 

2007

This is to certify that the
thesis entitled

An evaluation of large woody debris restorations on the Manistee
and Au Sable Rivers, Michigan

presented by

Matthew M Klungle

has been accepted towards fulﬁllment
of the requirements for the

 

 

Master of Science , degree in Fisheries and Wildlife

l
, ,1]

M “L ‘W‘

 

Major Professor’s Signature

é jam/my Z ”0%

f

 

Date

MSU is an Afﬁrmative ActiorVEqual Opportunity Institution

 

 

LIBRARY
Michigan State
University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd—p.1

 

AN EVALUATION OF LARGE WOODY DEBRIS RESTORATIONS ON THE
MANISTEE AND AU SABLE RIVERS, MICHIGAN

By

Matthew M Klungle

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
2006

ABSTRACT

AN EVALUATION OF LARGE WOODY DEBRIS RESTORATIONS ON THE
MANISTEE AND AU SABLE RIVERS, MICHIGAN

By

Matthew M Klungle

The addition of large woody debris (LWD) offers a practical management
technique for managers enhancing habitat in midwestem forest streams.
Although it is well established that LWD provides critically important habitat for
salmonids in conifer-dominated watersheds, little is known about the relationship
between LWD and ﬁsh habitat in northern hardwood forests. l evaluated the
effects of LWD restoration efforts on the Manistee and Au Sable rivers. I
evaluated ﬁsh response at different spatial scales, whole site, individual banks,
and individual rivers. There were no signiﬁcant differences detected in catch per
unit effort (CPUE) between treatment and reference reaches for any of the
commonly occurring gameﬁsh species at the whole site level. Brown trout and
rainbow trout showed signiﬁcant effects in CPUE between individual banks within
treatment and reference reaches while rock bass and smallmouth bass showed
no signiﬁcant differences. Gameﬁsh densities differed signiﬁcantly between the
individual rivers. The presence of LWD structures had no signiﬁcant effect on the
abundance of drifting aquatic invertebrates. Channel morphology and substrate
in the vicinity of LWD responded with a trend of aggradation and decreased

substrate size occurring at most sites.

This thesis is dedicated in memory of Ida May Klungle, an inspiration who

instilled in me an appreciation of nature.

iii

ACKNOWLEDGEMENTS

I would like to thank my advisor Dan Hayes for giving me this opportunity,
and committee members, Mary Bremigan and Rich Meritt for their insights and
advice. To the Hayes, Taylor, and Coon lab thank you for all you assistance,
insights, patience, entertainment and friendship. To Kevin Mann, Brian Bellgraph,
Kelly DeGrandchamp, Ryan Mann, Bryan Burroughs, Brad Thopmson, Aaron
Schultz, Cassy Meier and Tim Riley, thank you for all of your assistance in the
ﬁeld and lab. I would also like to thank Holly Jennings, Mike Joyce, Pat Fowler,
and Bob Stuber of the USDA Forest Service, Brian Benjamin, Brad Jensen, and
Dave Smith of Huron-Pines Resource Conservation and Development Council,
Mark Johnson of the Conservation Resource Alliance and Kyle Kruger, Steve
Sendek, and Mark Tonello of the Michigan Department of Natural Resources for
all their logistical support. Funding for this research was provided by the

Michigan Department of Natural Resources and the USDA Forest Service.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... vii
LIST OF FIGURES .............................................................................. ix
INTRODUCTION ................................................................................. 1
METHODS ........................................................................................ 4
RESULTS .......................................................................................... 9
DISCUSSION .................................................................................... 14
APPENDIX ........................................................................................ 19
LITERATURE CITED .......................................................................... 38

LIST OF TABLES

Table 1. Study site details within each river reach including ﬁsh sampling method
for each river reach, the year of LWD placement, number of placed trees
per treatment site, number of natural trees per treatment site, and location
of treatment site in relation to reference site.

Table 2. Size classes and codes used to classify substrate particles.

Table 3. Monthly summer invertebrate sampling schedule for each river reach
and site within the reaches.

Table 4. Best-ﬁt distribution as determined by NC patterns used for brown trout
(BNT), rainbow trout (RBT), rock bass (RB), and smallmouth bass (SMB)
at each scale of GENMOD analysis.

Table 5. P-values from a GENMOD analysis of variance modeling the effects of
LWD on overall mean CPUE for brown trout (BNT), rainbow trout (RBT),
rock bass (RB), and smallmouth bass (SMB). Mean and SE of CPUE
reported are back transformed least square means from GENMOD.

Table 6. P-values from a GENMOD analysis of variance modeling the effects of
LWD on individual river mean CPUE for brown trout (BNT), rainbow trout
(RBT), rock bass (RB), and smallmouth bass (SMB). Mean and SE of
CPUE reported are back transformed least square means from GENMOD.

Table 7. P-values from a GENMOD analysis of variance modeling the effects of
the river treatment interaction for CPUE of brown trout (BNT), rainbow
trout (RBT), rock bass (RB), and smallmouth bass (SMB). Mean and SE of
CPUE reported are back transformed least square means from GENMOD.

Table 8. Numbers of game ﬁsh tagged and recaptured during sampling events in
the Mio and Hodenpyl river reaches. Mio results from ﬁsh sampled in 2003
and Hodenpyl results from ﬁsh sampled in 2002 and 2003.

Table 9. P-values from a GENMOD analysis of variance modeling the effects of
LWD on individual bank mean CPUE brown trout (BNT), rainbow trout
(RBT), rock bass (RB), and smallmouth bass (SMB). Mean and SE of
CPUE reported are back transformed least square means from GENMOD.

Table 10. P-values from a GENMOD analysis of variance modeling the effects of

LWD structures on invertebrate behavioral drift density for overall,
Crustecea, Ephemeroptera, Tricoptera, Chironomidae, and other Diptera.

vi

Table 11. P-values from a paired t-test for longitudinal channel morphology in
each study site for sampling years 2002 and 2003 with mean change in
channel depth and SE. The letter represents each reach (Mio, Hodenpyl,
and Alcona) and the number designating which site within the reach.

Table 12. P-values from a paired t-test for substrate classiﬁcation in each study
site for sampling years 2002 and 2003 with mean change in classiﬁcation
and SE. The letter represents each reach (Mio, Hodenpyl, and Alcona)
and the number designating which site within the reach. No statistical
analysis for Hodenpyl site 3 due to lost 2002 data.

vii

LIST OF FIGURES

Figure 1. River reach and study site locations on the Manistee and Au Sable
rivers.

Figure 2. Split-plot design with paired treatment and reference reaches.

Figure 3. Length class (in millimeters) frequency distributions for brown trout (A)
and rainbow trout (B) in the Hodenpyl reach and Mia reach.

Figure 4. Overall estimates of mean (:t 28E) back-transfonned catch per unit
effort of the four most commonly encountered gameﬁsh in the treatment
and reference sections of the combined Hodenpyl and Mio reaches of the
Manistee and Au Sable rivers as determined by electroﬁshing.

