1

cu“

Nut .

 

. u
:7
9
.r

 

Waning:

it v.

 

 

 

MICHIGAN sn

IIIIIIII IIIIIIII II III'IIII' III III

 

 

This is to certify that the

thesis entitled

Distribution of In-lake Coarse Woody Debris Within
Old-growth and Second-growth Forest Settings

presented by

Bruce Martin Peffers

has been accepted towards fulfillment
of the requirements for

 

 

Master of Science degree in Fish. & Wildl.

mg / ﬁ{*’g/2/t/\/

Major professor

Date July 24, 1995

 

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

 

 

LIBRARY I
Michlgan State
University

 

 

PLACE III RETUM BOXto romowthb checkout from your record.
TO AVOID FINES Mum on or before data duo.

DATE DUE DATE DUE DATE DUE
APR 0 2 w);

!___,10 _- ~. m' ‘ II I
I I
DEC 0 5 2001 I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L4H:
II I l

MSU II An math/o Action/EM Oppommlly Inothwon
mm:

 

 

 

 

 

DISTRIBUTION OF IN-LAKE COARSE WOODY DEBRIS
WITHIN OLD-GROWTH AND SECOND-GROWTH FOREST SETTINGS

By

Bruce Martin Peffers

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

1995

ABSTRACT

DISTRIBUTION OF IN-LAKE COARSE WOODY DEBRIS WITHIN
OLD-GROWTH AND SECOND-GROWTH FOREST SETTINGS

By

Bruce Martin Peffers

The amount of coarse woody debris (CWD) found in lakes within old-
growth and second-growth forest settings was investigated to determine if
differences in distribution exist. Three lakes in each forest setting were selected
in the western Upper Peninsula of Michigan. CWD was quantiﬁed using the
line-intersect method, noting length, diameter, orientation to shore, and
orientation to bottom. Volume for each CWD piece was calculated. Terrestrial
variables (percent slope, stem density and diameter at breast height of live trees)
above each transect were also measured.

Estimated mean volume and mean number of pieces per hectare were
greater in old-growth lakes than in second-growth lakes. No set of
characteristics could be determined to adequately deﬁne "second-growth" CWD.
Basal area of the riparian vegetation may have some value for predicting in—lake
CWD volume. We conclude that the second-growth lakes contain less CWD

than the old-growth lakes, which has implications for ﬁsh habitat.

DEDICATION

This thesis is dedicated tomy parents, William and Laurette

Peffers, in thanks for their unconditional support and love.

iii

ACKNOWLEDGEMENTS

I wish to thank my major professor, Dr. Niles R. Kevern, for his guidance
and patience. I would also like to thank my committee members, Dr. Thomas
Coon, Dr. Richard Merritt, and Dr. Clayton Edwards. I have learned so much
from them, and have beneﬁtted from their friendship. I would also like to thank
my (soon-to-be) wife Rita for her love and support, especially during the hectic
writing phase.

This project was funded through a contract with the North Central Forest
Experiment Station, U.S.D.A. Forest Service, of Rhinelander, Wisconsin. My
travel was funded in 1992 by the Fish Habitat Relationship Program of the
U.S.D.A. Forest Service Rocky Mountain Region. I also wish to express my
gratitude to the University of Notre Dame Environmental Research Center,
especially Drs. Ron Hellenthal and Marty Berg, for the use of their beautiful
facilities and their hospitality.

I wish to thank my intern, DawnMarie Welcher, for keeping me in line
and always asking the right questions. Her delight in the out-ofodoors made me
smile again at so many things I had taken for granted.

Thanks are also due to my sister Lisa, her husband Don, and their

children Jessica, Peter, and Luke for their support (and many free meals).

iv

And last (but not least), I wish to thank the employees and "ologists" of
the various National Forests who were so helpful in this project. The National
‘ Forests they work in are the Chequamegon, the Chippewa, the Hiawatha, the
Huron-Manistee, the Nicolet, and the Ottawa. They are (in alphabetical order):
Jeff Allison, Chuck Bassett, Chantel Cook, Jerry Edde, Bob Evans, Beth
Gardner, Don Hair, Joe Kastenholz, Mary-Ann Kipper, Deyna Kuntzsch, Dave
Mandrella, Mark Prout, Sue Reinecke, Bill Reardon, Dave Roerick, Bob

Steuber, and "Black Bart" Tobin. Thanks to all.

TABLE OF CONTENTS

List of Tables ................................................ vii
List of Figures ............................................... ix
Introduction ................................................. 1
Description of Study Areas ...................................... 4
Methods .................................................. 14
Results ................................................... 18
Discussion ................................................. 45
Conclusions ................................................ 58
List of References ........................................... 60

. Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
Table 7.
Table 8.

Table 9.

LIST OF TABLES

Physical parameters of the study lakes: surface area, length
of shoreline, shoreline development (5d). and maximum

depth. .........................................

Mid-summer chemical parameters of the study lakes: pH,
alkalinity (mg/ L CaCO3). speciﬁc conductance (umhos/ cm

@ 25°C), and Secchi depth in meters. .................

Matrix of Chi-square probabilities for differences between -
lakes within settings (combined orientation and aspect count

data'). Probabilities 50.1 are in bold text. ..............

Matrix of Chi-square probabilities for differences between
lakes in orientation (perpendicular/parallel) count data.

Probabilities 50.1 are in bold text. ....................

Matrix of Chi-square probabilities for differences between
lakes in aspect (up/ down) count data. Probabilities 50.1

are in bold text. ..................................

Comparison of means signiﬁcant at 0: =0.05 level for
diameter (D), length (L), and volume (V) of CWD (Tukey’s

Studentized Range Test). ...........................

Signiﬁcant differences (a values) between lakes for
estimated mean volumes per hectare (one-tailed 2 test).

Alpha values 50.1 are in bold text .....................

Signiﬁcant differences (a values) between lakes for
estimated mean pieces of CWD per hectare (one-tailed 2

test). Alpha values 50.1 are in bold text. ...............

Estimated mean volume per piece of CWD for the study
lakes, obtained by dividing the volume per hectare estimate

by the estimated number of pieces per hectare. ..........

vii

Page

.19

.22

. 24

. 25

.26

.31

.33

.35

.37

Table 10. Signiﬁcant differences (a values) between lakes for
estimated mean volumes per diameter size class per hectare
(one-tailed 2 test). Only the ﬁrst three size classes are
shown: 1) 10-20cm, 2)21-30cm,and 3)31-40cm.
Alpha values 50.1 are in bold text. There is no 31 - 40 cm
estimate for Bergner Lake. .......................... 39

Table 11. Signiﬁcant differences (a values) between lakes for
estimated mean number of pieces per diameter class per
hectare (one-tailed 2 test). Only the ﬁrst three size classes
are shown: 1) 10 - 20 cm, 2) 21 - 30 cm, 3) 31 - 40 cm.
Alpha values 50.1 are in bold text. There is no 31 - 40 cm
estimate for Bergner Lake. .......................... 40

Table 12. Estimated mean volume per In2 , estimated mean pieces per
m2 , and actual number of CWD pieces measured for each
transect in the study lakes. ........................... 43

Table 13. Average characteristics of 15 m by 15 m shoreline plots
above lake transects. ............................... 44

viii

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.
Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.
Figure 11.
Figure 12.

Figure 13.

LIST OF FIGURES

location of old-growth and second-growth study areas,
showing study lakes. The lakes are: 1) Bergner, 2) Brown,
3) Crampton, 4) Corey, 5) Katherine, and 6) Snap Jack. . . .

Hydrographic map of Corey Lake, Sylvania Wilderness
Area. ........................................

Hydrographic map of Katherine Lake, Sylvania Wilderness
Area. ........................................

Hydrographic map of Snap Jack Lake, Sylvania Wilderness
Area. ........................................

Hydrographic map of Bergner Lake, UNDERC property. . .
Hydrographic map of Brown Lake, UNDERC property. . . .

Shoreline map for Crampton Lake, UNDERC property.
Known depths are shown in meters. ..................

Depth proﬁles along the longest axis of the study lakes. No
depth data available for Crampton Lake. ..............

Dissolved oxygen and temperature proﬁles of the six study
lakes at midsummer. Solid lines are temperature curves;
dotted lines are dissolved oxygen. ...................

Diameter distributions of CWD for each of the study lakes. . .

Length distributions of CWD for each of the study lakes. . .

Volume distributions of CWD for each of the study lakes. . . .

The distribution of A) second-growth and B) old—growth
CWD volumes per size class, and numbers of pieces for C)

second-growth and D) old-growth. ...................

Page

7

..11

.. 12

.. 13

.. 20

.. 21

28

.. 29

30

.. 32

Figure 14.

Figure 15.

Figure 16.

Estimated mean volume (i one standard deviation) of
coarse woody debris per hectare of littoral zone for each of
the six study lakes .................................. 34

Estimated mean number (i- one standard deviation) of
coarse woody debris pieces per hectare for all study lakes. . . . 36

Volume and piece estimates (at one standard deviation) per

m2 for the ﬁrst three size classes. Lake abbreviations are:

Be - Bergner, Br - Brown, Cr - Crampton, Co - Corey, Ka -
Katherine, and SJ - Snap Jack ......................... 41

INTRODUCTION

The importance of large woody debris (LWD) in terrestrial and aquatic
systems was often overlooked in natural resources management plans prior to
the 1970s, primarily because of the systematic removal of such debris by humans
for centuries, most notably in streams. Deforestation often indirectly reduces the
amount of LWD that enters a system (Fausch and Northcote 1992). Humans
have also altered the natural disturbance regimes by suppression of ﬁre and
ﬂood. Rather than remaining intact for 250 to 500 years, which is the interval
for natural catastrophic events (Triska and Cromack 1980), forests are being cut
on 80-year cycles. Younger, healthier forests produce much less woody debris
than older, mature stands. Thinning of forests decreases woody debris inputs by
suppressing tree mortality (Spies et al. 1988). As a consequence, natural inputs
and distributions of woody debris can only be estimated (Harmon et al. 1986).
There is concern that sufﬁcient LWD is no longer being produced and recruited
into aquatic systems.

