l4 di‘d‘mlflmm'«f¢:-—;

. \..:\‘:.~... : \T-‘NNH
~ x
u

 

u.

M". M.“ u ‘0:

:3. g;- ..:- ... \:;.f~h
“It 1

 

 

 

 

.W. . «any

 

 

 

 

 

 

 

 

av _.’,‘,"..,
, J, -.:,.--l

E. .
:f‘wuf .
., m . H“,
1. ,’ ,..\.I.

« .,,..,.
,:'_.,.,....“,
"gnaw...
'.'"

 

Huh}.-
",',H.,.,.
,1':'~».‘..,.A.,.,l,—..,I

....

‘( N u v.
m, '17. v; Elly: ¢ 313,7:

,
"Hits: "~
'1‘ “J
,,

 

‘II V .
“main—p,
.-.., nun.
ur- up -

s
‘ HEW...”

W”... ~. .-

:Il'm/ An! “f

 

 

 

 

uallllglll/ljlrgllllljllll

This is to certify that the
thesis entitled
Natural Reproduction of Pactific Salmon

In A Southern Michigan Stream

presented by

Holly Elizabeth Jennings

has been accepted totavards fulfillment
of the requirements for

Master of Science degree in Fisheries and Wildlife

 

 

Major professor

 

Date gammm 314/] /‘7‘6/ 2
U U .

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

LlBRARY
Michigan State
University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE

DATE DUE DATE DUE

 

 

 

m o 9 its?“

 

‘94u1292 ‘

 

h.-‘."+iGiC 2‘:

 

Hikes
.. -M ‘
00 4’13 F338

 

I
j) ‘T
I y“

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution

c:\ckc\dledm.pm3~p.1

NATURAL REPRODUCTION OF PACIFIC SALMON

IN A SOUTHERN MICHIGAN STREAM

BY

Holly Elizabeth Jennings

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Fisheries and Wildlife

1992

ABSTRACT

NATURAL REPRODUCTION OF PACIFIC SALMON
IN A SOUTHERN MICHIGAN STREAM

by
Holly Elizabeth Jennings

During 1989 and 1990 in Prairie Creek, fall spawning by
Pacific salmon was monitored by counting and measuring redds
at predetermined sites. Water velocity, depth and
temperature were measured; substrate type was estimated; and
redd locations were noted. With the use of fyke nets, the
number and timing of outmigrating smolts was observed in
spring 1990. Thirty-two redds were counted in 16 100-m
sites (7% of the stream) during Oct. 3 through Nov. 7, 1989;
the number of redds in the stream was estimated to be 366.

A dam on the creek may inhibit spawner migration. The mean
size redd was 2.3 m in length and 1.4 m in width. No redds
were found in fall 1990. Outmigration of smolts occurred
from 5/28 through 6/30. A total of 66 smolts were captured.
Water temperatures at night during outmigration were mainly

1 6°C and above .

ACKNOWLEDGMENTS

There are many people to whom I extend my deepest
gratitude for all the advice and support they've given me
over the course of my graduate studies. First of all, I
thank the Michigan Agricultural Experiment Station for
financing my research. I would like to thank my advisor,
Dr. Niles Kevern for taking an interest in furthering my
education and for persuading me to enter the masters program
in fisheries. His positive attitude and words of
encouragement always kept me working toward my goal, even
when problems arose and my frustration with my studies
became intense. I also appreciate all the help and
suggestions I received from the other members of my
committee, Dr. Tom Coon and Dr. Scott Witter. They gave me
a new perspective on my research and thesis, and I found
this very refreshing. A special thanks goes to Dr. Scott
Winterstein for helping me design my sampling method and
assisting me with the statistical analysis of my data. He
has no idea how much I appreciate his input.

In addition, I wish to thank my work study students,
Mike Morgan and Joe Patterson for all the help they provided

setting nets and collecting samples. I hope they learned

iii

something from the experience, but also enjoyed it. I also
thank other graduate students who assisted me with data
collection: Glenn Barner, Tim Watkins, Dave Zafft and Rob
Elliot, among others. And to all the graduate students who
participated with me in TGIF's and SAS meetings, thanks for
helping me adjust my attitude and keep my sanity. Former
employers Pam Tyning and Tony Groves have kept me laughing
and prevented me from taking myself too seriously; thanks to
the both of you. Of course, I thank my parents who always
supported my decisions and taught me to be my own person. I
thank Paul Hamelin for his friendship, love and deep
understanding. But most of all I wish to thank Brian Peceny
for his incredible support and consideration these past
othree years. He helped me through some very difficult times
and always reassured me when I was filled with doubt about
my ability to succeed in school and in the professional

world. His friendship and caring will stay with me forever.

iv

TABLE OF CONTENTS

Page
LIST OF TABLES.. ..... . ........ ........ .......... . ........ Vi
LIST OF FIGURES ..... ........... ........... . ............. Vii
INTRODUCTION.............................. ................ 1
DESCRIPTION OF STUDY SITE AND METHODS USED.. .............. 8
RESULTS.... ....... .... .............. .... ................. 22
DISCUSSION
I) Spawning returns: 1989................. .......... 29
II) Spawning returns: 1990............ .............. 31
III) Spawning habitat characteristics ............... 44
IV) Egg survival.......... .......................... 48
V) Fry and smolt migration .......................... 49
CONCLUSION.. ................................... . .......... 56
APPENDIX A .............................................. 59
LITERATURE CITED ........................................ 63

LIST OF TABLES

Table 1. Adult coho and Chinook returning to the
Platte River weir, Michigan, 1979-1990
(Pecor1991)OIOOOOOOOOOOOOOOOOOO... OOOOOOOOOOOOO 5

Table 2. Redd frequency in all sites, Prairie
Creek, Michigan, fall 1989........ ............. 23

Table 3. Monthly high gage levels (in feet) of
the Grand River at Ionia, Michigan during
fall 1985-1989 (R. Moore, NOAA,
unpublished data). ............. ............... 33

Table 4. Monthly high gage levels (in feet) of
the Grand River at Ionia, Michigan during
spring 1986, 1987, 1988, and 1990 (R.
Moore, NCAA, unpublished data). ................ 38

Table 5. Redd dimensions (in meters), Prairie Creek,
Michigan, fall 1989 ............................ 59

vi

Figure 1.

Figure 2.

Figure 3.
Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

LIST OF FIGURES

Location of study area: Prairie Creek,
MiChigan00.0. ..... 0.0.0.000... 00000 0' 00000000000 9

Sites selected for study in Section 1
(shaded areas). Prairie Creek, Michigan,
1989-19900.00.00.00.000...0.00.00.00.000 000000 11

Location of sample sites in Sections 2, 3,
and 4. Prairie Creek, Michigan, 1989-1990 ..... 13

Selection of sites in Sections 2, 3, and
4. Prairie Creek, Michigan, 1989-1990 ........ 14

Design of fyke net used to capture salmon
smolts in Prairie Creek, Michigan, spring
1990... ..... .......... ........................ 18

Placement of single fyke net to capture
salmon smolts in Prairie Creek, Michigan,
late April - early June, 1990......... ........ 19

Placement of two fyke nets to capture
salmon smolts in Prairie Creek, Michigan,
early June - early July, 1990 ................. 20

Migration of Chinook smolts, Prairie

Creek, Michigan, spring 1990. (* No

sampling on these dates; data estimated

from previous and subsequent data points) ..... 25

Morning and evening temperatures during
smolt migration sampling, Prairie Creek,
Michigan, spring 1990.................. ....... 26

Length frequencies of Chinook smolts,
Prairie Creek, Michigan, spring 1990 .......... 27

Redd frequency over a range of water

velocities. Prairie Creek, Michigan,
fall 1989.000...00.00.000.000... 0000000000000 6O

vii

Figure

Figure

Figure

Figure

Figure

12. Redd frequency over a range of water
depths. Prairie Creek, Michigan,
fall 19890000000.000000000000000 00000000000000 6O

13. Sand as percent of total substrate
associated with redds in Prairie Creek,
Michigan, fall 1989..................... ....... 61

14. Fine gravel as percent of total substrate
associated with redds in Prairie Creek,
MiChigan' fall 1989.00.0’0000000000 0000000000000 61

15. Coarse gravel as percent of total substrate
associated with redds in Prairie Creek,
MiChigan’ fall 198900.000...OOOOOOOOOOOO’ 0000000 62

16. Cobble as percent of total substrate

associated with redds in Prairie Creek,
Michigan, fall 1989.. .......................... 62

viii

INTRODUCTION

The first successful introduction of coho salmon
(Oncorhynchus kisutch) into Lake Michigan occurred in 1966
and that of Chinook (Q. tshawytscha) in 1967 (Westers et al.
1990). Since then, yearly stocking of these species has been
conducted by the Michigan Department of Natural Resources
(MDNR), along with other agencies in Indiana, Illinois and
Wisconsin (Westers et al. 1990). In 1988, 2.45 million
juvenile Chinook and 1.7 million coho smolts were stocked
into the Michigan waters of Lake Michigan. Originally, the
salmon were introduced to help control the abundant alewife
(Algga pseudoharengus) population and to provide further
sport fishing opportunity. The alewife is a marine species
that had invaded the Great Lakes probably via the St.
Lawrence River (Christie 1974). The alewife was first
observed in Lake Michigan in 1949 and by 1954, had been
found in all the Great Lakes (Brown 1968; Christie 1974).
Also during this time, the sea lamprey (Petromvzon marinus),
a parasite of large piscivores such as lake trout and
burbot, appeared in the Great Lakes and was observed in Lake
Michigan as early as 1936 (Christie 1974). Christie (1974)
suggests that because the lamprey prefer larger prey, they
were non-lethal parasites until harvesting pressure on the

1

2

larger fish caused the lamprey to attack smaller, less-
resistant fish. Consequently, lake trout populations in the
Great Lakes collapsed due to overharvesting, low
reproduction and predation by the lamprey (Christie 1974).
Additionally, as the larger lake trout disappeared, the
lamprey utilized other fish as prey, namely burbot and lake
Whitefish (Christie 1974). With the demise of these large
predators, the alewife population increased to great
proportions, peaking in Lake Michigan in 1966 (Brown 1968).
A massive die-off occurred in the spring of 1967 , and
alewife carcasses littered the lakeshores (Brown 1968).
Heavy mortalities of alewife in the Great Lakes during
spawning runs in the spring is not unusual; these deaths
have been attributed in part to the inability of the fish to
adjust to severe temperature changes as they travel from
their deep-water winter habitat to shallow spawning areas
(Brown 1968; Christie 1974). However, the sheer numbers of
alewife present in Lake Michigan created a nuisance,
producing the extensive fishkills and clogging intake pipes
during spawning runs (Christie 1974). The addition of
predator species, such as the Pacific salmon, reduced the
alewife populations and alleviated these problems.

