also
found that slimy sculpin density decreased as water temperature increased, and that
there was a most dramatic decrease at 22ºC. We found that densities of sculpin began
to drop off around 19ºC at forested sites, which may be indicative of a sublethal
threshold where avoidance becomes a favourable strategy and fish move out of the
85
warmer sites. At a few agricultural sites where sculpin densities were low in water
temperatures <19ºC, sediment depositions in the samplers were high and there may
have been other stresses associated with agricultural activities that were influencing the
fish. There may be a point where sediment, or other stressors, may affect sculpin
density, and then as temperature increases, thermal stress becomes the most
important limiting factor.
In the present study, all sites with temperatures over 25ºC were devoid of YOY
sculpin, and densities of sculpin declined with increasing temperatures among sites.
The two sites with temperatures above 23ºC and YOY sculpin were forested. It may be
that the fish are better able to seek out cool water refugia where there is cover
available. The warmest site (A9) was in the agricultural region and was devoid of
sculpin. Blacknose dace (Rhinichthys atratulus), which have an upper lethal limit at
around 29ºC (Wismer and Christie 1987), were the only fish species present.
Increased temperatures would predictably aid in facilitating a shift to more eurythermal
species, or those able to adjust to a wider temperature range. Although there was no
significant difference between forested and agricultural regions overall for the mean
daily temperature fluctuations, the greatest fluctuations were associated with smaller
streams in the agricultural region than the forested region. Temperature fluctuations
may play an important role in affecting the distribution for stenothermal species like the
sculpin that may not be able to adjust rapidly enough.
Growth is highly temperature-dependent in fish, with warmer temperatures resulting
in increased feeding, digestion, and respiration rates (e.g. Brett 1971; Ricker 1979).
YOY sculpin size was positively related, and sculpin density was negatively related,
with temperature in the forested streams. Although the data were variable in the
agricultural streams, the data from these sites suggest that temperature also plays a
role in determining size and density of sculpin. However, there are admittedly many
86
other factors that may be influencing fish at agricultural sites. There are multiple, nonexclusive hypotheses related to temperature and the increased growth of the YOY
sculpin at warmer sites. Before temperature becomes a limiting factor spawning may
be initiated earlier in the warmer streams, egg development may proceed at a faster
rate, emergence may occur earlier, and there may be a longer growing period at the
warm sites. Excluding the effect of temperature, decreased density may affect growth
through reduced competition (Anderson 1985), though we found only a weak
relationship between YOY sculpin size and sculpin density.
Similar to observations made in previous work with sculpin (Gray et al. 2002), young
sculpin were scarce or absent from almost all agricultural streams. It is not known
whether this decrease occurred as a result of decreased fecundity, decreased survival
at the egg or larval stage, or a combination of the two. The influence of increased
temperatures on body size generally also influences fecundity, as the two are positively
related (Vannote and Sweeney 1980). Female sculpin in the agricultural region tend to
have reduced fecundities, despite warmer temperature regimes (Chapter 2).
Observations during storm events of the rivers and streams adjacent to and
downstream from potato fields suggest that sedimentation should play a significant role
in reducing the availability and quality of spawning habitat. More research is needed to
investigate the relationship between sediment deposition and the spawning frequency
and success of sculpin.
The deposition of sediments differed between land-use regions both in the amount
and the type of sediment accumulated. Lisle and Eads (1991) reviewed methods to
assess sedimentation in streams and found that solid-walled containers were adequate
for measuring infiltration of sediment at the streambed level. The use of the cans as
sediment samplers precludes the ability to quantify intergravel transport of sediments,
but we were most interested in deposition of sediment from the water column. We do,
87
however, have important concerns regarding the relative impact that algal growth may
have had on sediment infiltration, and also the lack of change in sediment deposition
over time. While the use of the aluminium cans was primitive and possibly flawed, they
were adequate to show the relative increase in sediment deposition and accumulation
of fine materials in the agricultural streams compared with the forested streams.
Beyond that conclusion, however, it is unclear whether the lack of relation between
sedimentation and sculpin was due solely to methodological issues with the samplers,
or whether there really is not a relationship. Currently, there is no general consensus
for characterizing sediment in streams for biological surveys (Mebane 2001). Studies
are currently being developed to evaluate ways to assess sediment that allow for
application to biological data.
Farming practices have improved significantly from almost 30 years ago when
Welch et al. (1977) carried out their survey of New Brunswick streams. Discarded
chemical containers were routinely observed downstream of sites where pesticide
sprayers filled their tanks using stream water. Although this still remains an occasional
sight in this region of Canada, restrictions and regulations concerning pesticide use and
handling have reduced the likelihood of chemical contamination from this practice.
Sediment transport from fields has also decreased, though in most regions there are no
regulations requiring farms to adopt sediment-loss controls.
Promotion of Best Management Practices, the introduction of the Atlantic
Environmental Farm Plan Initiative in 1995, and effective producer awareness activities
by non-profit organizations like the Eastern Canada Soil and Water Conservation
Centre have meant that soil conservation practices have become an important element
of the potato production systems in New Brunswick's Upper St. John River Valley.
Undulating slopes, excessive runoff, highly erosive soils, soil compaction, stone
picking, organic matter depletion, loss of soil available water holding capacity and crop
88
productivity have compelled some potato producers to adopt and implement soil
conservation systems to ensure the long term sustainability of agriculture. While the
adoption of soil conservation practices has been difficult due to market pressures,
farming efficiency, land tenure and land availability issues, producers are becoming
increasingly aware of the need to adopt agronomically, environmentally and
economically sustainable production systems.
3.5.1
Conclusions
Our main objective was to investigate the relationships between temperatures and
sediment deposition with resident sculpin populations in streams draining agricultural
and forested lands. We confirmed previous observations of reductions to young-of-theyear sculpin densities and increased growth in warmer agricultural streams.
Comparisons with a similar study completed in 1974 suggest little change in the
populations of sculpin in agricultural streams, where YOY sculpin were absent. Our
results suggested that temperature was strongly related to fish abundance, density, and
growth. Although there was no relationship seen with the increased sediment
deposition in agricultural streams, this factor needs to be addressed more thoroughly.
A better understanding of the links between the abiotic and biotic factors, and how to
measure these, is required to assess the relative impacts of multiple stressors
introduced by agricultural activities.
89
3.6
Acknowledgments
Funding for this study was provided primarily by the Toxic Substances Research
Initiative (TSRI), with additional support from Crop Life Canada. MAG was funded by a
post-graduate scholarship from the Natural Sciences and Engineering Research
Council (NSERC), and KRM is supported as a Canadian Research Chair. The field
and lab work was conducted with the invaluable help of technicians from the New
Brunswick Cooperative Fish and Wildlife Research Unit. Water analysis was
conducted by the New Brunswick Department of Environment and Local Government,
and sediment analysis was conducted at the Laboratory for Forest Soils and
Environmental Quality at the University of New Brunswick. Jean-Louis Daigle and
Gordon Fairchild of the Eastern Canada Soil and Water Conservation Centre provided
information regarding regional agricultural practices.
90
Table 3.1. Location and site identification numbers for the streams used in the study. Streams are numbered in ascending
order from smallest to largest stream width in forested (F), and agricultural (A) regions. Asterisks indicate sites common to
Welch et al. (1977). The mean width and length of sites sampled for fishing survey are also listed.
Stream
W. Br. N. Monquart Brook
Big Forks Brook.
East Beaver Brook
Price Brook.
L. Hand Grindstone Brook.
Muniac Stream
Little River (@Irving)
Yellow Brook
Shikatehawk Stream
Site ID
*F1
*F2
*F3
F4
*F5
F6
F7
F8
F9
Latitude
46º 38’ 01”
47º 16’ 44”
46º 48’ 18
47º 18’ 40”
47º 17’ 20”
46º 38’ 35”
47º 13’ 49”
47º 20’ 07”
46º 28’ 39”
Longitude
67º 33’ 44”
67º 40’ 23”
67º 15’ 48”
67º 29’ 32”
67º 30’ 34”
67º 40’ 15”
67º 37’ 08”
67º 42’ 14”
67º 34’ 25”
Mean width (m)
1.6
2.5
3.0
3.2
4.5
4.9
7.1
7.5
14.1
Mean length (m)
20.5
13.7
13.0
7.9
8.6
12.8
6.3
8.0
9.1
Mooney Brook
Godin Brook
Hunters Brook
Harpers Brook
Black Brook
Whitemarsh Brook
Outlet Brook
Dead Brook
Holmes Brook.
Bulls Creek
Little River (@Dead Brook)
*A1
*A2
*A3
*A4
*A5
*A6
A7
A8
A9
*A10
A11
47º 04’ 03”
46º 59’ 56”
46º 26’ 16”
46º 12’ 27”
47º 05’ 24”
46º 29’ 32”
46º 59’ 29”
47º 06’ 05”
46º 34’ 51”
46º 06’ 11”
47º 05’ 02”
67º 37’ 23”
67º 38’ 11”
67º 39’ 55”
67º 35’ 22”
67º 43’ 46”
67º 39’ 42”
67º 35’ 49”
67º 43’ 25”
67º 37’ 48”
67º 36’ 51”
67º 42’ 52”
1.6
2.2
2.3
2.8
3.3
4.8
5.0
6.3
7.5
7.9
14.2
34.6
25.3
48.0
37.2
12.5
23.1
11.6
7.0
13.4
16.3
6.3
91
Table 3.2. Summary of temperature statistics for streams in forested and agricultural catchments of northwestern New
Brunswick. The values represent temperatures measured between 5 July and 8 October, 2001.
Site
Forested
F1
F2
F3
F4
F5
F6
F7
F8
F9
Median
Agricultural
A1
A2
A3
A4
A5
A6
A7
A8
A9
A10
A11
Median
Degree-days
Degree-days
>10ºC
Mean daily
difference
(maxT-minT)
#h >
23ºC
1192.8
1096.7
1216.4
912.7
1110.1
1343.4
1268.4
1320.5
1594.0
1216.4
1084.9
911.9
1111.0
460.0
920.7
1307.2
1138.4
1202.9
1558.6
1111.0
1.88
1.78
2.23
3.61
3.56
3.67
4.18
4.25
4.6
3
60
1645.8
1246.3
1616.5
1386.8
1271.6
1411.8
1531.4
1282.2
1708.4
1533.3
1280.0
1411.8
1609.4
1137.4
1589.1
1306.8
1167.5
1322.7
1488.6
1139.3
1681.9
1473.0
1191.2
1322.7
4.09
3.02
3.30
3.37
3.34
4.46
4.86
4.31
4.79
3.98
1.88
92
#h >
25ºC
181
34
81
6
3
36
260
25
99
2
Table 3.3. Presence/absence of fish species in forested (n=9) and agricultural sites
streams (n=11) in northwestern New Brunswick (October 2001). Values are the
number of sites with the particular fish species present. The values in brackets
represent the adjusted number of sites where Atlantic salmon could be present
(restricted in some cases due to hydroelectric facilities).
Species
Forested
Agricultural
Slimy sculpin (Cottus cognatus)
9
8
Brook trout (Salvelinus fontinalis)
8
9
3 (6)
0 (8)
Blacknose dace (Rhinichthys atratulus)
2
5
White sucker (Catostomus commersoni)
1
3
Creek chub (Semotilus atromaculatus)
1
4
Common shiner (Notropis cornutus)
1
1
Lake chub (Couesius plumbeus)
1
0
Burbot (Lota lota)
1
0
Atlantic salmon (Salmo salar)
93
Table 3.4. Summary of slimy sculpin abundance, density, and size of YOY sculpin collected in forested and agricultural
streams of northwestern New Brunswick (October 2001).
Site
# YOY
sculpin
Total density
(/m2)
YOY density
(/m2)
Maximum size YOY
(mm)
% YOY in
population
22
48
26
60
36
63
40
15
28
3
11
9
12
12
13
11
12
3
0.69
1.40
0.68
2.37
0.94
1.00
0.90
0.25
0.22
0.09
0.32
0.23
0.47
0.31
0.21
0.25
0.20
0.02
30
36
40
24
36
42
39
40
44
13.6
22.9
34.6
20.0
33.3
20.6
27.5
80.0
10.7
1
-3
10
13
-7
102
-41
79
-------40
--23
0.02
-0.03
0.10
0.32
-0.12
2.33
-0.32
0.89
-------0.91
--0.26
-------40
--44
-------39.2
--29.1
Total #
sculpin
Forested
F1
F2
F3
F4
F5
F6
F7
F8
F9
Agricultural
A1
A2
A3
A4
A5
A6
A7
A8
A9
A10
A11
94
Table 3.5. Water quality and sediment analysis in forested and agriculture sites.
Values represent mean (±SE), sample sizes are shown in brackets (n), and asterisks
indicate a significant difference between land-use regions (Mann-Whitney U test;
p<0.05).
Forested
Agriculture
Alkalinity (mg/L) *
37.11 ± 8.97 (9)
120.36 ± 12.72 (11)
Calcium (mg/L) *
15.15 ± 3.76 (9)
54.46 ± 6.24 (11)
Colour (1-100, light-dark)
8.33 ± 5.34 (9)
5.45 ± 1.96 (11)
90.57 ± 16.02 (9)
309.0 ± 34.14 (11)
Magnesium (mg/L) *
1.43 ± 0.25 (9)
4.87 ± 0.51 (11)
Nitrate (mg/L) *1
0.45 ± 0.09 (9)
2.59 ± 0.58 (11)
pH *
7.75 ± 0.10 (9)
8.22 ± 0.02 (11)
Total Organic Carbon (mg/L)
3.36 ± 1.05 (7)
3.47 ± 0.46 (11)
0.008 ± 0.002 (4)
0.012 ± 0.002 (8)
0.048 (1)
0.226 ± 0.128 (3)
Total Hardness (mg/L) *
43.71 ± 9.39 (9)
154.99 ± 16.75 (11)
Total Nitrogen (mg/L) *
0.64 ± 0.10 (5)
2.74 ± 0.60 (11)
Turbidity (NTU)
0.91± 0.26 (9)
1.04 ± 0.22 (11)
% carbon
9.43 ± 2.38 (6)
5.95 ± 1.15 (10)
% sulphur
0.12 ± 0.04 (6)
0.07 ± 0.02 (10)
% nitrogen
0.59 ± 0.15 (6)
0.42 ± 0.07 (10)
WATER
Conductivity (µS/cm) *
Total Phosphorus (mg/L)
Total Ammonia (mg/L)
SEDIMENT
95
Table 3.6. Ranked relative abundance, from lowest to highest, of brook trout and slimy
sculpin collected in 1974 and 2001. Blank cells indicate that fish were absent from that
site during the collection period. The shift in relative abundance from 1974 to 2001 is
indicated as an increase (+), a decrease (-), or no change (0).