A distinction has been made between LWD and coarse woody debris
(CWD) based on the size of the debris. The distinction appears to be relative,
and depends on the author. Large woody debris has been generally deﬁned as
greater than 10 cm in diameter and greater than 1 - 1.5 m in length (Murphy
and Hall 1981, Johnson and Heifetz 1985, Lienkaemper and Swanson 1987,

1

2
Murphy and Koski 1989, Van Sickle and Gregory 1990). Bilby and Ward (1989

& 1991) required the length to be >2 m, and Bryant (1985) defined LWD as
>30 cm diameter and >2 m in length. Coarse woody debris is smaller, from
>2.5 cm diameter (Harmon et al. 1986) to >10 cm diameter (Bryant 1985), with
Robison and Beschta (1990) deﬁning CWD as >02 m in diameter and > 1.5 m
in length. In west coast studies, where trees tend to be larger than 60 cm
diameter breast height (DBH), this may be a necessary distinction. For the
purpose of this study, where old-growth trees rarely exceeded 50 cm DBH,
coarse woody debris was deﬁned as having a diameter greater than or equal to
10 cm, with no limits on length.

Considerable work has been done on woody debris since 1970, especially
its role in stream systems. Gregory and Davis (1992) provide a detailed
overview of previous research with emphasis on channel management. Woody
debris can signiﬁcantly affect the morphology of the channel, retention and
movement of sediment and ﬁne organic matter, and the type of habitat available
for biological communities (Harmon et al. 1986, Andrus et al. 1988, Bilby and
Ward 1989 and 1991). The dynamic nature of stream systems affects the
distribution and movement of woody debris. Once the debris enters the system,
it is subject to currents and ﬂoods, and may not be retained. The role of woody
debris in lake systems is not as dramatic, however, and little work has been done
to document its importance in the lentic environment. Water movement is
limited in lakes, and in-lake CWD has a much greater chance of remaining in

the system. Submerged wood has a very slow decay rate, and is less susceptible

3

to invertebrate breakdown than terrestrial woody debris (Harmon et al. 1986),
leading to the possibility of in-lake CWD persisting for long periods of time if
undisturbed. This may mean that CWD deﬁciencies may be less severe in lake
systems than in stream systems.

It is believed that ﬁsh assemblages found in lakes surrounded by standing
timber have evolved with woody debris as a natural component of the available
habitat. With the advent of logging, this habitat component was assumed to be
in short supply, as has been found in streams (Angermeier and Karr 1984,
Bryant 1985, Sedell et al. 1984). A common ﬁsh management technique has
been the addition of woody structure (whole trees or artiﬁcial constructs) to
subsidize existing lake habitat. The U.S.D.A. Forest Service has used this
technique in many of its lakes in Michigan, Wisconsin, and Minnesota.

The objective of this study was primarily to determine if differences in the
amount and distribution of in-lake CWD existed when comparing lakes
surrounded by old-growth forest to similar lakes within second-growth forests.
Lakes found in old-growth settings should contain amounts of CWD close to
pristine levels, and provide a base-line; lakes surrounded by second-growth
would be CWD-deﬁcient. Secondarily, I wished to determine if attributes of the
surrounding shoreline such as slope, stem density, basal area, and DBH of live
trees could be used to predict the amounts of in-lake CWD, as found by Carlson
et al. (1990) for streams. Such a predictive model would be valuable to the
Forest Service biologists in preparation of the lake management component of

Forest Plans.

DESCRIPTION OF STUDY AREAS

Study Sites

The old-growth site selected for the study was the Sylvania Wilderness
Area (SWA), part of the Ottawa National Forest in Gogebic County of the
western Upper Peninsula of Michigan (centered at 46° 12’ N, 89° 18' W). The
SWA is one of the few forested regions in the upper Great Lakes area that has
remained relatively undisturbed by humans. The Sylvania tract comprises over
7400 ha (18,327 acres) of undisturbed old-growth hemlock-hardwood forest.
Lakes make up about 20 percent of that area. The majority of the Sylvania
lakes are seepage lakes with low chemical concentrations (Crumrine and Beeton
1975). Common tree species found in the SWA are eastern hemlock (Tsuga
canadensir), yellow birch (Betula alleghaniensir), sugar maple (Acer saccharurn),
paper birch (B. papyrifera), red maple (A. rubrum), balsam ﬁr (Abies balsamea),
and white cedar (Thuja occidentalis). Hemlock makes up roughly 55 to 80
percent of the trees >10 cm DBH, and stand ages have been estimated in excess
of 300 years (Tyrrell and Crow 1994). From prior to 1900 up to 1966, the SWA
was in private ownership as a hunting and ﬁshing preserve. In 1966, the tract
was purchased by the USDA Forest Service, and designated a semi-wilderness
area. The tract became the Sylvania Wilderness Area in 1987.

4

5
The second-growth site is the 2975 ha (7345 acres) University of Notre

Dame Environmental Research Center (UNDERC), located 6.5 km west of the
Sylvania Wilderness Area, centered at 46° 13 ' N, 89° 32' E. The property is
bounded on three sides by the Ottawa National Forest. The majority of the
property lies in Michigan’s Gogebic County, but a portion of the UNDERC
property is in Vilas County, Wisconsin. Lakes and bogs make up 16.5 percent of
the surface area (Anonymous 1991). The property has been owned by the
University of N otre Dame since the 19405, and the lakes have been studied since
the 19205. Access to all lakes on the property is restricted and requires the
permission of the UNDERC director. Common tree species found on the
shorelines of UNDERC lakes are black spruce (Picea mariana), eastern
hemlock, white cedar, red maple, balsam ﬁr, and red and white pines (Pinus
spp.). Calculating the forest stand ages using the average hemlock DBH and
regressions developed by Tyrrell and Crow (1994), the UNDERC stands may be
as much as 125 to 160 years old, although smaller hemlock are difﬁcult to age in
this manner. The last known logging occurred in the late 19405 (R. Hellenthal,
pers. comm. 1995).

Both sites lie in a geologically distinctive region, located on the
continental divide between the St. Lawrence and Mississippi river drainages
(Clady et al. 1975). The bedrock is mainly granites and gneisses common to the
Laurentian Shield, covered with glacial deposits of acid, sandy loam from 30 to
60 m thick (Crumrine and Beeton 1975, Spies and Barnes 1985). Nutrient

availability and productivity of these lakes are low, and the lakes are prone to

6

acid rain impact. Kettle lakes are a common feature, and can be found
relatively close-spaced. These adjacent lakes can often exhibit differences in
elevation of 2 - 3 m. Most of the lakes have no inlets or outlets. Underlying

outwash deposits result in steep shorelines along most of the lakes.

Study Lakes

Two lakes from each setting were selected for the study in 1992. A third
lake was added for both settings in 1993 (Figure 1). Lake size was limited to 30
hectares. Corey, Katherine, and Snap Jack lakes are found in the Sylvania
Wilderness Area. Corey Lake (Figure 2) is the smallest of the study lakes at 9.3
ha. Fish species found in Corey include largemouth bass (Micropterus salmoides)
and yellow perch (Perca ﬂavescens). Katherine Lake (Figure 3) is 19.4 ha in
area, located fairly close to the SWA campground. Katherine’s ﬁsh population is
dominated by smallmouth bass (M. dolomieu). Snap Jack (Figure 4) is the
largest old-growth lake at 23.9 ha. A portion of the south shoreline of this lake
was once the site of an old Civilian Conservation Corps camp, and is presently in
old-ﬁeld vegetation. Bluegill (Lepomis macrochirus) are plentiful in this lake, as
are largemouth bass. No-kill, artiﬁcial lure only regulations are in effect for
these lakes (Miller 1992). Bergner, Brown, and Crampton lakes are the
UNDERC lakes used in the study. Bergner (Figure 5) is a two-lobed lake
surrounded on three sides by bog-mat and leatherleaf, with an area of 16.8 ha.

Bluegill and largemouth bass were observed in this lake; ﬁshing is prohibited.

7

An intermittent outlet at the north end of the lake has been blocked by beaver
activity for an undetermined time (vegetation has taken root atop the darn).
Brown (Figure 6) is a wide, shallow lake with poor clarity, 14.6 ha in area. It is
the only lake with a true outlet. Northern pike (Esox lucius'), walleye
(Stizostedion vitreum ), and bluegill are reported in this lake. Crampton (Figure
7) is 29 ha in size, with fairly steep shorelines. largemouth bass are numerous.

Crampton Lake is located in Wisconsin; all other study lakes are in Michigan.

 

 

 

 

 

 

Figure 1. Location of old-growth and second-growth study areas, showing
study lakes. The lakes are: 1) Bergner, 2) Brown, 3) Crampton, 4)
Corey, 5) Katherine, and 6) Snap Jack.

.82 9.08033 «E23 .823 >980 we use 039336»: .N ouawwm

 

an .085: D

Hun” ii... is... it...
«u. was 8 98 o guﬂﬂmpamﬂm
82 82533 £93

93:50

.32 308033 «Enzmm .831— oﬁuoﬁaM «o 9:: Newsagent»:

2.1.52.8 .
«a. 8 o

ammughm

Co; 330 .3

I ". ;
‘ I.

. u... . I

.. .. m .., I...

_ .._,.._: _ .. .2... a3 38.8.5, 3:93

992. 88 ”9.2295
2 «.2 .82
nae: s .38

8.3 05.355.

.m 2&5

10

0,2,“.

2. .

.. .53.: .

,3.