Besides controlling forage fish populations, the
Pacific salmon in Lake Michigan have become a valuable
fishery. Between 1843 and 1966, the commercial fishery (for

species such as lake Whitefish, perch, walleye, lake trout

3
and chubs) had a greater impact biologically and
economically than the sport fishery (Keller and Smith
1990a). In fact, Keller and Smith (1990a) assert that many
of the changes in fish composition within the Great Lakes
and the associated decline or extinction of certain
populations were mainly due to poor management and
overharvesting by commercial fleets. In 1966, the MDNR
decided to improve its fisheries resources and to shift its
management focus from commercial to sport fishing; the coho
and the Chinook salmon were to become the base of this new
fishery (Keller and Smith 1990a). By 1986, the number of
Michigan charterboats, the majority of which are on Lake
Michigan, had increased to approximately 1,000 (Jamsen
1990). About $850 million are spent by anglers each year
while pursuing recreational fishing opportunities in
Michigan (Robertson 1990). The MDNR has estimated that the
sport fishery contributes $1.4 billion annually to
Michigan's economy (Robertson 1990). Throughout the 25
years since its inception, this fishery has generated
‘tremendous response from anglers.

After the initial introduction of coho salmon into Lake
Michigan, 17,000 adults were planted by the MDNR in fall
1967 into tributaries of lakes Michigan and Huron (Rybicki
et al. 1990). Since then any natural reproduction that
occurred was thought of as a bonus but was not managed (A.

Hilt, MDNR, personal communication). The MDNR has inhibited

4
spawning of Chinook on certain streams where they were
thought to compete with native salmonids (Rakoczy and Nelson
1990). Otherwise, natural reproduction of Chinook was
neither encouraged nor discouraged and has been documented
to have occurred in Michigan streams as early as 1973
(Rybicki 1973 and Taube 1974). However, few studies
quantifying actual production have been conducted.
Therefore, agencies are repeatedly stocking salmon into the
Great Lakes with little knowledge of natural recruitment.
Continuous hatchery-rearing and stocking of fish can be a
very expensive endeavor and may not provide the diversity
within populations that natural reproduction can. But the
ability of Michigan streams to support extensive salmon
natural reproduction is questionable.

Studies of Pacific salmon in northern lower Michigan
conducted by Carl (1982) in Baldwin Creek and by Seelbach
(1985) in the Little Manistee River report the occurrence of
successful spawning. Natural reproduction data for southern
lower Michigan streams, though, is severely lacking.

According to Rakoczy and Nelson (1990), the average
size of coho and Chinook harvested and the number returning
to spawn have declined in recent years. These lower returns
can be seen in recent data from the Platte River, Michigan
(Table 1). Patriarche (1980) reported a significant decrease
in mean lengths and weights of adult coho during his two-

year study (1978 and 1979). Female Lake Michigan coho seen

 

 

5

Table 1. Adult coho and Chinook returning to the Platte
River weir, Michigan, 1979-1990 (Pecor 1991).

 

XEAB gong CHINOOK
1979 36,404 4,702
1980 123,113 4,435
1981 168,049 3,563
1982 129,363 2,999
1983 156,358 6,114
1984 142,102 5,924
1985 80,354 4,865
1986 52,770 5,147
1987 55,144 7,787
1988 26,118 4,646
1989 49,793 1,899

1990 32,821 1,761

 

6
in fall 1979 had an average size smaller than female Pacific
Coast coho (Patriarche 1980). The MDNR hypothesized that
this drop in size and numbers returning was due to a
dwindling forage base, and it has decreased stocking efforts
in order to restore the prey populations and promote better
utilization of prey by individual fish (Rakoczy and NelSon
1990). Apparently, this reduction in stocking has worked.
The year class of 1984, the year of peak stocking, has
produced below average returns, whereas the 1985 and 1986
year classes, stocked in much lower numbers, have produced
above average returns (Rakoczy and Nelson 1990).

As stated previously, the Pacific salmon returns in the
Great Lakes have been fluctuating recently, as well as the
forage base on which they depend. Also, few studies of
natural reproduction of these salmon have been conducted in
southern lower Michigan. Therefore, the goal of this study
was to quantify this production in a particular southern
Michigan stream and provide these data as further evidence
of the present condition of these salmon populations.

The Grand River, located in southern lower Michigan, is
among those tributaries of Lake Michigan that are
consistently stocked with Chinook and coho salmon. Natural
reproduction probably occurs in this river but has not been
described or quantified to any great extent. To study
reproduction throughout the entire Grand River system seemed

impractical; therefore, this study concentrated on one

7
tributary of the Grand where salmon have spawned. Prairie
Creek was selected for this study for four main reasons:
early visual observations indicated potential spawning
habitat; anglers have observed salmon spawning in the creek
for years past; the Creek is assumed to be fairly
representative of 10 to 20 other Grand River tributaries
that have good water quality and suitable gravel and that
support similar fish assemblages; and the creek is easily
accessible.

The proposed objectives for this study included:
estimating the number of Pacific salmon nests, or redds,
produced in Prairie Creek during fall 1989 and 1990;
describing spawning habitat characteristics associated with
each redd; and estimating the recruitment of juvenile salmon
from Prairie Creek to the Grand River during spring 1990.

By accomplishing these objectives, I hoped to measure the
extent of Pacific salmon natural reproduction in this creek
and compare the measured habitat variables to data from
other studies conducted in Michigan and on the Pacific

coast.

8

DESCRIPTION OF STUDY SITE AND METHODS USED

Prairie Creek is located in Ionia County, in central
lower Michigan (Figure 1). It is approximately 25 km long
and no more than 23 m at its widest point during low water.
It flows south, entering the Grand River just southeast of
the town of Ionia. The watershed of Prairie Creek consists
of gently to steeply sloped hills with well—drained loamy
soils overlaying sand and gravel, where agriculture is the
main land use (Threlkeld 1967). The creek is fairly well-
buffered by a strip of upland forest containing hardwoods
such as oak, maple and beech; some reaches of the creek,
though, are immediately adjacent to residential and
industrial property in the downstream section, and to some
cropland and pastures farther upstream.

Prairie Creek is a marginal trout stream because it
retains fairly high temperatures throughout the summer in
some sections (C. Bay, MDNR, unpublished data; R. Moore,
National Oceanic and Atmosperic Administration, unpublished
data; Trimberger 1988; A. Hilt, MDNR, personal
communication). However, it does have better water quality
than many of the other marginal trout streams tributary to
the Grand River (Seelbach, 1991, MDNR, interoffice
communication). The river contains resident populations of
brown trout and juvenile steelhead (A. Hilt, MDNR, personal

communication). Rotenone treatments have been administered

 

   
 
  
 
  

STUDY SECTION
NUMBER

DAM

STUDY SECTION
BOUNDARY

 

 

:Z

 

 

 

 

 

 

 

 

 

 

 

 

MILES
1 2

I I
I I I

1 2 3
KILOMETEFIS

O—-—-°

 

 

 

Figure 1. Location of study area: Prairie Creek, Michigan.

10
in Prairie Creek on three separate occasions: once in 1966,
again in 1975 and recently in 1991 (A. Hilt, MDNR, personal
communication). The brown trout and steelhead have been
stocked into the creek periodically since 1966, especially
after each chemical treatment (A. Hilt, MDNR, personal
communication).

A dam with a fish ladder is located less than a mile
upstream from the mouth. The dam, located at its present
site for many years past, was refurbished and elevated in
1972 to prevent the passage of "rough fish" (i.e. non-game
species thought to compete with the salmonids) from the
Grand River (Weaver, 1981, MDNR, interoffice communication).
In this way, the MDNR hoped to reduce the number of chemical
treatments needed to control these unwanted species (Weaver,
1981, MDNR, interoffice communication). Spawning Chinook
and coho salmon have been observed upstream of this site;
therefore, they have been able to utilize the fish ladder or
swim over the dam.

To choose sites for fall sampling, I divided the river
into four sections (Figure 1). The smaller tributaries of
Prairie Creek were not included in this study because often
they were very narrow, intermittent or had a high proportion
of silt. I assumed spawning would not occur in these due to
insufficient habitat.

The southernmost section is situated between a dam and

the mouth of the stream (Figure 2). The first 50 m

11

.ommaummma .cmmflnoflz
.xwmuo ofluflmum .Ammmum Umcmnmv H cofluoom :fl >csum now couomaom wmufim .N wusvflm

AI >>O.._u_ E<mmhw

Ezcrmmnmmemmp

 

7:505.

 

 

 

 

 

 

 

 

_>_<D

12
downstream of the dam were ignored in case the bridge or
angling activity may somehow affect spawning habitat or
behavior. The remaining 1200-m stretch was divided into
twelve 100-m sites. Four of these sites were then chosen
randomly.

Each of the other three sections contained four bridges
(Figure 3). Sampling occurred near bridges because of ease
of access. The first 50 m on each side of each bridge were
ignored for the same reason as with the first section, to
avoid any influence of the bridges on spawning behavior or
habitat (Figure 4). The 300 m beyond that on each side were
divided into 100-m sites, six in all. One site was then
randomly selected.. Therefore, each of the four sections
contained four sites, sixteen sites in all, and these 1600 m
sampled totaled approximately 7% of the entire study area.

I investigated each site by wading in the stream, and
the sites were observed many times over the spawning period
in the fall of 1989 and 1990. All redds were located and
counted. Usually coho and Chinook redds could not be
differentiated from one another. An estimate of total
number of redds was made assuming a normal distribution.
Total redds in each section were calculated by multiplying
the mean density of sampled redds in a section (per 100 m)
by the total length of that section. The totals within all

sections were then summed. The variance of the estimate of

13

 

M-66

 

 

 

 

 

 

 

 

 

I M-20
IONIA ~

 

 

 

 

 

 

 

M-20 .31
M-66
MILES I SAMPLE SITE
0 1 2
I I I] #1 A DAM
o . 2 3 STUDY SECTION
KILOMEIERS — BOUNDARY

 

 

Figure 3. Location of sample sites in Sections 2,3, and 4.

Prairie Creek, Michigan, 1989-1990.

14

.Ommalmwmfl
.cmoflcoflz .xmmuu owufloum .v can .m .m mcofluoom :a woven mo cofluowamw .v ousmflm

All >>O.E S_<mm._.m

 

 

 

 

 

 

 

 

 

 

 

 

E omn E omN E amp E om E om E omw E omN E 0mm

NOD—mm

 

15

total redds within each section was calculated using the
following equation:

[(length of section)2/(length of site)2] * var u
where var u = the variance of the mean density of redds in a
section (based on sums of squares). These variances of.
total redds in each section were summed to produce a
variance of the total redds in the stream. The standard
deviation calculated from this variance was then multiplied
by the appropriate Student's t value to produce a 90%
confidence interval with 15 degrees of freedom.

