- p 5

it
whkfl. “WNW. , V .21.:
(Em. 3... , . V . . “3H,.“ ,
. 43.33”“!
n

.1 . .
«.1 w. 13.“. . c
K.” . 1. H . 1..
. 1

_
l, t. v33
. 17x4 7 I
:nfiaw.....nva ‘35:.
.. .. . .
vauafl‘ . 5..

k...“ s.

. u .. i
3.. éfiu m...

it

. .

r L...

.WJerfims
.

meau’ 5.
..

.- ~12.

in i?

.
I

y r v
v... (04.7.
.2. {3.1

s ..

3.
1.x;
i

r!

W t

gawk fl! .3. Ft. ..
. .. 4. w 3.. V 3.1 ...o.xc’..«. Wt . ~¢.:.u...u.r10.uv. . . he”. s!
5., » 6“ ewx may a: V $3meme . .12.?
, .3.

 

4 .mxwvaa .3

.4. . .vnt‘fiwfl.
Wing #23

3., . : v V
fiumfifiv

7

Apt:
c I !‘

g3... * .V
,

..
:Vtiil. .

. . . \
:Hunr.‘ . . . Em; .
. I

.p 3.. ‘

1 .
. 4,. an . :
Jahangfla?

11.1

:2: ‘
$3.on

\
\. 9.
van»...
(1).!

.. $.33}: 2.... _x.1. .
bu“ .turzlifxn.‘b .v
, .3. I: 343...”?

:3 fl, Lwnw
Ema: .

.
if): 9L.

 

13);; A u\p1..\~
. A .37}

 

 

 

This is to certify that the

dissertation entitled

An Ecosystem Approach to Assessing the Effects
of Forest Heterogeneity and Disturbance on Birds

of the Northern Hardwood Forest in Michigan's
Upper Peninsula.

presented by

Maya A. Hamady

has been accepted towards fulfillment ‘
of the requirements for

Ph.D degreein Wildlife Biology

5Wa QMA

Major professor

DagIJOJDO
~/] 7

MS U it an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

LIBRARY

Michigan State

Unlversity

 

 

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.

 

.9 DATE DUE

DATE DUE

DATE DUE

 

J
JUN 194102

 

 

 

 

 

 

 

 

 

 

 

 

 

11m mus-p.14

 

 

i\' ECOSYS
HETEROGE
HARE

AN ECOSYSTEM APPROACH TO ASSESSING THE EFFECTS OF FOREST
HETEROGENEITY AND DISTURBANCE ON BIRDS OF THE NORTHERN
HARDWOOD FOREST IN MICHIGAN’S UPPER PENINSULA
By

Maya A. Hamady

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Fisheries and Wildlife

2000

ABSTRACT
AN ECOSYSTEM APPROACH TO ASSESSING THE EFFECTS OF FOREST
I-IETEROGENEITY AND DISTURBANCE ON BIRDS OF THE NORTHERN
HARDme FOREST IN MICHIGAN’S UPPER PENINSULA
By

Maya A. Hamady

Forested regions such as the Upper Peninsula of Michigan, are considered
population sources for area sensitive forest birds which have disappeared from many
human-dominated landscapes because of forest loss and fi'agmentation. Northern
hardwood forest landscapes in the Upper Peninsula have been affected by past human
activities and continue to be periodically logged. The possibility that anthropogenic
activities would over time favor bird species that are tolerant of some levels of forest
fragmentation, at the expense of area sensitive species, needs to be evaluated. 1 developed
an ecosystem model for northern hardwood forest landscapes and explored relationships
among bird Species, and among bird densities, microhabitat variables and landscape level
variables to identify criteria necessary to make such an evaluation.

I modeled heterogeneity of the northern hardwood forest at the landscape level in
terms of historical change in land use and land type association. I considered period
elapsed after last logging to afi‘ect vegetation structure at the locality of the bird census
station. Habitat relationships among 10 bird species were derived fi'om individual bird
species associations with microhabitat variables. A forest fragmentation dimension was
evident in habitat relationships among 3 shrub nesting bird specieS, black-throated blue

warbler (Dentb'oica cemlescens), American redstart (Setophaga ruticilla) 311d chestnut-

sided mac lien"

 

at: first. igsrar
“Eb-£6: (Emir: 'iu’] _":.

, .
we - ‘ . .-.
them Cfa‘iit‘ .r. ..

timing is cis L;

 

“it: use loci and 1'25
buds (m: at: m the s?
amazed 1n some 12;.

Rev"
. U

tzncatson of 5.37

I -. vi- 4 ‘. '
an. of torts: sage “

dueiopmem in ~

:1 an
t he northern I’a'iu
torsion: that we c

Ti! northern ha;

3% O
and smut at difierr
hm mines on 5:
Wm Simple asses
mr‘liequate for etalua:

Si .

“it i ..
' emaieil'tilllbh'unr
tot} ‘ ~
mmfifimiifl SCI]

elm: ,
LII .. efi‘
are dis

sided warbler (Dendroica pensylvanica). Forest fragmentation aspects however could not
explain density disparities between the 2 conifer associated bird species, blackburnian
warbler (Dencb’oicafitsca) and black-throated green warbler (Dendroica virens).
Historical change in land use affected the density of late maturity conifers and forest loss
at landscape levels. Logging affected canopy cover and shrub layer development, effects
that were local and highly dynamic. Logging was responsible for increased densities of all
birds that nest in the shrub-sapling layer in managed forest landscapes but was also
associated in some landscapes with a relative greater increase of edge tolerant bird species.
Restratification of sampling units based on levels of forest cover within a radius of 1.6 km,
length of forest edge within a radius of 500m from bird census station, shrub layer
development and canopy opening elucidated what constitutes forest fragmentation effects
in the northern hardwood ecosystem. This restratifrcation permitted the quantification of
conditions that were difl‘erently favorable to 2 edge species and one area sensitive species.
The northern hardwood forest is a complex system that is both spatially heterogeneous
and dynamic at different scales. Multiple factors modify or aggravate the effects of

human activities on bird species that in turn respond differently to different dynamics in the
sYstem. Simple assessments of the effects of anthropogenic activities on bird species are
inadequate for evaluating the long term maintenance of area sensitive species in the
System. The inadequacies of treating the northern hardwood forest as a homogeneous
System at equilibrium, and of disregarding differences among forest bird species, relative
to their differential sensitivity to forest fragmentation, in current assessments of human

activity efi‘ects are discussed.

(this? .
H.332 knee Hera;

‘!

.
.l‘r
Cir-v

 

 

Copyright by
Maya Ameen Hamady
2000

ACKNOWLEDGMENTS

From its inception this research was bent on a tortuous path. I am indebted to my
advisor Dr. Scott Wmterstein for his efl‘orts in resolving many crises along the way.
Special thanks go to all my other graduate committee members, Drs. Henry Campa
Stephen Stephenson, John Hart and Jonathan Haufler for their individual contributions to
this research effort. I was inspired to present this work within an ecosystem context by my
earlier exposure to landscape ecology perspectives that were being progressively
promoted on the Hiawatha National Forest, and by opportunities to learn about the latest
modeling and ecosystem theories presented in course work and by guest lecturers in Dr.
Jianguo Liu’s class ‘Computer Simulation and Modeling’. This study was greatly
facilitated by ecological classifications of land units undertaken by Dr. E. Padley on the
West Unit of the Hiawatha and by J. Jordan on the Ottawa National Forest.

Funding and logistic support was provided by the Hiawatha National Forest
through the Neotropical Migrant Bird funding initiative. Funding was also received fi'om
North Central Forest Experiment Station. In kind matching grants were received fiom
Michigan State University. A College of Natural Resources grant of 5,000 8 was used to
defray some additional costs. Logistic support and information were provided by many
individuals on the Hiawatha National Forest, Ottawa Forest and Huron Mountain Club.

It was possible to collect extensive field data that were necessary for this research

because of the dedication and enthusiasm of many field assistants: Mike Berg, Amy

   
  

Prxhi Paul Drs..
Ton} 0.55.3. V;
Tzr. Nu .ic. Brai
broiog'szs “as can.
Tmrgsort Mic]
and for mm; a.

Mm min

Narional Forest pf

Gi

(I)

50"“386 COL.

P70} 6:1 I am era:

Proehl, Paul Thompson, Mark Parkinson, Jason Hoeksema, Mike Livingston, Kim Smith,
Tony Dufiny, Mark Ledhebur, Christine Hill, Sue Tangora, Laurie Necaise, Julie Car,
Tim Nuttle, Brad Kassuba and Robyn Oliver. Field work alongside this crew of young
biologists was enjoyable and motivating. I am especially indebted to Amy Proehl, Paul
Thompson, Mike Berg and Mike Livingston for preparatory work prior to field excursions
and for training and supervision of field crews.

Many individuals on the Hiawatha National Forest and individuals on the Ottawa
National Forest provided the necessary digital data, maps and aerial photos from which
GIS coverage could be derived. Rich Minnis prepared some GIS coverage for this
project. I am grateful to Andrea Schmidt, Abdoul Hayee, Jim Schopper and Beth Ursel
for undertaking the tedious task of reviewing and correcting 618 maps. Lauri Keen’s
tedious tasks of initial manuscripting of aerial photos and later of cleaning 618 maps
proved critical to the completion of this project.

This study would not have been undertaken if it wasn’t for the support I received
at a critical initial stage fi'om Stan Benes my then supervisor on the Hiawatha National
Forest. Stan’s vision and maverick attitude towards embarking on difficult tasks to foster
better understanding of the effects of forest management were crucial to study support and
initiation. The support I received from colleagues on the biology stafl‘, Kevin Doran,
Steve Sjogren and Lawson Gerdes was uplifting and encouraging. Returning to graduate
school after being employed for over a decade was excruciating. It was good to have a
trusted fiiend like Tim Van Deelen whose own familiarity and experience with the
administration of the Hiawatha National Forest and the Department of Fisheries and

Wildlife eased my transition fi'om work to graduate school.

Far from work and university, family and fiiends cushioned the many ups and
downs I encountered in carrying out this study. My mother Celia Hamady was unfaltering
in her advice to remain tough and face all what I had to encounter until the end. I greatly
appreciate the understanding and support I received from my brothers Omar and Nabeel
and my sisters- in- law Suzanne and Rolla. My long time fiiends Christopher Hull, Nancy
Armitage, Linda Semenko, Mary Hamilton and Terry Farrell lent a listening ear during
times when the going got rough. This research is dedicated to the memory of my 3 aunts
who passed away during the course of this study. I remember my aunt Helen Musfy for
her kindness and understanding, my aunt Jennie F rizelle-Hamady for sharing with me her
love of nature and the outdoors and my aunt Dr. Ruth Brinker-Hamady for

encouragement and interest in my career.

vii

usr or Its-LE
usr or nor-R;
(PAPER 1. m

ASK [5 DH U I‘.
] .-\ Nee-5'
2 XCOUO;

3 M ECOS
“513021.!

 

 

 

 

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES xiii
LIST OF FIGURES xxi
CHAPTER I; INTRODUCTION 1
ISSUES IN THE CONSERVATION OF FOREST BIRD SPECIES l
1. A Need for a Holistic Approach 1
2. Neotropical Birds in the Northern Hardwood Forest 3
3. An Ecosystem Approach to Relate Bird Species Densities to Forest
Heterogeneity g
OBJECTIVES 1 0
STUDY AREA 12
1. General Description of the Upper Peninsula 12
2. Description of the 10 Study Areas 16
CHAPTER 2; METHODOLOGY 22
BUILDING AN ECOSTYSTEM MODEL 22
1. Spatial Scales Pertinent to Assessing the Effects of Forest Heterogeneity
on Songbird Species 22

 

viii

" Segre:
that It

 

 

if.

his

In.

”21"

3 SCCC’ \
01- 5""

4 Sat" a
5 LITAT‘TE
PrOCE‘S:

6 [DCth-H‘
”UK.

] Sdfi‘tlc

DJ

2. Sources of Forest Heterogeneity in the Northern Hardwood Forest and

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

their Impacts at Scales Relevant to Forest Bird Species 29
Structure of forest heterogeneity 29
Historical events that added a human induced factor to forest
heterogeneity 32
Hierarchy of physical and climatic factors aflecting
heterogeneity 3 8
Incorporating human-induced and environmental heterogeneities
into model 39

3. Selection of Bird Species to Include in the Model and Representation
of their Habitat 42
4. Sampling Units and their Groupings 44
5. Linking Habitat Selection at the Microhabitat Level to Ecosystem
Processes —---46
6. Incorporating Landscape Scale Effects on Bird Species 47
7. Integrating Microhabitat and Landscape Level Effects 48
FIELD METHODS 53
1. Selection of Study Units 53
Selection of primary forest, managed forest and settled forest
lcmdscapes 53
Locating northern hardwood forest sites in primary
forest, managed forest and settled forest landscapes 54
2. Locating Bird Census Stations 59
3. Bird Census Methods 61
4. Sampling of Microhabitat Variables 64
5. Developing Geographical Information System (GI S) Coverage for Study
Areas 55

 

 

4 Er».-

‘\\k>

Forest 1

(Him-ER 3 RE

I Dmite
H11“ 0.
2 5550622.:
Chafame

Lo)

Vegetaiil

Lari

3”,,
"Or;

 

 

STATISTICAL TECHNIQUES 69

 

1. Estimation of Bird Densities 69

2. Assessing the Efi‘ects of Historical Divergence in Land use, Recent
Logging Disturbance and Land Type Association on Bird Species 69

 

 

 

 

 

 

 

 

Densities 7O
3. Habitat Relationships Among Bird Species 72
4. Efi‘ects of Interactions of Habitat F ragrnentation Related Aspects of
Forest with Habitat Availability on Bird Species 74
CHAPTER 3: RESULTS 77
1. Densities and Encounter Rates of Bird Species Across the Northern
Hardwood Forest 77
2. Association of Bird Species with Small Scale Vegetation
Characteristics 77
3. Vegetation Differences Among Northen Hardwood Forest Landscapes
at Difl‘erernt Spatial Scales. 81
Landscape level differences 81

Small scale differences in the conifer tree component among
northern hardwood forest strata 81

 

Canopy characteristics among strata of northern hardwood forest ------ 82
Shrub layer differences among northern hardwood forest strata --------86
Dynamics of canopy opening and shrub layer development within

15 years following [aging across dtflerent managed forest land
We associations 88

 

4. Density Patterns of 10 Bird Species Across Northern Hardwood Forest

 

 

 

Strata 88
Density patterns across land use categories 88
Density patterns of conifer associated bird species 100

 

the B:
Va: 3;:

 

D.”
‘\

CH‘J’TER 4 D:

4‘??th
“Rm-
BED U l\

CO SK 3

Density patterns of bird species affected by a shrub-layer --------------- 105

Habitat relationships among the black-throated blue warbler,
American redstart and chestnut-sided warbler. 107

 

5. Effects of Landscape Level and Microhabitat Variables on Density of
the Black-Throated Blue Warbler, American Redstart and the Derived
Variable BTARC S. 1 15

 

Effects of forest edge, forest cover and shrub layer development-—---1 15

Effects of extensiveness of the northern hardwood forest, canopy

 

 

opening, shrub layer development, and forest edge 124
CHAPTER 4: DISCUSSION AND CONCLUSION 132
ECOSYSTEM LEVEL RELATIONSHIPS I32

 

ASPECTS OF FOREST FRAGMENTATION IN THE NORTHERN HARDWOOD
FOREST 1 3 7

 

 

BIRD CONSERVATION AND MANAGEMENT CONSIDERATIONS I48

1. Function of Primary Forest in Present Northern Hardwood Forest
Ecosystems 148

 

 

2. Function of the Managed Forest 153

3. Viewing the Northern Hardwood Forest as an Open System vs a Closed
System, at Equilibrium or Undergoing Directional Change 155

 

4. Aspects to Consider in the Development of Bird Conservation Strategies-- 158

 

 

 

 

CONCLUSIONS 1 60
FOOTNOTES 163

Footnotes for Chapter 1; Introduction 163

Footnotes for Chapter 2; Methodology 163

xi

 

Footnotes for Chapter 3; Results

 

170

 

 

 

 

 

 

Footnotes for Chapter 4; Discussion and Conclusions 172
APPENDICES 175
APPENDIX A 176
APPENDIX B 180
APPENDIX C 186
APPENDIX D 197
LITERATURE CITED 244

xii

Ital

Ink:

TuEB

7&k4

Tats

Table 6.

Table 7.

 

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

LIST OF TABLES

Percent composition of land cover types in 10 study areas in the
Upper Peninsula of Michigan 1992- 1994. 17

 

How sources of variability in the northern hardwood forest were
incorporated into the study design. 35

 

Distribution of sampling points (bird census stations) in strata of
northern hardwood forest based on historical divergence in land use,
forest disturbance period and land type associations. 45

 

Probability values of regressions of densities of bird species on

vegetation variables. Vegetation was sampled within 1200 m2 at 321

bird census stations in northern hardwood forest. Data were collected

in the Upper Peninsula of Michigan 1992-1994. 78

 

P values of 2- way Anovas undertaken to compare the effects of land

type association, period elapsed after logging and interaction between
land type association and period elapsed after logging on CANOPEN,
WSDC<0.5, GAP>100. Vegetation variables were sampled 1992-1994

in the Upper Peninsula of Michigan. 92

 

Bonferroni adjusted P values of post test pairwise comparisons of
CANOPEN, WSDC<0.5, and GAP>100 (during the 15 years following
logging) among 3 land type associations. Vegetation variables were
sampled in 1992-1994 in the Upper Peninsula of Michigan. 93

 

P values of 2 way Anovas undertaken to compare the effects of land

type association, period elapsed after logging, and interaction between

land type association and period elapsed after logging on the
black-throated blue warbler BTBW, American redstart AMRE,

chestnut- sided warbler CSWA and the derived variable BTARCS

sampled 1992-1995 in the Upper Peninsula of Micigan 113

 

P values in Anovas undertaken to test the effects of EDGE and

WSDC<0.5 on BTBW, AMRE and BTARCS within 3 levels of

F CIMIL. Bird density and vegetation variables were sampled

1992-1995 in the Upper Peninsula of Michigan 119

 

xiii

Table :11

Table XI

labia Bl

Tabie B:

Table 03‘

 

 

Ger

of 1
Per

Table A1

Table A2

Table B1

Table B2

Table B3

Table C1.

Table C2

Table C3.

Table D1.

Table D2

Table D3.

Relationship of the land type association to ecological units above

and below its level in the national hierarchy of land units as given

in Ecomap, USDA Forest Service, Washington DC. (1993 ), and to
subdivisions identified in the regionalizationof Michigan landscapes.
(Albert et al. 1986). 177

 

Descriptions of the land type associations on the West Unit of the
Hiawatha National Forest as they had been delineated (1992) in a draft
ecological classification of the forest. 178

 

Scientific names of plant and animal species mentioned in text. ----------- 181

Availability of stands in difi‘erent strata of forest disturbance period
within managed forest and number of stands selected for locating bird
census stations. 182

 

Microhabitat variables measured in vegetation plots located at the bird
census stations. 183

 

Initial classification of cover types in the integration of land cover types
from different land classification systems. 187

 

 

Sources of geographical data covering study areas 192

Types of features in addition to primary and secondary roads, railroad
tracks, power lines and gas lines, included in the calculation of forest
edge within 500m circle areas around bird census stations. 194

 

Model selected in the program Distance Sampling (Lake et al. 1994)

and other specifications, in the estimation of a detection function

for each of 10 bird species. Distance data were obtained from bird
censuses conducted 1992-1994 in northern hardwood forest in the

Upper Peninsula of Michigan. 198

 

Mean density/km2 and encounter rate of 10 songbird species censused

at 321 bird census stations in northern hardwood forest. Northern
hardwood forest was stratified into 7 strata. A roughly equal number

of bird census stations were located in 1992-1994 in the Upper

Peninsula of Michigan. 199

 

P values of apriori contrasts in ANOVAs undertaken to test the
statistical significance of differences in means of TOTCON and
LMCON among land use and forest disturbance period strata.
These contrasts are for a hypothetical northern hardwood forest in
which the strata of primary forest, settled forest and 5 forest

xiv

labia 0-1

 

labia DE

5:0-r-‘(f5
'-

 

Table D‘

labie D3

Table [)9

able D11

Table D4.

Table D5

Table D6.

Table D7.

Table D8

Table D9

Table D10

Table D11

disturbance periods in managed forest were (finally represented.
Vegetation was sampled in 1992-1994 in the Upper Peninsula of
Michigan. 200

 

Mean of total number of conifer trees TOTCON, and total number

of late maturity conifer trees LMCON, in different strata of

forest disturbance period and land use. Vegetation was sampled in
1992-1994 in the Upper Peninsula of Michigan. 201

 

Overall P values of Anovas undertaken to compare TOTCON and
LMCON among different land type associations replicating forest
disturbance period and historical divergence in land use. Vegetation

was sampled in the Upper Peninsula of Michigan 1992-1994.-- ------- 202

Means of TOTCON and LMCON in different land type associations
replicating forest disturbance periods and historical divergence in land

use strata. Vegetation was sampled in the Upper Peninsula of Michigan
1991-1994. 203

 

Overall P values of Anovas undertaken to compare CAN OPEN,

GAPlOO, GAP>100 and WSDC<0.5 among different land

type associations replicating forest disturbance period and historical
divergence in land use strata of northern hardwood forest . Vegetation
was sampled in the Upper Peninsula of Michigan 1991-1994. - ------ 204

Means of CANOPEN, GAP>100 and WSDC<0.5 in different

land type associations replicating forest disturbance periods and land

use strata. Vegetation was sampled in theUpper Peninsula of

Michigan 1992-1994. 205

 

Mean of canopy opening CANOPEN, number of canopy gaps of

different size classes GAPSO, GAPIOO and GAP>100, and number

of woody stems <0.5" (1.26cm) WSDC<0.5 in different strata of

forest disturbance period and land use.Vegetation was sampled in
1992-1994 in the Upper Peninsula of Michigan 206

 

P values of apriori contrasts in Anovas undertaken to test the

statistical significance of differences in means of CANOPEN,

GAP100, GAP>100 and WSDC<0.5 among strata of northern

hardwood forest in a hypothetical northern hardwood forest in

which the land use strata of primary forest and settled forest, and the

5 strata of forest disturbance period in managed forest are equally
represented. 207

 

Mean density of 10 bird species, and mean value of the derived
variable BTARCS in different northern hardwood forest strata

XV

labia D12

Title D15

Table Dr 7

Table D12

Table D13

Table D14

Table D15.

Table D16

Table D17.

Table D18.

representing different forest disturbance period and land use. Bird
censuses were undertaken in 1992-1994 in the Upper Peninsula of
Michigan. 208

 

F and P values in Anovas undertaken to test the statistical significance

of differences in means of bird density (Table D11) among 7 strata

of northern hardwood forest representing land use and forest

disturbance period. Bird censuses were undertaken in 1992-1994 in the
Upper Peninsula of Michigan. 210

 

Mean densities of 10 bird species and mean values of Anovas of
differences in bird density means among land use strata in a

hypothetical northern hardwood forest in which strata of primary

forest, settled forest and 5 strata of forest disturbance period are

equally represented. Birds were censused in 1992-1994 in the Upper
Peninsula of Michigan. 211

 

P values of Anovas for planned apriori contrasts undertaken to test

the statistical significance of differences in bird density means among

land use strata of a hypothetical northern hardwood forest in which

strata of primary forest, settled forest and 5 strata of forest disturbance
period within managed forest are equally represented. Bird density data
are from bird censuses undertaken in 1992-1994 in the Upper Peninsula
of Michigan. 212

 

Bonferroni adjusted p values of Anova post tests undertaken to test

the statistical significance of bird density and (BTARCS) differences in
pairwise comparisons of northern hardwood forest strata. Bird data

are from bird censuses undertaken 1992-1994 in the Upper Peninsula of
Michigan. 214

 

Mean density of 10 bird species in 3 landscapes (land type associations)
replicating primary forest landscapes. Birds were censused in
1993-1994 in the Upper Peninsula of Michigan. 219

 

Mean density (birds / krnz) of 10 bird species in 3 land type associations
that are replicates of settled forest landscapes. Birds were censused in
1992-1994 in the Upper Peninsulaof Michigan. 220

 

Mean density (birds /km2) of 10 bird species in 3 land type associations
that are replicates of forest disturbance period strata in managed northern
hardwood forest landscapes. Birds were censused in 1992-1994 in the
Upper Peninsula of Michigan. 221

 

labia D2 9

labia DZ";-

labie 0:1

lable D2:

rm .
Lie D2)

able -
Dz)

 

 

P
‘3

f i

_‘ ‘n—O‘

Table D19.

Table D20.

Table D21.

Table D22.

Table D23.

Table D24.

Table D25.

P values of Anovas undertaken for comparing bird densities among land
type associations replicating forest disturbance periods within managed
forest landscapes and historical divergence in land use. 226

 

Mean LMCON, TOTCON, CANOPEN, and GAP>100 of rnicrohabitat
sites selected by the blackburnian warbler and the black-throated green
warbler and of each of 7 strata of northern hardwood forest. Data are
from vegetation sampling and bird censusing 1992-1994 in the Upper
Peninsula of Michigan. 227

 

P values in Anovas representing the statistical significance of

differences in LMCON, TOTCON, CANOPEN and GAP>100

between each of the blackburnian warbler and the black-throated green
warbler and individual strata of northern hardwood forest. Data are from
vegetation sampling and bird censusing at 321 bird census stations
undertaken in 1992-1994 in the Upper Peninsula of Michigan. ----------228

Mean TOTCON and LMCON of microhabitats selected by the
blackburnian warbler and the black-throated green warbler in different
forest disturbance period and land use strata of northern hardwood

forest and; p values of Anovas indicating the statistical significance of
differences in TOTCON and LMCON between the 2 bird species in the
different northern hardwood forest strata. Data are from vegetation

and bird censusing undertaken in 1992-1994 in the Upper Peninsula

of Michigan. 229

 

P values of regression coefficients in separate regressions of black
-throated blue warbler (BTBW), American redstart (AMRE) and
chestnut-sided warbler (CSWA) on woody stems<0.5" (1.26 cm) for

the 7 forest disturbance period and land use strata. Bird densities and
vegetation data were from bird censusing and vegetation sampling
undertaken in 1992-1994 in the Upper Peninsula of Michigan. --------- 230

Means of canopy and shrub layer variables in habitat selected by

the black-throated blue warbler, American redstart and chestnut-sided
warbler. Characteristics of habitat selected by a species were identified

by the characteristics of habitat around bird census stations at

which the density of the respective species was >80 birds / km2 . Data
were obtained from bird censuses and vegetation sampling undertaken in
321 bird census stations during 1992-1994 in the Upper Peninsula of
Michigan. 231

 

P values of Anovas of planned apriori tests undertaken to compare
canopy opening, shrub layer development and canopy gaps among
northern hardwood forest habitats selected by the black-throated

blue warbler BTBW, American redstart AMRE and chestnut-sided

..

late Dlé

lab-ls DZ?

labia D33

We 0:9

 

 

 

f“.

72;»???

 

L'T‘.

Table D26.

Table D27.

Table D28.

Table D29.

Table D30.

Table D31.

warbler CSWA. Data were obtained from bird censuses and
vegetation sampling carried out in 1992-1994 in the Upper
Peninsula of Michigan. 232

 

P values of Anovas undertaken to compare canopy opening and

canopy gaps among the black-throated blue warbler, American redstart,
and chestnut-sided warbler in 3 forest disturbance strata representing
current logging disturbance. Data were obtained from bird censuses and
vegetation sampling undertaken in 1992-1994 in the Upper Peninsula

of Michigan. 233

 

Means of CANOPEN, GAP100, GAP>100, and WSDC<0.5 in habitats
selected by the black-throated blue warbler BTBW, American redstart
AMRE and chestnut-sided warbler CSWA in northern hardwood forest
strata representing 3 forest disturbance periods after recent logging.

Data were obtained from bird censuses and vegetation sampling
undertaken in 1992-1994 in the Upper Peninsula of Michigan. --------- 234

Contribution of different land type associations and forest

disturbance periods to different groups of bird census stations

aggregated based on forest cover within 1 mile F ClMIL, length of

forest edge EDGE, and shrub layer development, WSDC<0.5. Data

were derived from vegetation sampling in 1992-1994 in the Upper
Peninsula of Michigan and GIS coverage of study areas current for

1992. 235

 

P values in a 3-way Anova, indicating the statistical significance of

effects of F ClMIL, EDGE, and WSDC<0.5 on density of black-

throated blue warbler BTBW, American redstart AMRE and the

variable BTARCS. Bird densities and vegetation were sampled in 1
1992-1994 in the Upper Peninsula of Michigan. 237

 

Mean densities of black-throated blue warbler BTBW, and American
redstart AMRE, and mean values of BTARCS of different groups of

bird census stations formed by a stratification based on 3 levels of
FClMIL, 2 levels of EDGE, and 2 levels of WSDC<0.5. Bird

densities and vegetation were sampled in 1992-1994 in the Upper
Peninsula of Michigan. 238

 

P values in 4 way Anovas in which the densities of the black-throated
blue warbler BTBW and American redstart AMRE and the Variable
BTARCS were compared among groups based on % of northern
hardwood forest within 1 mile radius NHWDIMIL, forest edge
EDGE, % canopy opening CAN OPEN, and number of woody stems

xviii

labia D}:

 

 

Table D32

WSDC<0.5. Data were obtained from bird censusing and vegetation
sampling undertaken in 1992-1994 and from GIS of study areas in
the Upper Peninsula of Michigan current for 1992. 240

 

Least squares means of black-throated blue warbler BTBW and

American redstart AMRE densities, and of BTARCS in 4 way Anovas
used to test the effects of NHWDIMIL, EDGE, CANOPEN, and
WSDC<0.5, least squares means of all groups of bird census stations

that represent main effects of factors are presented but only means of
groups that represent interactions among factors that were significant

at the 0.1 level of significance and that comprised more than 10 bird
census stations. Data presented are from bird censusing and vegetation
sampling1992-1994 in the Upper Peninsula of Michigan. 241

 

xix

 

First I

Flifhv 5
‘tthe .

. .I I
( Ix;-

1
0
la)

534': 4

Fight: 5.

Figure 6

figure 7‘

 

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

LIST OF FIGURES

Location of study sites in the Upper Peninsula of Michigan in which
bird censusing and vegetation sampling were conducted 13

 

Main features of conceptual model and study design showing
stratifications based on historical divergence in land use, land type
associations and time elapsed after logging; and comparisons of bird
densities and vegetation variables among strata. 33

 

Spatial scales considered in measuring environmental variables
around bird census stations. 49

 

Schematic representation of poststratification of sampling units

fiom a stratification based on land use, land type association, and

forest disturbance period to a classification based on levels

of % forest cover at the landscape scale, amount of edge within 500 m
radius and number of deciduous woody stems in the vicinity of the bird
census station. This postratification was used on vegetation data and
forest cover data collected in 1992-1994 in the Upper Peninsula of
Michigan 50

 

Means of total number of conifer trees TOTCON and of late maturity
conifers LMCON counted in 1200m2 vegetation plots at bird census
stations in northern hardwood forest strata of forest disturbance period

and land use. 83

 

Means % canopy opening measured within 1200 m2 vegetation plots
in different northern hardwood forest strata representing forest
disturbance period and land use.

{in
4}

Mean number of forest canopy gaps of different sizes sampled within
1200 m2 vegetatiOn plots in difi‘erent northern hardwood forest strata
representing forest disturbance period and land use 85

 

Means of woody stems <0.5" (1.26 cm) WSDC<0.5 and woody

stems >05" WSDC>0.5 sampled within 12 2x5m vegetation plots

located at bird census stations of different northern hardwood

forest strata of forest disturbance period and land use 87

 

ligdfe 9

lit-st It;

Fare it

Figure N

Figs“: ‘5

ESL-re ‘7

 

 

Figure 9.

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

Mean % canopy opening sampled within 1200 m2 vegetation plots

in strata of northern hardwood forest representing 3 forest disturbance
periods 1-5 years afier logging, 6-12 years after logging and

>12-15 years after logging, in 3 land type associations replicating

the land use strata of managed forest. 89

 

Mean number of woody stems <0.5"(1.26 cm) sampled within 1200 m2
vegetation plots in strata of northern hardwood forest representing 3 forest
disturbance periods 1-5 years after logging, 6-12 years after logging and
>12-15 years after logging, in 3 land type associations replicating the land
use strata of managed forest 90

 

Mean number of canopy gaps >100 m2 sampled within 1200 m2
vegetation plots in strata of northern hardwood forest representing

3 forest disturbance periods 1-5 years after logging, 6-12 years after
logging and >12-15 years after logging, in 3 land type associations
replicating the land use strata of managed forest 91

 

Mean densities of ovenbird, least flycatcher and red-eyed vireo in
northern hardwood forest strata of forest disturbance period and
land use. 95

 

Mean densities of the black-throated blue warbler BTBW, American
redstart AMRE and chestnut-sided warbler CSWA in northern hardwood
forest strata of forest disturbance period and land use. 96

 

Mean densities of veery VEER and scarlet tanager SCTA in northern
hardwood forest strata of forest disturbance period and land use. ------- 97

Mean densities if blackburnian warbler BBWA and black-throated
green warbler in northern hardwood forest strata of forest disturbance
period and land use. 98

 

Mean densities of the black-throated blue warbler in strata of northern
hardwood forest representing forest disturbance periods 1-5 years,

6-12 years and >12-15 years after logging in different land type
associations replicating managed forest. 109

 

Mean densities of the chestnut-sided warbler in strata of northern
hardwood forest representing forest disturbance periods 1-5 years,

6-12 years and >12-15 years after logging in difl‘erent land type
associations replicating managed forest. 110

 

 

Fare 18

Fare 20

Ilggre Bl

Figure B:

figure C1

figure C2

H1

 

In

Figure 18.

Figure 19.

Figure 20.

Figure 21.

Figure B1.

Figure B2.

Figure C 1.

Figure C2.

Mean densities of the American redstart in strata of northern

hardwood forest representing forest disturbance periods 1-5 years,

6-12 years and >12-15 years after logging in different land type
associations replicating managed forest. 1 11

 

Mean densities of BTARCS (Density difi‘erence of the black-throated

blue - the combined densities of the American redstart and the
chestnut-sided warbler) in strata of northern hardwood forest

representing forest disturbance periods 1-5 years, 6-12 years and

>12-15 years after logging in different land type associations

replicating managed forest. 1 14

 

Effects of forest edge, and shrub layer development and their

interactions on densities of the black-throated blue warbler and

American redstart and on BTARC S in different conditions of forest

cover within 1 mile. 120

 

Effects of interactions among NHWDIMIL, CANOPEN, and
WSDC<0.5, and among NHWDlML, CANOPEN and EDGE on
black-throated blue warbler and American redstart densities

and on BTARC S. Data were obtained from bird censusing and
vegetation sampling in 1992-1994 in the Upper Peninsula of

 

 

 

Michigan. 125
Vegetation plots centered at bird census stations in which

rnicrohabitat variables were measured. 184
Measurement of % canopy opening and canopy gaps. 185

Schematic representation of the contrast in the successional

relationship between sugar maple, hardwood dominated northern
hardwoods, hemlock dominated northern hardwoods and hemlock

in secondary and primary northern hardwood forest. 195

 

Relationship between vegetation cover and base features
(road, water courses), 1.6 km circle clips and 500 m clips
around bird census stations. 196

 

xxii

CHAPTER I

INTRODUCTION

ISSUES IN THE CONSERVATION OF FOREST BIRD SPECIES
l. A Need for a Holistic Approach.

As the number of imperiled species continues to spiral upward (LoveJoy 1979;
Wilson 1988,1992; Collar and Andrew 1988 cited in Marcot and Murphy 1996, Ehrlich
and Erlich 1981, Ehrlich and Wilson 1991, Soulé 1991, Gentry 1996), skepticism is
growing about the effectiveness of managing biological diversity on a species by species
basis (Hutto et al. 1987, Franklin 1993, Laroe 1993). The futility of saving species after
the processes that sustain them have disintegrated also points to the need for more
efficient conservation strategies. New efforts in species conservation consider
structures, processes and fiinctions existing in the natural world to be fundamental to the
persistence of species diversity and to its evolution (Allen and Starr 1982; Mooney and
Godron 1983; Delcourt et al. 1983; Wiens et al. 1986a, 1986b; Grumbine 1990,1994;
Wilson 1992; Scott et a1. 1993; Franklin 1993). In the holistic and ecosystem
approaches that are being proposed (Risser 1985, Agee and Johnson 1988, Grumbine
1990, Pickett et al. 1992, Frankfin 1993, Laroe 1993, Crow et al. 1994), there is strong
recognition that individual species cannot be separated from their system, because
complex relationships link species together and because species respond to the dynamics

of the systems that encompass them. In a holistic approach, the appropriate questions

 

thawedwb‘
ofspetrCS a'ef:
mumg‘.

The C01”.

 

 

and the uhoie ;“
(193$ Tangle?
mos-stem cm“
holistic aromas?
ms Interacting
may hie ar 1
lower Units

The cor
because recent 3
re

mighShipS be:

Cn‘vironmental c.

that need to be asked and the phenomena that need to be explained for the conservation
of species are relative to the behavior of the whole system (Allen and Star 1982, Naveh
and Lieberman 1994).

The concept of an ecosystem as a combination of the whole organism—complex
and the whole physical-complex forming the environment was first described by Tansley
(193 5). Tansley considered natural systems to be organized in a hierarchy; smaller
ecosystems embedded in larger ones, overlapping and interacting with one another. A
holistic approach likewise views the ecosystem as progressively subdividable into smaller
units. Interacting units at fine temporal and spatial scales produce patterns that are
recognizable at higher levels. Patterns and processes at higher levels in turn affect all
lower units.

The conservation of species through ecosystem approaches is becoming feasible
because recent advances in landscape ecology have provided a basis for establishing
relationships between the ecology of organisms and patterns of patchiness in
environmental conditions (F reemark et al.1992). Landscape ecology as a discipline has
emerged fi'om 2 different perspectives. One aspect of the discipline has focused on the
study of land heterogeneity and the underlying structures and dynamics that cause it
(Weinstein and Shugart 1933, Milne 1991). This facet of landscape ecology is based on
the physical factors (e. g.climate, geomorphology ) that operate at different scales.
Vegetation patterns that develop under natural disturbance processes reflect the patterns
of these hierarchical physical factors. This aspect of landscape ecology has provided the

basis for ecological land classification schemes (Barnes et al. 1982). An ecological land

classification stratifies a geographical area into a hierarchy of land units that comprises

 

the span sci—2

paciaicss m \t'
liill
. some

to raise pop-a:

itemize-res 1F

 

biological relatic
have an 60:14:10
cl amenities 0
to (Esteem are a
fine and Space
hm!) l0 reflec;
am: patterns

1991‘ Harper 1

the spatial scales at which major environmental physical factors affect patterns of
patchiness in vegetation (Bailey et al. 1978, Bailey 1985, Pregitzer 1981, Barnes et al.
1982)

Another more recent aspect of landscape ecology, which is still emerging, seeks
to relate population dynamics (e. g. productivity and dispersal) to spatial patterns of land
heterogeneity (Forman and Godron 1986, Dunning et al. 1992). The theoretical basis for
ecological relationships between the biology of organisms and landscape patterns can
have an evolutionary aspect (Wiens 1976, Riddle and Jones 1996). Life history
characteristics of species such as habitat selection, generation time, life span and ability
to disperse are adaptations to predictable disturbance events that generate patchiness in
time and space (Wiens 1976, Hansen and Urban 1992, Samson 1992). Because they are
likely to reflect evolutionary adaptations, relationships between population processes and

spatial patterns of patchiness are a potential basis for species conservation (Pickett et al.

1992, Harper 1992).

2. Neotropical Birds in the Northern Hardwood Forest.
Multi-scale patchiness of the vegetation is highly evident in the Upper Great
Lakes Region. The transition between the boreal forest to the north and the deciduous
forest biome to the south creates a pattern of large scale patchiness that is associated
with high bird species richness (Temple et al. 1979, Pastor and Brochart 1990). The
northern hardwood forest is a major component of the vegetation cover. Within it, a
finer scale of patchiness is evident as a mosaic of many difl‘erent tree species and

heterogeneous vegetation structure. This in turn is associated with ‘a within forest type’

high diversity in bird species (Noon et al. 1979). The large scale patchiness is a long-
lasting effect of landforms left behind from retreating glacials some 10,000 years ago.
The fine scale patches within the long lived northern hardwood forest, in contrast are due
to small scale forest disturbances such as blow downs fiom severe storms (F relich and
Lorimer 1991). Small scale patches also result from small scale variations is soils, local
topography and aspect that provide a diversity of local microsites for different plant
species to grow. Because processes of forest maturation and of frequent small scale
disturbance are continuous, the state of forest maturity in small scale patches is
continually shifting. Bird species must also continually track microhabitats created by
these small scale disturbances.
A large portion of the Great Lakes forest is second grth forest that was

heavily cut at the end of the 19th and early 20th centuries (Cunningham and White 1941,
Ahlgren and Ahlgren 1983 ). This second grth forest is largely managed for multiple
use and is subjected to logging. The Upper Great Lakes forest however still comprises
in several locations primary forest that is continuous over several square kilometers.
Such a forest provides the opportunity to assess how multi-scaled forest patchiness
resulting fi'om differences in environmental conditions and from forest management
induced disturbances afl‘ect the relative densities of bird species within the northern
hardwood forest ecosystem.

The majority, over 75% of songbird species at the latitude of the Great Lakes region
are neotropical migrant birds (Degraaf et al. 1992). Neotropical migrant bird species
winter in the tropics and subtropics but migrate to breed in North America (Koford et al.

1994). They make up a large proportion of the vertebrate biodiversity in many North

American ecosystems (Finch 1991; Freemark et al. 1992, 1995; Block et al. 1995). A
concern for these species in recent years stemmed from reports of declining populations
and species loss evident in some human dominated landscapes (Robbins et al. 1989a;
Askins et al. 1990). For many bird species, trends of declines and increases are not
consistent but vary by geographical region and by time period (Peterjohn et al. 1995).
Because of this variability, trends from specific localities cannot be extrapolated to the
general (James and McCulloch 1995). Reports of declines of some bird species (Holmes
and Sherry 1985, Robbins et al. 1989a, Askins et al. 1990, Hill and Hagen 1991,
Terborgh 1989, 1992) have nevertheless triggered a great concern for their welfare.
There is a general consensus that management of these species would be extremely
complex and the ecological consequences severe, if these observed declines were long
term and occurred over broad areas of species distributions’ (Senner 1988, Finch 1991,
National Fish and Wildlife Foundation 1994 ).

Limiting factors that have been outlined for neotropical migrant birds include
tropical deforestation, deterioration of habitat along migratory routes due to rapid urban
encroachment along coastlines, and habitat fragmentation. For the eastern deciduous
forest biome, declines in some bird species were blamed on forest fi'agrnentation. Forest
fiagmentation was represented by a pattern of forest patches irnbedded in a nonforest
matrix in early studies that have identified it to have a detrimental effect on forest interior
bird species (Whitcomb B.F. et al. 1977, Whitcomb RF. et al. 1981, Wilcove and
Whitcomb 1983). Forest interior birds nest in the interior of forest in contrast to those
that nest along an edge between openland and forest (Koford et al. 1994). Several

independent arguments have been presented to support forest fragmentation as a limiting

factor on the breeding grounds (\Vilcove 1987,1988; Finch 1991). Among these are the
high predation rates reported along forest edges (Gates and Gysel 1978; Wilcove 1985,
1987; Small and Hunter 1988; Yahner and Scott 1988) and high cowbird parasitism rates
in agricultural landscapes (Ambuel and Temple 1982, 1983; Brittingharn and Temple
1983). Forest interior bird species have disappeared from parks and forest reserves in
many human-dominated landscapes (Lynch and Whitcomb 197 8; Robbins 1979, 1989a;
Wilcove and Whitcomb 1983; Ambuel and Temple 1983; Blake and Karr 1984; Askins
et al. 1987; Terborgh 1989, 1992; Hill and Hagan 1991). Some forest interior bird
species that require large forest areas are considered area-sensitive species (Koford et al.
1994). Area-sensitive species such as the black-throated blue warbler and Canada
warbler (all scientific names are given in Table Bl) have become restricted to large forest
tracts in largely forested regions (Robbins et al. 1989b, Binford 1991).

