Response of dissolved nutrients and periphyton to spawning Pacific salmon in three northern Michigan streams

Abstract

The ecological effects of spawning runs of native Pacific salmon (Oncorhynchus spp.) on stream ecosystems of the Pacific Northwest and Northern Pacific Rim have been studied extensively. However, little is known about how nonnative Pacific salmon affect stream ecosystems in the Great Lakes Basin, especially given the difference in environmental context between the regions. Mechanisms by which salmon spawners alter stream ecosystems include nutrient enrichment from excretion by live adults, carcass decomposition, and physical disturbance of the substrate during redd construction. The objective of our study was to quantify changes in water chemistry and benthic periphyton in 3 streams in northern Michigan that have spawning populations of Pacific salmon. In each stream, dissolved nutrients (soluble reactive P [SRP], NH4+, dissolved organic C [DOC], NO3−), and periphyton on gravel were sampled before, during, and after the spawning run in reaches upstream and downstream of a salmon barrier. Nutrients increased in reaches downstream of the barrier when salmon were present, but the magnitude of increase was low relative to increases observed in Pacific Rim streams. During and after the spawning run, periphyton biomass declined significantly in reaches where high densities of salmon spawners were present. Our results suggest that disturbance by spawning salmon may override their enrichment effects in northern Michigan streams, but this pattern may in part be driven by environmental context, especially the presence of finer substrates.

Full text

Response of dissolved nutrients and periphyton to spawning Pacific
salmon in three northern Michigan streams
Scott F. Collins
1
AND Ashley H. Moerke
2
School of Biological Sciences, Lake Superior State University, Sault Ste Marie, Michigan 49783 USA
Dominic T. Chaloner
3
, David J. Janetski
4
,AND Gary A. Lamberti
5
Department of Biological Sciences, University of Notre Dame, Notre Dame, Indiana 46556 USA
Abstract. The ecological effects of spawning runs of native Pacific salmon (Oncorhynchus spp.) on stream
ecosystems of the Pacific Northwest and Northern Pacific Rim have been studied extensively. However,
little is known about how nonnative Pacific salmon affect stream ecosystems in the Great Lakes Basin,
especially given the difference in environmental context between the regions. Mechanisms by which
salmon spawners alter stream ecosystems include nutrient enrichment from excretion by live adults,
carcass decomposition, and physical disturbance of the substrate during redd construction. The objective
of our study was to quantify changes in water chemistry and benthic periphyton in 3 streams in northern
Michigan that have spawning populations of Pacific salmon. In each stream, dissolved nutrients (soluble
reactive P [SRP], NH
4+
, dissolved organic C [DOC], NO
3
2
), and periphyton on gravel were sampled before,
during, and after the spawning run in reaches upstream and downstream of a salmon barrier. Nutrients
increased in reaches downstream of the barrier when salmon were present, but the magnitude of increase
was low relative to increases observed in Pacific Rim streams. During and after the spawning run,
periphyton biomass declined significantly in reaches where high densities of salmon spawners were
present. Our results suggest that disturbance by spawning salmon may override their enrichment effects in
northern Michigan streams, but this pattern may in part be driven by environmental context, especially the
presence of finer substrates.
Key words: nutrient enrichment, resource subsidy, Pacific salmon, periphyton, Great Lakes, stream
ecosystems, water chemistry, disturbance.
Pacific salmon (Oncorhynchus spp.) have inhabited
the Laurentian Great Lakes for nearly 50 y. In 1966
and 1967, coho (Oncorhynchus kisutch) and Chinook
(Oncorhynchus tshawytscha) salmon from Oregon and
Washington hatcheries were introduced to and
became established in Lakes Michigan and Superior
(Crawford 2001). Salmon were first introduced in an
attempt to stabilize Great Lakes food webs after the
introduction of sea lamprey, which negatively affect-
ed the native lake trout (Salvelinus namaycush)
population and led to a strong increase in the alewife
population (Mills et al. 1994). Subsequently, the 2
Pacific salmon species were transplanted to the
remaining Great Lakes. They have since established
naturally reproducing populations (Carl 1982), but
populations continue to be supplemented by exten-
sive annual stocking.