A chi-square test was performed to determine if any
significant differences existed between number of redds
above the dam and number below the dam (the data from the
sections above the dam were pooled). The maximum length
and the width across the depression of each redd were
measured and averages were calculated. In addition,
measurements of streamflow and water depth were taken over
the streambed just upstream of each redd and over the ridge
and depression of each redd. The streamflow in m/s was
measured at the streambed and at 0.6 of the depth from the
bottom using a Marsh-McBirney digital meter. The actual
depth of each depression within the existing streambed was
found by calculating the difference between the water depth
over the streambed and the water depth over the depression;
these depths were averaged for all redds. Air and water

temperatures in degrees Celsius were measured with a

 

16
hand—held thermometer at the time of observations. The
proportion of each substrate size within each redd was
estimated using four main size classes: sand, which is .25
to 3 mm in diameter; fine gravel, 4 mm to 4 cm; coarse
gravel, 4.1 cm to 8 cm; and cobble, 8.1 cm to 30 cm.

Seven redds were excavated in winter 1990 to determine
egg presence and viability. The method was similar to those
described by Hobbs (1937) and by Braum (1978). The device
used to sample the eggs consisted of a 1 m x 1.3 m
rectangular frame of PVC pipe to which fine-meshed netting
was attached, and this was held perpendicular to the current
downstream of the redd. Starting at the downstream edge of
the redd and moving towards the ridge of the tailspill, I
shovelled the sediment to a depth of approximately 0.5 to 1
m and deposited it about 0.5 m upstream of the net. The
larger particles were allowed to settle back down to the
streambed while the finer particles were carried by the
current into the net. Disturbed eggs caught in the current
were also trapped in the net. The live eggs were identified
by their translucent, orange appearance; the dead eggs were
opaque and white.

During spring 1990, two fyke nets were used to block
the river and catch smolts moving downstream. Fyke nets
were used because they were inexpensive, easy to handle and
readily available. And other investigators (Davis et al.

1980; Milner and Smith 1985) have found that fyke nets are

17
effective for capturing smolts. The two nets in this study
were installed about 300 m upstream of the dam. The net
openings were 1.2 m x 1.5 m oblong frames made of wood and
were straight on one side so that they would rest flush with
the streambed when set (Figure 5). Attached to each frame
was a funnel approximately 3 m in length with three inner
funnels. Only one net was set between April 27 and June 5,
1990 in the area of greatest flow (Figure 6). Because the
net had only one long lead that could extend to the
streambank, a seine was used to block the other side of the
river. From June 6 through July 2, the funnels of the two
nets were situated next to each other perpendicular to the
current and again in the water of highest velocity (Figure
7). The lead-nets were attached to these and ran upstream
toward each shore at a 43’angle from the streambank,
blocking off the entire stream. Mesh size of all netting
used was 1.25 cm. Steel reinforcing rods with diameters of
0.6 cm and 1.25 cm were used to anchor all parts of these
nets.

The nets were set from four to six times a week, mainly
at night between 9:00 pm and 9:00 am because no smolts were
found when they were set during the day. Each morning
during sampling, the nets were checked for smolts. Any
smolts found were counted and measured for length. Stream

temperatures were measured on a Ryan thermograph and on a

18

3m

 

 

 

 

 

RODS

1.2m

 

 

‘L5ln
LEAD
FRONT NET
VIEW

Figure 5. Design of fyke net used to capture salmon smolts
in Prairie Creek, Michigan, spring 1990.

Z

 

 

 

STREAMFLOW
V v
45°
SEINE RODS
I I
I I
17 m

Figure 6. Placement of single fyke net to capture salmon
smolts in Prairie Creek, Michigan, late
April - early June, 1990.

 

Figure 7.

20

STREAM FLOW

 

 

I v
45°
...
\JO 5’
45 5%;
{’3'
4'
J
41"
I J
r” I
17th

Placement of two fyke nets to capture salmon

smolts in Prairie Creek, Michigan, early

June - early July, 1990.

Z

21
hand-held thermometer. The time of migration was noted. A
weighted mean migration date (weighted by the number of
smolts migrating per day) was calculated by assigning a
consecutive number for each date during the migration
period; for example, the number 614 corresponded with June
14, the number 615 with June 15, etc. Multiplying each
number by the number of smolts caught on that particular
day, summing over all the days, and dividing this result by
the total number of smolts gave the weighted mean migration

date. A normal distribution was assumed.

 

22

RESULTS

Between October 3 and November 7, 1989, a total of 32
redds were counted (Table 2). After this date, no new redds
were created. All fish seen associated with the redds were
positively identified as coho and Chinook. The total number
of redds in Prairie Creek with a 90% confidence interval was
estimated to be 366.65 i 398.75. This confidence interval
may seem very wide and is mainly due to the small sample
size, which produced a very high variance. The estimate of
total redds produced a density of approximately 14 redds per
km. The number of redds above the dam significantly
differed from the number below the dam (p < 0.05), based on
densities of 1.5 redds per 100-m site above and 3.5 per 100-
m site below. Many anglers were present, especially below
the dam, during this sampling.

The lengths of all redds measured ranged from 0.9 to
6.5 m with a mean length of 2.3 m i 0.49. Widths ranged
from 0.5 to 4.7 m with a mean width of 1.5 m i 0.36. And
redd depths ranged from 0.04 to 0.20 m with a mean depth of
0.1 m i 0.02. The mean depth of the streambed just upstream
of the redds was 0.375 m. The stream velocities at redd

locations ranged from 0.19 to 0.88 m/s. The mean velocity

 

23

Table 2. Redd frequency in all sites, Prairie Creek,
Michigan, fall 1989.

 

SITE SECTION
1 Z l A
A 0 0 3 O
B 2 3 3 O
C 9 1 0 O
D 3 2 2 4
TOTAL 14 6 8 4
MEAN 3.5 1.5 2 1

 

was 0.57 m/s : 0.064. The above estimates all include 95%
confidence intervals. The water temperature ranged from
6.5°C to 12°C during the spawning interval. Seventy percent
of redds sampled had fine gravel as the main substrate.
Almost 17% had equal amounts of fine gravel and coarse
gravel. Sand was abundant in only five redds and comprised
from 20 to 35% of the observed substrate.

Of the redds sampled for eggs, three of the seven had
live eggs, one contained only dead eggs and the remaining
three were apparently devoid of eggs. The diameter of 40
eggs was measured; egg sizes ranged from 6 to 7 mm. Anchor
ice was common over many of these redds. Two fry incubated
and hatched from eggs collected in Prairie Creek were

identified as Chinook.

 

 

24

Eight fry were caught between April 30 and May 31,
1990. Their lengths ranged from 35 to 55 mm. Outmigration
of Chinook smolts occurred from May 28 to July 1, 1990, with
peak migration occurring in mid-June (Figure 8). Data for
days not sampled were estimated based on data from several
days before and after those dates. No coho smolts were
caught. The weighted mean migration date was June 14. A
total of 66 smolts was captured. They were identified by
the loss of parr marks, a silvery appearance and a
streamlined body shape (Hoar 1976). With one exception,
water temperatures at night during smolt outmigration were
119C or above (Figure 9). Lengths of smolts were 48-95 mm,
with a mean of 76.2 mm 1 3.11. More smolts were in the 80-
89 mm size class than in any other particular size class
(Figure 10). No pattern was found linking lengths and
timing of migration; smolts of all sizes were present
throughout the migration period.

During high flows in spring 1990, drift would rapidly
collect on the lead nets, creating greater force on the
nets. This in turn bent the rods to the point that the
current was flowing over the float line of the leads. Many
fry may have escaped in this current. By the time the
smolts began migration, water levels were quite low, and
this was not a problem.

Only two redds were seen during fall 1990, one in

Section 1 and one in Section 4. Two fish carcasses were

A.mucflom dump ucmsvmwndm
can mSOH>ouQ Eouu coumfiflumo mumc “mount omega :0 @cflHmEmm 02 av
.omma mcflumm .cmmflnoflz .xomuo mfluwmum .muHOEm xoocflco Mo coaumuwflz

m._.<n_
an? QO ONE , mto _ 6:0 Em Em

__-___F___ o
4* 1

 

* Vw

In:

 

mw

.w musmflm

SI'IOWS :IO HSEWRN

SNNDET

 

26

-“ nm‘“mm“‘~
-“ mm‘“‘“mmmV

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

z
9
p.
‘<
I:
52
JE
. .
I l
I
I
‘ W
' x
I
I
00'
m.
put
I
. \
I I
I
I
m. t
.m
I
I
I
I
I
I
I
. I'"
l \\X\\\‘\—
[LXKK\\'\\\\\\\\\\LA
I I I I I
m c II: o In a
N N 1- 1-

(00) aunlvuadwal

6/20 6/30

6/10

5/25

5/15

5/5

6/25

5/30 6/5 6/15

DATE

5/20

5/10

4/30

EVENING TEMP

VIA

- MORNING TEMP

1990.

Creek, Michigan, spring

1rie

Morning and evening temperatures during smolt migration sampling,

Pra

Figure 9.

.04

27

.ommH
vcflumm .cwmwnofiz .xmmno mfluflmnm .muHOEm xoocflno mo mmflocmsvmum numcmq

 

 

1885:5sz
832 . 8-8 8-8 8-8 8-8 6.0

_ _ O
I m
4
I 2
I B
I om

 

.oa wudmflm

Sl'IOWS :IO HEIBWIIN

28
seen in Section 3, and these were identified as Chinook. At
the beginning of sampling, water levels were higher than the
previous year, and the water was very turbid. If spawning
did occur, it was difficult, if not impossible, to observe.
When the water levels eventually receded and turbidity

decreased, little evidence of spawning was present.

29

DISCUSSION

I) Spawning Returns: 1989.

Spawning of coho and chinook salmon was observed in
Prairie Creek in 1989. The estimate of 366 total redds
produces a density of approximately 14 redds per kilometer.
The mean numbers of redds per 100 m site in Prairie Creek
(3.5 below the dam and 1.5 above the dam) were much lower
than those described by Taube (1974) who found averages of
15.26, 14.98 and 13.34 coho redds per 100 m in 1969, 1970
and 1971, respectively, in a 2279.9 m stretch of the Platte
River in Michigan. However, the Platte River is
considerably larger both in discharge and cross—sectional
area.

Migration in Prairie Creek during 1989 began in late
September. By the beginning of November, spawning had
ceased. Taube (1974) noted that coho spawning on the Platte
River had ended by November 18, November 19 and November 2
in 1969, 1970 and 1971, respectively. During 1977—1978,
coho spawning peaked in early November in Bigelow Creek and
in late November in Pine Creek (Carl 1983), and chinook
spawning in Baldwin and Pine Creeks during those years

occurred from the end of September through mid-November,

30
with maximum spawning in mid- to late October (Carl 1984).
Studies in tributaries of Lake Ontario revealed that coho
and chinook moved into the creeks to spawn from late
September through early October (Keleher et al. 1985) and
that peak spawning of chinook occurred earlier than that of
coho (Johnson and Ringler 1981). Burner (1951) found
Columbia River fall chinook spawned from September through
mid-November. In British Columbia, Big Qualicum River
chinook spawned from mid-October to early November, peaking
in late October, and coho spawning peaked during the first
half of December (Lister and Genoe 1970). Peak spawning of
the coho and chinook in Prairie Creek could not be
determined; when counting redds, I often did not see fish
associated with them.