The immensity of the task of conserving the diversity of neotropical migrant
birds, and the huge administrative obstacles that would likely be faced, once these
species started facing perils, provided the impetus for the formation of Partners in Flight
(National Fish and Wildlife Foundation 1994, Martin and Finch 1995). Formed in 1990
to address the conservation needs of neotropical migrants, to educate and to coordinate
conservation and research efl‘orts, the organization has grown to include many
government agencies and private organizations from the United States, Canada, Mexico
and several countries in Central America. As a result of the Partners in Flight initiative,
support for it fi'om the US. Congress and public response to it, the conservation of these

species is beginning to dominate many aspects of land management on public lands,

including national parks, state parks, and federal and state forests (National Fish and
Wildlife Foundation 1993).

A hierarchical organization for addressing species conservation needs has been
proposed and followed by Partners in Flight. In this strategy, conservation issues and the
species that need prioritization are identified at the highest geographical levels, such as at
continental and national levels (Noss 1983, 1987, 1992; Probst and Crow 1991).
Because forest fiagmentation has been identified as an important issue in species
conservation and has increased with increasing human domination of landscapes
(Whitney and Somerlot 1985; Wilcove 1987, 1988; Wilcox and Murphy 1985; Faaborg
et al. 1995), many conservation strategies have focused on forest interior bird species
that are area-sensitive (National Fish and Wildlife Foundation 1993).

The percent of forest cover in the Upper Peninsula of Michigan and in general in
the Upper Great Lakes today is not very different than it was prior to settlement by non-
native Americans (McCann 1991 ). This is in contrast to the Lower Great Lakes area
where agriculture and urbanization have resulted in forest loss and fragmentation. The
highly forested northern Great Lakes region, stands out from a regional perspective to be
most promising for the conservation of forest interior and area-sensitive bird species.

The conservation of these species has also received public attention (Probst and Crow
1991, Crow et al.1994) in forest management of this region (Northern Wisconsin and
the Upper Peninsula of Michigan).

Because of the expansive area it occupies, the northern hardwood forest is an

ecologically important forest type in the Upper Peninsula of Michigan. It is an extension

of the eastern deciduous forest (Braun 1950, Pastor and Mladenofl‘ 1992). The northern

ll't‘iOOClDYt

trier llt‘JIJ.

 

times such as -

lliiixri e. i

“mm \atic
H9”

hardwood forest bird communities comprise many of the same species found in the
eastern deciduous forest, including many that have been identified as area-sensitive
species such as the black-throated blue warbler, Canada warbler, and northern parula

(Robbins et al. 1989b).

3. An Ecosystem Approach to Relate Bird Species Densities to Forest Heterogeneity.
The research reported herein, was motivated by the participation of the
Hiawatha National Forest in the ‘Neotropical Migrant Bird Initiative’ set by Congress
(1992) to advance ‘neotropical migrant bird’ conservation in national forests. I focused
on the northern hardwood forest ecosystem as this is the forest type that harbors many
bird species (e. g. ovenbird, black-throated blue warbler, northern parula and red-eyed
vireo) that have been identified as forest interior and area-sensitive species (Robbins et
al. 1989b). In a region that is still largely forested, the assembly of bird species that are
forest interior and area-sensitive would be more complete. In such a system the different
forest bird species are likely to greatly overlap in their habitat selection. Ifthere is a
gradient in the aspect of forest fiagmentation, bird species that are most forest interior
specialists could become afl‘ected by competitors that are less sensitive to habitat
fragmentation at some point along that gradient. To prioritize the conservation of the
most area-sensitive bird species at the level of a national forest, it is important to identify
what factors contribute to fragmentation effects. It is important to test whether variables
such as canopy opening, amount of forest edge and forest cover (i.e., variables that
could be related to a forest fiagmentation dimension) affect density differences between

highly area-sensitive and the less area-sensitive bird species. The heterogeneity of the

 

 

 

 

hairs 11’s: :7;
iii: the at?»
the bird 5pc;

Imam? forest

 

A come:
he iota‘ hetero!
also series to f:
species in zesza':
imitation to :
bird species are
Denim hall's
b." My facrors

lorce in nature?

WW feature

lid dmui’bance

northern hardwood forest would need to be assessed relative to these environmental
factors that create these difl‘erences. In this undertaking two related hypotheses were
made: the densities of forest bird species, and the relationship among forest interior and
edge bird species differed along the range of heterogeneity encountered at the scale of a
national forest.

A conceptual model was necessary to represent the hierarchy of strata into which
the total heterogeneity of the northern hardwood forest would be divided. The model
also serves to frame hypotheses about the importance of different factors on forest bird
species in testable formats. A theoretical basis however was needed to give biological
justification to the structure of the conceptual model. The ecologies of forest interior
bird species are related to processes of small scale disturbance and succession in the
northern hardwood forest. The pattern of small scale forest processes could be affected
by many factors including geographical location and forest management. An emerging
paradigm in ecology (Pickett et al. 1992) that considers disturbance to be an organizing
force in natural communities was considered. This paradigm justified considering
physical features of local geographic areas, past disturbance history, current land use,
and disturbance caused by current management as main factors affecting variability in
vegetation structure and composition and consequential bird responses. Bird densities
are considered to reflect bird population responses to that heterogeneity. The
conceptual model of the northern hardwood forest would serve to give a structure to
that heterogeneity and to provide a vehicle to reduce that heterogeneity to spatial scales
that are pertinent to the life histories of forest bird species. The model is an exploratory

tool that would provide new perspectives for looking at relationships between habitat

 

 

 

 

fl

clinician. 2
NM meow

mam ha: i u“.

 

Tm;

“Reich pff_
lorcsz on lOl'CS‘.
[57,9013 scales

MWOOG loves

 

Chilling

A“ 0\ er:
forest varied sir,
lamination d
fleet i

he relahi

'U‘

firm framem

I) Dfi'eIOp a CC
norm had‘*-l

Willem anlhtop,

charateristics and bird species densities and at habitat relationships among species.

Newly uncovered relationships would in turn be used to refine a future model of the
northern hardwood forest ecosystem. Such a refined model could be used to more
effectively predict the efl‘ect of different management strategies in the northern hardwood
forest on forest interior bird species. Justifications for variables, factors and spatial and
temporal scales of sampling units incorporated in this initial model of the northern

hardwood forest are discussed in the Methodology Section.

OBJECTIVES

An overall objective of this study was to test whether the northern hardwood
forest varied suficiently, along habitat variables which are related to a forest
fi'agmentation dimension, to afl‘ect the densities of forest bird species in general, and to
afl‘ect the relative densities of bird species that are known to be differently sensitive to

forest fragmentation. Specific objectives of this study were to:

1) Develop a conceptual model to represent and stratify the heterogeneity of the
northern hardwood forest at difl‘erent spatial and temporal scales, based on past and

current anthropogenic forest disturbance, and local geographical area.

2) Describe how factors incorporated in the model affect densities of 10 bird species in

the northern hardwood forest. More specifically:

lO

 

 

i) Test the effects of historical divergence in land use, time elapsed after
logging disturbance, and local geographical area on densities of 10 songbird
species.

ii) Identify microhabitat elements afl‘ecting densities of 10 songbird species.

iii) Compare the availability of major microhabitat elements for each of 10
songbird species across different land use trajectories of the northern hardwood
forest consisting of : primary forest, managed forest and settled forest and across
time periods elapsed after logging.

iv) Compare the effects of landscape level factors (availability of habitat at the
landscape level and amount of edge ) to the effects of microhabitat availability on
4 forest bird species densities.

v) Test how interactions among landscape and microhabitat factors affect
interspecific relationships of 3 bird species (black-throated blue warbler,
American redstart, and chestnut-sided warbler) that each nest in the sapling layer

but differ in their requirements of forest interior habitat.

3) Explain how forest disturbance processes, and land characteristics as might be

described in an ecological land classification, can be used to integrate the conservation of

forest interior bird species into forest management.

4) Evaluate the function of managed national forest relative to forest interspersed with

agriculture and forest reserve areas with respect to providing habitat for a diversity of

11

sonrhrc' s

KSSOCEIE‘C

Zamora a

ELYOpOg

él EQTSHI

Processes g

Sim iii
I General

ler

songbird species in general, and forest interior bird species in specific. Discuss aspects

associated with forest fiagmentation in the context of a largely forested region.

5) Evaluate the use of a holistic approach (e.g. multi species, hierarchy of different
temporal and spatial scales, species interrelationships) to assess the effects of

anthropogenic impacts on bird species.

6) Examine how a holistic approach to assessing human interference with forest

processes differs from comparing bird densities before and after a specific disturbance.

STUDY AREA

1. General Description of the Upper Peninsula

Ten study sites, located in the central and western Upper Peninsula of Michigan
(Figure l), spanned an area between 42° 50' and 46°52' N latitudes and between 86°
5' and 89° 50' W longitudes. The climate in the Upper Peninsula of Michigan is
dominated by the northern location of the Upper Peninsula and by the proximity to
Lake Superior and Lake Michigan. It is characterized as continental humid. The Upper
Peninsula is located in an active weather zone (F relich and Lorimer 1991). Some of the
most severe thunderstorms occur in the summer when polar air masses collide with sub
tropical air masses. Summers are cool and short and winters are cold with average July
temperatures varying fiom 19° C to 19.8 ° C and average January temperatures of

-7.S°C to -10.9° C (Frelich and Lorimer 1991, Albert et al. 1986). Annual precipitation

12

doc—-82 88:38

803 wee—9:8 cozfiow? ecu measmcoo BB 523 E 53:22 Le Esmficom 3&3 of E magnum—Ea— o. mo 5:801— ._ 235

2 <5 Semen—Jo.

         

e <5

55. ES.

25 9:75.52 beams;
e 55.2.0.

25 when... e
o

 

woes—=0
328.8932
C

85552 .83:
O

 

 

l3

318?“:ng Sid
Permit ta
Perirstla d1:

Phis
Skirts. 0:;
“file IOTTne:

all“: :e is de

1 1 .
pep,“ Wm:

maTile. bee,

White Pine 3

averages 80-90 cm and is evenly distributed throughout the year. The entire Upper
Peninsula was glaciated. Glaciers began retreating some 16,000 years ago but the Upper
Peninsula did not become ice free until 10,000 years ago.

Physiography is largely determined by features left by the retreating glacials: end
moraines, outwash plains, till plains, lake plains, kettle lakes, kames and eskers. Soils
were formed by erosion and weathering processes acting on glacial materials. Local
climate is determined by proximity to either Lake Superior or to Lake Michigan. The
vegetation of the Upper Peninsula falls under the hemlock-white pine -northern
hardwoods region described by Braun (1950). The Upper Peninsula occurs in a zone of
transition between the boreal forest occurring to the north and the deciduous forest
biome more to the south (Braun 1950, Pastor and Brochart 1990). Boreal tree species,
black spruce, white spruce, balsam fir, large-toothed aspen, trembling aspen, balsam
poplar, white birch and jack pine mix with eastern deciduous forest species sugar
maple, beech, basswood, yellow birch, red maple, American elm, white ash, hemlock,
white pine and yellow birch. The sequence of tree species establishment following the
retreat of glacial ice has been described by Davis et al. (1995). Sugar maple, a dominant
Species in the present day ecosystem did not get established in the Upper Peninsula until
about 6,000 years ago; black spruce established some 10,000 years ago.

Northern hardwood forest1 (‘) occurs over a large range fiom shallow soils over
bedrock to poorly drained soils. Graham (1941) and Braun (1950) have described in
detail the variability of the northern hardwood forest in the Upper Peninsula of Michigan.

¥

1‘ FOOtnotes are given by chapter beginning on page 163.

14

Mos timed
llbbts 19"

13351183367113

1381156 and c

Most typically this forest developed on fine-grained loamy soils that are well drained
(Tubbs 1977). The component species differ in shade tolerance and moisture
requirements and establishment ecology. Tree species composition varies based on soil
texture and content of organic material, topography, and disturbance history (Brubaker
1975, Tubbs 1977). Secondary northern hardwood forest is predominantly sugar maple
in all variations of the type (Tubbs 1977, Tubbs and Goodman 1983). The high diversity
of tree species and the ecotone of boreal and deciduous forest result in a
characteristically high bird species diversity; among the highest in North America
(Temple et al. 1979).

A large portion of the Upper Peninsula is in public ownership, mostly state and
national forest. Large tracts are also commercial forest. Human settlements occur in
dispersed small towns and clustered agrarian communities. Agriculture consists mostly of
pasture and hay production. Although the Upper Peninsula remains largely forested, the
present day forest is very different from the original forest that existed some 100 years
ago (McCann 1991). Much of the present forest is second growth and differs in the
extensiveness of forest types and in the relative tree species composition. The relative
importance of tree species within forest types has changed. The vegetation patchiness,
the interspersion of different forest types and the extensiveness of a particular forest type
depend on the constituent land forms, topography and soils, local climate and historic

diStlfl'bance. In some areas a particular forest type would stretch for kilometers, in others
“‘9 imerSpersion would occur at intricately small area scales (Braun 1950). The
r eSimmlization of landscapes of the Upper Peninsula (Albert et al. 1986) have delineated

different sections and subsections of the Upper Peninsula that differ ecologically because

15

 

refer it.

It
(I‘vefgec' 2‘

10 a difer

of difl‘erences in local climate and physiography. Land type associations (LTAs) 2,
smaller units in a hierarchy of spatial scales organizing ecological heterogeneity based on
physical factors and potential vegetation (Appendix A; Table A1), have also been

delineated on national forest lands (Appendix A; Table A2).

2. Description of the 10 Study Areas
Ten study areas were selected to replicate 3 different land use categories that had
diverged historically since the turn of the century. In general, a study area corresponded
to a difl‘erent land type association (LTA). Percent composition of different land cover
types for each of the 10 areas is given in Table 1. The Huron Mountains, McCormick
Wilderness Area and Sylvania Recreational Area, represented 3 different land type
associations replicating primary forest. Both the Huron Mountains and the McCormick
“filderness Area fall in the Superior Upland Subsection of the white pine -hemlock-
northern hardwoods region described by Braun (1950) and in the Michigamme
Highlands in the regionalization of Michigan landscapes (Albert et al. 1986). The Huron
Mountains study area was located within a 7,285 ha forest preserve area (Simpson et al.
1990). The Huron Mountains border the south shore of Lake Superior in Marquette
County. The McCormick Wilderness Area is part of the Ottawa National Forest and is
1owned south of the Huron Mountains in Marquette and Baraga Counties. The study
area was located within a designated natural area in the McCormick Wilderness Area.
This area is approximately 1550 ha within a 7000 ha tract (Pregitzer and Barnes 1984).
Areas along Lake Superior are under heavy lake effect but this effect declines

Freeipitously inland. The annual extreme minimum varies from -23 ° C along the lake to

16

:7. ....A VFAV _.

:2» v. vi. -.—.a v.—

e 7: v. V—A VP

cg-ZA~3~FF

Spires... >.-::V. qw~ :.-

gay. vv-A VLLAMF

aren’t e Feta-Pd ‘ :2:

s... ~= i=7; enhu~= v.0 :FUULUnN -

.I.~\< :. ts: N
1/

 

 

 

-~ fifths

 

 

 

mam e2 who one be mu:
2m o min 8.8 8.? x 89.20
8. _ owed e8 3:. won be one
8.: wm.~ Sow was. £8 on Sun: seem Boom
Es who own new s: be. mu:
:2 3m 8.8 8.8 8.8 x $54 Seance:
:L SN mom 3. a: 3 en:
83 a}. 2.2 e3: 8.8 x 72.5 2.352 means:
:3 so 3 wood :3 a. «no
So mam 32 8.3 8.8 x nose EeoEeooxm moses
8e 85 5e :5 Be a mu:
8.: 8.8 one 2.8 5% x 82 .eeoeeaoom seesaw
m3 m3 mi 8.2 a; be. Nu:
m: :e 5: 3.8 5.8 x 8:. aoEoEB soméoooz
Rd o3. as 3 no ea 3:
as 8.8 3% 2.3. me fix a: 83:
szmeebe M2022.509 6285... 35293:... 638.59

 

 

{l

.988 rose 2 E 89: 33o new. we 8.58888 .5ch 4 29¢

17

Vr. r» 2 .u». r... . It

 

 

e2...: :3... ..Z. e.3).._._.C._. .23.»...39 1322232... .CLZC-.....C.~.

 

 

-vrv==m=~afuv . N sexes-.- N

6:8 H 52:3 9:2an .82 u meOPOPo

.0qu 8231832an 98 Hoe—Eu: 28:26 5:32: 22 u ZOUEiHerFe
.399 628 How—Eco =m 2:280 Sacco :39 u 200.50,—a

.meooBEE 8058: 365828 pecans..— HSS N 33203:?

6:8 H 555 B>oo 628 H89 n OUdOnHHOHn

.538 2: Ho Hobo 92286 n 8..

.52: n xm

.cozmHoOmmm cab HES u <91:

.onBHo 2:: H wean—£53 no: He 0:8 H Ho 238:: n :.

 

 

e3 :3 8w of 2a om mu:
£2 :3 Rem $2 2.8 x eoneaosm
an new 3m 3% a one
2.2 c on? 8.2 3.8 an ego:
,zmaoeoe ozone—9:59 h28.5% eamzoBE 688.5...

 

 

6

985.83 .H 292.

18

-35‘ C rind
lips Penis
tie stiles 1c
filmed dc
and red pine,
r95 wit b;
delineated 5c
Ofgeomerp':
time (Sir:
“We n
mbmon (
Pores. in,
A 3,
National F0
b.“ a Stale,
{W91 mc
l9? comm
HUIOn M01
laid ape as

WHOM F0]

-35° C inland (Albert et al. 1986). The higher elevations relative to other regions of the
Upper Peninsula, are due to the underlying preCambrian bedrock. Much of these areas
are shallow to bedrock and bedrock is exposed in locations. Deeper mineral soils support
hardwood dominated northern hardwoods while shallow soils on ridges support white
and red pine, red oak and aspen. Swampy depressions support conifer forest, hemlock,
red maple, balsam fir and black ash. Although land type associations had not been
delineated for the Huron Mountains and the McCormick Wilderness Area, delineations
of geomorphological features (F arrand and Bell 1982) and detailed descriptions of local
climate (Simpson et al. 1990, Pregitzer and Barnes 1984) were suflicient to at least
recognize that these 2 areas would be in different land type associations within the same
subsection (E. Padley, US. Forest Service, Milwaukee, Wis; J. Jordan, Ottawa National
Forest, MI; pers. comm).

A 3rd primary forest area, Sylvania Recreation Area is part of the Ottawa
National Forest and is located in Gogebic County (Figure 1). The area is characterized
by a swale and ridge topography. The dominant landform is a rolling, sandy and loamy
terminal moraine. Depressions are occupied by lakes and bogs. The low relief, pitted
ice contact topography and heavy drift cover are in contrast to both the McCormick and
Huron Mountains. Sylvania Recreation Area has been classified under LTA-2 in the

land type association delineations of the Ottawa National Forest (J. Jordan, Ottawa

National Forest, MI; pers. comm) .

19

Forest in bit-cl In:

 

seated to nth-lair
illiriigbsm
plied Donia: hi.
to represent we
lie Huang )1.
lg: County ll
he rigor-am:
l7) like Supem
ltx‘tioped up’ii
toms of rela
Steuben Sega
norhem hard-
if llA ~7A l
When hirc
Elmer pan 1
Wed for
Area

“gel COUm

Three land type associations within the West Unit of the Hiawatha National
Forest in which the northern hardwood forest comprises a major forest type were
selected to replicate managed forest. The Munising moraines (LTA-l in Appendix Table
A1), a ridge-swale complex on the Rapid River Ranger District (LTA-7A) and a lake-
pitted northern hardwood area on the Manistique Ranger District (LTA-4) were included
to represent areas of northern hardwood forest within the core of the national forest.
The Munising Moraine (LTA-l) is located along the southern shore of Lake Superior in
Alger County (Table A1). It falls under the Luce District-Grand Marais subdistrict in
the regionalization of Michigan (Albert et al. 1986). The local climate is heavily affected
by Lake Superior with heavy snowfall in the winter. Soils consist mostly of well
developed upland sands and loamy sands of morainal origin. The natural vegetation
consists of relatively continuous northern hardwood forest. The Manistique LTA-4 is the
Steuben Segment of the Newberry Moraine. Loamy sands of morainal origin support
northern hardwood forest. Depressions are occupied by lakes. The ridge-swale complex
of LTA -7A formed by the retreating glacial Lake Algonquin. The upland areas supports
northern hardwood forest and the lowlands support a mixture of lowland conifers. The
greater part of the northern hardwood forest in all 3 land type associations is intensively
managed for timber.

Areas that comprised several square miles around the settlements of Chattharn in
Alger County, Trenary in Alger and Delta counties and on the Stonington Peninsula
(Table A2) in Delta County were included to represent settled forest landscapes. The
climate on the Stonington Peninsula is dominated by lacustrine influence. Severe

thunderstoms are not as frequent nor as intense because temperatures are moderated by

20

lale erletts I

mess ll‘; 2“

J

O

\

f

the It a"; it
misses (.3
lard user cl
l'p'pe: Pens:
the mt’hielalii
ml 1985)

The '.
area Iits in i]

Climbed ab

lake effects. Thin soils that originally supported northern hardwood forest are at present
mostly in pasture. Poorly drained soils support lowland conifers. Both the Chattam and
the Trenary study areas occur in the Dickinson district in the regionalization of Michigan
landscapes (Albert et al, 1986). The landscape comprises poorly drained sandy outwash.
Land use includes agriculture, mostly pasture. Both areas occur in the interior of the
Upper Peninsula. Air mass instability storms are more common here in the absence of
the moderating efl‘ects of lakes. Average annual minimum temperature is -29° C (Albert
et al. 1986).

The 10th study area consisted of a large unlogged (old growth) northern
hardwood area and adjacent logged areas within Dukes Experimental Forest. This study
area lies in the Hermansville subdistrict in the Dickinson District (Albert et al. 1986)

described above.

21

BiliDZXi'i
1 Spain

Species

CHAPTER 2

METHODOLOGY

BUILDING AN ECOSYSTEM MODEL

1. Spatial Scales Pertinent to Assessing the Effects of Forest Heterogeneity on Songbird
Species.

With advances in landscape ecology, the importance of using relevant spatial and
temporal scales to describe ecological processes is increasingly receiving attention (Wiens
1981, Wiens et al. 1986a, Meentemeyer and Box 1987, Wiens 1989, Golley 1989, Turner
1989, Allen and Hoekstra 1990, Dunning et al. 1992, Ruggiero et al. 1994). Wiens (1976,
1981) and Kotliar and Vlfiens (1990) discussed how different organisms perceive the
environment at different spatial resolutions, and how different ecological processes occur
at multi scales that differ from anthropocentric scales. Ongoing discussions about scale
suggest that the effects of environmental variables on forest bird species could be
predicted with more precision in the future, if these effects are studied at spatial scales that
are pertinent to the ecologies of forest birds.

Density relationships among forest bird species that might differ in their sensitivity
to forest fragmentation, would be best revealed at spatial and temporal scales at which
habitat selection and competitive interactions among bird species occur‘. There are
indications from bird census comparisons and fi'om field studies that bird species select

habitat at different spatial scales (Wiens and Rotenbeiry 1981, Anderson 1981, Gutzwiller

22

and Anders)
edit Pater:
513117.33 33.;
1931 ’ \lic:
are 5851 con
iii-Girlie; a
team 0c;
Sfieetion an:

M03
halo“ Ola
wished 1
a: me 5'13 IL
Egg ’3‘ Ha
511%qu an
19?]. We
rtiallonshjp
ell‘q'ronmen
“was”,

The
minimum).
kilns e: a]
Weep: to r

membeddec

and Anderson 1987). Macrohabitat variables such as size of habitat and distance from
edge, patterns of horizontal and vertical heterogeneity are correlated with bird community
structure and geographical distribution of species (W iens and Rotenberry 1981, Anderson
1981). Microhabitat variables that quantify vegetation structure and species composition,
are best correlated with habitat selection by individual species within a habitat type
(Holmes and Robinson 1981). Habitat variables measured in small areas reveal differences
between occupied sites vs unoccupied sites for individual species, as well as habitat
selection among different species within a habitat (Anderson and Shugart 1974).

Most forest songbird species have small nesting territories in the range of a
fraction of a hectare2 (many Parulidae species) to a few hectares. It is now well
established that bird species perceive patchiness of vegetation structure and composition
at fine spatial scales (James 1971, Anderson and Shugart 1974, Rotenberry and Wiens
1980 )3. Habitat selection of songbird species has been measured by sampling vegetation
structure and composition in small plots (0.04-0.08 ha) within nesting territories‘ (James
1971, Anderson and Shugart 1974, Whitmore 1977, Noon et al. 1980). Habitat
relationships among bird species have also been established from such fine scale
environmental measurements (James 1971, Anderson and Shugart 1974, Whitmore 1977,
Rotenberry and Wiens 1980)‘.

The recognition that environmental factors at several spatial scales affect bird
community structure and composition (Wiens and Rotenberry 1981, Anderson 1981,
Askins et al. 1987, Robbins et al. 1989b) has led to the consideration of a landscape
concept to refer to a mosaic of habitat patches in which a particular type of a habitat patch

is embedded (Dimning et al. 1992, Freemark et al. 1995). The landscape ° concept has

23

 

been usec 355
chi'aetenstic:
because the 5:
considered a l
aspeclcs. hor
preparation so
ocean) suzta‘:
irrigation a
dispersmg ind
al 1995 links
metapopnlazin
ntnchange in
locations thus
interchange of

PTOducrhny ’

[linings a]

been used as a theoretical basis for linking population processes of bird species to land
characteristics. From an organism perspective, no absolute size is given to a landscape
because the size of a habitat patch is specific to the organism. Dunning et al. (1992)
considered a landscape to occupy a spatial scale that is intermediate between the scale of
a species’ home range and that of its regional distribution. Landscapes differ as
population sources of a species and thus in their ability to supply individuals that would
occupy suitable habitat sites when these become vacant. Landscape characteristics affect
immigration and dispersal rates by the degree of environmental resistance they present to
dispersing individuals. Population source or sink theory (Pulliam 1988, Donovan et
al.1995) links landscape characteristics to population productivity. The concept of
metapopulation theory considers a population to be composed of subpopulations that
interchange individuals (Opdarn 1991). The abundance of a bird species measured at point
locations thus can be affected by landscape characteristics that control the rate of
interchange of individuals among subpopulations (immigration into area) and species
productivity 7 in a larger area surrounding the census point (Askins and Filbrick 1987,
Dunning et al. 1992, Freemark et al. 1992, 1995).

For the majority of songbird species however, little is known about the spatial and
temporal scales at which populations exchange individuals because of the many population
and environmental variables involved. Organism related variables can include the inherent
capability of the species to disperse, and numerous population processes that afl'ect yearly
fluctuations and population saturation. Land characteristics that could affect resistance to
dispersal and population productivity could include structural characteristics, availability,

and spatial configuration of patches at the smallest scale that bird species respond to

24

(patch that would illicit a response by the bird species to settle into a territory ); the spatial
configuration of small scale patchiness at larger spatial scales (the relationship of small
scale patchiness with higher level patchiness); characteristics of the physical structure
(landforrn complexes, soil complexes) at difl‘erent spatial scales that affect patchiness, and
spatial and temporal scales of disturbance processes (natural or human-induced) that afl‘ect
patchiness at different spatial scales (Kotliar and “Wiens 1990) .

There are increasing attempts and calls in bird ecology to relate bird population
characteristics and processes to landscape characteristics in spite of the complex ways in
which population and land characteristics could interact (Freemark et al. 1995, Petit et al.
1995, Donovan et al.1995). The effects of landscape characteristics on songbird species
have mostly been revealed fi'om empirical studies undertaken in landscapes in which loss
of forest cover was extensive (Ambuel and Temple 1983, Askins et al.1987, Freemark and
Merriam 1986 ). This aspect is largely due to the recentness of the landscape concept, and
the increasing concern with habitat fragmentation as a threat to species persistence
(\Vilcox and Murphy 1985). Several studies however have included a wide range of
regional forest cover that extends into landscapes in which forest loss has been extreme
(Lynch and Whigham 1984, Robbins et al. 1989b). Landscape level effects were often
represented by forest cover measured within a range of 1-3 km from the point at which, or
fiom the habitat patch in which, bird abundance was estimated (Lynch and Whigham
1984, Askins et al. 1987, Robbins et al. 1989b). Relationships of bird species abundances
to landscape characteristics, and to territory level characteristics (vegetation structure)
were considered linear and additive; represented by stepwise multiple regression and

multiple correlation models.

25

 

 

Vinita

1

iii metres Ci"
sttlll Smiles

nomination r...

resource os chi:

 

l'anabllns in pa
response to l at;
‘ “titties A :
llfit’lOgt’llClll c
chainensties
Micah: cha;
at diherent Sp
Species and la
Ines Penna
aPOSInliIe ii

That
Winnie
Wanted
WmICOmb l

1933, Askjr

Variation in landscape level characteristics however could result in variability in
bird species composition and in the relative abundance of species in different landscapes.
Several studies have shown that a difi‘erence in the assemblage of bird species (species
composition and the relative abundance of species ) afl‘ect the habitat utilization and
resource overlap among bird species (MacArthur 1958; Cody 1974,1978, 1981).
Variability in patchiness at the landscape level could cause variability in the species
response to (and therefore the species abundance in) similar small scale habitat in different
landscapes. A more general model of the efl‘ects of northern hardwood forest
heterogeneity on forest bird species, would be one that considers interactive effects of
characteristics of small scale habitat (spatial scale representing the territory level) and
landscape characteristics. Such a model would explicitly recognize that variability exists
at different spatial scales. Relationships between the relative abundance of forest bird
species and landscape characteristics for northern hardwood forest landscapes in the
Upper Peninsula are however not known. The interactive model is therefore presented as
a postulate in need of testing.

That forest fiagmentation (many human- induced changes have forest
fi’agmentation aspects) has affected forest bird species abundance has been well
documented in landscapes in which loss. of forest cover has been extreme (Lynch and
Whitcomb 1978, Whitcomb et al. 1981, Freemark and Merriam 1986, Ambuel and Temple
1983, Askins et al.1987). Comparative studies that considered a range of regional forest
cover' in which forest fragmentation is at advanced stages have shown that area sensitive
bird species are replaced by habitat generalist birds in fragmented landscapes. The

apparent area sensitivity of some forest bird species (those species only found in extensive

26

lOICSil is his
to forts: 5735'
literature tha‘
numerous spa
species is a s
198}!

The 2
pi ' nee of l
as sepegatéc
recenils the
integenei:
Songhird spe
Wihiers of 1
millde to
hit, rum rel;

Scale disrur'b

forest) is likely to be the result of competitive exclusion from species that are less sensitive
to forest fragmentation and to forest loss. This postulate is based on a large body of
literature that has described habitat occupancy by a species to be a process mediated by
numerous species interactions in contrast to the premise that habitat occupancy by a
species is a simple response to habitat conditions (Cody 1974, Morse 1976, Cody 1978,
1981)

The ability of ecologically similar bird species to coexist has been attributed to the
presence of coexistence mechanisms. Cody (1974 ) recognized some of these mechanisms
as segregation by geographical area, by altitude, between habitat and within habitat. Until
recently the limited spatial extent of single studies precluded the need to describe
heterogeneity in a hierarchy of spatial scales. The ability of many ecologically similar
songbird species to coexist within the northern hardwood forest (e. g. many species of
warblers of the family Parulidae; many even within the same genus Dendroica) has been
attributed to a high vertical and horizontal heterogeneity in the forest. This heterogeneity
is in turn related to small scale variations in topography, aspect, soils and frequent small
scale disturbances that continually reset forest maturation processes and provide many
different niches to different bird species. In this single scale perspective”, the northern
hardwood forest would be viewed as made up of units the size of bird territories (in the
range of a fraction of a hectare to several hectares) and occupancy of these units by
individual species to be solely determined by availability of the necessary vegetation
components.

In reality however the effects of northern hardwood forest heterogeneity at higher

scales could modify bird species response to small scale heterogeneity. In managed forest

27

there the 1
if i'ei sens
the mash}:
abundance

'

ml;

F

(I
('J
f;

species sin
assessed u
parental}
conditions
nether (c1
methane:
ml“ €01
appropriat-

mmattion

where the range of variability in forest cover among landscapes is narrow, it is not known
if area sensitive forest bird species are differently afleaed in different landscapes. Should
the variability of landscape characteristics be large enough to impact the relative
abundance of bird species in the larger landscape, then habitat occupancy of the
microhabitat sites could be affected. A spatial scale at the level of the patches that a
species would establish a territory in is the spatial scale at which habitat occupancy can be
assessed with the least amount of noise. Habitat overlap between 2 species (that are
potentially competitors) also needs to be measured at this level (under constant landscape
conditions) because this is the spatial scale at which the effects of one species replacing
another (competitive exclusion) can be observed”. Based on the frequency of small scale
disturbances in the northern hardwood forest and the small territory size of the bird
species considered, spatial scales in the range of 0.5 to several hectares would be most
appropriate to investigate the effects of vegetation structure on habitat occupancy and on
interactions among bird species (Wiens 1989).

Basic biological information concerning spatial scales at which population
processes of individual species occur (e.g. immigration into area, species productivity) are
lacking. Details about differences in the structure of bird communities (bird species
composition and their relative dominance) in different landscapes are also lacking. A
spatial scale of (1.6 km) was herein used to examine the effects of landscape scale
variables on bird densities based 0n the fact that it is within a range of spatial scales for
which correlations between bird abundance and landscape level variables were found to
be statistically significant for a majority of forest bird Species (Lynch and Whigharn 1984,

Askins et al. 1987, Robbins et al. 1989b)11 . The relationship between bird species

28

a,“

Pin

the

197

abundance and landscape level characteristics at these scales have only been emperically
derived, and have not been determined from actual examinations of species dispersal
patterns and productivities. The statistical significance of relationships between bird
abundance and landscape level variables at these spatial scales, for a majority of forest bird
species (Robbins et al. 1989b), suggests however that these scales must be relevant to at
least some aspects of forest bird species ecologies, their life histories and population

processes.

2. Sources of Forest Heterogeneity in the Northern Hardwood Forest and Their Impacts
at Scales Relevant to Forest Bird Species.
Structure of forest heterogeneity.

Landforms left behind from retreating glacials in the quaternary provided the
physical setting in which local ecosystems developed. Landforms differ in their soil
composition, slope, aspect and position in the landscape. Because they vary in their
physical structure and topography, and to some extent influence local climate, landforms
affect the development of vegetative communities and the rate at which succession and
forest disturbance occur (Swanson et al. 1988). The patterns of distribution of the
northern hardwood forest and its association with other forest types in the presettlement
forest of the Upper Peninsula is predominantly influenced by the pattern of glacial
landforms (Brubaker 1975, White and Mladenoff 1994). Large scale disturbance such as
fire and hurricanes have been hypothesized to be disturbance factors in pine forest and in
the more western extension of the northern hardwood forest (Graham 1941, Heinselrnan

1973, Canharn and Loucks 1984). For the northern hardwood forest in the Upper

29

ind;

Peninsula and more easterly, small scale disturbance in the form of tree blow down is the
most common form of disturbance (Nichols 1935; Lorimer 1977; Borman and Likens
1979a, 1979b; Runkle 1982; F relich and Lorimer 1991). The variability of the northern
hardwood forest in structure and tree species composition across its broad geographic
range as well as within the Upper Peninsula has been described“ by Braun (1950). The
landforrn is also likely to affect the structure and species composition of the vegetation at
a more local extent. The current state of the northern hardwood forest however is also
the result of human influences that have altered it from its mostly primaeval condition
before clearcutting began in the late 1800's. Under current forest management (under all
ownerships) a greater portion of the northern hardwood forest is periodically logged.
Large scale heterogeneity within the present northern hardwood forest is the result of the
overlaying of largely human-induced and recent multiple land disturbance events on
heterogeneity that had been laid down by very large scale geological events.

It is not possible to assess directly the amount of variability that the northern
hardwood forest of the Upper Peninsula presents in terms of available habitat fOr
individual bird species at the microhabitat level, or in terms of forest fragmentation aspects
at a landscape level. This information however is required to test if forest bird species
could be impacted by forest fiagmentation related variables12 within the range of variability
presented in the northern hardwood forest of the Upper Peninsula. An alternative to direct
assessment of songbird based variables and spatial scales, is an indirect methodology that
considers a hierarchical structure in the partitioning of the heterogeneity of the northern
hardwood forest. A first step in this methodology is to develop a conceptual model that

would break down the total variability of the northern hardwood at the spatial extent of

30

- 3
, is,

fill.“
hie:
ling

Pia

the Upper Peninsula, into the hierarchy of spatial and temporal scales that are related to
major environmental and human induced factors that have created this heterogeneity. In
this hierarchy, dominant environmental or human induced factors of higher levels would be
assumed to affect units at lower levels imbricated within them. A second step is to
consider the scale at which habitat occupancy of the bird species can be examined (at the
territory level) at the appropriate level in the hierarchy. This would be justified if there is a
basis for expecting factors of heterogeneity (at higher levels in the hierarchy) to affect
habitat structure at the territory level, and to modify characteristics at the 1.6 km radius
scale. Although the model would be based on known factors of heterogeneity, there is no
way of knowing apriori how higher levels of heterogeneity affect individual bird species,
their habitats at a territory level, and within a 1.6 km radius landscape level. Its utility
needs to be evaluated with respect to how well it explains heterogeneity in bird abundance
and in habitat structures they require at the appropriate scales. The test of the model also
lies in how well it uncovers relationships between human induced disturbance, ecosystem
processes and bird species ecologies.

A hierarchical ecosystem concept was taken by considering the northern
hardwood forest (within the range of the Upper Peninsula) as one large ecosystem that is
divisible into smaller interacting ecosystems. The structure of the model was based on a
hierarchy of spatial and temporal scales that correspond to climatic and geological factors,
known anthropogenic disturbance factors, current land use and forest management
practices. The structure of the model provided the higher stratification design for
sampling bird densities and habitat at the territory level and at the 1.6 km radius landscape

level. The structural components of the model and the among strata comparisons that this

31

structure makes possible are presented in Figure 2. Table 2 lists the major factors causing

variability among units at each level.

Historical events that added a human induced factor to forest heterogeneity

Large scale clearcutting that occurred over much of the Upper Great Lakes region
in the interval between the 1880's until the 1930's marked a historical change in the course
in which the existing forest was unfolding. The disturbance level that was created by this
event was unprecedented since the Upper Great forest began establishing 10,000 years
before, after the retreat of glacial ice (Ahlgren and Ahlgren 1983, Frelich and Lorimer
1991). Land use changes that followed clearcutting brought anthropocentric control over
much of this region’s forest (Ahlgren and Ahlgren 1983, Simpson et al.1990, Mladenoff et
al. 1993, White and Mladenoff 1994). In the aftermath of clearcutting, extensive areas
(and consolidated areas) on which a secondary forest was developing were incorporated
into national forests (Hiawatha National Forest and Ottawa National Forest in the Upper
Peninsula). On lands where topography and soils permitted agriculture, early homesteads
and farms replaced the original forest. Large areas (several square kilometers) that had
escaped clearcutting are today protected from logging in natural areas and forest
preserves.

Forest clearcutting, because of its catastrophic scale, altered many ecological
processes of different spatial and temporal scales that were ongoing in the preexisting
forest. The structure of the secondary forest was changed from that in the primary forest.
In the secondary forest that developed following clearcutting, slow growing, and long-

lived tree species, white cedar, white pine, and hemlock were reduced from their former

32

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

bird densities
BTBW BTGR BLWA AMRE REV] LEFL SCTA VEER OVEN
1 F!
H
vegetation variables
Canopy Opening
3:31.: W... W...
Woody Stemsl H
Woody Stem /
Dectrees<15.2cm _____
Dec trees<30.48cm 7“ .
% Leaf Mc common .1. ' ' “ ' . b TRENARY
% Fern _; ”@8110“ 52: ..
% Woody Debris ' ’ '
- 4 I
SYLVANIA smNmeN
Primary Managed Settled
Forest Forest F 01‘0“
Landscapes Landscapes landscape“

 

 

 

Figure 2. Main features of conceptual model and study design showing stratifications
based on historical divergence, land use, land type associations, and time elapsed after
logging; and comparisons of bird densities and vegetation variables among strata.

33

First -
leienc

. \RE:
BLMI
BTBW!
BlC-lll
lfll L1
OVEN
R£\l R
SClA S
ER 1

TH

.1..:€ air
55‘
We:
P3 Legit
Pi Seton

lRlO re
P5 Older.

l. Bll'd d,-
alld i011

3 Bird {it
ill ICE

3 Veins}
are com
md 1021
4 Win;
and Sir
2‘l’l‘ar
atroSS l

‘1’ refers

Figure 2. (Continued)
Legend

E' l S .
AMRE American redstart

BLWA Blackburnian warbler

BTBW Black-throated blue warbler
BTGW Black-throated green warbler
LEFL Least flycatcher

OVEN Ovenbird

REVI Red-eyed vireo

SCTA Scarlet tanager

VEER Veery

I. l 1 E I . 1' 1
P1 Logged 1-5 years ago
P2 Logged 6-12 years ago
P3 Logged >12 years ago
P4 Second growth forest not yet entered
into recent Logging cycle
P5 Older-growth forest

Analittisalfilsns
1. Bird density is compared among landscapes
and logging periods in managed forest
2. Bird density is related to vegetation variables
by regression analysis
3.Vegetation variables that are statistically significant in 2
are compared among land use (primary forest, settled forest )
and logging periods in managed forest
4. Within managed forest, bird density
and significant vegetation variables are compared in
a 2-way comparison across time elapsed after logging and
across landuse strata

-r refers to strata comparisons in a specific analysis

34

. . CCVALQQZ~C V H: \.0>rd-~
) 32:. C 99.2.}! :.:§( . r -
>H....7. :H H7:=.u..—- >a:—— 2:; . \)L

H‘ fl ' U > VI. 7 U AU F-- ii. I. .AUF-hlvr U... I. V) 1' U V U! UA' I ~ IV. .A'-- I'. n.‘ \' h ‘A. -;-.§’m\' AV nrlvlvv§fidfl V? -~IV..IV~‘~‘V >éflv- 1‘ 0‘ V—‘ ‘

888v €sz 28 8889.88
83 2.888. 3 8:85:23 <

.888 comma: 8m 88883.
on 8.88 N can H 8:853 .8 888w

888w 83m 38 888888

83 888:2 3 8:83:88 <

8888 88m 28 888888
83 388:2 .3 838588 <

553 28 8:828 83
8202.8. H88.H 8 889 8:85:28. <

.88. 823688 2598— 2:
8 28388 88280 E moo—Sega HH

828.888 83.83% H28 838088
2:88: 2: 8 2829.8 H8828».
8038: 828828 E magenta Hm

828888 HEB—3226
EocobHe 8:3 flown—88 8383
808b% H2: 2 Hoobo 03688 < AN

828888 2528— 88253
82:88 188280 8 808th5 AH

.88. 3888— 2: 8 28828
8828 H8828 E 8088an AH

883886
8. 028. 828888 H28 2.8
.82 + 02.3888 ozaobmfiao 83 HH

.823
8.38 855 89888_ =8>$om - 0

than: 888 can 8&2. 883 82m
292 Be $5.: $8.68 88...an
A285 8:888 8:852 888
8888 853» 88888 838m - m

.882er H28 8.2880on
£82882 88: 888.88 828
888 8223 88088.2 w88< .. <

828.8: H88H m 855 A 828888
25 .28: 888882 w88< AN

.828 8.23.
H28 828 H8888 .2888 58:8
motoHooweH H888 880.58 coozzom -H

 

83m 8 880:. 26:

£8; Ho bosom 82

8888.80 .8 82.5

 

8888 .83» 2: 98 888888 083 888 8368: E258: 2: 8 3:38? .8 8088 808b% 303

.N 033.