Figure 5. Overall estimates of mean (i 28E) back-transfonned catch per unit
effort of the four most commonly encountered gameﬁsh in the treatment
and reference sections on each the Hodenpyl reach of the Manistee river
and the Mio reach of the Au Sable river as determined by electroﬁshing.

Figure 6. Individual bank estimates of mean (i 28E) back-transformed catch per
unit effort of the four most commonly encountered gameﬁsh in the
treatment and reference sections of the combined Hodenpyl and Mia
reaches of the Manistee and Au Sable rivers as determined by
electroﬁshing.

Figure 7. Estimates of mean (i 28E) back-transformed catch per unit effort of the
invertebrate taxa occurring above and below LWD structures of the
combined Hodenpyl and Mio reaches of the Manistee and Au Sable rivers
as determined in monthly behavioral drift samples.

Figure 8. Survey transects from each of the Mio study sites, showing the
longitudinal river channel for 2002 and 2003. Water surface (2003) is
represented by the dotted top line. LWD structures were present along the
entire transect in 2003.

Figure 9. Substrate size percent frequency distribution for each Mio site in 2002
and 2003.

viii

Figure 10. Survey transects from each of the Hodenpyl study sites, showing the
longitudinal river channel proﬁle for 2002 and 2003. Water surface (2003)
is represented by the dotted top line with the shaded area representing
location of LWD structures in 2002. In 2003 LWD structures were present
along the entire transect.

Figure 11. Substrate size percent frequency distribution for Hodenpyl sites one
and two in 2002 and 2003. Hodenpyl site three missing 2002 data.

Figure 12. Survey transects from each of the Alcona study sites, showing the
longitudinal river channel proﬁle for 2002 and 2003. Water surface (2003)
is represented by the dotted top line. Location and size of LWD structures
did not change between 2002 and 2003.

Figure 13. Substrate size percent frequency distribution for each Alcona site in
2002 and 2003.

Figure 14. Mean daily water temperatures in °C during the summers of 2002 and
2003 in the Mio (A), Hodenpyl (B), and Alcona (C) river reaches. Shaded
area represents critical thermal limits at 19°C (upper thermal limit for
growth) and 247°C (lethal) for brown trout.

ix

INTRODUCTION

Large woody debris (LWD) is a natural component of streams and rivers
and plays a complex role in hydrological, chemical, and biological processes
(Gumell et al. 1995). Fish use LWD for foraging sites and as protection against
current, predators and competitors (Angermeiger and Karr 1984). Large woody
debris provides a stable substrate for aquatic organisms such as bacteria, fungi,
and invertebrates, all of which are involved in the decomposition of wood and
represent major components of the trophic pathways in lotic systems
(Angermeiger and Karr 1984; Lemly and Hilderbrand 2000). Consequently, LWD
placement has become a common technique for stream restoration and for
improving ﬁsh habitat to compensate for reductions in LWD due to various
anthropogenic activities (Kauffman et al. 1997)

During the late 1800’s and early 1900’s, timber companies used the
Manistee and Au Sable Rivers for their log drives. Historical accounts indicate
that before the log drives, these rivers were so full of woody debris they were
non-navigable (Miller 1963; Peterson 1972). To facilitate these drives, timber
companies removed all natural deadfall trees that once provided habitat for many
native ﬁshes. Continued large-scale land clearance and development has also
changed the landscape of the river corridors. Completion of Mio dam in 1916,
Alcona dam in 1924 and Hodenpyl dam in 1925 also hindered natural recruitment
as any LWD that travels through the reservoir and to the face of the dam is

blocked from downstream passage or removed. Consequently, natural

recruitment of LWD into many northern Michigan rivers has been reduced or
even eliminated for the last 100 years.

Stream habitat work by public management agencies in the US. was
pioneered by the Michigan Conservation Department (predecessor to the
Department of Natural Resources). Brush piles and cover structures were added
to Michigan trout creeks in 1927; prior to this, almost all habitat manipulations
were done privately by anglers (Hubbs et al. 1932). These manipulations were
initiated under the assumption that the amount and quality of physical habitat in
the stream channel is limiting.

Managers continue to recognize that habitat plays a critical role in the
dynamics of ﬂuvial ﬁsh stocks. Habitat restoration continues to be resource
managers’ preferred solution to increasing the number or size of ﬁsh in a
population as artiﬁcial propagation tends to be cost prohibitive and protective
regulations do not directly produce ﬁsh. The enhancement and restoration of
stream habitat for game ﬁshes, particularly salmonids, through the instream
placement of structures and LWD has increased dramatically in the Paciﬁc
Northwest (Roni and Quinn 2001). In the last 20 years, many studies have
emphasized the critical role that LWD plays in creating and maintaining ﬁsh
habitat in streams. LWD is widely recognized as a critical component of natural
streams and rivers and its role is ecologically complex.

Most LWD evaluations have considered small and medium-sized streams
but rarely larger rivers. Since the early 19808, the interest in habitat restoration

projects has focused on the Paciﬁc Northwest in an effort to mitigate degradation

and loss of habitat due to human disturbance and to stop or reverse signiﬁcant
salmonid population declines (National Research Council 1996). Steeper, high-
energy small streams common in the Paciﬁc Coast rain-forest mountains are
more vulnerable to logging practices and more sensitive to manipulations than
larger lower gradient ones. Large woody debris complexes create essential pools
and gravel deposits in steep, cool-climate streams of the Paciﬁc Coast.
Evaluations of these restorations have produced highly variable results. There
have been numerous reports indicating increased ﬁsh abundance in response to
restoration efforts (e.g., Angerrneier and Karr 1984; Fausch and Nortcote 1992;
Gowan and Fausch 1996; Flebbe 1999; Rosenfeld et al. 2000; Roni and Quinn
2001a) along with a number of reports with no signiﬁcant ﬁsh response and even
decreases in ﬁsh abundance in response to restoration efforts (e.g., Angerrneier
and Karr 1984; Thevenet and Statzner1999; Roni and Quinn 2001a).

In lower gradient streams and rivers of the Great Lakes region, LWD can
potentially slow ﬂow. Due to the concern that scarce spawning gravels may
become buried in sand and silt, and water temperatures may become too warm
for trout in the summer there has been little work on lower gradient midwestem
streams and rivers. This concern seems to have superceded many of the
potential beneﬁts of LWD and consequently, the effects of LWD on larger rivers
in the midwest have yet to be investigated. Based on the underlying ethos that
LWD is a critical component of a natural river system and that restoration is a
viable option, restoration efforts began in the fall of 1998 by the US. Forest

Service in conjunction with the local conservation districts (Conservation

Resource Alliance and Huron Pines Resource Conservation District) and the
Michigan Department of Natural Resources. Whole tree LWD structures were
placed with a helicopter in the Mio reach of the Au Sable river as a demonstration
project to determine the feasibility and economics. They continued this effort with
LWD placements in the Alcona reach of the Au Sable river and the Hodenpyl
reach of the Manistee river during the fall of 2000 with additional placements in
the Mio and Hodenpyl reaches in the fall of 2002.