It is interesting to note the significant difference
between numbers of spawners above the dam and those below
the dam. Possibly the dam is a definite barrier to these
fish. Banks (1969) refers to many studies showing that
light is needed by anadromous salmon to pass obstructions
when they would otherwise choose darker or more turbid
conditions when migrating upstream. Also, salmonids may
have a range of flows in which they can swim over dams.
Possibly the flows during fall 1989 were low enough to
impede some of the spawners. It should be noted that 30% of
the area below the dam was sampled; only 5% of the area

above the dam was studied. Areas above the dam with high

31
densities of redds could have been overlooked due to the

sampling method.

II) Spawning Returns: 1990.

Results for 1990 were very different than those in
1989. Only two redds were counted, and two carcasses were
the only salmon observed. They were not found with the
redds but may have been carried downstream. Many factors
may have influenced these results: floods during fall when
these fish were deposited as eggs inhibiting spawners from
upstream migration or washing out the eggs; heavy harvesting
of those spawners preventing them from depositing eggs;
anchor ice slowing intragravel flow and decreasing the
amount of oxygen transported to eggs; floods in early spring
washing the hatching and emerging fry downstream, causing
higher mortality; high temperatures in Prairie Creek
limiting survival of coho fry during their first summer
after hatching; intense predation on the fry or smolts that
made it to the Grand River; predation, harvesting and
disease, among other things increasing mortality of these
fish as they grew in Lake Michigan; high angling pressure on
spawners on the lower Grand River; and high flows and
turbidity in Prairie Creek during the fall of 1990 impeding
their upstream migration.

Of course, all these factors apply assuming the

returning salmon are wild. If these salmon were stocked

32
fish that had strayed into Prairie Creek, those factors
mentioned that limit the egg-to-presmolt phase would not
apply. Widespread straying of coho and chinook within the
Great Lakes is a well-known phenomenon (Rybicki 1973; Scott
and Crossman 1973; Patriarche 1980; Wenger et al. 1984).
Wenger et al. (1984) found that 51.4% of returning radio-
tagged coho and chinook in Lake Erie strayed to tributaries
other than those into which they were originally released.
Evidently, Great Lakes salmon are much more likely to stray
than Pacific Coast populations (Scott and Crossman 1973;
Wenger et a1. 1984). And the last year that chinook were
stocked in the Grand River upstream of Prairie Creek was
1984; those spawners returning in fall 1989 may have been
the last remnants of that population that happened to stray
into Prairie Creek (A. Hilt, MDNR, personal communication).
Therefore, it is possible that many of the salmon returning
to Prairie Creek in 1989 were of hatchery origin and were
mixed with wild fish.

Those naturally-produced chinook that would have
returned to Prairie Creek to spawn in 1990 were probably
progeny of fish that Spawned in fall 1985 or 1986. Although
no spawners were sampled for age in this study, it is likely
that they were mainly ages 0.3 (zero years in the stream and
three years in the lake) and 0.4, as reported by Hay and
Houghton (1990). According to Rybicki et al. (1990), coho

in the Great Lakes have a three-year life cycle and spawn

33
when they are age 1.1. Therefore, the spawning coho of 1990
were most likely deposited as eggs in fall 1987.

The precipitation data for Ionia indicates slightly
higher than normal rainfall in fall 1985 (R. Moore, NOAA,
unpublished data). Heavy precipitation in southern Michigan
in fall 1986 produced record high (loo-year) floods (Rakoczy
and Nelson 1990), and gage readings on the Grand River at
Ionia were twice as high as normal in September and October
of that year (Table 3). Fall 1987 saw no unusually high
water levels; they were similar to those of fall 1989.
Although no actual flow data are available for 1985 and
1986, the abnormally high water levels may have impeded the

upstream migration of spawning salmon, especially in 1986.

Table 3. Monthly high gage levels (in feet) of the Grand
River at Ionia, Michigan during fall 1985-1989 (R.
Moore, NOAA, unpublished data).

 

 

 

 

MONTH 1353

1285 1299 1291 1988 1989
Sept. 14.96 23.22 9.27 13.36 8.75
Oct. 11.89 22.97 8.79 8.60 15.41
Nov. 17.40 11.07 12.10 12.64 *
MEAN 14.75 19.09 10.05 11.53 12.08

 

* Data

not available.

 

 

34

Although Banks (1969) reviewed many studies that show
no correlation between discharge and migration, he concluded
that the rate of flow is probably the main factor
influencing migration. Light intensity and temperature may
also influence migration when coupled with discharge.
Huntsman (1948) reported that Atlantic salmon began to
ascend a stream as a freshet started, with the majority
ascending just after peak discharge as water levels fell.
However, Hunter (1959) stated that chum and pink salmon had
a maximum flow above which migration would be impeded, and
Wenger et al. (1984) stated heavy rainfall that produced
unusually high water levels inhibited the migration of coho
and chinook in tributaries of Lake Erie. This may have been
the case during the record flows of fall 1986. The water
levels of fall 1985, though, were about average and probably
did not impede migration of spawners.

Stream levels that are too low can also inhibit
migration, at least until the salmon are very mature, at
which time they will migrate regardless of streamflow (Banks
1969). However, in a study by Keleher et al. (1985),
chinook and coho failed to move far upstream, possibly due
to low discharge. It is possible that the flows during fall
1987 were low enough to inhibit the passage of salmon over
the dam and would have produced similar results to those

found in 1989. If this is the case, suitable spawning

 

35
habitat above the dam, if present, would not have been fully
utilized, and therefore less production would have resulted.

High stream flows in 1986 may have not only prevented
spawner migration, but also washed out redds and the eggs
contained therein before they are covered by the female
(Braum 1978; Quinn and Tallman 1987). If not washed out,
the eggs could have been damaged if gravel in the redd was
shifted by the strong flows; in addition, these flows could
have deposited silt on the redds, suffocating the eggs
(Wickett 1958; Foerster 1968).

Harvesting of spawners is probably a major factor
limiting natural production of salmon in Prairie Creek. The
MDNR has determined that the Grand River produced over
34,000 angler hours of salmonid fishing in 1986 (Rakoczy and
Rogers 1987) and over 81,000 in 1987 (Rakoczy and Rogers
1988). On the Kalamazoo River, Michigan, almost 22,000
angler hours were fished between September and November,
1985 (Rakoczy and Lockwood 1988). Within the tributaries of
Lake Ontario studied by Keleher et al. (1985), 78 per cent
of the returning tagged salmon were taken by angling and
snagging. Carl (1982) reported heavy harvesting of spawning
chinook on the Manistee and AuSable rivers, with females
taken off their redds. He presumed that this led to lower
egg deposition thereby limiting production in these streams.
This is likely the case with Prairie Creek. Many anglers

were observed there in fall 1989, and the creek has been

 

36
very popular for salmon fishing in years past (A. Hilt,
MDNR, personal communication). No angler-use data have been
collected on Prairie Creek, though.

Of those eggs actually deposited in falls 1985, 1986
and 1987 and not washed out by high flows, some-may have
been susceptible to mortality during the winter months.
Anchor ice, observed over redds in winter 1989/1990,
probably was present in earlier years and may have decreased
the intragravel flow of water, important for bringing oxygen
to eggs and washing away waste products (Vaux 1968).
However, even if flows were significantly diminished, the
water probably contained sufficient oxygen to sustain the
eggs; those eggs found during winter 1989/1990 were alive
and appeared healthy.

Silver et al. (1963) studied the effect of different
water velocities and oxygen concentrations on the
development of chinook eggs and found that changes in
oxygen had more effect on timing of hatching and on size of
embryo than changes in velocity. In the above study, at
dissolved oxygen (D.O.) concentrations below 2.5 mg/l, no
hatching occurred. At 2.5 mg/l or above, the hatching rate
was high, although post-hatching survival was lower at lower
concentrations. Increasing velocity tended to increase
post-hatching survival, but with high D.O., not as high a
velocity is needed to transport sufficient D.O. to the

developing embryos. Embryos reared at a high D.O. were

 

37
larger than those reared at a low D.O., and embryos reared
at a high velocity were larger than those in low velocity.
Differences in embryo size were more pronounced with changes
in D.O. concentrations than with changes in velocity.
Therefore, even if anchor ice lowered intragravel water
velocity in the redds, the very cold water was probably
saturated with oxygen and provided more than enough D.O. to
produce high egg and post-hatching survival.

Another source of mortality includes flooding in the
early spring that may displace the emerging fry and wash
them out of the stream. Extensive flooding of Prairie Creek
in the spring is not unusual. High water levels were
observed in spring 1989 and 1990. Gage height readings
taken in the Grand River at Ionia during April and May of
1986 and 1987 show water levels approximating average annual
levels (Table 4). March of 1986 and March through early
April, 1988, however, had much higher than average levels,
and March 1986 readings were 75% higher than those of the
same month in 1987 and 1988.

According to Scott and Crossman (1973), coho salmon
emerge from the gravel during March and April. Ottaway and
Clarke (1981) determined that salmon fry are more vulnerable
to flow at younger stages, and they agree with McDonald
(1960) that upon losing visual contact with the substrate
and other stationary objects at night, salmon may lose

positioning in the stream. The emerging fry in Prairie

38

 

Table 4. Monthly high gage levels (in feet) of the Grand
River at Ionia, Michigan during spring 1986, 1987,
1988, and 1990 (R. Moore, NOAA, unpublished data).
MONTH YEAR
1986 1987 1988 1990
March 21.62 12.38 13.90 20.74
April 12.96 10.83 18.51 13.30
May 12.71 9.98 9.81 13.41
June * * 8.13 10.09
MEAN 15.76 11.06 12.59 14.39

 

* Data not available

Creek could easily have been displaced by the swift current,
especially in 1986, and carried downstream, possibly to the
Grand River. These fry might have then encountered less
available food, higher predation and overall harsher
conditions, all leading to higher mortality.

Many predators of salmonid fry and smolts inhabit the
Grand River and its tributaries, including the steelhead (Q.
mykiss), brown trout (galmg trutta), coho yearlings and
largemouth bass (Micropterus salmoides) found in Prairie
Creek during August 1986 (C. Bays, MDNR, unpublished data)
and during spring 1990 when trapped in my nets. Other
possible predators I found at that time were northern pike
(Egg; lucius), smallmouth bass (Migrgptgrus gglgmigui) and

crayfish (Families Astacidae and Cambaridae). These

39

predators, among others, may have caused extensive mortality
on the fry and smolts in the riverine habitat and during the
first year of lake residence. Predation has been postulated
to be a major limiting factor on juvenile Atlantic salmon
(MacCrimmon 1954) and the main determinant of juvenile to
smolt survival (Symons 1979). Larsson (1985) reported that
predation produced 35 to 70% mortality of a smolt run.