35

"Il ‘III‘.I -

~III\- -

>HH:£H:,:~.) .‘C 9.22.37. :.:.€<

tavzttnbtgvrv \C ~b>0§

A~vrv-=wu=avVViI -- . N1 .Uvnsluslkl

 

.8 :000 .800 0 :.....3 80.8.0 :02. 00..
>000. .88.. 00m0:08 0. .83 .8 .00.).
.:0..00....0..0 0 an. 00.0.0000 0.03 .88..
...30.w-.00.0 0:0 .88.. ...30.m 0:000m

.2030
0.:. 00:w.80 003 8.80. .3 00:08.0
0080 00000.0 08.. 8.. :0..00c..0..0 <

82.0.0 888 0.... 0000... 2.8008. .3
000.8008. 0. :0..0..0> 0.000 ..080 0... ...

800.8088 .08. .0..m... 0 8..
000.8008. 0.03 0.00.00 0...... 00.00.88
.0: 0.03 .08. 0.... .0 800000800

.:0..0.00000 2.28:0.
0 :.....3 02.0 808 ..0 8.8.0288 2..
000.8008. 00800 0800 0. 0. 80.00 0......

0.389 58.5. .0

.mem MC OwNHm Hus—Cm Cm WOOCOHOEmQ AN

.0080. .00. 0080 00000.0 08. ... A.

8880.0
..00 20008000.. 80:08.80 0..m AN

830:... .0: 0. 8:00.320
.00. 0080 00000.0 08.. 0... .88.. 8.08.80
8. 08:20.00 0.00.0 .0... 00.000 80.0.0.0
.0 0808 00:00.30... 0..00..00.m A.

. :0..0.00000 003000.
0... 30.0.. 8.000 .0 8080858 0:0.
:. 80:08.80 .8 0.00.00 03.0.0800 AN

2.00008. .88..
08808 0:0 :.....3 02.0 :003.0m-m

2.00008.
.88.. 808.... 0 :.....3 8..0 :003.0m - .0.

2.00005. 0:0 :.....3 02.0 .88..
00030.0... 80....0: 80.0.0.0 8084 -m

.88..
02.80 :.....3 8000008.. :83.0m. -0

 

>005 :. 00.00.... 300.

.....£E> .o 8.80 :02

80.80800 .8 .08..

 

.3288. .~ 2...:

36

.11...

-\I1 Il‘I.I.-.

\I“‘~ -

>...:......:...\) k: 8.0.237. 2.2;).

..Cv..:.2.x:.c..b k.» T0) .0 . N

«\wb~.:.~..:v..vv N

 

.93... 0.

 

00.00... 0.... .8 :0..00.0.
8.80.888 .80....0 80.3 80.8..
.0800 0.... 0... .0 00.0808. 0.0.00...
:0..0.0m0> :. :0..0..0> 0808.00. 0... ..

800.8088 .08.
.0..w... 0 8.. 0:0...0:8 0.00....0. 0. .0...
808 8.00.08 .3 000.8008. 0.03 N
0:0 . 000.000 8 0.00.8 0...... 00.00.88
.0: 0.03 .08. 0.... .0 80080800

8.00...
00.. :. :0..0..0> 0:0 .000w 80:00 .8 00..
0:0 00:08... :. :0..0..0> .000. 8.80."...
A00000:. 880.. 0:0 .8085 00:00.0...0
0.00. =08. 0:0 0.8. 0:0 08:88
5.008000. :. :0..0..0> .0001. A.

0:0. 08>... 8.880..
.0 .88.. .8000: 08. 8.8.0088.
8 08.. .0 0... .8 0.0.0 .008. Am

8.0.8..
0.....0:30 0:0. .3 00.00....0 0.. 0.08 8...
8080 0:0 0... 0 :.....3 080.88.. A.

80.8.. .0800 0.... 0.0808... w:08<- 0

00000000.
.88.. 00.80 . 0.....3 .0... :003.0m- 0

 

>005 :. 00.8.... 30:

0.08; .o 85o. :02

000000800 .8 .08.

 

.8288. .0 00...

37

m

extent both as forest types or as components within forest types (Mladenoff et al. 1993,
White and Mladenofl‘ 1994, Mladenofl‘ and Steams 1993, Crow 1995, Davis et al. 1995,
Mladenoff 1995). Thus for bird species associated with a conifer component, habitat
availability at the territory level, and habitat continuity at landscape levels, would have
been impacted by such changes in tree species composition.

With the interspersion of human settlements and agricultural areas, and
management of early successional forest, forest cover was also reduced in some
landscapes. In addition human related activities introduced edge into areas that were
devoid of it, in the form of induced interspersion of forest openings, early successional
timber types that are periodically clearcut, pine plantations, roads and logging

trails (Krummel et al. 1987, Robinson 1988). Because land use has had a dominant
influence on the spatial configuration of forest cover at landscape scales and on forest
composition and structure at finer scales, land use would be expected to account for a

large amount of forest heterogeneity”.

Hierm'chy of physical and climatic factors qffecting heterogeneity.

Regardless of anthropogenic influence, an underlying heterogeneity exists in the
forest that can be attributed to multiple scales of physical and climatic factors (Allen and
Star 1982, Turner 1987, Milne 1991). This multiple scaled heterogeneity forms the basis
of an ecological classification system that has been adopted by the US. Forest Service
(and recently by many other land management agencies) to classify land into a hierarchy of
land units. Ecological classification systems are based on relationships between

heterogeneity at large scales (regional climate) and heterogeneity at progressively smaller

38

E.

1”.

Sen

scales (landforrn, local climate; forest types) that are imbricated within larger units.
Ecological classification systems reflect natural hierarchies of temporal and spatial
dimensions of ecosystem processes unique to specified geographical locations. As such
they can only deal with potential vegetation cover and do not incorporate well the
patchiness imposed by human induced disturbance.

At the level of the national forest, the largest unit in the hierarchy of units is the
land type association. The spatial extent of a land type association is in the range of
hundreds to thousands of hectares (Table A1). The delineation of land units at the level of
land type association emphasizes the effects of local climate and geomorphology.
Different forest types are comprised within a land type association. Difi'erent
microhabitats in turn are incorporated within each forest type. The configuration of the
northern hardwood. forest with respect to other forest types differs in different land type
associations. Processes of forest maturation and disturbance, their intensity, frequency
and the patchiness they produce are likely to also differ in the different land type
associations since these are determined by higher level environmental factors. The
difference in the configuration of the northern hardwood forest, in microhabitat
conditions, and in modifications imposed by human induced disturbance that songbirds
encounter in the different land type associations could be of a magnitude large enough to

affect individual species densities, productivity, and the relative abundance of bird species.

Incorporating human-induced and environmental heterogeneity into model
In the Upper Peninsula, land use categories of primary forest, managed forest and

settled forest can be delineated by large patches that cover several square kilometers. This

39

| l
“fien.

A‘L‘

'

Li

gin

results from the perhaps relatively unique situation” (compared to a majority of eastern
deciduous forest regions), in which large contiguous areas had been incorporated under
national forest management, large tracts of remnant primary forest exist, and human
settlements and agriculture are mostly concentrated where topography and soil
productivity can support agriculture. Land use categories of primary forest, managed
forest and settled forest were incorporated into the northern hardwood ecosystem model
(Table 2, Figure 2) at the highest level in the hierarchy of factors affecting northern
hardwood forest heterogeneity in the Upper Peninsula". Each land use category was
replicated by 3 different land type associations" (Fig. 2).

All 3 land type associations replicating the land use category of primary forest
(Huron Mountains, Sylvania Recreation Area and McCormick Wilderness Area) are forest
reserves. In primary forest not subjected to logging, most small scale forest disturbances
such as tree fall result from natural processes”. The frequency and spatial extent of small
scale disturbance are likely to be influenced by local topography, soil and climate but the
actual events within a specified area and time period are stochastic. The emerging pattern
of these small scale disturbances are likely to be reflected in differences in vegetation
structure measured at small scales among the 3 primary forest areas. The objective of
including difl‘erent land type associations was to capture a broader range of the variability
existing in the northern hardwood forest. The model was not designed to elaborate on the
specifics of the different areas. Such an objective is better relegated to an ecological land
classification or to the development of specific management objectives for these areas.
Comparisons of bird densities and of microhabitats among the 3 land type associations

however would test model assumptions concerning the effects of land type associations in

40

.1
(5;:

m

f 7"
r\

Mg

ha
Reg
lI-l

SIR] g

cum

general on bird species densities and their habitats. These comparisons would address
whether the land type association level should be considered as a level of stratification for
bird conservation or for firrther bird research (Table 2).

Managed forest was also replicated by 3 different land type associations (Munising
Moraine, Manistique section of the Newberry Moraine, and an interior sand ridge -swale
complex in the Rapid River District). Although the spatial configuration of forest types is
affected by topography, slope, aspect and soils, in managed forest it is also modified by
forest practices such as management for early successional forest and pine plantations. In
national forest, logging is a predominant forest disturbance factor for much of the northern
hardwood areas. Management guidelines for northern hardwood forest in the Great Lakes
Region specify a selective cut every 8-10 years (Ohmann 1979, Tubbs 1977), a cycle of
12-15 years is given in the Hiawatha National Forest Management Plan (1986). The
structure of a northern hardwood forest after logging is related to a large extent on the
time elapsed since that disturbance, and on ecological conditions. Periodic selective
cutting in northern hardwood forest interferes with forest maturation and disturbance
processes. Stochastic events such as severe storms and tornadoes would still occur but
their efl‘ects could be confounded, influenced or overshadowed by logging disturbance.
Another source of variability within the northern hardwood forest under national forest
management can be attributed to the initial patchiness of the northern hardwood forest
when it was entered under national forest management. Although the majority of the area
of the northern hardwood forest had been clearcut around the turn of this century
(Cunningham and White 1941), some areas were not completely logged and these areas

currently represent older growth within the national forest. Other areas that had been

41

‘Y‘“

L7

Kl

y...
(I)

but difi-

Mach}

clearcut had not yet been entered into the logging cycle. The pattern of small scale
patchiness is likely to differ based on the time elapsed after logging”. To account for this
variability, the northern hardwood forest within managed forest was stratified into 5
classes representing forest that was logged 1-5, 6—12, and >12 years ago, second growth
forest not logged since the initial cut at the turn of the century, and older-growth forest19
in areas which had not been completely clearcut and therefore contain at least some
features of the original forest (e.g. lack of white pine or hemlock stumps, presence of large
sugar maple and hemlock trees).

The landscapes on the fiinges of the National Forest where human settlement
intersperse forest lands are subjected to both natural and anthropogenic disturbance (e.g.
logging, road building). Because of the multitude of ownerships, no systematic form of
human-related disturbance can be identified at the landscape scale. Within these
landscapes the northern hardwood forest was not stratified any firrther. Similar to the
situation in the primary forest category, replication would recognize difl‘erences between

these areas on the basis of local environmental conditions and randomize their effects. .

3. Selection of Bird Species to Include in the Model and Representation of their Habitat
The forest bird community is multidimensional in that it is composed of many
different species that would respond differently to changes in the forest. To examine a
contrast in bird species response to difl‘erent ecosystem processes, it was necessary to
simultaneously consider a number of bird species that share the northern hardwood forest,
but differ in theirindividual ecologies. Ten bird species were selected for study: ovenbird,

black-throated blue warbler, American redstart, chestnut-sided warbler, blackburnian

42

l' 63:11::

1153: iii

leez'x (

warbler, black-throated green warbler, least flycatcher, red-eyed vireo, veery and scarlet
tanager (scientific names are given in Table Bl). These bird species are generally
associated with the northern hardwood forest (Brewer et al. 199]). As a group they
require different vegetational components: the black-throated blue and American redstart
nest in the shrub layer, the scarlet tanager nests at mid-canopy level, the ovenbird and
veery on the ground, and the red-eyed vireo at various heights in the subcanopy and
canopy. The blackburnian and black-throated green warblers are associated with a
coniferous forest component. These species differ with respect to their area sensitivity
and to the degree to which they are habitat specialists associated with deciduous and
deciduous-coniferous interface. Many have been linked to large tracts of forest (Bond
1957, Ambuel and Temple 1983, Lynch and Whigham 1984, Robbins et al.1989b). Area
sensitive species were purposely included to allow examination of some forest landscape
level effects such as the reduction of forest cover, and the increased amount of forest
edge. These aspects have been associated in general with forest fragmentation but have
not been addressed within largely forested regions (Ambuel and Temple 1982, 1983;
Freemark and Merriam 1986).

It was also desirable to have species with geographical ranges that extend well
beyond the geographical area spanned by the study sites. A difference in the locations of
study areas with respect to the geographical distribution of a species would afi‘ect densities
of a species among different study units. Consideration of geographical distribution was
important to remove a possible confounding effect of geographical distribution with the
response of the species to habitat conditions caused by forest disturbance. Current

distributions of a number of these species, the blackburnian, black-throated green and the

43

 

m'dliooc.

Cent; '25(
“me: he
menSlXe
“CERT”.
redSim is
the 10M
simulates
{dame d1

You] hO‘

4. $61ij

I
and We
landmp

‘
32

1 Poin

black-throated blue warbler, correspond to the geographical range of the northern
hardwood-hemlock forest. This correspondence could indicate a closer link between the
ecologies of these species and the northern hardwood forest ecosystem. The black-
throated blue warbler and the American redstart were included because of similarities and
contrasts of their individual ecologies. The distributional range of the black-throated blue
warbler has shrunk in southern Michigan, a change that has been attributed to loss of
extensive forest (Binford 1991). The American redstart however is well distributed across
Michigan and eastern North America (Peterson 1980, Reinoehl 1991). The American
redstart is more tolerant of forest edges and might benefit from them. Both species nest in
the lower understory vegetation and likely respond positively to disturbance that
stimulates dense undergrowth. They were included because of the possibility that their
relative distribution within the range of variability of the northern hardwood forest might

reveal how forest changes might difl‘erently affect ecologically overlapping species.

4. Sampling Units and their Groupings

The basic sampling units considered were point locations at which bird censusing
and vegetation sampling were undertaken. Variables representing characteristics of the
landscape within a 1.6 km were also detemiined for these point sampling units. A total of
321 points were located in the northern hardwood forest following the stratification
considered in the ecosystem model (Figure 2, Table 2). The distribution of sampling units
into the various strata considered in the model is given in Table 3. An initial stratification
of 321 sampling units into model strata permitted comparisons among categories of land

use, land type association, and the various states of the northern hardwood forest after

44

Z...- Z.

a... ~....~..UCme.fl 090?... 3.3... 3:... 003.:an
99:91.9).3 22:52.:— Z: 3.4%.... -:.~:... a... AZ...V.:=~?. I~...m:av..u uv....£v [0:0.—Xx Trisha—:03..-th -Avm~30=L~2mfiN

.M. .U‘CRN.

.8000... .0800 0...: ...00 8..080. .0 .3800 n 0.

 

c. o. o.
NV”:
0. m. 0.

00.0..
ovn0
“008.000.. .muuam .008... m

0 o . . v.

0 cm 0 m.

N .N 0N . .

0 an 0. N.

N v. 0. N.
0 .n0 8H0 008 SHE

$33387

.0320 .020
032. 0:. 000.000
08 .0. SA 0080.

08 .0. N70 0080.
08 .0. m. 0080.

 

.80..0.00..0 00... 000. 000 00..00
00000.0...0 ..0.0.. 0.0 000. 0. 0008.020 .0..0..... 00 00.00 0.0... 0. .80..0.. .0800 0.... $0.00 80080. .0 003005.... .m 0.00...

45

 

dimrm

bird dens

5, Link

disturbance by logging, with respect to vegetation characteristics of small area patches and
bird densities. Sampling units were also poststratified into new groups that were based on
specific ranges of values for vegetation characteristics (i.e. density of woody stems and
canopy opening) measured at the bird census stations and landscape level characteristics
(i.e. % forest cover, length of forest edge). Such poststratification permitted the
comparison of differences in the density patterns among the black-throated blue warbler,
American redstart and chestnut-sided warbler along a forest fragmentation dimension at

spatial scales that are relevant to their interactions.

5. Linking Habitat Selection at the Microhabitat Level to Ecosystem Processes
Vegetation structure and composition are both multidimensional, composed of
many interrelated components. The importance of vegetation structure to birds has been
well established in the current ornithological and wildlife management literature (James
1971, Noon et al.1979, Robbins et al. 1989b). Ecosystem processes and human related
disturbance can be linked to bird species ecologies by assessing how they affect habitat
features at the spatial scales bird species respond to. It was important to identify
vegetation features that each bird species would most likely respond to and to test how
these features difi‘ered among the difl‘erent levels of stratification incorporated in the
ecosystem model (Figure 2). Vegetation variables were selected to describe forest
canopy, tree species composition and structure, shrub layer and ground layer structures”.
Bird densities were related to vegetation variables through regression analysis. Figure 2

shows how the density of each individual bird species, and each vegetation variable

46

61:

91 A?
Md
lien:
31 in
1B3:
1982
Ram
50m:
Spec.
toler

sped

important to that bird species would be compared across the diflerent strata of variation in

the northern hardwood forest.

6. Incorporating Landscape Scale Efl‘ects on Bird Species

There is increasing theoretical basis linking population dynamics of bird species to
habitat continuity at the landscape level (Pulliam 1988, Kotliar and Wiens 1990, Opdam
1991, Freemark et al. 1992, 1995). Many human related activities reduce habitat
continuity, a process that has been referred to as habitat fragmentation. At this time
vagueness still surrounds concepts of what constitutes fragmentation, its quantification,
and its progressive nature as a process (see Godron and Forrnan 1983, Haila 1986,
Freemark and Merriam 1986, Krummel et al. 1987, Kotliar and Wiens 1990, Hansen and
Urban 1992, Harris and Silva-Lopez 1992, Frankfin 1993, Freemark et al. 1995, Faaborg
et al.1995). Edge has been associated with many biotic processes such as increased
predation, brood parasitism by the brown-headed cowbird and interactions with habitat
generalist species. Perhaps it is because most studies have addressed forest fiagmentation
at its end stages, that predation and brood parasitism have received more attention
(Brittingham and Temple 1983, Wilcove 1985, Yahner and Scott 1988, Small and Hunter
1988, Donovan et al. 1995, Pearson et al. 1996). One study however, Rosenburg and
Raphael (1986) found that in the recently disturbed primary forest of Northern California,
some bird species completely avoided edges while others utilized them. This indicates that
species of a bird community previously undisturbed by human interference have different
tolerance to edge. At advanced stages of forest fragmentation, the most area sensitive

species would be expected to have completely dropped out. It is in largely forested

47

r *0!”
.35....
v.

depic

9‘4an

regions, where the assembly of area-sensitive forest birds is more complete, that a true
assessment of the relative sensitivity of different forest bird species to edge can be made.
The possibility that the introduction of edge by human activities into largely contiguous
forest areas would provide suitable habitat for edge-interior bird species at the expense of
highly specialized, area- sensitive forest bird species, was the justification for exploring the
effects of increased edge in the northern hardwood forest. The theoretical indication that
habitat continuity and edge affect bird species warranted exploring the nature of these
landscape level effects. This exploration might lead to a better conceptualization of the
process of habitat fiagmentation in a highly forested region.

Landscape level effects were incorporated by considering forest edge21 within
500m-radius areas, and total canopy cover and northern hardwood forest within a 1.6 km
radius circular areas centered at bird census stations. Thus to each bird census station
corresponded variables that quantified the environment at different scales. Figure 3
depicts the spatial relationship between the area scale at which vegetation structure was

quantified and area scales at which landscape level aspects were quantified.

7. Integrating Microhabitat and Landscape Level Efl‘ects

Bird species could respond to landscape level conditions (areas within 500m and
1.6 km radius), and to microhabitat conditions which were measured at the locality of the
bird census station. Sampling units (bird census stations) were poststratified into new
strata defined by ranges of values for the selected microhabitat and landscape level
variables. Figure 4 represents the poststratification of sampling units initially grouped by

land use category, forest landscape and period elapsed since logging, into strata .

48

 

 
    
   
   
    
   
     
   
 
 
    
     

  

Bird Census Station
Bird Density Estimates
4 10 x 30 vegetation plots

Estimates of campy, shrub layer
and ground attributes

Estimates of amount
of edge between
northem hardwood
forest and nonforest

  

1.6 km radius circle
Estimates of total forest

cover and % composition
of forest types.

 

 

 

Figure 3. Spatial scales considered in measuring environmental variables associated with a
bird census station.

49

Figure 4. Schematic representation of poststratification of sampling units fi'om a
classification based on land use, land type association, and forest disturbance period to a
classification based on levels of % forest cover at the landscape scale, amount of edge
within 500m radius and number of deciduous woody stems in the vicinity of the bird
census station. This poststratifrcation was used on vegetation data and forest cover data
collected in 1992—1994 in the Upper Peninsula of Michigan.

50

l .und use

l-and use

.-... Aa-lt‘t‘I1I'In I‘

prim (I ry forest

nm nagcd (‘2 was t

$80? 058 02o...

..\.00A 0260 “.0000

 

0.85%? 0:8 use.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8633 o. as;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 0 00.0
002005.950.—
28. 2.. 00000
08800.0 02...
. . .H. .. Y 08. E 3.50
.. .5. . 30m
/ \ ./
EOE abidi— flogug gas— :Oflnmg gas“ a 33082 £33
0.000.. .0800 «.80.. 009.008
0.0 000.

83 fig

51

 

represent:
The posts
sea-es m
core: a".
census 9
and ill te
of the 32
he 2 5;
L56 here
be relay
meaSUri
Conifer.
efiei‘ts
misc:
M de
home,

“Dim

representing a range of microhabitat conditions and a range of landscape level variables.
The poststratification of sampling units based on their characteristics of vegetation at small
scales (measured around the bird census station) and on their characteristics of forest
cover and amount of forest edge (measured at 1.6 km radius and 500m radius from bird
census stations) is equivalent to reducing the heterogeneity of the forest to spatial scales
and in terms of variables that are relevant to bird population processes. Poststratification
of the sampling units based on a number of variables that relate to fragmentation effects at
the 2 spatial scales (vicinity of territory and within 1.6 km radius ) permit the viewing of
the heterogeneity of the forest along a forest fragmentation dimension. Variables that can
be related to fragmentation include number of large canopy gaps and % canopy opening
measured around the bird census station”, and length of forest edge, % forest cover, %
coniferous forest, % deciduous forest measured at 1.6 km radius landscape level. If the
effects of forest fragmentation at the vicinity of the territory and those in a broader
landscape were linear and additive then multiple regression could be used to show how
bird density is affected by fiagmentation aspects at the 2 scales. Bird ecology studies
however of competitive interactions and coexistence mechanisms point to the need to
explore the interactions among variables measured at the 2 scales since it is likely that the
effect of fragmentation at the larger scale would affect community structure. If there is a
difference in the relative abundance of bird species in difi‘erent landscapes, bird census
stations fi'om difl‘erent landscapes would be exposed to different bird species population
abundance from the broader landscape even if they had similar characteristics at the
breeding territory level. Thus the sampling units would need to be grouped based on

Values of fiagmentation related variables at the 2 scales. Both the range of values of the

52

In:

Ya:

no

Be.

$8

“0

fiagmentation variables and the total number of sampling units placed a limit on how fine
of a gradation to consider for each variable and therefore on the number of new strata.
Two- way analysis of variance was used to elucidate the interactions between
fragmentation related variables in the vicinity of the territory and those in the larger area.
In effect, with the 3- way and 4- way analyses of variance that were performed using
variables fi'om the 2 scales”, the total range of variability of the forest captured in the
model was reassessed in terms of fragmentation related variables at new spatial scales.
Because the whole range of northern hardwood forest was considered in these
poststratifications, the original association of sampling units with land use and landtype

association was lost“.

FIELD METHODS

l-Selection of Study Units
Selection of Primary Forest, Managed Forest and Settled Forest Landscapes.

The gathering of field data was undertaken from May-August in each of 1992,
1993 and 1994. An endeavor to locate all possibly suitable study sites was attempted in
1992. The land type association was the largest land unit, within the hierarchy of spatial
scales considered in the model that needed to be selected. The focus of the study on
northern hardwood forest ecosystems required that this forest type be a dominant forest
cover type within the selected land type associations representing a category of land use.
In the settled forest category, where some forest cover was lost to agricultural use, it was
required that the northern hardwood forest would have historically been a dominant cover

type. This requirement was easily met because of the preponderance of northern

53

(I?

hardwood forest. The selection of land type associations in the land use category of
primary forest was dictated by the availability of large tracts of primary forest that
comprised northern hardwood forest. The Huron Mountains, McCormick Wilderness
Area and Sylvania were selected since these were among the few remaining primary forest
landscapes in the Upper Great Lakes region. Delineations and descriptions of land type
associations on the West Unit of the Hiawatha were used to select land type associations
comprising northern hardwood forest in managed forest and in settled forest (Table A1).
The suitability of hardwood forest soils for agriculture accounted for the patchiness of the

northern hardwood forest in settled forest landscapes.

Locating northern nardwood forest sites in primary forest, managed forest and settled
forest landscapes.

Locations of northern hardwood forest within a land type association were
identified from US. Forest Service data bases (CDS), from aerial photos and fiom land
cover maps resulting from an ecological classification of the land base. Forest cover
typing in managed forest was available from compartment maps, and fi'om a pilot GIS
project for parts of the Hiawatha National Forest. Forest cover maps for the Huron
Mountains and for the McCormick Wilderness Area were produced during past studies in
those 2 areas that involved an ecological land classification (Pregitzer and Barnes 1984;
Simpson et al. 1989). These vegetation cover maps were available from the Ottawa
National Forest and from the Huron Mountain Club respectively.

The forest types and their delineations in managed forest are the result of many

decades of silvicultural exams that had emphasized assessments of forest resources. There

54

are iii
resin:
exams
in the
empire
tier:
reed:
ecolo;
tree 3;
Differ
and 1h
(com;

heme

are differences in the forest typing produced by silvicultural examinations and that
resulting from ecological studies. The recognized forest types resulting from silvicultural
examinations emphasize finer differences in commercial tree species to reflect a difl‘erence
in the suitability of different areas for timber management. An ecological study in contrast
emphasizes subtle differences in tree species composition when these reflect ecological
differences such as those due to soil variations, slope and aspect; variations that oflen are
recognized by ground flora (Pregitzer and Barnes 1982, 1984). Forest typing in
ecological studies included finer divisions of nonforest cover types and noncommercial
tree species associations, than those recognized by US. Forest Service forest types.
Differences between forest cover types recognized in forest cover maps of the McCormick
and the Huron Mountains, and those recognized in forest cover maps of managed forest
(compartment maps), also reflect differences in the patchiness of hemlock and white pine
between primary forest and secondary growth forest. Table C1 matches the forest types
of both systems.

In primary forest landscapes, forest cover types along a gradient of hemlock -
northern hardwoods include the cover types of hemlock, hemlock dominated hemlock-
northern hardwoods, hardwood dominated hemlock-northem hardwoods and sugar
maple-hardwoods. For the McCormick Wilderness Area, that range was fiirther divided
into finer differences based on ground vegetation. Although in the presettlement forest,
natural forest succession of northern hardwoods was to hemlock, I did not include forest
cover typed as "hemlock" in my sampling. Because of time and budget limitations, the
main focus was on the managed northern hardwood forest. Primary forest landscapes and

settled forest landscapes were included only to put the managed forest in the perspective

55

of less and 1
areas that a
pnman for
disturbed i
ha'dn 00d
corer ma;
and Ham:
Sites \0
(I992 lea
norr‘rrem
Wm r
equVale

and 05)

broken
Units Cal
WQH def

Which Ill

of less and more disturbed forest. For this purpose, considering northern hardwood forest
areas that are hardwood dominated hemlock-northern hardwood, a type prevalent in
primary forest landscapes would suflice to extend the managed forest towards the less
disturbed forest end. All types that could be included in the broader categories of
hardwood dominated hemlock-northern hardwoods, and sugar maple-hardwoods, in the
cover maps compiled by Simpson et al. (1989) for the Huron Mountains, and by Pregitzer
and Barnes (1984) for the McCormick Wilderness Area, defined the availability of suitable
sites. No cover type map was available for Sylvania at the time of the study. Aerial photos
(1992 leafoff, infrared scale 1: 15,840) were used to delineate areas that were hemlock-
northern hardwood forest dominated by hardwoods. Areas of heavy eastern hemlock or
eastern hemlock-white pine cover were avoided since these were considered to be the
equivalent of hemlock or hemlock-white pine forest types (U. S. Forest Service types 04
and 05).

In all primary forest landscapes, accessibility due to lack of roads had to be
considered in the selection of suitable sites. Only those northern hardwood areas that
could be accessed within 1 hour of travel whether on foot or by motor vehicle were
included. In the Huron Mountains few areas of hardwood dominated, hemlock-northem
hardwood forest could be reached within an hour’s travel.

All lands on national forest lands that exclude wilderness and recreation areas are
broken down into management units called stands which are comprised within larger land
units called compartments. As a management unit, a stand is not exactly congruent to any
well defined ecological factor”. This is in contrast to ecological classification systems in

which the ecological factors that difi‘erentiate between land units at a particular level are

56

thr

defined (USDA Forest Service—National Hierarchical Framework of Ecological Units)2".
An ecological classification of land units presents a consistent basis of ecological
heterogeneity, and the pattern of lower units occurrence is repeatable within a higher
unit”.

The US. Forest Service data base ”CDS" was used to locate northern hardwood
forest in the 5 categories of time elapsed after disturbance within the 3 different land type
associations in managed forest. In the data base, management units are identified by
compartment and stand number. Descriptive information on a stand in general consists of
forest type, age-class, silvicultural treatment, year of silvicultural treatment, basal area,
average diameter of trees, survey‘year, and for some stands the year of origin”. Northern
hardwood forest types are included in the range of US. Forest Service forest types 81-89
(all forest cover types used on the Hiawatha National Forest are given in Table Cl).
Within each of the selected managed forest landscapes, a first cut for suitable locations
was all the forest types in the range 81- 89 in US. Forest Service classification.

In managed forest, the northern hardwood forest was further stratified based on
when it had been last disturbed by logging. More complete information in the recent
computerized database (CDS) was available for forest management activities undertaken
within the previous 15 years (from the beginning of field sampling in 1992). For the 3
categories of forest logged 1-5 years ago, logged 6-12 years ago and logged >12-15 years
ago, the number of locations in each of the 3 land type associations used to replicate
managed forest was well defined”.

In the data base there was a category of northern hardwood forest areas for which

there was no record of logging. This category actually represented different conditions:

57

areas that were logged prior to 15 years ago (fi'om the beginning of field sampling in
1992), unlogged second growth, and older grth areas”. Mean average diameter at
breast height (DBH) and stand origin (when it was available) were used as a guide to
locate second growth northern hardwood forest, and older grth forest. Information
fi'om old compartment folders was used to further identify areas that were logged prior to
1979, and eliminate these from unlogged second growth, and potentially older growth
forest. Because of the rarity of old-growth forest outside of research natural areas and
wilderness areas, the knowledge of US. Forest Service personnel was also relied upon to
suggest forest areas showing old-grth characteristics. Size of trees, an uneven canopy,
and canopy gaps, lack of tree stumps, and tree species composition", as revealed by field
checks further refined the separation of older grth fi'om second growth forest.

Because the placement of timber sales across the forest is not random, the
location of northern hardwood forest patches in the different classes of time elapsed since
logging 1-5, 6-12 and >12-15 years would not be random. In many instances logging also
did not involve the whole stand unit. Patches that were logged were identified from timber
sale contractor files and from field checks. A large enough area is required to be
representative of forest interior conditions and of conditions of northern hardwood forest
in the different stages of time elapsed after disturbance. Recommendations by Morrison et
al. (1981) that a 20 ha area is the smallest area that could be considered continuous habitat
were followed in the selection of suitable northern hardwood forest locations

In settled forest landscapes the northern hardwood forest had become patchy and
distinct from the surrounding matrix. Potential suitable patches of 20 ha or larger were

rare but easily located from aerial photos. After giving consideration to the shape of the

58

patch and getting permission fiom private land owners few patches were available. In
some land type associations (Stonington), all available patches were used for the

placement of points in settled forest landscapes.

2- Locating Bird Census Stations

The model structure specified 21 strata of northern hardwood forest needed to be
sampled which included 3 land type associations in each of primary forest and settled
forest, and 5 strata of forest disturbance in each of 3 land type associations within
managed forest (Figure 2)”. A larger number of sampling units would be required to test
for differences in bird densities and vegetation variables among strata of the northern
hardwood forest than among strata representing different forest types 33. In a previous
bird study, a number of sampling units in the range of 10-15 was sufficient to show
differences between habitat occupied by each of 10 bird species and random points in the
northern hardwood forest (Hamady 1983). In the absence of specific apriori data, this
latter study was used as a rough approximation of the number of bird census stations that
would be required to test for differences in bird species mean densities and vegetation
variable means among northern hardwood forest strata (among land type associations and
forest disturbance periods). Because bird density and vegetation variables were sampled at
a spatial scale close to the territory level, each bird census station was an independent
sampling unit and not a replicate of bird density or vegetation conditions of a larger area.
It was important not to sample the same individual birds from 2 adjacent bird census
stations. Because the bird census methodology used (Distance Sampling) estimated the

number of birds fiom their singing and calling, a minimum distance of 150m was

59

required be
ones llms
rhar can ix
possible rr
available
consrrarrr'
ola 100
the minir

reduced

Mime”
Strarum

bird Q31

required between bird census stations so that birds could not be heard from 2 adjacent
ones. Thus although theoretically the sampling space would be the number of territories
that can be fitted within the space available, technically the maximum number of samples
possible would be the number of 150m radius areas than can be fitted into the space
available. In order to evaluate habitat conditions in the interior of the forest another
constraint on the placement of bird census stations was that they be placed at a minimum
of a 100 m from any forest edge. The distance between adjacent stations greatly exceeded
the minimum of 150m because of irregular shape and avoidance of edge, which fiirther
reduced the number of possible samples.

The number of bird census stations that can possibly be fitted within a strata of
northern hardwood forest was determined to a large extent by the total area in that
stratum that was 100m away from a forest edge. A completely random design for placing
bird census stations would subdivide the area available in each category of northern
hardwood forest into units that had a 150m radius and then would select 15 points at
random. For most strata of time elapsed after logging within a land type association, the
number of locations available were less than the number of bird census stations required
(Table B3). Thus a random selection of points would result in several bird census stations
being placed in one location of a category of forest. The methodology of locating points
within the strata of northern hardwood was modified from a completely random design in
order to accommodate logistic constraints. Several bird census stations were located in
one patch depending on the size of the patch. The successive order of locations to be
considered for placement of bird census stations was determined from a randomly ordered

list of available locations prepared for each class of time elapsed after logging. For

60

I
Fr

Lo)
1

5P:

northern hardwood forest logged <15 years ago, this list was well defined. New locations
were not considered if the required number of bird census stations had already been
assigned. For other less well defined classes (unlogged second-growth forest and older
growth), new entries were placed on the list if the locations already on it were not found,
following a field check, to be representative of the class of forest disturbance period. The
actual number of bird census stations placed in a location depended on the size of the
patch and the length and spatial distribution of forest edges that needed to be avoided. A
total of 321 bird census stations were placed in a total of 78 patches.“

Layouts of bird census stations were drawn on US. Forest Service compartment
maps when available, or on aerial photos prior to field location. Fifteen different
individuals were involved in locating bird census stations on the ground. This reduced the
potential for bias in the placement of bird census stations. Edges avoided were those
created by roads that are maintained as system roads (Table C2), by management created
openings, clearcut stands where the forest canopy had not recovered, stands with canopy
openings >50%, streams, and lakes. Edges around canopy gaps that were the result of
tree blow down, or logging were not considered. Logging skid trails and 2-track roads
within the stand that are not maintained after a logging operation were not considered in
the quantification of edges. Their effects on canopy opening was evaluated in the

quantification of canopy gaps.

3- Bird Census Methods
Distance sampling (Buckland et al. 1993) was used to estimate densities of bird

species. In distance sampling, distances are measured from randomly placed points to

61

ohm
den;
sam
“Cl

rhes

deer

micro
aflbcr
“35C:

duerc

objects. Distance sampling uses these measured distances to make unbiased estimates of
density, given that certain assumptions are met. Underlying the theory of distance
sampling is a detection firnction g(y) = prob {detectionldistance y}. Bird census points
were considered the point transects in distance sampling. Distances were estimated from
these randomly placed points to singing or calling birds. Detectability of birds decreases
as a function of distance from the census point. Distance sampling fits curves of known
firnctions to the distances. The theory allows for the estimation of bird densities without
requiring that all individuals be detected or that the size of the sample area be defined.
These 2 aspects of distance sampling are advantageous in multispecies studies when
detectability differs among species.

The objective of estimating bird densities was to assess how birds respond to
microhabitat and landscape differences in different forest trajectories. Many factors can
affect bird species censusing and introduce noise into the data (Ralph and Scott 1981). It
was crucial therefore to reduce the amount of variability in bird density estimates that is
due to other sources than the intended comparisons. The methodology followed was that
recommended for point sampling. A number of measures were taken to reduce undesired
effects that include observer bias, weather, sampling bias and time of day. Each bird
census station was censused 4 times during a field season. A few bird census stations
were censused only 3 times because of a combination of logistic problems and bad census
conditions. Censuses were scheduled at intervals of 10 days beginning the first week in
June. Censuses began at 630 AM Daylight Saving Time and ended by 10 AM Daylight
Saving Time. Different observers censused the same plot during the course of the field

season. No censuses were taken on windy and rainy mornings. Censusing was terminated

62

lithe M

Variable
as criri-
and r0
accorrr
idenrif
560m
indepr
comp;
during
Seine
Oensu
Censu
Speci.
on of
ObSer

Spear

if the wind velocity became greater than 10 mph.

The field methodology for point transects in distance sampling is equivalent to the
variable circular plot method (Reynolds et al. 1980). Training of observers was identified
as critical to the proper use of the variable circular-plot method (Kepler and Scott 1981)
and to reduce among observer differences. Training of observers was a major task in
accomplishing the field work. All personnel involved in censusing began learning the
identification of bird songs and calls beginning several months prior to the respective field
season. Unexperienced censusers were trained for a period of a month prior to censusing
independently. Observers were trained in a group situation where the opportunity to
compare field data was possible. Six field personnel were involved in censusing birds
during the 1992 field season. Of those 6, 3 individuals were experienced birders with
several years of bird censusing experience. Nine and 8 individuals were involved in the
census in 1993 and 1994 respectively. More than half of the individuals involved in
censusing were experienced censusers. Training centered on proper identification of bird
species songs and calls and on estimating the distances to birds heard or observed. Checks
on observer skills were maintained throughout the census season by having 2 or 3
observers periodically census the same points together”. More effort was spent on
species that were harder to locate aurally and on the species prioritized in this study.

Statistical inference in distance sampling is based on the validity of certain
assumptions that impose requirements on field methodology. An important requirement is
that birds at or very near the census point be detected with a probability of 1. Birds
should also be detected at their initial location, prior to any movement in response to the

observer. A third requirement is that distances from the census point to the bird should be

63

measur
require
bird ca
more

know

43a

meas:
cardi-
break
plors

Tetor

Wirh
COVe

cinc

measured accurately. Personnel training discussed above was crucial to meet these
requirements. In addition censusers were required to wait 2 minutes after arrival at the
bird census station to allow birds to settle back and stop responding to the observer’s
movements. To aid in the estimation of distances to detected birds, flags were tied at

known distances from the bird census station.

4- Sampling of Microhabitat Variables

Variables describing microhabitat structure and composition (Table B3) were
measured at the bird census station. Four 10 x 30 m plots were setup, aligned along the 4
cardinal directions from the bird census station (Figure B1). Tree species, diameter at
breast height and height were recorded for all trees within the North and South 10 x 30 m
plots. Several tree heights were measured using a clinometer and visual estimates were
recorded for the rest of the trees in the vegetation plots. The number of woody stems
<O.5" (lcm) and >O.5- 2" (>lcm-3cm) was counted at breast height in 12 5 x 2 m plots
within the 4 10 x 30 m plots. Canopy cover was recorded by 2 measures, percent canopy
cover and number of canopy gaps in 4 canopy gap size classes. The quantification of
canopy cover followed a modification of the Canfield method (Canfield 1941). Percent
canopy cover was obtained by stretching a measuring tape along the 30 m length of the
vegetation plot, and recording the stretches of tape above which the canopy was open or
closed. This involved visually projecting points in the canopy onto the tape at the position
where the canopy changed fi'om open to closed and fiom closed to open (Figure B2A).
The edges of each canopy gap within the vegetation plots were visually projected onto the

forest floor. The area of a canopy gap was estimated based on its width and length

64

{Figure
>100-r
was re

class <

5- De

areac
SCCOH

heme

“llhir
Cours.
ACCCS

Nallo:

Table

land C(

Was has
Atrial p]

manusm‘

(Figure BZ). Four classes were considered: >10-25m2, >25-50m2, > 50-100m2, and
>100-<150m2. When the canopy was so open that gaps could not be defined the canopy
was recorded to be >5 0% open. Training of personnel involved the estimation of the size

class of canopy openings.

5- Developing Geographical Information System (GIS) Coverage for Study Areas

Geographical information system (GIS) coverage was constructed during the years
1995-1998 for each study area. The first objective of a GIS coverage was to quantify the
area of the difi‘erent cover types within a 1.6 km radius of a bird census station. The
second objective was to quantify the length of edge between specified cover types, and
between cover types and roads, in areas within 500 m from bird census stations.

Digital coverage for forest cover types, was available for part of the public land
within the boundaries of the Hiawatha National forest. The roads, water bodies and water
courses layers were available for most USGS quadrangles covering the study areas.
Access to these data was available through the US Forest Service (Hiawatha and Ottawa
National forests). Geographical digital data were created from the initial stage of
manuscription of cover types for lands that did not have an already available coverage.
Table C3 lists all the 7.5 minute quadrangles which covered study sites, and the sources of
land cover types of the corresponding spatial data.

Map manuscripting of forest cover available fi'om the Hiawatha National Forest
was based on infrared, ortho-photos for 7 .5 minute USGS quadrangles (scale 1:24000).
Aerial photography covering the Upper Peninsula was taken in 1992 during leaf-off. The

manuscripting of land cover on public lands began in 1992. Much of the manuscripting

65

MS C

r'erifi

phO‘
\ar

mor

310
incl

des:

was contracted out by the Hiawatha National Forest. Manuscripts of cover types were
verified on site by US. Forest Service personnel using the same photos, infra red photos
(1992 infrared, leaf-off, scale 1115,840), compartment maps, field checks and other data
bases.

During the period between 1992 to 1996, digital coverage became available for
only a portion of the public land. Spatial data had to be created starting from the
manuscripting stage, for all public lands that had not yet been covered and for all private
lands. The same procedure was followed to create these coverages as was used by the
US. Forest Service, when ortho photos of USGS quadrangles were available. Ortho
photos were available for lands within and adjacent to the boundaries of the Hiawatha
National Forest and for Sylvania Recreation Area but were lacking for the Huron
mountains and for the McCormick area.