-3.) .
.3,

ans

v I‘

.82 808023» «magam .33 x02. 98m we use oﬁmﬁwougm

2a 38:3 D

2.5.52.8
"nu“
«u. 3o 8 08 o

«M 5

2.28. 53 ”£22m
2 QR ”82

as»: 5 £95
8952? 3:93
9.3 300... nucw

lab; -I

. .5»
,rd--. ..

of ..

NII .

.v 8&5

11

_ "N

to

. , 1 6.1 I

T , 446;

or , ‘ 1 1 3 .
' i 1.5

Bonner Lake

UNDERC Properly
Depth In Motors

l1 9 Area: 16.8 ha
Shonllno: 1899 mun

0"!

L10

5
.
\

4 u - 7.6
4.0”
1f.5 ‘

'2

CI Tramct sun

5

o 61 122
Salolnmohn

a [I

Figure 5. Hydrographic map of Bergner Lake, UNDERC property.

12

 

 

 

Brown Lake
UNDERC Property
Depth In Meter:
Area: 14.6 he
Shoreline: 1703 meters
0 4o e11.5
each In Metere
C] Troneeot Site
Inlet
m 'I p
"2
*- 8w
. L1
I 5.5
a' e
4.6
3 .
to It .1

st"

or" ,

Figure 6. Hydrographic map of Brown Lake, UNDERC property.

13

Crampton Lake
UNDERC Property
Area: 24.7 he
Shoreline: 3312 meters

Figure 7.

6 H

'J

“:7 m 13.3 m

EEO

o-ll

rs

E] Traneeot Site

ScaleinMobn

Shoreline map for Crampton Lake, UNDERC property. Known
depths are shown in meters.

METHODS

Field Sampling

Standard limnological parameters were measured at surface, mid—depth,
and bottom for each study lake. Samples were taken in midwinter, spring (after
turnover), and midsummer (after stratiﬁcation). Samples for alkalinity were
stored in 500 ml polyethylene bottles, refrigerated, and tested as soon as possible
(usually 3 - 5 hours) after sampling. Alkalinity was determined using the acid
titration method and brom-cresol green indicator dye (APHA 1985). Endpoints
were checked periodically with a pH meter. Differential titrations (Novak et al.
1975) were used to determine the proportion of actual CaCO3 buffer. Lake pH
was measured in the field with an Orion model 230A pH meter. Dissolved
oxygen and temperature proﬁles were constructed using a YSI model SOB
oxygen meter. Surface conductivity was measured with a YSI model 33 SCT
conductivity meter. Secchi transparency was measured using a standard 20 cm
black and white quartered disk.

In-lake coarse woody debris was sampled with the line intersect method
(DeVries 1974, Wallace and Benke 1984). Transects 15 m in length were placed
parallel to and approximately 2-3 m from shore. All debris intersecting the
transect line was measured for length, diameter of the larger end, orientation to
shoreline QJarallel or perpendicular), and orientation to bottom (down on the

14

15

bottom or suspended above). Measurements were made to the nearest
centimeter. A piece was considered perpendicular to shore if it lay >45° to the
shoreline; debris was considered off the bottom if any portion of its length was
suspended above the substrate. The number of transects per lake was based on
the area of the littoral zone (1.25 transects per hectare). Since no hydrographic
map could be found for Crampton Lake, the littoral zone area was estimated at
5 ha. The littoral zone was deﬁned as that depth 53 m (10 ft).

On-shore (terrestrial) variables were measured in a 15 m by 15 m plot
immediately above the in-lake transect line. Percent slope of this terrestrial plot
was taken with an inclinometer at two points within the plot, and an average
taken. Diameter at breast height (DBH) was measured for all living trees within
the plot using a DBH tape. Counts of trees by species in the plots were made in ,
conjunction with DBH measurements. Basal area per plot was calculated using
the DBH and stem density per plot data.

Data Analysis
Volume and surface area for each piece was calculated using formulas for

the frustum of a cone:

IE“?! (r12+r1r2+r22) (l)

SA=1t(rl+r) 12+(rl-r7)2 (2)
where r1 is the larger radius, r2 is the smaller radius, and I is the length of the

CWD piece. The smaller radius was calculated with a regression equation

16

derived from actual measurements of fallen trees. The volume and surface area
values were used to calculate volume per hectare and surface area per hectare

estimates with DeVries’ (1974) equation:
X
1?: 1‘. .1 3
2L 2 I, 0

where L is transect length, Xi is the volume or surface area of the ith piece, and
Ii is the length of the ith piece. By substituting 1 for all Xi values, the number of
pieces per hectare can be estimated. Equation 3 is an unbiased estimator of the
true mean quantity per unit area (DeVries 1974). Variance of the estimate can

be given by
2 x 2
var X: —“— —‘ 4
(2.) 2i.) 0
DeVries states that the estimates are independent of width W, and can be
applied to the total littoral zone area.

Frequencies of perpendicular and suspended CWD were tested from lake
to lake and setting to setting using chi-square contingency analysis. Estimated
means of CWD volume per hectare and number of pieces per hectare were
compared using the z-test for differences between means, since Equations 3 and
4 provide only one estimate and variance per lake. To examine distributional

patterns in more detail, CWD was divided into ﬁve 10 cm diameter size classes,

and subjected to the same analysis. Volume, length, and diameter relationships

17

were analyzed as general linear models with analysis of covariance (PROC
GLM, SAS Institute, 1985); the same procedure tested differences between
actual means of the variables with AN OVA for unbalanced samples (Tukey’s
Studentized Range Test). linear and multiple regressions were used to
determine if relationships existed between shore variables and volume and piece

estimates for each transect.

RESULTS

Morphology of the study lakes varied widely (Table 1). Shoreline lengths
ranged from 1310 m for Bergner to 3312 m for Crampton. Maximum depths
ranged from 5.5 m for Brown Lake to 18.3 m for Crampton Lake. Shoreline
development (Sd), an index indicating deviation of the shoreline from the
circumference of a circle having the same area as the lake, showed the study
lakes to be near the normal range (1.5 to 5.0) for natural lakes (Kevem 1989).
Depth proﬁles along the longest axis of each lake (Figure 8) showed that the
bottom gradients for the SWA lakes were relatively steep compared to the two
UNDERC lakes having hydrographic maps. From sonar soundings I have done
on Crampton Lake, I believe that lake to be very similar to Snap Jack in basin
shape.

On-shore slopes above the lakes were usually steep because of their
glacial ice block origin (Crumrine and Beeton 1975). Corey Lake had the
highest average slope at 44.6 percent, which means that the shoreline would rise
44.6 cm for every 100 cm of linear distance away from the lake. Crampton Lake
was also steep-sided, averaging 41.6 percent slope. Katherine and Snap Jack
lakes were intermediate, with average slopes at 20.9 and 31 percent. Bergner
and Brown lakes, with average slopes of 15 and 10 percent respectively, reﬂected
the shallow shape of their basins.

18

19

The study lakes exhibited thermal and dissolved oxygen characteristics
typical of north temperate lakes. Stratiﬁcation occurred by mid-June in all lakes
(Figure 9), and continued through final sampling in September. Dissolved
oxygen levels were usually near zero in the hypolimnion of each study lake.
Metalimnions were found from 2 to 3 m in Bergner, Brown, and Corey lakes.
Metalimnions occurred from 4 to 6 m in Crampton, Katherine, and Snap Jack

lakes.

Table 1. Physical parameters of the study lakes: surface area, length of
shoreline, shoreline development (Sn). and maximum depth.

 

LAKE AREA SHORELINE sd MAXIMUM
(ha) (m) DEPTH (m)

 

Old Growth Lakes - Sylvania Wilderness Area

Corey 9.3 1310 1.21 10.0
Katherine 19.4 2705 1.73 16.8
Snap Jack 23.9 2431 1.40 16.8
Second Growth Lakes - UNDERC Property

Bergner 16.8 1899 1.31 10.0
Brown 14.6 1703 1.26 5.5

Crampton 24.7 3312 1.88 18.3

 

20

.03 noEESU e8 035mg“ 83 ﬁmop oz doe—.2 33m 05 «o mp8 30mg— oﬁ mac? 8an ﬁnen—

Second Growth

 

Old Growth

xoﬂ. 35

.w 05mm

91613111 ul qtdea

 

 

8h

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Bergner Corey
12 r I r . 12 . . .
0 5 10 15 20 25 0 5 10 15 20 25
Dleeolved Oxygen (mg/l.) and Temperature ('C) Dleeolved Oxygen (mg/L) and Temperature (‘0)
0 . 0 .
2 4 ' ' 2 a
4 _. 0' ‘ .—
E ‘ E *
go a E5 ‘:
a — a a:
10 _ 1o _
12 Brown 12 1 Katherine
0 5 1‘0 115 20 25 0 5 10 1'5 20 25
Dleeolved Oxygen (mg/L) and Temperature ('C) Dleeolved Oxygen (mg/L) and Temperature (‘0)
o . ' O
2 ‘* i 2 _
4 - 4 _
£5 1 ﬁe -1
8 ‘ r3 5
a - a 4 °
10 ~ 10 —
12 Crampton ‘2 SnapJack
o is 1‘0 1'5 20 25 o 3 10 {a 20 25
Dleeolved Oxygen (mg/L) and Temperature (‘0) Dleeolved Oxygen (mg/L) and Temperature ('C)

Figure 9. Dissolved oxygen and temperature proﬁles of the six study lakes at
midsummer. Solid lines are temperature curves; dotted lines are
dissolved oxygen.

22

Chemical characteristics of the study lakes were also consistent with other
lakes of this region, with the exception of Brown Lake (Table 2). Excluding
Brown Lake, average pH was 5.7 and speciﬁc conductivity averaged 22
umhos/ cm at 25C. Alkalinity averaged 3.5 mg/L CaCO3. Differential titration
performed on samples from Bergner, Crampton, Corey, and Katherine lakes
demonstrated that most of this buffering was in the form of organic acids.
Brown Lake was consistently higher for all parameters. Most of the lakes were
stained to some degree, possibly due to the organic acid content. Secchi depth

averaged 2.8 m for all lakes. The lowest value was for Brown Lake, with a

Table 2. Mid-summer chemical parameters of the study lakes: pH, alkalinity
(mg/ L CaCO3), speciﬁc conductance (umhos/ cm @ 25°C), and
Secchi depth in meters.