Brook trout
Site
1974
2001
Slimy sculpin
Shift in relative
1974
2001
abundance
Shift in relative
abundance
F1
1
3
+
1
1
0
F2
3
1
-
3
3
0
F3
2
4
+
2
2
0
F5
4
2
-
4
4
0
A1
2.5
4
+
1
+
A2
1
1.5
+
A3
2.5
1.5
-
2
+
A4
4
6
+
3
+
A5
5
5
0
4
+
A6
7
7
0
1
A10
6
3
-
2
0
96
5
0
NEW
BRUNSWICK
Maine-USA
Fredericton
St. John River
F8
F4
F5
F2
F7
1
A5
Grand Falls
A8
A11
A2
A1
A7
Plaster Rock
Maine-USA
F3
F6
2
F1
A9
A6
A3
N
0
10
20
F9
Florenceville
30km
A4
A10
3
Woodstock
St. John River
Figure 3.1. Map of study sites located throughout western and northwestern New
Brunswick. Streams were located in areas dominated by forested (F1-F9) and
agricultural (A1-A11) land-use. Circles identify climate stations used for precipitation
data.
97
21
Agriculture
Forested
Mean daily temperature (°C)
19
17
15
13
11
9
7
4-Oct
27-Sep
20-Sep
13-Sep
6-Sep
30-Aug
23-Aug
16-Aug
9-Aug
2-Aug
26-Jul
19-Jul
12-Jul
5-Jul
5
Date
Figure 3.2. Mean (±SE) daily temperatures for agricultural (black line, n=11), and
forested (grey line, n=9) streams of northwestern New Brunswick, 5 July to 8 October
2001.
98
50
St. Leonard-1
40
30
20
10
0
Precipitation (mm)
50
Bon Accord-2
40
30
20
10
0
50
Woodstock-3
40
30
20
1-Oct
1-Sep
1-Aug
0
1-Jul
10
Figure 3.3. Daily rainfall (mm) from 1 July to 1 October, 2001 at weather stations in St.
Leonard (1), Bon Accord (2), and Woodstock (3) (Environment Canada 2001). The
numbers (1-3) represent the location of the weather station on Figure 3.1.
99
1200
Dry weight sediment (g)
1000
800
600
400
200
0
F1
F2
F3
F4
F5
F6
F7
F8
F9
A1
A2
A3
A4
A5
A6
A7
A8
A9 A10 A11
Site
Figure 3.4. Mean (±SE) dry weight of total sediment (g) deposited in samplers placed
in forested (grey bars) and agricultural (black bars) streams. Data from three samplers
collected at 1, 2, and 3 months were combined. Line at 200g indicates conservative
estimate of natural sediment deposition in the region based on the forested sites.
100
100
other>2000um
Percent of sediment grains
80
500-2000um
0-500um
60
40
20
0
F3
F8
A1
A2
A4
A5
A7
A8
A11
Site
Figure 3.5. Grain size separation of sediment deposited in solid-walled containers with
>200g dry sediment accumulation in forested and agricultural streams. Silt and very
fine to medium fine sand: 0-500µm; coarse to very coarse sand: 500µm-2mm; and
gravel: ≥2mm (Cummins 1962).
101
Figure 3.6. Median size of young-of-the-year (YOY) sculpin (A) and sculpin density (B)
versus maximum mean daily water temperatures, and median size of YOY sculpin
versus sculpin density (C). Triangles represent forested (white) and agricultural (black)
sites.
102
Median size of YOY sculpin (mm)
45
A
40
35
30
25
20
12
14
16
18
20
22
Maximum mean daily water temperature (°C)
24
Sculpin density (per m2)
2.5
B
2
1.5
1
0.5
0
Median size of YOY sculpin (mm)
10
12
14
16
18
20
22
24
Maximum mean daily water temperature (°C)
26
45
C
40
35
30
25
20
0
0.5
1
1.5
Sculpin density (per m2)
103
2
2.5
3.7
References
Allan JD. 1995. Stream ecology: Structure and function of running waters. Chapman
and Hall, London, UK.
Anderson CS. 1985. The structure of sculpin populations along a stream size gradient.
Environ. Biol. Fishes. 13:93-102.
Balon EK. 1975. Reproductive guilds of fishes: A proposal and definition. J. Fish. Res.
Board Can. 26:1429-1438.
Berkman HE, and Rabeni CF .1987. Effect of siltation on stream fish communities.
Environ. Biol. Fish. 18:285-294.
Brett JR. 1971. Energetic responses of salmon to temperature. A study of some
thermal relations in the physiology and freshwater ecology of sockeye salmon
(Oncorhynchus nerka). Amer. Zool. 11:99-113.
Cummins KW. 1962. An evaluation of some techniques for the collection and analysis
of benthic samples with special emphasis on lotic waters. J. North Amer. Benthol.
Soc. 8:331-341.
Curry RA, and MacNeill S. 2004. Early life history success of brook trout charr in
agricultural watersheds. J. North Amer. Benthol. Soc. In press.
Edwards P. 2001. An investigation of the potential effects of natural and anthropogenic
disturbance on the density and distribution of slimy sculpin (Cottus cognatus) in
Catamaran Brook, New Brunswick. MSc Thesis. University of New Brunswick,
Fredericton, NB, Canada.
Environment Canada. 2001b. Canadian daily climate data. CD-ROM.
Gray MA, Curry RA, and Munkittrick KR. 2002. Non-lethal sampling techniques for
assessing fish populations for environmental assessment. Water Qual. Res. J. Can.
37:195-211.
Kuehne RA. 1962. A classification of streams illustrated by fish distribution in an
eastern Kentucky creek. Ecology. 43:608-614.
Lisle TE, and Eads RE. 1991. Methods to measure sedimentation of spawning gravels.
Note PSW-411. Pacific Southwest Research Station, Forest Service, US
Department of Agriculture, Berkley, CA, USA.
Mebane C. 2001. Testing bioassessment metrics: macroinvertebrate, sculpin, and
salmonid responses to stream habitat, sediment, and metals. Environ. Mon. Assess.
67:293-322.
104
Munkittrick KR, McMaster ME, Van Der Kraak G, Portt C, Gibbons WN, Farwell A, and
Gray M. 2000. Development of methods for effects-driven cumulative effects
assessment using fish populations: Moose River project. Society of Environmental
Toxicology and Chemistry, Pensacola, FL, USA.
Mutch JP, Savard MA, Julien GRJ, MacLean B, Raymond B, and Doull J. 2002.
Pesticide monitoring and fish kill investigations on Prince Edward Island, 19941999. In Effects of land use practices on fish, shellfish, and their habitats on Prince
Edward Island. Cairns DK (ed). Can. Tech. Report Fish. Aquat. Sci. No. 2048. pp.
94-115.
NB DAFA. 2000. Annual report 1999-2000. New Brunswick Department of Agriculture,
Fisheries, and Aquaculture. Fredericton, NB, Canada.
Ricker WE. 1979. Growth rates and models. In Fish Physiology, Vol. 8. Hoar WS,
Randall DJ, and Brett JR (eds). Academic Press, New York, NY, USA. pp. 677-743.
Stauffer JC, Goldstein RM, and Newman RM. 2000. Relationship of wooded riparian
zones and runoff potential to fish community composition in agricultural streams.
Can. J. Fish. Aquat. Sci. 57:307-316.
Symons PEK, Metcalfe JL, and Harding GD. 1976. Upper lethal and preferred
temperatures of the slimy sculpin, Cottus cognatus. J. Fish. Res. Board Can.
33:180-183.
Otto RG, and Rice JO. 1977. Responses of a freshwater sculpin (Cottus cognatus
gracilis) to temperature. Trans. Am. Fish. Soc. 106:89-94.
Van Vliet WH. 1964. An ecological study of Cottus cognatus (Richardson) in Northern
Saskatchewan. MA Thesis, University of Saskatchewan. Saskatoon, SK, Canada.
Vannote RL, and Sweeney BW. 1980. Geographic analysis of thermal equilibria: A
conceptual model for evaluating the effect of natural and modified thermal regimes
on aquatic insect communities. Am. Naturalist. 115:667-695.
Waters TF. 1995. Sediment in streams: Sources, biological effects and control.
American Fisheries Society Monograph 7, Bethesda, MD, USA.
Welch HE, Symons PEK, and Narver DW. 1977. Some effects of potato farming and
forest clearcutting on small New Brunswick streams. Technical Report No. 745.
Fisheries and Marine Service, St. Andrews, NB, Canada.
Wismer DA, and Christie AE. 1987. Temperature relationships of Great Lakes Fishes:
A data compilation. Special Publication No. 87-3. Great Lakes Fishery Commission,
Ann Arbor, MI, USA.
105
CHAPTER 4
The use of stable isotope analysis to assess the site fidelity of slimy sculpin
(Cottus cognatus)3
4.1
Abstract
The primary concerns regarding sentinel species for assessing environmental
impacts are residency, abundance and suitability for measuring responses. These are
important if effects are to be attributable to local conditions. We have been using the
slimy sculpin (Cottus cognatus) to investigate the impacts of agricultural activity on fish
populations, and were interested in evaluating the use of the slimy sculpin as a sentinel
species for investigating site-specific environmental impacts. Stable isotope analysis
was used as a tool to investigate site fidelity and mobility in order to establish residency
and exposure for the sculpin. We predicted that sculpin collected from sites adjacent to
agricultural activity would show higher δ15N values than those collected from sites in
forested areas due to isotopic enrichment by fertilizers in the former. The predominant
use of chemical fertilizer applications resulted in no specific enrichment of δ15N in
sculpin collected in the agricultural region. We found an incremental enrichment in the
fish muscle tissue of approximately 5‰ in δ13C values in a downstream direction,
irrespective of surrounding land-use. A dual-isotope comparison was successful at
demonstrating site-specific isotopic signatures across sites along 30km of the river
system. The smallest distance between sites with a significantly different mean isotope
value was 360m. The site-specific signatures suggest that slimy sculpin are not moving
3
This chapter is currently submitted to the Canadian Journal of Fisheries and Aquatic
Sciences under joint authorship with Richard A. Cunjak and Kelly R. Munkittrick.
106
considerable distances among sites, and are incorporating an isotopic signature over a
narrow spatial scale.
4.2
Introduction
Aquatic environmental effects monitoring programs have generally favoured largebodied fishes for their programs because of ease of collection, tissue size requirements
for analysis, the value placed on species due to public or commercial importance, and
comparability with historical monitoring programs. Some of the limitations of largerbodied fish species include their potential for movement over large distances, and the
challenges with identifying the stressors associated with changes when there are
multiple discharges that occur in relatively close proximity. Recently there has been
increasing use of small-bodied fishes that may sometimes be more appropriate
monitoring species (Gibbons 1997, Fitzgerald et al. 1999, Galloway et al. 2003,
Tetreault et al. 2003), and the Canadian Environmental Effects Monitoring program has
seen a progressive increase in the use of forage fish species in fish surveys
(Munkittrick et al. 2002). Many small-bodied fish species have smaller home ranges
than traditional monitoring species and may overcome some of the challenges and
problems associated with the larger fish species. Two of the major issues associated
with choosing an appropriate monitoring species are establishing the ability to reflect
local conditions, and its exposure and residency within a zone of impact (Gibbons
1997; Munkittrick et al. 2000).
The slimy sculpin (Cottus cognatus) is a small-bodied, benthic fish species. The
slimy sculpin lacks a swim bladder and maintains a largely cryptic lifestyle beneath
cobble and other instream cover, precluding significant mobility. The male guards a
nest during spawning and rearing of the young (Van Vliet 1964), and mark-recapture
107
studies involving freshwater sculpin species have found that home ranges and
movements of recaptured fish were generally < 50 m (McLeave 1964; Hill and
Grossman 1987; Morgan and Ringler 1992). Tracking of individual slimy sculpin over a
period of 10 months with passive integrated transponder (PIT) tags indicated that 75%
of recaptured fish moved < 30m from release points (Chapter 5).
Site fidelity can be defined as the tendency of an animal to return to an area it
previously occupied or to stay within the same area for an extended period of time
(White and Garrot 1990). Stable isotopes offer a unique, and potentially less labour
intensive, way to look at site fidelity and residency by assuming that if an animal
resides in a particular area, it will be feeding in that area and thus acquire an isotopic
signature from the food sources from that area. The food sources in turn will obtain
their isotopic signatures from the local organic matter which will have specific 13C:12C
(δ13C) and 15N:14N (δ15N) ratios (Hershey and Peterson 1996). Thus, animals feeding in
a particular area will integrate a unique isotope composition as a function of the
localized biogeochemical processes (Fry et al. 1999). Fry et al. (1999) were able to
discriminate fish of the same species that had been feeding in offshore lotic habitats
versus inshore marsh areas in the same lake.
Harrington et al. (1998) used δ15N values to distinguish natal streams of Atlantic
salmon (Salmo salar) in agricultural catchments. They found a positive correlation
between the percentage of a catchment under agriculture and the stable nitrogen
isotope ratio in fish tissues. Hebert and Wassenar (2001) also found a strong positive
relationship between δ15N in juvenile mallard feathers and the percent of land in
agriculture. Due to intense agricultural cultivation in the lower part of the Little River
catchment (northwestern New Brunswick, Canada), we predicted that there would an
enrichment of δ15N in streams draining agricultural regions due to fertilizer application
108
and subsequent run-off. Based on the assumption of limited mobility for this fish
species, we hypothesized that fertilizer inputs into the stream would be reflected in the
δ15N signature in the sculpin and could be used to discriminate fish from different land-
use regions. The alternative hypothesis would be that if sculpin were moving
significantly throughout the system, they would have overlapping isotopic signatures
reflecting the incorporation of food sources from across a wider spatial and temporal
scale.
4.3
Methods
Study Site
As part of a larger study looking at the health of fish populations in agricultural
regions, a sculpin monitoring study (1999-2001) was designed in the Little River
catchment, north of Grand Falls, New Brunswick. The Little River is a 4thorder stream,
with a catchment area of approximately 340 km2, which originates in a forested
landscape and drains predominantly agricultural lands in its lower reaches. Sites
selected for inclusion in the stable isotope study were classified based on the dominant
surrounding land-cover that was forested (sites 1-5), a transition area between forestry
and agriculture (site 6), and the area where agriculture is the predominant land-use
(sites 7-10; Figure 4.1). Approximately 15% of the total catchment area is under potato
cultivation, with all agricultural activities located below site 5.