Coverages were created for the McCormick Wilderness Area and the Huron
Mountains from available forest cover maps. New cover maps were created that also
included land lying just outside the boundaries of these areas, and to retype cover
designations in a way that it could more easily be tied back to US. Forest Service typing.
The new cover manuscripts were based on the original cover maps and on black and white
photos obtained fiom the Michigan DNR (1986; scale: 1/15,840). For the McCormick
Wilderness Area, the relative dominance of tree species in the different local ecosystems,
given in Pregitzer and Barnes (1984) were used, to match cover types with US. Forest
Service types. The cover type map compiled by Simpson et al. (1989) for the Huron
Mountain Club recognized types along the successional gradient of northern hardwoods as

'hardwoods', 'hardwood dominated hemlock-northern hardwoods', 'hemlock dominated

66

herrloc
recogrr

prepare

using 1
off) rlr
hemio
POSllIr
gradie
comp

cenru

hemlock-northern hardwoods' and 2 types of hemlock. Table C1 lists all cover types
recognized in the classification of different study areas and the corresponding classification
prepared for landscape-level comparisons.

For the Sylvania Recreation Area, delineation of cover types was accomplished by
using 1992 ortho photos of 7.5 minute USGS quadrangles (scale: 1/24000; infi'ared leaf
ofi) that covered the area. A distinction was made between hardwood areas, hardwood-
hemlock and pure hemlock areas. Figure C1 is a schematic representation of the relative
position of forest types, in managed and primary northern hardwood forest, along the
gradient of forest succession from hardwood to hemlock. Much of the hemlock
component in the original forest was lost when the forest was clearcut at the turn of the
century. The US. Forest Service typing of northern hardwood forest, as was previously
discussed, does not separate the second growth forest that has retained some component
of hemlock from that which has completely lost it. The presence of a coniferous
component however as likely to be important for some bird species. It was therefore
important to separate northern hardwood forest that had a coniferous component from
forest that did not. An additional type 819-hemlock was added to US. Forest Service
typing of forest cover to achieve this separation. In U. S. Forest Service typing, patches of
hemlock are often recognized as separate cover types within the matrix. The data base
shows patches as small as 1 ha, as separate hemlock stands. Because the delineation of
hemlock patches was not consistent, the existing cover typing could not alone be relied
upon to insure that the presence of hemlock was delineated. This separation could only be
achieved on a qualitative basis. In the USPS stand data base (CDS) the presence of

hemlock is only noted qualitatively. Quantification of the actual amount of hemlock could

67

theoretically be obtained from silvicultural examinations but these data are not
computerized. A coniferous component could also be recognized from aerial photos taken
with leaf-off; this separation was possible for private lands not covered in the USPS
database. Digitizing of maps of areas not covered by an already available GIS was
accomplished by various individuals using the program Arc/Info. Public and private land
coverages and adjacent coverages were 'edge-joined' if the study area overlapped 2
difierent coverages.

Attribute data were entered that identified each polygon in the coverage. For
private lands the land cover type of a polygon was noted at the manuscripting stage.
When the polygon represented national forest land, the identification it was given
corresponded to the compartment and stand numbers that it represented. The land cover
type then could be obtained fiom US. Forest Service data bases. Areas of polygons
within 1.6 kin-radius circles centered on the location of bird census stations can be
identified by the capability of Arc/Info (ESRI-Environmental Systems Research Institute
1993) to generate a buffer zone around a geographic feature. These 1.6-km radius circle
clips overlaid over the original coverage, created new coverages with new polygon
relationships (Figure C2). The initial design of spatial data analysis however called for a
1.6-km radius, for each bird census station. Many technical difficulties were encountered
after 1.6-km radius clips were obtained. Because digital information was obtained from
the US. Forest Service while the agency was still putting together this GIS system, there
were many instances of incomplete digital data discovered within the 1.6 km circle clips.
Compartment maps provide maps of forest cover types. Circle clips that contained

polygons that were not well defined in the digital coverage were redrawn on compartment

68

maps and the areas of the different polygons measured with a planimeter. For bird census
points that were only a distance of 150 m fiom one another, 1.6 km circles around these
points overlapped extensively. When a visual inspection of the area outside of the area of
overlap showed no difference in forest cover pattern, the large effort it would take to
remeasure portions of missing polygons justified consideration of one circle for adjacent
points. The difference in percent cover of different cover types would be of a magnitude
that it would not afi‘ect the classification of bird census stations with respect to landscape
variables.

Roads, trails, water courses and water bodies were separate coverages provided to
the US. Forest Service by the US. Geological Survey. Base maps for these features were
available for most quadrangles covering study areas. In the road coverage, roads and
structures were classified into types following US. Forest Service specification. These are
given in Table C2. All road types that were maintained, fi'om primary to dirt roads were
considered in the calculation of edge. Because the amount of edge within a 500 m of a
bird census station was of interest, clips of 500 m circles were created in which edges

were quantified.

STATISTICAL TECHNIQUES

1. Estimation of Bird Densities

The program Distance for point transects (Distance, Version 2.1; Laake et al.
1994) was used to estimate bird densities from bird distance data. Analyses of distance
sampling data were repeated for each of 10 bird species to obtain a density estimate for

each species at individual bird census stations. Distance sampling (Buckland et al. 1993)

69

estimates bird density from observed bird distances by modeling a detection function.
Several models and adjustments to these models were initially tried for fitting a detection
function to distance data. The model with the best fit was selected by comparing
alternative model fits with the Akaike’s Information Criterion (AIC). Details of sample
size (number of time the species was encountered), model selected, cutoff points and
intervals applied in the estimation of each species density are given in Table D1. In
addition to bird density, I defined encounter rate to be the °/o of total bird census stations
at which a species was encountered. Density and encounter rate (Table D2) were used

to describe the relative patterns of distribution and abundance among species.

2. Assessing the Effects of Historical Divergence in Land Use, Recent Logging
Disturbance and Land Type Association on Bird Species Densities.

Mean density of each of the 10 bird species herein considered was calculated for
the 2 strata of landuse primary forest and settled forest, and for 5 strata of forest
disturbance within managed forest (Figures 5-8, Table D3). Univariate ANOVA (MGLH;
SYSTAT 5, 1992) was used to test for statistical significance among the 7 means. The
overall F test indicated the statistical significance of the overall difl‘erence among all strata
(Table D4). Mean density and variance of each species was also calculated for managed
forest landscapes. Anova in the procedure Test in MGLH (SYSTAT 5, 1992) was used
to test hypotheses of no difference in the mean density of individual bird species between:
primary forest landscapes and managed forest landscapes; between managed forest
landscapes and settled forest landscapes; and between primary forest landscapes and

settled forest landscapes (Table D5). The statistical significance between means of bird

70

 

d1

density of the strata older grth forest within managed forest landscapes, and northern
hardwood forest in primary forest landscapes was tested. A number of apriori tests,
applied to contrast density of a bird species among specific model strata, were also

planned once habitat availability patterns in northern hardwood forest were identified
(Table D6). These habitat availability patterns were revealed by comparisons of habitat
variables among model strata and are presented in a following section". In the absence of
a logical basis for apriori planned contrasts, the procedure Bonferroni in MGLH was used
as a Anova post test to determine the statistical significance of difl‘erences in bird density
means between all possible pairs of forest disturbance periods and historical divergence in
landuse strata (Table D7).

Mean bird density for each species was estimated for each land type association
within the land use strata primary forest and settled forest and for each forest disturbance
period stratum in managed forest. Univariate AN OVA (Manova in MGLH; SYSTAT 5,
1992) was used to test for the effect of land type association on bird densities (Table D8).

Bird density means were compared among the 3 replicate land type associations, Huron
Mountains, McCormick and Sylvania within primary forest landscapes (Table D9); among
the replicate land type associations, Munising, Manistique and Rapid River for each of the
5 forest disturbance periods in managed forest landscapes” (Table D 10); and among the 3
replicate land type associations, Chattam, Trenary and Stonington within settled forest
(Table D11).

Within the managed forest, recent logging disturbance was represented by 3
periods of forest disturbance that spanned 15 years preceding the herein reported

collection of bird and vegetation data. A Two-way ANOVA (MGLH; SYSTAT 5, 1992)

71

was used to test for the effects of the interaction between land type association and forest
disturbance period on mean densities of the black-throated blue warbler, American
redstart and chestnut-sided warbler (Table 5). Two-way ANOVA (MGLH; SYSTAT 5,
1992) considered the 3 land type associations of Munising, Manistique, and Rapid River
within managed forest landscapes and the 3 forest disturbance periods, logged 1-5 years

ago, logged 6-12 years ago, and logged >12 years ago (Figure 13).

3. Habitat Relationships Among Bird Species.

Univariate regression38 (MGLH; Systat version 5, 1992) was used to identify
habitat variables that afi‘ected densities of the 10 bird species (Table 4). The data set used
in these regressions included all habitat variables and densities sampled at 321 bird census
stations. Variables which had a statistically significant regression coefficient in the
regressions of bird density on habitat variables, in regressions involving the black-throated
blue warbler, American redstart, chestnut-sided warbler, blackburnian warbler and black-
throated green warbler were considered in firrther ANOVA tests of means. These
variables included °/o canopy opening (CAN OPEN), number of canopy gaps of size
>100m2 (GAP>100), shrub layer development measured as number of deciduous woody
stems <0.5" ( WSDC<0.5), number of conifer trees (TOTCON), and number of late
maturity conifer trees (LMCON). Means of these habitat variables (TOTCON, LMCON,
CANOPEN, GAP>100, WSDC<0.5) were calculated for each of the 5 strata of forest
disturbance period within managed forest, and for each of the 2 land use strata of primary
forest and settled forest (Figures 9-12; Tables D13, D20). Univariate ANOVA was used

to test habitat variable means among various strata of northern hardwood forest. These

72

comparisons elucidated the effects of forest management and environmental variables on
these habitat variables (Table D12). Testing the means of habitat variables between
managed forest and primary forest, between managed forest and settled forest, and
between primary forest and settled forest elucidated how different land uses had affected
habitat variables”. Comparisons also tested means of habitat variables between managed
forest that is under active management (logged<12 years ago) and other strata within
managed forest. Univariate ANOVA (MGLH; Systat 5, 1992) was also used to test means
of habitat variables among land type associations within each forest disturbance period
category in managed forest, and within primary forest and settled forest respectively
(Tables D14 and D15).

To delineate the habitat selected by the black-throated blue warbler, American

redstart and chestnut-sided warbler, bird census stations at which the density of the
respective species wasz 8O birds/km2 were considered representative of the habitat

selected by the species“. Univariate ANOVA (MGLH; Systat 5, 1992) was used to
compare the means of microhabitat variables among the 3 species across the whole range
of the northern hardwood forest sampled (Table D23) and within individual strata of forest
disturbance period (Table D24).

Univariate ANOVA was also used to test for habitat variable differences among
the blackburnian and the black-throated green warbler. Because of the low density of the
blackburnian warbler, habitat variable means for this species and the black-throated green
warbler were habitat variable averages of bird census stations at which the respective
Species was at or above a density of 30 birds/km2“. Differences between the 2 species

Were examined with respect to the total range of northern hardwood forest that was

73

sampled, as well as within specific ranges of forest conditions: managed forest excluding
older growth, in older growth forest within managed forest landscapes and in primary

forest landscapes (Tables D16, D18).

4. Eflects of Interactions of Habitat Fragmentation Related Aspects of Forest with
Habitat Availability on Bird Species.

Approximations of total forest cover, and land cover types for the different land
type associations were calculated by averaging land cover types for within 1.6 km radius
circles centered in the different land type associations (Table 1). The range of forest
cover within 1.6 km FC lMIL in the settled forest landscapes stratum was below 70%. A
general absence of the black-throated blue warbler and the blackburnian in settled forest
landscapes precluded testing for effects of fiagrnentation within this stratum. Sampling
units from settled forest landscapes were as a result dropped from further analyses
involving the effects of forest fragmentation related aspects on densities of blackburnian
warbler and black-throated blue warbler. Bird census stations aggregated from primary
forest landscapes and managed forest landscapes were stratified into 3 groups with respect
to forest cover within a 1.6 km radius F C lMIL, into 2 groups with respect to length of
forest edge within a 500m- radius EDGE, and into 2 groups with respect to shrub layer
development WSDC<0.5. The resultant data set comprised 181 bird census stations fiom
managed forest landscapes and primary forest landscapes. The cutoff point of 2000m,
was selected for the variable EDGE to create 2 groups <2000m and >2000m; the median
for 181 bird census stations for which edge data were available was 1888m. FClMIL

ranged from 70% to >90% in the aggregation of bird census stations from managed forest

74

and primary forest landscapes. Bird census stations were stratified into 3 groups of
FCIMIL >90°/o, FC lMIL >80%-<90%, and FClMIL <80%->70°/o. The number of bird
census stations, as well as land type association and forest disturbance period strata
represented in each of the 12 groups resulting from this 3 variable classification are given
in Table D28.

Densities of the black-throated blue warbler, American redstart and a derived
variable BTARCS (Table D30) were compared among the 12 groups of bird census
stations formed in the stratification based on FClMIL, EDGE and WSDC<0.5. The
variable BTARCS was derived from the difference between the density of the black-
throated blue warbler BTBW and the combined densities of the American redstart AMRE
and the chestnut-sided warbler CSWA. Univariate ANOVA (Manova in MGLH Systat
5,1992) was used to test the statistical significance of the efi‘ects FClMIL, EDGE,
WSDC<0.5 and to test for effects of various interactions among these variables, on
BTBW, AMRE, and BTARCS (Table D29).

The effects of interactions among FClMIL, EDGE and WSDC<0.5 were fiirther
investigated by a series of Anovas. Within each F C lMIL group, separate Anovas were
used to test the effects of EDGE and of WSDC<0.5 on densities of BTBW and AMRE
and on BTARCS. In considering EDGE and WSDC<0.5 separately, the effect of the
variable not considered (W SDC<O.5 in the first case and EDGE in the second case) was
randomized (Figure 15, Table 8). The efl‘ects of interactions between WS<0.5 and EDGE
were also assessed within each of the 3 groups of F ClMIL.

To identify landscape and microhabitat aspects in highly forested landscapes

(>80°/o forest cover) that favored the more edge tolerant American redstart and chestnut-

75

sided warbler, ANOVA was used to test the efl‘ects of 2 additional variables, amount of
northern hardwood forest within a 1.6 km radius NHWDIMIL, and % canopy opening
CANOPEN, and their interactions with EDGE and WSDC<0.5. This analysis excluded
the group of bird census stations for which FClMIL in the 3-way stratification was >70%-
80%. This new 4-way stratification produced 16 difl‘erent bird census station groups
based on 2 levels of NHWDIMIL, 2 levels of CANOPEN, 2 levels of EDGE and 2 levels
WSDC<0.5. For EDGE and WSDC<0.5 the cutoff points were as previously established
in the 3-way stratification. Under natural disturbance regimes (in primary forest
landscapes) CANOPEN ranged from 0 to 18%; in northern hardwood forest subjected to
logging, canopy opening ranged from 18% to 26%, 1-5 years afler logging. A cutoff
point of 20% canopy opening was selected for CANOPEN. This value of canopy opening
exceeded the average of naturally occurring disturbance but was within the range of
recently logged forest. Within the data set considered, a natural break in the data set at
roughly 30% occurred in the continuum of percent of northern hardwood forest within a
1.6 km radius (NHWDIMIL). A cutoff point of 30% was established to reflect the actual

occurring difference among the land type associations included in this study.

76

CHAPTER 3

RESULTS

1. Densities and Encounter Rates of Bird Species Across the Northern Hardwood Forest.
Across all combined strata of the northern hardwood forest, the 2 least common
bird species of the 10 herein considered, the blackburnian warbler and the chestnut-sided
warbler, were encountered in respectively 19% and 24% of all bird census stations visited
(Table D2). All other 8 bird species were common with encounter rates > 40%. The red-
eyed vireo, ovenbird, least flycatcher, and black-throated green warbler had high densities
(Tables D2, D3) and high encounter rates >75% (Table D2). The black-throated blue
warbler and the American redstart were about equally encountered at respectively 49%
and 44% of bird census stations. The veery and scarlet tanager were common (encounter

rate>40 %) but occurredat low densities of <15 birds / km2.

2. Association of Bird Species with Small Scale Vegetation Characteristics

Densities of all 10 bird species had significant regression coeflicients on several
vegetation variables that described characteristics of shrub, canopy, tree and shrub layers.
Basal area, canopy opening, deciduous and coniferous woody stems (representing shrub
layer development), and the presence of coniferous tree species affected densities of the
most number of species (Table 4). Seven of the 10 bird species were significantly related

to basal area. Densities of the American redstart, chestnut-sided warbler, and veery were

77

 

 

.0 3:00 a. 3.00 380.0 a. a 3.80 _00.0v 000.0v .3 80 .50

3000.0 3.80 a a a. 3.000 000.0v 80.0v .NV 80 80

.000 .0 .0 .0080 .0 .0 .0 .0 .0 80.0v =0v 000 000

.0 .0080 .0 .0 0N00 .0 .0 .0 .0 .0 0000 01.0000 .m

.0 .0 .0 .080 .0 .0 .0 .0 .0 800 0000 .>

30.00 a. 800.0v a. 3.00.? 3.00.? .0 a .0 a? 80 80

a. .0 :00 .0 $0000 3.000.? 08.0 a. .0 0.-.? 80 80

.0 .0 .0 .0 .0 .0 .iva00 0000 .0 .0 :07: 000 000

.0 .0 .0 .0 .0 .0 000.0 .0 .0 .r:000v ..N.v 000 000

.0 .0 .0 .0 .0000v .0 0.0.0 .0 .0080 gov 00.0 000
gal—0d

3.000 30000 .0 8 a .0 0.0.0 .0 a. 000.0 ....0A 82. .50

$000.0 $000.0 .0 .0 .0 .0 3.0 .0 20.0 .0 ....0v as. .50

.0 000.0v .0 .0 .0 R00 800.0v 30.0 3000.0 .0 ....0A 52. 80

.0 .000v 000.0 .0 0.0000 000.0 800.9 0000.0v $000.? a. ....0v 0.2. 80
00343.

.0 .0 .0 .0 .0 80.0v 0000 .rv.0 .0 .0 0000 0000A

.0 :00 .0 3.0 3.0.0 80.? 000.0 .0 330.0 a. .5007 000

.0 .0 .0 .0 .0300 .0 .0 .0 .0300 .0 N500— 000

.0 .0 .0 .0 .0 .0 .0 .0 70030 .0 N500 000

.0 .0 200 .0 .0 .0 .0 800 .0 .0 N05. 000

.000 3:00 a. a. .000 $000.? $000.? a. .000 000.0 82 .800

a. :00 .0 .0 $000.? 000.0v 80.0v .0 0.000 a. .0008 0&0 e
gnaw

0000.00.00 .0005 0000000 80> 000005 .3 00..00. .3 00.0 .3 000.» .3 03.003
0.00. .0008 00.0 000.. 0000002 00.000. 00.0000 0.00000 00_.0.0mo>
.000 r3585 r0000 .0020 -0020

 

00 .000000. .0 N000N. E003 00.000. 80.00.00.000 0000003.. 00 .0000. 000 00 0.0.0.000 00 800.00%”: 00 000—9, 0.0—00000.0—

.003082 00 0.0.5000 00003 00. 00 0.0.0.. 00030000 500000 E .0008. 00.000 000 N.”

0. 20.:

00
7

000.0009. .0 .05. 00— .. .000—.0.0»? .00 n .0N
000.0508 00.3200. 00.0000 u .0.

 

m:

 

.0 .0 .0 .0 .0.0000v .0 .0 00.0 .0 .0000 .0003 .03
.0 .0 .0 .0mvo0 .0 .0 .0 .0 .0 .0 00.. .000 33
.0 000.0 .0 .0 .0 S00 .0 .0 .0300 .0 00.0 .000 33
.0 .0 .0 .0 .0 .0 0000 R00 .0 0.00. .000
.0 .0000v .0 .0 .0 .000v .0 800.0v .080 .0 00.00
30.00 .0 0.0.0 .80 30.0 .0.0 .0 .0 .0 3.80 .0008.
.0 .0 .0 «N00 .0 v0.0 Nm00 .0 .00000v .0 .200
.0 3.80 .0 .0 .c 3.00.? ~.0.0 0.00 .000 .= .80
0.0.00.0 3.8.0 .0 .0 .0 3.8.0 .0 .0 .000 0.0.0 .82 .0
.0 .0 .0 .0 .0 .0 .0 .0300 .0 .0 000..— 00
000.0 .0 .0 .0 .000 .= .0 $0000 .0800 .0 £00 0
Jill... 3 00005..
0.0.8.0 00..00 .0 3.8.0 .0 .0 .0 .0 .000v .000v .0. 80 .0 0.00.02
.0 .0 .0 .0 .0 .0 .0 .0 .0 .0 =0? .wa0m
.0 .0.0 .0 .0030 .0.0 .0 .0 .0 .000v .0 00000 .3
.0 .0 .0 .0 .0030 .0 .0 .0 .000v .000v :NW 00..—
.-.S00 .0200 .0 .0 $000 .0 .= 3.80 .00.0v .000v .0? 08.000
.0 .0 .0 .0 .0 .0 .0 .0 03.0 000.0. .30.. .m
.3080 I000 3'02...
000.000.... b00> 000.00. 00.... 00.0005 .3 02.00. .3 0.00 .3 000.» .3 030.03
500.. .0080 00.0 000.. 000.0002. 00..00... 00.8.0. 00.0000 00..0.000>
:00”. 005.000 0.0.5 0.005 0.00:.

 

80000080

0 2......

79

negatively related to basal area and positively related to % canopy opening. The
blackburnian warbler, black-throated green warbler and ovenbird were positively
associated with basal area; the ovenbird and the black-throated green warbler were
negatively associated with % canopy opening and number of large canopy gaps. Densities
of the blackburnian warbler, black-throated blue warbler, scarlet tanager, least flycatcher
and red-eyed vireo however were unaffected by the range of variability in % canopy
opening and number of canopy gaps that was sampled.

Densities of 5 of the 10 bird species, the black-throated blue warbler, American
redstart, chestnut-sided warbler, scarlet tanager, and veery were positively associated with
deciduous woody stems (Table 4). The black-throated green warbler and ovenbird were
negatively associated with deciduous woody stems (p<0.00l). Only densities of the
blackburnian warbler, red-eyed vireo and least flycatcher were not affected by the
development of the shrub layer.

Most species showed either a strong association with either deciduous or
coniferous trees. The black-throated green warbler and the blackburnian warbler were
highly associated with large coniferous trees. The black-throated blue warbler, ovenbird,
red-eyed vireo, veery and least flycatcher were negatively associated with a coniferous
tree component. Only the American redstart and the least flycatcher did not show any
strong amnities to either deciduous or coniferous trees. Both species were however

negatively associated with large deciduous trees.

80

3. Vegetation Differences Among Northern Hardwood Forest Landscapes at Different
Spatial Scales

Landscape level differences

Forest cover within 1.6 km radius circles centered in settled forest landscapes
ranged from 40 to 60 %, compared to forest cover of > 70% to >90% in managed and
primary forest respectively (Table 1). Loss of forest cover between primary forest
landscapes and managed forest landscapes was gradual and the 2 land use categories
overlapped in this respect. Forest cover declined sharply in settled forest landscapes
relative to managed forest. Other differences among landscapes included differences in
the percentage of total land cover that is in late successional upland conifer forest
(hemlock and hemlock-white pine) and in northern hardwood forest (Table 1). All
managed forest and settled forest landscapes had lower levels of late maturity conifer
forest types than did primary forest landscapes. Percent composition of late successional
conifers and hardwood dominated northern hardwood forest however also varied among
primary forest landscapes (Table 1). Land type associations replicating a land use also

differed with respect to percent total forest cover that is northern hardwood forest.

Small scale dtflerences in the conifer tree component among northern hardwood forest
strata

Small scale vegetation characteristics differed among different land type
associations and difl‘erent land uses. Both the number of total coniferous trees, TOTCON
(includes all conifers red and white pine, hemlock, cedar, balsam fir and white spruce),
and the number of late forest maturity conifers (hemlock and white pine) were significantly

difl‘erent among the 3 land uses primary forest, managed forest and settled forest (Table

8]

D3). TOTCON and LMCON were significantly higher in primary forest than in either
managed forest or settled forest landscapes (Tables D3, D4). Managed forest landscapes
however did not differ with respect to TOT CON and LMCON from settled forest
landscapes. TOTCON and LMCON did not significantly differ among different land type
associations within the land use categories primary forest or settled forest. Within
managed forest, neither TOTCON nor LMCON varied among land type associations
replicating a forest disturbance period (Tables D5, D6). LMCON was significantly higher
in the older growth forest category than in the combined 4 other strata of forest
disturbance period (Figure 5; Table D3). TOTCON and LMCON however did not differ
between the older grth category within managed forest and primary forest landscapes

(Figure 5, Table D3).

Canopy characteristics among strata of northern hardwood forest
Canopy characteristics represented by % canopy opening (CAN OPEN), and

canopy gaps of different size classes (GAPSO, GAP100, and GAP>100) did not vary
among primary forest landscapes (Tables D7, D8). Within managed forest however,
significant differences among means of CAN OPEN, GAPSO, GAPlOO and GAP>100 were
common among landscapes replicating forest disturbance period strata (Tables D7 , D8).

In managed forest, canopy characteristics reflect the time elapsed afier logging (Figures 6,
7). Percent canopy opening is highest in managed forest that had been logged less than

12 years prior to sampling (Fig. 6; Tables D9, D10). Mean canopy opening within 12
years following logging is significantly higher than means of canopy opening in primary

forest and in settled forest (Table D10). It is also significantly higher than means of

82

 

    
  
 
 

 

 

 

 

 

 

 

 

 

 

 

 

 

l .
t Hemlock and Total Conifer
I 8 f__
+ l
‘ l
A } n=25
E i $1.4 1.2
0 I
V 6 g
in l
:2 1., _ W - - -
i0 ‘ n=46
ii se=l.l 0.3 n=55
m 4 se=0.9 0.7 “:58
Q) sewB 0.7
:9. ._._
“5
L- 7
8 2 73*
E if;
3 *1: 1:1
c a ..
I
0
I Disturbance period & land use strata
i
I i « irorco~ LMCON
l m.

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older growth forest; 6.Primary forest; 7.Settled forest.

Figure 5. Means of total number of conifer trees (TOTCON) and of late maturity conifers
(LMCON) counted in 1200 m2 vegetation plots at bird census stations in northern
hardwood forest strata of forest disturbance period and land use.

83

 

n=43
542320.29 ‘

 

l
l
l
l
20‘
l , , ,, ,
l g) l 11:46
-— 1 33:12.49

‘ g ":25
‘ g- ‘ se=8.41
’ 810
l °\° ‘

5 .

0 , -.

1 2 3 4 5 6 7

Disturbance & land use strata

I CANOPEN

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older growth forest; 6.Primary forest; 7.Settled forest.

 

 

Figure 6. Means % canopy opening measured within 1200 m2 vegetation plots in different
northern hardwood forest strata representing forest disturbance period and land use.

84

 

n=46 l

se=0.l9 0.13 0.15 Canopy Gaps g

 

 

 

 

 

 

         

 

 

 

 

 

 

n=43
‘ 1.6 se=0.l9 0.13 0.16
\ n=55
§ sc=0.l7 0.11 0.13
1 4 - \
' \ n=S7 \§
1 \ \
§ —. se=0.l7 0.11 0.13 -___ __ fl, 7 “=46 \
i a 1.2 “:40 sc=0.19 0.13 0.15 E g
‘ (0 «=02 0.13 0.16 . \ , 2 ,, l
U) \ ‘
J “A \ \ 1
l a 1 § a 3
° \ \
C \ér-t. \ l
808 V .m a; I
‘5 § se=0.25 0.17 0.2 i
" § u
E § \\\ \ij i
3 1:,
\ \
F 0'4 § § l
. s s».
l 0'2
i o g 13?} {11: , '1. \\ :: § :\\ ._ jg;"_
i 1 2 3 4 5 6 7

Disturbance period & land use strata

I GAP>100

_fi

1 GAP100

 

 

§ GAPSO

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older growth forest; 6.Primary forest; 7 .Settled forest.

Figure 7. Mean number of forest canopy gaps of difl‘erent sizes sampled with 1200 m2
vegetation plots in difl‘erent northern hardwood forest strata representing forest
disturbance period and land use.

85

canopy opening in unlogged second grth and older growth strata in managed forest
(Table D9, Table D10).

Within managed forest, mean number of gaps of size classes >25- 50m2, >50-
100m2, and >100-<150m2 were all higher within 12 years afier logging than in the strata
of unlogged second grth forest or older grth forest (Fig. 7; Tables D9, D10). Thus
within managed forest the effect of logging was an increase in the mean number of gaps of
all 3 size classes. Differences in % canopy opening between strata of managed forest
logged <12 years ago and primary forest, and between strata of managed forest logged
<12 years ago and settled forest, were largely due to an increase in the mean number of
canopy gaps of size class 50m2, GAPSO (Tables D10, D9). Mean number of gaps of the 2
size classes mom and >100m2, GAPIOO and GAP>100 respectively, were not
significantly different between primary forest and managed forest logged < 12 years ago
(Table D10). Mean number of canopy gaps of size class >100m2 was significantly greater

in settled forest than it was in managed forest logged <12 years ago.

Shrub layer differences among northern hardwood forest strata

Shrub layer development represented by deciduous woody stems <0.S" (1.26cm),
WSDC<0.5, in general showed significant differences among different landscapes
replicating a northern hardwood forest category (Table D7). Primary forest landscapes
differed significantly with respect to mean number of woody stems <0.5" (p=0.023).
Within managed forest, mean WSDC<O. 5 differed among landscapes replicating the same
forest disturbance period category (Table D7). In managed forest the mean number of

woody stems increased 6-12 years afier logging (Fig. 8). It was lowest in unlogged

86

 

I n=55 I

 

 

      

 

 

 

             

80 I se=6.5 2.3 _se;6_r;:§1IT_Shmb Layer g
I
E I § § n=25 I
:60 I § § se=9.8 2.6 “F43
I '8 i — — se—7.4 3.,7 I
o n—43 \ \ n—46 I
; se=7.2 2.3 \ se=7.2 3.4 I
a) \ I
§4o _ ~ -, I
'8 I . § I\ 3537017 I
8 a = \ ~
“5 I \ § \ I
L : § \ \ I
a \ \ \ a
0 s s a \
1 2 3 4
Disturbance pen‘od & land use strata

 

E2

a WSDC<0.5 I WSDC<0.5 I
I

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older grth forest; 6.Primary forest; 7.Settled forest.

Figure 8. Means of woody stems <0.5" (1.26 cm) WSDC<0.5 and woody stems >O.5"

WSDC>O.5 sampled within 12 2x5m vegetation plots located at bird census stations of
different northern hardwood forest strata of forest disturbance period and landuse.

87

second growth forest reflecting the low % canopy opening (Figure 6). Mean WSDC<0.5
in the strata of managed forest logged <12 years ago was significantly higher than it was in
either primary forest, settled forest or in the older grth forest within managed forest

(Table D10).

Dynamics of canopy opening tmd shrub layer development within 15 years following
logging across diflerent managed forest land type associations

Dynamics of canopy characteristics and shrub layer development within 15 years
following logging differed among the 3 land type associations replicating managed forest.
These differences were revealed from patterns that means of % canopy opening, number
of gaps in the different size classes, and number of woody stems <05" (1 .26 cm) exhibited
across forest disturbance periods in the different landscapes (Figures 9,10,11).
CANOPEN, WSDC<0.5 and GAP>100 differed among the 3 different land type
associations Munising, Manistique and Rapid River (Table 5, Table 6). A statistically
significant interaction at the 1% level between land type association and period elapsed

after logging indicates that dynamics of shrub layer development differ among landscapes.

(Figures 9,10,] 1).

4. Density Patterns of 10 Bird Species Across Northern Hardwood Forest Strata
Density patterns across land use Categories

The density of each of the 10 bird species differed significantly among strata of
land use and time elapsed after disturbance (Tables D11, D12). The density pattern that a

species exhibited across strata of northern hardwood forest heterogeneity, its association

88

 

Canopy characteristics

 

 

 

 

 

 

 

 

 

 

l
I n=12
I se=55 " __
35I ° I I
I I F “=11 Manistique Rapid River I
30'; If “—12 se=29 -s- 777777 7 g iiiiiiii -.I I
I I I «2:34 11:; 46 11:15 I I
I - I- se=. s = I
I 25+. 1; IiI :~ ---------- _-. ———————— n=22~~~I
I g I I n=l7 I ,-— ”22‘” I
'E I I I 56:2.3 I.
I §20 W I """ I """"" """"""" I
r l I Pi
I g 1' I I.— _ n=24 I: ---I I
I g 15 TI I » OI »»»»»» -. - I - 58:39 - I ~ .-'.l-I - n=20 I
I 3 I I I : I .-. se=l.3 I
10 :I ' -I - - I -- W ..... I. -1 .
II I I . I> I -. . 11 '~ I I . I
I . ————— I ----- II .
. 03‘ ‘f ILA II“ II. I' -.II”I
I 1-5y >12-15y 1-5y >12-15y 1-5y >12-15y I
I 6-12y 6—12y 6-12y ‘
I Forest disturbance period I

 

Figure 9. Mean % canopy opening sampled within 1200 m2 vegetation plots in strata of
northern hardwood forest representing 3 forest disturbance periods 1-5 years after
logging, 6-12 years after logging, >12 -15 years after logging, in 3 land type associations
replicating the land use strata of managed forest.

89

 

Shrub layer development

 

 

 

 

 

 

 

I I
I 120 I Manistique Rapid Riva I
I I Munising I I
I I “‘2‘ I I
I A100 + - n=ll ----------- “W89 ““““““““ I I
E I
O
I m I I
I N ;
I 5 80 I 77777777 I
I in
I $5 In=12 I I
I g 60 TISta—12.3 I I
I 9 I I
m T
I '5‘ I __. —99
I O 40 I I 55—
2 I I I I
.._ I I
I 0 II I
I ‘I‘ 20 I I I
II _I
I o I" II“
1-5y
I 6-12 y 6-12 y 6'12 Y
I Forest disturbance period
I

 

 

Figure 10. Mean number of woody stems<0.5" (1.26 cm) sampled within 1200 m2
vegetation plots in strata of northern hardwood forest representing 3 forest disturbance
periods 1-5 years after logging, 6-12 years after logging, >12 -15 years afier logging, in 3
land type associations replicating the land use strata of managed forest.

90

 

 

1.6 a

Canopy Gaps >100m2

 

f Munising

i

1.4 +-

T
1.2 if” ——————

# of Canopy Gaps/1200m2
p o

A

.0
N

.0
or

1

'oo

-A—————+——*—fv ——-—-+—-— —+——~— -+—~ #+"'"—‘ —o-- “1L —— +———-

O

___-_—
_- .~ .~--e'
57.r._.
. arr".

 

 

 

 

n=ll

 

__....._._...._a—u-N...-_.__~.._—.—_——m—_——_—.——q

 

 

____.. —__L‘_N‘._ ._ _ _._._

  

 

 

 

 

 

1-5y

6-12 y

:i .r - ~;'~ f?

if 2;. FEE-E3

if a: u: :4 . - ;

‘ ' :2: l " ' L -‘-- .i. l l
‘ T I

r % ._r_-;.':;:_; + T _:_V:;;.__?:.
>12 '15 Y 1-5 y >

 

   

I I
5 y 1-5 y
6-12 y 6—12 y
Forest disturbance period

 

Figure 1 1. Mean number of canopy gaps>lOOm2 sampled within 1200 m2 vegetation
plots in strata of northern hardwood forest representing 3 forest disturbance
periods 1-5 years after logging, 6-12 years after logging, >12 -15 years afier logging,
in 3 land type associations replicating the land use strata of managed forest.

9]

 

 

Table 5. P values of 2 way ANOVAs undertaken to compare the efl‘ects of land type
association, period elapsed afier logging, and interaction of land type
association X period elapsed after logging on means of CANOPEN,
WSDC<0.5, GAP>100. Vegetation variables were sampled 1992-1994 in the

 

 

Upper Peninsula of Michigan.
CANOPENl WSDC<0.52 GAP>1003
Land type association <0.001 0.031 0.38
Period elapsed after logging <0.008 0.001 <0.001
Land type association x 0.208 0.058 0.112
Period elapsed after logging

 

lCANOPEN = % canopy opening.
2WSDC<0.5 = number of woody stems <05".
3GAP >100 = number of canopy gaps >100m2.

92

Table 6. Bonferroni adjusted P values of post test pairwise comparisons of
CANOPEN, WSDC<0.5, and GAP>100 (during the 15 years following
logging) among 3 land type associations. Vegetation variables were
sampled in 1992-1994 in the Upper Peninsula of Michigan.

 

 

CANOPEN WSDC<0.5 GAP>100
Munising vs Manistique 0.001 1.00 <0.001
Manistique vs Rapid River 1.00 0.201 1.00
Munising vs Rapid River 0.001 0.037 <0.001

 

lCANOPEN = % Open canopy.
2WSDC<0.5 = number of woody stems <05".
3GAP>100 = number of canopy gaps >100m2.

93

with microhabitat variables, and the pattern of vegetation characteristics across northern
hardwood forest strata, suggested which underlying factors and their spatial scales
afi‘ected densities of the different species. Bird species response varied from an increase in
density to a near absence across land use strata fi'om primary forest to settled forest
landscapes.
Density of the ovenbird did not differ among land use categories (p>0.3, Table

D13, Figure 1). Three species, the least-flycatcher (Fig. 12), the chestnut-sided warbler
(Fig 13), and the veery (Fig. 14) increased in settled forest landscapes (Table D13). The
American redstart ( Fig. 13) density did not differ between managed forest and settled
forest landscapes (p=0.996, Table D13). Densities of the red-eyed vireo (Fig. 12), black-
throated blue warbler( Fig. 13), scarlet tanager (Fig. 14), blackburnian warbler and black-
throated green warbler (Fig. 15) were lowest in settled forest landscapes (Table

D13) which suggested that these 5 species would be the most likely species to be
sensitive to forest loss or fragmentation.

Density patterns of the 10 bird species among primary forest, managed forest and
settled forest landscapes also served to demonstrate habitat relationships among bird
species and to ordinate them along a disturbance intensity gradient. Blackbumian
warbler density declined sharply between primary forest landscapes and managed forest
landscapes (p=0.003) but to a much lesser degree than between managed forest and
settled forest landscapes (p=0. 157) (Fig. 15, Table D13) . In contrast, densities of the
black-throated green warbler and scarlet tanager declined only between managed forest
landscapes and settled forest landscapes (p=0.002 for each species, Table D13). Densities

of the veery (Fig. 14) and the American redstart (Fig. 13) increased significantly between

94

 

 

-se=7.9 47.6. 7.4

1
1 600 n=43
1
I
1
i
1

    
 
  
  
 

500 1 —

f _ i
1

N 400 s

E l

Rx. i n=25

% I se=10.462.6 9.7

c l

g _ n=46

"‘46 F55 “’58 “=42 sc=7.7 46.2 7.1

'2 ”=17 46.2 7W7o 422 6 .ac=6.9 41.1 6.3

'- ' " se=8.1 48.3 7.5

m 200 7 1

 

  

 

1 100 . — ‘—-~~~—-

 

 

 

O
.a.
N
(A)
-h» .
01
(D
‘1

Disturbance period and landuse strata

. Ovenbird Least Flycatcher L Red-eyed vireo

 

 

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older grth forest; 6.Primary forest; 7. Settled forest.

Figure 12. Mean densities of ovenbird, least flycatcher and red-eyed vireo in northern
hardwood forest strata of forest disturbance period and land use.

95

’

 

 

 

n=55

 

 

 

     
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i
i
1 _
i i se=10.2 10.4 5.8
1 l
i
i 120 1 n=58
l l sc=9.9 10.2 5.6
l
100 L
N
g n=42 25
\ sc= . . . n:
2‘ 80 ”7119 66 se=15.2 15.5 8.6
'55 n=43
5 $11.6 11.8 6.55
IO 60 i " [1;46H—h ' "—7" “ "
1E! se=ll 2 1L4.6 3
'55 M
i .
i 20 r h —~ r~
'1
1 o w . _ . . . 1 +__
i 1 2 3 4 5 6 7
I Disturbance period and landuse strata
1 \
i I Black-throated blue warbler American redstart
l . .
i Chestnut-sided warbler

 

 

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6—12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older growth forest; 6.Primary forest; 7.Settled forest.

Figure 13. Mean densities of black-throated blue warbler (BTBW), American redstart
(AMRE) and chestnut-sided warbler in northern hardwood forest strata of forest
disturbance period and land use.

96

 

 

 

   

n=42

:1
'00
N
U)
I

 

Bird density Ikm2

    

 

 

 

     

°° /////////////////////////////////,
”‘ /////////////////////////////////

°”///////////////////////%/,

 

Disturbance period and landuse strata

I Veery § Scarlet Tanager

 

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older growth forest; 6.Primary forest; 7 .Settled forest.

Figure 14. Mean densities of the veery (VEER) and scarlet tanager (SCTA) in northern
hardwood forest strata of forest disturbance period and land use.

97

 

I n=42 sc=5.6 9.4

 

 

  
  

            
         

 

  
   

 

 

       

 

 

|
I

I 100 I I
I I ”“73689 S “‘46 I
I I His-{1 $5.4 8.9 1
I 80 I R33. 0“” by: r25 ”=73 122
E I E331 so=4.8 a Iii-iii '
I I 7‘99 "3:515
I ‘é‘ I Iii: 75% .

e 9:1 x. ‘~ to

u \..‘. {SE-v3 l

E 40 I E\\ 3:9. § \ “’43

E I g ‘ \ .~ sc=5.6 9 3
I \ IKE I
I 20 \ a: \ I
I i
I § § § I
I o .\\ .\ § I
I 1 4 5 6
I Disturbance period & landuse strata
I . . .
I I Blackburnian w. rmn Black-throatsd green w. I

 

Legend: Forest disturbance and landuse strata of northern hardwood forest:
1.Managed forest logged 1-5 years ago; 2.Managed forest logged 6-12 years ago;
3.Managed forest logged >12 -15 years ago; 4.Managed forest unlogged second
growth; 5.Managed forest older growth forest; 6.Primary forest; 7. Settled forest.

Figure 15. Mean densities of blackburnian warbler (BBWA) and black—throated green
warbler in northern hardwood forest strata of forest disturbance period and land use. F

98

primary forest landscapes and managed forest landscapes (p=0.03, p=0.006 respectively;
Table D13), but did not significantly difl‘er between managed forest landscapes and settled
forest landscapes. Although common in primary and managed forest landscapes, the least
flycatcher sharply increased in settled forest landscapes (p=<0.001; Table D13). The
chestnut-sided warbler (Fig. 13), was uncommon in primary and managed forest
landscapes but increased sharply in settled forest landscapes (Table D13).

In general bird species showed a linear decline or increase along the range from
primary forest landscapes to settled forest landscapes. The black-throated blue warbler
(Fig. 13) was the only species that showed a significantly higher density in managed forest
than in either primary forest landscapes or settled forest landscapes. Density of the black-
throated blue warbler increased significantly between primary forest landscapes and
managed forest landscapes (Table D13), and declined significantly between managed
forest landscapes and settled forest landscapes (p=0.001). The density of the black-
throated blue warbler was low in both primary forest and settled forest landscapes. The 2
land use strata did not differ (p=0.36, Table D13 ). [In actuality the black-throated blue
warbler differed significantly between primary forest landscapes and settled forest
landscapes when the 2 land use categories were compared in an Anova that included only
these 2 categories. The least square means for the black-throated blue warbler was 28.28
birds/km2 (se = 4.85) in primary forest and 8.4 birds/km2 (se=5.017) in settled forest
(F=8.162; df=1 and 87; p=0.005). The high increase in black-throated blue warbler
density with logging in managed forest resulted in the difl‘erence between primary forest
and settled forest to become insignificant relative to total variance in the AN OVA that

also included all strata within the managed forest landscape. ]

99

Overall the density patterns of all 10 bird species was as follows. Blackburnian
warbler density declined sharply from primary forest to managed forest and continued a
slight decline in settled forest landscapes. The red-eyed vireo, common throughout,
declined gradually from primary forest landscapes to settled forest landscapes. The black-
throated green warbler common in all landscape categories maintained similar densities
over primary forest landscapes and managed forest landscapes but then declined sharply in
settled forest landscapes. The scarlet tanager maintained similar low densities (relative to
other bird species) across primary and managed forest landscapes but declined sharply in
settled forest landscapes. The black-throated blue warbler density peaked in managed
forest relative to both primary forest landscapes and settled forest landscapes. The
ovenbird maintained high densities in all landscape categories. The American redstart was
uncommon in primary forest but maintained similar densities in managed and settled
landscapes. The veery which occurred at low densities throughout, had similar higher
densities in settled forest landscapes and managed forest landscapes relative to its
significantly lower density in primary forest. The chestnut-sided warbler was uncommon
in both primary forest and managed forest landscapes but increased sharply in settled

forest landscapes.