The introduction of Pacific salmon has resulted in
re-establishment of a top predator in the Great Lakes
food web and a popular sport fishery. However,
Pacific salmon are not native to the Great Lakes. Their
effects on native species, food webs, and habitats
within the Great Lakes and connecting waterways are
extensive (Crawford 2001), and many ecological
questions exist regarding their effects on stream
ecosystems in the Great Lakes basin. Pacific salmon
in the Great Lakes rarely reach spawning densities as
high as those in the Northern Pacific Rim (.2
individuals/m
2
; Janetski et al. 2009) during the late
summer and early autumn. However, studies outside
their native range suggest that effects on stream
1
E-mail address: scollins@lssu.edu
2
To whom correspondence should be addressed. E-mail:
amoerke@lssu.edu
3
E-mail addresses: dchalone@nd.edu
4
djanetsk@nd.edu
5
glambert@nd.edu
J. N. Am. Benthol. Soc., 2011, 30(3):831–839
’2011 by The North American Benthological Society
DOI: 10.1899/10-164.1
Published online: 5 July 2011
831

ecosystems similar to those observed in their native
range can occur (Denison and Meier 1979, Schuldt
and Hershey 1995).
Salmon in their native range are an important
vector of marine-derived nutrients to freshwater
ecosystems (Gende et al. 2002, Naiman et al. 2002).
Approximately 95%of the biomass of Pacific salmon
is obtained from the marine ecosystem and is
deposited annually in freshwater streams during
spawning. This subsidy is important for the produc-
tivity of these freshwater ecosystems (Naiman et al.
2002). This material is delivered via multiple path-
ways, such as excretion by living salmon and con-
sumption and decomposition of carcasses, and is
beneficial to organisms (Schindler et al. 2003).
However, significant variation in the extent and
outcomes of resource delivery has been observed
over space (Chaloner et al. 2004) and time (Chaloner
et al. 2007). These differences have been attributed to
variation in environmental context and methods
(Janetski et al. 2009).
One important aspect of the environmental context
is nutrient availability and use by producers and
consumers. Salmon spawners contribute nutrients to
the surrounding ecosystem via 2 main pathways:
1) direct consumption of carcass tissue and eggs by
consumers (terrestrial and aquatic), and 2) recycling
of nutrients through excretion, decomposition, and
leaching (Gende et al. 2002, Naiman et al. 2002). In
southeastern Alaskan streams, both NH
4+
concentra-
tions and soluble reactive P (SRP) in the water column
increase during spawning runs (Chaloner et al. 2004,
Mitchell and Lamberti 2005). Once leached into the
water column, dissolved nutrients can be taken up by
autotrophs and heterotrophs. This fertilization, in
turn, stimulates algal or bacterial growth and has
direct and indirect effects on higher trophic levels,
including invertebrates (Wipfli et al. 1998) and fish
(Wipfli et al. 2004). Nutrient additions may increase
periphyton biomass via resource subsidies, but redd
construction can induce bedload transport, alter
substrate characteristics, and reduce periphyton bio-
mass (Moore et al. 2004, Moore and Schindler 2008).
Environmental factors, such as sediment size, play a
role in the magnitude of disturbance caused by redd
construction, which can position salmon in the role of
ecosystem engineers because of alteration in benthic
processes (Jones et al. 1994).
The effects of anadromous salmonid spawning
have been well studied in their native ecosystems,
such as the Pacific Northwest, but the effects of Pacific
salmon on streams where they have been recently
introduced are unclear. Differences exist between
native and introduced ranges of salmon in terms of
abiotic factors, such as water chemistry, topography,
and geology. For example, Pacific Northwest systems
are often high-gradient, dominated by large substrate,
and nutrient-poor, whereas Great Lakes tributaries
are often low-gradient, dominated by fine sediments,
and subject to increased nutrient loading and reduced
nutrient limitation. Such differences may play a
critical role in how freshwater ecosystems respond
to Pacific salmon subsidies. A better understanding
of how these potential nutrient pulses alter stream
nutrient dynamics is needed. The objective of our
study was to quantify the effects of spawning
salmonids on stream nutrient concentrations and
periphyton biomass in Great Lakes tributaries. We
measured dissolved nutrients and chlorophyll aabove
and below barriers, and before, during, and after
spawning in 3 northern Michigan streams.
Methods
Study sites and design
Three wadeable streams in Michigan’s Upper
Peninsula were selected for this study (Fig. 1).