 

LAKE pH ALKALINITY SPECIFIC SECCHI
CONDUCT AN CE DEPTH

 

Old Growth Lakes - Sylvania Wilderness Area

Corey 5.8 4.2 33.8 2.5
Katherine 5.4 3.1 21.4 2.25
Snap Jack 6.3 3.0 22.0 3.5
Second Growth Lakes - UNDERC Property

Bergner 5.6 3.2 13.6 2.5
Brown 7.6 65.0 127.8 1.5

Crampton 6.1 3.8 18.9 4.75

 

23
Secchi depth of 1.5 m. This lake was normally very turbid, and Secchi depths as

low as 0.4 m have been recorded (Anon. 1991). Crampton Lake was the most
transparent, with a Secchi depth of 4.75 meters.

A total of 134 pieces of CWD was measured in 17 transects in the SWA
lakes. Of these pieces, 117 (87.3%) lay perpendicular to the shoreline. Only 25
pieces (18.7%) were off the lake bottom, 23 of which were also perpendicular to
shore. A total of 59 pieces were measured along 23 transects in the UNDERC
lakes, with 47 pieces (79.7%) being perpendicular to shore. Only 19 of these
pieces (32.2%) were off the bottom, and also perpendicular to the shoreline.

By-lake comparisons (chi-square) of the frequencies of the combined
orientation (perpendicular to shore/parallel to shore) and aspect (up off
bottom/ down on bottom) distributions are detailed in Table 3. Data were
separated into four categories: perpendicular-up, perpendiculapdown, parallel-
up, and parallel-down. Brown Lake differed signiﬁcantly from both Bergner and
Crampton lakes, with a greater frequency of perpendicular-down CWD. Snap
Jack lake was signiﬁcantly different from Katherine Lake, with greater
frequencies of parallel-down and perpendicular-down CWD. Due to the low
sample size for the parallel-up category, X2 was inappropriate when including
this category. Only two pieces were recorded in this category, both in Corey
Lake. No signiﬁcant difference (p =0.343) was found between Corey and Snap
Jack lakes after removal of this category. Between settings, combined second-
growth distributions differed signiﬁcantly from old-growth (p=0.012). A greater

percentage of CWD in the perpendicular-up classiﬁcation was found in the

24

second-growth setting lakes, while a larger percentage of CWD in the

perpendicular-down classiﬁcation was found in the old-growth setting lakes.

Table 3. Matrix of Chi-square probabilities for differences between lakes
within settings (combined orientation and aspect count data ).
Probabilities 50.1 are in bold text.

 

 

 

 

LAKE Brown Crampton Second-growth combined
Bergner 0.042 0.209
Brown 0.041
Katherine Snap Jack
Corey 0.381 0.150
Katherine 0.046
Old-growth combined 0.012

 

U
Orientation: perpendicular or parallel; Aspect: up off bottom or down on bottom.

Orientation and aspect were tested separately to determine which
characteristic contributed most to the differences in Table 3. Brown Lake
differed signiﬁcantly in orientation from both Bergner and Crampton lakes, with
a higher percentage of perpendicular CWD (Table 4), consistent with the
differences seen in Table 3. However, Table 5 showed the reverse of Table 3
for upright CWD frequencies of the second-growth lakes: Crampton Lake had a

higher frequency of upright CWD than Bergner Lake, while Brown Lake was not

25
signiﬁcantly different from either Bergner or Crampton lakes. Old-growth lake

results in Tables 4 and 5 agree with Table 3: Snap Jack Lake’s percentage of
perpendicular CWD was lower than Katherine’s (p =0.100). The frequency of
CWD down on the bottom in Snap Jack Lake was much greater than Katherine
Lake (p =0.034). Between settings, second-grth differed from old-growth in

both orientation (p=0.170) and aspect (p=0.039).

Table 4. Matrix of Chi-square probabilities for differences between lakes in
orientation (perpendicular/parallel) count data. Probabilities 50.1
are in bold text.

 

 

 

LAKE Brown Crampton Second-growth combined
Bergner 0.012 0.351
Brown 0.070
Katherine Snap Jack
Corey 0.255 0.497
Katherine 0.100

 

Old-growth combined 0.170

 

26

Table 5. Matrix of Chi-square probabilities for differences between lakes in
aspect (up/ down) count data. Probabilities 50.1 are in bold text.

 

 

 

 

LAKE Brown Crampton Second-growth combined
Bergner 0.378 0.078
Brown 0.280
Katherine Snap Jack
Corey 0.387 0.195
Katherine 0.034
Old-grth combined 0.039

 

Regression analysis of the diameters of downed trees demonstrated an
exponential relationship between the distance from the base (d) and percent of
base diameter remaining(%R): %R =0.867e'°'°°°9&‘ (rz=0.998). This relationship
was used to estimate the terminal diameter of each piece for volume and surface
area calculations. .

Box-and-whisker plots (boxplots) described by Hartwig and Dearing
(1979) are used to illustrate the actual distributions of diameter, length, and
volume CWD measures. The boxplots show important percentiles and mean,
maximum, and minimum values at a glance. The rectangle of the plot is
bounded by the 25th and 75th percentiles, the central horizontal line is the
median, the small square is the mean, the whisker ends are the 5th and 95‘“

percentiles, and the small circles are the extreme values.

27

Diameter distributions by lake (Figure 10) show Brown Lake to have
more larger diameters than all other lakes; the other second-growth lakes are
skewed to the smaller diameters. Length distributions by lake (Figure 11)
likewise show CWD was longer in Brown Lake compared to all other lakes,
while the other second-growth lakes were skewed to shorter lengths. Volume
distributions by lake (Figure 12) are skewed to smaller values in all lakes, but
Brown lake again had a larger proportion of large volumes. Brown Lake
distributions were dissimilar from those of all other study lakes for all three
measures.

Tukey’s test of differences between means found differences between _
several lakes (Table 6). For the diameter and volume means, Brown lake was
signiﬁcantly larger than all lakes but Katherine lake (a=0.05). For the variable
length, the Brown lake mean was signiﬁcantly larger than those of Bergner and
Crampton lakes (a =0.05) only. Mean length for Bergner Lake was signiﬁcantly
smaller when compared to Katherine lake (a=0.05).

Analysis of covariance (general linear models) was performed using
diameter and length as covariates, with lake and interaction of lake and setting
as main effects. Brown lake was found to be signiﬁcantly different from both
Bergner and Crampton lakes (p =0.0004) within the second-growth setting.
Crampton lake was not signiﬁcantly different from the old-growth lakes
(p =0.3096). Katherine Lake was signiﬁcantly different from Snap Jack lake

0, =0.0858) within the old-growth setting.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.6
0.5
0.4
E
g
'5
0.3 =
0.2 '
0.1 l . | .
n=13 n=21 n=25 n=42 n=26 n=66
Bergner Brown Crampton Corey Katherine Snap Jack
Figure 10.

Diameter distributions of CWD for each of the study lakes.

29

20

15—

 

 

 

 

Length (m)
3
l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I O

Bergner Brown Crampton Corey Katherine Snap Jack
n=13 n=21 n=25 n=42 n=26 n=66

Figure 11. Length distributions of CWD for each of the study lakes.

30

Volume (1113)
i

 

 

0.5 —

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 I; o . -
Bergner Brown Crampton Corey Katherine Snap Jack
n=13 n=21 n=25 n=42 n=26 n=66

 

 

 

 

 

 

. l .

A; n

 

Figure 12. Volume distributions of CWD for each of the study lakes.

31

 

 

Table 6. Comparison of means signiﬁcant at a=0.05 level for diameter (D),
length (L), and volume (V) of CWD (Tukey’s Studentized Range
Test).
LAKE Bergner Crampton Corey Katherine Snap Jack
Brown D,L,V D,L,V D,V D,V
Katherine L

 

For portions of the analysis, CWD was divided by diameter into ﬁve size
classes in 10 cm intervals. Smaller diameter pieces constituted the majority of
CWD in all lakes. Over 62% of the total number of CWD pieces in second-
growth lakes was contained in the 10 - 20 cm size class (Figure. 13). The
percentage increased to 79.6% when the next larger size class was added. For
the old-growth setting, the 10 - 20 cm size class accounts for 51.5% of the total
number of pieces, increasing to 82.1% when including the next size class. The
contribution of total CWD volume for the first size class was 10% for second-
growth and 14% for old-growth. Adding the next size class increased the
percentage to 28% and 43% for second-growth and old-growth, respectively.
The single size class with the greatest contribution to second-growth total
volume was the 31 - 40 cm size class, with 28% of the CWD volume accounted
for by 11.9% of the pieces. The 21 - 30 cm size class, with 30.6% of the pieces,

accounted for 29.2% of the old-growth CWD volume.

32

.goewéo An end goEﬁoSB 6 new
moooa mo €383 on“ .320 emu Hum 8829 950 £38338 3 can 53056508 2 we newsstand 2F

3&0 3m 5an

"Wm mo; :0 rum-d

 

:oqwnN 131°er rum-d

 

 

.3 25mm

OWHIOA M01. 1° W'Oﬂd

 

 

owniOA M01. 1° tum-d

 

33

Estimates of the mean volume of CWD per hectare calculated with
Equation 3 (Figure 14) are shown with one standard deviation. All second-
growth lake volume estimates were signiﬁcantly different from each other (Table
7). Bergner lake and Crampton lake, with the lowest volume estimates, were
signiﬁcantly different from all other study lakes. Corey and Snap Jack estimates
for mean volume of CWD were signiﬁcantly higher than all second-growth lakes.
The volume estimate for Katherine lake was similar to Brown lake (or =0.3446),
signiﬁcantly higher than Bergner (a=0.0005) and Crampton (a=0.0368) lakes,

and signiﬁcantly lower than Corey (a=0.0885) and Snap Jack (a=0.1021) lakes.