The purpose of this study was to evaluate the effectiveness of these large
woody debris restoration efforts implemented in the Manistee and Au Sable
rivers. These data will provide a base line for possible future studies on
established LWD, and will aid in the design of future evaluations of LWD
implementation projects.

The objectives of this evaluation were to determine whether the addition of
LWD structures produced a signiﬁcant change in physical habitat, gameﬁsh
abundance, and invertebrate abundance in two northern Michigan rivers.
Speciﬁcally we tested the null hypothesis that paired treatment and reference
reaches would not differ in (1) gameﬁsh abundances, (2) survival and retention of
gameﬁsh, and (3) invertebrate drift. In addition we documented channel
morphology changes in proximity to LWD being added.

METHODS
Study Reaches
Three study sites were selected in the Manistee River from Hodenpyl Dam

to Red Bridge, Manistee County, four sites on the Au Sable River from Mio Darn

to Cummins Flat pullout, Oscoda County, and three sites on the Au Sable River
from Alcona Darn to Thompson's Landing, Alcona County (Figure 1). Hodenpyl
and Mio dams are the ﬁrst barriers encountered heading down stream on these
rivers. All of these dams are operated on a run-of-river ﬂow scheme. Gradient in
these reaches is somewhat higher, with water quality and biological communities
typical of large, cold-cool rivers.
Fish Response to LWD structures

A split-plot design with paired treatment and reference reaches was used
to determine the response of gameﬁsh to LWD treatments. Treatment and
reference sites 100 m long were selected in each reach and broken into three
sections (left bank, middle, right bank; Figure 2). Treatment is deﬁned as the
artiﬁcial placement of LWD within the active stream channel. This design involves
comparison between paired treatment and reference reaches at the 10 study
sites. Sites were chosen by proximity to tree harvest sites to minimize turnaround
time and maximize the number of trees placed. At all sites, paired treatment and
reference sites were also chosen so that pairs would be as similar to each other
as possible and so that treatment sites would represent a range of amounts of
LWD (Table 1). Whenever possible, I paired treatment and reference sites
adjacent to one another to maintain site similarity. When this was not possible,
treatment and reference sites were within 200 m of one another. I also tried to
alternate the upstream site to avoid confounding from natural down stream

migrations (Table 1).

All ﬁeld sampling occurred during the summers of 2002 and 2003. A
Smith-Root Cataraft electroﬁshing boat was used for all ﬁsh sampling efforts on
the Mio and Hodenpyl reaches. The electroﬁshing boat was set at pulsed DC
(40% cycle duty) on low range (50 - 500) volts at 4 — 6 amps. Fish were sampled
at the three sites along the Hodenpyl reach, once per month (May to August), in
2002 and 2003. The four sites in the Mio reach were sampled once per month
(June and July), in 2003. Sites on the Mio reach were shocked at night, due to
much greater recreational use during the day than the Hodenpyl or Alcona
reaches .

Gameﬁsh populations were sampled using single pass electroﬁshing. All
captured gameﬁsh were measured for total length to the nearest 1 mm. In order
to assess site ﬁdelity trout under 150 mm were marked with a bank speciﬁc ﬁn
clip, and those over 150 mm were marked with Visual Implant alpha-numeric
tags. Other gameﬁsh were marked with Floy T-bar tags.

Fish communities on the Alcona reach of the Au Sable river were sampled
once per month (June to August) in 2002 and 2003 by visual observations, as
electroﬁshing was not possible due to the remoteness and inaccessibility of the
reach. Gameﬁsh populations were sampled using single pass snorkel surveys.
Swimmers with a mask and snorkel sampled individual banks within the paired
treatment and reference sites noting species encountered in each sampling pass.
Channel response

Localized changes in river channel morphology and substrate were

monitored by ﬁxed longitudinal survey transects located around the LWD

structures. Benchmarks were established by driving an aluminum nail into the
base of a healthy tree with an assumed elevation of 30.5 m. One meter lengths
of 1.6 cm rebar were used to delineate endpoints of the 100 m long study sites.
Channel morphology was measured at all 10 treatment sites. Standard
longitudinal proﬁle surveying techniques were used in 2002 and in 2003
(Harrelson et al. 1994). Morphology was measured every 0.6 m along the
longitudinal proﬁle using a stadia rod and surveyor’s level to map channel
elevation in proximity to trees being added. Substrate was evaluated using a
pebble count code classiﬁcation method (Harrelson et al. 1994). One hundred
random streambed particles were classiﬁed along each transect.
Invertebrate response

Drifting aquatic invertebrates were collected at monthly intervals (Table 3),
as river conditions allowed, from May through August during early evening hours
to determine behavioral drift response to LWD placements. Drift samples were
collected for 1 hour in 2001 and 15 minutes in 2002 after sunset, above and
below LWD structures. Invertebrates were captured with 200 um mesh drift nets
14.5-cm wide placed at an average depth of 65 cm, and then strained through a
500 um sieve. Samples were stored in appropriately sized nalgene® bottles or in
zip closure freezer bags (doubled bagged) and preserved in 90% ethyl alcohol. I
sampled a minimum of six sites per month in this evaluation (Table 3).

To identify other potential limiting factors, water temperature was recorded
every 1 hour with Onset Stowaway temperature loggers at each study site in

2002 and 2003.

Statistical analyses

I used data collected from the Hodenpyl reach in 2002 and 2003
and data from the Mio reach in 2003 to test differences in mean catch per unit
effort (CPUE) between treatment and reference reaches. All data sets were
analyzed using SAS version 8.2. Data analysis focused on the four most
common gameﬁsh species encountered, brown trout Salmo trutta, rainbow trout
Oncorhynchus mykiss, rock bass Ambliopites rupestris, and smallmouth bass
Micmpterus dolomieu. Differences in gameﬁsh CPUE between treatment and
reference reaches were compared using a generalized linear model ANOVA
(GENMOD) with user deﬁnable distributions to meet necessary assumptions. An
alpha level of 0.05 was used to determine statistical signiﬁcance for each
analysis. A negative binomial distribution and a Poisson distribution were applied
to the ﬁsh data to meet the basic assumptions of GENMOD and to account for a
large number of zero’s in the data sets. Akaike’s information criteria (AIC) was
used to determine the most appropriate model for each data set.

Fish can respond to LWD additions at different spatial scales. Thus, I
evaluated their response at the scale of whole site (i.e., entire 100 m treatment or
reference sites) as well as the scale of individual banks (i.e., river banks within
treatment sections that did not receive LWD treatments were considered
reference banks). I also evaluated whether the response to LWD varied between
the Mio reach of the Au Sable River and the Hodenpyl reach of the Manistee
River.