Streams were selected because they supported natu-
rally spawning populations of Chinook and coho and
were representative of streams in this region. The
study sites were 2
nd
- to 3
rd
-order streams with similar
wetted-channel widths, discharge, and overhead
canopy cover (Table 1). Substrate in Haymeadow
Creek (Lake Michigan basin) consisted mostly of
shale, gravel, and bedrock. Thompson Creek (Lake
Michigan basin) substrate was mostly sandy with
patches of spawning gravel, and Pendills Creek (Lake
Superior basin) substrate was a cobble and gravel
mix. Dams or waterfalls in each stream were
impassible barriers that separated upstream and
downstream locations and allowed comparison of
water chemistry and periphyton biomass with and
without spawning salmon (cf. Chaloner et al. 2004,
2007, Mitchell and Lamberti 2005). An effect of
barriers on nutrient dynamics and periphyton bio-
mass was unlikely in our systems (see Haymeadow
results). In each stream, riffles were haphazardly
selected in one 200-m reach above and one below the
in-stream barriers. Riffles were sampled biweekly
from September through November 2007. This sam-
pling schedule included dates before, during, and
after salmon spawning runs. At each site, water
samples (n=3) were collected for dissolved nutrient
analysis and periphyton samples (n=3) were
collected for chlorophyll aanalysis. Spawning density
in each stream was characterized by counting the
number of spawning fish, carcasses, and redds in the
200-m reaches. Mean masses of Chinook salmon
832 S. F. COLLINS ET AL. [Volume 30

(4.08 kg) and coho salmon (2.26 kg) were used with
direct fish counts to estimate wet mass and fish/m
2
at
the peak run of each stream (Chaloner et al. 2004,
Hubbs and Lagler 2004). Canopy cover was quanti-
fied using a convex spherical densiometer.
Dissolved nutrients
Water samples were collected during autumn 2007,
filtered through 0.45-mm glass-fiber filters (Millipore,
Billerica, Massachusetts), and stored frozen until
analyzed. SRP and NO
3
2
concentrations were mea-
sured on a Lachat QC8500 Flow Injection Autoanalyzer
(Lachat Instruments, Loveland, Colorado) with the
ascorbic acid and Cd-reduction methods (APHA 1995),
respectively. NH
4+
samples were analyzed on a
Genesys 2 spectrophotometer (Thermo Spectronic,
Rochester, New York) with the indophenol blue
method (Aminot et al. 1997, Holmes et al. 1999). DOC
samples were analyzed on a Shimadzu TOC-5000A
(Shimadzu Scientific Instruments, Inc., Colombia,
Maryland; Sharp et al. 1993).
Periphyton
At each site and on each sampling date, 3 replicate
periphyton samples were collected. For each replicate,
5 random pieces of gravel were individually selected
from riffles, placed in a Whirl-PakHbag, and trans-
ported on ice to the laboratory. In the laboratory, the
gravel was scrubbed in deionized water for 2 min/
bag to remove periphyton. Surface area was deter-
mined by covering each rock with aluminum foil,
weighing the foil, and calculating ½of the total rock
surface area based on the area–mass relationship. The
total volume of the periphyton slurry was recorded,
and a subsample of known volume was vacuum-
filtered through a 0.45-mm glass-fiber filter. The filters
were placed in opaque film canisters and stored
frozen until analysis. Chlorophyll a(chl a) was
extracted using 10 mL of 90%buffered acetone held
in canisters for 24 h at 4uC. Chl aand pheophytin were
measured on a Genesys 2 spectrophotometer (Ther-
mo Spectronic, Rochester, New York) according to
Steinman et al. (2006).
Statistical analysis
An upstream–downstream, before–during–after ex-
perimental design (modified BACI; Stewart-Oaten et al.
1986) was used to parse the associated effects of
environmental factors, such as water temperature, dis-
charge, and canopy cover, from effects of the presence/
FIG. 1. Locations of the 3 study streams in northern
Michigan: 1) Haymeadow Creek (Lake Michigan basin),
2) Thompson Creek (Lake Michigan basin), and 3) Pendills
Creek (Lake Superior basin).
TABLE 1. Characteristics of the 3 northern Michigan streams used in our study. Cr. =creek, L. =lake.
Characteristic Pendills Cr. Thompson Cr. Haymeadow Cr.
Great Lakes drainage L. Superior L. Michigan L. Michigan
Mean discharge during study (m
3
/s 61 SE) 0.42 60.04 0.31 60.04 0.42 60.03
Anadromous Pacific salmon species Coho Chinook, coho None observed
Mean channel width (m) 3.2 2.6 3.6
Mean canopy cover 86%75%73%
Median substrate size (mm) 23.5 0.9 NA
Peak count of salmon observed (no. of fish/200 m
stream length)
Alive: 12 Dead: 1 Alive: 224 Dead: 60 None observed
Estimated wet mass at peak salmon run (kg/m
2
)
a
0.04 2.09 0.00
Estimated peak salmon density (fish/m
2
) 0.02 0.54 0.00
a
Mean Chinook =4.082 kg, mean coho =2.267 kg
2011] RESPONSE OF NUTRIENTS AND PERIPHYTON TO SALMON 833

absence of salmon (Janetski et al. 2009). Response
variables included concentrations of dissolved nutrients
(SRP, DOC, NO
3
2
,NH
4+
)andchla(mg/cm
2
). Indepen-
dent factors were location (upstream and downstream of
barriers to fish migration) and time. A repeated-
measures analysis of variance (rmANOVA, a=0.05;
Gotelli and Ellison 2004) was used to establish whether
dissolved nutrient concentrations and periphyton bio-
mass differed between upstream and downstream
locations in each stream over time. Data that violated
ANOVA assumptions were appropriately transformed
(e.g., log[x]-transformation). All analyses were done with
R(version2.8.0;RCenterforStatisticalComputing,
Vienna, Austria). If a significant location 3time
interaction was detected, an a posteriori Tukey’s
Honestly Significant Difference test was used to identify
specific dates when significant differences existed
between upstream (salmon absent) and downstream
(salmon present) reaches.