Table 7. Signiﬁcant differences (a values) between lakes for estimated
mean volumes per hectare (one-tailed 2 test). Alpha values 50.1
are in bold text.

 

 

LAKE Bergner , Brown Crampton Corey Katherine
Brown 0.0020
Crampton 0.0047 0.0885
Corey <0.0001 0.0401 0.0005
Katherine 0.0005 0.3446 0.0368 0.0885

Snap Jack <0.0001 0.0427 0.0001 0.4139 0.1021

 

34

 

 

 

 

300
250 -
Corey ?
_ Snap Jack .
200 _ _ —r-—
g —l——
I _ __
é Katherine 4
«E 150 -
E Brown 6
2 _
9 __
100 —- __
— . Crampton
50 ._..
E Bergner
0

 

 

Figure 14. Estimated mean volume (i one standard deviation) of coarse woody
debris per hectare of littoral zone for each of the six study lakes.

35

Estimates of the mean total number of pieces of CWD per hectare
calculated with Equation 3 (Figure 15) are shown with one standard deviation.
Brown lake piece estimates were signiﬁcantly lower than all other lakes (Table
8). Bergner lake piece estimates were signiﬁcantly lower than Crampton
(a=0.1021) and Corey (a =0.0808) lakes, but not Katherine or Snap Jack lakes.
The Crampton Lake piece estimate was similar to the old-growth lake estimates.

There were no signiﬁcant differences in total piece estimates between old-growth

 

 

lakes.
Table 8. Signiﬁcant differences (a values) between lakes for estimated
mean pieces of CWD per hectare (one-tailed 2 test). Alpha values
50.1 are in bold text.
LAKE Bergner Brown Crampton Corey Katherine
Brown 0.0376
Crampton 0.1021 0.0062
Corey 0.0808 0.0003 0.4169
Katherine 0.2878 0.0736 0.3300 0.3745

Snap Jack 0.2389 <0.0001 0.1698 0.1492 0.4052

 

36

 

 

 

 

 

 

5,000
4,000 — ‘-
b Crampton T
0 Corey e
g 3,000 ~—
8. Katherine 0
g r
E -J— Snap Jack
o
g 2,000 ——
OBergner
1,000 — -L
* Brown i
0

 

Figure 15 . Estimated mean number (i one standard deviation) of coarse woody
debris pieces per hectare for all study lakes.

37

Division of the estimated mean volume per hectare by the estimated
mean number of pieces per hectare yields the estimated mean volume per piece
of CWD for each lake (Table 9). Bergner and Crampton lakes had very small
values. The old-growth setting lakes had slightly larger volume per piece
estimates. The highest estimate was obtained for Brown lake, in the second-

growth setting.

Table 9. Estimated mean volume per piece of CWD for the study lakes,
obtained by dividing the volume per hectare estimate by the
estimated number of pieces per hectare.

 

 

 

SETTING LAKE

Second- Growth Bergner Brown Crampton
volume (m3) 0.0131 0.2861 0.0226
Old- Growth Corey Katherine Snap Jack
volume (m3) 0.0768 0.0606 0.1013

 

Comparisons of volume estimates by diameter size class provide greater
detail into differences in distribution. Due to the small sample size for the

larger diameter size classes (>40 cm), only the ﬁrst three size classes are used in

38

Table 10 and Figure 16. Within-setting comparisons of volume estimates show
that all second-growth lakes were equivalent for the 21 - 30 cm size class, where
total volume estimates were signiﬁcantly different in Table 7. The 10 - 20 cm
size class volume of Crampton lake was signiﬁcantly higher than both Bergner
and Brown lakes (0: =0.003). The 31 - 40 cm size class volume of Brown lake
was signiﬁcantly higher than Crampton Lake (a=0.041). No estimate was made
for Bergner lake for the 31 - 40 cm size class because no CWD in that class was
measured.

Comparisons of mean piece number estimates by diameter size class also
provides increased resolution of differences in distribution (Table 11, Figure 16).
Within-setting comparisons of piece estimates show that all second-growth lakes
were again equivalent for the 21 - 30 cm size class, where total piece estimates
were signiﬁcantly different (Table 8). The estimate of the mean number of
pieces in the 10 - 20 cm size class for Brown Lake was signiﬁcantly lower than
those for Bergner (a=0.027) and Crampton (a=0.089) lakes. Within the old-
growth setting, no signiﬁcant differences were found between lakes, conﬁrming

the mean total piece comparison results of Table 8.

39

 

 

 

 

 

 

Table 10. Signiﬁcant differences (a values) between lakes for estimated
mean volumes per diameter size class per hectare (one-tailed 2
test). Only the ﬁrst three size classes are shown: 1) 10 - 20 cm, 2)
21 - 30 cm, and 3) 31 - 40 cm. Alpha values 50.1 are in bold text.
There is no 31 - 40 cm estimate for Bergner lake.
Size Class Bergner Brown Crampton Corey Katherine

1 0.456

Brown 2 0.195
3
1 0.003 0.003

Crampton 2 0.421 0.245
3 0.041
1 0.0001 0.0001 0.068

Corey 2 0.001 0.005 0.002
3 0.319 0.054
1 0.062 0.058 0.119 0.007

Katherine 2 0.014 0.061 0.019 0.104
3 0.309 0.051 0.488
1 0.0001 0.0001 0.252 0.157 0.026

Snap Jack 2 0.001 0.011 0.001 0.136 0.371
3 0.475 0.019 0.390 0.379

 

40

 

 

 

 

 

 

Table 11. Signiﬁcant differences (a values) between lakes for estimated
mean number of pieces per diameter class per hectare (one-tailed
2 test). Only the ﬁrst three size classes are shown: 1) 10 - 20 cm,
2) 21 - 30 cm, 3) 31 - 40 cm. Alpha values 50.1 are in bold text.
There is no 31 - 40 cm estimate for Bergner lake.
Size Class Bergner Brown Crampton Corey Katherine
1 0.027
Brown 2 0.295
3 .
_ 1 0.089 0.004
Crampton 2 0.224 0.281
3 0.020
1 0.201 0.002 0.236
Corey 2 0.042 0.003 0.001
3 0.184 0.056
1 0.360 0.100 0.252 0.437
Katherine 2 0.215 0.039 0.020 0.138
3 0.239 0.062 0.402
1 0.452 0.0001 0.078 0.184 0.375
Snap Jack 2 0.084 0.003 0.001 0.275 0.271
3 0.285 0.008 0.268 0.352

 

.xoﬂ. mam - B was 6:553— - «M $0.80 - oU doEBEU - 5 .85on - em .eonmeom - um ”03

8033.659“ 03 .8820 on? ooh: mew 05 e8 «8 Sq Acetagoc 93525. 28 3 83850 80E was 083.3 .3 onE

 

 

41

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

@ H @ s o . . c
a E _ E e u
a _ a a a , a i a
x a , , s...
, hi no i 4 l ..
® @ good
i O i
i it «.0 H MM mW i @@ A- 08.0 W
a w. .- a - W.
-- m e .
1.. no .. .ﬁ 08...
av -
m0 8...
to
l -e . «66
.5915 808.8 58.9 609.5 58.8 .889

 

 

42
Care should be taken when considering the calculated volume per hectare.

and piece per hectare estimates. The calculated values give the impression of
evenly-distributed CWD across the littoral zone. This is not the case, as the
CWD more frequently displayed clumped distributions. The number of pieces
within Snap Jack lake ranged from 1 to 16 pieces per transect; half the transects
in Bergner lake contained no CWD at all. The volume and piece estimates per
transect (Table 12) show that some second-growth transects contain as much
CWD as old-growth transects, and some old-growth transects were as sparse as
second-growth transects. The quality of the CWD available as habitat may be
equal between the two settings; old-growth lakes simply have more of this quality
habitat available.

Average on-shore stem densities of the live trees 210 cm DBH for each
second-growth lake ranged from 3.2 to 10.1 stems per 15 m by 15 In pilot (Table
13). Old-growth lake stem densities ranged from 6.6 to 11.8 stems per plot.
Averages by setting were 6.3 stems per plot for second-grth and 9.6 stems per
plot for old-growth. Total basal area (TBA) of trees 210 cm DBH ranged from
0.19 m2 to 0.61 in2 per plot for the second-growth lakes, and 0.83 m2 to 1.51 in2
per plot for the old-grth lakes. Average TBA for each setting was 0.32 m2 per

plot for second-growth and 1.18 m2 per plot for old-growth.

43

 

 

 

o 036 386 o
H HH66 3666 m
HH 3N6 836 w v 366 6866 v
a HH.N6 886 h m 5N6 $.86 m
N $66 H566 o 6 66 66 N
w 83 38.: m m SS 885 H 82.56
H 866 6866 H. N >86 6866 H.
H 086 6866 m a $2.. 386 6
NH me6 836 N v >86 m§6 m
o 6666 886 H H03. 98m H $66 6686 H.
c 866 886 m m 3.66 3666 m
m 686 386 v H 686 H.866 N
n on.6 2.86 m H 366 H366 H .565
H 866 N866 N m «No.6 886 6H
6 >36 NNN66 H 0&2:an v $6 3666 w
o 366 6686 v . m NHN6 $666 6
NH H36 6686 m N 3.66 9666 v
3 wwm6 09.66 N 6 66 66 mﬁde
o NNm. 6 386 H .HoeoO H 63.6 8666 H eonwnum
60:50on « E\000E « E\._o> 6003...; \ 03 605302 a E \ 000E « E\._o> Sum—HEP \ 0013
530-20 5.80-283

 

.803 >25 05 E 80985 :80 e8 605808
800E D30 Ho 0956 H.033 98 .NE .06 $003 508 603850 .«8 H06 0839» 508 60:25me

:2 03g.