Channel analyses

I used data collected in 2002 and 2003 to test differences in substrate
classiﬁcations and longitudinal channel transects in treatment reaches. Channel
morphology and substrate change between years were determined using a
paired t-test.

Invertebrate analyses

I used data collected in 2002 and 2003 to test differences in behavioral
drift densities above and below LWD structures. Data analysis focused on the
following taxonomic groups; Crustacea, Ephemeroptera, Tricoptera, and Diptera.
which I broke down to family Chironomidea and other Diptera. Differences in
behavioral drift above and below LWD structures were compared using a
generalized linear model ANOVA with deﬁnable distributions to meet necessary
assumptions. An alpha level of 0.05 was used to determine statistical
signiﬁcance for each analysis. A negative binomial distribution was applied to the
overall and individual taxonomic group invertebrate data to meet the basic
assumptions of GENMOD to account for a large number of zero’s in the
individual taxon.

RESULTS
Fish Response

Brown and rainbow trout consistently had the highest mean CPUE of
gameﬁsh encountered in the Mio and Hodenpyl reaches. This is not surprising as
these reaches are heavily stocked yearly by the MDNR. Length frequency
distributions of these species (Figure 3) show a lack of ﬁsh less than 180mm,

indicating few if any young of year or 1+ ﬁsh are produced in these river reaches.

There were no signiﬁcant differences detected in catch per unit effort
(CPUE) between treatment and reference reaches for any of the commonly
occurring gameﬁsh species as indicated by the GENMOD analysis (Table 5).
Point estimates of mean CPUE for brown trout, rainbow trout, and rock bass
were higher in treatment reaches (27.7%, 10.4%, and 40.0% respectively) than in
reference reaches but were highly variable resulting in no statistical signiﬁcance
between the treatment and reference reaches (Figure 4). Smallmouth bass
showed a slightly higher mean CPUE point estimate for the reference reach
(4.2%), but the difference was not signiﬁcant. CPUE of other species was too low
to permit statistical analysis.

Brown trout and rainbow trout showed signiﬁcant difference in CPUE
between individual banks within treatment and reference reaches (river banks
within treatment sections that did not receive LWD treatments were considered
reference banks) while rock bass and smallmouth bass showed no signiﬁcant
differences (Table 9). Point estimates of mean CPUE were consistently higher in
the treatment reaches (30.0% brown trout, 41.9% rainbow trout, 58.5% rock
bass, and 16.5% smallmouth bass) than reference reaches (Figure 6).

Mean CPUE for the different species differed between the Manistee and
Au Sable rivers (Table 6). Overall mean CPUE was consistently higher (ave.
80%) in the Mio reach for each game ﬁsh species (Figure 5). In the Mio reach,
gameﬁsh mean CPUE was consistently higher in treatment sections with the
exception of smallmouth bass (Figure 5). In the Hodenpyl reach, mean CPUE

was higher in the treatment sections for brown trout and smallmouth bass and

10

higher in the reference sections for rainbow trout and rock bass (Figure 5). Brown
trout, rainbow trout, and smallmouth bass showed highly signiﬁcant differences
between the rivers, with rock bass only marginally signiﬁcant (Table 6). Within
the river reaches, rainbow trout were the only species to have a signiﬁcant
interaction of river*treatment (Table 7). Thus rainbow trout responded differently
to the LWD additions between rivers.

In the Alcona reach, visual observation was used to assess ﬁsh response
to LWD additions. Within treatment sites, the following species were observed
walleye (Stizostedion vitreum), smallmouth bass, bluegill sunﬁsh (Lepomis
macrochirus), common shiner (Luxilus comutus), logperch (Percina caprodes),
blackside darter (Percina maculata), silver redhorse (Moxostoma anisunrm),
crappie (Pomoxis nigromaculatus), northern hog sucker (Hypentelium nign'cans),
rock bass, white sucker (Catostomus commersonir), and johnny darter
(Etheostoma nigmm). Within the treatment reaches, ﬁsh were primarily observed
in banks receiving treatments, often with no ﬁsh observed in non-treatment
banks. Species observed in reference sites were smallmouth bass, logperch
(Percina caprodes), blackside darter (Percina maculata), and bluegill sunﬁsh
(Lepomis macrochinrs). Fish observed in treatment banks were often found in
high densities compared to non-treatment banks where often single individuals
were found. Redhorse suckers and walleye were occasional observed in the
deepest areas of the river. All ﬁsh moved substantially (possibly due to snorkeler
presence) making it very difﬁcult to count. Overall counts were generally very

low, thus precluding statistical analysis

11

The results of ﬁsh tagging were consistent with a low sampling efﬁciency,
as few tagged ﬁsh were recaptured. Fish recapture rates for the 2002 Mio reach
sampling were 4.9% in the treatment site and 11.1% in the reference site (Table
8). In the Hodenpyl reach, ﬁsh recapture rates were 2.1% in the treatment site
and 0% in the reference site (Table 8) from combined 2002 and 2003 sampling.
All of these recapture rates were considered too low for statistical analysis.
Invertebrate response

No effects of LWD structures on overall mean invertebrate behavioral drift
density were detected (Table 10). The presence of LWD structures had no
signiﬁcant effect on invertebrate behavioral drift density or on the density of any
particular taxon captured in the drift nets. There were no observed differences in
mean density of the taxon Crustacea, Ephemeroptera, Tricoptera, and Diptera
between above and below LWD structures (Table 10). CPUE of Lepidoptera and
Plecoptera was too low to permit statistical analysis. Invertebrate drift numbers
were dominated by Diptera (84.1%, 77.6% Chironomidae and 6.6% other
Diptera) followed by Crustacea (6.8%), Ephemeroptera (5.1%), Tricoptera
(3.3%). Coleoptera (0.33%), Plecoptera (0.24%), and Lepidoptera (0.05%)
combined to make up less than 1% of the drift net composition (Figure 7).
Channel response

Channel morphology changed in all of the Mio sites (Table 11) and
substrate classiﬁcation differed statistically for three of the four sites (Table 12).
In the Hodenpyl reach, sites one and two showed signiﬁcant channel changes

(Table 11) and sites two and three showed signiﬁcant substrate differences

12

(Table 12). In the Alcona reach, only site three showed signiﬁcant morphology
differences (Table 10) while sites one and three showed signiﬁcant substrate
differences (Table 11). It is important to note here that paired t-tests are sensitive
statistical analyses, and it may be more insightful to compare yearly morphology
ﬁgures. Mio site one showed channel aggradation (Figure 8) of ﬁner sediments
(Figure 9). Mio site two showed consistent channel morphology (Figure 8) and
stable substrates (Figure 9) between the sample years 2002 and 2003. This is
expected because this site only received one treatment tree in 1998, in the bank
surveyed, and no treatment in 2002. Mio site three showed substantial
aggradation (Figure 8) while substrate seemed to stay similar from year to year
(Figure 9). Mio site four showed some channel incision with a shift towards ﬁner
substrates. The Hodenpyl channel morphology showed minimal change from
year to year (Figure 10), the only changes appear to be some substrate
redistribution at site 1 and a shift to ﬁner substrate at site 2 (Figure 11). The three
Alcona sites also appear to stay relatively unchanged (Figure 12) with a slight
redistribution of substrates that appear to shift to ﬁner substrates (Figure 13).
Temperature

Daily mean water temperatures in 2002 and 2003 for the Mio reach reached
critically high temperatures in mid June and stay there until mid August. There
was a dip in mid June in 2002, which was probably due to a large rain event. In
the Hodenpyl reach, water temperatures for 2002 went into the critical zone in
mid May and stayed there at least until August. In 2003, the water temperature

did not reach the critical zone until early July and stayed there until at least

13

August. In the Alcona reach, water temperatures were in the critical zone the
entire temperature recording duration.