Results
Salmon spawning runs and site characteristics
Pacific salmon were observed during the study
period in Pendills and Thompson Creek, but not
in Haymeadow Creek (Table 1, Fig. 2). Coho were
observed in Pendills and Thompson Creek, but
Chinook were found only in Thompson Creek.
Thompson Creek had a much higher proportion of
Chinook (86%) than coho (14%) salmon. Salmon were
more abundant in Thompson than Pendills Creek
(at the peak of the run: 224 live salmon/200 m in
Thompson Creek, 12 live salmon/200 m in Pendills
Creek). Wet mass and density of salmon were 52 and
273greater in Thompson than in Pendills Creek,
respectively (Table 1). Canopy cover did not differ
between upstream and downstream reaches at any of
the streams (t-tests, all p.0.05). Water temperature
did not vary .1uC between upstream and down-
stream reaches.
Dissolved nutrients
Dissolved NO
3
2
concentrations were higher in
upstream than in downstream reaches of Thompson
and Pendills Creek, but in the presence of salmon
spawners, NO
3
2
increased in downstream reaches
relative to in upstream reaches (rmANOVA, time 3
site, Thompson Creek: p=0.024, Pendills Creek:
p,0.001; Fig. 3A, B). Prespawning nutrient concen-
trations were not available in Thompson Creek, so our
ability to draw inferences from the data is limited.
However, salmon abundance in Thompson Creek was
low (,25/200 m) when the first nutrient samples
were collected. In Thompson Creek, NO
3
2
concentra-
tions were significantly higher in the downstream
than in the upstream reach on the last sampling date
(after the peak in live salmon and carcasses) (Fig. 3A).
In Pendills Creek, concentrations of NO
3
2
in the
downstream reach were 51%higher during peak
salmon runs than before the runs (Fig. 3B). In
contrast, concentrations in the upstream reach de-
clined by 18%over the same time period. In Pendills
Creek, NO
3
2
concentrations differed significantly
between reaches on all sampling dates, but patterns
differed between 16 September and 28 October
during peak salmon abundance (Fig. 3B). In Hay-
meadow Creek, changes in dissolved NO
3
2
concen-
trations were similar in upstream and downstream
reaches over the study period (rmANOVA, time 3
site, p=0.350; Fig. 3C).
Dissolved NH
4+
concentrations increased in Pen-
dills Creek during the salmon run (rmANOVA, time
3site, p,0.001; Fig. 3E). Over the study period,
downstream NH
4+
concentrations increased .300%,
whereas they decreased 300%in the upstream reach.
In Thompson and Haymeadow Creeks, changes in
concentrations of NH
4+
were similar in upstream and
downstream reaches throughout the salmon run
(rmANOVA, time 3site, Thompson Creek: p=0.067,
Haymeadow Creek: p=0.218; Fig. 3D, F).
SRP increased during the salmon run at Thompson
Creek (rmANOVA, time 3site, p=0.045; Fig. 3G) and
Pendills Creek (rmANOVA, time 3site, p,0.001;
Fig. 3H). In Thompson Creek, downstream concentra-
tions of SRP were 22%higher at the peak of the salmon
run (14 October) than on the first sampling date (,25
fish/200 m). Over the same period, SRP concentrations
FIG. 2. Counts of spawning salmon in downstream
reaches during autumn 2007. No salmon were observed in
Haymeadow Creek. Only 1 dead salmon was observed in
Pendills Creek.
834 S. F. COLLINS ET AL. [Volume 30

in the upstream reach increased only 3%. In Pen-
dills Creek, SRP concentrations increased 31%in the
downstream reach and decreased 19%in the upstream
reach over the study period (Fig. 3H). Changes in SRP
concentrations at Haymeadow Creek were similar in
upstream and downstream reaches over the study
period (rmANOVA, time 3site, p=0.055; Fig. 3I).
In Thompson Creek, DOC concentrations increased
significantly during the salmon run (rmANOVA, time
3site, p=0.020; Fig. 3J). During the peak salmon run,
DOC increased 30%in the downstream reach and
only 18%in the upstream reach. In Haymeadow
Creek, changes in DOC concentrations were similar in
upstream and downstream reaches throughout the
sampling period (rmANOVA, time 3site, p=0.234;
Fig. 3K).