Fish and sample collections
Sculpin were collected using a backpack electrofishing unit (Smith-Root C-15). The
primary objective at each site was to collect 100 sculpin to assess site-specific
population distributions (Gray et al. 2002). All captured fish were measured for length
109
(±1mm) and weight (±0.01g) and the 30 largest sculpin were kept for lethal sampling
(Chapter 2). The remaining fish were returned to the site of collection after
measurements were taken. The collections for stable isotope analysis reported here
were made in November 1999 (10 sites), with additional sampling in March and
November 2000, and April 2001 (3-4 sites) to investigate the temporal stability of the
stable isotope signatures in the sculpin. The number of sites where fish were sampled
was reduced after November 1999 as sites were focused for the study of agricultural
impacts (Chapter 2).
Sculpin carcasses were frozen until a sample of epaxial muscle tissue
(approximately 40-400mg) was dissected ventral to the dorsal fin. Tissues samples
were air dried in glass scintillation vials at 50-60ºC for 48h, and then ground to a fine
powder using a mortar and pestle. Subsamples of 200µg were analyzed with a
continuous-flow isotope-ratio mass spectrometer (Finnigan Mat Delta Plus) equipped
with a ThermoQuest (NC2500) elemental analyzer. Isotope ratios in the samples were
determined using international reference materials; Pee-Dee Belemnite for 13C:12C
(δ13C), and atmospheric nitrogen for 15N:14N (δ15N). Results were expressed as parts
per thousand (‰) deviation from the reference materials.
Evaluation of the precision and variability of the machine over time with replicate
analyses of the same sample (n=35) resulted in an average δ13C value of –
27.82±0.13‰ (SD), and δ15N value of 9.04±0.16‰ (SD). The coefficient of variation for
standards run concurrently with the samples were always <2% for δ13C and <10% for
δ15N.
110
Data analysis
Data were assessed for normality by visual examination of normal probability plots,
and homogeneity of variance by visual examination of residual plots. Deviations from
normality were not strong and were considered negligible when using analysis of
covariance (ANCOVA), which is considered moderately robust to small deviations.
ANCOVA was used to assess isotopic differences between sites using fish length as a
covariate to account for variability due to fish size. Tukey’s test was used for post-hoc
analysis for detecting differences between sites (p<0.05). Statistical analyses were
completed using SYSTAT® (ver.10, SPSS Inc., Chicago).
4.4
Results
There was no differentiation of sculpin attributable to enrichment of δ15N in the
agricultural region (Figure 4.2-top). The three most upstream forested sites (sites 1-3)
were the most variable with mean δ15N values differing by approximately 2‰. Moving
downstream, adjacent sites were similar with mean δ15N values within 1‰. Despite
such a narrow range, there was a statistical difference among sites (ANCOVA; F=59.0,
df=9, p<0.0001). Sculpin at site 3 had significantly lower δ15N values than at any other
site along the river, most likely due to groundwater inputs from a tributary immediately
upstream of the collection site. There was slight enrichment of δ15N in the lowermost
agricultural sites (sites 8-10), but these were not significantly different than the
uppermost forested site (site 1).
Differences in δ13C among sites were more apparent, with increasing enrichment
downstream (Figure 4.2-bottom). δ13C values in sculpin muscle were significantly
enriched from approximately –33‰ in the forested region to –28‰ at the lowermost
111
agricultural site (ANCOVA; F=84.0, df=9, p<0.0001). The highest δ13C value was found
at the lowermost agricultural site (site 10), which was significantly higher than all other
sites.
Plotting the δ15N and δ13C data together provides evidence of site-specific isotopic
signatures that were not as clear when the isotope data were plotted separately (Figure
4.3). With the exception of a small switch with sites 4 and 5, and sites 8 and 9,
adjacent sites are more isotopically similar to one another than more geographically
distant sites along the river continuum.
Temporal analysis of isotope signatures at selected sites shows that carbon and
nitrogen isotopes did not vary within sites (Figure 4.4). Over a period of eighteen
months, average isotope values differed by approximately 0.3-0.9‰ for carbon, and
0.3-0.5‰ for nitrogen. In general, the spring samples were slightly depleted in carbon,
and slightly enriched in nitrogen.
4.5
Discussion
The δ15N isotopic composition of slimy sculpin collected from various sites along the
Little River differed significantly, but the difference was not related to any consistent
enrichment in the agricultural region as predicted. Sculpin tissues showed a
progressive enrichment in δ13C along the river gradient from the forested to the
agricultural region. Using a dual isotope approach, site-specific signatures were
evident with adjacent sites showing the most similarity, suggesting that sculpin were not
moving extensively across sites. For the purposes of this study, we define the
residency of the slimy sculpin as dwelling within the same ‘site’ over time. Because
sites ranged from 0.1-7.8km apart, the resolution of “site” was a minimum of 100m in
this study. The smallest distance over which there was a significant difference in
112
isotope signature detected was 360m. This and the nature of the site-specificity of the
carbon and nitrogen isotope signatures suggest that the sculpin are reflecting local
conditions, likely due to a high degree of residency.
Stable isotopes have been used extensively to investigate food web structure and
possible food sources, but have more recently gained attention as a tool for studying
movement and behaviour of animals. Hansson et al. (1997) used a point-source δ15N
input from a sewage treatment plant to assess movement patterns of fish in the Baltic
Sea. They found that more mobile species tended to integrate the characteristics of
many areas and thus became more isotopically homogeneous. Kline et al. (1998) used
natural stable carbon isotopes to indirectly determine migratory behaviour of
coregonids based on marine versus freshwater carbon signals. Harvey and Kitchell
(2000) investigated movement patterns of many fish species in western Lake Superior
using the adjacent cities of Duluth and Superior as a point source of anthropogenically
enriched δ15N. The slimy sculpin was the only fish species that reflected the δ15N
enrichment, with site as a significant factor for its isotopic signature, indicating its
sedentary nature. Conversely, rainbow smelt (Osmerus mordax) showed no
differences in δ15N values across all sites confirming its high vagility (Harvey and
Kitchell 2000). Farwell (University of Waterloo, unpub. data) looked at δ13C and δ15N
signatures in white sucker (Catostomus commersoni) at three sites in an undeveloped
northern Ontario river over a distance of 67km and found values differed by 0.6‰,
suggesting that the suckers were integrating their surroundings over a broad spatial
scale.
The prediction that δ15N values would be enriched in the agricultural region, due to
fertilizer inputs from adjacent potato fields, was based on previous stable isotope
research. Harrington et al. (1998) showed that δ15N in Atlantic salmon parr was
113
strongly related to the amount of agricultural land within a catchment. Hebert and
Wassenaar (2001) also showed that δ15N in feathers of juvenile mallard ducks were
highest in regions of high agricultural development. The inconsistent finding in our
study was likely due to the type of agricultural activity and the type of fertilizers applied
to the fields. Isotopic δ15N values of animal manure range from +10 to +20‰, whereas
synthetic commercial fertilizer has a range of 0 to +3‰ (Kendall 1998). The studies by
Harrington et al. (1998) and Hebert and Wassenaar (2001) were both carried out in
regions with livestock feedlots and manure-fertilized fields, suggesting those were the
sources of enriched δ15N. That the fields in the Little River catchment were almost
exclusively fertilized with synthetic commercial fertilizer blends (J-L Daigle, Eastern
Canadian Soil and Water Conservation Centre, pers. comm.), may explain the lack of
significant nitrogen enrichment downstream of agricultural inputs.
By contrast to δ15N, there was a progressive downstream enrichment of δ13C in
sculpin muscle. Adjacent sites were most similar with the largest difference seen
between the upper forested sites and the lowermost site in the agricultural region.
Doucett et al. (1996) observed similar longitudinal δ13C enrichment for algae and fish
among sites following the stream continuum in a forested central New Brunswick
stream. The authors proposed that the enrichment was due to a progressive
enrichment in dissolved inorganic carbon from δ13C-depleted groundwater sources in
the headwaters to incorporation of δ13C-enriched atmospheric CO2 in downstream
regions where algal-derived carbon sources predominated in the food chain.
Slimy sculpin isotope signatures appear to be stable over time, with site-specific
deviations less than 0.5 and 0.9‰ for δ15N and δ13C, respectively. Harrington et al.
(1998) observed a similar temporal stability in δ15N values of 0+ and 1+ salmon parr,
with a difference of between 0.4-0.9‰ over one year. Coupled with the narrow spatial
114
isotopic values at each site, the temporal stability provides additional support that the
fish are not moving significantly within the system and are resident within a small
spatial scale.
Variations in the carbon and nitrogen stable isotope constitution arise from food
sources and accumulate in tissues of consumers as they are passed up the food web
(Peterson and Fry 1987). In turn, there are many factors that can influence δ13C values
in food sources; from seasonal changes in primary production (Leggett et al. 1999) to
differences in water turbulence (France 1995), or influence δ15N from catchment
denitrification (Kendall 1998) to anthropogenic inputs. Numerous studies have shown
that a multiple isotope approach can be a powerful and informative way to study
differences between populations of animals (Harrington et al. 1998). Fry et al. (1999)
described trophic relationships in a food web in Lake Okeechobee, Florida, using δ15N
values, but in combination with δ13C, habitats with site-specific isotopic signatures were
differentiated over a wide variety of fish species. Wayland and Hobson (2001) used
stable nitrogen and sulphur isotopes to distinguish biota from upstream or downstream
of pulp mill and sewage effluents. Galloway et al. (2003) found very different δ13C and
δ15N signatures in slimy sculpin collected downstream of effluents from a paper mill and
a pulp mill but on opposite banks of the Saint John River. A distance of only 200m
between the two sites suggested that the sculpin were not moving across the river and,
instead, were likely reflecting the conditions of the habitat in which they were collected.
Isotopic differences tend to be more apparent when there is an external and unnatural
source of isotopes entering the system, and such is the basis for natural tracer studies
(Hershey and Peterson 1996).
115
4.5.1
Conclusion
Our primary objective was to use stable isotopes as a tool to evaluate the site
fidelity of slimy sculpin. With a relatively sedentary species like the sculpin, we were
able to show that it is not necessary to have an anthropogenic signal to observe
differences among sites on a relatively small spatial scale. There were no significant
differences in δ15N associated with land-use activity. This was likely due to the use of
commercial fertilizers with low δ15N values. Using a multiple-isotope approach we were
able to identify site-specific isotopic signatures showing low spatial and temporal
variation. The low variability and geographically distinct signatures suggest that the fish
were displaying a high degree of site fidelity and limited instream movement.
Therefore, we consider the slimy sculpin to be a suitable fish species for environmental
monitoring when residency and exposure are important factors.
4.6
Acknowledgements
This project was funded by the Toxic Substances Research Initiative (TSRI), Crop
Life Canada, the Canadian Water Network NCE and a NSERC Discovery Grant to
KRM. KRM and RAC are both supported under the Canada Research Chair program,
and MAG was supported by a post-graduate scholarship from the Natural Science and
Engineering Research Council (NSERC). The authors would like to acknowledge the
work done by Kyle Vodjani in helping with sample preparation, and additional technical
support from the Stable Isotopes in Nature (SIN) Lab at the University of New
Brunswick. Technical field help was provided by the New Brunswick Co-operative Fish
and Wildlife Research Unit (NBCFWRU), and graduate students from the Canadian
Rivers Institute.
116
Little River
Grand Falls
Flow
NEW BRUNSWICK
Little River
St. John River
Fredericton
Saint John
1
2
3
4
5
6
7
8
0km
9
5km
10
St. John River
Figure 4.1. Schematic map of the Little River, New Brunswick catchment indicating
sites where fish were collected for stable isotope analysis. Sites 1-5 were located
within the forested region, site 6 was within a transitional region influenced by both
forest and agricultural inputs, and sites 7-9 were within the agricultural region, and site
10 was in an urban area.
117
11
δ15N (‰)
10
A
BD
9
8
BD
BE
AEF EF
D
AF
BD
C
7
6
-26
E
δ13C (‰)
-27
-28
-29
BC
-30
D
8
9
C
B
-31
-32
BC
D
A
A
A
2
3
-33
-34
-35
1
4
5
6
7
10
Site
Figure 4.2. Mean (±95%CI) δ15N (top) and δ13C (bottom) isotope values (‰) in sculpin
muscle tissues. Site numbers progress downstream from the forested region (1-5), the
transition region (6), into the agricultural region (7-9), and at the urban site (10).
Statistical differences between sites are denoted with different letters (Tukey’s HSD
test, p<0.05).
118
10
9.5
δ15N (‰)
1
9
9
5
2
8.5
4 6
8
10
7
8
7.5
3
7
-34
-33
-32
-31
-30
δ13C
-29
-28
-27
(‰)
Figure 4.3. Mean (±95%CI) δ15N and δ13C plotted together. Site numbers progress
downstream from the forested region (1-5), the transition region (6), into the agricultural
region (7-9), and at the urban site (10). Site-specific isotopic signatures are more
evident and overlap generally only occurs between adjacent sites.
119
10
9
1
9.5
δ15N (‰)
7
9
8.5
8
3
7.5
7
6.5
6
-35
-34
-33
-32
-31
-30
-29
-28
δ13C (‰)
Figure 4.4. Temporal variation in mean (±95%CI) δ15N and δ13C isotope composition of
sculpin collected at particular sites (1, 3, 7, and 9) on the Little River; November 1999 –
black triangles, March 2000 - white diamonds, November 2000 – black circles, and
April 2001 – white squares.
120
-27
4.7
References
Doucett RR, Power G, Barton DR, Drimmie RJ, and Cunjak RA. 1996. Stable isotope
analysis of nutrient pathways leading to Atlantic salmon. Can. J. Fish. Aquat. Sci.
53: 2058-2066.
Fitzgerald DG, Lanno RP, and Dixon DG. 1999. A comparison of a sentinel species
evaluation using creek chub (Semotilus atromaculatus) to a fish community
evaluation for the initial identification of environmental stressors in small streams.
Ecotoxicol. 8:33-48.
France RL. 1995. Critical examination of stable isotope analysis as a means for tracing
carbon pathways in stream ecosystems. Can. J. Fish. Aquat. Sci. 52:651-656.
Fry B, Mumford PL, Tam F, Fox DD, Warren GL, Havens KE, and Steinman AD. 1999.
Trophic position and individual feeding histories of fish from Lake Okeechobee,
Florida. Can. J. Fish. Aquat. Sci. 56:590-600.
Galloway BJ, Munkittrick KR, Currie S, Gray MA, Curry RA, and Wood C. Examination
of the responses of slimy sculpin (Cottus cognatus) and white sucker (Catostomus
commersoni) collected on the Saint John River downstream of pulp mill, paper mill,
and sewage discharges. Environ. Toxicol. Chem. In press.
Gibbons WN. 1997. Suitability of small fish species for monitoring the effects of pulp
mill effluent on fish populations. PhD thesis. University of Waterloo. Waterloo, ON,
Canada.
Gray MA, Curry RA, and Munkittrick KR. 2002. Non-lethal sampling methods for
assessing environmental impacts using a small-bodied sentinel fish species. Water
Qual. Res. J. Can. 37:195-211.