Density patterns of conifer associated bird species

Densities of the blackburnian warbler did not differ between the land use stratum
of primary forest and the older-growth stratum of forest disturbance period within
managed forest landscapes (p=0.583; Tables D14, D15). Within managed forest,

blackburnian warbler mean density in the older growth stratum was significantly greater

100

than the mean of its densities in the 4 other strata of forest disturbance period (p=0.01;
Table D14). The pattern of blackburnian warbler densities in the 7 northern hardwood
forest strata herein considered (Fig. 15, Table D14) closely matched the pattern of means
of total number of conifer trees (TOTCON) and of number of late maturity conifer trees
(LMCON) in these strata (Fig 5; Table D3). A regression of means of blackburnian
density on means of late maturity conifers, LMCON, in the 7 categories of land use and
forest disturbance periods explained >80% of the total variability in blackburnian warbler
density (p=0.005, r2= 0.82).

The precipitous declines that blackburnian warbler density exhibited between
primary forest landscapes and managed forest m=0.003, Table D13 ) and between primary
forest and settled forest landscapes suggested that a shift fi'om primary forest to second-
growth was the major factor affecting it. More exactly, blackburnian density declines
reflect a loss of the coniferous tree component that had occurred as second grth
northern hardwood forest replaced primary forest. In addition to representing second
growth forest conditions, however settled forest also presents a decline in forest cover and
in late maturity conifer forest cover at a landscape scale (Table 1) . Because of the
confounding of 3 factors, loss of conifer tree component with the northern hardwood
forest, loss of forest cover at the landscape level, and loss of late maturity conifer forest
types at a landscape leveL it is not possible to attribute the almost absence of blackburnian
warbler fi'om settled forest landscapes to any one specific factor. The fact however that
the density decline between managed forest and settled forest was not statistically

significant at the 1% level (p=0. 157; Table D13) indicated that loss of the coniferous tree

101

component that came about with the growth of the second growth forest was the more
immediate limiting factor‘.

The density pattern that the black-throated green warbler exhibited across land use
and forest disturbance period strata difi‘ered fi'om the blackburnian’s in that a significant
decline occurred between managed forest and settled forest (Table D13). In contrast to
the blackburnian warbler density pattern, the greater availability of late maturity conifer
and total number of conifer trees in the older growth stratum (Fig. 5, Table D3) within
managed forest, did not result in a greater density of black-throated green warbler relative
to other forest disturbance strata (Fig. 15; Table D15). Regressions of means of black-
throated green warbler density on means of total conifer trees, TOTCON, and means of
late maturity conifers, LMCON, in the 7 categories of land use and forest disturbance
periods were not statistically significant (p=0.69 and 0.84, respectively). For the black-
throated green warbler, a density decline between managed forest and settled forest
(Tables D11, D13) is more likely to be due to a loss of forest cover at a landscape scale.

In general neither blackburnian warbler nor black-throated green warbler mean
densities differed significantly among land type associations that replicated a land use
(Tables D16, D17), or among land type associations within managed forest that replicated
a forest disturbance period category (Table D18; p values of Anovas undertaken to test
difi‘erences in mean densities among land type associations ranged from 0.11 to 0.9, Table
D19). All 3 land type associations replicating primary forest landscapes (Huron
Mountains, Sylvania, and McCormick) showed consistently higher densities of
blackburnian warbler than did all other land type associations replicating managed forest

landscapes (Munising, Manistique, and Rapid River) and settled forest landscapes

102

(Chattam, Trenary and Stonington; Tables D16-D18).

Habitat relationships between the 2 species were fitrther explored for individual
forest strata. Means of TOTCON and of LMCON at bird census stations in which the
blackburnian warbler and the black-throated respectively occurred at a density >30
birds/km2 were considered to represent the means of these 2 variables in habitat selected
by each of the 2 bird species. Means of TOTCON and LMCON for each of the
blackburnian warbler and the black-throated green warbler were compared with means of
TOTCON and LMCON of each of the 7 northern hardwood forest strata. In these
comparisons, means of TOTCON and LMCON at bird census stations in which the
blackburnian warbler occurred at a density >30 birds/km2, was greater than means of
TOTCON and LMCON in all strata of the northern hardwood forest except for primary
forest landscapes and the older growth within managed forest landscapes 2 (Table D20) .
As expected from its high encounter rate and high density in managed forest, the black-
throated green warbler did not differ from any of the strata except primary forest
landscapes. In primary forest, mean LMCON in the habitat selected by the black-throated
green warbler was lower than the mean LMCON for that stratum (Table D20).

Mean CANOPEN for the blackburnian warbler was larger than mean CANOPEN
of each of the northern hardwood forest strata except logged 1-5 years ago and logged 6-
12‘ years ago in managed forest landscapes3 (Table D13, D16). This indicated that the
blackburnian warbler was more limited by its selection for late maturity conifers and total
conifers in the northern hardwood forest, rather than by any fi'agmentation related effect
occurring at the micro habitat because of canopy opening. The additional decline in

blackburnian warbler fi'om managed forest landscapes to settled landscapes could not be

103

due to a more severe loss of micro-habitat availability since the LMCON and TOTCON
did not differ between the managed forest landscapes and settled forest landscapes. The
efl‘ects brought about by a reduction in canopy opening would also not be a likely factor
since the blackburnian warbler mean on CANOPEN did not significantly differ fiom the
mean for settled forest landscapes at the 0 .1 level (p=0. 169, Table D17). Blackburnian
warbler density was also greater in forest disturbance period 1-5 years afier logging than
all other forest disturbance periods in managed forest except the older grth forest
(Table D11). This stratum had the highest mean CANOPEN of all northern hardwood
forest strata considered‘.

The blackburnian warbler and the black-throated green warbler differed with
respect to TOTCON (p=0.001) and LMCON (p=0.002) (Tables D20 and D21). The
black-throated green warbler was associated with habitat that was relatively more
deciduous. The 2 species did not differ in terms of CANOPEN (p=0.93 7) or in terms of
GAP>100 (p=0.22, Tables D20, D21). This indicated that the relationships between the
2 species at the micro habitat level did not have an aspect of habitat fragmentation.
Habitat separation between the blackburnian warbler and the black-throated green warbler
was only significant in managed forest excluding older growth (p=0.062). The 2 species
overlapped the most in primary forest landscapes (p=0.365, Table D22), and the least in
managed forest excluding older-growth (p=0.062‘)’. In regressions of blackburnian
warbler on LMCON in difi‘erent strata of the model, only in the strata of primary forest
landscapes and in older-grth were the regressions statistically significant. This indicated
that for most of the northern hardwood forest in managed landscapes the amount of

LMCON was uniformly too low to affect the habitat selection of the blackburnian warbler.

104

Density patterns of bird species aflected by a shrub-layer

Densities of the black-throated blue warbler, American redstart, chestnut-sided
warbler, veery and scarlet tanager showed a significant response to the increase in shrub
layer development as characterized by the number of woody stems WSDC<0.5 in forest
disturbance periods 6—12 and >12-15 years following logging disturbance in managed
forest (Figures 13, 14, 8). The black-throated blue warbler, chestnut-sided warbler and
veery sharply increased 6-12 years after logging, but declined >12 years after that
disturbance (Figures 13, 14, Table D11). The 3 species had mean densities in the period
6-12 years after logging that were significantly higher than their respective mean densities
for the combined 4 other remaining strata (p=0.001 for the black-throated blue warbler,
p=0.004 for the chestnut-sided warbler, and p=0.003 for the veery; Tables D11, D14).
The American redstart increased (6- 12 years afier logging) but attained even higher
densities in the forest disturbance period >12-15 years after logging; (Figures 13 and 14;
Tables D11, D14). Mean density of the American redstart during the period of >12-1 5
years after logging was significantly higher than mean density in the remaining combined
other strata of managed forest (Table D14). Although the scarlet tanager showed a similar
pattern of increase as the redstart in the period of 6-12 and >12-15 years afier logging, it
also had high densities in older growth forest within managed forest. Density increases of
the scarlet tanager in forest disturbance periods 6-12 years and >12-15 years were not as
distinct as were densities of other bird species associated with a shrub layer development
(Tables D13, D14 and D15 ). Densities of the black-throated blue warbler, American
redstart, chestnut-sided warbler and veery difl‘ered significantly among land type

associations replicating either a land use (primary and settled forest) or a forest

105

disturbance period in managed forest (Table D19). The pattern these species exhibited
across land type associations reflected the variability of WSDC<0.5, canopy gaps and
canopy opening across landscapes (Tables D7& D8).

The black-throated blue warbler’s significant density increase between primary
forest and managed forest (p=0.023, Table D13) is most likely a response to the increase
in shrub layer development in managed forest landscapes following logging disturbance.
The black-throated blue warbler’s significant association with WSDC<0.5 (p=0.001; Table
4) and its significant density increase in disturbance period 6-12 years after logging (p=
<0.001, in Anova comparing density in period 6-12 years after logging with density in the
other pooled strata; Table D13 ) support this argument. Compelling arguments however
exist that preclude shrub layer development from being a major factor in the precipitous
decline of the black-throated blue warbler in settled forest landscapes. The American
redstart and the chestnut-sided warbler, 2 species equally associated with WSDC<0.5
increased in settled forest landscapes. Additionally regressions of American redstart
density, and of chestnut-sided warbler density, on WSDC<0.5 in ‘settled forest
landscapes" were highly significant (p= 0.018 for each regression, Table D23). The
significance of these regressions indicated that settled forest landscapes included some
heterogeneity in the development of the shrub layer that the American redstart and
chestnut-sided warbler were tracking, even though the level of shrub layer development
when aggregated over the category of settled forest landscapes was significantly lower

than in managed forest landscapes (p=0.03; Table D10).

106

Habitat relationships among the black-throated blue warbler, American redstart and
chestnut-sided warbler

Microhabitat relationships among the black-throated blue warbler, American
redstart and chestnut-sided warbler were quantified to test the hypothesis that forest
fragmentation aspects and not simply the presence of a well developed sapling layer could
be responsible for observed density and distribution patterns of the 3 species. Means of
WSDC<0.5, CANOPEN and GAP>100 were calculated for locations in the forest where
each respective species was encountered at a density >80 birds /km2 (Table D24) in order
to elucidate how the 3 species partitioned the availability of microhabitat. The 3 species
did not difl‘er with respect to mean number of WSDC<0.5 (p=0.8) but differed in terms of
CANOPEN (p=0.001) and GAP>100 (p=0.006,Table D25)7.

The increase in American redstart and chestnut-sided warbler densities following
logging corresponded to what would have been predicted based on their habitat
association with reduced canopy cover and deciduous woody stems8 (large canopy gaps
>100m2 did not significantly increase with logging). The relative density patterns that the
black-throated blue warbler, American redstart and chestnut-sided warbler exhibited in
managed forest however, across forest disturbance periods affected by logging 1-5 years
after logging, 6-12years after logging, and >12-15 years after logging, differed from what
would have been expected based on their habitat ordination along a gradient of canopy
cover. The black-throated blue warbler increased in forest disturbance period 6-12 years
after logging when the canopy was still highly open, but declined in the forest disturbance
period >12-15 years when the canopy was already closing. The American redstart on the
other hand maintained a high density in forest disturbance period >12-15 years after

logging when the canopy was already closing surpassing black-throated blue warbler

107

density during that forest disturbance period (Figure 14; Tables D11, D9).

Basing habitat relationships among the 3 species on their distribution across the
whole range of northern hardwood forest, the 3 species were well separated along the
gradients of canopy opening and large canopy gaps. Mthin the managed forest however
there was not a clear separation among the microhabitats selected by the black-throated
blue warbler, American redstart and the chestnut-sided warbler. The 3 species overlapped
with respect to canopy opening (p=0.74), canopy gaps>100m2 (p=0.324) and shrub layer
development (Tables D26, D27 )9. The fact that micro habitat difl‘erences among the
black-throated blue warbler, American redstart and chestnut-sided warbler could not
account for the pattern of their relative density increases after logging, suggested that
factors other than micro habitat differences afl'ected their density patterns.

Since the habitat relationships among the black-throated blue warbler, American
redstart and chestnut-sided warbler across the whole range of northern hardwood forest
sampled reflected a fragmentation component (different response to % canopy opening
and canopy gaps and different patterns of density from primary forest to settled forest), it
was of interest to elucidate their relative densities in the different land type associations
following disturbance by logging. The efl‘ects of interactions of land type association and
forest disturbance period (l-5 years, 6-12 years and >12-15 years following logging) on
densities of the black-throated blue warbler, American redstart and chestnut-sided warbler
were explored in 2- way AN OVAs. Although all 3 species responded positively to forest
disturbance, the pattern of their relative densities and responses differed in the 3 land type
associations (landscapes of Munising, Manistique and Rapid River) replicating managed

forest landscapes (Figures 16,17,18). The efl‘ects of land type association, period of

108

 

Black-throated blue warbler
I Rapid River

 

200 + _____

 

150+-_n:]2- "‘11---- ___________ n=12_-__

Density birds/km2

' -~—-—:- 11:14 :3:
100 I “=12 ————— I ~~~~~~~~~~~ se=33.5 ————————————— ~~~~~

 

 

_. r. I
:- _...:.'
. .3. I
_. any-r“
n— L'...'u'..v-.‘. ..
. v 333 r-
.‘ . ”3

. --
. ..~-. . a
3. ..I-.'~‘_"d
"'"""""_.-. .4.- . 3c: 1 2 8 :t-tz;
' I ~(o. no"; . ' e':"*'i:;‘ .‘
“.1.“ “.7" ‘ff‘ .5?

 

o... H...

. III-:93 _;:.~:::".‘;=;'; “=17 “ ‘ .
50 '..'>— — w _ .. a ._ _ _. t..- I _ _ _ _se—102.

 

 

     

 

 

 

 

 

 

 

 

.s
.. -.
.. 1:. .
._ ~..- ..
.. ‘3' “V
. L .3. ~.. ~.-. .
I . .. .:.r.u
~.-~.~ -. v
__ H _ .,. .., ... __
r. .m .. .t _ . . . ..-..._-:g.‘ ‘. . '-.'.-
~- .,‘.7; ~~-v~..- . r . .~~. .... , . _.~.c.-4 -~
v- . . ,.t.-..~o-. > ~. .. - P .. ... d...’ ' t‘~“.~a .
j y. “-2- ‘_:. » ~ ~ . -vv:-: ‘ ~ 1 .a. .
cm c ~ .u . r .x .. ..~-.:.
' r‘ '. ‘r :77...'._ u '-t~ _-'_ a it L ~-
.:. Lnr r—vr -. ~- x- . ._ _- . - .0 --.~ A
l ~r- ~71 .- .-_.p -'v r- r - - -.-vr -.-, . h'"
‘ on. 10“ .,... .2. ~ x231 .-.--. a
nu- _ .. _. q :.: ._._..r.
l ,. 1 , .‘.- .~.;, I- .- <
' 'V' .-. n I L.‘.'l‘ - wn-r a
I I "warn t; r; .;,. ".6 c It L.
. .n . - . .315.“ . « ~
~_u an- .04 _n . »- _ n
‘ .'.: u ~.., -. was '-
T v .. 7‘ "r.~.I.-.:~ - u
.. u . . . u .. . .
,.. 7 : . . ,
I | 1': .c. #4.:- r. '--
’- . H“. v 7 >...< .
I - - .. ... L4” ~ '- 't
I . tuna; ms! -. ~-
' 3...: .l now» an... _ ,, . - .- . . . . zvurn-
I r- r' . .2 "v - -> . \w- ._~..;1 . ~ ~ . . l '.;z..-~.. - .2-
v . v— 73":1‘. .._ . 5.x.) . . --._.a- . -. . - '.-.:‘ 5'".
an”. .«r. .'.1~ .~ . . _-. - -~ .. .-~ . .. '-v.. . p .
K - -.~.. ~.-.': ’ITV'" .- . -.. . ' . ~.~-~.~ ~~ . -r-- 1 :- . .v‘nr. " _I
-.:n: . -. -- -._g¢.- ~~ -. a ------ . r-w ..-..-u 2. .z.-.-.-.-:_-: .- .. ._ ..-. “A I...
l 1.1—- ~ ' . L.-._-:.-:- .-_ .. . . -, :- -. .-... ,. g.‘_-.;1 . .. . . a - . ”war—r? .- -"-r.
- n--.--—v Q.‘I‘.l"a."r - --. n-H ~.—-. -. - .n..r ._-- - ‘5 ...n..u..- r
- n.1- lag—J j .g . .-. -- r . '. . .1: : ---1- ---r I .-.'__-"': ; ..r.l.-..-.r.n..-u
A -.-- l .—m~-gr 1 >— _r- -. - 1 L3. 1 ::-..... - -.-.'. j .. .. I n
I I I r

 

 

 

.. . .. I '1 l I T . ? 7 : . Filter.
1-5y >12-15y 1-5y >12-15y 1-5y >12-15y
642 y 6-12 y 612 y
Forest disturbance period

 

 

Figure 16. Mean densities of the black-throated blue warbler in strata of northern
hardwood forest representing forest disturbance periods 1-5 years, 6-12 years and >12-15
years after logging in different land type associations replicating managed forest.

109

 

I Chestnut-sided warbler I

I Munising

80'

O)
0

Bird density/km2
rs
o

   

 

 

  

 

 

~ s: I I I
1-5y >12-15y 1-5y
6'12 V 6-12 y 6-12 y
Forest disturbance period

 

 

r,_____#’_’44 _4; V g _._#44
A0

Figure 17. Mean densities of the chestnut-sided warbler in strata of northern hardwood
forest representing forest disturbance periods 1-5 years, 6-12 years and >12-15 years after
logging in different land type associations replicating managed forest.

110

 

I American redstart

Munising Manistique

 

 

I 160 a— ,
i. I fen I
I

I II
fl

 

 

 

Bird density/km2 ‘—
(n
o
I

 

 

;‘ A: '_, ;i’:j:.:f;: se=10.l7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1-5y >12-15 1-5y >12 -15 1-5y >12-1
6-12 y 6-12 y 6-12 y
Forest Disturbance period

 

 

 

—-II——

0
(flat...

 

 

Figure 18. Mean densities of the American redstart in strata of northern hardwood forest
representing forest disturbance periods 1—5 years, 6-12 years and >12-15 years after
logging in different land type associations replicating managed forest.

1]]

disturbance and the interaction of land type association x period of disturbance were
statistically significant for the 3 species (Table 7) . The 3 landscapes differed with
respect to the magnitude of a bird species density increase in any of the forest disturbance
strata, and with respect to the relative densities of the 3 bird species across the 3 forest
disturbance periods. The black-throated blue warbler density in forest disturbance period
6-12 years after logging in the Rapid River landscape, was 2-3 times its levels in that same
forest disturbance period in either the Manistique or Munising landscapes (Figure 16
p=<0.001 for both pairwise comparisons by Tukey’s post test). In contrast increases in
chestnut-sided warbler and American redstart densities in the Rapid River landscape in
forest disturbance periods 6-12 years afier logging and >12 -15 years after logging were
minimal compared to these same 2 forest disturbance periods in Manistique and Munising
(Figures 17 and 18; p<0.001 for all 4 pairwise comparisons by Tukey’s post test) .

The variable BTARCS was considered as the difference between the black-throated
blue warbler density and the combined densities of the American redstart and chestnut-
sided warbler. This variable expressed a contrast between the forest interior black -
throated blue warbler and the more edge tolerant chestnut-sided warbler and American
redstart. Both the Manistique and Munising landscapes showed a negative (BTARCS) for
almost all 3 forest disturbance period strata. In contrast, BTARCS was positive for all 3

strata of forest disturbance period in the Rapid River landscape (Figure 19) .

112

.2<Bmo+m§<0-35m0 3333 800-3536 20 05
taumfie anew—054x 05.3 momzmcon tor—3500 05 98 .5333 0:3 toadofiarxofln of mo bigot 05 050339 ggbbmfi $3.56 N mU§Hm~

 

eotfioomma 095 984...

 

~000v :00 000.0 80.0v 0.0002 coca 08020 00:60 0000885

50.0.7 _000v 000.0v 000.0v 05002 600 06826 0080

:00 _000 ~00.0v 000.0 0000088 00.0 003
0925 <38 $92 Beam scam

 

$0222 00 23580 5&3 2: 5 0000-002 Baas $002.80
03053 003000 05.00 508 05 98 23mg 020003 00200350050 0:0 ASH—>20 0080000 505:2 A3908 0333
0:3 00000209003 05 .00 0020000 5008 co @0me bda 0030—0 0050 0:0 0200000000 095 052 000.509 003000005
0:0 $00me— codm 000020 00:00 £25083 0%: 0:2 00 300.000 05 089.80 8 08.000095 0922 >03 N mo 0.0303 A .0 033.

113

 

‘ BTARCS nanny...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I n=22
I 12C“) . se=23L3
I .
I I
I e150 e, -----
i <1: I ‘ -_~;
1 g i Munrsmg ‘‘‘‘‘
I a 100 I - - ..............................
I g I Manistique
I n=12 n=21
I E 50 + se=3l.5 _ 0:11 ___________ “=15 r. _sc=23 3
3' ":12 se=32.9 11:” n=l4 n=24 “Z282
m I W28 ”226-5 $6292 se=22.3
g.
I; 0 IE1 T7 I“? I 7 : +22- ‘—
I c I I I q I 2:413:53
s I I .. :~
in -5040”; ...__ ................
g I I Ei'.iff;'_":::"
2: I , ' r
"" '- I I
8-100I--——§ ._:-:- —————————————————
a I
-150 1 0% I I I I I I I I
1-5y >12 —15y 1-5y >12-15y 1-5y >12-15y
. 6-12 y 6—12 y 6-12 y
I Forest disturbance period
I

 

 

Figure 19. Mean densities of BTARCS (Density difference of the black-throated blue
warbler - the combined densities of the American redstart and the chestnut-sided warbler)
in strata of northern hardwood forest representing forest disturbance periods 1-5 years, 6-
12 years and >12-15 years after logging in difi‘erent land type associations replicating
managed forest.

114

5. Effects of Landscape level and Micro habitat Variables on Density of the Black-
throated Blue Warbler, American Redstart and the Derived Variable BTARCS.
Effects of forest edge, forest cover and shrub layer development

The effects of landscape level variables on the densities of the black-throated blue
warbler and American redstart, and on the variable BTARCS were explored as micro-
habitat level variables alone failed to provide a solid explanation for the variability of
BTARCS among the 3 land type associations replicating managed forest. Landscape
factors that were considered included 'forest cover within a 1- mile radius' ( FC lMIL),
and ‘amount of forest edge within a 500m -radius from the bird census station' (EDGE).
These effects were explored in managed forest and primary forest onlylo .

A 3-way ANOVA considered 3 levels of F ClMIL, 2 levels of EDGE and 2 levels
of WSDC<0.5". [The contribution of sampling points to each resultant group of forest
cover, forest edge and woody stems fi'om different land type associations, land use and
forest disturbance period strata is given in Table D28]. The density of the black-throated
blue warbler was mostly affected by WSDC<0.5 (p<0.003, Table D29) and to a lesser
extent by FClMIL (p= 0.028) and EDGE (0.10). The black-throated blue warbler density
was also affected by interactions between FClMIL and EDGE (p=0.056), between
FClMIL and WSDC<0.5 0): 0.025) and among FClMIL, EDGE and WSDC<0.5
(p=0.022). The efi‘ects of WSDC<0.5 on the black-throated blue warbler could be
appreciated after the effects of EDGE and FC lMIL were randomized. The effect
corresponded to a difl‘erence in density between 44.56 birds/km2 (n=90, $9.8) for the
group representing WSDC<0.5' = <50 stems and 85.32 birds/km2 (n=91, se=9.55) for the

group of ‘WSDC<0.5' = >50$tems (Table D30).

115

Highest densities of the black-throated blue warbler were for the group of
FClMIL >80-90% and not as would be expected for F ClMIL >90°/o (Table D30). The
seemingly contradictory decrease in density from 89.48 birds/km2 in FClMIL >80-90%
(n=68, se=10.47) to 53.86 birds/km2 (n=74 se=l 1.1 in FClMIL >90°/o (Table D30) for a
forest interior bird species was the result of a confounding of 2 effects (an artifact of this
specific data set): the actual effect of forest cover within 1 mile and intensity of forest
disturbance. Forest cover >90°/o was represented largely by primary forest and older
grth forest within managed forest, while forest cover in the range of 80-90% was
highly represented by managed forest”. The effect of EDGE on density of the black-
throated blue warbler as a main factor was not as significant as its interactions with
FClMIL and WSDC<0.5. Interactions among FClMIL, EDGE, and WSDC<0.5
however were assessed in more detail and are discussed later.

Density of the American redstart was significantly affected by FClMIL, EDGE,
and WSDC<0.5 (p = 0.001, <0.001, and 0.001 respectively, Table D29) in the ANOVA
analysis that included all 3 variables as factors. Interactions between FClMIL and
EDGE and that between EDGE and WSDC<0.5 were also statistically significant
(p=0.001 and 0.012 respectively; Table D29). When the efl‘ects of EDGE and WSDC<0.5
were randomized (one-way Anova with F C lMIL as factor), the least squares mean for
American redstart density in FClMIL >90°/o was 29.30 birds/km2 (n=74; se=8.62) in
contrast to 81.43 birds/km2 (n=39; se=10.72) in FClMIL > 70-80%. When the efl‘ects of
FC lMIL and WSDC<0.5 were randomized (One way Anova with EDGE as factor), the
effect of EDGE could be recognized as an increase in density from 22.66 birds/ka

(n=108 se=7.04) for EDGE <2000m to 73.95 birds/km2 (n = 73, se =8.0) for

116

EDGE>2000m. In a similar randomization of the other 2 factors (one way Anova
WSDC<0.5 as factor), the effect of WSDC<0.5 was an increase from 31.01 birds/km2
(n = 90, so = 7.63) for ‘WSDC<0.5' = <50 to 65.6 birds/10112 ( n = 91, se = 7.4) for
‘WSDC<0.5' = >50 (Table D30). The interaction of FClMIL and EDGE increased
American redstart density from 25.9 birds/km2 (n=53 se=9. l l) for FClMIL>90°/o,
EDGE<2000m to 136.56 birds/km2 (n=20, se=l4.9) for FClMIL>70-80°/o and EDGE
>2000m.

The derived variable BTARCS was highly affected by forest cover FC lMIL
(p<0.001, Table D29), by EDGE (p<0.001) and by the interactions between EDGE and
FClMIL (p=0.001) and among FClMIL, EDGE, and ‘WSDC<0.S'. BTARCS was
negative for most groups of bird census stations for which FC lMIL was <80°/o, except for
combinations of F C1MlL<80°/o and EDGE <2000m (Table D30). The least squares
mean for BTARC S was negative (~34.18) for EDGE >2000m (n= 73, se=l3.9) when the
effects of FClMIL and ‘WSDC<0.5' were randomized (one way Anova with EDGE as
factor). The highest (positive) value of BTARCS was for the group FClMIL >80°/o -
90%; EDGE <2000m, and ‘WSDC<0.5' = >50, which delineates the conditions under
which the black-throated blue warbler fared best. The lowest value of BTARCS was for
FClMIL <80°/o; EDGE >2000m; ‘WSDC<0.5' = >50. BTARCS was also negative for
FClMIL >80 - 90%, EDGE >2000m, ‘WSDC<0.5' = >50 indicating that edge increases
edge bird species even in relatively high forest cover.

The significant effect of the interaction between FClMIL, EDGE, and WSDC<0.5
and the relevance of this relationship to management of forest bird species, warranted

exploring the effects of edge and shrub layer development separately within each level of

117

forest cover. Within forest cover >90%, the effect of EDGE and the effect of the
interaction of EDGE with ‘WSDC<0.5' were not significant for either the black-throated
blue warbler or the American redstart (Table 8) and as a result these 2 effects were not
significant for BTARC S. Densities of both the black-throated blue warbler and the
American redstart increased at the higher level of WSDC<0.5. An increase in
WSDC<0.5 afi‘ected the black-throated blue warbler more than it afl‘ected the American
redstart. BTARCS remained positive for all combinations of EDGE and WSDC<0.5
within FClMIL >90°/o (Figure 20).

Within forest cover 80-90%, the higher level of forest edge but the efl‘ect of EDGE
was not significant (p=0.194). The American redstart density more than tripled between
the lower and higher levels of EDGE (Figure 20a), an increase that was statistically
highly significant (p=0.004, Table 8). The interaction between EDGE and ‘WSDC<0.5'
was significant for the black-throated blue warbler, and American redstart and was
reflected in a high statistical significance for BTARCS. The American redstart densities
remained relatively low for all combinations of WSDC<0.5 and EDGE except for the high
levels of both EDGE and ‘WSDC<0.5' (>2000m, >50 respectively). The black-throated
blue warbler increased with a higher level

of ‘WSDC<0.5' (p=0.006). That increase was greatest for the condition of EDGE
<2000m (p=0.053 for the interaction between EDGE and ‘WSDC<0.5)’. The Black-
throated blue warbler density declined from 181 birds/km2 for the group WSDC<0.5 =
>50; EDGE <2000 to 83 birds/km2 for the group ' WSDC<0.5' = >50; EDGE >2000m.
This decline in the density of the black-throated blue warbler suggested a negative efl‘ect

of high density of American redstart on the black—throated blue warbler. BTARCS was

118

Table 8. P values in Anovas undertaken to test the effects of EDGE' and WSDC<0.52
on mean densities of BTBW3, AMRE4 and mean of BTARCS5 within 3 levels
of FClMIL“. Bird density and vegetation variables were sampled in 1992-
1995 (Upper Peninsula, Michigan).

 

 

BTBW AMRE BTARCS

Within FC1M1L>90

EDGE 0.194 0.8 0.382

WSDC<0.5 0.003 0.017 0.619

EDGE*WSDC<0. 5 0.127 0.289 0.406
Within FClMIL>80<90

EDGE 0.39 0.004 0.024

WSDC<0.5 0.006 0.011 0.512

EDGE*WSDC<0.5 0.053 0.015 0.005
Within FClMIL>70<80

EDGE 0.038 0.002 0.001

WSDC<0.5 0.687 0.364 0.436

EDGE*WSDC<0.S 0.448 0.309 0.782

 

'EDGE = group based on length of forest edge within 500m radius.

2WSDC<0.5 = group based on number of woody stems within vegetation plots.

3BTBW = Black-throated blue warbler.

‘AMRE = American redstart.

’BTARCS = difference in density between the Black-throated blue warbler and the combined
densities of the American redstart and the chestnut-sided warbler.

“FClMIL = group based on Forest cover within 1 mile radius.

119

2:: F £53, .98 628 .x.

 

- . I I o2-

 

 

 

gowvoefiuzaa goovowézaa $8356;

 

 

 

 

 

 

.m
... a...
. ..
_ a...//
Lu...
T
hflJJ JAM);
411a./.r.....

. . . . . 1
.. a . a .— r 4

. a . . I
.......

. a . . ..
s . a . r

lllll .5”I:..4O‘
I v I ’4 r e

Zuni/5911191199 DJ!8

om

 

 

 

 
 

 

at. , .- . . - i .- _- I 2:
om”: . emu:
, mm _m

I: I - -IIIII.-. one
8 8mm;

 

 

 

 

120

 

 

 

 

 

 

 

 

 

2E. _. £55, .530 $92 3

In . , _- i - ..|--.¢

gowVOhAA—Z _ Um

 

 

own:

 

 

mm

 

l .‘ , - .‘ .' z , a .
r .’ .‘ . .- I _. f . I f l p

f 4 - . . .~ g , r .

g . .. . , . , ' . r ,1 , , x

I ,' . . . r - - . . . I .

4.4424 444.4%.4.

 

 

aboavoqu—E _ Um

mm

 

II; ir-.+-¢il .r+ I» -Ixiig one-
- .. cos-
I oml
$83G; _ Um

 

CD
Zuni/39951199 PJ!8

on

 

 

_Nu: mm":

1 . I - - I 02
mm _m

I I- I I o2

 

 

120

 

 

mum—(hm

 

mm2<
>55
4A.. Z
//M
/

 

/.

/
/

 

 

 

 

: V

 

 

... .
.—..

 

 

omVONA 4:2 F0”.

 

omvomA r_=>:.o.._

 

_mwxxr ,

on":

 

omA 4.530”.

 

 

--f,

.om

 

co..-

ZUDIMJBUGD pm

cow

 

121

 

I

 

 

 

 

 

.3
.

J
w, r.
a - I

on:

Nm>>_m~

 

,
.
a J.
. Ir.

 

c.

__u:

_w>>mm

A a .- l
943.,
I x _. C.
I . a 4 —.
2,um
If?» , . .

.
,JJ 4 3,“...
.. .,rr.-
:9“ - r

_m>>_m

 

wwnnwn
2:: F £55, .98 522 o\..

ANT?! ,_ _ a- a1“

soovomézau $83.58 .

/ .
. r
.i .1 I
a 1.‘ . I
. r. I.
.. . r/r I .
v “ If .ra
.. 1 1r
.. z I
;.
fl . /, .
,, r .r.
a
a
.r
I.
.

_Nn:

  
 

Bi: 5.35 .- or: $3;

_m>>mm Nm>>mm

_m>>_m

 

 

om—

: com

com-

omF-

cor-

zwx/Kusuap one

com a:

 

122

Figure 20 (continued)

Legend.
Grou b' census sta ' w '
FC 1MI1>90% Forest cover within 1 mile >90%.
FC lMII)80-<90% Forest cover within 1 mile >80 <90°/o
FClMIL>70-80% Forest cover within 1 mile >70% -80%
El Forest edge within 500m <2000m
E2 Forest edge within 500m >2000m
WSl Number of woody stems<0.5” <50
W82 Number of woody stems<0.5" >50
ElWSl Forest Edge<2000m & Number of woody stems<0.5" <50
E2 WSl Forest Edge>2000m & Number of woody stems<0.5" <50
EIWSZ Forest Edge <2000m & Number of woody stems<0.5" >50
EZWSZ Forest Edge >2000m & Number of woody stems<0.5" >50

 

Figure 20. Efl‘ects of forest edge, and shrub layer development and their interactions on
densities of the black-throated blue warbler and American redstart and on BTARCS in
different conditions of forest cover within 1 mile.

123

positive (black-throated blue warbler winning) for all combinations of ‘WSDC<0.5 and
EDGE except for the combination of higher levels for ‘WSDC<0.S' and EDGE.

In forest cover >70-80%, EDGE was the only factor that affected densities of the
black-throated blue warbler and American redstart, and therefore BTARCS (p=0.03 8,
0.002 and 0.001 respectively). At this forest cover level, the black-throated blue warbler
density showed no significant response to an increase in shrub layer development. Any
development of shrub layer was preempted by the American redstart as its density
increased with the increase in WSDC<0.5 (Figure 20 b). The combination of high EDGE
and high ‘WSDC<0.5' favored the American redstart as it did in PC lMIL >80-90%
(Figure 20 c.). The effects in this case were however even more accentuated for all
responses, for the increase in the American redstart and for the decrease of the black-
throated blue warbler and for the more elevated negative value of the variable BTARCS.
Eflects of extensiveness of the northern hardwood forest, canopy opening, shrub layer
development, and forest edge.

The effect of the extensiveness of the northern hardwood forest within highly
forested landscapes FC lMIL >80°/o on black-throated blue warbler and American redstart
densities and on the variable BTARCS was explored. In 4-way ANOVAs, % of northern
hardwood forest cover within a l mile-radius NHWDIMIL was considered along with
length of forest edge within a radius of 500m EDGE as landscape factors, and the factors
CANOPEN and WSDC<0.5 as micro habitat factors. In highly forested landscapes most
combinations of the 4 factors favored the black-throated blue warbler as was evident by

the positive values of BTARCS (F igure21a, 21b). The black-throated blue warbler density

124

 

 

 

 

 

 

 

 

 

 

|
I
l
|

 

l

n=3l

    

18

I]

  
 
 

|
l

WWW/

‘ a

l
l

l
l
l
l

 

 

 

 

 

 

Ecnjp3 (HKCHJP4

Habitat characteristics

CHUDLHXZ

 

(3RLNJP1

 

 

 

-4o

 

 

 

 

 

 

 

 

 

 

 

zwwkusuap pm

 

126

Figure 21 (continued )

Legendfigmtella

Group 1
% Northern hardwood forest within 1 mile ( NHWDIMIL) <30%

% Canopy opening (CANOPEN) < 15%
Number of woody stems <05" (WSDC<0.5) <50

Group 2

% Northern hardwood forest within 1 mile ( NHWDIMIL) >30%
% Canopy opening (CANOPEN) <15%

Number of woody stems <05" (WSDC<0.5) <50

Group 3

% Northern hardwood forest within 1 mile ( NHWDlMIL) >30%
% Canopy opening (CANOPEN) <15%

Number of woody stems <05" (WSDC<0.5) >50

Group 4

% Northern hardwood forest within I mile ( NHWDlMIL) >30%
% Canopy opening (CANOPEN) >15%

Number of woody stems <05" (WSDC<0.5) <50

Group 5

% Northern hardwood forest within 1 mile ( NHWDIMIL) >30%
% Canopy opening (CANOPEN) >15%

Number of woody stems <05” (WSDC<0.5) >50

Wilt:

Group]

% Northern hardwood forest within 1 mile ( NHWDIMIL) <30%
% Canopy opening (CANOPEN) <15%

Edge within 500m <2000m

Group2
% Northern hardwood forest within 1 mile ( NHWDlMIL)>30%

% Canopy opening (CANOPEN) <15%
Edge within 500m <2000m

127

(Fig 21 continued)

Group3

% Northern hardwood forest within 1 mile ( NHWDIMIL)>30%
% Canopy opening (CANOPEN) <15%

Edge within 500m <2000m

Group4

% Northern hardwood forest within 1 mile ( NHWDIMIL)>30%
% Canopy opening (CANOPEN) >15%

Edge within 500m <2000m

Groups

% Northern hardwood forest within 1 mile ( NHWDIMIL)>30%
% Canopy opening (CANOPEN) >15%

Edge within 500m >2000m

 

Figure 21. Efl‘ects of interactions among NHWDIMIL, CANOPEN, and WSDC<0.5, and
among NHWDIMIL, CANOPEN and EDGE on black-throated blue warbler and
American redstart densities and on BTARCS. Data were obtained from bird censusing
and vegetation sampling in 1992-1994 in the Upper Peninsula of Michigan.

128

was not affected by NHWDIMIL (p=0.69 ) or EDGE (p=0.816). The effects of EDGE
on density of the American redstart bordered on the 1 % significance (p=0.11). BTARCS
was also not affected by NHWDIMIL (p=0.513) or by EDGE (p=0.23). This indicated
that in highly forested landscapes the effect of EDGE ‘in general’ and alone did not favor
edge species over the more forest interior black-throated blue warbler. Both CAN OPEN
and WSDC<0.5 were highly significant for the black-throated blue warbler (p<0.006 and
<0.001 respectively). An increase in CANOPEN and WSDC<0.5 increased black-
throated blue warbler density fi'om 24.65 birds/km2 ( n= 49, se=16.29 for CANOPEN
<15% and WSDC<0.50 = <50; tol7l birds/ km2 (n=30, se=22.58). WSDC<0.5 was
significant for the American redstart (p = 0.018) but the effects of CANOPEN were less
pronounced than for the black-throated blue warbler. When the effects of all the other 3
variables were randomized, an increase in ‘WSDC<0.5' favored the black-throated blue
warbler as represented by the difl‘erence between 29.97 birds/km2 for ‘WSDC<0.5'= <50
(n= 73, se =15.65) and 115.61 birds/km2 for ‘WSDC<0.5' = >50 (11 = 69, se = 15.20;
Table D32 ). Density of the American redstart increased to a lesser magnitude with this
range of increase in woody stem density (17.96 birds/km2, se =8.26; to 45.61 birds/km2,
se = 8.02 respectively; Table D3 2).

The interaction between NHWDIMIL and WSDC<0.5 was significant for the
black-throated blue warbler (p=0.021) and very significant for BTARCS (p=0.006). The
highest density of the black-throated blue warbler was encountered in landscapes with a
lower % of northern hardwood forest with a high level of WSDC<0.5”. The low value of
3.66 (n= 57, se = 14.44) for BTARCS for the group of bird census stations in the 2 factor

combination NHWDIMIL >3 0%; WS<0.5 >50 indicates that overall (when the efiects of

129

EDGE and CAN OPEN were randomized), the black-throated blue warbler fared only
slightly better than the combination of AMRE and CSWA in extensive northern hardwood
areas in landscapes where total forest cover was high ( >80%) and where there was good
shrub development.

The interaction of NHWDIMIL x CAN OPEN x EDGE was highly significant for
BTARCS (p= 0.005). Of the 8 possible combinations of NHWDIMIL, CANOPEN,
EDGE, 5 groups consisted of more than 10 bird census stations (Table D32). The higher
level of EDGE favored the edge bird species when combined with the higher level of
CANOPEN (American redstart density = 84.66 birds/km2 for EDGE >2000m,
CANOPEN >15%, n=l4; BTARCS = -50.09 ). EDGE was not associated with a high
density of American redstart when canopy opening was low (CANOPEN <15%)
(American redstart density = 31.28 birds/km2 for Edge>2000m, CANOPEN<15%; and
BTARCS= 30.05 for that combination of edge and canopy opening). The difference in the
variable BTARCS was significant (p=0.05; Bonferroni post test of pairwise comparisons)
between the group NHWDIMIL>30; EDGE >2000m; and CANOPEN <15% (n=31) with
the group NI-IWDlMIL =>30; EDGE= >2000m; CANOPEN >15% (n=14). The effect of
a high level of edge and a high level of canopy opening in increasing edge species was
even more accentuated with the combination of the high level of shrub layer development.
For the group NHWDIMIL >30; EDGE >2000m; CANOPEN >15; WSDC<0.5 >50,
the American redstart reached its highest density of 133.68 birds/km2 (n = 8, se =16.71 ),
and BTARCS its lowest value of (~127.25, w 36.72 ) with respect to all other 15 groups
of 4 factor combinations. The black-throated blue warbler density for this combination of

4 factors was 70.99 birds/km2 (se 31.66) . This density for the black-throated blue in the

130

combination of 4 factors NHWDIMIL >30; EDGE >2000m; CANOPEN >15;
‘WSDC<0.5'= >50, was lower than its density of 11 birds/kmz 8.75 for the same
combination of factors except for the lower level of EDGE although the (Bonferroni post
tests of pairwise comparisons was not significant). The density pattern that the black-
throated blue warbler exhibited in highly forest landscapes indicated that it does not avoid
edges and high canopy openings. A high level of canopy opening with the combination of
a high level of edge however increased the density of edge species to a level that they

surpassed the density of the forest interior black-throated blue warbler.

l3]

CHAPTER 4

DISCUSSION AND CONCLUSIONS

ECOSYSTEM LEVEL RELATIONSHIPS

Data analyses provided information on bird species associations with habitat
variables and on patterns of bird density and habitat variable differences among strata of
northern hardwood forest. Synthesis of this information permitted an ecosystem level view
of relationships between bird species populations and the northern hardwood forest and
among bird species. Ecosystem level properties affecting bird species in northern
hardwood forest and resulting perspectives on forest fiagmentation and forest bird species
conservation are discussed below.