Periphyton biomass
Periphyton chl adecreased significantly during the
salmon run in the downstream reaches of Thompson
Creek (rmANOVA, time 3site, p,0.001; Fig. 4A)
and Pendills Creek (rmANOVA, time 3site, p=
0.017; Fig. 4B). In Thompson Creek, a decline in chl a
was observed at the onset of the spawning run, and
over the entire sampling period, chl adecreased
.8-fold in the downstream reach, whereas it in-
creased .2-fold in the upstream reach. In Pendills
FIG. 3. Mean (61 SE) NO
3
2
(A, B, C), NH
4+
(D, E, F), soluble reactive P (SRP) (G, H, I), and dissolved organic C (DOC) (J, K)
concentrations in Thompson (A, D, G, J), Pendills (B, E, H), and Haymeadow (C, F, I, K) Creeks, Michigan. No DOC data were
available for Pendills Creek. Shaded regions indicate the presence of salmon. Site 3time interactions were evaluated using
repeated measures analysis of variance (rmANOVA, a=0.05). Asterisks indicate a significant difference at specific time periods
(Tukey’s Honestly Significant Difference test, a=0.05).
2011] RESPONSE OF NUTRIENTS AND PERIPHYTON TO SALMON 835

Creek, chl aincreased nearly 2-fold in the upstream
reach, but declined 2-fold in the downstream reach
(Fig. 4B). In Haymeadow Creek, chl adecreased in
both reaches over the sampling period (time, p,
0.001), with no significant difference in the change in
chl abetween upstream and downstream reaches
(rmANOVA, time 3site, p=0.538; Fig. 4C).
Discussion
Ecological effects of salmon on Great Lakes tributaries
Salmon increased concentrations of NH
4+
, NO
3
2
,
and SRP concentrations in Thompson and Pendills
Creeks. The pattern of nutrient responses to salmon
spawners in Great Lakes streams was consistent with
patterns in many previous studies (Janetski et al.
2009), but the magnitude and duration of increases
were lower than in Northern Pacific Rim streams (e.g.,
Mitchell and Lamberti 2005, Chaloner et al. 2007) and
varied markedly among streams. In general, varying
nutrient responses depend on environmental context
(Janetski et al. 2009) including discharge, geomor-
phology, nutrient limitation, size of the spawning
run, and biological communities (e.g., N fixation)
(Hastings 1990, Dent and Grimm 1999).
Background nutrient concentrations in the Great
Lakes tributaries were within or near the range of
concentrations in Alaskan streams (Chaloner et al.
2004, Mitchell and Lamberti 2005), but increases in
NO
3
2
and SRP concentrations (,50%) during the
salmon run were lower than in streams in the
Northern Pacific Rim (Chaloner et al. 2004, Mitchell
and Lamberti 2005) and in a Lake Superior tributary
(Schuldt and Hershey 1995) (.200%). NH
4+
increased
most in response to salmon, but only in Pendills
Creek, which had a relatively small run of salmon.
The density of spawners was highest in Thompson
Creek, but the study reaches were downstream from a
fish hatchery, and any nutrient subsidy provided by
the salmon spawners probably was negligible relative
to the nutrients released from the hatchery. Thus,
local land use may mask or override the enrichment
effects of salmon spawners.
Concentrations of DOC tracked salmon spawner
density, but only in Thompson Creek where DOC
peaked in the downstream reach when salmon
carcass density increased. Sporadic and inconsistent
responses of DOC to salmon have been reported in
other studies (Hood et al. 2007, Janetski et al. 2009).
The inconsistencies might be related to the distur-
bance associated with redd construction, which can
reduce the biomass of the heterotrophic community
and lead to a decrease in uptake of DOC. Alterna-
tively, the heterotrophic community might become
DOC-saturated, such that it is unable to use all of the
available DOC, causing concentrations to increase.
The effects of salmon spawners on DOC concentra-
tions have received little attention and should be
studied further.
The increases in dissolved nutrients we attributed
to Pacific salmon did not stimulate increases in
periphyton biomass, which either remained the same
or decreased over the course of the study. Autotrophs
might not have been nutrient limited in some of our
streams (i.e., Thompson Creek). However, similar
responses have been reported many times for streams
in the Northern Pacific Rim (Chaloner et al. 2004, 2007,
Mitchell and Lamberti 2005). Periphyton responses
FIG. 4. Mean (61 SE) chlorophyll aconcentrations in Thompson (A), Pendills (B), and Haymeadow (C) Creeks, Michigan.
Shaded regions indicate presence of salmon. Site 3time interactions were evaluated using repeated measures analysis of variance
(rmANOVA, a=0.05). Asterisks indicate a significant difference at specific time periods (Tukey’s Honestly Significant Difference
test, a=0.05).