Hansson S, Hobbie JE, Elmgren R, Larsson U, Fry B, and Johansson S. 1997. The
stable isotope ratio as a marker of food web interactions and fish migration.
Ecology. 78:2249:2257.
Harrington R, Kennedy BP, Chamberlain CP, Blum JD, and Folt CL. 1998. 15N
enrichment in agricultural catchments: field patterns and applications for tracking
Atlantic salmon (Salmo salar). Chem. Geol. 147:281-294.
Harvey CJ, and Kitchell JF. 2000. A stable isotope evaluation of the structure and
spatial heterogeneity of a Lake Superior food web. Can. J. Fish. Aquat. Sci.
57:1395-1403.
Hebert CE, and Wassenar LI. 2001. Stable nitrogen isotopes in waterfowl feathers
reflect agricultural land use in Western Canada. Environ. Sci. Technol. 35:34823487.
Hershey AE, and Peterson BJ. 1996. Stream food webs. In Methods in Stream
Ecology. Hauer FR, and Lamberti GA (eds). Academic Press, San Diego, CA, USA.
pp. 511-530.
121
Hill J, and Grossman GD. 1987. Home range estimates for three North American
stream fishes. Copeia. 1987:376-380.
Kendall C. 1998. Tracing nitrogen sources and cycling in catchments. In Isotope tracers
in catchment hydrology. Kendall C, and McDonnell JJ. (eds). Elsevier, New York,
NY, USA. pp 519-576.
Kline TC Jr, Wilson WJ, and Goering JJ. 1998. Natural isotope indicators of fish
migration at Prudhoe Bay, Alaska. Can. J. Fish. Aquat. Sci. 55: 1494-1502.
Leggett MF, Servos MR, Hesslein RH, Johannsson O, Millard ES, and Dixon DG. 1999.
Biogeochemical influences on the carbon isotopic signatures of Lake Ontario biota.
Can. J. Fish. Aquat. Sci. 56: 2211-2218.
McCleave JD. 1964. Movement and population of the mottled sculpin (Cottus bairdi
Girard) in a small Montana stream. Copeia. 1964:506-512.
Morgan CR, and Ringler NH. 1992. Experimental manipulation of sculpin (Cottus
cognatus) populations in a small stream. J. Freshw. Ecol. 7:227-232.
Munkittrick KR, McMaster M, Van Der Kraak G, Portt C, Gibbons W, Farwell A, and
Gray M. 2000. Development of methods for effects-based cumulative effects
assessment using fish populations: Moose River project. SETAC Press, Pensacola,
FL, USA.
Munkittrick KR, McGeachy SA, McMaster ME, and Courtenay SC. 2002. Overview of
freshwater fish studies from the pulp and paper environmental effects monitoring
program. Water Qual. Res. J. Can. 37:49-77.
Peterson BJ, and Fry B. 1987. Stable isotopes in ecosystem studies. Annu. Rev. Ecol.
Syst. 18: 293-320.
Tetreault GR, McMaster ME, Dixon DG, and Parrott JL. 2003. Using reproductive
endpoints in small forage fish species to evaluate the effects of Athabasca Oil
Sands activities. Environ. Toxicol. Chem. 22:2275-2282.
Wayland M, and Hobson KA. 2001. Stable carbon, nitrogen, and sulfur isotope ratios in
riparian food webs on rivers receiving sewage and pulp-mill effluents. Can. J. Zool.
79:5-15.
White GC, and Garrot RA. 1990. Analysis of wildlife radio-tracking data. Academic
Press, San Diego, CA, USA.
Van Vliet WH. 1964. An ecological study of Cottus cognatus (Richardson) in Northern
Saskatchewan. MA Thesis, University of Saskatchewan. Saskatoon, SK, Canada.
122
CHAPTER 5
Measuring small-scale movements of the slimy sculpin (Cottus cognatus) to
assess site fidelity in a small river4
5.1
Abstract
Passive integrated transponder (PIT) tags were used to assess mobility of slimy
sculpin (Cottus cognatus), a small benthic fish of interest as a sentinel in environmental
monitoring. The objectives of the study were to assess mobility of sculpin in order to
address the issue of residency of this fish species, and to determine the feasibility of
PIT tag technology for monitoring slimy sculpin in natural systems. Preliminary lab
studies showed that tag loss and mortality were rare after tagging. Sculpin were
tagged in the field on two occasions; July 2001 (Group I) and October 2001 (Group II)
and monitored through to May 2002. Tracking events corresponded to 11, 48, 93, 129,
and 306d post-release for Group I (n=57), and 36 and 212d post-release for Group II
(n=55). Of 112 tagged fish, a total of 41 (37%) were recaptured. Movements ranged
from 0.5-101m within the study area (140m in length), with an overall median
displacement of 11.3m for 56 individual movements. Using encounter information from
tagged sculpin, a model was produced which estimated apparent survival of 78-99%
and a constant recapture rate of 27% during the study. Our results are consistent with
the hypothesis of limited mobility for slimy sculpin, and provide evidence of both spatial
and temporal residency. PIT tags proved to be a valuable marking technique that
allowed us to gather detailed information on survival and movement of individual
sculpin over time.
4
This chapter is currently submitted to the journal Copeia under joint authorship with
Richard A. Cunjak and Kelly R. Munkittrick.
123
5.2
Introduction
Establishing residency in the area of possible exposure is a significant challenge
when using fish in environmental monitoring (Fish Survey Expert Working Group 1997).
The mobility of the species in question will affect the facility of using reference sites
adjacent to or immediately upstream of the impact area when no natural barriers are
present. The use of a fish with both high spatial and temporal residency for monitoring
ensures the highest degree of confidence that fish were exposed to conditions at the
site under study. Seasonal mobility (e.g. migrations, spawning movements) will greatly
influence the suitability of using a particular fish species collected at both exposed and
unexposed sites.
Many small-bodied fish exhibit characteristics that overcome some problems
encountered when using large-bodied fish for environmental assessments (Munkittrick
et al. 2002). Small-bodied fish are generally found at high enough densities for
collection of sufficient samples, their reduced mobility aids in establishing exposure,
short life-spans relate to more rapid responses to environmental changes, confounding
factors such as fishing pressure are usually absent, and life history characteristics can
still be easily measured (e.g. age distribution, energy storage and expenditure) (Fish
Survey Expert Working Group 1997; Gibbons et al. 1998; Gray et al. 2002).
We are currently studying non-point source impacts of agricultural activities on
populations of the slimy sculpin (Cottus cognatus) (Gray et al. 2002). As with point
sources of environmental stressors, we must establish that the sculpin is a suitable
study species and will reflect the environmental conditions of the site from which it was
collected. The primary objective of this study was to assess the mobility of the slimy
sculpin in a natural system in order to assess residency of this fish species.
Secondarily, we wanted to evaluate the feasibility and utility of PIT (passive integrated
124
transponder) technology for tagging slimy sculpin. PIT tags are radio frequency
identification (RFID) products that have been used in a variety of research programs
with fish and wildlife where individual identification is desired. Commercially available
glass encapsulated tags range in size from 11 to 50mm, and can relay a unique signal
code after induction by an electromagnetic field sent from the detector, with no need for
a battery within the tag itself. PIT tags have been used extensively in fish culture to
evaluate individual performance (e.g. Baras et al. 2000), and to monitor movements
through fish passageways (e.g. Prentice et al. 1990a; Castros-Santos et al. 1996), and
in the wild (e.g. Prentice et al. 1990b; Roussel et al. 2000).
5. 3
Methods
Tags used in this study were 11.5mm long x 2mm diameter, glass encapsulated
passive integrated transponders (Destron Fearing Corp., South St. Paul, MN), with a
frequency of 125 kHz for Group I fish and 134.2 kHz for Group II. Separate PIT tag
readers were used to detect each frequency. There were two phases of the PIT tag
study; Phase I involved sculpin tagged in the laboratory to develop implantation
methods and assess the probability of tag loss and mortality, and Phase II involved a
field study of sculpin tagged in the Kennebecasis River, New Brunswick to study
movements and survival of tagged fish.
Phase I Lab Study
Eight sculpin (63-80mm total length (TL)) were captured in the wild and acclimated
for one week in fibreglass tanks at the University of New Brunswick. The tagging
procedure involved first anaesthetizing sculpin with clove oil. Clove oil was dissolved in
ethanol (1:10), and then added to water to yield a 30-35µL clove oil/L water solution
125
(Peake 1999). Fish were placed in the anaesthetic solution for roughly 10min or until
they were completely inactive. A 3-4mm incision was made along the mid-ventral line
anterior to the urogenital papilla and the tag was carefully inserted into the body cavity.
A small amount of Polysporin® antibacterial ointment and VetBond® glue were applied
to the incision to prevent infection and to seal the incision. Sutures were not used
because the sculpin abdominal wall is very thin.
After one month of observation, tag loss and mortality were both 0%. The incision
marks were visible but faint and there was no sign of infection. Although sculpin as
small as 63mm were surgically implanted with 11mm tags, there was moderate
resistance as the tag was inserted. After observing the tagged sculpin in the lab and
due to the limited space within the body cavity of the sculpin, the minimum tagging size
was set at 65mm.
Phase II Field study
Sculpin were tagged on two occasions for the field study; 27 July (Group I) and 29
October (Group II) 2001. All sculpin were collected using a backpack electrofishing unit
(Smith-Root Model C-15) in the Kennebecasis River, southern New Brunswick (45° 49’
37”N, 65° 13’ 9”W). The study area was a 3rd-order stream ranging from 5 to 12m wide
with a mean depth of 20cm. Riffle and run habitats each represented approximately
50% of the study area (Figure 5.1). Gowan et al. (1994) recommended that fish
mobility studies use multiple sections in order to cover larger spatial scales. Therefore,
our initial study design incorporated five 10m long treatment sections within a longer
140m reach of river (A-E; Figure 5.1). Treatment sections were separated from each
other by 15m long “buffer zones” (Figure 5.1), in which no fish were disturbed on the
tagging date. Small flags were placed at 1-metre intervals on both sides of the
126
riverbank for the entire 140m study area to aid in identifying the position of individual
fish within the stream. Sculpin were initially removed from treatment sections using one
pass with the electrofisher through each section. All sculpin >65mm TL were weighed
(±0.01g) and tagged (Table 5.1).
The tagging procedure was identical to the lab study. Tags were surgically
implanted and fish were allowed to recover in an aerated bucket for at least 30min
before being placed in flow-through live boxes for a minimum of 24h to monitor postsurgery effects. After 24h, the incision of each fish was examined, and fish were
subsequently released at the mid-point of the treatment sections where they were
captured. There were no mortalities of tagged fish observed after 24h.
A second set of sculpin, Group II, was tagged on October 29, 2001 to increase the
sample size of tagged fish, and to monitor survival and movements over winter. These
sculpin were collected during the third tracking event for Group I fish (93d postrelease), when any sculpin >65mm TL in the lowest three buffer zones (Buffers 1-3)
were tagged and released in the same areas (Table 5.1). Group II were captured for
tagging in only the lower three buffer zones due to the observation of low recapture of
Group I tagged sculpin in the riffle habitats. The same tagging procedures were
employed as for Group I fish. There were no 24h delayed mortalities, and fish were
released at the mid-point of their respective collection area.
Fish movement and recapture
Attempts to relocate Group I fish were made on 5 separate occasions (11, 48, 93,
129, and 306d post-release). Group II fish were monitored only for the last two tracking
events, which corresponded to 35 and 212d post-release. The entire study area
(treatment sections and buffer zones) was electrofished with one complete pass in the
127
upstream direction. All sculpin >65mm TL were collected, scanned using a handheld
PIT tag reader, and examined for an incision scar. There was only one case (306d
post-surgery) where a fish was scanned and a tag detected without seeing some visual
evidence of the incision. There was never a case where a fish with an incision mark
had no detectable tag.
When a tagged sculpin was captured, the tag number and location of capture was
noted by visual reference to flags located at each metre along the stream bank, and
then by estimating position within 5 strata subdividing the stream width (Figure 5.1).
Given the design of the study area, the minimum and maximum measurable distance of
movement was 0.5m and 120m for Group I fish, and 0.5m and 130m for Group II fish.
For the first tracking event of Group I fish (11d post-release), all tagged sculpin
were identified, then immediately released at their site of recapture. For all subsequent
tracking events, the recaptured tagged sculpin were kept in live boxes outside of the
study area until they were measured for length and weight. Fish were then released
within 20m from where they were captured and the release point noted. At 93d and
129d (35d; Group II) post-release, an additional 15m of river above and below the study
area were electrofished. At 306d post-release (212d; Group II), an additional 15m
below the entire study area was electrofished, and weather conditions did not permit
electrofishing in the final section (Buffer 6) and further upstream. One fish was
captured 10m downstream of Buffer 1 212d post-release (a Group II sculpin), the only
case where a tagged fish was found outside of the study area.
The term ‘movement’ within this paper refers to the spatial displacement between
the last known release point and the recapture point. Distance between release and
recapture points was adjusted using Pythagoras’ theorem to incorporate lateral (i.e.
across the flow) movements within the study grid.
128
Fish encounter data
The program MARK (v. 3.1; G. White, Colorado State University) was used to
develop a model to estimate apparent survival and recapture rates based on the
encounter histories of tagged fish using the standard Cormack-Jolly-Seber (CJS) model
for live recaptures (White and Burnham 1999). Apparent survival is defined as the
probability that a fish is alive and remains in the study area, and is thus available for
recapture (White and Burnham 1999). This means that although the CJS model does
not make a distinction between a fish that has died and a fish that has emigrated out of
the study area, a recapture at a later date means that a fish was alive at least until the
last recapture and is therefore included in the apparent survival estimate. Models were
developed containing constant or variable survival and recapture probabilities over
time, for both groups together or separately (model notation: constant (.), variable over
time (t), and by groups (g)). The unequal intervals between sampling events were
incorporated into the models using log-transformed values of the time interval (i.e. log(#
days between sampling events)).
Statistical analysis and modelling
Fish size and distance data were not normally distributed (expected values
assessed visually) and were tested using the non-parametric Kruskall-Wallis test or
Mann Whitney U-test where appropriate. Recaptures were evaluated among study
sections using the Chi-square test. Statistical analyses were performed using Systat®
(ver.10, SPSS Inc., Chicago).
Model selection was completed using Akaike’s Information Criterion (AIC) to select
the best candidate models (Burnham and Anderson 2002). AIC model selection is
based on an information-theoretic approach, and can be more appropriate than
129
traditional hypothesis testing for observational data by allowing for the comparison of
more than two models at once while balancing precision and bias (parsimony)
(Burnham and Anderson 2002). The models were adjusted for overdispersion (Q) and
small sample size (c), and the models were ranked by their QAICc weights to indicate
which models were given the most support. The best model was then used in a
bootstrap analysis that simulated encounter histories to produce maximum likelihood
estimates with confidence intervals for apparent survival and recapture rates (100
iterations).