It is impossible to reconstruct the chronology of density changes that the
blackburnian warbler and the black-throated green warbler underwent after much of the
forest was clearcut and a second growth forest replaced the primary forest. That this
change was detrimental to the blackburnian warbler and might have benefited the black-
throated green warbler was indicated by habitat relationships between the 2 species that
were apparent along a conifer-deciduous forest gradient and density patterns in primary
forest and secondary forest. A prospect that the black-throated green warbler will invade
blackburnian warbler habitat was suggested by its high selectivity for late successional

forest conifer trees in settled forest1 where it occurred in the absence of the blackburnian

132

warbler and at a lower density than it did in managed forest. The fact that the black-
throated green warbler reached its highest abundance in second-growth forest and not in
primary forest or older growth in managed forest (strata that had a higher conifer
component) suggested that the blackburnian warbler could have suppressed it in these
northern hardwood forest strata.

Although land type associations that replicated managed forest and settled forest
categories of land use were uniformly low in hemlock and hemlock-white pine that
represent late successional forest, they varied with respect to total amount of conifer and
to amounts of various other coniferous forest types (i.e. lowland conifers and pine forest).
Landscape level efl‘ects relative to amount and types of coniferous forest at a land type
association level were somewhat evident for the black-throated green warbler. In
managed forest, the black-throated green warbler showed consistent higher densities
(although not always significant) in the Rapid River landscape perhaps reflecting the
higher amounts of lowland conifer forest types. The rarity of late sucessional forest
conifers (at local sites) and the associated low encounter rate of the blackburnian warbler
precluded the analysis of the effects of different landscape level conditions on its density.
The density pattern that the blackburnian warbler exhibited across forest disturbance
periods (Tables D11, D15 ) clearly demonstrated that it was not affected by fi'equent small
scale forest disturbance that is representative of canopy and shrub layer dynamics 2. An
explanation why the blackburnian ’warbler was not affected by selective logging is that
under present forest management (Hiawatha National Forest) patches of hemlock in
northern hardwood forest are preserved. Thus without the decrease in sites containing a

coniferous tree component, habitat availability at small scales would not be affected .

133

The increase in shrub layer development and densities of bird species (the black-
throated blue warbler, American redstart, chestnut-sided warbler, scarlet tanager, and
veery) associated with a shrub layer during time intervals of 6-12 and >12-15 years after
logging indicated the close association between the dynamics of the shrub layer and bird
populations. There are indications fiom density patterns that this increase in bird density
following logging could be responsible for their greater densities in managed forest
relative to primary forest. Higher densities of the black-throated blue warbler, American
redstart and chestnut-sided warbler in unlogged second grth and older grth forest
(Figure 14), and higher densities of all 5 species (black-throated blue warbler, American
redstart, chestnut-sided warbler, veery and scarlet tanager, relative to primary forest
(Figures 14, 15) that could not be accounted for by a shrub layer development suggested
that habitat packing was occurring in the managed forest3 .

Although all species associated with a shrub layer responded to the increase in
shrub layer development following logging in managed forest, the density pattern of each
species across all northern hardwood forest strata was unique. The bird species that were
associated with a well developed shrub layer differed with respect to the degree to which
they were affected positively or negatively by percent canopy opening and number of
canopy gaps. If these species were ordinated along a gradient that combined canopy
cover (at local site) and forest cover (at a landscape scale), the black-throated blue and
scarlet tanager would receive highest scores, the veery and American redstart intermediate
scores and the chestnut-sided warbler lowest scores on this gradient. The increase in
scarlet tanager and black-throated blue warbler between 6-12 years and >12-15 years after

logging relative to all other strata of northern hardwood forest indicated that poor shrub

134

layer development was a limiting factor to shrub layer associated species in the interior of
the forest. The increase in bird species that were favored by settled forest conditions
within these strata after logging suggested that conditions were created that simulated
settled forest conditions within the managed forest.

Significant difl‘erences in mean densities of shrub associated forest bird species
were common among land type associations replicating a land use or a forest disturbance
period. These difl‘erences reflected difi‘erences in shrub layer development. A question
pertinent to forest management would be whether difi‘erences in the dynamics of the shrub
associated bird species among different landscapes4 reflected some inherent characteristics
of the landscape or whether they stemmed fi'om human induced disturbance. Increases of
the American redstart and chestnut-sided warbler relative to the black-throated blue
warbler after logging in the Munising and Manistique landscape as compared to the Rapid
River landscape could be explained by their higher densities prior to logging. There were
no known (previous) disturbance factors that would account for higher densities of edge
species in either landscape. There were indications however that habitat conditions in
general, in strata other than just those describing conditions following logging disturbance,
were more favorable to these species in the Munising and Manistique landscape. The
American redstart and chestnut-sided warbler had higher densities in forest disturbance
strata >12-15 years and in unlogged second growth in the Munising and Manistique
landscapes than they did in the Rapid River landscape. Although the Munising landscape
overall had higher forest cover than either of the Rapid River and Manistique landscapes
(many bird census stations in the Munising landscape had forest cover within a 1.6 km

radius >90%, Table D28) which would favor the black-throated blue warbler, a more open

135

forest canopy (in all forest disturbance period strata) and, a greater number of large
canopy gaps, along with a better shrub layer development in all forest disturbance strata
might have allowed edge species to persist in between logging events. [ More favorable
habitat to the American redstart and chestnut-sided warbler in the Manistique landscape
was due to a higher amount of edge and lower forest cover in that landscape (Table
D28)].

The greater number of large canopy gaps (100m2 and >100m2) in forest
disturbance strata depicting recent logging activity (1-5 years; 6-12 years and >12-15
years) would at first suggest an efl‘ect of logging activity. Number of canopy gaps across
these 3 forest disturbance strata however did not show a decrease in successive forest
disturbance periods following logging as would be expected if the number of large canopy
gaps was due directly to logging. The number of large canopy gaps also difi‘ered among
land type associations within strata representing the same forest disturbance periods,
possibly reflecting a difl‘erence in local climate. The higher number of gaps in the strata
several years after logging (6-12 and >12-15 years afier logging) could suggest an
interaction of logging disturbance with weather activity in a process that involved standing
trees lefi afier the removal of others becoming more susceptible to toppling. Although
large canopy gaps were also numerous in primary forest landscapes they were not
associated with high densities of the American redstart and chestnut-sided warbler. As
was shown in the Anova analysis of the efl‘ects of interactions between edge, canopy gaps,
shrub layer development and forest cover, large canopy gaps alone were not associated
with an increase in forest edge species. Although the Munising landscape did not have any

more edge than the Rapid River landscape (it had lesser edge than the Manistique

136

landscape), that amount of forest edge combined with a better shrub layer development
and a more open canopy created more favorable conditions for edge species. The high
density of black-throated blue warbler in the Rapid River landscape several years after
logging could be due to habitat saturation. In the Rapid River landscape, shrub layer
development was poor in all forest disturbance periods other than that of 6- 12 years
following logging’. The amount of northern hardwood forest was also limited in this
landscape. [In this ridge swale land type association, the northern hardwood forest
occurred on ridges surrounded by lowland conifer; the land type association itself was

surrounded mostly by others that lacked northern hardwood forest].

ASPECTS OF FOREST FRAGMENTATION IN THE NORTHERN HARDWOOD FOREST

Examination of bird species densities along a range of forest conditions that
included aspects of forest fi'agmentation, provided perspectives on habitat fragmentation
fi'om the end closer to intact ecosystems than those resulting fiom studies of highly
disintegrated landscapes. The range of forest conditions that was herein spanned, included
areas in which the northern hardwood forest had become separated fi'om other forest by
openland. Although among the 10 bird species considered several had declined and 2, the
blackburnian warbler and the black-throated blue were largely absent from landscapes in
which the forest was patchy, isolation was ruled out as an underlying cause of this decline
on several grounds. All study sites in settled forest landscapes were close to the core of
the Hiawatha National forest; many were less than 1 km away fi'om contiguous northern
hardwood forest. Since several forest interior bird species were abundant, there was no

cause to believe that the dispersal ability of bird species that showed high densities difi‘ered

137

fi'om that of bird species that had declining densities. Another reason why dispersal was
not an issue is that the black-throated blue warbler and the blackburnian warbler already
had low densities under certain forest conditions within the managed northern hardwood
forest.

Often cited causes of decline of forest bird species in fiagmented landscapes are
increased levels of predation and cowbird parasitism (Donovan et al. 1995). In the eastern
deciduous biome declines of forest interior bird species with fiagmentation have been
blamed on their life history traits of nesting on or close to the ground in an open cup nest.
Several bird species that remained abundant in settled forest or that increased, the
ovenbird and veery nest on the ground and the American redstart and chestnut-sided
warbler most often nest in the sapling layer close to the ground. The mostly absent
blackburnian however nests in the canopy of conifer trees and the red-eyed vireo that
showed declining densities in settled forest (although still abundant) nests in the canopy of
deciduous trees. From these density and life history patterns there was no basis to suggest
that either predation or nesting habits were major causes for bird species density declines.

Several studies have associated loss of habitat elements ( i.e. microhabitats) fi'om
forest fiagments as the cause of decline in the number of bird species (F reemark and
Merriam 1986). The fact that the black-throated blue warbler was nearly absent in settled
forest but the equally shrub layer associated American redstart and chestnut-sided warbler
occurred at high densities suggested that the 2 latter species rather than a lack of habitat
might have kept the black-throated blue warbler out of settled forest landscapes. The near
absence of the black-throated blue warbler from settled forest landscapes however

precluded the examination of any competitive interactions among the 3 species within this

138

land use category. The relationship among these species however was revealed by their
relative densities within a higher range of forest cover in primary and managed forest
landscapes. Mthin a forest cover range of >70-90%, there was enough variability in the
northern hardwood forest with respect to forest cover, edge and canopy gaps, to
significantly affect the relative densities of American redstart, black-throated blue warbler
and chestnut-sided warbler. The fact (shown in the results of difl‘erent analyses), that
densities of black-throated blue warbler were lower in the presence of high densities of
the American redstart and chestnut-sided warbler suggested that an interaction was
occurring among these species.

It has been suggested that bird species select habitat at different spatial scales
(Gutzwiller and Anderson 1987; Steele 1992). It is possible that the black-throated blue
warbler avoids landscapes with reduced forest cover as an inherent habitat selection
behavior since these landscapes would have higher densities of edge and edge-interior bird
species. Given the proximity of settled forest to contiguous forest in this case, it is more
probable that the black-throated blue warbler was kept out of suitable microhabitat
patches within settled forest landscapes through competitive interactions with the 2 other
species that prevented it fi'om settling‘. The high densities of edge and habitat generalist
bird species in these landscapes could discourage the black-throated blue warbler from
establishing territories’.

The blackburnian warbler Was already rare in most of the managed forest even
though this was contiguous forest and landscapes varied between 70- > 90% forest cover.
The near absence of this species fi'om settled forest could not be assessed in terms of loss

of forest cover or forest fragmentation because of the confounding factor of loss of late

139

forest maturity conifer. If Lord and Norton’s (1990) definition of habitat fiagmentation as
simply the disruption of continuity at any spatial scale is considered, loss of late forest
maturity conifers would be a habitat fragmentation aspect because this loss would increase
distances between sites that have suitable habitat and the chance that some of them will
not be discovered. There were aspects in the relationship between blackburnian warbler
density and the occurrence of late forest maturity conifers that suggested that there were
no effects beyond those of habitat availability. The fact that densities of the blackburnian
warbler were proportional to amounts of late forest maturity conifers in difl‘erent northern
hardwood forest strata (regression explained >80% in blackburnian warbler variability)
suggested that the population was in equilibrium with its habitat, that is the occupancy rate
of habitat was even across the managed forest. The similarity of older growth forest and
primary forest sites with respect to blackburnian warbler densities that reflected their
similarity in the amount of late maturity conifers suggested that the more discontinuous
habitat in managed forest did not reduce the blackburnian warbler’s ability to locate
suitable sites.

There are several arguments that could be contemplated as to why with this severe
loss of hemlock and white pine there was no evidence of fragmentation efi‘ects. One
consideration is that both white pine and hemlock are long lived and slow growing
species, such that in the absence of cutting, many decades would have to pass before there
would be any change in their availability. Following the initial selective logging of white
pine and hemlock which reduced the types in the present managed forest, the availability
of hemlock or old pine in the landscape remained constant under forest management

practices that preserved these species within the northern hardwood forest. In contrast to

140

the shrub nesting species that have to track constantly changing patches of shrub growth,
the blackburnian faces less stochasticity in its search of late forest maturity hemlock and
white pine.

In forested habitat the blackburnian could also be utilizing other coniferous forest
patches which would allow it to persist in the landscape“. In contrast to more simplified
habitat in which most structural components would have been already lost and the species
might not find a difi‘erent patch type as a substitute (Karr 1982). In forest areas different
types of coniferous forest patches would have different habitat quality but would still
'provide alternative habitats for some individuals which could then establish in the higher
quality habitat when these are vacated. In highly simplified ecosystem this gradation of
habitats might not exist.

The presence of competitor bird species appears to be an important factor affecting
the density and distribution of the black-throated blue warbler. Because the shrub layer
was highly dynamic, by sampling forest at different time intervals after a logging
disturbance, a temporal dimension for the densities and relative densities of the black-
throated blue, American redstart and chestnut-sided warbler could be detected. Thus the
relationship could be viewed along temporal and spatial dimensions. Although the black-
throated green could be a potential competitor for the blackburnian", it was impossible to
see their interactions in either temporal and spatial dimensions. How the species relate to
each other along a temporal scale was precluded because of the length of time it takes
conifer forest to establish. There was no sampling procedure that would allow
observation of habitat dynamics, and the effects of the surrounding landscape on the

interactions of the 2 species within the short time span of this study”.

141

The most abundant bird species in managed forest landscapes, the red-eyed vireo,
least flycatcher, ovenbird and black-throated green warbler, remained abundant in settled
forest landscapes although the black-throated green warbler and red-eyed vireo showed
statistically significant declines. The ovenbird, red-eyed vireo and black-throated green
warbler can be considered forest interior bird species (Ambuel and Temple 1982, 1983;
Lynch and Whigham 1984; Askins and Philbrick. 1987; Robbins 1989; Askins et al. 1990)
showing declines in regions where the forest has been severely reduced or fi'agmented
(Wilcove 1985; Askins and Philbrick 1987; Hill and Hagan 1991). Because density alone
is not sufficient to indicate the productivity status of the population or reflect habitat
quality (Van Horn 1983 ), other population parameters would need to be sampled before
habitat suitability for forest interior bird species in settled forest can be assessed. Because
of the proximity of settled forest landscapes to managed forest landscapes, it is possible
that high densities of these species were in part due to overflows from the more
contiguous managed forest. Another possibility would be that high densities were due to
habitat saturation. Population levels in primary and managed forest landscapes could have
been at saturation levels and all other landscapes would have poorer habitat quality for
forest interior bird species.

The ovenbird, red-eyed vireo, and black-throated green warbler are associated
with a closed campy forest that characteristically has poor shrub-layer development. This
association might explain their ability to persist in closed canopied northern hardwood
fi'agments in the settled forest. This type of habitat is abundant in the interior of second
growth northern hardwood patches in settled forest. Shrub nesting bird species such as the

black-throated blue warbler, have to depend on some form of canopy disturbance to open

142

it and stimulate the development of the sapling layer. In situations in which edge and edge
-forest interior bird species are abundant such as occurred in the settled forest, edge bird
species would be the species to first preempt the newly available habitat in the interior of
the forest. Under closed canopy conditions, selected by the ovenbird, red-eyed vireo and
black-throated green warbler, the interior of the forest patch is protected fi'om edge bird
species simply because these species do not find their necessary microhabitats under a
closed forest canopy. That is the assemblage of species that are edge and edge interior
birds that require a well developed shrub layer would not comprise any species that could
invade the closed canopy forest fragments.

In bird conservation studies, the concepts involving forest fragmentation have
evolved fi'om landscapes in which the forest is reduced to fragments in a matrix that is
nonforest. Models that have addressed the behavior of bird populations in such
fragmented landscapes were largely influenced by the theory of island biogeography
equilibrium proposed by MacArthur and Wilson (1967) in considering the spatial
arrangement of fragments (for metapopulation models) or the distance from source
populations, extinction and immigration rates as major factors affecting the persistence of
a species in these environments.

Forest ecosystems however are. subjected to periodic disturbance events that create
temporary conditions in which fiagmentation aspects could become elevated. In managed
forest the northern hardwood forest is often interspersed with early successional forest
that is periodically clearcut. \Vrthin the northern hardwood forest both selective logging
and natural disturbance affect canopy cover and canopy gaps. The response of species to

these disturbances would depend on their population levels at the time of disturbance (as

143

was indicated by the response of the black-throated blue warbler, American redstart, and
chestnut-sided warbler in different land type associations in managed forest). Population
levels in turn reflect the events and conditions leading up to the time of disturbance. There
is undoubtably an element of stochasticity associated with the coincidence of population
levels of species with logging disturbance that would afi‘ect the outcome of which species
will preempt the sudden availability of habitat. Disturbance patterns and processes that are
characteristic of a specific environment however affect the frequency with which certain
events occur and their coincidence with other events. Conditions and events that tend to
favor one species at the expense of another would themselves become characteristic of
certain landscapes. In forested ecosystems the assemblages of bird species are more
complete such that species overlap in their habitat. Bird species that overlap in their
habitat can persist in the northern hardwood forest because of heterogeneity of
landscapes, and complexinteractions among dynamics of landscape characteristics, habitat
availability, bird species populations and relationships among species. A directional change
(one that persists) applied over a large part of an ecosystem would move it in a direction
of forest conditions that could favor one species (which also could be a forest associated
species) over another.

It should also be noted that once a disturbance event has unfolded, it can lead to a
difi‘erent scenario. For example if a disturbance by logging introduces an edge species in
the interior of the forest then the next cycle of logging might further increase the
population of edge species especially if the species is able to persist until the next
disturbance event. The fiequency of canopy disturbance and the time it takes population to

decline afier suitable habitat produced by canopy disturbance begins to decline are likely

144

factors determining whether edge species populations would remain in the interior of the
forest at the beginning of another cycle.

In the second growth northern hardwood forest, the cycle of periodic logging
started fi'om habitat conditions that consisted of closed canopy, low edge and low shrub
layer development. Although the density of the black-throated blue warbler would be
limited by the lack of a good shrub layer development, the species would be protected
from forest edge species. In the first cycle of logging shrub layer development would favor
the black-throated blue warbler. Once edge species have established in the interior of the
forest after the initial disturbance, it is likely their densities would build up with successive
logging cycles until their densities and the density of the black-throated blue warbler reach
an equilibrium. Because at this time most the second grth northern hardwood forest has
only received one cycle of logging it is likely that the relative densities of the black-
throated blue warbler, American redstart and chestnut-sided warbler have not yet reached
a state of equilibrium. Even if the intensity of logging and the amount of forest remain
constant, the relative densities of these 3 species will likely change in favor of edge
species.

Not all aspects of change in forest conditions and subsequent declines in species
involve forest fiagmentation. The reduction in late forest maturity conifers is a long term
efl‘ect that cannot be reversed without a well directed long term effort aimed at
reestablishing this component baCk into the second growth northern hardwood forest. Yet
in today’s forest, this loss of the conifer component is not associated with a forest
fi'agrnentation dimension. The discontinuity in habitat of the blackburnian has occurred

along a temporal dimension and not along a spatial one.

145

At this time vagueness still surrounds the concept of forest fragmentation
(F aaborg et al. 1995) with controversy focusing on what biological processes to link to it
(Fahrig 1997). Because habitat loss must necessary occur for habitat to become
fiagmented, debate has revolved on whether fi'agmentation should be restricted to efl‘ects
that occur beyond the effects of habitat loss or whether fragmentation should also include
habitat loss (Haila 1986, Fahrig 1997). Such debate is considered necessary because
conservation strategies that would need to be implemented would difi‘er between the 2
different premises.

Some contentions about fragmentation would be relaxed if fi'agmentation was
considered a process that had a temporal dimension. Fragmentation aspects in the northern
hardwood forest which consisted of factors operating at different spatial scales suggested
that fiagmentation should be considered a complex process that involves different spatial
scales and progresses with time to difi‘erent levels. As a process fi'agmentation could be
followed from initial stages starting with a highly forested system and would end in a
highly fragmented ecosystem. This contrasts with most current conceptual models of
fiagmentation that have been developed for the end state of fiagmentation and have had to
progressively include more complexity to increase their representation of reality. Highly
forested ecosystems, even when they are no longer representative of pristine forest
conditions, as was herein described contain a compliment of forest bird species, and a
diversity of landscape structures, far greater than what is typical in the more simplified
landscapes consisting of woodlots irnbedded in a landscape of nonforest. As was described
for the black-throated blue warbler, processes that affect the decline of a bird species

begin at such fine spatial scales that they are not detected as involving forest fragmentation

146

effects until the process has advanced to a level at which it becomes detectable at larger
scales. At what point in the disintegration process would the term fi'agmentation be
applied is a matter of definition. At initial stages, only a slight change in habitat conditions
can cause an edge species to begin preempting habitat that previously was occupied by a
forest interior bird species. Thus the replacement of a forest interior bird species by
another species (a process associated with fragmentation) can occur before a forest
fiagmentation process is recognizable.

Lord and Norton’s definition of discontinuity of habitat that could occur at any
scale overlaps the effects of habitat loss because any habitat loss would result in some
discontinuity. Discontinuity of habitat in a spatial dimension would only have biological
significance however if it starts to exceed the species’ dispersal ability, or the increase in
search time starts to impact its productivity. The efl’ect of habitat discontinuity in a spatial
dimension alone was not evident for any of the species that were declining across strata of
northern hardwood forest. Discontinuity in a temporal dimension could be viewed from
the perspective of habitat becoming more suitable to a competitor species for a long
enough time that the competitor gains ground over the original species.

That many changes in deciduous forest ecosystems can appear to be related to
forest fi'agmentation stem from the fact many bird species in this biome are associated with
shrub layer development and their densities affected by shrub layer dynamics. These
species also show habitat separation along a gradient of forest cover. However species can
begin to decline before forest cover is lost. It should be evident that the multitude of
events that lead to change in the relative densities of bird species and the eventual local

extinction of species cannot be recovered fi'om the stage of a fragmented landscape as was

147

indicated in the case of the blackburnian warbler and black-throated blue warbler within
settled forest. This aspect is one of the strongest arguments for the need to shift the
emphasis in bird conservation from fi'agmented landscapes to forest ecosystems that are
still functioning as forest ecosystems and to understand how these systems can disintegrate
along temporal and spatial dimensions. It follows that while predation and cowbird
parasitism could be important factors afi‘ecting the viability of forest bird species in highly
fragmented landscapes, it is presuming to suggest that it is these factors that have caused
the disintegration of the system and are the underlying cause of their current population

levels.

BIRD CONSERVATION AND MANAGEMENT CONSIDERATIONS

1. Function of Primary Forest in Present Northern Hardwood Forest Ecosystems

The shortcoming of not having baseline data for forest disturbance processes, bird
densities and habitat relationships among species, is the absence of models against which
present day forest disturbances and forest temporal and spatial patterns can be assessed
with respect to bird species adaptations. It is against such baseline data that directional
change resulting fi'om human induced disturbance can be assessed. There is a temptation
to consider primary forest as a control to which human impacted forests can be
compared. The amount of variability that was herein detected among northern hardwood
landscapes within a single category of land use warns however against accepting
differences between primary forest and secondary forest landscapes to be solely

directional changes.

148

As results of comparisons of landscapes within a category of land use revealed,
landscapes differed with respect to many small scale habitat elements within the northern
hardwood forest. Difi'erences in the dynamics and densities of the black-throated blue
warbler, American redstart and chestnut-sided warbler among the 3 land type associations
replicating managed forest were previously discussed in detail. The Munising landscape
for example was more similar to primary forest landscapes with respect to number of large
canopy gaps than it was to the Rapid River and Manistique landscapes. The amount of
edge differed among primary forest landscapes, while it was non- existent in the
McCormick Wilderness Area, shoreline in Sylvania and the Huron Mountains produced
edge habitat. It should be evident that landscapes could vary in a multitude of ways
because a multitude of habitat conditions can be produced by the interaction of different
features that are inherent to geophysical characteristics of landscapes and disturbance
activities superimposed over them.

Ecological classification systems such as the regionalization of Michigan
landscapes (Albert et al. 1986; landscapes described by Pregitzer et al 1984; Simpson et al.
1990, Spies and Barnes (1985) and the land classification systems underway on the
Hiawatha National Forest and the Ottawa National Forest, reveal a high level of variability
among landscapes with respect to geomorphological features, and local climate. These
differences in geomorphological features also produce differences in natural disturbance
frequencies (Turner 1987, Simpson et al. 1990) which in turn affects the balance between
northern hardwood forest succeeding to hemlock vs its persisting as a sugar maple
dominated forest. For landscapes dominated by northern hardwood forest, the fiequency

of disturbance will determine the relative composition of the landscape in terms of late

149

maturity conifers or more sugar dominated deciduous forest". Because only a tiny
fraction of the area formerly occupied by the original northern hardwood forest is left in
the Upper Peninsula and across the Great Lakes Region, the range of variability that
existed in the primary forest among landscapes has been reduced. Thus to infer that
management induced changes have occurred from general comparisons between managed
and remaining primary forest would be inadequate because the fiill range of variability of
the primary forest would not have been captured.

Several facts can be used as arguments that the primary forest does not represent
pristine conditions”. From a regional level (scale of the whole Upper Peninsula or
Northern Great Lakes) a well documented change is the greater interspersion of forest
with openland; a directional change that has favored edge species. The density of the
white-tailed deer has increased many fold in the Upper Peninsula of Michigan after the
large scale clearcutting and the growth of the secondary forest (Doepker et al. 1995).
Deer in the Upper Peninsula of Michigan concentrate in hemlock forest during the winter.
Because distances that white-tailed deer in the Upper Peninsula travel in their movements
fiom summer range to winter yarding areas can be upwards of 30 km (Van Deelen 1995),
all primary forest landscapes comprising extensive areas of hemlock forest are vulnerable.
Thus even though large areas (several square miles) were selected to represent primary
forest landscapes these would still be below the spatial scales of deer movements.

A large body of information exists concerning the impact of deer browsing in
primary forest on the shrub layer (Anderson and Loucks 1979, Alverson et al. 1988, Miller
et al. 1992). Heavy impacts of deer browsing in the Sylvania Recreational Area have been

a long time concern (Webb et al. 1956). Directional changes in forest conditions that have

150

afl’ected shrub cover in primary northern hardwood ecosystems would have negatively
afi‘ected the black-throated blue warbler and other shrub nesting bird species. Early
accounts of the nesting habits of the black-throated blue warbler describe it as a
specialized woodland bird (Pearson 1936). Bent (1963) also described it as commonly
nesting in Canada yew in deep mature woods. Among the plant species that have nearly
been decimated from changes in forest conditions is Canada yew (Braun 1950). The large
scale decimation of Canada yew has been associated with changes in ecological conditions
with the establishment of secondary forest following large scale clearcutting (Braun 1950).
In primary forest however, deer browsing has been blamed for vegetation structure and
compositional changes in understory vegetation including a broad scale decimation of
Canada yew (Braun 1950, Anderson and Loucks 1979, Allison 1990).

An historical reduction in the shrub layer within the primary forest implies that
densities of shrub layer associated bird species that were herein reported for the primary
northern hardwood forest would not be representative of their former densities in pristine
forest (former primary forest before the large scale ecological changes). The association
of these species with the sapling layer that is largely renewed by logging in managed forest
might also be a phenomenon that is part of the directional change that the northern
hardwood forest has undergone after the large scale clearcutting.

Regional declines of forest interior bird species would affect their immigration
rates into the primary forest. CoOk (1904) described the black-throated blue warbler as
one of the most common migrants at several migration passage points in the US. Early
records describing the distribution of the black-throated blue warbler as abundant and its

range extending well into southern Michigan (Pearson 1936). The distribution of the

151

black-throated blue warbler has been shrinking in Michigan such that its present
distribution is mostly restricted to the northern part of the state (Binford 1991). The
reduction of the species distribution and early accounts of its nesting habits should caution
about describing this species niche and habitat relationships from its present distribution
and habitat. This point is important because not realizing the fiill potential of a species
habitat use and the extent of the different environmental conditions under which it existed
and associated with other bird species would limit management options because of
forgotten information.

In today’s forested landscapes primary forest landscapes adds to the total
heterogeneity of the northern hardwood forest by stretching the gradients of forest cover
at the landscape scale, amount of total late maturity conifers and forest edge (lower end).
Significantly higher amounts of late forest maturity conifers presents the greatest
opportunity to reserve habitat for the blackburnian warbler. Long term viability of the
black-throated blue is most probable in primary forest because a high forest cover and in
general lower amount of forest edge provide the greatest safeguard against edge species.
Even when shrub layer development is limiting, the black-throated blue warbler will occur
at low density but will be protected from high densities of edge species.

While there are limitations to what the primary forest could reveal about historical
changes in the relative densities of bird species, primary forest landscapes can be setup as
controls to monitor ongoing changes in the forest. Since less human disturbance occurs
within primary forest reserves, long term monitoring of these landscapes would elucidate
how changes in forest conditions at regional levels (or in surrounding landscapes) would

affect species inflow into them and impact bird densities.

152

2. Function of the Managed Forest

Major characteristics of the managed forest that were uncovered by comparisons
with primary forest and settled forest were related to shrub layer dynamics. Logging
disturbance was a dominant organizing force in second growth forest at the level of a land
type association that affected densities of shrub layer associated bird species. The efl‘ects
of land type association acted in conjunction with logging disturbance to afi’ect the relative
response of bird species and the patterns of their response following logging. The fact
that densities of both forest interior and area sensitive species as well as edge associated
species peaked several years following logging disturbance revealed as was previously
discussed that shrub layer development was limiting forest interior bird species densities in
managed forest while logging disturbance enhanced the suitability of the interior of the
forest to edge related bird species. Thus the managed forest can be looked upon as
providing an environment that blended habitat suitability for edge species with that of
forest interior bird species.

Heterogeneity of forest conditions in the managed forest became apparent through
contrasts of habitat variables and bird species densities across its different forest
disturbance period and land type associations strata. Heterogeneity of forest conditions
with respect to how the black-throated blue warbler, American redstart and chestnut-sided
warbler perceive the environment however became only apparent through contrasts of
these species densities among strata that consisted of defined ranges for a combination of
factors that included canopy opening, shrub layer development, length of forest edge and
forest cover at a landscape basis. The fact densities of each of these 3 species differed

significantly (at 0.05 level) among the strata delineated based on these specified criteria

153

established that heterogeneity in forest conditions along these derived dimensions was of
biological significance to a forest interior bird species and to its edge species potential
competitors. The strata were not represented by geographically defined areas as
stratification by land type association would be. The relative contribution of difi’erent
geographical locations of the forest to these derived strata was not determined by this
study". The spatial pattern of different types of patches that define difi‘erent forest
conditions (that are difi’erently suitable to the black-throated blue, American redstart and
chestnut-sided warbler) thus could not be viewed spatially. A number of situations
establish that the grain of such patchiness would be finer than the patchiness associated
with the forest patch delineated based on the last logging operation event. Canopy
reduction (one of the variables affecting habitat suitability that differently affected edge
and area sensitive species) by selective logging produces finer patchiness than the
delineated area for selective logging . In selective logging either single or several trees are
removed from a small area. Thus because canopy reduction is patchy, bird census
stations from within the same logged site would fall within different strata of canopy
opening. Point samples from within a single site would also fall in different strata
delineated on the basis of a mount of forest edge (within 500m) depending on whether
they were close to forest edges or a great distance away. Forest cover patchiness was
coarse grain in that it included patches of different forest types. Bird census stations from
the same sampling site were frequently included within the same stratum defining an
interval of forest cover.

The managed forest that was herein considered consisted of northern hardwood

ecosystems within the core of the Hiawatha National Forest. The need to consider

154

managed forest as an integral part of lands available for the conservation of forest bird
species stems from the fact that most forested areas in the US. are managed. The range of
forest cover that was available within the managed forest on the Hiawatha National Forest
would likely overlap ranges of other managed forest areas across the northern Great
Lakes. The heterogeneity in forest conditions encountered within the managed forest
included ranges of forest conditions that were suitable to bird species that declined in
settled forest. This confers upon the managed forest a potential for the conservation of
forest interior and bird species that like the blackburnian warbler require specific habitat
elements or like the black-throated blue warbler require large forest areas. The
heterogeneity of managed forest however also increases the risk for these same species
because it exposes them to increased levels of competition fi'om more habitat generalist
species. For the blackburnian warbler, the managed forest presented more suitable habitat
to the black-throated green warbler. With respect to the black-throated blue warbler the

managed forest presented suitable habitat to edge bird species.

3. Viewing the Northern Hardwood Forest as an Open System vs a Closed System, at
Equilibrium or Undergoing Directional Change.

A hierarchical approach was used to organize the heterogeneity that exists in the
northern hardwood forest. This northern hardwood forest ecosystem however is itself
imbricated within higher level systems along with other ecosystems with which it
exchanges materials, by which it is affected and which in turn it affects (see Tansley 1935,
Reiners 1983, Delcourt et al. 1984, Noss 1992). Viewing the northern hardwood forest as

an open system that exchanges populations with other hardwood forest and other forest

155

systems has very different consequences with respect to bird conservation than viewing it
as a closed system which exists by itself. In an open system view, the value conferred
upon the northern hardwood forest depends upon its capacity to produce an overflow of
species that have declined in other ecosystems. Thus the attributes that increase the
capacity of the system to produce such species would be enhanced. Management of the
ecosystem would seek to keep or restore a greater part of the system to a condition that
can produce these species. In a closed system view, the emphasis would be in maintaining
the greatest diversity of species. Forest management would emphasize enhancing
heterogeneity in the system through human induced disturbance. This heterogeneity
would include providing a range of conditions that would comprise the habitat suitability a
large number of species. Differences in the 2 approaches can be easily illustrated with
respect to the black-throated blue warbler that has declined in many ecosystems. In an
open ecosystem view, properties of the ecosystem that enhances the protection and
productivity of the black-throated blue warbler such as reduction of edge, maintenance of
a low canopy cover, and high forest cover would be applied over most of the system. In a
closed view of the northern hardwood forest ecosystem, heterogeneity of forest conditions
would be maintained to enhance all species even if this consisted of applying disturbance
levels and disturbance frequencies that deviate from intensities and frequencies of naturally
occurring events.

The northern hardwood ecosystem was herein described within a minuscule time
span (90+ years) with respect to the time it took the system to develop (see Davis 1981,
Pielou 1991, Tyrell and Crow 1993 ). This time frame however was suficient to provide

some insight into some dynamics, patterns and structures within the ecosystem. It can be

156

inferred that the system is progressing into the firture, but the direction it is taking to some
extent is determined by the structures, patterns and firnctions of present forest.

There are thus different ways to view a northern hardwood forest along a temporal
dimension. Viewed on a short term basis management would only be concerned with
immediate changes in the ecosystem. Efi’ects of disturbance imposed by management
would be assessed by monitoring changes at the individual project level by assessing the
effects just prior and just after disturbance events. On a short term management basis little
consideration is given to the state of the ecosystem decades into the future. It is evident
that the assessment of disturbance effects would differ greatly between the 2 methods. In
the short term basis any disturbance imposed by logging would appear to have a positive
effect. The black-throated blue warbler would appear to be positively associated with
disturbance of logging. However as has been described of its interactions between the
American redstart, chestnut-sided warbler over the long term, the black-throated blue
warbler might be loosing ground to the edge tolerant species. The long term effect of
applying frequent disturbance on the northern hardwood forest would depend on the
interaction of many factors and cannot be known from simple monitoring of before and
after monitoring. Only modeling of population dynamics, habitat dynamics under different
assumptions of forest disturbance level and under difi’erent assumptions of spatial and
temporal arrangement of disturbance events and landscape conditions can a view emerge

of what the long term future condition of the ecosystem would be like.

157

4. Aspects to Consider in the Development of Bird Conservation Strategies

1. A first step to the conservation of forest birds in the northern hardwood forest is
to establish which bird species should receive management priority. Bird species that need
to be prioritized are those that have suffered population declines from known directional
changes in the northern hardwood forest ecosystem that have occurred at a regional level.
The blackburnian warbler and the black-throated blue warbler should receive priority in
the herein described northern hardwood forest ecosystem .

2. Assess the suitability of different land type associations comprising a component
of northern hardwood forest with respect to their potential for providing large areas of
habitat suitable for the long term sustainability of black-throated blue warbler and
blackburnian warbler populations. For the black-throated blue warbler, suitable land type
associations would comprise continuous forest that is dominated by northern hardwood
forest and older grth northern hardwood forest. For the blackburnian warbler land type
associations that contain large areas of older growth northern hardwood forest comprising
components of hemlock and white pine and that is intermixed with hemlock and hemlock -
white pine forest types would be most suitable.

3- Delineate core areas to provide areas for the long term sustainability of the
black-throated blue warbler and blackburnian warbler. In these areas management
strategies include the protection of black-throated blue warbler from edge species and
enhancement of the hemlock component for the blackburnian warbler. Within a land type
association, older grth northern hardwood forest areas and adjacent northern hardwood
forest areas should be delineated as core areas for both the blackburnian and black-

throated blue warbler. Early successional forest and forest openings within the core areas

158

(area should be in the extent of a 1,000 ha) should be allowed to succeed to northern
hardwood forest. Shrub layer development could be enhanced by light canopy disturbance
in areas adjacent to older growth, while the older growth itself would act as a bufi’er
against edge species. Enhancement of hemlock regeneration should occur within older
growth forest where the blackburnian is already present and in forest areas adjacent to
older growth. The reintroduction of Canada yew or its propagation should be encouraged.
Management guidelines for areas adjacent to older growth should include maintenance of
low edge, relatively low canopy disturbance (removal of single trees rather than group
selection and shelterwood methods), and longer intervals between canopy disturbance
events ( >20 years). Canopy disturbance should occur in small areas at one time (5-10
hectares) separated by larger undisturbed areas. Logging operations that covers 1003 of
hectare at one time should not be undertaken in these core areas.

4. Develop general forest management guidelines for all northern hardwood forest
and guidelines for all logging operations within northern hardwood forest. The
interspersion of early successional forest, pine plantations and forest openings close to
northern forest should be discouraged. Areas already in these states should be reverted
back to northern hardwood forest. Smaller less intensive logging operations that are
undertaken over a longer period of time should be encouraged over intensive logging
operations. The treatment of 105 of hectares every 5-6 years in an area of several hundred
hectares would be less likely to establish forest edge species within the interior of the
forest than operations that treat hundreds of hectares at once every 15 years. Smaller less
intensive operations spaced over a long time will maintain availability of a shrub layer in

the interior of the forest without providing the opportunity for edge bird species to invade.

159

Logging roads should be maintained to the narrowest possible widths and openings in
northern hardwood forest should be eliminated.

5. Develop sampling strategies to monitor long term changes in relative abundance
of edge species and the black-throated blue warbler in core areas. Sampling strategies
could also be used to evaluate densities of the black-throated blue warbler in core areas
compared to its densities in other northern hardwood forest areas not slated specifically
for its management. This sampling should however also be carried out over a long term
basis because the black-throated blue warbler densities are likely to peak and then fall
sharply with periodic logging. The long term relative response of populations of black-
throated blue warbler and American redstart with successive logging cycles cannot be
known from short term sampling. (It would take many decades for hemlock planting or
regeneration to begin enhancing habitat availability for the blackburnian warbler.
Monitoring of blackburnian warbler and black-throated green warbler populations
however could be used to assess fluctuations in the relative population densities of the 2

bird species which over time would indicate if directional change is occurring).

CONCLUSIONS

1- Bird species declines in a northern hardwood forest ecosystem can occur long
before the effects are recognized as forest fiagmentation (involving loss of forest cover).
When the degradation involves a habitat fragmentation dimension, the process can begin at
small spatial scales and does not emerge at larger scales until it has become relatively
advanced. In a forest ecosystem in which the assemblages of bird species is more

complete, directional change in forest conditions is mediated by interactions between

160

species. Species that are the most habitat specialist would likely be replaced by forest
species that are more habitat generalist.

2- One reason why degradation of the deciduous forest biome has been associated
with forest fragmentation might involve the association of many bird species with the
shrub layer and their habitat distribution along gradients of canopy cover and forest cover.
Changes of forest canopy cover and forest cover (at the landscape scale) exposes the most
forest interior bird species to edge species.

3- Bird species that are associated with a closed canopy, low shrub layer
deve10pment are less likely to be outcompeted by edge species. These species are likely
to persist in closed canopied second growth, northern hardwood forest fiagments as long
as the canopy remains closed.

4. Degradation of the system (loss of species or habitat elements) can be along a
gradient that is not related to forest fiagmentation as the loss of late forest maturity
conifers would be. For the blackburnian a discontinuity in availability of habitat along a
temporal scale is more limiting that the spatial arrangement of habitat fragments. A severe
loss of a habitat element that takes a long time to recover favors a more habitat generalist
species (e. g. the black-throated green warbler). The presence of high densities of a habitat
generalist might itself instill a further change in habitat conditions for the original species.

5. Because the northern hardwood forest is still recovering from the large scale
clearcutting at the turn of the century and because the cycle of logging of the second
growth forest is recent, the northern hardwood forest would not have likely reached an
equilibrium with respect to the relative densities of the black-throated blue warbler,

American redstart and chestnut-sided warbler. Long term effects of successive cycles of

161

logging in the northern hardwood forest on the relative abundance of these species cannot
be determined from short term bird censuses that are designed to sample densities just
before and just after a disturbance event.

6. The complex interaction of factors that create heterogeneity in the northern
hardwood forest should warn against predicting the firture state of bird species
populations at an ecosystem level from forest disturbance effects observed within a narrow

time interval and within a limited area.

162

FOOTNOTES

Footnotes for Chapter 1: Introduction

1. The term "northern hardwoods" is given to an association of many different tree
species that includes sugar maple, American beech, yellow birch, hemlock, red maple,
American basswood, white pine, black cherry, American elm, white ash, and balsam fir
(Nichol 1935, Graham 1941, Braun 1950, Tubbs and Goodman 1983). Characteristic
species however are sugar maple, American beech, yellow birch and hemlock.

2. A land type association is a land unit delineation in a hierarchical land classification
system adopted by the US Forest Service. At the level of a land type association,
delineations of the different units are based on local climate and landforrn. A more
complete explanation of land type association and how it fits into a hierarchical land
classification system is given in the chapter ‘Methodology’ under the heading ‘Hierarchy
of physical and climatic factors afi‘ecting forest heterogeneity’.

Footnotes for Chapter 2: Methodology

1. Wiens (1989) discussed how habitat relationships among species are dependent on
the spatial scale at which these relationships are being investigated. An example he gave
was of habitat relationships between the American redstart and the least flycatcher: a
competitive relationship appears at a scale of 4 ha and that of positive association appears
at larger scales (regional scales).

2. Schoener (1966) gives the size of feeding territories of several warbler species to
be a fiaction of a hectare. Morse (1976) gives a range of nesting territory sizes for black-
throated green warbler and blackburnian to be in the range of 0.2 -0.5 ha and for the
yellow—rumped warbler and magnolia warblers to be in the range between 0.7-0.9 ha.
Cody (1978) gives an approximation of the size of territories of old world sylviid warbler
species (counterparts to the parulid warblers in the new world) to be 0.5 ha.

3. The fine scale at which birds select habitat has been supported by many field

studies (Hilden 1965) in which removal of an individual fi'om its territory results in another
of the same species replacing it.

163

4. An area around a singing male bird was often used to approximate the location of
the nesting territory and to measure habitat selection by the species during the breeding
season (James 1971).