836 S. F. COLLINS ET AL. [Volume 30

appear to be a function of the environmental context
(Janetski et al. 2009). For example, redd construction
may disturb periphyton communities and prevent
them from taking advantage of the nutrient subsidy
associated with the salmon (Peterson and Foote 2000,
Moore et al. 2004, Moore and Schindler 2008). In the
case of our study, the low proportion of quality
spawning gravel in Thompson Creek may have
concentrated disturbances associated with redd con-
struction in the riffles we sampled for periphyton.
However, although the spawning run was smaller and
disturbance was more broadly distributed in Pendills
than in Thompson Creek, periphyton biomass also
decreased in Pendills Creek. This result suggests that
spawning density and substrate availability alone may
not determine a stream’s response to spawning salmon.
Other factors, such as seasonal changes in stream
temperature and insulation (canopy and ice) also
could be important in determining the periphyton
response in Great Lakes tributaries and may help
explain differences between responses in streams
where salmon are native and streams where they
are introduced. Ice became prevalent on all streams as
daylight and temperature decreased near the end of
our study. The ice might have limited periphyton
recovery from disturbance by redd construction, but
such factors are likely to be more important at higher
latitudes where salmon are native, such as the Pacific
Northwest and Northern Pacific Rim. When com-
pared with periphyton responses in streams with
native salmon spawners, the variability in periphyton
responses among Great Lakes tributaries further
emphasizes the importance of considering environ-
mental context when interpreting effects of salmon
spawners.
The role of the Great Lakes environmental context
Environmental factors in the Great Lakes region may
provide a context that predisposes stream communi-
ties to respond more to the disturbance created
by spawning salmon than to the nutrient subsidies
they deliver. In study streams where spawners were
present, dissolved nutrients increased but periphyton
decreased. Other investigators have reported positive
periphyton responses to nutrient enrichment via sal-
mon spawners (Mitchell and Lamberti 2005, Chaloner
et al. 2007). Salmon nutrients stimulated periphyton
growth in manipulative studies of periphyton respons-
es to carcasses and salmon analogs (Wipfli et al. 1998,
2004, Kohler et al. 2008), but the disturbance associated
with salmon spawners was missing from these studies.
Our results are consistent with the suggestion of some
investigators that the disturbance caused by salmon
spawning can control benthic responses to salmon
spawners (Peterson and Foote 2000, Moore et al. 2004).
In particular, sediment particle size might explain why
we observed strong disturbance effects (Janetski et al.
2009). Sediments in Great Lakes tributaries consist of
smaller particles (silt, sand, gravel) than in most Pacific
Northwest streams and are more likely to be agitated
during salmon runs.
Salmon spawner density probably is another factor
that modifies ecological responses in Great Lakes
tributaries. The density of spawning salmon generally
is much higher in the Pacific Northwest and Northern
Pacific Rim than in Great Lakes tributaries, in part,
because of the species composition of the salmon runs.
Alaskan runs are often dominated by high-density
spawning species, such as chum, pink, and sockeye
salmon, whereas Great Lakes runs consist of low-
density spawners, such as coho and Chinook salmon
(Mills et al. 1994, Hubbs and Lagler 2004). However,
salmon densities in Thompson Creek were within the
range of densities observed by Mitchell and Lamberti
(2005) and Minakawa and Gara (2003) in Alaskan and
Pacific Northwestern streams, respectively. The size of
annual runs may vary, but smaller-order Great Lakes
tributaries can receive runs with densities similar to
those in the Pacific Northwest and Alaska. However,
the influence of species composition on nutrient and
benthic responses in stream ecosystems is unclear and
requires further investigation.
Considerations for future study of Pacific salmon subsidies
in the Great Lakes region
Understanding the balance between enrichment
and disturbance effects of Pacific salmon on Great
Lakes tributaries is important for several reasons.
First, Pacific salmon occur in tributaries throughout
the Great Lakes basin and may have variable effects
across the landscape. Second, the ranges of both
native and introduced Pacific salmon encompass
many regions and associated environmental contexts
(e.g., geomorphic, climatic). Thus, spatial variability
in the structure and function of freshwater ecosystems
in which salmon spawn is likely to be enormous. The
variability in ecological responses to salmon is more
likely to reflect the environment than the biology of
salmon, and assessment of those responses might
provide valuable insights about the relationship be-
tween subsidies and recipient ecosystems.