44

Table 13. Average characteristics of 15 m by 15 m shoreline plots above lake

 

 

transects.
LAKE PERCENT STEM TOTAL BASAL
SLOPE DENSITY AREA (m2)
Corey 44.6 11.8 1.51
Katherine 20.9 6.6 0.83
Snap Jack 31.0 10.4 1.24
Old-growth Average 31.2 9.6 1.18
Bergner 15.1 3.2 0.19
Brown 10.4 10.1 0.61
Crampton 41.6 7.2 0.20
Second-growth Average 20.6 6.3 0. 32

 

Basal area and estimated volume exhibited a weak linear relationship
(r2=0.26). A larger sample size may show that basal area has predictive
potential for estimating in-lake CWD volume. Multiple regression with
estimated volume as the dependent variable, and percent slope, basal area, and
stem density as the independent variables resulted in an r2 of 0.34; larger
samples may be able to fine-tune these relationships as a predictive model for

CWD volume.

DISCUSSION

The difference in morphologies of the study lakes between settings is
most probably an artifact of random chance rather, than an actual distinction.
The old-growth lakes selected tend to be steeper-bottomed and deeper, but an
examination of available hydrographic maps show that not all old-grth lakes in
the SWA are similar. Likewise, the second-growth lakes selected tend to be
shallow-sloped and less deep, but maps of other UNDERC lakes show a variety
of morphologies. The depth of a lake may have an inﬂuence on the amount of
CWD remaining in the littoral zone: Moring et al. (1986) described the gradual
migration of submerged logs to deeper portions of a reservoir due to wave and
ice action. Such logs would be less important ecologically because summer
dissolved oxygen levels are usually very low in deeper water. In addition, the
low dissolved oxygen levels in the hypolimnions of the study lakes in midsummer
indicates the presence of oxygen-consuming materials in the deeper portions of
the lakes (Crumrine and Beeton 1975), which may include woody debris that had
migrated.

Shoreline slopes above the transects averaged higher for old-growth
settings, but this would again be due to random chance. Crampton Lake, in the
second-growth setting, averaged 41.6 percent slope, at the upper end of the

45

46

range for slopes in this region; the average slope found in the SWA has been
reported as less than ten percent (Spies and Barnes 1985). The average slopes
of the study lakes were consistently higher than this. The two lowest average
slopes, 15.1% (Bergner) and 10.4% (Brown), are found in the second-grth
setting.

The seCond-growth setting shoreline plots averaged lower for both stem
densities 210 cm DBH and total basal area. This is not surprising, since the old-
growth trees are larger diameter, and have had more time to attain this size.
What was not recorded were the number of trees below 10 cm DBH observed,
which would have increased the second-grth stem densities to numbers
surpassing the old-grth densities. Dense stands of balsam ﬁr 3 - 7 cm DBH
were common along the second-growth shorelines, but were absent along the
old-growth shorelines. These saplings would not increase the total basal area
signiﬁcantly.

The study lakes all shared similar chemical characteristics, with the
exception of Brown Lake. Low pH, low alkalinities, and low conductivities were
the rule for the ﬁve other lakes. The differential titrations performed on
samples from Bergner, Crampton, Corey, and Katherine lakes demonstrated that
most of the alkalinity measured in these lakes is actually organic acid buffering,
and not CaCO3 alkalinity. Organic acids are by-products of the decomposition
of organic compounds (such as woody debris), and are also responsible for
stained (colored) water, which occurs in all the study lakes except Brown Lake.

Brown Lake, with consistently higher pH, alkalinity, and conductivity (Table 2)

47

was the only study lake with surface inlets and an outlet. Groundwater
throughput is probably the cause of these higher limnological values (Crumrine
and Beeton 1975) and may have been a factor in the differences in CWD
distributions to be discussed later. The balance of the study lakes had chemical
character similar to precipitation, suggesting little groundwater input.

Recruitment of trees into the water occurs primarily at or near 90° to the
shoreline due to a variety of factors: phototropism, bank undercut, slope of the
shoreline, and the tendency to produce more biomass on the unshaded side
(Robison and Beschta 1990). When trees enter the water naturally, the trunk
usually remains associated with the rootwad, anchoring it to shore (Bryant 1985),
and props a major portion of the trunk off the lake bottom. Branches extending
from the trunk can also hold the debris off bottom. After some time in the
water, a combination of decay of branches, biotic activity, sedimentation, and
water movements can alter the position or elevation of the CWD. An
assumption can be made that pieces of CWD found parallel to the shoreline
have moved into this orientation at some point since entering the system (Bilby
and Ward 1989), and that CWD lying flat on the bottom has been in the system
for a long time.

In the old-growth setting, Corey Lake had the highest values for the
terrestrial variables. This lake was the smallest of the study lakes, entirely
surrounded by steep (nearly 45%) slopes covered in relatively dense (averaging
11.8 stems per plot) stands of very large trees. The estimate of total volume per

hectare was the highest for all study lakes, with the 10 - 20 cm and 21 - 30 cm

48

size classes contributing the largest portion of that volume. The estimated total
number of pieces was also high, but not signiﬁcantly different from other old-
growth lakes. {The majority of the piece estimate came from the 10 - 20 cm and
21 - 30 cm size classes. The relatively dense stem densities above this lake may
account for the greater volume per hectare estimate within Corey Lake.

Percent slope, stem density, and basal area values for Katherine Lake
were the lowest of the old-growth lakes. Estimated total volume reﬂects this,
being the lowest estimate for old-growth lakes, signiﬁcantly lower for the 10 - 20
cm size class. Estimates of number of pieces per hectare were similar to other
old-growth lakes. Frequencies of upright and perpendicular CWD suggest that
recruitment of the sampled debris was recent. The basin shape of Katherine
Lake is the steepest of the study lakes (Figure 8), and older CWD may have
migrated to the deeper areas of the lake. Attempts to verify this hypothesis by
SCUBA were unsuccessful; visibility below the thermocline was essentially zero.

The CWD in Snap Jack Lake had the highest percentage of pieces
parallel to the shoreline, and the least percentage of pieces up off the bottom,
suggesting that the CWD measured had been in the lake for some time. Snap
Jack’s terrestrial stem density and basal area values were higher than Katherine
lake, but lower than Corey Lake. The estimated volume per hectare was
similar to that of Corey Lake, but the estimated number of pieces was lower
than the other old-grth lakes. The basin shape of this lake deepens gradually
(Figure 8), and migration of CWD may not be as easily accomplished as in

Katherine Lake.

49

Bergner Lake in the second-growth setting had the lowest percentage of
perpendicular CWD, and a low percentage of upright CWD as well, suggesting
that the CWD within the lake had been present for some time, and had possibly
relocated from its original entry point. Bergner Lake, with the lowest basal area,
stem density, and low percent slope values for the shoreline plots, would be
expected to have very low recruitment rates for CWD. The estimated CWD
volume value suggests this low rate. The lack of larger, slower-to-decay CWD in
this lake suggests a long time interval since old-growth trees were found close
enough to the lake to be recruited into the system, or that any residual CWD
may have been covered by the encroaching bog mat. The intermittent outlet (at
present dammed by beavers) may also have provided emigration for some CWD
in the past. The debris found in Bergner Lake probably was brought in via
beaver activity, as most of the CWD was found in transects near the dam.

Crampton Lake (second-growth) had a low percentage of perpendicular
CWD, but the highest percentage of suspended CWD of all the study lakes. All
of the up-off-bottom CWD was also perpendicular to shore. This suggests recent
recruitment of CWD. The shoreline surrounding Crampton Lake has a very
high percent slope, but the shore plot stem density and basal area values were
relatively low. Any recruitment of CWD would necessarily be in the smaller
diameters. The volume and number of piece estimates bear this out: the major
contribution to CWD was found in the 10 - 20 cm diameter class (Figure 16).

Brown Lake, a second-growth lake, is more similar to old-growth lakes in

shoreline and CWD characteristics. Stem density and basal area values were

50
higher than the other second-growth lakes. Over 95% of the CWD found in the

lake was perpendicular to shore, and 28.6% was suspended off the bottom,
nearly identical to Katherine Lake in the old-grth setting. yAll of the
suspended CWD was perpendicular. The other second-growth lakes had low
frequencies of perpendicular CWD, signiﬁcantly different from old-grth
frequencies (a=0.072), but the frequency of perpendicular CWD in Brown Lake
was not signiﬁcantly different from old-grth (a=0.185). The CWD sampled in
Brown Lake had the highest average volume of all the study lakes. Estimated
volume per hectare in Brown Lake was higher than the other second-growth
lakes, but the estimate of CWD pieces per hectare was much lower than all
other lakes. This would correspond to an estimated mean volume per piece of
0.2861 m3, much higher than all other lakes (Table 9). For this to be true, the
CWD in Brown Lake must be consistently longer for any given diameter. The
average dimensions of CWD sampled in Brown Lake (Figures 10 and 11) were
consistently greater than the other study lakes; the AN COVA results confirmed
that the relationship of diameter and length for Brown Lake was different from
all other lakes. Bilby (1984) and Benke and Wallace (1990) found that the
longer the piece of CWD, the less likely it is to move from the system. The low
number of pieces and relatively large volume of each discrete piece of CWD in
Brown Lake would suggest that smaller CWD in the lake has been moved from
the lake. Brown Lake does have inlets and an outlet; movement of water
through this lake may be a factor in this distinctive distribution of CWD.

Becauseof the higher pH and lower organic acid content in this lake,

51

decomposition of CWD may occur at a faster rate (Wetzel 1983) than the other
study lakes, with the smaller pieces breaking down ﬁrst. The shape of the lake

basin and relative lack of wind shelter causes a strong fetch effect (Anonymous

1991), which may also move the CWD away from its original entry point.