Discussion

The beneﬁts of LWD as a habitat restoration technique have been well
established and well documented; LWD provides current breaks, cover from
predators, and forage sites for ﬁsh (Angermeiger and Karr 1984; Gowan and
Fausch 1996; Flebbe 1999; Fausch and Nortcote 1992; Lemly and Hilderbrand
2000; Rosenfeld et al. 2000; Roni and Quinn 2001a). Most recent studies on
habitat restoration are from the western United States and Canada and have
reported signiﬁcant physical responses to restoration; but, ﬁsh response to
habitat restoration, speciﬁcally LWD placements, and the mechanisms
responsible have been less well researched and produced more variable results.
Also, these recent studies are often on debris dams spanning the stream
channel; few evaluations of gameﬁsh in streams where LWD complexes (60+
trees fully in the stream channel) only encompass 30% of the stream channel are
published. Because the actual response of ﬁsh to LWD additions is less well
understood and studies from the Midwest are lacking, I conducted this study to
determine response and help ﬁll in gaps in understanding of LWD restorations on
larger midwestem streams.

LWD structures were placed in the Manistee and Au Sable rivers to
improve habitat for valued gameﬁshes and help restore the natural riverine
ecosystem. These LWD structures were installed under the premise that physical

habitat may be a limiting factor due to historic land use practices. With the

14

apparent direct inﬂuence of LWD on the quantity and quality of habitat available
to ﬁsh, it was expected that the LWD additions would increase gameﬁsh density
in these reaches.

At the whole site level, treatment sites generally had higher densities, but
the highly variable data yielded no signiﬁcant results. This was also true when
the response was evaluated for individual rivers. However, a test of ﬁsh density
at the individual bank scale showed a signiﬁcant increase in mean CPUE for
brown trout and rainbow trout in treatment banks, indicating that habitat may be a
limiting factor for these species in these reaches. The impact of LWD structures
in these reaches, however, appears to produce a very localized effect.

Because natural LWD was lacking in much of these river reaches, I
expected a larger response from game ﬁshes. Angerrneier and Karr (1984),
Gowan and Fausch (1996), and Roni and Quinn (2001a) all found higher ﬁsh
densities in areas where LWD was present or added. There are several potential
reasons why a greater response was not observed. First, the Hodenpyl reach
has holes >6 m deep (T onello 2003), which combined with the high summer
water temperatures, may prove to be preferred refugia over the shallower,
warmer areas with LWD structures. These deep holes may concentrate ﬁsh and
explain the lower mean CPUE for game ﬁsh in this reach, as the holes are too
deep to sample with electroﬁshing gear. Although sampling efﬁciency was low
based on mark-recapture analysis, the relative comparison between treatment
and reference sites is still valid. Another potential problem is that hatchery trout

are known to have difﬁculties efﬁciently adapting to natural habitats when

15

stocked (Bettinger and Bettoli 2002). From the trout length frequency charts
(Figure 3) it is obvious that hatchery ﬁsh dominate these populations, with little
yearly hold over. This limited hold over may be due to the poor adaptation to
natural environments, the high summer water temperatures (Figure 14) or a
combination of these factors.

Non-salmonid response was puzzling given the cool summer water
temperature regime and the physical beneﬁts provided by the LWD structures.
Studies have indicated that warmwater species are more abundant in the
presence of LWD than its absence (Angerrneier and Karr 1984; DuFour 1989). I
expected a response from species like smallmouth bass and rock bass. Although
these were some of the most abundant gameﬁsh species in these reaches, there
was no response to LWD structures at any scale.

Large woody debris also provides substrate for aquatic organisms such as
bacteria, fungi, and invertebrates (Bilby and Likens 1980; Lemly and Hilderbrand
2000), all major components of riverine ecosystems. By increasing substrate for
invertebrates colonize and food for invertebrates by increasing bacteria, fungi,
and detritus inputs, LWD complexes can increase invertebrate production
(Angenneier and Karr 1984; Lemly and Hilderbrand 2000). Thus, I expected to
see an increase in invertebrate drift density below LWD structures. This
expectation was not supported by the data, however. There are several potential
reasons for this observation. First, drift may not be representative of density nor
production in these reaches. Further, only some taxa drift. Thus, the density and

production of inverts as a whole could vary without showing substantial changes

16

in drift. Also, variability in the data may obscure any real patterns. Finally, the
coarser substrates of these river reaches may provide invertebrates with refugia
from currents and predators keeping invertebrate densities naturally high in
reference reaches.

In the six sites in which substrate differed between 2002 and 2003,
substrate size decreased (Figures 9,11, and 13). This suggests that the LWD
structures are collecting ﬁner sediments. Overall though, the substrate of these
river reaches consists largely of cobble and gravel, which are relatively resistant
to redistribution. This may be due to past dam management (i.e., daily peaking
ﬂows until the early 80’s) that may have armored the substratum making it
difﬁcult for any scouring effects to be observed. So, it is not surprising that
observed morphometry changes are generally aggradations and substrate
changes tended towards ﬁner substrates being deposited in the current breaks
. provided by the LWD structures. It is important to note that in Mio site 3 the
survey data suggest channel aggregation of three to four feet. It is possible this
could be due to sampling error, a shift in one of the ﬁxed points the readings are
based on, or extensive point bar development.

The most striking outcome of this study is the summer water temperatures
in these river reaches. For brown trout, the critical thermal limit for growth and
feeding is 19° to 20° C and the critical thermal lethal limit is approximately 25° C
(Elliot 1994). As shown by Figure 14, mean daily water temperatures in these
river reaches is consistently above critical thermal limits during the summer. The

consistently high water temperature regimes, minimal yearly holdover, and lack

17

of young of year suggest that high summer temperatures may be the primary
factor limiting trout populations in these reaches. Consequently, a put, grow, and
take ﬁsh stocking practice may be the only way to maintain a trout ﬁshery in

these river reaches unless water temperature can also be managed.

18

APPENDIX

Tables and Figures

19

Table 1. Study site details within each river reach including sampling design, ﬁsh
sampling method for each river reach, the year of LWD placement, number of
placed trees per treatment site, number of natural trees per treatment site, and
location of treatment site in relation to reference site with A representing
upstream and B representing downstream.