A growing body of literature highlights the ecolog-
ical role of anadromous species, primarily native
species, as providers of an important subsidy or pulse
of nutrients and energy to streams. Pacific salmon are
not native to Great Lakes tributaries, so this subsidy is
2011] RESPONSE OF NUTRIENTS AND PERIPHYTON TO SALMON 837

of ecological interest. The extent to which the pulsed
resource subsidy generated during salmon spawning
runs in Great Lakes tributaries is used and transferred
throughout the stream and associated riparian food
web is unclear. Nonnative organisms can play signif-
icant roles in ecosystem functioning in both aquatic
and terrestrial ecosystems (Crooks 2002). Thus, intro-
duced salmon might cause critical changes in ecosys-
tem structure and function in Great Lakes tributaries
by subsidizing an entire network of streams and
providing alternative food sources. In the Great Lakes
basin, potadromous species (e.g., suckers [Campostoma
spp.] and sturgeon [Acipenser fulvescens]) are part of the
native fauna (Hubbs and Lagler 2004), but whether
native species reach the large spawning densities of
semelparous fishes, such as Pacific salmon, and thus,
have similar associated ecological effects deserves
further attention.
Attention also should be given to how disturbance
associated with salmon spawners influences stream
biotic communities (e.g., bottom-up effects, nutrient
cycling). In our study, disturbance by salmon spawn-
ers decreased periphyton biomass in spawning riffles.
Short-term suppression of basal food resources or
downstream export could have indirect bottom-up
effects on benthic insects and resident fishes. Thus,
positive enrichment effects associated with nonnative
salmon introductions might be balanced by nega-
tive disturbance effects. Both types of effects should
be considered by fisheries managers. Longer-term
studies are necessary to understand the full range
of potential effects, including interannual variation
(Chaloner et al. 2007) of salmon spawners within
Great Lakes tributaries.
Acknowledgements
Data collection and analysis would not have been
possible without the assistance of Drew Afton, Mike
Brueseke, Brandon Gerig, Kate Harriger, and Aaron
Ohrn. This research was funded by the Great Lakes
Fishery Trust (Project 2007.857). Support also was
provided by Lake Superior State University’s Aquatic
Research Laboratory and the University of Notre
Dame’s Center for Environmental and Science Tech-
nology.
Literature Cited
AMINOT, A., D. S. KIRKWOOD,AND R. KEROUEL. 1997.
Determination of ammonium in seawater by the
indophenol-blue method: evaluation of ICES NUTS
I/C 5 questionnaire. Marine Chemistry 56:59–75.
APHA (AMERICAN PUBLIC HEALTH ASSOCIATION). 1995. Stan-
dard methods for the examination of water and
wastewater. 19
th
edition. American Public Health
Association, American Water Works Association, and
Water Environment Federation, Washington, DC.
CARL, L. M. 1982. Natural reproduction of coho salmon and
Chinook salmon in some Michigan streams. North
American Journal of Fisheries Management 4:375–380.
CHALONER, D. T., G. A. LAMBERTI, A. D. CAK, N. L. BLAIR,AND
R. T. EDWARDS. 2007. Inter-annual variation in response
of water chemistry and epilithon to Pacific salmon
spawners in an Alaskan stream. Freshwater Biology 52:
478–490.
CHALONER,D.T.,G.A.LAMBERTI,R.W.MERRITT,N.L.MITCHELL,
P. H. OSTROM,AND M. S. WIPFLI.2004.Variationinresponses
to spawning Pacific salmon among three south-eastern
Alaska streams. Freshwater Biology 49:587–599.
CRAWFORD, S. S. 2001. Salmonine introductions to the
Laurentian Great Lakes: an historical review and
evaluation of ecological effects. Canadian Special
Publication of Fisheries and Aquatic Sciences 132.
CROOKS, J. A. 2002. Characterizing ecosystem-level conse-
quences of biological invasions: the role of ecosystem
engineers. Oikos 97:153–166.
DENISON, D. L., AND P. G. MEIER. 1979. Effects of salmon
spawning activity on macroinvertebrates in a small
Michigan stream. Great Lakes Entomologist 12:57–61.
DENT, C. L., AND N. B. GRIMM. 1999. Spatial heterogeneity of
stream water nutrient concentrations over successional
time. Ecology 80:2283–2298.
GENDE, S. M., R. T. EDWARDS, M. F. WILLSON,AND M. S. WIPFLI.
2002. Pacific salmon in aquatic and terrestrial ecosys-
tems. BioScience 52:917–928.
GOTELLI, N. J., AND A. M. ELLISON. 2004. A primer of
ecological statistics. Sinauer Associates Inc., Sunder-
land, Massachusetts.
HASTINGS, A. 1990. Spatial heterogeneity and ecological
models. Ecology 71:426–428.
HOLMES, R. M., A. AMINOT, R. KEROUEL, B. A. HOOKER,AND B. J.
PETERSON. 1999. A simple and precise method for
measuring ammonium in marine and freshwater eco-
systems. Canadian Journal of Fisheries and Aquatic
Sciences 56:1801–1808.