Older stands of trees have been proven to produce more woody debris
than younger, healthier stands (Andrus et al. 1988), and the trees of the second-
growth setting of this study contained a higher percentage of deciduous species
which have been shown to produce less CWD than coniferous stands (Harmon et
al. 1986). The lakes in the old-grth setting of this study contained more CWD
pieces and volume than the second-growth lakes, which agrees with several
studies comparing streams in old-growth vs. second-growth settings (Bilby and
Ward 1991, Murphy and Koski 1989, Murphy and Hall 1981), but the percent
difference was much less severe than seen in streams. The average second-
growth volume per m2 was 37.6% of the average old-growth volume for the study
lakes, compared to an average of 25% for western stream studies; the average
number of pieces of CWD per m2 in second-growth was 70% of the old-growth
average for the lakes, contrasted with 50 - 56% for stream systems.

Corey and Snap Jack lakes were the most similar in CWD characteristics
of the old-growth lakes studied. Shoreline characteristics were also similar for
these two lakes. Katherine Lake, with lower values for shoreline variables, had
'less potential recruitable volume than the other old-growth lakes, and had less
estimated total volume. The estimated number of pieces per hectare was not

signiﬁcantly different among old-growth lakes, meaning the volume per piece for

52
Katherine Lake would also be less than the other old-growth lakes. However,

due to the small number of lakes sampled, it would be inappropriate to conclude
that Katherine Lake was "less old-growth" in nature than the other lakes; rather
it may be that Katherine Lake was found at one end of a range of characteristics
for old-growth.

The lakes in the second-grth setting were more dissimilar than similar
for CWD characteristics. No combination of in-lake CWD characteristics or
size-class distribution could be used to classify a lake as "second-growth" in
nature. Brown Lake, with inlets and an outlet, low-sloped shoreline, and shallow
basin shape demonstrated that the CWD content of a lake depends on more
than the composition and age of the surrounding forest stands. The small
number of lakes used in this study increase the difﬁculty of making any

meaningful generalizations about either setting.

Importance of CWD as Habitat

Although invertebrate and ﬁsh abundance were not measured directly in
this study, a basis exists to make inferences from the habitat potential of the
CWD (Carlson et a1. 1990). In stream systems, many studies have shown that
increases in CWD resulted in increased secondary production (Benke and

Wallace 1990), increased invertebrate abundance (Angermeier and Karr 1984),

53
increased invertebrate diversity (Murphy and Hall 1981), increased ﬁsh

abundance, diversity, and numbers of larger ﬁsh (Angermeier and Karr 1984,
Murphy and Hall 1981, Sedell et al. 1984).

Benke and Wallace (1990) estimated snag surface area in streams to
exceed benthic habitat area by 20 to 50 percent. This increased area provides
more chemically and biologically active surfaces and attachment sites important
for grazing, scraping, and ﬁlter-feeding aquatic invertebrates (Triska and
Cromack 1980, Harmon et al. 1986). This implies that the old-grth lakes in
the study, with greater amounts of CWD, can potentially produce more
invertebrate biomass than second-growth lakes.

The function of CWD itself as a food source for aquatic invertebrates is
less important. Very few insects ingest or bore into submerged or waterlogged
wood when compared to terrestrial woody debris (Harmon et a1. 1986), and this
is one of the reasons that CWD persists so long in aquatic systems. It is this
stability of CWD in stream systems that makes it attractive as habitat (Benke
and Wallace 1990). In lentic systems, where water movements do not supply a
constant particulate food source, or require ﬁrm attachment to prevent being
carried off by the current, CWD may be less important for invertebrate
production. Moring et al. (1986) found invertebrate densities to be less in areas
of log accumulation than adjacent bare sediment, and suggested this as a reason
for ﬁnding no signiﬁcant difference in ﬁsh abundance between log and non-log
areas of the lake. Invertebrate populations in the littoral zone of ﬁve of the

study lakes may be reduced due to the low pH found in those lakes (Dermott 1985).

54

For ﬁsh, the three-dimensional arrangement is the most important
feature, rather than increased invertebrate food supply (Sedell et al. 1984). Fish
abundance in streams has been shown to increase with increasing structural
complexity (Harmon et al. 1986, Sedell et al. 1984). This phenomenon was also
apparent in the study lake systems. In the lakes with good visibility (Crampton
and Snap Jack lakes), ﬁsh were usually seen associated with the CWD being
measured. The debris was being used primarily as cover, although some
largemouth bass spawning beds were located adjacent to debris. No feeding
behaviors were observed at all, even after allowing ﬁsh sufﬁcient time to become
used to my presence. In Bergner Lake, where CWD was scarce, the ﬁsh
populations had switched to the overhanging bog mat for cover requirements,
and were not associated with the woody debris in the lake.

It has not been determined whether the presence of CWD is required for
ﬁsh production, or merely concentrates the ﬁsh already present in the system.
Triska and Cromack (1980) stated that "wood debris seems to have a direct
effect on the size of ﬁsh populations but only an indirect inﬂuence on their
metabolism". I would paraphrase this as "wood debris seems to have a direct
effect on the numbers of ﬁsh present in an area but only an indirect inﬂuence on
their metabolism". Some species, such as the Micropterus species and crappies,
do utilize woody debris when spawning by locating nests next to, or broadcasting
eggs over, woody debris (Scott and Crossman 1973). Insufﬁcient woody debris
could be limiting in this case, and I have observed smallmouth bass in a clean-

bottomed lake constructing beds next to discarded beer cans when CWD was

55

absent. For other species, it is more likely that CWD simply provides a stable
visual reference in an otherwise featureless environment, or at most provides

ambush Cover or refuge from predators.

Implications for Management

In general, abundance of in-lake CWD does differ between study lakes in
old-growth and second-growth settings, with fewer and smaller pieces occurring
in the second-growth setting seepage lakes. The protected nature of the second-
growth lakes studied may not be representative of other deforested or reforested
settings where human impacts are less controlled. Clearing of CWD for
navigation, aesthetics, and swimming beaches will lessen the amount of CWD
available as habitat. In these lakes, I believe the addition of woody structure
would be indicated and necessary.

The three important questions posed by Sedell et al. (1984) are:

1) what kind of structure should be used?
2) where should the structure be placed?
3) how much structure is enough?
I would add a fourth question:
4) how has the change from old-growth to second-growth forest

changed the dynamics of the ecosystem?

56

The answers will depend on the lake and the goals of the manager. The
characteristics of the second-growth study lakes demonstrate that each of these
lakes was unique in CWD content. Any naturally-occurring CWD will require
no management, and provide a template for supplementation. Restoration of
CWD to pre-existing conditions is impossible, and should not be attempted
(Bryant 1983); rather the focus of CWD management should be to maximize the
beneﬁcial aspects of structural additions while offsetting the adverse effects (if
any) and the costs of installation. Structures most like natural inputs (tree-
drops) at 90° to the shoreline will provide cover at a range of depths, from the
land-water interface to the near-pelagic. More branching (complexity) will
increase the attractiveness of the structure, and provide a range of interstice
sizes for various ﬁsh to select from. The importance of the microhabitat size
within the macrohabitat of woody structure was emphasized by Johnson et a1.
(1988). I have observed that pine trees recently added to the water, complete
with dense needles, were more attractive to a variety of ﬁsh species and sizes
than bare trunks. Coniferous trees persist longer in aquatic systems (Andrus et
al. 1988) and would be the preferred species for tree-drops when available.
Considering how structure-poor most recreational lakes are, as many of these
structures as feasible should be used.

A great deal of emphasis is placed on manipulation of structural habitat
by ﬁsheries managers today. Addition of structural habitat is relatively easy and
inexpensive, visually impressive, and psychologically satisfying to the users of the

resource. I believe that managers should step back and consider the productivity

57

potential of the system before improving habitat.

The old-growth lakes of the Sylvania Wilderness support small
populations dominated by older, larger ﬁsh (Miller 1992). This is not due to
higher growth rates; rather it is due to longevity of the individual ﬁsh (Clady et
al. 1975). I observed no ﬁsh utilizing CWD in either Corey or Katherine lakes;
this may be because of small numbers present or poor visibility in these lakes.
Recruitment in old-growth lakes of the SWA may be reduced due to the low
productivity of the lakes (Clady et a1. 1975), and the low pH. I think it is
doubtful that the presence or absence of CWD in these lakes would affect either
ﬁsh numbers or production. Many ﬁsh were observed around the CWD in Snap
Jack Lake, but this lake had a higher pH and more light penetration.

i The ﬁsh populations observed in the second-grth lakes appeared
healthy and abundant, in spite of the reduced CWD. The productivity of these
lakes must have been altered in some way by the shift from surrounding old-
growth to second-grth forest. Replacement of hemlocks with hardwoods has
been shown to decrease soil acidity and increase the cycling of nutrients (Hix
and Barnes 1984). It is possible that this change has increased the availability of
nutrients in the second-grth lakes, and increased overall production. If this is

true, increasing woody habitat may beneﬁt the ﬁsh population in this case.

CONCLUSIONS

The study lake systems within old-grth forest settings were similar in
CWD attributes, although CWD in Katherine Lake was found to be slightly less
than Corey or Snap Jack lakes. The CWD distribution in lakes within the
second-growth forest settings were not similar to each other, and Brown Lake
was dissimilar from all other study lakes. The presence of inlets and an outlet,
and subsequent movement of water through the lake, higher decomposition rates
allowed by higher pH, and fetch may all attribute to the high degree of
difference in Brown Lake’s CWD. The CWD distribution (volume and number
of pieces) in the second-growth study lakes was less than that of the old-growth
study lakes, indicating that the second-growth systems may be CWD-deﬁcient.
Lack of CWD may be limiting for secondary production and/ or ﬁsh habitat
requirements. However, the differences in this study between the two settings is
not as extreme as those found in similar studies in stream systems.