Because chinook smolt and migrate to the lake the same
season they hatch, they are not affected by the stream water
temperatures during the warmer periods of the year. Coho,
however, spend their first summer and winter in the stream
before smolting the following spring. According to Brett
(1952) , coho prefer temperatures between 12 and 14°C with an
upper lethal limit of 25.0TL The highest temperature I
recorded during June 1990 was 23%5 which is near that upper
limit. It is likely that Prairie Creek reaches or even
exceeds 29%:during the summer months. No water temperature
data were available for Prairie Creek during these months
for any year, except that collected in August 1986 (C. Bays,
MDNR, unpublished data). During this time, no temperatures
over 20°C were reported. These measurements were few,
though, and were collected over a short two—day sampling
period. If water temperatures do rise above this 25%:
limit, if only for a few days, coho production would be
severely limited. Reduced growth and increased mortality

could result. However, coho juveniles in Prairie Creek can

40
escape this thermal stress by moving to pools or migrating
farther upstream. Apparently, the brown trout and steelhead
(other species with restrictive upper lethal temperature
limits) that are stocked in the creek have been able to
withstand the summer water temperatures. Perhaps these coho
juveniles can also survive them.

Competition with brown trout, steelhead or even chinook
juveniles could also push the coho young out of Prairie
Creek early. Consequences of this, mentioned previously,
include increased predation and harsher conditions.

Definite size classes between coho and steelhead juveniles,
though, may cause them to choose different food sources.
While coho fry hatch in March or April, steelhead do not
hatch until late May or June. And steelhead that had
hatched the previous year would be much larger than those
coho fry. Stein et al. (1972) found little interaction
between trout and salmon in Sixes River. Rainbow
(steelhead) fry inhabit faster water while coho fry prefer
the slower water of pools and stream margins (Hartman 1965;
Johnson and Ringler 1980). Bustard and Narver (1975) agree
with this consensus; they also found that during winter,
steelhead fry occupy rubble areas, whereas coho fry prefer
roots, logs and bank cover areas. These studies did not
relate this to time of emergence and growth of fry. But

Carl (1983) states that competition is present between coho

41
and steelhead and between coho and chinook and that it is
minimized behaviorally through spatial segregation.

According to Lister and Genoe (1970), coho and chinook
juveniles tend to be spatially segregated because the
chinook hatch earlier and are always larger. ,In another
study, though, coho emerged two weeks before chinook (Stein
et al. 1972). Nevertheless, they were spatially segregated;
coho preferred upstream areas and tributaries while chinook
more often occupied the main river channel (Stein et al.
1972).

Populations of these salmonids in Prairie Creek are
probably too low to induce much competition. And what
little competition exists probably forces some individuals
to migrate. Steelhead coevolved with coho and chinook on
the Pacific Coast, and behavioral differences have evolved
to minimize competition between these species (Rybicki et
a1. 1990). Everest and Chapman (1972) found little
competition between chinook and steelhead due to disparate
times of emergence. Additionally, coho-brown trout
competition in the Great Lakes has been found to be fairly
insignificant (Rybicki et al. 1990).

Once the coho and chinook reach Lake Michigan, they
begin to grow at a rapid rate (Scott and Crossman 1973).
During this time (eighteen months for coho and 3.5 to 4.5
years for chinook), predation, harvesting and diseases such

as bacterial kidney disease can cause increased fatalities.

 

42
Mortality of young salmon in the lake is probably caused
mainly by predation; as the salmon grow, harvesting becomes
a more important mortality factor. Nevertheless, the
percentage of any year class taken by harvesting is very low
(Rybicki et al. 1990). Patriarche (1980) estimated that
only 7.4% of the entire 1977 coho year class stocked in Lake
Michigan was taken by the sport fishery.

Bacterial kidney disease (BKD) has become an important
issue lately with agencies stocking coho and chinook in the
Great Lakes. BKD occurs naturally in salmonids but usually
is activated only by some environmental stress. The
incidence of BKD in coho and chinook populations has
increased in recent years, and a large die-off of chinook
occurred in Lake Michigan during spring 1988 (Keller and
Smith 1990b). Due to an environmental stress, these fish
had much lower overall resistance and were stricken by BKD
as well as other infections and stresses (Keller and Smith
1990b). The year classes of salmon that spawned in 1989 and
1990 Were more than likely affected by this disease.

Although these mortalities of coho and chinook in Lake
Michigan can be significant, it is generally hypothesized
that the egg-to-smolt stages of salmonids undergo the
heaviest mortality. For example, Foerster (1954) showed
that high mortality of sockeye salmon occurs before seaward
migration. Upon review of many studies, Braum (1978) also

concludes that mortality among freshwater fishes is highest

 

43
in the first year of life. Furthermore, Kocik and Taylor
(1987) assert that factors limiting year-class strength
affect pink salmon during the early part of their life while
in the stream environment. Basically, "...it is the
riverine ecosystem that plays the principal role in
determining the recruitment of oncorhynchid populations"
(Kocik and Taylor 1987).

If juvenile survival of those salmon spawning in 1990
had been high and effects of mortality factors in Lake
Michigan were not significant, those fish may have
encountered intense harvesting in the lower Grand River
before they reached Prairie Creek. This is unlikely; the
higher water levels and turbid conditions would lessen
angler efficiency (Rakoczy and Nelson 1990). I may have
missed the spawning altogether, if it did not occur during
the same period as in 1989 (early October through early
November). But spawners were observed climbing the dams on
the Grand River during this time, although in greatly
reduced numbers (A. Hilt, MDNR, personal communication).
The higher flows and turbid conditions probably did not
inhibit spawners in 1990. The flows were not unusually
high, and, as stated previously, salmon prefer increased
turbidity during migration (Banks 1969). Again, this higher
turbidity may prevent passage of spawners over the dam, but
no salmon were seen below the dam, either. Therefore, it

appears that those year-classes of salmon that should have

 

44
spawned in 1990 were very weak and that this poor spawner
turnout was caused primarily by factors that affected these

fish when they were juveniles.

III) Spawning Habitat Characteristics

Measurements of redd size and certain habitat
characteristics associated with each redd were taken in fall
1989 in order to compare with those of other studies and
also to discern the preferences of these salmon. The same
measurements were planned for fall 1990, but because only
two redds were observed and flows were very high and
difficult to wade, I decided not to proceed with them.

Length measurements of all redds observed in the study
sites averaged 2.3 m i 0.49, and widths averaged 1.5 m i
0.36. Therefore, these redds were probably less than 3.45
If mean area. Fall spawning Columbia River chinook were
observed to produce an average redd size of 5.1 n?, which is
quite a bit larger than those found in Prairie Creek (Burner
1951). In the same study, coho redds averaged 2.84 n?. A
study of sockeye salmon in the Taku River, British Columbia
reported redd sizes with a mean width of 1.7 m, a mean
length of 2.3 m and a mean area of 4.0 n? (Lorenz and Eiler
1989). The redds produced in Prairie Creek may have been
produced mainly by coho; this would explain the smaller redd

size. However, chinook in Prairie creek are probably

45
smaller than those in the Columbia River and would build
smaller redds.

In Prairie Creek, the mean depth of the water just
upstream of the redds was approximately 0.38 m, and mean
depth of each redd depression within the streambed was 0.1 m
i 0.02. Smith (1973) reported an average depth at the
streambed of 0.39 m for fall chinook and 0.22 m for coho in
Oregon. Chinook redds in Kalama and Toutle Rivers,
Washington had mean depths of 0.37 and 0.31 m, respectively,
and coho redds in the Toutle had a mean depth of 0.18 m
(Burner 1951). Those chinook redds observed in a California
study had a mean depth of 0.34 m, while the coho redds had a
0.15 m mean depth (Briggs 1953). Overall, the depths of
redds in Prairie Creek were similar to those reported for
chinook redds in these earlier studies.

In my study, water velocity was measured at 0.6 of the
depth from the bottom and at the bottom of the water column
over the ridge and the depression of each redd and over the
streambed immediately upstream of each redd. The mean
velocity over the streambed at 0.6 depth was 0.57 m/s i
0.064. Smith (1973) also measured velocities just upstream
of redds but always took measurements at 0.12 m from the
bottom. Mean velocities were found to be 0.43 m/s for
chinook and 0.44 m/s for coho (Smith 1973). Burner (1951)
reported a surface water mean velocity of 0.61 m/s for

chinook, and Briggs (1953) found mean surface velocities

 

46
over chinook and coho redds of 0.61 and 0.58 m/s,
respectively. Mean velocities over chinook redds in two
Michigan streams were 0.42 m/s in Baldwin Creek and 0.50 m/s
in Pine Creek (Carl 1984). Again, the data for Prairie
Creek are comparable to those already reported.

Water temperature may be an important factor
influencing salmon spawning. Snyder and Blahm (1971),
contend that upstream migration of salmon is blocked by
temperatures equal to or exceeding 20-2FTL However, Banks
(1969) cites many conflicting studies of the relationship
between temperature and migration. For example, Hayes
(1953) found temperature did not trigger Atlantic salmon
runs; on the other hand, Stuart (1953) concluded that brown
trout would not migrate until the water temperature had
fallen to 6 or Tkthr the first time each fall. Banks
(1969) feels streamflow is the main factor influencing
migration and that temperature, as well as other factors,
may be secondary factors.

During salmon spawning in fall 1989, temperatures in
Prairie Creek ranged from 6.5 to 12°C, well below the 20-21°C
upper limit proposed by Snyder and Blahm (1971). Stream
temperatures in a study of Lake Erie coho and chinook never
rose above 13miand did not seem to affect migration. When
coho and chinook began upstream migration into certain
tributaries of Lake Ontario, stream temperatures were

already below 19-20T!(Keleher et al. 1985). Lorenz and

 

 

47
Eiler (1989) measured much colder stream temperatures (4.5—
6.UTO within sockeye salmon redds in British Columbia and
Alaska. These redds were probably influenced more by
groundwater and springs than by runoff, an important factor
in Michigan streams.

The majority of redds sampled in Prairie Creek (70%)
had fine gravel (4 mm to 4 cm diameter) as the main
substrate, and another 16-17% had equal amounts of fine and
coarse gravel (4.1 to 8 cm diameter). Only five redds had
an appreciable amount of sand (0.25 to 3 mm diameter),
ranging from 20 to 35% of the observed substrate. In the
Entiat, a tributary of the Columbia River, summer chinook
avoided gravel with silt and clay binders, probably due to
insufficient water flow through the gravel (Burner 1951).
Lorenz and Eiler (1989) found that the probability of use of
habitat by sockeye spawners was greatest when the substrate
was less than 15% fine sediment (in their study, less than 2
mm diameter).