5.4
Results
There were no significant differences in the lengths or weights of fish tagged in
different sections for Group I (Kruskall- Wallis (KW) test stat=2.30 and 5.55, df=4, and
p=0.68 and 0.24, respectively) or Group II (KW test stat=2.35 and 2.14, df=2, and
p=0.31 and 0.34, respectively) (Table 5.1).
Fish movement and recapture
Of 37 movements detected for Group I fish, two were lateral movements, 24
movements were upstream and 11 downstream (Table 5.1). The median movements
were 16.8m upstream and 6.8m downstream, but there was no statistical difference
due to the variability within the data (U=95, df=1, p=0.19). The frequency of tagged fish
recaptured from Group I appeared higher at the two boundaries of the tagging sections
within the study site (Figure 5.2). Over the total study period of 10 months (306d), the
25 recaptured Group I fish moved a minimum distance of 0.5m, and a maximum
distance of 101m (Figure 5.3). Seventy-five percent of recaptured fish moved less than
38m.
130
Of 19 movements detected for Group II fish, 13 were upstream and 6 were
downstream (Table 5.1), with no statistical difference in the magnitude of the directed
movements (U=34, df=1, p=0.50). The median movements were 21.2m upstream and
11.7 downstream. The recapture locations were also bimodally distributed within the
tagging sections for Group II sculpin (Buffer 1-3; Figure 5.2). Over a period of 7 months
(212d), the 16 recaptured Group II fish moved a minimum distance of 2.5m, and a
maximum distance of 55.3m (Figure 5.3). Seventy-five percent of recaptured fish
moved less than 24m.
There was no statistical difference in degree of movement between the two groups
of tagged sculpin (U=323, df=1, p=0.86). When all data are combined for recaptured
Group I and II fish, the median absolute distance moved from release point for all 56
movements was 11.3m.
For Group I fish, 25 of 57 tagged sculpin were recaptured (44%); 13 were
recaptured once, 10 were recaptured twice, and 2 were recaptured three times during
five tracking events (11-306d post-release). A significantly higher proportion of fish
from Section A and B than in other sections were recaptured (χ2=12.78, df=4, p=0.012)
(Table 5.1). Group II fish were tracked on only two occasions, 35d post and 212d postrelease. Of the 55 fish originally tagged, 16 fish were recaptured (29%); 13 were
recaptured once, and 3 fish were captured twice.
Fish encounter data
Six models were developed that produced plausible variations of group- and timedependence of the two parameters (apparent survival and recapture probability), and
incorporated the unequal sampling intervals (Table 5.2). Generally, models with
∆QAICc values ≤ 2 have ‘substantial’ support, 4-7 have ‘moderate’ support, and 10
131
have essentially no support (Burnham and Anderson 2002). The strongest model
(QAICc weight = 0.51; all weights sum to 1) supported survival being a function of the
time interval between sampling events, a constant recapture probability of tagged
sculpin of 0.27 over the entire study, and no group-dependence of the estimates (Table
5.3). There was substantial variability associated with the estimates for survival and
recapture probabilities, likely due to the small sample sizes, and so the rates provided
serve more to illustrate what the model is capable of estimating than as absolute
values. The next two models also had moderate support with one model supporting a
difference in survival for group I and II sculpin, the other supporting a constant survival
over the 10 months, and both kept recapture probability constant over time. The model
likelihood for the best model, however, is more than double that of the second and
third-ranked models.
5.5
Discussion
Adult slimy sculpin tagged with passive integrated transponder (PIT) tags showed
limited movements in the Kennebecasis River. Of the 112 sculpin tagged on two
separate occasions, 41 individuals (37%) were recaptured over a total of five tracking
events. For 56 separate movements of tagged sculpin, the overall median distance
travelled was 11.3m. Although there was moderate to high variability in the degree of
individual movements (0.5-101m), results from this study provide evidence that slimy
sculpin display a relatively high degree of both spatial and temporal residency. There
were minor differences between the movement patterns of slimy sculpin tagged in the
summer (Group I) and autumn (Group II) and these were likely due to the difference in
the number of tracking events for each group. The apparent concentration of recapture
locations at upper and lower ends of the study area may have been habitat-related
132
because release locations were generally distributed throughout the study site. Dry
conditions in the summer resulted in very low water within the riffle habitats (middle
section of study area). This may have caused fish to move into the deeper run habitats
where more tagged fish were recaptured.
Previous research on freshwater sculpin movements has indicated that cottids are
not likely to move long distances. Using a variety of marking and collecting techniques,
most slimy sculpin, and the closely related mottled sculpin (Cottus bairdi), were found
to stay within stream areas ranging from 1-50m (McLeave 1964; Hill and Grossman
1987; Morgan and Ringler 1992). In our study of individually tagged sculpin, it was
interesting to note that the single sculpin from Group I responsible for the longest
downstream movement (101m) was also responsible for the longest upstream
movement (99.5m) resulting in an absolute displacement of only 0.5m from its original
release point 10 months earlier. More than half of the fifteen PIT-tagged sculpin that
were recaptured more than once were found less than 5m from their previous recapture
points.
The probability of capturing an individual sculpin depends on many factors
including, but not exclusive to, the technique, experience of the collectors, substrate
type and distribution, water velocity, and water clarity. Although in many habitats
electrofishing is a very effective way to collect live sculpin, it has its drawbacks.
Because sculpin lack a swim bladder and choose cryptic habitats, sculpin may be
shocked beneath rocks but not be displaced for collection (Morgan and Ringler 1992).
From previous electrofishing experiments, Gray (unpub. data) found that the
percentage of the sculpin population captured in successive electrofishing passes was
similar in several rivers with different habitat types and sizes of study areas
(average=0.35 for single pass). Clément (1998) conducted similar studies in small
streams in New Brunswick and found an average sculpin capture probability of 0.31 on
133
the first pass. Utzinger et al. (1998) reported one-pass capture probabilities for a
similar species, the European bullhead (Cottus gobio), ranging from 0.17 to 0.34. The
recapture probability we estimated by the best model was 0.27 (0.17-0.41; 95%CI), and
although accompanied by a wide range of error, it falls within what researchers have
found independently.
One of the main problems to overcome with mark-recapture studies is accounting
for those fish not recaptured after the initial release. There are four possible
explanations: (a) the fish died, (b) it moved out of the area (i.e. emigrated from the
study area), (c) it remained in the study site but was not captured during that particular
tracking event, (d) or the tag was lost. The Cormack-Jolly-Seber (CJS) model was
unable to distinguish between mortality and emigration and combines the two factors
into the estimate for apparent survival between tracking events (i.e. apparent survival =
1-apparent mortality). Based on the encounter histories of the PIT tagged sculpin,
apparent survival ranged between 0.78 and 0.98 (0.46-1.00; 95% CI, all estimates
combined). Again, the uncertainty was high but suggests that the combined mortality
and emigration for sculpin during the 10-month study period was relatively low.
Over a 7-month period, McLeave (1964) estimated an emigration rate of 12% for
marked mottled sculpin using movement data, but did not estimate mortality due to the
associated difficulty. Studies by Van Vliet (1964) and Morgan and Ringler (1992)
suggest that most adult sculpin mortality occurs during the late summer and fall. The
apparent survival estimates produced from our model support this position. However,
there remains the issue of determining a reasonable estimate of mortality, without
which there is no way to decouple emigration from mortality. Estimates of apparent
mortality for adult sculpin within the literature are substantially higher than our
estimates (52-92% 1 year, Smith 1972; 43% June-September, Morgan and Ringler
1992). Possibly the sculpin in the Kennebecasis River are experiencing a higher
134
survival or lower emigration rates than populations studied by other researchers. The
highest apparent survival rate estimated from our model coincided with the longest
sampling interval, during which one could expect more opportunities for dispersal out of
the study area or natural mortality of sculpin. This may merely be a function of the
diminishing power of the model as the sample size of recaptured fish decreased over
time. The survival estimates conformed more closely to what was observed in the field
over the first three sampling events, suggesting that the contribution of the second
group of tagged fish with only 2 post-release sampling events did not significantly help
the model.
One of the main strengths of using a CJS approach to assess mark-recapture data
is that it deals with the third possibility of undetected individuals by allowing for tagged
individuals to be considered alive and available for recapture if they are recaptured
during a subsequent tracking event. The recapture probability calculated by modelling
encounter histories, despite the range of uncertainty experienced here, is then more
appropriate and informative than calculating absolute recapture success (#tagged fish
recaptured / #originally tagged). Finally, from our lab study, we concluded that tag loss
was not a significant factor in accounting for unobserved fish.
Gowan et al. (1994) reviewed numerous studies that investigated stream salmonid
movements and proposed that the paradigm related to restricted fish movement may
be an artifact of the study designs employed. Although there are admittedly similar
biases in the current study, attempts to reduce overall bias were made by increasing
the number of smaller sites within a larger area and through analyzing encounter
histories. The slimy sculpin, by its benthic nature, was not expected to make long
movements like those of semi-pelagic and pelagic fishes. This does not negate the
possibility of long-range movements, but the area encompassing possible movements
of sculpin was assumed to be relatively small.
135
The secondary objective of this study was to assess the feasibility and utility of PIT
tagging slimy sculpin. Relative to other marking techniques available for sculpin,
advantages of the PIT tag include individual identification, easy-to-read tags, low
mortality from tagging, and with no internal battery and high retention, an increased
duration of a tracking study. In the past we have attempted other methods for marking
sculpin and found difficulties that were overcome by PIT tagging: (1) panjet dye
marking – mark not always retained and visible only short-term, (2) otolith dyeing –
mark detection requires lethal sampling, (3) fin-clipping – increased 24h delayed
mortality and morbidity observed. Bruyndoncx et al. (2002) evaluated the use of PIT
tags and visible implant elastomer (VIE) to mark European bullhead. They found
similar survival rates and no significant differences between recapture rates of fish
marked with either technique. The disadvantage of PIT tagging is the restriction of
marking sculpin >65mm, thus introducing an age- and size-bias to the study. Another
disadvantage with PIT tagging small-bodied fish like the slimy sculpin is the reduced
detection range because of the small tags, and the need for the fish to be repeatedly
electrofished for detection. Roussel et al. (2000) developed a portable, passive
tracking system for 23mm PIT tags in Atlantic salmon (Salmo salar) parr that could
detect tags 70-100cm from the antenna. A passive detection method is currently under
development for 11mm tags that may enable detection of tagged sculpin within 1520cm (Cunjak, UNB, unpub). Such an assessment of movements would minimize
disturbance to the fish and would give PIT tags a major advantage over other marking
techniques.
5.5.1
Conclusions
In summary, recaptured sculpin provided information of direction and distances
136
travelled from release points of individual fish over a period of ten months. Movements
of individual fish were variable, but with 75% of recaptured fish moving less than 30m
over a period of ten months, tagged slimy sculpin displayed a moderate to high degree
of site fidelity. Although overall recapture of tagged sculpin was 37%, analysis of
encounter histories provided mean apparent survival estimates ranging between 78
and 98%. Though there is currently no strict definition for describing a sedentary
versus mobile fish species (i.e. degree of emigration, distance travelled, etc), the
results from this study provide support for the hypothesis of limited mobility and high
residency rates for the slimy sculpin. Thus, the slimy sculpin appears to be an
appropriate candidate for environmental monitoring studies when residency is an
important factor. Increasing the number of tagged individuals and post-release
sampling events should be incorporated into future sculpin movement studies to
improve the ability of the model to estimate survival and recapture rates. Further PIT
tag research is already underway to assess fundamental questions related to the
factors that influence displacement behaviour of sculpin.
5.6
Acknowledgements
This project was funded by the Toxic Substances Research Initiative (TSRI), Crop
Life Canada, and a Natural Sciences and Engineering Research Council (NSERC)
Operating Grant awarded to KR Munkittrick. MA Gray was funded by a post-graduate
scholarship from NSERC, and KR Munkittrick and RA Cunjak are both supported by the
Canada Research Chairs program. Fish encounter data was analyzed using the
program MARK by AR Breton. Technical field help was provided by the Canadian
Rivers Institute (CRI). Sculpin were kept under laboratory conditions with the UNB
137
Animal Care Protocol #00019, and collected in the field under the Department of
Fisheries and Oceans Collection Permit #2001-431.
138
Jul 2001
Oct 2001
Group II
Group I
Table 5.1. Median length (min-max) for sculpin collected and PIT tagged in the Kennebecasis River (Group I A-E, Group II
Buffer 1-3). Number of sculpin recaptured from each section and mean ±SE distances [number of movements] are given for
upstream and downstream movements.
Number of
Median fish
Number
Upstream
Downstream
Section
sculpin
length (mm)
recaptured
movements (m)
movements (m)
A
9
67 (65-84)
7
27.4 ± 19 [5]1
31.0 ± 23 [4]
B
11
70 (65-85)
8
32.7 ± 11 [9]
7.3 ± 1 [4]
C
11
73 (65-84)
2
43.8 ± 14 [3]
[0]
D
11
68 (65-77)
3
30.5 ± 7 [4]
[0]
E
15
72 (65-83)
5
6.2 ± 2 [3]
14.8 ± 11 [3]
Buffer 1
13
77 (67-87)
2
33.2 ± 10 [3]
55.3 [1]
Buffer 2
27
72 (65-89)
7
13.1 ± 5 [4]
11.2 ± 4 [4]
Buffer 3
15
73 (65-91)
7
16.8 ± 3 [6]
11.2 [1]
1 - there were 2 lateral movements (2m and 4m)
139
Table 5.2. Full model set was assessed using Akaike’s Information Criterion (AIC).
Survival, φ, and recapture probability, p, were kept constant (.), varied by group (g),
time (t), or group and time (g+t). Lowest QAICc values are given to the most
parsimonious model, ∆QAICc = model QAICc - best model QAICc, and QAICc weight
indicates the degree of support given to the model.
Model
QAICc
∆QAICc
QAICc weight
φ (log interval) p (.)
216.4
0.00
0.506
φ (g) p(.)
218.2
1.87
0.199
φ (.) p (.)
218.4
1.99
0.187
φ (log interval) p (t)
220.7
4.29
0.059
φ (t) p (.)
221.6
5.28
0.036
φ (g+t) p(.)
223.7
7.31
0.013
140
Table 5.3. Maximum likelihood estimates, with standard error (SE) and 95%
confidence limits, for apparent survival (φ) and recapture probabilities (p) obtained
using the program MARK for the best model (see Table 5.2). TI-TV represent the
tracking events.