HOOD, E., J. FELLMAN,AND R. T. EDWARDS. 2007. Salmon
influences on dissolved organic matter in a coastal
temperate brown-water stream: an application of
fluorescence spectroscopy. Limnology and Oceanogra-
phy 52:1580–1587.
HUBBS, C. L., AND K. F. LAGLER. 2004. Fishes of the Great
Lakes region. Revised edition. University of Michigan
Press, Ann Arbor, Michigan.
JANETSKI, D. J., D. T. CHALONER, S. D. TIEGS,AND G. A.
LAMBERTI. 2009. Pacific salmon effects on stream ecosys-
tems: a quantitative synthesis. Oecologia (Berlin) 159:
583–595.
JONES, C. G., J. H. LAWTON,AND M. SHACHAK. 1994. Organisms
as ecosystem engineers. Oikos 69:373–386.
KOHLER, A. E., A. RUGENSKI,AND D. TAKI. 2008. Stream food
web response to a salmon carcass analogue addition in
two central Idaho, U.S.A. streams. Freshwater Biology
53:446–460.
838 S. F. COLLINS ET AL. [Volume 30

MILLS, L. M., J. H. LEACH, J. T. CARLTON,AND C. L. SECOR. 1994.
Exotic species and the integrity of the Great Lakes.
BioScience 44:666–676.
MINAKAWA, N., AND N. I. GARA. 2003. Effects of chum salmon
redd excavation on benthic communities in a stream in
the Pacific Northwest. Transactions of the American
Fisheries Society 132:598–604.
MITCHELL, N. L., AND G. A. LAMBERTI. 2005. Responses in
dissolved nutrients and epilithon abundances to spawn-
ing salmon in southeast Alaska streams. Limnology and
Oceanography 50:217–227.
MOORE, J. W., AND D. E. SCHINDLER. 2008. Biotic disturbance
and benthic community dynamics in salmon-bearing
streams. Journal of Animal Ecology 77:275–284.
MOORE, J. W., D. E. SCHINDLER,AND M. D. SCHEUERELL. 2004.
Disturbance of freshwater habitats by anadromous
salmon in Alaska. Oecologia (Berlin) 139:298–308.
NAIMAN, R. J., R. E. BILBY, D. E. SCHINDLER,AND J. M. HELFIELD.
2002. Pacific salmon, nutrients, and the dynamics of
freshwater and riparian ecosystems. Ecosystems 5:
399–417.
PETERSON, D. P., AND C. J. FOOTE. 2000. Disturbance of small-
stream habitat by spawning sockeye salmon in Alaska.
Transactions of the American Fisheries Society 129:
924–934.
SCHINDLER, D. E., M. D. SCHEUERELL, J. W. MOORE, S. M. GENDE,
T. B. FRANCIS,AND W. J. PALEN. 2003. Pacific salmon and
the ecology of coastal ecosystems. Frontiers in Ecology
and the Environment 1:31–37.
SCHULDT, J. A., AND A. E. HERSHEY. 1995. Effect of salmon
carcass decomposition on Lake Superior tributary
streams. Journal of the North American Benthological
Society 14:259–268.
SHARP, J. H., R. BENNER, L. BENNETT,AND C. A. CARLSON. 1993.
Reevaluation of high-temperature combustion and
chemical oxidation measurements of dissolved organic-
carbon in seawater. Limnology and Oceanography 38:
1774–1782.
STEINMAN, A. D., G. A. LAMBERTI,AND P. R. LEAVITT. 2006.
Biomass and pigments of benthic algae. Pages 357–380
in F. R. Hauer and G. A. Lamberti (editors). Methods in
stream ecology. 2
nd
edition. Elsevier, Amsterdam, The
Netherlands.
STEWART-OATEN, A., W. W. MURDOCH,AND K. R. PARKER. 1986.
Environmental impact assessment: ‘‘pseudoreplication’’
in time? Ecology 67:929–940.
WIPFLI, M. S., J. HUDSON,AND J. CAOUETTE. 1998. Influences of
salmon carcasses on stream productivity: response of
biofilm and benthic macroinvertebrates in southeastern
Alaska, U.S.A. Canadian Journal of Fisheries and
Aquatic Sciences 55:1503–1511.
WIPFLI, M. S., J. HUDSON,AND J. CAOUETTE. 2004. Restoring
productivity of salmon-based food webs: contrasting
effects of salmon carcass and salmon carcass analog
additions of stream-resident salmonids. Transactions of
the American Fisheries Society 133:1440–1454.
Received: 14 December 2010
Accepted: 12 May 2011
2011] RESPONSE OF NUTRIENTS AND PERIPHYTON TO SALMON 839