As stated previously, redds containing eggs need
sufficient intragravel flow to transport dissolved oxygen to
the eggs and carry away their waste products; excessive silt
and sand can block this flow. According to Lorenz and Eiler
(1989), sockeye salmon chose areas with upwelling
groundwater because those areas provide loose, uncompacted
substrate and a more constant water temperature, both of

which aid incubation of eggs. They concluded that substrate

 

48
composition and surface water velocity are not as vital to
survival of eggs and fry if spawning areas are not dependent
on surface water to provide flow through the gravel.
Although no data are available concerning the extent of
upwelling groundwater in Prairie Creek, it is assumed to be
of less importance than surface water velocity and substrate
composition. Apparently the salmon chose spawning areas that
had low proportions of sand and silt and that were

subsequently less embedded.

IV) Egg Survival

Seven of the redds counted in fall 1989 were sampled
for eggs during early winter 1989-1990. No eggs could be
found in three of those redds. This does not necessarily
mean that no eggs were present; rather they may have been
overlooked due to problems with the sampling method.
However, the absence of eggs may signify heavy fishing
pressure in which females are pulled off redds before
depositing eggs. Or it may denote egg retention, in which
the female does not release all her eggs when spawning.
Coho adults transferred to the Platte River Michigan in
1969—1971 were estimated to have between 44 and 67% egg
retention which is abnormally high compared to West Coast
rivers (Taube 1974). Kocik and Taylor (1987) feel egg
retention is caused mainly by high spawner density. Taube

(1974) offered handling as another possible cause and

 

49
assumes there are other causes that have not been
identified. No salmon were examined in fall 1989 to measure
egg retention, but because spawner density was so low in
Prairie Creek, any egg retention was probably due to some
other stress factor.

The presence of live eggs and their apparent hearty
condition, even when in redds covered by a thick layer of
anchor ice, supports the theory that dissolved oxygen
concentrations and the amount of intragravel flow within
Prairie Creek are suitable for the incubation of salmon
eggs. The eggs found were bright orange and ranged from 6
to 7 mm in diameter with a mean of 6.45 mm i 0.819. This
range coincides exactly with sizes of chinook eggs reported
by Scott and Crossman (1973). They are also quite similar
to coho eggs sampled in Lake Ontario tributaries (6.6 to 7.1
mm) and somewhat larger than those found on the Pacific

Coast (4.5 to 6.0 mm) (Scott and Crossman 1973).

V) Fry and Smolt Migration

Between April 30 and May 31, 1990, eight salmonid fry
were sampled. Most likely, at night these fry lost visual
contact with the substrate and were displaced by rising
flows which were occurring at this time according to R.
Moore (NOAA, unpublished data). The peak of the chinook fry
migration in a British Columbia river took place in early

April and that of coho fry migration in early May (Lister

 

50
and Genoe 1970). This agrees with data from a tributary of
Lake Ontario showing chinook emerging earlier than coho
(Johnson and Ringler 1981).. In another study, however, coho
emerged before chinook in an Oregon stream (Stein et al.
1972). Carl (1984) found chinook fry drifted during early
April through mid-June, 1978 and 1979 in Baldwin Creek and
during May, 1978 in Pine Creek. More than 98% of all
drifting fry sampled in both creeks were caught between
dusk and dawn. Also, Carl (1984) estimated that of the
emergent chinook fry produced in Baldwin Creek during 1978
and 1979, 25-30% drifted each year.

Lengths of Prairie Creek fry were 38-55 mm. Lister and
Genoe (1970) reported coho fry between 38.4 and 41.6 mm and
chinook fry between 42.9 and 69.3 mm moving downstream.
Carl's study (1983) of Pine Creek, Michigan during 1977-1979
revealed mean coho underyearling lengths between 40.7 and
62.0 mm in May and June. In another study by Carl (1984),
involving both Pine and Baldwin Creeks, Michigan during the
same time period, chinook fry were 44-60 mm mean length.
Those fry caught in Prairie Creek were coho or chinook or
both; they were not identified to species. Nonetheless, the
majority of fry drift of both species probably occurred
before the net was set in late April.

Fry that resist displacement downstream continue to
grow until they reach the smolt stage. At this time, these

fish undergo physiological changes that prepare them for

 

 

51
life in the ocean, and they proceed to migrate out of their
streams (Bley 1987; Hoar 1976; Rybicki et al. 1990; Scott
and Crossman 1973). They become thinner, lose their parr
marks and appear silvery and more streamlined (Bley 1987;
Hoar 1976). During spring 1990, 66 chinook smolts were
captured in Prairie Creek. This is an estimate of the total
number of smolts above the dam. The recruitment of smolts
from the area below the dam is unknown.

Smolts were caught between May 28 and July 1, 1990,
with 97% migrating between June 8 and June 24. The peak of
coho smolt migration occurred in late May in the Big
Qualicum River, British Columbia (Lister and Genoe 1970).
Coho smolts in a Michigan stream were reported to migrate in
April (Carl 1983), whereas the chinook smolts migrated
mainly in June (Carl 1983, 1984). The coho in a study by
Seelbach (1985) smolted in mid-May, and the chinook smolted
in June.

The weighted mean migration date of smolts in Prairie
Creek was June 14. Calculated mean migration dates for coho
in the Little Manistee River were May 15 in 1982, May 16 in
1983 and May 21 in 1984 ; chinook mean migration dates were
June 13 in 1982, June 19 in 1983 and June 20 1984 (Seelbach
1985). The results for chinook smolt migration from Prairie
Creek correspond well with these other Michigan studies.

No coho smolts were caught in 1990. They may have

migrated in April, before the net was installed. Or they

 

52
may have migrated during sampling and somehow avoided the
net. Davis et al. (1980) found that coho smolts up to 150
mm fork length were collected in fyke nets at higher
velocities (z 0.7 m/s), but at lower velocities, the net
seemed to select smaller fish. In early May 1990,
velocities in Prairie Creek were quite high, and although
this may have decreased the efficiency of the net, migrating
coho smolts would not have avoided it as well. Later in the
spring, the entire stream cross-section was sampled
effectively during the lower water levels. Coho smolts,
though large, could not have avoided the nets at this time.

Movement of smolts in 1990 occurred mainly at night,
which corresponds well with the findings of Carl (1984) in
his study of chinook in Pine Creek, Michigan. Migration of
Atlantic salmon has also been reported to occur at night
(Thorpe and Morgan 1978) as well as at dusk and dawn (Fried
et al. 1978).

Lengths of the chinook smolts ranged from 48 to 95 mm
with a mean of 76.2 i 3.11. This is similar to chinook
smolts in the Little Manistee River, Michigan which range
from 70 to 90 mm (Seelbach 1985) and to those in the
Sacramento-San Joaquin Estuary, California, which are 70-80
mm (Kjelson and Raquel 1984). Fessler and Wagner (1969)
propose that the winter steelhead parr-smolt transformation

is dependent on size. Refstie et al. (1977) also found that

 

53
the timing of Atlantic salmon smolting is also based on size
and not on age.

During the nights that smolts were caught, temperatures
were mainly 16°C or above. Increases in temperature at the
time smolts began migrating might indicate that temperature
plays a role in triggering smolting and migration. Other
factors, such as stream flow and photoperiod, may also
affect the onset of migration, alone or in conjunction with
temperature. White (1936), Fried et al. (1977) and Solomon
(1978) all agree that the main stimulus for Atlantic salmon
smolting is increasing water temperature, with migration
peaks occurring when water is 10°C or more. Becker (1973)
correlated migration of young chinook with an increase in
flow followed by a decrease with corresponding warmer
temperatures. Kwain (1983) determined that rainbow trout
smolting was caused by lower stream flow and rising water
temperatures. Hoar (1976) noted water temperature and
photoperiod, together with size of fish, triggered smolting,
and photoperiod had a greater influence than temperature.

However, other studies declare that water temperature
does not induce smolting but does affect the length of time
of migration (Zaugg and Wagner 1973; Wagner 1974). In
another study, chinook smolts moved out of streams only
during high water after heavy rains (Carl 1984). Seelbach
(1985) found that daylight length alone corresponded with

peak migration dates of coho and chinook smolts. Daily mean

54
water temperatures in his study were 7°C or higher during
migration, but no relationship was established between
temperature and smolt movement. Also, peak migrations were
not associated with the few fluctuations in flow that
occurred (Seelbach 1985). Finally, Bjorrn (1971) noted that
migration of smolts "...frequently coincided with changes in
water temperature and streamflow..." However, Bjorrn
"...could not establish a consistent causal relationship and
concluded that photoperiod and perhaps growth must initiate
the physiological and behavioral change associated with
seaward migration." The stimuli of chinook smolting and
migration in Prairie Creek might very well be photoperiod
and growth, but this could not be confirmed.

Very few smolts were captured in spring 1990. This
could be due to debris clogging the opening of the net(s)
during high water. It could also be due to the fact that
only one net was used to block the river for a majority of
the sampling period and during most of the periods of high
flow and was not as efficient at capturing the fish as two
nets. A total of 66 smolts was captured on Prairie Creek,
and because the nets were effectively sampling the entire
river at this time, that total is probably a good estimate
of total chinook smolt production above the dam. However,
total production of the stream's 1990 chinook year class and
its supplement to the Lake Michigan salmon fishery are

unknown.

 

 

 

55

Although many redds were seen during fall 1989, not all
of them may have been productive and yielded smolts the
following spring (due to heavy harvesting, egg retention,
mortality of eggs and fry, etc.). Wright (1981) feels that
spawning counts are good for assessing past management
practices. In his study of Baldwin and Pine creeks, Carl
(1984) found that "number of smolts produced in both streams
appeared to be independent of the adult density the previous
spawning season." Foerster (1954) determined that 60% of
the variation in sockeye spawner returns were a function of
smolt size and numbers together, and he and McLemore et al.
(1989) agree that the numbers of returning spawners can be
predicted by the actual numbers of outmigrating smolts.
Therefore, this evidence suggests that for estimating a
stream's total production of an anadromous salmonid, it may
be more beneficial to enumerate smolt yields rather than
spawner returns and that the latter can be used to evaluate

previous management strategies.

56

CONCLUSION

In summary, natural reproduction of coho and chinook
was found to occur in Prairie Creek in fall 1989.
Apparently, good spawning habitat is available in Prairie
Creek. However, the dam could be a barrier to utilization
of upstream spawning habitat and could limit salmon
production. The lack of reproduction in fall 1990 was
probably due more to the poor year-classes of the spawners,
the absence of hatchery fish returns, and the overall
decline in salmon populations in the Great Lakes than to
inadequate spawning habitat conditions.

Previous studies and actual conditions on Prairie Creek
during this time period suggest that highest mortality of
the wild salmon probably occurred in their early life
history, through the time of smolt migration. Rearing
habitat, especially for coho fry in the summer, may be
limiting. However, it is very likely that a majority of the
fry are washed downstream during increased spring flows and
sustain considerable mortality.