Parameter
Estimate
SE
95% CI
Recapture probability (p)
0.27
0.06
0.17-0.40
TI (August)
0.82
0.11
0.51-0.95
TII (September)
0.91
0.03
0.84-0.96
TIII (October)
0.93
0.02
0.86-0.96
TIV (December)
0.91
0.03
0.84-0.95
TV (May)
0.99
0.06
0.81-1.00
Apparent survival (φ)
141
140m
D
E
Buffer 6
C
RUN
Buffer 5
10m
B
Buffer 3
15m
Buffer 2
Buffer 1
A
Buffer 4
RIFFLE
RUN
FLOW
1
2
3
FLOW
4
5
10m
Figure 5.1. Schematic of the study area on the Kennebecasis River. Fish were tagged
and released in treatment sections A-E. Upon recapture, fish location was determined
within 0.5m longitudinally and 1-2m laterally using flags placed along the riverbank and
by dividing the river width into 5 equal strata. Generalized habitat types (run versus
riffle habitats) are indicated.
142
9
Group I
Group II
Number of sculpin
8
7
6
5
4
3
2
1
0
B1
A
B2
B
B3
C
B4
Study section
D
B5
E
B6
Figure 5.2. Frequency of recapture locations throughout the entire study (by section) of
Group I (black bars) and Group II recaptures (grey bars) in the study area (refer to
Figure 1 for study layout).
143
120
GROUP I
GROUP II
Distance (m)
80
40
0
-40
-80
-120
Figure 5.3. Box and whisker plots of movement (m) upstream and downstream for
recaptured PIT tagged sculpin in Group I (11-306d), and Group II (35-212d). Vertical
lines represent the maximum and minimum movement, and the upper and lower limits
of the boxes represent the 75th and 25th percentiles, respectively.
144
5.7
References
Baras E, Malbrouck C, Houbart M, Kestemont P, and Mélard C. 2000. The effect of PIT
tags on growth and physiology of age-0 cultured Eurasian perch Perca fluviatilis of
variable size. Aquacult. 185:159-173.
Bruyndoncx L, Knaepkens G, Meeus W, Bervoets L, and Eens G. 2002. The evaluation
of passive integrated transponder (PIT) tags and visible implant elastomer (VIE)
marks as new marking techniques for the bullhead. J. Fish Biol. 60:260-262.
Burnham KP, and Anderson DR. 2002. Model selection and multi-model inference: a
practical information-theoretic approach. 2nd ed. Springer-Verlag, Inc., New York,
NY, USA.
Castro-Santos T, Haro A, and Walk S. 1996. A passive integrated transponder (PIT)
tag system for monitoring fishways. Fish. Res. 28:253-261.
Clément M. 1998. The effects of electrofishing and the accuracy of the “removal
method” to estimate population size of juvenile Atlantic salmon (Salmo salar) and
slimy sculpin (Cottus cognatus). MSc Thesis. University of New Brunswick,
Fredericton, NB, Canada.
Fish Survey Expert Working Group. 1997. Recommendations from Cycle 1 review.
EEM/1997/6. Environmental Effects Monitoring Program. Environment Canada,
Ottawa, ON, Canada.
Gibbons WN, Munkittrick KR, McMaster ME, and Taylor WD. 1998. Monitoring aquatic
environments receiving industrial effluents using small fish species 1: response of
spoonhead sculpin (Cottus ricei) downstream of a bleached-kraft pulp mill. Environ.
Toxicol. Chem. 17:2227-2237.
Gowan C, Young MK, Fausch KD, and Riley SC. 1994. Restricted movement in
resident stream salmonids: a paradigm lost? Can. J. Fish. Aquat. Sci. 74:26262637.
Gray MA, Curry RA, and Munkittrick KR. 2002. Non-lethal sampling techniques for
assessing fish populations for environmental assessment. Water Qual. Res. J. Can.
37:195-211.
Hill J, and Grossman GD. 1987. Home range estimates for three North American
stream fishes. Copeia. 1987:376-380.
McCleave JD. 1964. Movement and population of the mottled sculpin (Cottus bairdi
Girard) in a small Montana stream. Copeia. 1964:506-512.
Morgan CR, and Ringler NH. 1992. Experimental manipulation of sculpin (Cottus
cognatus) populations in a small stream. J. Freshw. Ecol. 7:227-232.
145
Munkittrick KR, McGeachy SA, McMaster ME, and Courtenay SC. 2002. Overview of
freshwater fish studies from the pulp Environmental Effects Monitoring program.
Water Qual. Res. J. Can. 37:49-77.
Peake SJ. 1999. Sodium bicarbonate and clove oil as potential anesthetics for nonsalmonid fishes. N. Am. J. Fish. Man. 18: 919-924.
Prentice EF, Flagg TA, McCutcheon CS, Brastow DF, and Cross DC. 1990a.
Equipment, methods, and an automated data-entry station for PIT tagging. Amer.
Fish. Soc. Symp. 7:335-340.
Prentice EF, Flagg TA, McCutcheon CS. 1990b. Feasibility of using implantable
passive integrated transponder (PIT) tags in salmonids. Amer. Fish. Soc. Symp.
7:317-322.
Roussel J-M, Haro A, and Cunjak RA. 2000. Field test of a new method for tracking
small fishes in shallow streams using passive integrated transponder (PIT)
technology. Can. J. Fish. Aquat. Sci. 57:1326-1329.
Smith W L. 1972. The dynamics of brook trout (Salmo trutta) and sculpin (Cottus sp.)
populations as indicators of eutrophication. PhD Thesis. Michigan State University,
MI, USA.
Utzinger J, Roth C, and Peter A. 1998. Effects of environmental parameters on the
distribution of bullhead Cottus gobio with particular consideration of the effects of
obstructions. J. Appl. Ecol. 35:882-892.
White GC, and Burnham KP. 1999. Program MARK: survival estimation from
populations of marked animals. Bird Study. 46(suppl):S120-139.
Van Vliet WH. 1964. An ecological study of Cottus cognatus (Richardson) in Northern
Saskatchewan. MA Thesis, University of Saskatchewan. Saskatoon, SK, Canada.
146
CHAPTER 6
GENERAL DISCUSSION
The overall objective of the thesis research project was to evaluate the biological
effects in streams subject to non-point source agricultural stressors. I began by
initiating an effects-based assessment in the Little River using the slimy sculpin as the
monitoring species. After changes were observed in the downstream agricultural
reaches, it was necessary to establish whether the responses were a result of an
upstream-downstream stream continuum or were influenced by inputs from surrounding
agricultural land-use. The effects-based assessment was expanded to also include a
continuum of sites (upstream to downstream) within a catchment uninfluenced by
agricultural activities, and a follow-up study was designed to confirm responses in other
streams and rivers subject to agricultural stressors.
A secondary objective of the overall thesis research was to evaluate the slimy
sculpin as a sentinel environmental monitoring species. Previous research had
suggested the slimy sculpin have a small home range and therefore was not expected
to make long distance movements. Novel techniques were used to provide evidence of
the ability of slimy sculpin to reflect local conditions, high spatial and temporal
residency, and low mobility. The following discussion will summarize the significant
results from the thesis research project while considering the application of an effectsbased assessment to study potential effects of agricultural stressors and the suitability
of the slimy sculpin in environmental monitoring. Additionally, the issue of intense
potato production in the region will be considered with respect to the timing of input of
potentially important stressors (soil runoff, pesticides, etc), that could be responsible for
influencing the YOY sculpin and the adult sculpin organ development. Suggestions for
future research needs and conclusions from the research will finish the discussion.
147
6.1
Effects-based assessment
Slimy sculpin populations performed differently at sites in the agricultural region
than sculpin populations in the forested region of the Little River. Slimy sculpin
population structures were altered in the agricultural region, with fewer YOY sculpin
comprising the local populations. Mature sculpin tended to be larger, male sculpin had
smaller gonads, and female sculpin had smaller livers, gonads, and reduced fecundity
and egg size. Based on the comparison of sculpin collected at agricultural sites with
one or more forested sites, there was evidence that sculpin were responding to
changes in their environment and consistent responses were confirmed over time.
Population and individual differences between agricultural and forested sites were
less marked in the fall of 2001, which followed the driest summer during the study
period. Based on the generalized fish population response patterns proposed by
Gibbons and Munkittrick (1994), the responses of the adult slimy sculpin in the lower
catchment of the Little River may be a result of metabolic disruption or metabolic
redistribution (Figure 6.1). A change in the ability of the fish to adequately process
energy leads to conflicting performance in indicators of energy storage and expenditure
(Gibbons and Munkittrick 1994).
It is clear that the responses were not a consequence of natural downstream
changes, and that the impacts are not restricted to the Little River catchment. In a
survey of 19 New Brunswick rivers (Chapter 3), only 2 of the 11 agricultural streams
had evidence of successful reproduction, while YOY were present at all forested sites.
This study confirmed responses seen with YOY sculpin in the agricultural region of the
Little River but over a much broader geographic scale and on many different river
systems. The multi-river study also suggested that there may be multiple stressors
exerting influences on fish in agricultural areas as temperature and sediment were
148
unable to explain all of the variability within the data. The study demonstrated a strong
negative relationship between sculpin density and temperature, and a strong positive
relationship between median YOY sculpin size and temperature. The variability in the
density of YOY sculpin at the agricultural sites suggests that there are more factors
responsible than temperature alone.
It was also clear that agricultural areas were exposed to increased levels of
sediment deposition, especially fine material. However, there were no direct
relationships of sediment deposition with sculpin size or density in either the agricultural
or forested regions. Sediment deposition was measured at each site from the
beginning of July to the beginning of October, using relatively small sediment traps. It
is possible that the sediment traps did not accurately reflect the characteristics of
sediment deposition, or that the relationship of sediment levels to fish performance is
complex. Sediment may function as an important transport mechanism for
contaminants, and the consequences of sediment deposition to fish may depend on the
contaminant load of the sediment and not necessarily the physical properties. Studies
have already been initiated in agricultural regions to extract chemicals associated with
sediments deposited after storm events, and expose them to fish under laboratory
conditions to monitor possible acute and chronic effects (Hewitt LM, Environment
Canada; Munkittrick KR, University of New Brunswick; and Teather KL, University of
Prince Edward Island).
It is also possible that sediment influences sculpin populations outside of the period
evaluated. Sculpin are benthic lithophilous fish, requiring clean, stony substrates for
spawning (Balon 1975). The slimy sculpin exhibits a mating pattern where males
clean, maintain, and defend open spaces beneath rocks (Mousseau and Collins 1987).
Based on observations of egg masses, sculpin spawned in mid-May in the Little River,
coinciding with mean water temperatures ranging between 6 and 10ºC. Multiple
149
females may lay their eggs in one nest (Figure 6.2), which the male cares for until
shortly after the incubation period of about 29d (Van Vliet 1964; Mousseau and Collins
1987). This means that the ‘reproduction’ period, from nest preparation to hatch to
larval emergence, occurs from about the beginning of May until the middle of June.
Depending on the timing and degree of spring runoff and subsequent rainfall in the
spring, sediment may have biological impacts earlier than was measured in our study.
Stressors often have the greatest impact during critical early developmental stages
(McKim 1977).
6.2
Suitability of the slimy sculpin
Gibbons (1997) suggested that in some cases small-bodied fish species might be
more suitable than larger bodied fish species for environmental monitoring. The limited
movement and sometimes territorial behaviour of smaller fish species both increase the
probability that responses will reflect the area from which they were collected. Smallbodied fish are generally expected to be present at higher numbers due to their position
in the food chain, and with shorter life spans they may show responses to changes in
their environment faster than long-lived species (Gibbons 1997). Finally, the lack of
commercial importance of many small fishes removes any confounding effects that
exploitation pressures may exert (Gibbons 1997).
It was important to determine whether the sculpin responses were reflective of local
conditions, and detailed studies were designed to examine the issue of spatial and
temporal residency. Stable isotopes were used to address the questions of site fidelity
and the ability to reflect local conditions (Chapter 4). The basis for our study was a
previous study that found enriched δ15N in fish residing in steams receiving fertilizer
input from agricultural activity (Harrington et al. 1998). Although we did not find a
150
similar δ15N enrichment in sculpin in the agricultural region of the Little River, likely due
to the predominant use of commercial synthetic fertilizers, stable isotope analysis was
still a useful technique for identifying sculpin from different sites. Using δ15N and δ13C
isotopes together, slimy sculpin exhibited site-specific signatures along the continuum
of the Little River. The variability in isotopic values was low both spatially and
temporally. Because the isotope values are based on what fish eat, movement into
other locations or migratory behaviours would result in a more variable signature.
Stable isotopes proved to be a useful technique to suggest a moderate to high degree
of site fidelity.
Movements of individual sculpin were followed using PIT-tagged fish monitored over
a 10-month period to quantify movements and determine the feasibility of using PIT
tags in sculpin (Chapter 5). Throughout the study, 37% of the 112 tagged sculpin were
recaptured at least once, showing a median movement of 11.3m from release points.
Although there was moderate variability in individual movements, 75% of recaptured
sculpin moved <30m. There were no differences between movements in the fall and in
the spring, and there were still new fish recaptured on the last tracking event,
suggesting that the sculpin do not move out of the area in the winter or to spawn in the
spring.
Results from the studies on site fidelity and mobility provided strong evidence for
the high spatial and temporal residency of the slimy sculpin. Both factors increase the
probability that responses in fish will reflect the local environments from which they
were collected. This may be particularly important when assessing the impacts of nonpoint source pollution when the pollution is difficult to quantify and characterize. The
responses in YOY and adult sculpin performance observed in the agricultural region of
the Little River (Chapter 2) and in other streams influenced by surrounding agricultural
151
activities (Chapter 3), were identified by comparison to sculpin residing in forested
regions uninfluenced by agricultural stressors. Based on the ability to respond to, and
reflect, changes in its local environment and the high spatial and temporal residency we
are confident that the slimy sculpin is a suitable sentinel monitoring species.
6.3
Important agricultural inputs
Potato farming can affect fish through a variety of mechanisms, including altering
water temperatures by reducing cover, changing the timing and patterns of surface
runoff, increasing the amount of sediments that enter streams, adding nutrients to the
system, and discharging a variety of chemical control agents. My study also
demonstrated increased water temperatures and sediment deposition in potato-growing
areas. Any of these factors acting alone, or in combination could be responsible for the
biological responses that were documented in the Little River catchment and in the
other agricultural sites studied.
The potato belt located along the upper St. John River valley in northwestern New
Brunswick has some of the most serious soil erosion problems in Canada (Chow et al.