 

 

Reach Hodenpyl Mio Alcona
Design After placement After placement After placement
Sampling Method Electroﬁshing1 Electroﬁshing2 Snorkel

Year of sampling 2002 and 2003 2003 2002 and 2003
Year of placement 2001 and 2002 1998 and 2002 2001

Treatment sites 1 2 3 1 2 3 4 1 2 3
No. placed trees 12 35 35 21 62 28 31 4 6 3
No. natural trees 6 0 8 0 0 0 0 0 0 0
Treatment location A B B A A B B A A B

 

1Electroﬁshing conducted during the day
2Electroﬁshing conducted at night

Table 2. Size classes and codes used to classify substrate particles.

 

 

Size Code Size Class (mm) Particle
0 Trash
1 Organic
2 0.00024 - 0.004 Clay
3 0.04 — 0.062 Silt
4 0.062 - 2 Sand
5 2 - 4 Very ﬁne gravel
6 4 - 8 Fine gravel
7 8 - 16 Medium gravel
8 16 - 32 Coarse Gravel
9 32 — 64 Very coarse gravel
10 64 — 128 Small Cobble
11 128 - 256 Large Cobble
12 256 - 512 Small Boulder
13 >512 Medium Boulder

 

20

Table 3. Monthly summer invertebrate sampling schedule for each river reach
and site within the reaches sampled.

 

 

Reach Hodenpyl Mio Alcona
May 1,2 1,2 1,2
June 2,3 3,4 2,3
July 1,3 1,3 1,3
August 1,2,3 2,4 1,2,3

 

Table 4. Best-ﬁt distribution as determined by AIC patterns used for brown trout
(BNT), rainbow trout (RBT), rock bass (RB), and smallmouth bass (SMB) at each
scale of GENMOD analysis.

 

Species Overall Individual Rivers Individual Banks
BNT Negative Binomial Negative Binomial Negative Binomial
RBT Negative Binomial Negative Binomial Poisson

RB Poisson Poisson Negative Binomial
SMB Poisson Poisson Poisson

 

Table 5. P-values from a GENMOD analysis of variance modeling the effects of
LWD on overall mean CPUE for brown trout (BNT), rainbow trout (RBT), rock
bass (RB), and smallmouth bass (SMB). Mean and SE of CPUE reported are
back transformed least square means from GENMOD.

 

 

 

Treatment Reference
Spp. df Mean SE Mean SE X2 P-values
BNT 55 7.94 1.17 5.65 1.17 2.49 0.115
RBT 55 2.11 1.23 1.89 1.26 0.12 0.731
RB 55 0.19 1.58 0.11 1.79 0.49 0.484
SMB 55 0.84 1.24 0.87 1.234 0.02 0.886

 

Table 6. P-values from a GENMOD analysis of variance modeling the effects of
LWD on individual river mean CPUE for brown trout (BNT), rainbow trout (RBT),
rock bass (RB), and smallmouth bass (SMB). Mean and SE of CPUE reported
are back transformed least square means from GENMOD.

 

 

Hodenpyl Mio
Spp. df Mean SE Mean SE X2 P-values
BNT 54 3.61 1.14 12.43 1.20 14.37 <0.001
RBT 54 0.64 1.25 5.65 1.21 44.38 <0.001
RB 54 0.07 1.84 0.25 1.75 4.42 0.036
SMB 54 0.33 1.31 2.18 1.18 14.67 <0.001

 

21

 

 

 

 

Evd de mm; and 89 8e om._. mud Nee and vm me
Rd 5;. «EN aﬁo no; and mod oﬁo and mod 3 mm
Rod «mum 93 Ed mm; mmd Fm; mad 3e med vm .Em
oomd mod omé m2: one m3... mm; cod owe and 3 ._.2m
mm:_m>i Nx mm :82 mm cams. mm cams. mm cam—2 he dam
85.3% 20.58:. 8:90.61 EmEummc.
2.2 .358...

 

.092sz So: mcmmE mama—cm Smog 39:0ch9. xomn 0.5 3:89 man—o

3 mm 2m :85. 3.29 83. saga...» as... .3”: 82 .8. .56 as. 262.9 .szv 52. $65 a mac
.2 8.8995 EcEﬁob 52. c5 3 gusto 05 9580:. 85:9 Co war—Em 0022mm m E0: mm:_m>i N mink

22

Table 8. Numbers of game ﬁsh tagged and recaptured during sampling events in
the Mio and Hodenpyl River reaches. Mio results from ﬁsh sampled in 2003 and
Hodenpyl results from ﬁsh sampled in 2002 and 2003.

 

 

 

Mio Hodenpyl
Treatment Reference Treatment Reference
Tagged 163 108 96 85
Recaptured 8 12 2 0

 

Table 9. P-values from a GENMOD analysis of variance modeling the effects of
LWD on individual bank mean CPUE brown trout (BNT), rainbow trout (RBT),
rock bass (RB), and smallmouth bass (SMB). Mean and SE of CPUE reported
are back transformed least square means from GENMOD.

 

Treatment Reference
Spp. df Mean SE Mean SE X2 P-values

 

 

BNT 112 4.09 1.16 2.74 1.14 4.09 0.043
RBT 114 1.29 1.15 0.75 1.16 8.63 0.003
RB 114 0.11 1.63 0.05 1.82 1.2 0.273
SMB 114 0.46 1.26 0.40 1.22 0.21 0.646

 

Table 10. P-values from a GENMOD analysis of variance modeling the effects of
LWD structures on invertebrate behavioral drift density for overall, Crustecea,
Ephemeroptera, Tricoptera, Chironomidae, and other Diptera.

 

 

Family df Above mean SE Below mean SE X2 P-values
Overall 40 46.34 1.24 53.67 1.25 0.22 0.638
Crustacea 40 4.99 1.60 1.81 1.63 2.23 0.135
Ephemeroptera40 2.33 1.41 2.81 1.41 0.15 0.702
Tricoptera 40 1.71 1.36 1.57 1.36 0.04 0.841
Chironomidae 40 34.38 1.26 43.19 1.26 0.48 0.488
other Djatera 40 2.43 1.40 4.14 1.38 1.28 0.257

 

23

Table 11. P-values from a paired t-test for longitudinal channel morphology in
each study site for sampling years 2002 and 2003 with mean change in channel
depth and SE. The letter represents each reach (Mio, Hodenpyl, and Alcona) and
the number designatingwhich site within the reach.