Timing of spawner migration, redd sizes and habitat
characteristics such as substrate size, water temperature
and water depth and velocity in Prairie Creek were

comparable to other studies of salmonids. Of the redds

 

57
sampled in winter 1989/1990, some contained live, apparently
healthy eggs, while others had no eggs at all. The eggs may
have been overlooked in these redds due to crude sampling
methods. The lack of eggs, however, may be indicative of
the occurrence of heavy angling pressure on spawners before
the females can deposit their eggs. A few chinook smolts
were caught during spring 1990, indicating that at least
some of the spawning habitat was satisfactory for the
incubation of eggs and the rearing of fry. Again, most of
the fry had probably been displaced earlier in the spring,
before the sampling period.

Salmon production in Prairie Creek and in the Great
Lakes are probably limited by a number of factors working in
conjunction. This trend of decreasing production may just
be a temporary situation in a fluctuating population. But
because the salmon sport fishery is so important
economically, it behooves us to closely monitor these
fluctuations and also the forage base supporting the salmon.
Any dramatic decrease of forage fish populations may
necessitate stronger management initiatives, such as heavier
restrictions of the commercial harvest of these species.

If Prairie Creek were to be managed for natural
reproduction, angling pressure on spawners should be
monitored and possibly restricted to ensure egg deposition.
Also, the dam and fish ladder should be checked to assess

the ease with which salmon ascend them during upstream

 

58
migration. Dams such as this have been used in the past for
trout management by preventing passage of non—game species
that may compete with salmonids. Other dams on streams in
Michigan are used for the production of hydroelectric power.
Still others are presently not used for any known purpose.
Managers and the public at large must decide if these dams
are serving their needs and are cost effective. If people
decide that their streams should be managed for salmon, it
may be necessary to dismantle the dams for easier passage.
Or if the dams are serving a purpose, but salmon
reproduction in the stream is desired, fish passage
structures such as fish ladders should be installed to allow
upstream migration. The key is access by spawners to
upstream areas where habitat may be suitable.

In order to evaluate and quantify production in the
stream, it would probably be better to sample smolt
outmigration rather than redd density, as number of smolts
has been found to be independent of number of redds. The
results of this study suggest that sufficient spawning
habitat may be present in many of these southern Michigan
marginal trout streams and that they may be managed for

salmon production.

 

APPENDIX A. Supplemental data.

59

APPENDIX A. Supplemental data.

Table 5. Redd dimensions (in meters), Prairie Creek,
Michigan, fall 1989.

 

 

 

REDD SECTION LENGTH WIDTH DEPTH
l 1 2.0 1.3 0.07
2 3.4 1.5 0.13
3 1.4 0.8 0.20
4 1.4 1.2 0.12
5 1.9 1.0 0.14
6 3.7 2.1 0.16
7 3.3 1.7 0.10
8 6.5 4.7 0.17
9 2.6 3.0 0.10

10 1.6 1.3 0.04
11 2.8 1.3 0.18
12 2.8 1.4

13 2.3 1.3 0.11
14 1.5 1.1 0.11
15 2 5.1 2.5

16 1.7 1.1 0.12
17 1.6 1.0 0.03
18 5.1 3.2 0.10
19 2.2 1.7 0.10
20 3.2 1.5 0.07
21 3 2.4 2.7 0.05
22 0.9 0.5

23 1.8 1.1 0.15
24 2.4 1.1 0.08
25 1.6 0.8 0.10
26 4.0 3.8 0.17
27 0.9 0.7 0.07
28 1.1 0.7 0.04
29 4 1.2 0.9 0.04
30 1.8 0.8 0.06
31 1.0 0.6 0.06
32 1.3 0.6 0.06

 

 

APPENDIX A. Supplemental data.

Figure 11.

NUMBER OF REDDS

Figure 12.

NUMBER OF REDDS

 

0-0.18 0.26-0.32 0.40-0.46 0.54-0.60 0.08-0.74 0.82-0.08
0.19-0.25 0.33-0.39 0.47-0.53 0.61-0.67 0.75-0.81 >030

VELOCITY (m/s)

Redd frequency over a range of water velocities.
Prairie Creek, Michigan, fall 1989.

 

0-0.19 02600.31 0.38-0.43 0.50-0.55 0.62-0.67 0.74-0.79
0.20-0.25 0.32-0.37 0.44-0.49 0.56-0.61 0.68-0.73 >0.79

WATER DEPTH (m)

Redd frequency over a range of water depths.
Prairie Creek, Michigan, fall 1989.

APPENDIX A. Supplemental data.

NUMBER OF REDDS

Figure 13.

NUMBER OF REDDS

Figure 14.

25-

20—

 

 

No 11-20 ”'30 31-40 “'50 51-60 61'" 71-80 “1'90 91-100

% SAND

Sand as per cent of total substrate associated
with redds in Prairie Creek, Michigan, fall 1989.

 

°'1° 11-20 ”'30 31-40 “'50 51-60 61'" 71-80 a1"9091-100

% FINE GRAVEL

Fine gravel as per cent total substrate
associated with redds in Prairie Creek,
Michigan, fall 1989.

APPENDIX A. Supplemental data.
12

10-

NUMBER OF REDDS
OI
I

 

 

I I I I
0-1 0 21 -30 41 -50 61 -70 81 -90
11-20 31 -40 51 -60 71 -80 91-100
% COARSE GRAVEL

Figure 15. Coarse gravel as per cent of total substrate
associated with redds in Prairie Creek,
Michigan, fall 1989.
25

20-
15—

10-

NUMBER OF REDDS

 

 

I I l

l ‘- I I I
0-1 0 21 -30 41 -50 61 -70 81 -90
11-20 31 -40 51 -60 71-80 91 -1 00
% COBBLE

Figure 16. Cobble as per cent total substrate associated
with redds in Prairie Creek, Michigan, fall 1989.

 

 

LITERATURE CITED

63

LITERATURE CITED

Banks, J.W. 1969. A review of the literature on the upstream
migration of adult salmonids. J. Fish Biol. 1:85-136.

Bays, C. and B. Ewing. Fish collection data, Prairie Creek,
Michigan. Unpublished data, 1986. Mich. Dep. Nat.
Resour., Fish. Div.

Becker, C.D. 1973. Food and growth parameters of juvenile
chinook salmon, (Oncorhynchus tshawytscha), in
central Columbia River. Fish. Bull. U.S. 71:387—401.

Bjornn, T.C. 1971. Trout and salmon movements in two Idaho
streams as related to temperature, food, streamflow,
cover, and population density. Trans. Am. Fish. Soc.
100(3):423-438.

Bley, P.W. 1987. Age, growth, and mortality of juvenile
Atlantic salmon in streams: a review. U.S. Fish Wildl.
Ser., Biol. Rep. 87(4). 25 pp.

Braum, E. 1978. Part II. The eggs and larval phase. Pages
7-47 in Methods for assessment of fish production in
fresh water. T. Bagenal, ed. Blackwell Scientific
Publications, London.

Brett, J.R. 1952. Temperature tolerance in young Pacific
salmon, genus Oncorhynchus. J. Fish. Res. Board Can.
9:265-323.

Briggs, J.C. 1953. The behavior and reproduction of salmonid
fishes in a coastal stream. Calif. Dep. Fish Game, Fish
Bull. 94. 62 pp.

Brown, E.H. 1968. Population characteristics and physical
condition of alewives, Alosa pseudoharengus, in a
massive die-off in Lake Michigan, 1967. Great Lakes
Fishery Commission Tech. Rep. 13. 20 pp.

 

Burner, C.J. 1951. Characteristics of spawning nests of
Columbia River salmon. U.S. Fish. Wildl. Serv., Fish.
Bull. 61:97-110.

64

Bustard, D.R. and D.W. Narver. 1975. Aspects of the winter
ecology of juvenile coho salmon (Oncorynchus kisutch)
and steelhead trout (Salmo gairdneri). J. Fish. Res.
Board Can. 32(5):667-680.

Carl, L.M. 1982. Natural reproduction of coho salmon and
chinook salmon in some Michigan streams. North Am. J.
Fish. Manage. 4:375-380.

Carl, L.M. 1983. Density, growth, and change in density of
coho salmon and rainbow trout in three Lake Michigan
tributaries. Can. J. Zool. 61:1120-1127.

Carl, L.M. 1984. Chinook salmon (Oncorhynchus tshawytscha)
density, growth, mortality, and movement in two Lake
Michigan tributaries. Can. J. Zool. 62:65-71.

Christie, W.J. 1974. Changes in the fish species composition
of the Great Lakes. J. Fish. Res. Board Can.
31:827-854.

Davis, S.K., J.L. Congleton, and R.W. Tyler. 1980. Modified
fyke net for the capture and retention of salmon smolts
in large rivers. Prog. Fish-Cult. 42:235-237.

Everest, F.H. and D.W. Chapman. 1972. Habitat selection and
spatial interaction by juvenile chinook salmon and
steelhead trout in two Idaho streams. J. Fish. Res.
Board Can. 29:91-100.

Fessler, J.L. and H.H. Wagner. 1969. Some morphological and
biochemical changes in steelhead trout during the
parr-smolt transformation. J. Fish. Res. Board Can.
26:2823-2841. -

Foerster, R.E. 1954. On the relation of adult sockeye salmon
(Oncorhynchus nerka) returns to known smolt seaward
migrations. J. Fish. Res. Board Can. 11(4):339-350.

 

Fried, S.M., J.D. McCleave, and G.W. LaBar. 1978. Seaward
migration of hatchery-reared Atlantic salmon, Salmo
salar, smolts in the Penobscot River estuary,

Maine: riverine movements. J. Fish. Res. Board Can.
35:76-87.

Hartman, G.F. 1965. The role of behavior in the ecology and
interaction of underyearling coho salmon (Oncorhynchus
kisutch) and steelhead trout (Salmo gairdneri). J.
Fish. Res. Board Can. 22:1035-1079.

 

 

65

Hay, R. and W. Houghton. 1990. Stream management. Pages
46-55 in Review of salmon and trout management in Lake
Michigan. M. Keller, K.D. Smith, and R.W. Rybicki, eds.
Mich. Dep. Nat. Resour., Fish. Special Rep. 14. 254

PP.

Hayes, F.R. 1953. Artificial freshets and other factors
controlling the ascent and population of Atlantic
salmon in the LeHave River N.S. Fish. Res. Board Can.
Bull. 99. 47 pp.

Hoar, W.S. 1976. Smolt transformation: evolution, behavior,
and physiology. J. Fish. Res. Board Can. 33:1234-1252.

Hobbs, D.F. 1937. Natural reproduction of quinnot salmon,
brown and rainbow trout in certain New Zealand waters.
New Zealand Mar. Dep. Fish. Bull. 6:1-35.

Hunter, J.G. 1959. Survival and production of pink and chum
salmon in a coastal stream. J. Fish. Res. Board Can.
16:835-886.

Huntsman, A.G. 1948. Freshets and fish. Trans. Am. Fish.
SOC. 75:257-266.