5. Anderson and Shugart (1974), and Rotenberry and Wiens (1980) found differences
between vegetation characteristics measured within sites occupied by individual bird
species and unoccupied sites within study plots that would be considered homogeneous in
vegetation structure and composition (at the study plot level). The spatial resolution at
which most songbird species select their habitat is finer than the spatial resolution used for
managing forest resources: that is birds can perceive patchiness within land areas that are
considered uniform for management purposes. The averaging of vegetation measurements
over a space that appears heterogeneous to bird species would mask vegetation
characteristics that individual bird species select and habitat difi’erences among bird
species. This point was used by “hens and Rotenberry (1981) to argue for the need to
understand species specific habitat needs at relevant spatial scales in the management of
bird communities rather than depend on bird community characteristics (species diversity
and richness) or broader based habitat characteristics.

6. The concept of a landscape in organism-based biology is relatively recent. The
concept of a landscape has long before been used in land classification. A landscape
concept has been recently introduced into forest management to be a level above the unit
of management which is the stand. A spatial scale in the range of 1005-10005 of hectares
presents a perspective of higher level patchiness. At this level the spatial configuration of
different management schemes and of different forest types (e.g. early successional forest
vs long-lived forest types, old-grth vs managed forest) emerges.

7. The importance of population sources have been best revealed by the
disappearance of forest bird species from woodlots in which the vegetation structure (at
the microhabitat level) had not changed but forest had been lost from the larger landscape
(Lynch and Whitcomb 1978; Askins and Philbrick 1987, Askins et al. 1987). Much of
what is known about dispersal and the greater landscape has been revealed from
landscapes in which the forest had become patchy invoking the theory of island
biogeography (McArthur and Wilson 1967). In forested regions, environmental resistance
and patchiness of source populations would be more dificult to assess because
environmental patchiness would be perceived difl’erently by difi‘erent species.

8. The spatial scale of a region as used here is at a level of a large part of a state
(level of a ‘Section’ in the definition of ecological units as presented in Ecomap (U SDA-
Forest Service 1993).

9. With the introduction of a hierarchy perspective (Allen and Star 1982),
heterogeneity in the forest could be assigned to different scales. Many bird community
studies were restricted to the level of a habitat and did not consider heterogeneity that
would be due to difi‘erences in the spatial configuration of habitat patches at broader
spatial scales. With the increase in spatial extent of more recent studies (covering difi’erent

164

regions) the need to consider a hierarchy becomes more evident. This is because the
relationships found within lower levels (habitat types) might be affected by higher levels in
the hierarchy. For example the relationship between a species and a habitat type might not
be constant across regions.

10. The relationship of 2 species can be examined by assessing the relative occupancy
of patches (territory level) of known characteristics by 2 species within a habitat type in a
landscape of known characteristics; then moving to a different landscape (of different
characteristics) and investigating the relative occupancy of same (territory level) habitat
patches in the same habitat by the 2 species. Alternatively within the same habitat in the
same landscape one could examine the occupancy of (territory level) patches of different
characteristics by the 2 species. In either case looking at the occupancy of the 2 species in
larger areas (larger than the scale at which discrimination between patches is occurring)
would mask difi’erences between the species. The 2 species will appear correlated ( to the
habitat type and not to individual types of smaller patches).

11. In the Robbins et al. (1989) study, of the 3 measures of % forest cover within 1
km, 2 km and 3 km fiom the point of bird censusing, the abundance of bird species were
most correlated with forest cover within 2 km.

12. Readily available delineations of the forest based on habitat availability of different
bird species at the rrricrohabitat level, and forest fiagmentation variables at both the
microhabitat and landscape level (here considered within 1.6 km and 500 m) do not exist.
Ideal delineations of strata that would be organism defined (that show either density or
population productivity differences) could start fiom the bottom up by classifying land
units the size of bird territories based on some vegetation structure characteristic
associated with a population parameter of that particular species. Adjacent small size units
that are similar with respect to the defined variable would be combined to form a patch. A
spatial pattern of like patches becomes apparent at some higher spatial scale. Different
patterns of patchiness could become obvious at increasingly higher levels of spatial scale.
It might be possible in this manner to relate the pattern of patchiness at one level to some
underlying environmental variables. For example the patches of coniferous trees within a
patch of northern hardwood forest patch could be related to topography and aspect

13. The efi’ects of forest changes on bird species have been described qualitatively in
various treatises (Bond 195 7, Brewer 1991) but these descriptive accounts do not provide
the quantification needed for strong causal inference.

14. The Upper Great Lakes region has only recently been affected by anthropogenic
activities and these have been minimal relative to human impacts in other geographical
regions in the world.

15. Landuse is an extraneous factor to naturally occurring environmental factors that

determine the hierarchical nature of ecosystems in the natural world as described by Allen
and Star (1982). It is only the naturally occurring environmental factors which have

165

provided the theoretical basis for the hierarchy in ecological land classification systems.
Land use however has a dominant influence in modifying the multi-scaled heterogeneity
that has developed naturally at this geographic location. Forested ecosystems in the
Northern Great Lakes provide unique opportunities to study the effects of human-caused
disturbance because of the large patchiness in which different land uses occur and because
of the fact primary forest still exists in several large areas.

16. The land use and land type association categories are large enough (in area ) that
bird density patterns that emerge at their scales would be expected to be affected more by
intrinsic factors (i.e. habitat suitability and population productivity) afl‘ecting bird species
operating within these patches and less by extrinsic factors (immigration from other source
areas). [Information from biological reserves indicates that the larger the area the less it
would be afi‘ected by outside processes such as bird immigration into the area (Blake and
Karr 1984; Lord and Norton 1990, Harris and Silva-Lopez 1992, Noss 1983, Ambuel and
Temple 1982, Robinson 1988)].

17 . Because forest landscapes have been drastically altered outside of forest reserves,
processes that occur over broad scales (e. g. deer migration) could still affect structure and
composition of the forest inside forest reserves.

18. Logging creates canopy openings in locations where trees are removed. In these
locations forest maturation processes are set back. The state of the vegetation structure
therefore depends on the time that had elapsed since logging. The patchiness created by
logging is laid over preexisting patchiness caused by other factors such as small scale soil
variation, topography, aspect and past disturbance. Factors that have contributed to
previously existing patchiness are likely to have different spatial scales than that in which
logging is undertaken such that the overlaying of cutting creates a variety of new
patchiness. Northern hardwood forest grows on mesic soils so it occupies those areas in
the land type association in which mesic conditions dominate. The patches of northern
hardwood forest within a land type association would match the level of ‘land type’ in the
hierarchy of ecological units (Ecomap-USDA Forest Service 1993). Patchiness at this
scale represents very fine variation in aspect, slope, soils, and plant associations. Small
scale patchiness within the northern hardwood forest across a land type association would
be similar because they occur on the same soil type, landform, rock type and are exposed
to the same geomorphic processes and local climate. For the same silvicultural treatment
(logging methodology) this small scale patchiness would only differ based on the time
elapsed after logging disturbance.

19. The term older growth is used to refer to forest that forest managers recognize to
have an age class and tree species composition that are more typical of the primary forest
than of the secondary forest that developed after intense logging or clearcutting. The term
older grth is however used rather than primary forest because it is not with absolute
certainty that the area was not logged and the intensity or selectivity of logging (other than
clearcutting) when it did occur is unknown.

166

20. Specific vegetation variables measured and methodology of sampling vegetation
are described under Sampling of Microhabitat Variables in the Methodology Section of
Field Methods.

21. A list of features that were considered forest edge is given in ‘Locating Bird
Census Stations’ in the section of Field Methods in the chapter Methodology.

22. The specific variables that quantified fragmentation at different spatial scales and
the statistical analyses used are discussed in ‘Etfects of Interactions of Habitat

F ragrnentation Related Aspects of Forest with Habitat Availability on Bird Species’ in the
Section Statistical Techniques in Methodology. Here only a broad overview of the study
design is presented.

23. The details of these analyses are given in Statistical Techniques section of the
Methodology chapter.

24. Alternatively some initial stratification of the sampling units could have been
retained and the sampling units from each of the strata post stratified separately to new
spatial scales. Sampling units from different strata of land use, managed forest, primary
forest and settled forest, or from difi‘erent land type associations could have been analyzed
separately. The individual land use would not provide the full range of values for some
fiagmentation variables; for example settled forest might not have any sampling units that
would fall in a stratum of high % forest cover. There would however be a management
interest in comparisons of difi’erent land type associations within a land use category.

Only the comparison of land type associations in primary forest would reveal the effects of
land type association on forest fiagmentation. For managed forest what would be revealed
by a comparison of land type association is the intertwined effects of forest management
with inherent land type association characteristics . The number of sampling units fi'om
each strata limited the possibility of carrying out separate analyses. Contribution of
difi’erent land uses to groups representing difi’erent ranges of fragmentation values at the 2
spatial scales can however be uncovered by assessing the contribution of each land type
association within a specific land use to the difi’erent groups.

25. The present delineation of stands has developed gradually. Through the years,
difi’erent foresters have made decisions on where to place stand boundaries. Aerial
photographs have been used to delineate boundaries of difi’erent forest types. Streams,
rivers, roads have been used to subdivide areas that appear homogeneous in a photograph
into smaller units. Because the delineation of stands have not followed consistent
methodology through the years, it is now difficult to derive a well set of criteria to define
how stands difi‘er from one another or to find a consistent spatial resolution that went into
the delineation of stands. Stands widely differ in size, usually between several hectares to
hundreds of hectares. Some stands could comprise high heterogeneity in the vegetation
structure, topography and could comprise many easily recognizable patch types.

167

26. The hierarchy of units below a spatial scale of a national forest include (as
presented in USDA-Forest Service National Hierarchical Framework of Ecological Units
1993)

1) land type association in which landform complexes and local climate are
dominant factors. These factors produce repeatable patterns of soil complexes and
plant communities (USDA. 1993)

ii)land types differences in the units is based on soils, landform (example Kalkaska
soil, end moraine). They are mapped in the field based on local topography and
vegetation.

iii)Land type phase. Finer division of land types, based on very local site
conditions, slope, aspect, topographic position.

27. Because factors causing the heterogeneity at difi’erent spatial scales, and the
relationship between units at difi‘erent scales are well defined, there is justification for
relating the ecology of organisms to the physical-vegetative environment than do stands as
they are delineated for management purposes.

28. Although parameters such as diameter at breast height and basal area are averaged
over the area comprising a management unit (stand), this should not be considered an
indication that the stand is homogeneous with respect to these parameters. Since the
range of values and the variance of ecological parameters (soils, slope, vegetation
communities) are not apriori set for their delineations, stands as management units do not
represent consistent ecological units.

29. Although silvicultural treatments are usually applied over a whole management
unit (stand), occasionally only a portion of a stand receives a treatment. In such a case
only the area of the stand that has received the logging treatment would be suitable to
represent the respective stratum of time elapsed after logging. Alternatively when several
adjacent stands are treated at the same time, the total area receiving the treatment was
included in the respective stratum of time elapsed after logging.

30. The ‘CDS database’ is incomplete for management prior to 1979 and for areas that
had not received active management. Such information was found in old compartment
records.

3]. Presence of black cherry often indicates past logging and burning.

32. There were 2 instances in which a class of time elapsed since last disturbance was
not represented in a land type association.

33. This is especially true for comparisons of vegetation variables and bird densities
among 3 replicate land type associations within a land use category. For the category of

168

managed forest, separate comparisons among the 3 land type associations were
undertaken for each class of time elapsed after disturbance.

34. In managed forest, each of the patches of continuous forest that represents a
stratum of northern hardwood forest (e. g. logged 1-5 years ago) represented an area
logged at one time. Traditional studies addressing the effects of logging on bird densities
often consider a patch (stand or plot) as a replicate of that logging treatment. The
perspective herein taken is not that of testing the efi’ects of logging on areas of a specified
size and shape, and of particular characteristics. The objective was to determine the
general effects of (selective) logging when superimposed over many other factors of
heterogeneity, on land units the size of bird territories and when viewed at the level of the
entire northern hardwood forest ecosystem. Little was known about the individual logging
treatments (only that it was selective logging and undertaken within a specified period of
time). Furthermore, the size and the shape of the patches varied, thus the effects of
individual treatments on different points within a patch varied depending on the larger
vicinity of the point.

One aspect of spatial data is autocorrelation. It indicates points that are closer to
each other to be more similar than points that are farther apart. Although the northern
hardwood forest model did not include the efi’ects of the patch (site), spatial
autocorrelation and individual patch effects were later examined by testing for significance
among different patches that were included within a stratum specified by the model . An F
test examined if variability of bird census stations within a patch was less than the
variability among patches. \Vrthin patch variability was compared to among patch
variability of bird census stations in each of 2 strata of forest disturbance periods within
each of 2 land type associations. All bird variables (10 bird species), and the vegetation
variables representing % canopy opening, number of gaps >50-100n12, % ground cover by
shrub layer, and number of coniferous trees were compared. Of 40 F tests of significance
that compared within patch and among patch variance of bird densities, 4 were statistically
significant at 0.05 level. None of the vegetation variables showed a significant difference
among patches. A conclusion was therefore reached that the patch had no effect beyond
the stratification already included in the model (that is lumping together of bird census
stations from a single forest disturbance period within a land type association was
justifiable).

35. In most cases, censuses by 2 individuals were treated as practice runs and were
not part of the data set. In the cases where the census data became part of the data set,
only the data fi'om an assigned censuser (and not the companion’s data) were considered.

36. The matching of bird density patterns with habitat availability patterns across
model strata was part of the analytical pathway followed to assess the relative efi’ects of
habitat availability and habitat fiagmentation aspects on bird density.

37. The strata of 'unlogged second growth forest' and 'older growth forest' were each

represented in only 2 of the 3 replicate managed forest landscapes due to their
unavailability in all 3 landscapes.

169

38. I used univariate rather than multivariate regression because the relationships
among the habitat variables were not linear across strata of the northern hardwood forest
system sampled. This non-linearity was later demonstrated by interactions among variables
in different strata of the model that affected bird densities. Some variables affected a
species density only within certain 'periods of forest disturbance' and certain ranges of
forest cover. Variables that had an effect in one strata, failed to have an effect in the
instance of another variable becoming more limiting under difi’erent forest conditions.

39. In these comparisons of managed forest with settled forest and primary forest,
managed forest is assumed to comprise an equal proportion of the 5 forest disturbance
periods. In reality the older growth stratum is over represented in the data since little older
growth remains in the managed forest.

40. In the program Distance Sampling used to census birds, there was no way to
identify whether the bird recorded at a station was breeding or simply occurred there as a
vagrant. To reduce the probability of recording vagrants, only those bird census stations
in which the species occurred at densities above 80 birds/km2 were selected to represent
the breeding habitat of a species. The black-throated blue warbler was commonly
encountered. Habitat conditions at bird census stations at which the species occurred
above its mean density of (5 1 .2 +/- 4.65) were considered representative of its habitat
selection.

41. Because of the blackburnian warbler’s low density in the forest, its presence even
at very low densities could reflect its habitat selection. The assumption made here is that
the probability of encountering the blackburnian as an overflow into marginal habitat is
lower than it is for the more abundant American redstart and black-throated blue warbler.

Footnotes for Chapter 3: Results.

1. An Anova of mean density differences between settled forest and a hypothetical
forest in which all 5 strata of forest disturbance periods were equally represented
(including the older growth stratum) yielded a p value of 0. 1 57 which is not significant at
the 1% level. This p value would be even greater if the older grth forest strata was
considered to be < 1/5 of the managed forest. That is there would be a smaller difl’erence
between the density of blackburnian warbler in managed forest than in settled forest.
Under present conditions the older growth forest makes up only a very small percentage
of the total managed forest.

2. These comparisons are not completely independent from the regression in which
bird density and LMCON were averaged for the 7 strata of ‘forest disturbance periods’
and ‘land use’. In regression however different patterns of the data can result in the same
r2 value and regression coefficient. These comparisons therefore give slightly more detail
about the relationship of micro habitat availability and habitat selected by the blackburnian
warbler and black-throated green warbler separately for each stratum.

170

3. In the stratum ‘logged 1-5 years ago' the blackburnian density was higher than in
the other strata of ‘ managed forest landscapes’ excepting the older-growth stratum but
these density differences were not statistically significant

4. It should be understood that a reduction of canopy cover in the strata of 1-5 years
after logging does not imply that canopy cover was reduced in the habitat patches selected
by the blackburnian. Habitat selection occurs at a spatial scale smaller than the area
representing a stratum. The blackburnian can find patches of coniferous trees which
differed in mean canopy cover from the mean canopy cover of the stratum. Mean density
of the blackburnian warbler however aggregated over the stratum of disturbance period
reflected the slightly higher mean for coniferous trees in this stratum in spite of the
reduction in canopy cover.

5. The blackburnian warbler was only encountered once in settled forest landscapes
and as a result it was not possible to compare its micro habitat selection with that of the
black-throated green warbler in settled forest landscapes.

6. A separate regression was applied to regress bird densities on values of woody
stems for each different strata of northern hardwood forest.

7. Differences between the black-throated blue warbler, American redstart and
chestnut-sided warbler are for the whole range of northern hardwood forest heterogeneity
covered in this study. As was previously presented, over the variability that appeared in
the whole range of northern hardwood forest that was sampled, both the American
redstart and chestnut-sided warbler were positively associated with % canopy opening
CANOPEN, and with number of large canopy gaps >100m2 (GAP>100) in contrast to
the black-throated blue warbler that showed no response to either CAN OPEN or
GAP>100 (Table 4) .

8. Logging was not associated with an increase in the number of large canopy gaps
>100m2
9. Mean number of canopy gaps>100m2, % canopy cover, and number of deciduous

woody stems <0. 5 were compared among the 3 species groups, black-throated blue
warbler, American redstart, and chestnut-sided warbler. To calculate each of the 3 bird
species means for these variables, only bird census stations at which the density of the
respective bird species was > 80 birds / km2 were used

10. As previously discussed, the black-throated blue warbler dropped out in 'settled
forest'. The fact micro habitat effects and landscape effects were confounded in settled
forest also precluded further detailed investigation of individual variable effects.

11. Please see the section ‘Statistical Techniques’ and Table D28 for details about this
3-way ANOVA.

171

12 The effects of forest disturbance were incorporated by considering canopy gaps,
shrub layer development and canopy opening. These however are micro-habitat level
effects. Forest management effects at the landscape level such as edge were also
incorporated. Factors that could have important effects on the density of a bird species at
the local site but could not be incorporated were quantifications at the landscape level of
the amount of habitat that contained the species specific micro habitat, and of that species
population levels. Population levels in the landscape however could also be lags from
previous events in the landscape prior to data collection.

13. The group of 12 bird census stations representing the combination of
NHWDIMIL <30%, WSDC<0.5' >50 included 5 bird census stations from the land type
association of Rapid River. Where shrub layer was developed the Rapid River LTA had
unduly high densities of black-throated blue warblers. This high density might be
explained by factors other than conferring a high habitat suitability for the species on a
landscape which is highly forested but in which the northern hardwood forest makes up a
low percentage of the total forest cover. The alternative explanation is that the
population of black-throated blue warbler begins to saturate areas where the sapling layer
has developed due to logging because of the limited sapling layer development elsewhere.

Footnotes for Chapter 4: Discussion and Conclusions.

1. The black-throated green means on total number of conifer trees and number of
late forest maturity conifer trees for 19 sites in settled forest at which it occurred at a
density>30 birds/ km2 were: TOTCON = 5.26 (se=0.72) and LMCO = 4.8 (se=0.62).
This is compared to TOTCON= 2.8 (se=0.95) and LMCO = 2.6 (se=0.38) within the
stratum of forest disturbance period 1-5 years ago. The 95% confidence limits for means
of TOTCON and LMCON for the black-throated green in settled forest overlapped the
95% confidence intervals for TOTCON and LMCON means of blackburnian calculated
across the whole northern hardwood forest. This shows that in settled forest in the
absence of the blackburnian and where it occurred at low densities the black-throated
green warbler could be more selective for conifers. In managed forest the black-throated
green occurred in more deciduous habitat and showed less selectivity. (Means of
TOTCON and LMCON for all northern hardwood forest strata and for the mean of the
blackburnian and black-throated green across the whole range of northern hardwood
forest are given in Table D 20).

2. The blackburnian warbler was not affected by the increased canopy opening
caused by logging. Within the stratum of forest disturbance 1-5 years after logging, the
stratum that has the greatest canopy cover reduction, % canopy opening at sites selected
by the blackburnian did not difi‘er from mean canopy opening for that stratum. Mean
CANOPEN= 30% (sd= 19.5, n=12) at bird census stations at which the blackburnian
density was >30 birds/km2; CANOPEN=22.55 (sd=15.9, n=46) for forest disturbance
period 1-5 years after logging.

172

3. In the Rapid River land type association, the black-throated blue warbler had a
relatively high density in unlogged second growth in spite of the very low number of
woody stem. This situation might be reflecting an overflow into marginal habitat fi'om
proximate more suitable habitat because of the limited northern hardwood forest habitat in
this landscape into which individuals could disperse.

4. Details of dynamics of the American redstart, black-throated blue warbler and
chestnut-sided warbler in Munising, Manistique and Rapid River were presented in the
Results section.

5. High deer populations are blamed for the poor shrub development on the Rapid
River District on the Hiawatha National Forest. The study area receives high numbers of
deer moving fi'om lowland conifer areas fi'om both the Whitefish River and Stonington
deeryards. See Van Deelen (1995).

6. In some narrative accounts there are descriptions of the black-throated blue
warbler using recently cut brushy forest areas in Canada (Godfrey 1986). This suggests
that the black-throated blue warbler is adapted to a wider range conditions than what was
revealed by its density pattern in the Upper Peninsula of Michigan.

7. All encounters of the black-throated blue warbler in settled forest occurred early in
the season. That it was not encountered later in the breeding season in settled forest
suggested that it did not establish territories. T. Donovan (pers. com) also observed that
some bird species are prevented fi'om establishing territories. Distance Sampling, the
methodology used to estimate bird densities, does not differentiate between mated males
and unmated males, or males that have actually established territories or simply passing
through. All black-throated blue warblers that were encountered during bird censusing
regardless of their true breeding status, were used to derive a density estimate .

8. In the Upper Peninsula the blackburnian warbler has been recorded in habitat types
that include shrub with scattered conifers, northern white cedar, young black spruce and
mature black spruce (Dawson 1979); Robert Doepker (Michigan DNR; Pers.Com)
however found the blackburnian warbler to be most abundant in hemlock patches within
the northern hardwood forest relative to other conifer forest habitat during his bird
censusing for the Escanaba State Forest Plan development.

9. Several studies have indicated competition among the blackburnian and black-
throated green warbler (MacArthur 1958; Morse 1976).

10. Since the microhabitat selected by the 2 species is relatively stable, their densities
could be monitored at selected sites to establish interactions between them up along a
temporal dimension.

11. This process has been described by Graham (1941). The fiequency of disturbance
for 2 primary forest landscapes Upper Peninsula has been described by F relich and

173

Lorimer (1991 ). Davis (1995) describes present day stands of sugar maple that had no
indication of ever including a hemlock component.

12. A stratum of primary forest landscapes was included in the conceptual model to
stretch the range of forest conditions with respect to forest cover at landscape scales,
presence of a coniferous component and low edge. This stratification was used to test if
there were significant difi’erences in the densities of individual forest interior bird species
(black-throated blue warbler) and forest edge bird species (American redstart and
chestnut-sided warbler) and in the relative abundance of the forest edge and forest interior
bird species within this forest heterogeneity range. With respect to describing the
sensitivity of bird species to habitat fragmentation aspects or to the presence of a
coniferous component and to describing relationships among species along any
environmental gradient, the primary forest did not need to be representative of pristine
forest conditions that existed prior to the large scale disturbance at the turn of the century.
It should however be emphasized that without strong evidence that the primary forest
sampled in this study is representative of pristine forest conditions all relationships
established in this study, are limited to describing relationships among species and between
species density and habitat conditions existing in the present forest.

13. It is conceivable (although not practical and very tedious) that with a large quantity
of data collected at small intervals to map isoclines of canopy opening, distances from
edge, forest cover to produce different maps or GIS layers that will reveal the spatial
arrangement of patches falling in specified ranges of a variable. The intersection of the
different layers would produce a map in which patchiness of habitat conditions will
delineate difi’erential suitability of forest conditions to the black-throated blue warbler,
American redstart and chestnut-sided warbler.

174

APPENDICES

175

APPENDIX A

176

 

823 0959.3

 

88a 2
55 as a 82 25 23 as as:
8.8 we moo. 8 38: 533088 25 284 38353
stun—52m 8:335
3:8 Bee—cm mo Bria
a: 2 38.. E .= ._ geese Seam Seesaw
8535 Ecomwom
.8635 3:83:00
8:8 0323
ac 88.2 8 88.83 cacao 320
eofiouoom
owe; cum .8280 952 .._e 6 ton—ac SE53 388m 5 520 8 03% we»?
a cegaaoaom as: 338m 23 mesa:

 

.Gwe 9? «0 ton—3 83898. gwfioaz mo 5:32.82on
2: a 852% aoeseaa a as .38: on acaczaa 622$ aaoa <8: .aaesm a 52» a an:
flame—o8 me 3882: 3:332 05 E .02: m: 322. 93 30% 3E: flame—coo 9 some; 25 new. 05% unicorn—om .~< 033.

177

 

80...:00 0:0.30. 3.082 .002 00 5m 0.3.1. 08. 05880 0.5
.3085 0:03:80: 0030.. 0083 08:3 8.0: 0 00 0.05. .0003 000.8 80.0..
8.505 00.0». 05 :. 00. .0.0—.3 8.00000: 00:3 00:80". 098.80 0.0 0.00

 

80+. 00:00.8. 0.88:0 80:0. 5.3 .83”. 5.08.55 05 .8 803 088.89 8:230]. :88<-fia£.£5 0 <0..
.05. :.=.0:0m.< 00.3 05:0 8.0... 08380 088003 0. 8:85 80.5. 00.04 80m m «a...
008.08.... :.:.0:0w.< 00.0.. 0500 0. 0.00:... 8.85.0 m: 80.8.83
88000.3 003 ,8 000. 8.800% 0:: 5m :35 05 50300: 0880,.
0800.88. :0 0: 0. Emaofi 0. 80800: 0...... «0.0.8.8800 000308.. 8058:
3.300 .30. 8.08003. .:.w.8 .0:.088 .8 00:0: >800. 0000.050 :03 00.082 E00302 838% 80:05 *0 <0...
.:.0.0:0w.<
00.3 3. 000.830: 0:0: £00380 :. 00:80.. 5.0800 05.003 0801. :0 n. :08.w:._.m m «>5
.008 0.5 :. 8000 0.000308... 3.03.0 30. 0. 80.008 0:: 3:08.. 080 00002
.:.w.:0 88:00-00. 0:: 50380 .8 00:8 0:20: 5:380 0000.050 :03 80... 50380 088003 N «A:
8.805 00.3 80... 00:038. 0.88:0 meobm 0.00.00: 086088
00 30:20 000:: 080m 00.808800 00030.8: E058: 8.08003
:.m.8 £00380 8 08.088 .8 00:8 >800. 0:0 00:00 05.0: 0000.050 :03 08082 8.0.8.2 *. «a...
88.080 8.3083 25 83 082 <5 3:...

 

as: 20.8 8:80:30 8.0280 :80 a s .82.
00808.00 :00: 0.2. 55 m: .0008”. 080.87. 2.830... 05 .8 .83 0003 05 .8 80.8.0800 095 0:0. 05 .8 8.8.8009

.2 use.

178

000:0 0.030 00 0080.0... 80.00.000.00 003051....

 

5008.. 0:0.—0.800 5008

 

 

.030. 00.03 .3... 300880 0:358: :0800 5.3 0:358: :05 :5 8.... 0000080 88885 .. 0......
00.88.8800 00030.00 E058: 5.080000
00:0.00 0:0 3008.. 000000800 5008 8.580000 000:0 .03 5.3 00:0.0: 0:0
55.53 .8 E0500 0000005. 030.00., 6.00.000 0:858. :05 0.80 :5 8;... 088028088005 o. <.....
.0:.0 0.08. 380030 5:00.50
00:0 05 808.050 000.585 0:0 0008.00 003 :0.00.0m0> 050: 0....
08233 83.5883 20 5:25 008.20 30:0 as. 200 050 200 8:223 0 <8
0880:30 8050 8 5003.
000.. 00m. :005 05 .8 :08000 308.580 05 5 008.8,. 08088 ...... 08083. 8.000 m 5....
00888800 000308.. 5058: 5.3 00.000w0> 0:0 005503
5.3 00800808. 00:00 0:0 m0m0.: 80000 .8 80:00 0000005. 030.:0> 5.0800 0.03m-0w0.d 0:0m 8.008. 0.. <91.
88880 80008.2 25. 0:3 252 <8 0<S
.8288. .2 2.8.

179

APPENDIX B

180

 

MammaLSnecies

White-tailed deer

21mm

American beech
American elm
Blasam fir

Black cherry

Black spruce

Canada yew

Eastern hemlock
Eastern hophornbeam
Jack pine

Northern white cedar
Paper birch

Red oak

Red maple

Red pine

Sugar maple

White ash

White pine

White spruce

Yellow birch

Table B 1. Scientific names of animal and vegetation species.

E. l S .
American redstart Setophaga ruticilla
Black-throated blue warbler Dendroica cerulescens
Black-throated green warbler Dendroica virens
Blackburnian warbler Dendroicafusca
Canada warbler Wilsom'a canadensis
Northern Parula Parula americana
Brown-headed cowbird Molothms ater
Chestnut-sided warbler Dendroica pensyhmica
Indigo bunting Passerina cyanea
Least flycatcher Empidonax minimus
Ovenbird Seiurus aurocapillus
Red-eyed vireo Vireo olivaceus
Scarlet tanager Pirangea Iudoviciana
Swainson's thrush Catharus ustulatus
Veery Cathamsfizscescens

Odocoileus virginicmus

F agus grandifolia
Ulmus americana

A bies balsmnea
Prunus serotina
Picea marirma

T axus canadensis

T saga cmmdensis
Ostrea virginiana
Pinus banksiana
77mja occidentalus
Bemal papynfera
Quercus rubra

Acer rubrum

Pinus resinosa
Acer sacchamm

F raxinus americana
Pinus strobus

Picea glauca
Betula alleghaniensis

 

181

Table 82. Variables considered in describing vegetation structure and composition at the
microhabitat level.

 

1 - tree species

2 - diameter of trees and snags

3 - tree height and snag height

4 - % canopy cover

5 - ground cover (% forb, grass, litter, base of shrub, base of coniferous shrub or tree)
6 - average height of shrub or sapling layer

7 - number of canopy gaps in 4 size categories
8 - distance to road, trail, opening, and other forest types
9 - number of deciduous stems <1cm and >lcm-3cm<

10 - number of coniferous stems <1cm and >lcm-3cm<

 

182

 

m

o

n

v

o

w

v

v

3‘

o

2

o

53% 320

538m 288 comma—=3

own E3» m _A @233 8 33225

own 38» m TEA comma: 820m “gowns:
emu mac» ~70 Emmo— umohou Bwncaz

own who» 2 commo— aeom now—282

 

838$ 233$
luafluaumll. .

83:3 03233..
Idaafizl. .

835% 03232

mamazz

 

.2253.

3550 ES .«0 Eon—5:988 .8 3.85 E338: of SE 85 22369.8 25 Ba— 93 mecca 02.3536 3.28 me

SE; Eobbfi of .8 3.28 womafiE 2: E 03283 203 85 N2: 5 883.3% mum: Eat coma—BE 35% mo 38:2

.mm 035.

183

 

10 x 30 m plot
campy cover
tree species
tree diameter
2‘5““ MN, tree height
woody stems campy gaps

and ground cover
Center of vegetation

lsamplingplot
r lfijl lfilr J.

I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure Bl. Vegetation plots centered at bird census stations in which microhabitat
variables were measured.

184

 

  

 

 

 

 

 

 

' as,
~ .-
z 24 .26 30m
o 10 12 I
closed :gd
campy open py open
campy

canopy

A Measurement of % canopy opening along a line

 

 

 

 

 

 

Canopy Gap Projection
B. Canopy gap measurement

 

 

 

Figure B2. Measurement of % canopy opening and canopy gaps.

185

APPENDIX C

186

 

 

555085 :Em 6.552 :03 65.33 63:289-._0.:=ou 02 35268365600250
5:05.085 35m 6352 :03 65:5..— 6=o=6_009.00.=:ou ‘5. "— Z 035E :85 0059:3535 55m 2.
5225085 25m 6.0.— 303 65.33 630320095550 1E. m 2 0.2. :23 00595-5 55m 2 _
5.33 53m ago 562 was: :25 gaaaaé 835 2 _
5:52 5552 0.3 6:52 6502 65:5 ”mac—6.085.500 2: 68365: 550838.80: 60556 0.02803
.5552 5552 5..— 6552 6502 65:5 60.3on 2: 0.58 535050—005:
.9552 5952 0.3 6:52 6802 65:5 60.250 2: 0.00:5:
.232 Eta“: 23 M25: 682 6.33 5:80 "_z 0.8.5: as
.552 55:55:. 6552 ME 65:5 655280-5350 U2 5365:-50 62-0.5 32>»
_eaaz €855: m2.33;. Ea 68:5 582803068 2: can 33.8 30-05325 635
.55 Z 5552 33 6:52 ”5 65:5 630.625.00.500 U2 58055522600365.205— 823
.852 m€35 23 M25: in 68:: ”38289580 2: 683.950.8535 223
.552 55205;. 6552 6502 65:5 60.530 m2 0.8—80:05:. 823 ovo
.852 6822.5 m23.5 En 65:5 5:80 02 2886538
5552 5520.5; 6552 ME 65:5 60.250 02 355.805; 0023
855: 6825: mean: So 65:: 588 m2 232 8:. 323 as
.555: 6855: 3.8 En asap ”5:80 ".2 28 2i 233 08
5:5 55 5&0 “_z 05.88 2i 323 m8
.332 68225 M£52 ube m22.5 €0.80 2: 22380 Ea 223-05 8:
5:52 55:55.... 6:52 “5 65:5 60.259 2: .55: 05m 60m
8:35; 682.05; m23.5 Ea 65:5 5:80 ".2 255 Ba 3. a8
858: 6855; 68 Ba 68:: 5:80 52 22. 8&3. 08
5:: 35 M58 52 05.38 Ba 3. MS
.8552 5.85885 25m 6252 ”5 65:5 60.250 2m. Saabfiq 0.5..
8:55: 683385 25 M25: Min 653: ”3:80 "_z 23.2: one 0.8. as
56855 556885 beam 66m “5 65:5 60.250 m2 0:5 0::— 303. 08
5:5 55 M58 52 ”553 Ba 0.8. 5
£955 053654 com seasoning—U 083m 65. 0535 65 Egon 355005”.— 85 068
35 252% 5 wanes; 65802 .2352 was
:85: awoken 5:052 89a 8%. 550 =<
.5055 55355.0 “05— Eobbfi Bob 00%. 0960 65. be 5550:: 05 E 8.5 0030 we 5:50:65? 525 .5 033‘

187

 

08...: 88.0 82.0 0.....2 05°38? 82082 mm.

.0532 5:06.05... 6532 6302 65:5 636.009.0650 02 03260036365500

0.. 5.632....

3006.05... 6532 60 3 65.33 6.506.000.0250 2.. 00300555056000 600363..-.0..:00
3:06.05... 6532 603 65.39. 6056.89-36.80 4.....2 F.8... 0050-50.00.50 00.
080.5... 8.0.. 003 85.33 m2.030.000....80 0...... 8.... 8.0-8.8880 .8.
08.33 88.0 .8.0 0.....2 8.... 2.0-8.8880 a.

3:06.05... 6532 603 65.33 6.506.000-3650 2... 0:...5 62-3600 86.38.005.650... 30..0>
3:06.05... 6532 60 3 65.33 60.500 5......2 053:. 00.500 0530 6052 0...
300635.... 6.0.. 603 65.39. 60.500 5......2 0.0.. 00.500 53... 60.02 6..
08.33 .830 82.0 0...... 05.... 80.80 .53.. 8...: 3.

3:06.05... 6.532 .3 3 65.39. 60.500 2.. 00:50. 0.6.3-... 503.
3006.05... .532 603 65.33 60.500 1......2 053:. 6. 50.00.0085 0....3 06.
.558; 8.0.. 00>. 85.33 .8680 0.....2 0.0.. .0 5.5.8:... 3...... 00.
08.33 8.8.0 0.2.0 0.....2 was. .0 0.83-88... 0......» 8.
3:06.05... 6532 603 65.33 60.500 4.....2 053... 0.0.35... on .
3006.05... 6.0.. 60 3 65.39. 60.500 .2 0.0.. 0.0.55... on.

3.532 0.3 6532 603 65.33 60.500 2... 3600 0.6.3 6.0.652

.0532 0.3 6532 603 65.33 60....00 22 5600 0......» 80.6.02
0.5.2 3.... .83.: .83 88.33 00.80 0...... 238. .88 3...... 82.82 a...
5.532 03.. 6.0.. :03 65.33 60.500 4......2 0.0.. 3600 06...5 80.6.02 6...
6.80.2.5 .20.. 003 88.33 .8050 0.....2 0.5... .88 3...; 820.02 m...

300.8085 3.3.. 6532 603 65.33 60.500 02 5604.805... 0.02m.

300300005 3.3m. 6532 60 3 65.33 60.500 2.... 00:50 0.005
300300026 3.3m 6532 .0 3 65.39. 60.500 4......2 058 00.5.0 0.005 mm.
6.56885 2.... .06.. .83 85.33 00....50 0...... 0.0.. 8...... .8... 00.
08.33 88.0 82.0 0.....2 055. 8a... .8... MN.
0.0.55. 050065.. .0. 006006.056 350m 065. 03...... 65 30.0.... 3.005.090. 00065 0600
3.0 282% .82 «858...... .0.880.... .8555: 0.30

:05: .3000". 300.32 50.. 00..... .260 .60.

 

5.8.88.

..0 0:3...

188

 

3.552 83 6552 6.82 65...: 65056.3 3......2 A255. 683683-255 human 3..
6:52 33 28 5.82 65.5 68.6.8: 1:62 3.8. 6855-25. 5.5 es
23 233 Ba :23 32.3.2.5
3.552 33 6552 6.82 65...: 6556.89 22 62-25:. human. 668365.. wasp—:26 6a..
5 552.2. 258.55
3:06.65... 6552 6.82 657.3 6505...:oo-m:o:6.8n. U2 58.05.58.255 555
552 33 M25.). 6.82 65:5 68.6.8: 562 5:2 8 xsafiéooagiaaa 5%
3.552 83 6.552 6.82 65.9: 685.500-536.89 2: 68365.. 558: 28.502-6055506 68365:
3.552 83 6552 6.82 65...: 6556.89 02 20.2255 5m5m
3.552 83 6552 6.82 6553 ”5056.68. 02 5555552255 5w5m
3.552 83 6552 6.82 65...: 6:96.02. 02 5526509255 655
552 23 M252 6.8: 65...: M255.8: 562 .258. as... 32.03.283-25: 5% 2 m
55.2 33 6.8 3.82 653: 68:28: 562 .22.. .53 32.03.58-035 5% 2 a
5:06.65; 6:52 303 65.33 65:56.89 3......2 68365: 65.33 35
5:06.656... 6.0.. .63 65.33 6556.82 3.6.22 668365.. 65.33 was
6553 5.5 5A5 5.62 5835.. 553 ME
65255:... 6552 :85 65.33 65056.89 02 «655605.58225 3263.255 63.
650.55.; 6.552 :03 65.33 6556.3 U2 «585.05.252.85 30:03.25:— 63.
656.55; 6.552 .63 65.33 6556.3 U2 368.05.500.22... 32.63.25:— 63.
650.55.... 6552 so? 6:233 6556.89 imdflmz A3...w 63. 255: 255 63. $5
3:66.655. 6.0:. 63 65.33 6556.02. 3......2 A05 63. 2o: 255 63. 62.
650.55; 6.552 .63 65.33 6556.89 22 5.6.5355
650.55.... 6552 .63 65.33 6:96.89 3.6.22 255 62.5.0 52.854.256.55 a. a
1:06.65; 6552 so? 6:233 6556.89 AREA 255 65.5.“v 56.55-256.55 6;
65.33 656m 62.0 3......2 255 66.75.0 50.55.256.55 m. h
6855: 552 So 63...: 5828: 562 25:. :83. 82.82 an
6855:.— 68 E: 65.5 58.6.8: 5....2 22. V.8 62 .5552 62
292.65 058653 5.. 5685820 850m 665. 3.3.5 65 680.2 62:65.5:me 82.5 068
Sun. 252% 62¢. 8652.3 225.6062 658552 mum:

:65... .696... 656.52 58.. 8.3 :38 ..<

 

.3583

.. U 25...

189

 

 

35.33 .32.. .58 3...... .385 a 53.2. 82% 333-832 33
3.5.6085 3.3m. 6532 :03 6.3.39. 6.5.6.009 4......2 0.53... 3.5.0.. .533: 95
3.5.6085 3.3m. 6.0.. .33 6.3.39. 6.5.6.009 4......2 0.0.. 3.5.5.. .533: 63
6.3.33 66...: 62.0 4.....2 $59. 3.5.5.. 5363: 95
3.5.6085 3.3: 6532 6.602 6.3..5 6.5.6.000 4......2 053... 60.63 600%.: omo
686883 35.. m2.... 3.82 65...: m255.85 3.32 0.2. 8.... 58.03 38
6.353 66...: 62.0 A... n. Z 950.. 60.6.. 600%.: mmo

3.56.6.3; 6532 .33 6.3.33 6.5.0....50356300 U2 36::Emo\.0.....00.6oo363:

3.56.6.3; 6.532 .33 6.3.33 65.0....5065056300 2: 0.9.... 62-030.50.668... 3...?

3.5.6085 3.3m. 6.532 6.602 6.3.53 6.5.6.02. 2: 60.... 30:0» 6.3 60.... 0......»
383883 35m .23.: 6.32 68...: 38:28.. 3.32 23.... .2... E... 33
686883 .50 .29. 6.82 38...: 38:28.. 0...“... 22. .2... .2... 38
35...: M32.. .88 3.32 was. .2... 5%.. 8.

3.5360003 3.3.. 6.532 6.602 6.353 65.6.80 2: =0.6<
3.5.6085 3.3M 6532 6.602 6.3...D 6.5.6.000 0......2 0.53... 60...... 9.6.5.0 6;
3.5.6085 3.3m 6.0.. 6.602 63...: 6.5.6.00n. 1.... n. Z 0.0.. 60.6.. 9.6350 6.6
35...... m32.. 3&0 5...... 5%. 3.3.25 6.3

3.5.5.3.... 6532 .33 6.3.33 6.5.6.009 2: 683639-050... 363-5...

3:0...658h 6.532 .33 6.3.39. 6.5.6.02. 2: .33 0....3

22... 28-..... 3...;
3.56.6.3; 6.532 303 6.3.33 6.5.6.009 1......2 6836.36.06.35. 6600363.. 6052 3..
3....2 £2... 28-..... 3...;

3.56.6.3... 6.0.. .33 6.3.33 6.5.6.08 4......2 683636.055... 6600363.. 6052 63
3.532 03. 6532 6.602 65...... 6.5.6.00n. A... ... Z 0.9.... 3m=m mm»
.33.: a... .0.9. 3.82 M3...... 6.8280 5.....2 2%... .36 33
3.56.6.3... 6532 6.602 6.35: 6.5.6.02. 4......2 A0535. 0.3 36 0.9.... 60: 9...
3.56.6.3; 6.0.. 6.602 6.3.3. 6.5.6.003 1......2 6.03 0.... 6 0.9.... 60y. 6.3
6.6%: 0.60665... 5.. 6230:3635 0050m 66.3. 03>...“ 6.3 305.. 3.55me: 600.59 0600
a... 252.6 .82 .858; U.0550“... 353:8: 63:

.55: .305... 3.5.32 :5... 60.5 .050 =<
60.5.58. .0 0.3...