1995). The rolling topography results in high runoff rates and soil loss, with soil loss in
the region estimated at about 20 t/ha/yr (Chow et al. 1995). Although there has been
little direct monitoring of runoff into the Little River, Agriculture and Agri-Food Canada
has been monitoring the Black Brook Experimental Watershed since 1990 to monitor
runoff and evaluate different cropping systems in soil loading, nutrient enrichment, and
chemical contamination. Black Brook is a sub-catchment of the Little River, measuring
about 14.5km2 (joins Little River at site 9, Figure 2.1). Although Black Brook drains only
approximately 5% of the land area of the Little River, an estimated 65% of the
catchment is under potato production (Chow et al. 1995). This relatively small, 3rd order
152
stream experiences significant sediment and nutrient loading, with peak discharge and
inputs occurring in April during spring runoff (Table 6.1). If the inputs of April are
subtracted, 85% of sediment, and 45-60% of nutrient annual loading occurs during the
cropping period.
Agriculture and Agri-Food Canada ranked pesticides used in the Black Brook
catchment based on frequency and extent of application (Table 6.2; H Rees,
Agriculture and Agri-Food Canada, pers. comm.). Based on toxicity to fish and
chemical properties, mancozeb and chlorothalonil represent the greatest acute risk to
fish. Both are active ingredients in fungicide formulations that are generally applied
throughout the cropping season (Table 1.1). Diquat is the most persistent in
sediments, but is not considered a risk to fish and other aquatic organisms.
Chlorothalonil and imidacloprid are highly toxic to aquatic invertebrates (EXTOXNET
1999) and could impact the food base in a receiving stream. Of the five pesticides,
mancozeb and metribuzin have both been identified as possible endocrine disruptors
(Kaminuma 2001). Depending on the mechanism of action, exposure to endocrine
disrupting chemicals may have impacts on reproductive success of fish and even the
health of offspring from exposed individuals (Gray et al. 1999).
High water solubility values and adsorption coefficients are both properties that
increase the likelihood of input of agrochemicals into streams, via rainfall or associated
with soil particles, if there is not sufficient time for pesticides to break down before the
next significant rainfall event. Pesticide spray programs vary during the potato growing
season depending on the conditions, but June to September are peak periods for
herbicide, insecticide, and fungicide applications (Table 1.1). The emergence of YOY
sculpin in mid to late June means the potential for exposure to agricultural stressors
during sensitive early life stages is quite high. The 2001 year- class was the strongest
153
year class seen during my study, and represented the driest year, though the reduced
precipitation was not significantly different from the other years studied.
6.3.1
Recent legislation related to potato farming
Following an increase in the frequency and magnitude of fish kill events, Prince
Edward Island became the first Maritime Province to introduce legislation to mitigate
aquatic impacts of agricultural cultivation. In 2001, the government of PEI created a
Watercourse Buffer Zone amendment to the province’s Environmental Protection Act,
which prohibits planting of an agricultural crop within 10m of a watercourse or
designated wetland (Environmental Protection Act 2003 RSPEI c. E-9 s.11). In 2002,
the PEI government introduced the Agricultural Crop Rotation Act stating that the same
crop could not be planted more than once in three years, and restrictions were placed
on certain crops planted in fields with slopes greater than 9% (Crop Rotation Act 2002
RSPEI c. A-8.01).
No similar legislation currently exists in New Brunswick, though there are programs
that exist to promote best management practices (BMPs) by farmers. BMPs include
special attention to crop rotation, tillage, winter cover, cross slope farming, strip
cropping, and erosion control structures (PEI DAF 1998). Chow et al. (1999) compared
the runoff from a terraced potato field with a grassed waterway to that of a potato field
planted up-and-down the slope and found mean annual sediment losses of 1 001 and
20 760 kg/ha, respectively. With a possible 20-fold difference in soil loss it seems
logical to implement BMPs, but the greatest factor that discourages farmers from the
implementation of full BMP is the lag time for benefits to be realized (PEI DAF 1998).
Farmers weigh long-term benefit with short-term financial gains if under high farm debt
154
pressures. In addition, much agricultural land today is rented, so farmers have less
incentive to consider the long-term sustainability of the land resources (PEI DAF 1998).
6.4
Conclusions
The objectives of this PhD research, in order of appearance in the thesis, were first
to conduct an effects-based assessment of the potential impacts of non-point source
agriculture on the slimy sculpin. A three-year effects-based monitoring study on the
Little River was successful at demonstrating changes in slimy sculpin population
structure and changes in energy storage and expenditure for mature slimy sculpin in
the agricultural region. The changes may be indicative of metabolic disruption and
could be due to exposure to multiple stressors. This is one of the first field studies to
document the biological effects of chronic exposure of fish to multiple agricultural
stressors.
Secondly, in order to establish whether these responses were system-specific or
due to land-use inputs, a multi-system study was conducted in agricultural and forested
streams throughout western and northwestern New Brunswick. The relative influence
of increased sediment deposition and water temperatures were also assessed in
relation to YOY sculpin size and sculpin density. This study confirmed the responses
seen with YOY sculpin in the agricultural region of the Little River over a much broader
geographic scale, on many different river systems. The multi-river study also confirmed
that there were multiple stressors exerting influence on sculpin performance in
agricultural regions.
Thirdly, stable isotope analysis and PIT tags were used to assess the spatial and
temporal residency of slimy sculpin to provide evidence of its suitability as a sentinel
species, able to reflect its local environmental conditions. Both techniques provided
155
strong support for the hypothesis that sculpin have limited mobility and should reflect
the environment from which they were collected.
6.5
Future considerations
Many questions and possible directions for research have arisen during the course
of this research project. Research on the biological effects from non-point source
agricultural inputs has just begun, and there are still many questions related to the
slimy sculpin that could be explored further. The following cursory list represents
significant questions and research needs that could be addressed in the future.
•
More information on the basic ecology and biology of the slimy sculpin and
other small bodied fish species will aid in their use in environmental monitoring
and the interpretation of results and responses. Broader baseline data on
growth rates, reproductive strategies and biology, and habitat requirements will
facilitate use of small fish species used in environmental monitoring (Gibbons
1997).
•
Comparison of effects-based responses of other fish species in the same
system to determine whether the slimy sculpin is adequately reflecting the
responses of other biota exposed to the same stressors.
•
Development of methods to evaluate sedimentation to allow for meaningful
comparisons to resident biota. The timing and methodology needs to be
relevant for the responses observed. Deposition should be monitored
throughout the year and associated habitat alteration also quantified.
•
More study could be done on the mobility of slimy sculpin. Environmental
factors influencing the movement of sculpin would be an interesting research
study if related to factors including temperature, habitat alteration, and dissolved
156
oxygen could be measured or manipulated.
•
The impact of nutrients from fertilizer runoff in streams adjacent to agricultural
fields could be studied using isotope tracers and assessing biota at each level
of the food web in order to trace effects on growth and productivity.
•
Gibbons and Munkittrick (1994) recommended that an appropriate follow-up
study for fish showing signs of metabolic redistribution would be a detailed
physiological investigation of energy expenditure and storage, and hormonal
regulation of reproduction and growth.
•
The timing of the inputs of agricultural stressors needs to be assessed more
closely with respect to the changes in fish populations and individual responses.
Detailed studies at the biochemical level may help in determining when the
effects are initiated in mature fish.
157
Table 6.1. Average monthly discharge, and sediment and chemical loading for the
Black Brook Catchment (1992-93) (modified from Chow et al. 1995). The non-cropping
period is 1 November-30 April, and the cropping period is 1 May- 31 October.
Month
Discharge
Sediment
N03-N
P205
K
Mg
Ca
(‘000m3)
(ton)
(ton)
(ton)
(ton)
(ton)
(ton)
January
718
95
2.7
<0.1
0.9
1.7
33.3
February
215
14
1.1
<0.1
0.2
0.8
10.0
March
216
30
0.9
<0.1
0.3
0.6
10.3
April
4851
4712
12.5
0.2
10.8
10.8
148.8
May
982
334
2.9
<0.1
1.4
4.7
46.2
June
417
217
1.5
<0.1
0.7
2.5
18.2
July
424
932
1.3
<0.1
0.8
2.2
18.2
August
400
82
1.3
<0.1
0.9
2.6
18.3
September
188
13
0.6
<0.1
0.3
1.2
9.1
October
891
143
2.7
<0.1
1.6
5.5
43.9
November
688
66
3.5
<0.1
1.1
4.5
44.1
December
773
96
3.9
<0.1
1.3
4.9
48.5
Non-cropping
7461
5013
24.6
>0.2
14.6
23.3
295
Cropping
3302
1721
10.3
<0.2
5.7
18.7
153.9
Total
158
Table 6.2. Toxicity classification, expected toxicity to fish, chemical properties, and persistence1 of the top five pesticides
used in the agricultural region of the Black Brook catchment. Chemical ranking was based on the number of applications
multiplied by the area the chemical was applied (H Rees, Agriculture and Agri-Food Canada, pers. comm.).
Water solubility
Adsorption
Persistence
Chemical
Pesticide
Toxicity to fish
(mg/L)
coefficient
1. Mancozeb
fungicide
moderate to high
6
>2000
1-7d
1-2d
2. Chlorothalonil
fungicide
high
0.6
1380
1-3mths
n/a2
3. Diquat
herbicide
practically non-toxic
700 000
1000000
>1000d
<48h
48-190d
>31d
30-120d
7d
4. Imidacloprid
insecticide
5. Metribuzin
herbicide
2
moderate to low
0.51
n/a
slightly
1 050
60
1 - Chemical information summarized from pesticide information profiles (EXTOXNET 1999).
2 - Information not available.
159
Soil
Water
0 0 0
Possible recovery
+ 0 0
Disruption removed
+ - +/-
0 - +/-
Starting point
0 0 0
Figure 6.1. Progression of a fish population response pattern showing metabolic
disruption with possible recovery pattern after the disruption is removed. Response is
summarized based on age distribution, energy expenditure, and energy storage, [x x x],
respectively showing no change [0], an increase [+], or decrease [-]. Modified from
Gibbons and Munkittrick (1994).
160
Figure 6.2. Slimy sculpin egg nest on the underneath of a rock found in the Little River.
The colour difference in the egg masses is indicative of multiple females laying their
eggs in nests maintained by male sculpin.
161
6.6
References
Balon EK. 1975. Reproductive guilds of fishes: A proposal and definition. J. Fish. Res.
Board Can. 26:1429-1438.
Chambers PA, Guy M, Roberts E, Charlton MN, Kent R, Gagnon C, Grove G, Foster N,
DeKimpe C, and Giddings M. 2001a. Nutrients. Section 6. Threats to Sources of
Drinking Water and Aquatic Ecosystem Health in Canada. NWRI Scientific
Assessment Report Series No.1. National Water Research Institute, Burlington, ON,
Canada. pp 23-36.
Chambers P, DeKimpe C, Foster N, Goss N, Miller J, and Prepas E. 2001b.
Agricultural and forestry land use impacts. Section 13. Threats to Sources of
Drinking Water and Aquatic Ecosystem Health in Canada. NWRI Scientific
Assessment Report Series No.1. National Water Research Institute, Burlington, ON,
Canada. pp 57-63.
Chow TL, Rees HW, Ghanem I, and Daigle J-L. 1995. Effects of intensive potato
production on surface water quality in the Black Brook Watershed, St. André Parish,
New Brunswick. Proceedings of the 48th Canadian Water Resources Association
Annual Conference, Managing the Water Environment. June 20-23. Fredericton,
NB, Canada. pp. 303-312.
Chow TL, Rees HW, and Daigle J-L. 1999. Effectiveness of terraces/grassed waterway
systems for soil and water conservation: A field evaluation. J. Soil Water Cons.
54:577-583.
EXTOXNET. 1999. Extension Toxicology Network. Cooperative initiative between
University of California-Davis, Oregon State University, Michigan State University,
Cornell University, University of Idaho. www.ace.orst.edu/info/extoxnet.
Downloaded 15 August 2003.
Gibbons WN. 1997. Suitability of small fish species for monitoring the effects of pulp
mill effluent on fish populations. PhD thesis. University of Waterloo. Waterloo, ON,
Canada.
Gibbons WN, and Munkittrick KR. 1994. A sentinel monitoring framework for identifying
fish population responses to industrial discharges. J. Aquat. Ecosys. Health. 3:227237.
Gray MA, Teather KL, Metcalfe CD. 1999. Reproductive success and behaviour of
Japanese medaka (Oryzias latipes) exposed to 4-tert-octylphenol. Environ. Toxicol.
Chem. 18:2587-2594.
Harrington R, Kennedy BP, Chamberlain CP, Blum JD, and Folt CL. 1998. 15N
enrichment in agricultural catchments: field patterns and applications for tracking
Atlantic salmon (Salmo salar). Chem. Geol. 147:281-294.
Kaminuma T. http://www.nihs.go.jp/hse/endocrine-e/paradigm/pesticide/pest-list-e.html.
Updated 17 November 2001, downloaded 2 August 2003.
162
McKim JM. 1977. Evaluation of tests with early life stages of fish for predicting longterm toxicity. J. Fish. Res. Board Can. 34:1148-1154.
Mousseau TA, and Collins NC. 1987. Polygyny and nest site abundance in the in the
slimy sculpin (Cottus cognatus). Can. J. Zool. 65:2827-2829.
PEI DAF. 1998. Best management practices: Soil conservation for potato production.
PEI Department of Agriculture and Forestry, Charlottetown, PE, Canada.
Van Vliet WH. 1964. An ecological study of Cottus cognatus (Richardson) in northern
Saskatchewan. MA Thesis, University of Saskatchewan. Saskatoon, SK, Canada.
163
VITA
Candidate's full name:
Michelle Anya Gray
Universities attended:
1992 - 1996
1996 - 1998
Trent University, Peterborough, ON
BSc – Biology and Environmental Science
Trent University, Peterborough, ON
MSc – Watershed Ecosystems Graduate Program
Articles Published in Refereed Journals
Galloway B, Munkittrick KR, Currie S, Gray M, Curry RA, and Wood C. Examination of
the responses of slimy sculpin (Cottus cognatus) and white sucker (Catostomus
commersoni) collected on the Saint John River downstream of pulp mill, paper mill,
and sewage discharges. Environ. Toxicol. Chem.22:2898-2907.
Gray MA, Curry RA, and Munkittrick KR. 2002. Non-lethal sampling techniques for
assessing fish populations for environmental assessment. Water Qual. Res. J. Can.
37:195-211.
Teather KL, Harris M, Boswell JL, and Gray MA. 2001. Effects of Acrobat-MZ ® and
Tattoo-C® on Japanese medaka (Oryzias latipes) development and adult male
behaviour. Aquat. Toxicol. 51:419-430.
Teather KL, Boswell JL, and Gray MA. 2000. Early life history parameters of Japanese
medaka (Oryzias latipes). Copeia. 2000:813-818.