 

 

 

 

Site df mean SE t-value pvalue
M1 105 -1.00 0.0546 -18.38 <.0001
M2 99 -0.13 0.0274 -4.70 <.0001
M3 105 0.11 0.0340 3.31 0.0013
M4 106 -0.09 0.0119 -7.70 <.0001
H1 105 0.38 0.0495 7.64 <.0001
H2 117 -0.17 0.0413 -4.04 <.0001
H3 108 -0.05 0.0850 -0.62 0.5370
A1 105 0.01 0.0297 0.42 0.6764
A2 92 0.05 0.0552 0.87 0.3855
A3 72 -0.42 0.0452 -9.24 <.0001

 

Table 12. P-values from a paired t-test for substrate classiﬁcation in each study
site for sampling years 2002 and 2003 with mean change in classiﬁcation and
SE. The letter represents each reach (Mio, Hodenpyl, and Alcona) and the
number designating which site within the reach. No statistical analysis for

Hodenpyl site 3 due to lost 2002 data.

 

 

 

 

Site df mean SE t-value p-value
M1 99 0.76 0.3699 2.05 0.0425
M2 99 -0.33 0.2349 -1.41 0.1631
M3 99 0.58 0.2610 2.22 0.0285
M4 99 1.62 0.2485 6.52 <.0001
H1 99 -0.78 0.4282 -1.82 0.0715
H2 99 0.71 0.2687 2.64 0.0096
H3
A1 99 1.94 0.3187 6.09 <.0001
A2 99 -0.03 0.3647 -0.08 0.9346
A3 99 1.88 0.3295 5.71 <.0001

 

24

 

 

Hodenpyl Dam ‘

 

   
 

Au Sable River
1)

L.»
at“, ‘1

Manlstee Rlver

  

  
     

 

 

+6. 51 7"
Alcona Dam

 

 

 

Mlo Dam

 

Figure 1. River reach and study site locations on the Manistee and Au Sable

rivers.

25

 

 

 

    

s

rug-r
3 Reference tk
:. .

Figure 2. Split-plot design with paired treatment and reference reaches.

26

9
0|
J

I Mio
I3 Hodenpyl

Proportion
.0 .o .o
N on b

.0
c-I
l

 

 

O
l

100 140 180 220 260 300 340 More

.0 .o .o
00 # 01
l J

Proportion
.0
N

 

 

100 140 180 220 260 300 340 More
Length class (mm)

Figure 3. Length class (in millimeters) frequency distributions for brown trout (A)
and rainbow trout (B) in the Hodenpyl reach and Mio reach.

27

14 I Treatment
12 El Reference

Mean CPUE

 

Brown Rainbow Rock Smallmouth
trout trout bass bass

Species

Figure 4. Overall estimates of mean (i 28E) back-transformed catch per unit
effort of the four most commonly encountered gameﬁsh in the treatment and
reference sections of the combined Hodenpyl and Mio reaches of the Manistee
and Au Sable rivers as determined by electroﬁshing.

28

Hodenpyl

I Treatment
E] Reference

Mean CPUE

 

 

 

._.L.L., Liz.

 

 
 

 

 

 

 

 

 

 

 

Brown Rainbow Rock Smallmouth
trout trout bass bass
Species
Mio
I Treatment
El Reference
I.I.l
:
o.
o
I
I
O
E
I
Brown Rainbow Rock Smallmouth
trout trout bass bass
Species

Figure 5. Overall estimates of mean (i: 28E) back-transformed catch per unit
effort of the four most commonly encountered gameﬁsh in the treatment and

reference sections on each the Hodenpyl reach of the Manistee river and the Mio
reach of the Au Sable river as determined by electroﬁshing.

29

 

 

 

14 _ I Treatment
12 _ El Reference
5‘
3 1O -
c 8 _
8 6 -
E
4 _
2 _
o j—LI ﬁ—L. i I _L T
Brown Rainbow Rock Smallmouth
trout trout bass bass

Specles

Figure 6. Individual bank estimates of mean (i 23E) back-transformed catch per
unit effort of the four most commonly encountered gameﬁsh in the treatment and
reference sections of the combined Hodenpyl and Mio reaches of the Manistee
and Au Sable rivers as determined by electroﬁshing.

IAbove
60
El low
50 Be
,1 40
8 3o
5 2o
10
0
ﬁ «6 e e g e
a; 8 2 2 :9 .9
a 8 3 8 E a
E 0’ E C L
O E l— 9 a)
2 E 5
o. O 0
Lu

Taxon

Figure 7. Estimates of mean (:I: 28E) back-transformed catch per unit effort of the
invertebrate taxa occurring above and below LWD structures of the combined
Hodenpyl and Mia reaches of the Manistee and Au Sable rivers as determined in
monthly behavioral drift samples.

30

Assumed Elevation (feet) Assumed Elevatlon (feet) Assumed Elevation (feet)

Assumed Elevation (feet)

(9
m
I

97-

88

8 E
l l

' " ' 2003

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

 

CD
N

  
 

 

O
(A)
O

o
m
I

97‘

8288
I‘ll

 

‘0
N

eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee

 

 

U

90

120

V

150

U

180

I

210

T

240

270

 

‘3. 8
J

.l

88
CF

8 9
l I

 

(D
N

O
(a)
O

90

I

120

150

180

180

210

210

Distance Down Stream (feet)

31

I

240

270

Figure 8. Survey transects from each of the Mio study sites, showing the
longitudinal river channel for 2002 and 2003. Water surface (2003) is

represented by the dotted top line. LWD structures were present along the entire
transect in 2003.

 

330

 

 

 

 

340 -0—2002
C
330 I.‘ ""2003
5 ' ' .—
u,20 ’ ‘ x
510 A; VI. \
11’ 0 _ ~0—-e’ ‘I. .PI-z?‘
1 2 3 4 5 6 7 3 9 1o 11 12 13 14
2
,3-40-
C
330- 1"
g A‘ -‘
“20" 4 x \
E
1« ‘ ‘
a 0M . . . Pumas—4
1 2 3 4 5 6 7 a 9 1o 11 12 13 14
3
5‘40
0
gap
“20
E10
& o

 

 

 

4
E40
830
“20 ‘. \
S10 ‘C-t-‘h
0 . . . . W
11’; 7 a 9 1o 11 12 13 14

Pebble Count Code
(2=clay, 4=sand, 8=coarse gravel, 12=small boulder)

Figure 9. Substrate size percent frequency distribution for each Mio site in 2002
and 2003.

32

——2002 ' ' " 2003

I
8

(feet

92

 

90~
88‘

Assumed Elevation

 

 

 

86 I I I I I I I I I

 

 

96 .
94 . W J‘
92 ‘W

90-

 

88-

Assumed Elevstlon (feet)

 

 

86 I j 'I I I I I
O 30 60 90 120 150 180 210 240 270 300 330

 

 

 

8

”no.--.--uu......,--..uunn-“nun-I-u-un-u- -. .-
. u

88?

Assumed Elevatlon (feet)
on
on

 

on
O)

180 210 240 270 300
Dlstance Down Stream (test)

 

Figure 10. Survey transects from each of the Hodenpyl study sites, showing the
longitudinal river channel proﬁle for 2002 and 2003. Water surface (2003) is
represented by the dotted top line with the shaded area representing location of
LWD structures in 2002. In 2003 LWD structures were present along the entire
transect.