The dynamic nature of stream systems can drastically affect the amount
and distribution of CWD within the banks; changes in CWD input through
changes in the riparian forest can decrease the amounts of CWD in the stream
system. Lake systems, with their relative lack of dynamic forces, are not as
susceptible to movements or emigration of CWD within the lake. Inundated and
waterlogged wood decay rates are extremely slow, especially for larger diameter

58

59
debris and low pH waters. If left undisturbed, in-lake CWD will persist for a

long time, even in second-growth settings.

The characteristics of the riparian shoreline vegetation and slope have
been useful in modelling CWD inputs to stream systems, and may prove useful
as predictors of in-lake CWD content as well. The basal area of live trees along
the shoreline may have the most potential for predicting natural CWD volume
within the lake. In lakes where human inﬂuence has altered or removed most
naturally-occurring CWD, analyzing the riparian vegetation could suggest a base
guideline for supplementing in-lake woody structure. Placement of this structure
to mimic naturally occurring inputs will increase the chance of the added
structure being used as habitat. I

The small number of lakes and subsequent minimal sample size obtained
in this study may not have been sufﬁcient to adequately demonstrate the
differences or similarities among lake or settings. Future studies utilizing a
larger number of lakes and greater percentage of littoral zone sampled may
result in better resolution of CWD characteristics within the lakes of both

settings.

LIST OF REFERENCES

LIST OF REFERENCES

Andrus, C.W., B.A. Long, and HA. Froehlich. 1988. Woody debris and its
contribution to pool formation in a coastal stream 50 years after logging.
Canadian Journal of Fisheries and Aquatic Sciences 45:2080-2086.

Angermeier, PL and J .R. Karr. 1984. Relationships between woody debris and
ﬁsh habitat in a small warrnwater stream. Transactions of the American
Fisheries Society 115:716-726.

A.P.H.A. 1985. Standard methods for the examination of water and wastewater.
16th ed. American Public Health Association, Washington, DC. 1268 pp.

Anonymous. 1991. Guide to University of Notre Dame Environmental Research
Center. Environmental Research Center, Department of Biological
Sciences, University of Notre Dame, Notre Dame, Indiana.

Benke, AC. and J.B. Wallace. 1990. Wood dynamics in coastal plain blackwater
streams. Canadian Journal of Fisheries and Aquatic Sciences 47:92-99.

Bilby, RE. 1984. Removal of woody debris may affect stream channel stability.
Journal of Forestry 82:609-613.

Bilby, RE. and J .W. Ward. 1989. Changes in characteristics and function of
woody debris with increasing size of streams in western Washington.
Transactions of the American Fisheries Society 118:368-378.

Bilby, RE. and J.W. Ward. 1991. Characteristics and function of large woody
debris in streams draining old-growth, clear-cut, and second-grth forests

in southwestern Washington. Canadian Journal of Fisheries and Aquatic
Sciences 48:2499-2508.

Bryant, MD. 1983. The role and management of woody debris in west coast
salmonid nursery streams. North American Journal of Fisheries
Management 3:322—330.

60

61

Bryant, MD. 1985. Changes 30 years after logging in large woody debris, and its
use by salmonids, pp. 329-334 in Johnson, R.R., C.D. Ziebell, D.R. Paton,
P.F. Ffolliott, and RH. Hamre (coords.), Riparian Ecosystems and their
management: reconciling conﬂicting uses. U.S.D.A. Forest Service
General Technical Report RM-120, Rocky Mountain Forest and Range
Experiment Station, Juneau, Alaska.

Carlson, J .Y. C.W. Andrus, and HA. Froehlich. 1990. Woody debris, channel
features, and macroinvertebrates of streams with logged and undisturbed
riparian timber in northeastern Oregon, U.S.A. Canadian Journal of
Fisheries and Aquatic Sciences 47:1103-1111.

Clady, M.D., D.E. Campbell, and GP. Cooper. 1975. Effects of trophy angling
on unexploited populations of smallmouth bass, pp. 425-430 in Stroud,
RH. and H. Clepper (eds.), Black bass biology and management. Sport
Fishery Institute, Washington, DC.

Crumrine, JP. and AM. Beeton. 1975. Lirnnology of the lakes of the Sylvania
Recreation Area, Ottawa National Forest. Special Report Number 24,
Center for Great Lakes Studies, University of Wisconsin, Milwaukee. 34
pages plus appendices.

Dermott, RM. 1985. Benthic fauna in a series of lakes displaying a gradient of
pH. Hydrobiologia 128:31-38.

De Vries, PG. 1974. Multi-stage line intersect sampling. Forest Science 20: 129-
133.

Fausch, KD. and T.G. Northcote. 1992. Large woody debris and salmonid
habitat in a small coastal British Columbia stream. Canadian Journal of
Fisheries and Aquatic Sciences 49:682-693.

Gregory, K.J. and R]. Davis. 1992. Coarse woody debris in stream channels in
relation to river channel management in woodland areas. Regulated
Rivers: Research & Management 7:117-136.

Harmon, ME. and 12 co-authors. 1986. Ecology of coarse woody debris in
temperate ecosystems. Advances in Ecological Research 15:133-302.

Hartwig, F. and BE. Dearing. 1979. Exploratory Data Analysis. Sage University
Paper series on Quantitative Applications in the Social Sciences, series
number 07-016. Beverly Hills and London. Sage Publishing. 83 pp.

62

Hellenthal, Ronald A. 1994. Personal communication. Professor and Gillen
Director, Environmental Research Center, University of Notre Dame,
Notre Dame, IN.

Hix, D.M. and B.V. Barnes. 1984. Effects of clearcutting on the vegetation and
soil of an eastern hemlock dominated ecosystem, western Upper
Michigan. Canadian Journal of Forest Research 14:914-923.

Johnson, D.L., R.A. Beaumier, and WE. Lynch,Jr. 1988. Selection of structure
interstice size by bluegills and largemouth bass in ponds. Transactions of
the American Fisheries Society 117: 171-179.

Johnson, SW. and J. Heifetz. 1985. Methods for assessing effects of timber
harvest on small streams. NOAA Technical Memorandum NMFS
F /NWC-73. US. Department of Commerce NCAA/NMFS. 33 pp.

Kevern, NR. 1989. A manual of limnological methods. Department of Fisheries
and Wildlife, Michigan State University, East Lansing, MI. 212 pp.

Lienkaemper, G.W. and F .J . Swanson. 1987. Dynamics of large woody debris in
streams in old-grth Douglas-ﬁr forests. Canadian Journal of Forest
Research 17:150-156.

Miller, BR. 1992. Results of voluntary catch surveys on Sylvania lakes in 1989
and 1991. Michigan Department of Natural Resources, Fisheries Division,
Technical Report #92-7. 15 pp. .

Moring, J .R., P.D. Eiler, M.T. Negus, and KB. Gibbs. 1986. Ecological
importance of submerged pulpwood logs in a Maine reservoir.
Transactions of the American Fisheries Society 115:335-342.

Murphy, ML and JD. Hall. 1981. Varied effects of clear-cut logging on
predators and their habitat in small streams of the Cascade Mountains,
Oregon. Canadian Journal of Fisheries and Aquatic Sciences 38:137-145.

Murphy, ML and KV. Koski. 1989. Input and depletion of woody debris in
Alaska streams and implications for streamside management. North
American Journal of Fisheries Management 9:427-436.

Novak, J .T., A8. Goodman and D]... King. 1975. Aquatic-weed decay and color
production. Journal of the American Water Works Association 65 :134-
139.

63

Robison, E.G. and R.L. Beschta. 1990. Characteristics of coarse woody debris for
several coastal streams of southeast Alaska, U.S.A. Canadian Journal of
Fisheries and Aquatic Sciences 47:1684-1693.

SAS Institute, Inc. 1988. SAS/ STAT user’s guide, release 6.03 edition. Cary,
NC: SAS Institute, Inc. 1028 pp.

Scott, W.B. and EJ. Crossman. 1973. Freshwater ﬁshes of Canada. Fisheries
Research Board of Canada Bulletin 184. 966 pp.

Sedell, J .R., F.J. Swanson, and S.V. Gregory. 1984. Evaluating ﬁsh response to
woody debris. pp. 222-245 m T.J. Hassler (ed.), Proceedings of the Paciﬁc
Northwest stream habitat management workshop. Humboldt State
University, Arcata, California.

Spies, TA and Barnes. 1985. A multifactor ecological classiﬁcation of northern
hardwood and conifer ecosystems of Sylvania Recreation Area, Upper
Peninsula, Michigan. Canadian Journal of Forest Research 15:949-960.

Spies, T.A., J .F . Franklin, and TB. Thomas. 1988. Coarse woody debris in
Douglas-ﬁr forests of western Oregon and Washington. Ecology 69:1689-
1702.

Tyrrell, LE. and TR. Crow. 1994. Structural characteristics of old-growth
hemlock-hardwood forests in relation to age. Ecology 75:370-386.

Triska, F.J. and K. Cromack, Jr. 1980. The role of wood debris in forests and
streams. pp. 171-190 in R.H. Waring (ed.), Forests: fresh perspectives
from ecosystem analysis. Proc. 40th Biology Colloquium (1979). Oregon
State University Press, Corvallis, Oregon.

Van Sickle, J. and S.V. Gregory. 1990. Modeling inputs of large woody debris to
streams from falling trees. Canadian Journal of Forest Research 20:1593-
1601.

Wallace, J .B. and AC. Benke. 1984. Quantiﬁcation of wood habitat in
subtropical coastal plain streams. Canadian Journal of Fisheries and
Aquatic Sciences 41:1643-1652.

Wetzel, R.G. 1983. Limnology. Second edition. Saunders College Publishing,
Philadelphia, PA. 767 pp.

I t -

 

NICHran STATE UNIV, LIBRARIES
llllllWlWllHlllllmIll”llNIWI”llllllllllmlllll
31293014110450