Munkittrick KR, McMaster ME, Van Der Kraak G, Portt C, Gibbons WN, Farwell A, and
Gray M. 2000. Development of methods for effects-driven cumulative effects
assessment using fish populations: Moose River project. Society of Environmental
Toxicology and Chemistry. Pensacola, Florida.
--- French translation: National Water Research Institute Publication No. 01-064,
Environment Canada, Burlington, Ontario, Canada.
Gray MA, Teather KL, and Metcalfe CD. 1999. Reproductive success and behaviour of
Japanese medaka (Oryzias latipes) exposed to 4-tert-octylphenol. Environ. Toxicol.
Chem. 18:2587-2594.
Gray MA, Niimi AJ, and Metcalfe CD. 1999. Factors affecting the development of
testis-ova in medaka, Oryzias latipes, exposed to octylphenol. Environ. Toxicol.
Chem. 18:1835-1842.
Gray MA, and Metcalfe CD. 1999. Toxcity of 4-tert-octylphenol to early life stages of
Japanese medaka (Oryzias latipes). Aquat. Toxicol. 46:149-154.
Metcalfe CD, Gray MA, and Kiparissis Y. 1999. The Japanese medaka (Oryzias
latipes): An in vivo model for assessing the impacts of aquatic contaminants on the
reproductive success of fish. In Impact of hazardous aquatic contaminants:
Concepts and approaches. Rao S (ed). CRC Press, Boca Raton, USA. pp. 29-52.
Gray MA, and Metcalfe CD. 1997. Induction of testis-ova in Japanese medaka (Oryzias
latipes) exposed to p-Nonylphenol. Environ. Toxicol. Chem. 16:1082-1086.
Articles Submitted to Refereed Journals
Gray MA, Cunjak RA, and Munkittrick KR. Measuring small-scale movements of the
slimy sculpin (Cottus cognatus) to assess site fidelity in a small river. Submitted to
Copeia. May 2003.
Gray MA, Cunjak RA, and Munkittrick KR. The use of stable isotope analysis to assess
the site fidelity of slimy sculpin (Cottus cognatus). Submitted to Canadian Journal of
Fisheries and Aquatic Sciences. July 2003.
Gray MA, and Munkittrick KR. An effects-based assessment of slimy sculpin (Cottus
cognatus) populations in potato agriculture regions of northwestern New Brunswick.
Submitted to the Water Quality Research Journal of Canada. January 2004.
Gray MA, Curry RA, and Munkittrick KR. Investigating the impacts of sediment and
temperature on slimy sculpin (Cottus cognatus) populations in agricultural
catchments. Submitted to Environmental Toxicology and Chemistry. January 2004.
Non-Refereed Contributions
Gray MA, and Curry RA. 2002. Stream fish responses to blueberry production:
searching for indicators of human-induced stress. NBCFWRU Report #02-06. New
Brunswick Cooperative Fish and Wildlife Research Unit, University of New
Brunswick, Fredericton, NB, Canada.
Gray MA, Teather KL, Sherry J, McMaster M, Hewitt M, and Mroz RE. 2002. Potential
endocrine disruption in freshwater systems near agricultural areas on Prince
Edward Island. In Effects of land use practices on fish, shellfish, and their habitats
on Prince Edward Island. Cairns DK (ed). Can. Tech. Rpt. Fish. Aquat. Sci. No.
2408. pp. 116-118.
Gray MA, Teather KL, Sherry J, McMaster M, Hewitt M, and Mroz RE. 2000. Endocrine
disrupting potential in freshwater ecosystems near agricultural areas on Prince
Edward Island. EPS-5-AR-99-6. Environment Canada Surveillance Report.
Environment Canada, Dartmouth, NS, Canada.
Savard M, and Gray M. 1998. Assessment of the use, handling and disposal of
antifouling paints at ship repair facilities in the Atlantic region. EPS-5-AR-98-2.
Environment Canada Surveillance Report. Environment Canada, Dartmouth, NS,
Canada.
Conference Presentations – Platform
Gray MA, and Munkittrick KR. 2003. An effects-based assessment of slimy sculpin
(Cottus cognatus) populations in agricultural regions. Society of Environmental
Toxicology and Chemistry. 24nd Annual Meeting. Nov 8-11. Austin, TX, USA.
Gray MA, and Munkittrick KR. 2003. Slimy sculpin (Cottus cognatus) population
responses in agricultural regions of northwestern New Brunswick. Aquatic Toxicity
Workshop. 30th Annual Meeting. Sep 28-Oct 1. Ottawa, ON, Canada.
Brasfield SM, Gray MA, and Munkittrick KR. 2003. Use of fish populations in an effectsbased assessment to evaluate non-point stressors associated with agriculture.
Society of Environmental Toxicology and Chemistry Asia-Pacific. Sep 29-Oct. 1.
Christchurch, New Zealand.
Gray MA, and Munkittrick KR. 2003. Overview of a 3-year monitoring study looking at
fish responses in agricultural watersheds in northwestern New Brunswick. 3rd
Annual Canadian Rivers Institute Meeting. Mar 27. Saint John, NB, Canada.
Munkittrick KR, Hewitt LM, Teather K, MacLatchy D, Van Der Kraak G, Brasfield S,
Gray M, Jardine C and Gormley K. 2003. Quantification of sediment-associated
EDSs in agricultural areas, and their potential biological impacts on fish. Canadian
Network of Toxicology Centres Annual Research Symposium, Mar 25-26. Ottawa
ON, Canada.
Gray MA, Curry RA, and Munkittrick KR. 2003. Investigating the impacts of agricultural
stressors on sculpin populations. Canadian Conference for Fisheries Research
(CCFFR). Jan 2-5, Ottawa, ON, Canada.
Gray MA, Doucett RR, Cunjak RA, and Munkittrick KR. 2003. Stable isotope analysis
of slimy sculpin (Cottus cognatus) to address questions of limited mobility.
Canadian Conference for Fisheries Research (CCFFR). Jan 2-5, Ottawa, ON,
Canada.
Brasfield S, Galloway B, Gray M, Peters L, Curry A, and Munkittrick K. 2003.
Identification of an upstream source of contamination on the Saint John River near
Claire, New Brunswick. Canadian Conference for Fisheries Research (CCFFR).
Jan 2-5, Ottawa, ON, Canada.
Aguilar C, Gonzalez-Sanson G, Faloh I, Gray M, and Curry A. 2002. Trophic structure
among coral reef fishes adjacent to Habana City, Cuba with special reference to the
bicolour damselfish, Stegastes partitus (Poey, 1868). Stressors in aquatic
environments. Nov 28-29. Havana, Cuba.
Gray MA, Doucett RR, Cunjak RA, and Munkittrick KR. 2002. The use of stable isotope
analysis to assess the site fidelity of the slimy sculpin (Cottus cognatus).
Applications of Stable Isotope Techniques to Ecological Studies. 3rd International
Conference. Apr 29-May 1. Flagstaff, AZ, USA.
Gray MA, Cunjak RA, and Munkittrick KR. 2002. You are what you eat: Using stable
isotopes to assess movements of the slimy sculpin. 10th Annual Graduate Student
Association Conference on Student Research. Feb 21-22. University of New
Brunswick, Fredericton, NB, Canada.
Gray MA, Cunjak RA, and Munkittrick KR. 2002. Assessing the mobility of slimy sculpin
in rivers. 2nd Annual Toxic Substances Research Initiative (TSRI) Public Workshop.
Jan 25. University of New Brunswick-Saint John, Saint John, NB, Canada.
Gray MA, and Munkittrick KR. 2002. Investigating the effects of sediment and
temperature on wild fish. 2nd Annual Toxic Substances Research Initiative (TSRI)
Public Workshop. Jan 25. University of New Brunswick-Saint John, Saint John, NB,
Canada.
Munkittrick KR, Gray M, Galloway B, and Curry A. 2001. Identifying areas of concern in
the freshwater portion in the Saint John River using fish performance. Society of
Environmental Toxicology and Chemistry. 22nd Annual Meeting. Nov 11-15.
Baltimore, MD, USA.
Gray MA, Cunjak RA, and Munkittrick KR. 2001. The use of PIT tags to establish
movements of the slimy sculpin (Cottus cognatus) in small streams. Atlantic
International Chapter of the American Fisheries Society. 27th Annual Meeting. Sep
23-25. Lake Winnipesaukee, NH, USA.
Gray MA, and Munkittrick KR. 2001. The utility of the slimy sculpin (Cottus cognatus) in
environmental monitoring. Northeast Wildlife Graduate Conference. Mar 2-4.
Durham, NH, USA.
Gray MA, Cunjak RA, and Munkittrick KR. 2001. The ability of slimy sculpin (Cottus
cognatus) to reflect local conditions. Canadian Conference for Fisheries Research
(CCFFR). Jan 4-6. Toronto, ON, Canada.
Gray MA, and Munkittrick KR. 2000. Agricultural influences on fish ecology and
performance near Grand Falls NB. 1st Annual Toxic Substances Research Initiative
(TSRI) Public Workshop. Dec 18. Wu Conference Centre, University of New
Brunswick, Fredericton, NB, Canada.
Munkittrick KR, Currie S, Gray M, Galloway B, and Curry A. 2000. Development of a
cumulative effects strategy for the freshwater portion of the Saint John River. 1st
Annual Toxic Substances Research Initiative (TSRI) Public Workshop. Dec 18. Wu
Conference Centre, University of New Brunswick, Fredericton, NB, Canada.
Curry RA, Doherty C, Gray M, and Munkittrick KR. 2000. Studies of fish mobility in the
Saint John River. 1st Annual Toxic Substances Research Initiative (TSRI) Public
Workshop. Dec 18. Wu Conference Centre, University of New Brunswick,
Fredericton, NB, Canada.
Gray MA, and Munkittrick KR. 2000. Agricultural influences on fish ecology and
performance in New Brunswick. Aquatic Toxicity Workshop. 27th Annual Meeting.
Oct 1-4. St. John’s, NF, Canada.
Munkittrick KR, Currie S, Gray M, Galloway B, and Curry A. 2000. Development of a
cumulative effects strategy for the freshwater portion of the Saint John River.
Aquatic Toxicity Workshop, 27th Annual Meeting. Oct 1-4, St. John’s, NF, Canada.
Gray MA, and Munkittrick KR. 2000. Population responses of slimy sculpin in
agricultural regions of the Upper Saint John River. Canadian Conference for
Fisheries Research (CCFFR). Jan 6-8, Fredericton, NB, Canada.
Gray MA, Teather KL, Metcalfe CD, and Niimi AJ. 1997. Reproductive success of
Japanese medaka (Oryzias latipes) exposed to 4-tert-octyphenol. Aquatic Toxicity
Workshop. 24th Annual Meeting. Oct 19-22. Niagara Falls, ON, Canada.
Metcalfe TL, Gray MA, Metcalfe CD, and Niimi AJ. 1997. Effects of estrogenic
compounds on the gonadal development of Japanese medaka (Oryzias latipes).
Aquatic Toxicity Workshop. 24th Annual Meeting. Oct 19-22. Niagara Falls, ON,
Canada.
Gray MA, and Metcalfe CD. 1997. Factors affecting the development of testis-ova in
Japanese medaka. Society for Environmental Toxicology and Chemistry. St.
Laurent/Laurentian Chapter Meeting. May 30-31. Montreal, PQ, Canada.
Metcalfe CD, and Gray MA. 1996. Induction of testis-ova in Japanese medaka exposed
to 4-nonylphenol. International Association for Great Lakes Research. 39th Annual
Meeting. May 26-30. University of Toronto. Missisauga, ON, Canada.
Conference Presentations – Poster
Gray MA, and Munkittrick KR. 2003. Monitoring the impacts of agricultural activities
using a small-bodied fish species in New Brunswick. Canadian Water Network
Meeting. Mar 23-26. Saint John, NB, Canada.
Gray MA, and Munkittrick KR. 2002. Monitoring the impacts of agricultural activities
using a small-bodied fish species in New Brunswick. Toxic Substances Research
Initiative (TSRI) National Conference. Mar 5-8. Ottawa, ON, Canada.
Gray MA, and Munkittrick KR. 2002. Population responses of slimy sculpin (Cottus
cognatus) in agricultural regions of New Brunswick, Canada. Toxic Substances
Research Initiative (TSRI) National Conference. Mar 5-8. Ottawa, ON, Canada.
Gray MA, Cunjak RA, and Munkittrick KR. 2002. Evaluation of the ability of the slimy
sculpin (Cottus cognatus) to reflect local conditions for environmental monitoring.
Toxic Substances Research Initiative (TSRI) National Conference. Mar 5-8. Ottawa,
ON, Canada.
Galloway B, Munkittrick KR, Currie S, Gray M, Curry RA, and Wood C. 2002.
Examination of the responses of slimy sculpin (Cottus cognatus) and white sucker
(Catostomus commersoni) collected on the Saint John River downstream of pulp
mill, paper mill and sewage discharges. Toxic Substances Research Initiative
(TSRI) National Conference. Mar 5-8. Ottawa, ON, Canada.
Gray MA, Cunjak RA, and Munkittrick KR. 2001. Evaluation of the ability of the slimy
sculpin (Cottus cognatus) to reflect local conditions for environmental monitoring.
Society of Environmental Toxicology and Chemistry. 22nd Annual Meeting. Nov 1115. Baltimore, MD, USA.
Munkittrick KR, Galloway B, Currie S, Gray M, Curry A, and Wood C. 2000.
Examination of the responses of slimy sculpin (Cottus cognatus) and white sucker
(Catostomus commersoni) collected on the Saint John River downstream of pulp
mill, paper mill and sewage discharges. Society of Environmental Toxicology and
Chemistry. 21st Annual Meeting. Nov 12-16, Nashville, TN, USA.
Teather KL, Duffy E, and Gray MA. 2000. What Japanese medaka can tell us about
Prince Edward Island streams. 27th Annual Aquatic Toxicity Workshop. Oct 1-4. St.
John’s, NF, Canada.
Gray MA, and Munkittrick KR. 2000. Population responses of slimy sculpin (Cottus
cognatus) in agricultural regions of New Brunswick, Canada. Society of
Environmental Toxicology and Chemistry. Third World Congress and 10th Annual
European meeting. May 21-25. Brighton, UK.
Metcalfe CD, Gray MA, Metcalfe TL, and Niimi AJ. 1997. Induction of testis-ova in
medaka by exposure to estrogenic compounds. Society of Environmental
Toxicology and Chemistry. 18th Annual Meeting. Nov 16-20. San Francisco, CA,
USA.
Gray MA, and Metcalfe CD. 1996. Development of testis-ova in Japanese medaka
(Oryzias latipes) exposed to 4-nonylphenol. Society of Environmental Toxicology
and Chemistry. 17th Annual Meeting. Nov 17-21. Washington, DC, USA.