Jamsen, G. 1990. Economics. Pages 195-209 in Review of
salmon and trout management in Lake Michigan. M.
Keller, K.D. Smith, and R.W. Rybicki, eds. Mich. Dep.
Nat. Resour., Fish. Special Rep. 14. 254 pp.

Johnson, J.H. and N.H. Ringler. 1981. Natural reproduction
and juvenile ecology of Pacific salmon and rainbow
trout in tributaries of the Salmon River, New York.
N.Y. Fish and Game J. 28(1):49-60.

Keleher, C.J., J.M. Haynes, D.C. Nettles, R.A. Olson and
J.D. Winter. 1985. Fall movements of Pacific salmon in
Lake Ontario and several tributaries. N.Y. Fish and
Game J. 32(2):167-175.

Keller, M. and K.D. Smith. 1990a. Introduction. Pages 1-13
in Review of salmon and trout management in Lake
Michigan. M. Keller, K.D. Smith, and R.W. Rybicki, eds.
Mich. Dep. Nat. Resour., Fish. Special Rep. 14. 254

PP-

Keller, M. and K.D. Smith. 1990b. Management of salmonid
fisheries. Pages 210-239 in Review of salmon and trout
management in Lake Michigan. M. Keller, K.D. Smith,
and R.W. Rybicki, eds. Mich. Dep. Nat. Resour., Fish.
Special Rep. 14. 254 pp.

66

Kjelson M.A. and P.F. Raquel. 1984. The life history of fall
run juvenile chinook salmon, Oncorhynchus
tshawytscha, in the Sacramento - San Joaquin Estuary of
California. Estuaries 4(3). 285 pp.

Kocik, J.F. and W.W. Taylor. 1987. Effect of fall and winter
' instream flow on year-class strength of Pacific salmon
evolutionarily adapted to early fry outmigration: a
Great Lakes perspective. Amer. Fish. Soc. Symp.
1:430-440.

Kwain, W. 1983. Downstream migration, population size, and
feeding of juvenile rainbow trout. J. Great Lakes Res.
9(1):52-59.

Larsson, P.-O. 1985. Predation on migratory smolt as a
regulating factor in Baltic salmon, Salmo salar L.,
populations. J. Fish Biol. 26:391-397.

 

Lister, D.B. and H.S. Genoe. 1970. Stream habitat
utilization by cohabiting underyearlings of chinook
(Oncorhynchus tshawytscha) and coho (Q. kisutch)
salmon in the Big Qualicum River, British Columbia. J.
Fish. Res. Board Can. 27:1215-1224.

Lorenz, J.M. and J.H. Eiler. 1989. Spawning habitat and redd
characteristics of sockeye salmon in the glacial Taku
River, British Columbia and Alaska. Trans. Am. Fish.
SOC. 118:495-502.

MacCrimmon, H.R. 1954. Stream studies on planted Atlantic
salmon. J. Fish. Res. Board Can. 11:362-403.

McDonald, J. 1960. The behavior of Pacific salmon fry during
their downstream migration to freshwater and saltwater
nursery areas. J. Fish. Res. Board Can. 17:655-676.

McLemore, C.E., F.H. Everest, W.R. Humphreys, and M.F.
Solazzi. 1989. A floating trap for sampling downstream
migrant fishes. U.S. For. Serv., Pac. Northwest Res.
Stn. Res. Note PNW-RN-490. 7 pp.

Milner, A. and L. Smith. 1985. Fyke nets used in a
southeastern Alaskan stream for sampling salmon fry and
smolts. North Am. J. Fish. Manage. 5:502-506.

Moore, R. Record of river and climatological observations,
Grand River, Ionia, Michigan. Unpublished data, 1985-
1990. U.S. Dep. of Commerce, National Oceanic and
Atmospheric_Administration.

 

67

Ottaway, E.M. and A. Clarke. 1981. A preliminary
investigation into the vulnerability of young trout
(Salmo trutta) and Atlantic salmon (g. salar) to
downstream displacement by high water velocities. J.
Fish Biol. 19:135-145.

Patriarche, M.H. 1980. Movement and harvest of coho salmon
in Lake Michigan, 1978-1979. Mich. Dep. Nat. Resour.,
Fish. Res. Rep. 1889. 52 pp.

Pecor, C.H. 1991. Platte River harvest weir and coho salmon
egg-take report. Mich. Dep. Nat. Resour., Fish. Tech.
Rep. 91-1. 42 pp.

Quinn, T.P. and R.E. Tallman. 1987. Seasonal environmental
predictability and homing in riverine fishes. Environ.
Biol. Fishes 18:155-159.

Rakoczy, G.P. and R.N. Lockwood. 1988. Sportfishing catch
and effort from the Michigan waters of Lake Michigan,
and their important tributary streams, January 1,
1985 - March 31, 1986 (Appendices). Mich. Dep. Nat.
Resour., Fish. Tech. Rep. 88-11b. 52 pp.

Rakoczy, G.P. and D. Nelson. 1990. Salmonid harvest. Pages
56-137 in Review of salmon and trout management in Lake
Michigan. M. Keller, K.D. Smith, and R.W. Rybicki, eds.
Mich. Dep. Nat. Resour., Fish. Special Rep. 14. 254

PP.

Rakoczy, G.P. and R.D. Rogers. 1987. Sportfishing catch and
effort from the Michigan waters of Lake Michigan,
Huron, and Erie, and their important tributary streams,
April 1986 - March 31, 1987. Mich. Dep. Nat. Resour.,
Fish. Tech. Rep. 87-6a. 58 pp.

Rakoczy, G.P. and R.D. Rogers. 1988. Sportfishing catch and
effort from the Michigan waters of Lake Michigan,
Huron, and Erie, and their important tributary streams,
April 1, 1987 - March 31, 1988. Mich. Dep. Nat.
Resour., Fish. Tech. Rep. 88-9a. 65 pp.

Refstie, T., S.A. Torstein, and T. Gjedrem. 1977. Selection
experiments with salmon. II. Proportion of Atlantic
salmon smoltifying at one year of age. Aquaculture
10:231-242.

Robertson, J.M. 1990. Preface. Pages i-iii in 1980-1990
Michigan fisheries: a foundation for the future. Mich.
Dep. Nat. Resour., Fish. Special Rep. 13. 130 pp.

68

Rybicki, R.W. 1973. A summary of the salmonid program
(1969-1971). Pages 1-17 in Michigan's Great Lakes trout
and salmon fishery 1969-1972. Mich. Dep. Nat. Resour.,
Fish. Manage. Rep. 5. 105 pp.

Rybicki, R.W., P.W. Seelbach, and W. Wagner. 1990. Biology
of salmonids. Pages 138-194 in Review of salmon and
trout management in Lake Michigan. M. Keller, K.D.
Smith, and R.W. Rybicki, eds. Mich. Dep. Nat. Resour.,
Fish. Special Rep. 14. 254 pp.

Seelbach, P.W. 1985. Smolt migration of wild and
hatchery-raised coho and chinook salmon in a tributary
of northern Lake Michigan. Mich. Dep. Nat. Resour.,
Fish. Res. Rep. No. 1935. 19 pp.

Scott, W.B. and E.J. Crossman. 1973. Freshwater fishes of
Canada. Fish. Res. Board Can. Bull. 184. 966 pp.

Silver, S.J., C.E. Warren, and P. Doudoroff. 1963. Dissolved
oxygen requirements of developing steelhead and chinook
salmon embryos at different water velocities. Trans.
Am. Fish. Soc. 92(4):327-343.

Smith, A.K. 1973. Development and application of spawning
velocity and depth criteria for Oregon salmonids.
Trans. Am. Fish. Soc. 102(2):312-316.

Snyder, G.R. and T.H. Blahm. 1971. Effects of increased
temperature on cold-water organisms. J. Water Pollution
43(5):890-899.

Solomon, D.J. 1978. Some observations on salmon smolt
migration in a chalkstream. J. Fish Biol. 12:571-574.

Stein, R.A., P.E. Reimers, and J.D. Hall. 1972. Social
interaction between juvenile coho (Oncorhynchus
kisutch) and fall chinook salmon (Q. tshawytscha) in
the Sixes River, Oregon. J. Fish. Res. Board Can.
29:1737-1748.

Stuart, T.A. 1953. Spawning migration, reproduction and
young stages of loch trout (Salmo trutta). Freshwat.
Fish. Res. 5.

 

Symons, P.E.K. 1979. Estimated escapement of Atlantic salmon
(Salmo salar) for maximum smolt production in rivers of
different productivity. J. Fish. Res. Board Can.
36:132-140.

 

69

Taube, C.M. 1974. Transfer releases of coho salmon and trout
into an upper part of Platte River and observations on
salmonid spawning. Mich. Dep. Nat. Resour., Fish. Res.
Rep. No. 1815. 28 pp.

Thorpe, J.E. and R.I.G. Morgan. 1978. Periodicity in
Atlantic salmon (Salmo salar L.) smolt migration. J.
Fish Biol. 12:541-548.

 

 

Threlkeld, G. 1967. Soil survey, Ionia County, Michigan.
U.S. Department of Agriculture, Soil Conservation
Service.

Vaux, W.G. 1968. Intragravel flow and interchange of water
in a streambed. Fish. Bull. 66(3):479-489.

Wagner, N.H. 1974. Photoperiod and temperature regulation of
smolting in steelhead trout (Salmo gairdneri). Can. J.
Zool. 52:219-234.

 

Wenger, M.N., R.M. Lichorat, and J.D. Winter. 1984.
Pre-spawning and spawning behavior of coho salmon and
chinook salmon in eastern Lake Erie. N.Y. Fish and Game
J. 31(2):146-164. ‘

Westers, H., W. McClay, C. Pecor, V. Bennett, and J. Driver.
1990. Hatchery production and planting. Pages 14-45 in
Review of salmon and trout management in Lake Michigan.
M. Keller, K.D. Smith, and R.W. Rybicki, eds. Mich.
Dep. Nat. Resour., Fish. Special Rep. 14. 254 pp.

White, H.C. 1936. The food of kingfishers and mergansers on
the Margaree River, Nova Scotia. J. Biol. Board Can.
2:299-309. .

Wickett, W.P. 1958. Review of certain environmental factors
affecting the production of pink and chum salmon. J.
Fish. Res. Board Can. 15:1103-1126.

Wright, S. 1981. Contemporary Pacific salmon fisheries
management. North Am. J. Fish. Manage. 1:29-40.

Zaugg, W.S. and H.H. Wagner. 1973. Gill ATPase activity
related to parr-smolt transformation and migration in
steelhead trout (Salmo gairdneri): influence of
photoperiod and temperature. Comp. Biochem. Physiol.
45:955-965.

 

"- i” 57.3; :if'ijryi? :9 5. 5' ; j- t] . a s
-. +. .   IIIIII
, “”*/”2¢?113=""' 7' 3129

TRTE UNIV , LIBRRRIES

I I II I I

   

3008775615