33

 

 

 

 

 

#1
o
I

8

Percent Frequency
8 B

 

 

 

O

1 2 3 4 5 e 7 a 9 1o 11 12 13 14
Pebble Count Code
(2=clay, 4=sand, 8=coarse gravel, 12=small boulder)

Figure 11. Substrate size percent frequency distribution for Hodenpyl sites one
and two in 2002 and 2003. Hodenpyl site three missing 2002 data.

34

—2002- I I 2003

828

 

m
m
I

E

‘E

3 LA, ‘

g AW vL

a W
1: 90

0

E

3

0

2 I I I I I I

30 60 90 1 20 1 50 1 80 21 0 240 270 300

 

&
O)

 

O

98

 

Assumed Elevation (feet)
3 8 B

S

 

 

0 30 60 90 120 150 180 210 240 270

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

    

98

N
I

m
I

Assumed Elevation (feet)

(13 on (O (O
O
l

 

 

c»
1

I

0 30 60 90 120 150 180 210
DIstance Down Stream (feet)

Figure 12. Survey transects from each of the Alcona study sites, showing the
longitudinal river channel proﬁle for 2002 and 2003. Water surface (2003) is
represented by the dotted top line. Location and size of LWD structures did not
change between 2002 and 2003.

35

Percent frequency

 

0|
0

A
0

Percent Frequency
N 0)
O O

.I
O

O

 

Percent Frequency

 

Pebble Count Code
(2=clay, 4=sand, 8=coarse gravel, 12=smal| boulder)

Figure 13. Substrate size percent frequency distribution for each Alcona site in
2002 and 2003.

36

 

 

 

——2002
25‘ I I I 2W3
\
o 20 - v -. ./‘.» 2
a v‘
g " \va

 

10 . . . . . . . . . . .

 

6/6 6/13 6I20 6/27 7/4 7/11 7/18 7/25 8/1 8/8 8/15 8/22 8129 9/5

 

 

6/19 7/3 7l17 7/31 8114 8/28

 

    
 
  

.”.——- - -—---_ - -
_ _-.-— I-f

 

0204

§15<

 

 

7/13 7/27 8/10 8/24

Figure 14. Mean daily water temperatures in °C during the summers of 2002 and
2003 in the Mio (A), Hodenpyl (B), and Alcona (C) river reaches. Shaded area
represents critical thermal limits at 19°C (upper thermal limit for growth) and
24.7°C (lethal) for brown trout.

37

LITERATURE CITED

38

LITERATURE CITED

Angerrneier, PL and JR. Karr. 1984. Relationships between woody debris and
ﬁsh habitat in a small warm water stream. Transactions of the American
Fisheries Society 113:716-726.

Bettinger, J.M. and P.W. Bettoli. 2002. Fate, dispersal, and persistence of
recently stocked and resident rainbow trout in a Tennessee tailwater. North
American Journal of Fisheries Management 22(2):425-432.

Bilby, RE. and GE. Likens. 1980. Importance of organic debris dams in the
structure and function of stream ecosystems. Ecology 61(5):1107-1113.

DuFour, J.A. 1989. Evaluation of half-logs as habitat improvement structures for
smallmouth bass Micropterus dolomieui and rock bass Ambloplites mpestris.
Master's Thesis. Michigan State University, East Lansing.

Elliot, J.M. 1994. Quantitative ecology and the brown trout. Oxford University
Press, New York, New York.

Fausch, K.D. and TC. Nortcote. 1992. Large woody debris and salmon habitat
in a small costal British Columbia stream. Canadian Journal of Fisheries and
Aquatic Sciences 49:682-693.

Flebbe, PA. 1999. Trout use of woody debris and habitat in Wine Spring Creek,
North Carolina. Forest Ecology and Management 114:367-376.

Gowan, C. and KO. Fausch. 1996. Long-term demographic responses of trout
populations to habitat manipulation in six Colorado streams. Ecological
Applications 6(3):931-946.

Gurnell, A.M., S.V. Gregory, and GE. Petts. 1995. The role of course woody
debris in forest aquatic habitats: implications for management. Aquatic
Conservation: Marine and Freshwater Ecosystems 5:143-166.

Harrelson, C.C., C.L. Rawlins, and JP. Potyondy. 1994. Stream channel
reference site: an illustrated guide to ﬁeld technique. General Technical Report
RM-245. Fort Collins, CO: US. Department of Agriculture, Forest Service, Rocky
Mountain Forest and Range Experiment Station. 61p.

Hubbs, C.L., J.R. Greely, and CM. Tarzwell. 1932. Methods for the

improvement of Michigan trout streams. Bull. No. 1. Institute for Fisheries
Research, University of Michigan, Ann Arbor, MI, USA.

39

Lehtinen, R.M., N.D. Mundahl, and JC. Madejczyk. 1997. Autumn use of woody
snags by ﬁshes in backwater and channel border habitats of a large river.
Environmental Biology of Fishes 4927-19.

Lemly, A.D., R.H. Hilderbrand. 2000. Inﬂuence of large woody debris on stream
insect communities and benthic detritus. Hydrobiologia. 421:179-185.

Mehan, W.R. 1991. Inﬂuences of forest and rangeland management on
salmonid ﬁshes and their habitats. American Fisheries Society Special
Publication 19, Bethesda, MD, USA.

National Research Council. 1996. Upstream: salmon and society in the Paciﬁc
Northwest. National Academy Press, Washington, DC.

Roni, P. and TR Quinn. 2001a. Density and size ofjuvenile salmonids in
response to placement of large woody debris in western Oregon and Washington
streams. Canadian Journal of Fisheries and Aquatic Sciences 58:282-292.

Roni, P. and TR Quinn. 2001b. Effects of wood placement on movements of
trout and juvenile coho salmon in natural and artiﬁcial stream channels.
Transactions of the American Fisheries Society 130:675-685.

Rosenfeld, J., M. Porter, and E. Parkinson. 2000. Habitat factors affecting the
abundance and distribution of juvenile cutthroat (Onchorhyncus clarkr) and coho
salmon (Onchorhynchus kisutch). Canadian Journal of Fisheries and Aquatic
Sciences 57:766-774.

Thevenet, A. and B. Statzner. 1999. Linking ﬂuvial ﬁsh community to physical
habitat in large woody debris: sampling effort, accuracy and precision. Archive
der Hydrobiologica 145:57-77.

Tonello, MA. and A.J. Nuhfer. 2004. Status of the ﬁshery resource report
Manistee River: Hodenpyl Darn to Red Bridge, No. 2004-2.

White, R.J.. 1996. Growth and development of North American stream habitat
management for ﬁsh. Canadian Journal of Fisheries and Aquatic Sciences.
53(Suppl. 1):342-363.

Platts, W.S., and R.L. Nelson. 1988. Fluctuations in trout populations and their
implications for land-use evaluation. North American Journal of Fisheries
Management 8:333-345.

40

   

lljllllljijliljlljjllllljlll