190

 

 

:25 2: >522: :30
58 4.22 38 8

333 mswam m=25 2: 32

2.233 imam E25 2: moi—mé—uw .025

333 imam m=25 2: 3.5534
333 Mfiém “Ego 4.22 fies 3.33 8
3.86385 bam 6532 so? 6533 638559303289 Jam—Z A2355 5 Sam—3-38% BEB-=o%< o3
3.55885 23m 8.0“— 83 6533 ”32035933280 17:2 395 E Bum—3-82% 3_:3-=o%< wma
max—SE 239.654 .8 593385 830m 35. BPEQ v5 “meow Eggxm 8x5 38
5.5 253% ,82 @5033 U35:03,. 33:82 mum:

:83.— Jmusom .3232 89a 8gb ~38 =<

 

Asaécsv

AU 03:.

l9]

 

 

 

mOmD 88.... 1.8.. :8... 33.8852 8...... 158822... 1.8.. 52. 38.88.52 5.50
£23 88.... 1.8.. 88.. 3.....8552 32.. 85588.... .38.. :8... 38.88.52 55.5
582.0
mOmD 88.... 38.. 88.. 88.8....52 85... 8 0.: 5-888... .2852 2.5.5... 89.0 88m.
mOwD 88.... 38¢ 88.. 88.58.52 858 .8 o... 5.808“. .2852 «£95... 5.8....
Dace-h.
E... .2282.
mOwD 88.... 1.8.. :8... 808.8352 88.... 38¢ 88... 8:58:52 8.1. 0....
wow: 8.2... .2... 22. 8.2.8.82 8.... 5.2.8.2... .2... 28. 8.88.8.2 .25. 2...
£2: 88.... .38.. 88.. 3.9.8552 32.. 85.3.8.2... .38.. :8... 38.88.52 5....3.
mom: 8.2... .2... .2... 8.2.3.2.... 8.... 5.88.2... .2... 22. 88.8.8.2 2.... .88..
mOmD .88.... 38.. 28.... 888352 8...... 858.788.... 1.8.. 89... 38.28952 8.5. .83.
£2: 88.... 38.. 88.. 88.88.52 8.2.. 858808.... .38.. 82.. 38.88.52 55.5
.522...
mOwD 88.... 1.8. 88.. 88.28952 356 .8 om. .5388. .3857. .595... 02.... o...
£2: 88.... 1.8.. 82.. 388.352 £0 8....-.88... .5852 2.5.5... 3...... 8:80
£2: 88.... .38.. 82.. 88.8852 m5 8....-.88... .2852 2.5.5.: 882m
£23 88.... 38.. 88.. 88.8....52 85.... 8 o... .8388". .2852 2.5.5... 8.3
wow: 8.2... .2... 28. 8.2.8.82 m... 8.8.88“. .32.... .528... 288.80
mOmD 88.... 1.8.. 82.. 822.8852 26 8....-598... .3852 2.5.5... 0...... 8.8.2....
2.2.25.2
mOmD 88.... 38.. 88.. 38.88.52 858 .8 o... 5.508... .8357. 2.5.5... 0...... 5.8..
mOmD 88.... 1.8a 82.. 3828252 .85... .8 o... 3.88.... .3852 2.5.5... 2.1. o...
8... 8.2... .2... 22. 8.88.82 8.... .8 058...»... .8222 .55.... 8.3
wow: 8.2... .28. 28. 8.88.82 8.... .o 058...»... .882 5.3.... .3... 2......
.8... 2.3.
83.80 853 .883. 853 85.. 29...... .85.. 8.8m .88... m: Nun-.830
88... 8.. 833.880 mom:

 

.55 >85 w£8>3 3a.. 8035.58» .8 82.5w .mu 2......

192

 

 

mOmD 8:5: 3.5g 88: 3:58:52 8:2: 2:538:32: 3a 88: 3:58:52 .3555?

wOmD 8:5: :58 82: 3:58:52 «:2: 333882.: 53: 82: 3:58:52 .853:

m8: 88...: Es. 82: 83:85: 2...: 2.38.2... :5. 32: Baraga: 83.0 23

88: 8.2a 1:2 as: 83:85: 2:8 Barnes: 3.8 Spa 85852 23 :5 5.:

25me

$22.: 33 3:58:52 8:5: :55 88: 3:58:52 :2: 569329.: 3.5. 85: 3:58:52 8:5 wow—3m—

Amofia: 133 3:58:52 8:5: :55 82: 3:58:52 :2: 859833: :55 88: 3:58:52 225 8:3 35m

v3.....5092

98:5: .503 3:58:52 :3: 350892: :58 88: 3:58:52 :5. 255: =< 5:5:2 :95:

90:5: .503 3:58:52 :2: 852.823: :53 88: 3:58:52 :5. 225: =< a: 80>—

Amoaa: 1505 3:58:52 :2: 526823: 3.8a 85.. 3:58:52 5. 295: =< 8:5 oBoI

A823: .503 3:58:52 :2: 350822.: 53.. 8:5 3:58:52 5. 225: =< 8:5 5332

2.12592 :95:

mOmD 88:: 13a 82.. 3:58:52 :2: :5.m+mo:£: 13a 88.. 3:58:52 «move:

88: 8.2: 3...... 82.. Baraga: 9:: 5388...: :5. 82. Bags: 585

m8: 8.2: 3.3 82: 35852 9:: 538.2: 1.5. 52: 3585: 5.8.3::

33::

8: 8.2.: 3:2. as: 35852 was: 5.. 55-308”. 1822 .53: 3:5

mOmD 8:5: 53a 88: 3:58:52 8:58 mo 01.25808: 1:852 253.5: 58:0

mOwD 8:5: 5.52 88: 3:58:52 8:58 :o 0:: 2.82:: 1:052 2:255 39352

83.588

88500 :53 .382: 553 8:5.— 055: 8:54 33.—um .88: mb 0353.30
833: 3: ems—950 560 3.3530 mOmD

 

 

 

3:538:

do 03:.

193

Table C3. Types of features in addition to primary and secondary roads, railroad tracks,
powerlines and gaslines, included in the calculation of edge within 500m circle

areas around bird census stations.

 

 

Code of Feature
Used in CFF Files“ Description
106 Unimproved road class 4
402 Stream perennial
410 Lake or pond
515 Road, light duty class 3

 

*added to USGS quadrangle maps per US Forest Service specification.

194

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__ L.

T_‘ 1——
Sugar Maple T Hemlock Hemlock Sugar Hardwood Hemlock Hemlock

 

dominated maple dmtinated dominated
Hardwoods northern hunlock- hemlock-
hemlock
D Relative area of type

<= Forest Disturbance
=> Forest succession

 

 

 

Figure C1. Schematic representation of the contrast in the successional relationship
between sugar maple, hardwood dominated northern hardwoods, hemlock dominated
northern hardwoods and hemlock in secondary and primary northern hardwood forest.

195

 

500m circle
around bird census station

 

 

 

 

 

 

 

 

 

 

l-mile circle
., ‘-,. use; F'- , around bird
Roads, water courses m;:“‘"'" ‘ census station
water bodies layer * ...... >/
E . Q... . I.
,‘3 /
0 . ......‘ M ...... M .“O

Vegetati on cover 5" 5
types “'~~- ........m ........... '

 

 

 

 

Figure C2. Relationship between vegetation cover and base features (roads, water
courses), 1.6 km circle clips and 500 m clips around bird census stations.

196

APPENDIX D

197

.86on :8 833203: mo 89:3: n :_

 

 

N _ m 8 _ 8:88.88 8_ 8:82:38: .825 .238
88 818.8 8.8.8 2: 0:88:58: bog
Ea 8.8.28.8 8 8:883: 5:288 as:
v _ 88 88.8.8.8 8 888823: 023 83-8:
88 8.8.8 8 8:52 :3: 258.6
:8 8.8.88 8 58:5 3333 828-5535
88 8.8.8.8 8 0580338: 5.82 50:05.
88 8.2.8.8 8 8888.3: 3333 83 38233.:
E. 8.8.8.8 8 :88: 3333 :88 3:82:85
_2 8.8.8.8 8 82 ES: .0333 58:88.:
_: £9685 gauge... E52 86on BE
.8 as:

 

.5??va we Sagan $85 05 :m 820.: 80322 E058:
:_ Ema—-32 :_ 3:26:00 8388 EB 80¢ 3:83: 803 8:: 8585 .86on BB 2 he some 5: 530:3 5:88:
a .8 :23838 05 :_ 8:03.850on :05: :5 3.02 J: 8 838: wE—qpam 3585 8:38: 2: :_ vogue—om .302 .5 035.

198

Table D2. Mean density/km2 and encounter rate of 10 songbird species censused at 321
bird census stations in northern hardwood forest. Northern hardwood forest
was stratified into 7 strata. A roughly equal number of bird census stations
were located in each stratum. Bird censuses were conducted in 1992—1994

 

 

in the Upper Peninsula of Michigan.
Mean Encounter Rate
Bird Species Density sel (% of bird census stations)

Blackburnian warbler 14.18 2.15 18.7
Black-throated green warbler 64.45 3.60 75.3
Black-throated blue warbler 51.2 4.65 49.5
American redstart 50.09 4.52 43.9
Chestnut-sided warbler 18.54 2.52 24.3
Ovenbird 96.02 2.96 96.9
Red-eyed vireo 101.11 2.76 97.8
Least flycatcher 201.11 18.59 57

Scarlet tanager . 10.23 0.87 40.8
Veery 14.95 1.09 53.9

 

lse = standard error of the mean.

199

Table D3. P values of apriori contrasts in ANOVAs undertaken to test the statistical
significance of differences in means of TOTCON and LMCON among land
use and forest disturbance period strata. These contrasts are for a
hypothetical northern hardwood forest in which the strata of primary forest,
settled forest and 5 forest disturbance periods in managed forest were equally
represented. Bird density data are from bird censuses conducted 1992-1994

 

 

in the Upper Peninsula of Michigan.
Strata Compared‘ TOTCONZ LMCON“

Managed forest 0.001 <0.001
vs primary forest

Managed forest 0.42 0.7
vs settled forest

Primary forest 0.001 0.001
vs settled forest

Managed forest logged <12 years ago 0.64 0.45
vs all other managed forest strata

Managed forest logged <12 years ago <0.001 <0.001
vs primary forest

Managed forest logged < 12 years ago 0.85 0.55
vs settled forest

Managed forest older growth 0.143 0.31
vs primary forest

Managed forest older growth 0.068 0.001

vs all other managed forest strata

 

lEach stratum consisted of a random sample of 25 bird census stations from each stratum.

2TOTCON = total number of conifer trees.
3LMCON = total number of late maturity conifers (hemlock and white pine).

‘TOTCON and LMCON were sampled within 1200m2 vegetation plots set at bird census

stations.

200

Table D4. Mean of total number of conifer trees TOTCON, and total number of late
maturity conifer trees LMCON, in difi‘erent strata of forest disturbance period
and land use. Vegetation was sampled in 1992-1994 in the Upper Peninsula

 

 

of Michigan.
lQICQN.

Forest Disturbance n1 x2 se3 x se
Managed forest logged 1-5 years ago 46 3.09 1.05 1.28 0.87
Managed forest logged 6-12 years ago 55 2.38 0.96 0.8 0.79
Managed forest logged >12-15 years ago 58 2.46 0.94 1.37 0.78
Managed forest unlogged second growth 42 2.9 1.13 1.3 0.93
Managed forest older grth 25 5.84 1.43 5.52 1.18
Primary forest landscapes 46 7.56 1.05 6.65 0.87
Settled forest landscapes 43 2.88 1.09 1.74 0.9

 

1n = number of samples (bird census stations at which vegetation plots were set up).
2x = mean
3se = standard error of the mean.

201

.mcomwmum mama—00 was a“ boquO— mac—Q comuwwomgr NEOON— E Bufifig 0.53 20024 Us ZOUHOHn

woo: msouomioo bran:— 82 do .8982: _83 n 2002‘:
.80.: 39333 .«o 698:: _82 n ZOUHO?

 

 

dmdd mmod v.28 dad—tom 5&5:on .. Eco; .. 888:0
wad mood £3ch 320 0:36:32 _. maids:
mdwd wood comma—.5 538w vacuum .521 Edam _. mamas:
wwdd dmod own as?» SA Emma; 53¢ Edam .. osgflcfiz .. mfimeaz
med E. .d owa Ea?» N70 comma; :32 Baum .. 0:35:82 _. @6332
Rod End emu E8» m; dammed .25— Eaam .. gags—2 .. waving
dmvd wood 528 SETS 6.03.5002 _. 252$ _.. 32 552
20024 ZOUHOH datum 359.525 3.3950 828682 33. v3

 

32-82 8322.6

£355“. 5%: 2: S BEES 83 =o_§owo> ..o«: 98. E accombzc .8582. 23 boron 359.56% “8.8%. magic.
.223683 25 98—. “Sheba mecca 20024 93 ZOUHOH 8888 8 coo—«tong 8>o=< mo mos—g m =80>O

.mD 03:.

202

$0256 313...: 82 mo .382. :38 n 2002..“
.80.: m:o.o.=..8 do 338:: .59 u ZOUHOH.
.50.: 2.. do .oto Paved... n own

due: MO hon—En:— fldve ||.. urn

.328... .5280 E... «m 3.32 32: catfiomo.’ «Edda. 55.3 BEEmm 203 mesmtg 8.38%; .292: mEEEam do .382. u :.

 

203

 

8.. ohm m: on... m. .85.. 8.2% somceem
8m 8.. 3m 8.. o. .33. Beam .536
2.. Rm M: 2m 2 .88.. 8.8m 55:
Em 8... fim «mm : £32m 52o .88.. 8mg: 3.0.2.5:
o. m .2. Sm 6mm 4. 532m .8... .85. 8mg: mamas:
8.. 8.. o. .. o. .m cm comma? £32m 2m .88.. 8mg: 23.
m2 8m 5 . .. mm a. Bmmoas 538m 2m .85. 8mg: mamas:
8.0 3.. mm... o. .m om emu e8: SA Emma. .88. 8mg: 23.
m3 .6. E... t m cm cm“ as; «X Emma. use commas. Bowman:
hm... mm. o. .. 8m .. omm 3...% SA Emma. .88.. 8mg: mamas:
on. 2.... 2.. mm... mm omm 38» 2.... Emma. .88.. 8mg: 23.
8.. a; 8.. Sn 3 cm“ $9 2... Emma. .88. 8mg: 33.....52
3.. 8m om. mom m. om“ 3% 2... Emma. .86.. 8mg: mamas:
3. 8m E. a... m. oma :8» m. Emma. .88. Emacs. 2.5.
mm... 2.. 8... on. : om“ 3% W. comma. .33. Emma“: 33352
8.. 2m 8.. mm... m. cm... $9 9. Emma. .38. 8mg: mamas:
«mm Fm mom 3.0 t .88.. SEE «52$
2m m3 5 . .M 8m m. .88.. $8.... moacooos.
Sm ems m2 .3. E .83. ESE as. 85:
om x Pom Nx .: totem 0059.5“me «neon— comaumoamd‘
“2.824. .quHQH. .o 303:5 omD v.84 09C. :53

 

€02-32 :meoaz do 235:0.— ..oaqa 65 a. 3383. mm? :o_.m.omo> .383. .8: 98. E casewbzd

.355... as. .388 3......5.....V use... massage 223.8% 25 23. scam... a. 282.. Ea 288.. .o as»: .8 2%..

80:3... 0:800 .099 ... .0280. .0... 5:808.» 882 a 89:3 8.8:... 083 8.988... 59808.. =<

...mdv m8... on.» 8.08:... ..0 «880. €003 ..0 898:: n ndvuam?
.NEdd.A 08.0 on... ..0 2...» 30:8 ..0 898:: H 87.20..
.88. m8... 0:... ..0 8..» 30:8 ..0 898:: u cor—<0N

.890 m. 89. .8088 .88.. ..0 .x. u 2mmoz<9

 

 

8. ... .8... .8... .8... .8.... 8.80m aomsasm .. Es... .. 58......

m8... . . ... 8.... 8.... 532m .8... 82... 8mg: 088.52 .. 385:.

.8... 9...... m8... 8.... 382.5 538m .888 .88. 8mg: 55. 2.5. . 8.9.52

.8... .8... .8... .8... om. e8. 2A Emma. .8.... 8mg: .05. .....s. .. 088.52 . m8.88.).

2:. 8. ... 3...... mm. ... om. 23. 2-0 ..ommo. .8.... Emma... .25. .....5. ._ 088.82 . 8.95.2

2.... 8.... N. m... 8.... om. as. 3 8mm... .8.... Emma... 85. 2.5. . 3......82 _. 8.9.3:.

m8... m8... 5.... .2... .85. 9a.... 8:88.). . 252% . as. :23.
“.8008. .... . A<0 8. .20 8.025 8.3.. 358...... 3258 38:88 2...: as...

 

+8.8... Smut... ... «32.5.. .8.... 2.. ... Baa... a;

:0:Som0> :88. 80026:... 805.0: ..0 .8.... .8: 0:... 8 3:88.90 80:889. 0:: .908: 359:3...1.8..0...m::3=:2 80.3088
2... us... .588... mean... .3808. ...a 8.88. .8. .20 ..zm...oz<0 08:8 o. 8.5.0.0.... 832.. .o 8:...> .. .85

KG 0.9:...

204

3:008... 23:00 0..9 .: 08:00. 30... :0..80m0> N803. 8......» 00.38:... 203 8.9.2.8.. 800.09.. ..<
.225... ... ......v 25... .08.. ... 3.2.... u 2.9002...
~80... A «8.0 023 ..0 2...» 30:8 ..0 898:: u oo.A..<O..
N80... 2.3.0 02.... ..0 3% 308.0 .0 898:: H 02.20.

$522.0 .3050 ..\.. u meOZ<UN
2.5. 8.3.5... ... 328.... u ...

 

 

205

N3. 3.: 8... SN 2... 2.. 2... 8.8 2 .33. 3.32 5.22.52
8.2 8.2. ...... ...... 8... 8.. 8..“ 2.2 ... 3.23.32 .53....
...... ......N mm... 8... 2... ...... 2N ...... 2 3.23:3 .55...
No.8 2.... ...... ...... 8... 8... 8... SN .. 2.3...» 3233.... 3352 2.2.252
8.». 8.2 .... .N... 2... 8.. N...N 8.2 2 53033233.... 3352 8.22.2
$2 8.8 2... .N... 8... .N... R... .....m ..N 332.... 5.82 2.23.2 3352 2......
...... 8.8 ...... ..N... 2... .... ....m NN... 2 382.... 2.3...» 2.23.... 32.52 23.5.2
8.. 2...... ...... .N... ..N... .N... 8.. 8.8 ..N 2... 2333382383352 2......
2.... R2. ...... ...... .... .N... .....m NN.N. 5N 8.. 23373823.... 3352 32.252
2.8 8.2. .m... 8.. ...... 8.. 8N 8.8 .. 8.. 23373823.... 3352 8.22.2
.3. 8.2 2.. NM... .... 8... RN ...... NN o3 23.N...332.33..32.52 2......
8.2 .N.... ..N... .N... 2... .N... 32 8.8 2 8.. 23.2-.. 322.33. 3352 2.2.252
8.0 .22 ...... P... 8... ...... .....m 8.8 N. 8.. 23.2-. 332.33. 3352 23.5.2
N2. 8.... 2... 2... .N... ..N... 8... 2.8 2 8.23. 238233.3352 2......
2... N22. ...... ...... ..N... 8... RN 8.: 2 32332382323352 2.2252
8.2 2.... .N... N... ..N... 8... N3 ..Nm N. 833.. 2382.33. 3352 2.2.2.2
.2 ......N ..m... ...... NM... ...... .... ....o 2 3.2252... 252%
8.. 8.... mm... 2.... 8... 8... NWN 2.2 2 .3232... 22.3802
8.. 8.8 .N... 3... ...... 2.. SN 82 2 .3235... .2852 52...
0m x 0m x 0m ... 0m x .: 00.8.. 38.98.20 .88... 8.8.0082.
. 2.282.. .. 8.2.... . 22.0240 .... 2823 2.. ..5. 22. ..5.

 

.000. .32 88.90.: ..0 8.0.8.8.. .033 0... :. 00.2.88 as.» 8080m0> .823 .02. 0:... ... 0059020 0.0.8.2...

..5 .303... 852.3... 222. 8.323. 2.2.28.2 2805. .332... ... 2.9002: ..5 83...... .8...... 22.02.... ... 2.32 .... 2...;

508 " Xv

a: .8 203 $03 5:80»? 523 an 323% 25:8 EB mo 53:55 moi—ham .«o 896:: u no

..2v 8.3.». 383° .38.... n 25%.?

~52: A an... Ba .0 an» .o .38.... u 8. A29.

~58. 3% as we we.» go .38.... u 8:29
N88 mun—o cum mo 8% ..o .383: u 3.20"
8% a. as 22.8 328 ..o .x. u 2.9825.

 

 

3s 8.: 2 .o «.2 2 .c :3 2 .o a. Z 38 a. 893.5. .85 Beam
2 3.2 2 .o S... m. .e no... a .o 8.. 8.. 3.2 3 masmufi :28 98F.
o; 2 .2 S 3 2 .o moo no 3. .2 .3. n 532m 20 .85. 39.82
N: a: 03 N3 25 a; do no EN 23 8 3325. 539» €88 .83. 8352
5% 2 .2 2 .o 3... :.o and S .o 3 .. a: one 3 own 33» 2&7 Emma. .85 3mg:
”3 is 3 .o 2. :.o S... : .o ”.2 m: 3. 2 cm“ 9% NE 832 .83. 89.52
2 5mm 2 .o as 2 .o one 2 .o 8.. 8.. 2.2 3 own 38» 3 Emma. .83.. 3mg:
8 x on x 8 x on x 8 x a 3§n§m5 “mono..—

nflg . Nag—Mug awn—8.340

mafia .wfimdddu

 

2a .252 .35 .2525. was: as figs... 8% 388 Mo .36.... SEES. 9:88 €88 5° so:

.3222 Ho 2355.. 5&2 2a a 32 -82 a 33.8.". a; 83%; 3.. Ba. 2...
8:2. 858%... .33.. .3 sea Esme s 698%» 95 8. : .09 28.... €83 3828.. .o .32.... 2a .GSSAE

.8 035.

206

 

.. -. - —
mum“..- ..'_-. -.

Table D10. P values of aprion' contrasts in ANOVAs undertaken to test the statistical
significance of differences in means of CANOPEN, GAP100, GAP>100
and WSDC<0. S in a hypothetical northern hardwood forest in which the
land use strata of primary forest and settled forest, and the 5 strata of forest
disturbance period in managed forest are equally represented.‘

 

 

Strata Compared CANOPEN 2 GAPIOO 3 GAP>100 ‘ WSDC<0_5 5
Managed forest vs
primary forest 0.14 0.11 0.094 0.042
Managed forest vs
settled forest 0.54 0.16 <0.001 0.031

Primary forest vs
settled forest 0.11 0.85 0.08 0.93

MF-logged <12 y ago vs
all other MP strata <0.001 0.31 0.48 0.11

MF-logged <12 y ago vs
primary forest <0.001 0.37 0.31 0.08

MF-logged <12 y ago vs
settled forest 0.041 0.48 0.003 0.031

MF-older growth vs

primary forest 0.63 0.085 0.045 0.38
MF-older growth vs
all other MF strata 0.01 0.46 0.006 0.26

 

125 samples of bird density were taken from each of the 7 strata of primary forest, settled
forest and 5 strata of forest disturbance period in managed forest.

2CANOPEN = % canopy that is open.

3GAP100 = number of gaps in size class 100m2.

‘GAP>100 = number of gaps >100m2.

5WDSC<0.S = number of deciduous woody stems <0.5".

207

 

no. 3.».
3: SM
m.m~ omdm
5.2 58.
as mac-
N: 34.
$2 owN-
laml Inl
Ind—Hal
EN 3;
mmd 26
mod 3.2
mmd 36
m6.— no.2
cod 3:.—
MNN m2 .w
luwl Inl
uJuanqu
818m

«3 8.3
48 a
3. a...
8.8 2.2
8.0 2s
.2 $2..
48 8...
luml INI
Ififlgl

8%-...586
a: 8.:
”3 2.8
3m 3:
:2 z...
92 NE:
2..~ 2.8
8a ...:
low! Ia!
|aa>|

5.: 2.2 0.: ..3 2. 8.2 Beam
3.: ~02 N: 8.8 2. 2885: .85.. ESE
8.2 8.2 N2 2.8 8 £32» 82.. .88 B852
3.: N62 5: won a. 3825-532» 289. 322 39.52
: .2 3 3o ”.2. an own 8% :A 382 see 3882
3.2 no: 2: ..2 2 cm“ 8% NE 882 .82 B882
3.: 3.: N: 8.8 2. own as» 2: B82 :22 8852
Ian: IMI IuW. In: In!
legal IaEual
5055. 8.85285
88 2.2. can NS 2. .82 Beam
8.: 8.2. new :25 2. @885: .82 SE
2 .2 8.8 2 m2: 2 532» 8% .82 3852
...o 3.8 Sn 8.” 2. 8822-53...» 288 use 8852
w 8.8 on... _3. an 8.. 3% 2A B82 .82 3852
a... $22. 3.... o. s 2 own 28» NE 382 .83.. 8852
8.” :8 ”an 42: 2. om“ 28» 2 B82 8.2 B85:
luau In: luau HI Idl
8385.83 legal 43.8833
38.5285 5.82388

 

395m .85.: “.8328: Eustoe “88.2.26 5 _m02<._.m 033...; v3.5a 05 he 2:? :88 v5 .88on ES 2 no .088: :82

.5353 ..o 3858,— 833 05
E woo—-Noo. E coo—atop...» 083 8888 BE .9265— 5 8:88.50 .8:an Ea 80:3 3§€3mfi «88.. “88:6 $88852

._ _Q 035.

208

.8286 25:8 E5 .8 83:5: u :N

.8033; 822358.: + :58. 5055.: - 53a; 2.: 38.5-2822 885% 22% n 8325.

 

 

 

22. 2.8 2.2. 2.22 8.2 2.2: 2. 28.2 82%
2: 3.: 2.2. 22: E 2.; 2. 203035382555
2: 2: 8.8 :22 2: 2.2 2 530232283: 3252
2.2 $8 2.2. :22 2; 3: 2. 82252302288 .85: B252
2..» 2.2 :.:. 2.2.. 23 23 2 omagasAmeozfiengz
23 $2.. 2.9. 32 2:. 0.22 2 cm“ 8.22-8322328252
22 22 2.2. 3.2 E 2.2 8 2.228323258223252
luml IMI Iuml IMI luml IMI Ial
8:: lag Ignquflol lgfiflaaqgl
322.3. .23
232283 .2 E 0:2

209

Table D12. F and p values in Anovas undertaken to test the statistical significance of
differences in means of bird density (Table D3) among 7 strata of northern
hardwood forest representing landuse and forest distribution period. Bird
censuses were undertaken in 1992-1994 in the Upper Peninsula of

 

 

Michigan.
F p
Blackburnian warbler 6.28 <0.001
Black-throated green warbler 6.49 <0.001
Black-throated blue warbler 10.36 <0.001
American redstart 5.15 <0.001
Chestnut—sided warbler 5.69 <0.001
Ovenbird 2.4 0.021
Red-eyed vireo 3.13 0.003
Least flycatcher 6.92 <0.001
Veery 7.82 <0.001
Scarlet Tanager 3.78 0.001
BTAMCS 3.27 0.002

 

lBTAMCS = density difference of: (Black-throated blue warbler) - (American redstart +
chestnut-sided warbler).

210

AR: $28 335:. E 8588 .83 o

5 n: a 8% And «29:8 «o .5855 05 m1:

M35“. 388%... “wee “o sues. m 05 mo :8... a base 860% 3.8% 23 .8 85:: 2.. a «a qsommfiownht> .85 8955
E grow 83 :38 05 ..o 3553 65 .8 22:8 «530:8 05 mam: Baa—=28 $3 on 628 vows—BE 5m .52: 05 .80 Ste 8358 u can

.3?ch =38 u x“

.8323 :80 E 3 a £023 8388 Ho 8858 u a.

 

 

 

:— .o mooo 03 .o 9.. Nvd vmé 3.2 SN ow.» Smooch steam
Eoo a. mo omoo 3; 3.2 so.”~ mm: mod own E8>
oooo of .o o3.o 3mm wmom mmo nooo— cud. “WW—.2 88> 36.3.».—
.ooov _ooov homo 3.8— omdov moo». vodm _ RNm Zoo 5.888% .33
vomo 2 mo ego 2: 36o. omo moNo om.” omwo 39:95
Moo 2 o.o mmmo 3.: oh; «mg. 3.2 36 end 83.33 “cacao—253:0
emoo oooo oooo own _ omom nod 3.3 2 .m hoo— tvaou gotoE<
ammo _ooo mmoo oz. :o— 32 no.3 $6 5.3 5333 6:3 3395-385
mooo Nooo mmvo «2. ovdm mi. 2.8 3.2 3.3 5333 52» 3885-:an
_ooo 52 .o mooo m2 m2 2 s :22 3.3 3.3 8.895 §_E:nxo£m
a m a on x 0m x mom «on
woo—sow m> 330m? 3352 m> lnNM=| win: IIwNMaII
bus—t.— oomafiz SEE “meow 3.30m “mucou— vomaswz 320m ban—to

 

.5322 go 2358.. 3&3 2: a 28782 a Bass 203 3.5 38852

3133 2a .305.“ 85556:. $28 .80 Spam m 88 .328 3.58 .328 .9253 .80 88% £033 5 328 303.5: E055:

30505093 a .«o 887. 8.65— 388 base“. BB 5 moococobfi .80 8>o=< ..o 82.3 a v5 860% ES 2 .80 823% :32

.2Q 03a...

211

 

628 0035:. E 322 550 :0 m>

 

o 2 .o v2 8:: own as» 2-2 A 322 08.8 3252 53:; 32“-...586
$28 vamp—~20 E 826 650 :0 m>
u so? ”3 2;: own £2 a -0 B22 328 3252 5333 83-3586
$28 Emacs: 5 305m 650 :0 m>
u :5? 8.2 2;: $2 2A7 B22 use 3252 552 5055.
$28 vowacafi E 826 850 :0 m>
o 2 .o ”.2 2 _ a: 02 e8» 2.... 322 020.. 3252 525. 8085
828 00358 E 322 680 :0 m>
a so? no? 8 _ 2 cm“ .52 2-0 323 522 3252 5333 33 0282385
:28 woman»:— E 323. .050 :0 m>
a :3 m3 2;: aaew 52285. 3252 535, £55.85
3.028 Sara m>
a ”no one 8:: 532» 522.88. 8252 53:3 3530.85
”.2350 .8 A. m E 38:80 325 28% Em
00:88.33.

 

.5322 00

235:2 3%: 2: ___ 32 .82 s :23 8388 25 220 pa 38 £88 Em 383.8. 20:8 2a .85 825:. as? 8.5..
3:02:66 228 .80 222 n 05 628 3.38 .828 basin .80 326 £033 5 328 303.20: 505.5: 3.350%: a .80 826 8:05.

9.95 282: .0an 23 5 80:20.8% .80 8:88.99.“ Boomoflm 05 “m8 0. 503.53: 895000 £030 305... «>22 .80 323 .—

.ED 03:.

212

.oma flag 2&7

Emmo— 98 .03 E3» 2.: comma: 328 coma—88 mo 38% 05 5 A80 Rd ..mdv £23m €83 30:28: mo :26. 6:3: bug—mama u o
5386 539m 620 5 E85 32% 820.: 39:38 __a E 35a BEBéoBEoé £0380 bias: 8.: .«o 3:33:26 32 u n
.88.: raga 5 BB 820.: comacaE £55 £38» 820 05 E «Hob—So 3:32: 3.: no cognac u a

Engage“. .5ch 8.: sauce—mums?
8353 mm B 38858 :80 803 620.: 3935: 505.5: mo Saba h 0:?

 

68:: Emacs: 5 895m 550 :5 m>

 

0 23 m: ”B g cm» as: SA 3&2 .88: 8mg:
628 “Homage 5 338 5:8 .3 m>
0 8d 3m 3 _ a: oma e3: NE 332 328 3382 $23 8.5m
$28 Emacs: E 33% .38 E" m>
o .23 2; $32 own as: BAX Emma. “wee B352
388 Ema—88 E 886 850 =a m>
0 so? a; a: a: own 33: «3 B32 use: commas 90>
"fiasco é a. m “E 38:80 385 928% Em
cosmocuman

 

ABEESV in 03.;

213

 

mooo

moo

mooo

oooo

~o.o

co

Nvo

_ooov

Nooo

_ooov ovoo

_ _ooo

_oo.ov moo

_ _

_ooov

oooo

_oo.ov

_ooo

cm: 0.80% 2&7 commo— aea Bang—z m>
ow: 300» m; Emmo— um0..on_ Bwaszz

ow: 30% ~— 6 332 605: vowngE m>
cm: 930% m; Emma «008.: woman—z

“008.: 330m m>
60.5.: 38::

538w :02: - “85.: V0930: m>
“00.5.: SEC:

538» 20:88 Emma—S: “008.: 2009302 m>
«00.5.: 3.5.5

om: 0.30» 2&7 comma: 30.8.: 33:02 m>
:83: 95::

ow: 8:0» 2-0 comma. 60:8 Bang: m>
“00.5.: bani:

cm: 030» 2: Emma. 398 3m052 0>
“003.: >381:

 

.0333
33-5535

5202
505i

.0333 03: 53:95 50%

33:03

3:95:85 028.508.: £5532:

3:800 Sabm

 

.5022: ..o 2358: 5&3 05 a 32-82
5:38—05. 803:8 EB 88.: 0:: 8:: 3m .3050 “008.: 30320: 50:95: .20 0:85:88 003:: :_ 050.: .AmUM<Hmo :5
0:00.: 38:0: EB :_ 000:0:0b6 ..o 00::0E:m_m 30:03.30 05 a0. 8 5:552: 300: .8: < >0 Z< oo 00:?» : 2.20:3: 28:09:32 .2Q 035.

214

 

_o.o

_o.o

onoo

_oo.o

vmoo

coo

mooo

_ooov

_ooov

_oov

ooo

bmo

hooo

_

Go

moo

vooo

m_o.o

mwo

cm: 3» m; 332 30:2 33:22 m>
60:2 330m

526.3 .23 60:2 33:22 m>
5303 380m 332:: 80:2 33:52

5303 :020 “00:2 33:52 m>
cm: 30» 2&7 332 80:2 33:22

5263 3:88 332:: 60:2 33:22 m>
cm: 30» 2&7 332 60:8 3332

530.3 :02: 60:2 33:22 m>
om: 2:0» &_ -0 332 30:2 33:22

5263 380m 332:: “00:2 33:52 9»
om: 30» ~73 332 60.2 33:02

om: 3:0» 2&7 332 60.5.2 33:22 m>
om: 30» 2.3 332 “00.0.2 33052

5303 .020 30.2 33:22 m>
ow: €00» m; 332 60:8 33:22

526.3 388 332:: 30.2 33:22 m>
om: «.30» m; 332 “meow 33:52

 

33:3»
B:_m-.=58._0

53:2
503:?

33.33 0:3
3385-285

3383 :83
3:85-285

$3.33
5253285

3:22:80 8.5m

 

A3::u:8o m .0 03:2.

215

 

_

_

cod

mood _

_

Ed

_

mad

_

538m .588 332:: “no.5.“ towns“: m>

828 SEE

_owu E8» 2&7 Emmo— umohow Emacs: m>

.88. ESE

own mac» NTe Emmo— um8£ Ema—32 m>

use SEE

own who» n; wao. .88.“ 3352 m>

$28 basin

 

mug—[Elm

hows—ah
8.30m

b8>

023
83.3.

6:833
.83

3985

33950 8.5m

 

mod

mmd

mood

ovd

mmd

vmd

Sodv

mmd

_oodv

v8.0

mmwd

mood

538m 520 .88.. 39322 m>

see 3.3

532m 388 comma—S. 628 39282 m>

gee 33%

cm“ who» 2&7 Emmo— aeoh Bang—z m>

use 33%

0mg E3» .26 “ammo— amohoh 3352 m>

“8.8 8.38

 

.0333
83-3585

tflmuo.
50toE<

33.33 33
38.5285

.0383 2
3385-;an

85.33
Esafisa

3.39.80 395m

 

ABssaoe

.m _D 05:.

216

 

 

mod

26

Ed

_

mwod

nod

ood

m_o.o

..o

one

5.0

Nmmd

5o

_oo.ov

2

mad

nmd

532m .020 888 town—82 m>
own E8» &_-c commo— umouou Ema—£2

538» 6:88 332:: “88.“ Ema—as. m>
emu mac» & _ -c Emmo— amouom Ema—£2

own 38» 2&7 “ammo— .moue woman—z m>
on: who» N70 .8me— 628 coma—82

538m .620 $28 vows—£2 m>
emu 9.8» m; .8wa 628 3352

539m 383 Emma—E. 328 39:82 m>
own 930% m; Emmo— umohe Emacs).

own 33» 2&7 Emmo— 388 vows—“2 m>
own E8» 9. commo— umoha 3352

own «3% &_ -c Emmo— umea “gown—32 m>
own who.» 2. Bmws 320.2 39232

$28 330m m>
$28 SEE

530..» 820 328 woman—z m>
$20.. SEE

 

mow—(Fm

Emacs
«230m

boo>

023

86.3.

5:883

.33

3955

gnfiou Scam

 

ABEESV .20 2.3

217

A3333 Bvfiésfimozu + 3332 588:2: - €03.83 0:3 339—53026 .«0 bacon 05 fl mumflwmr

 

 

532w 820 328 Ema—«2 m>

_o.o $5 _ _ Sod _ 628 3.80m
£39m 2808 332:: $28 coma—£2 m>

and _ _d _ _oodv _ 328 3.30m
own e3» WTSA Emmo— 623 woman»: 9»

_ mod _ _ _oo.ov _ 528 330m
own who; 26 commo— aoho.‘ towns“: m>

2 .o :3 _ _ _oo.ov _ 620.. 3.30m
own «he» m; Emmo— aBQ 3352 m>

8d _ mmd cod Sodv _ 328 3.30m
:52» .320 $28 3352 m>

_ _ _ _ _ E .o 539» 2.88 comma—a: $98.. town—32
538m 520 $.28 coma—as. m>

as _ _ _ _ _ own 3% m3? 3&2 sea gowns:
532w 9.80m Emma—S. 820.. vows—£2 m>

_ Ed 2 _ _ _ own 3% 2&7 Emma. use Bans:

Saab 33:3 50> 8E 238»: 2355 8388 sea
33% BrvBm .33

 

AB=§=8V .mE 2.3

218

 

02 M: 8.0 «3 3m com 50>

 

 

 

 

mod :6 25 2a: 3.». 8.2 00mg“ 850m
09$ 3.3 na— gi: 3.2 5.91 000$ onéod
9.; $02 :3 3.: 00.? IS 228% .33
05w mm. _ m 2.0 3.02 00.0 0060 23:30
fin 9; men 3% vmd m _ .m 3 003.83 00.08-550.030
wmd ~02 00.0 mm.m as chow tam—00. 505:2
3.0 3.2 3.0 3.3 on.» 3.2 003003 0:3 003933—005
3.2 «3:. 8.3 Quon 8.2 0? 003.83 50% 30885-0.an
mm: NR 5.2 No.3. 0.2 3.3 003003 530303005
0m x 0m x 0m x
NT: Eu: 3 n:
252$ 0.03—.8002 25:32 .853

 

.5w332 .3 03330.— 8an 05 5 32-32 E 303050 0003 3.5
00000005. 0008.3 ESE €903.00. $830683 09$ 28: monaomc§_ m E 860% 23 2 no N80: £803 502 .29 038.

219

.52: of .«o St» 2.85% N am
56:8 :88 n xN
.Amcoafim 3.88 38 3:: @5928 ac .388 u 5

 

 

 

 

 

3m 2m 2 .m 8.: as $3 38>
SN 3 3— 5a «3 .2 E25 3.an
:1: 3w 2: £8 2 .2 t .9: 8% 83.3
N _ a: SN? 2 g: 2.20 2 .E «$3 5:283 :33
m2 2 .E S 0.2: one 3.0: 23:26
8.2 92‘ 2 .2 8w 8. _N mm. 3 5533 8232586
2.: SS 2 .2 3.2 Sam 5% Save 505:2
2 5m mo.” § 2: 3: has; 3: Baasiosm
a; 3.2 2 .w mm. _ m $2 8.2 .2933 52m 332:3an
o o o o o: 8.». caesfiga

0m x 0m x mom Nx

2”: SJ:
:2??on base; guano

 

 

.Swwioaz Mo 28:29— BQQD 05 5 32-32 3333 803 mgm
..moaaomv§_ $.88 85%. ,«o 830:3. 8a .2: 823083 09: 93. m E 360% BE 9 mo Agata 38:3 :82

.SO 033.

220

 

 

 

... 1.5 I .awlhmm . wk

 

 

 

 

 

 

 

Ba. 2:. 31“ a; :2. 2.2 28>
mam 3.2 Ba 3;. 3m 9% amass 8.30m
32 3.2: 8. : :2; B .2 3.2: 8.3 85-8:
8.x 38m 8.8 :32 $8 :2: .2882: :83
a? 3% mg 3.2: 3. : 2% 2585
8:. a." 8+ 23. N2 2. a 5:53 83-5585
:2: $2 2.: 22m 82 no? Eagfiofifi.
v.2: 3.: 8d 3.2 a... _ 8. s .0533 2.3 8282385
9;: 2.22 $2 2.2 3;: 33 2:33 58» B§§-§.m
OS: 38 8.2 3% 8.: £2 5:83:85
0m x mom x 0m «x
2": 2“: 2J4
20>:— Eand 03E§52 $3252
dawanz

mo 2855.: Saab 2: E $9-32 3328 803 2:5 .mogaomu5_ $88 @8323 505.5: coma—BE E 88%

boron 853536 288.2. .«o $3238 8: 35 28:28me on»: can. m E 86on BB 2 mo 3:55:28 $26: :32

.wR— 035.

221

 

 

 

 

 

 

 

m3 3. : 03 no.3“ 3.3. 3.3 50>
G3 8.2 ”an 5. : m3 2.: amass 333
mg SN: 8.: 222 3.2 :3 02$ 83-8”—
92 2 .8“ 2 .8 3+2 8. K 032 .238»: “33
3.2. 8.2 S .2 3.3 8.2 3% 2555
3: £2 3.: 8.3 8.8 8.2 5:33 Bavéaogo
3.2 9.: 32 3.3; $2 8.3 Save 5055.
9.8 8&8 3.2 $2 8.3 3.3 3533 2.5 38%-“.an
8.. _ 3.3. :i a: 2.2 :9. 3:53 82m Baoégosm
N3“ :3 N3 5.: o o §E=§§m
Om X new X om mx
vfina «fin:
ESE Baud Sana—82 mfimmcaz

@8583 .w 5 2.3

222

 

 

 

 

 

 

2% N; 88 3.8 8.2. 8. _N 28>
88 :2 8.". 8.: 2%. 8.” 32:3 238
8.8 2.8 :8 8.2: 8.2 2.8 8.3 85-83.
8.8 8.88 8. G 8.8 8.8 2.8 8:388 «83
2; 8.8 8.22 302 2 .2 8.8 2555
8+ 2.. 8; 8.8 :3 8.8 5.283 888-5585
2.2 3.: 8.2 2 . _2 8.8 8.22 9282 53:22
8.2 2 .3. :2 8.2.. $2 8.2: .0333 2.3 3833.85
8.2 8.: 8.2 8.8 3.2 :8 .0333 :88 B8§¢§E
2 .N E. 82 8.8 3.8 8.8 5283285
0m x 0m x 0m x
v~u= 22":
$2M 2&3 0:28:82 2:252

 

 

885203 .29 038

223

 

 

 

 

 

8.: 8a 2.2 N? 28>
8.2 z: 3.2 2.8 smug: 23%
8.2 8.2.“ 2.8 8.3 08$ 82.82
2.8 3am 5.2 a: 2.28% 283
8.2 8.2: mm. : 2%2 2385

o 38 8.2.. 3333 82.2-3585
3.x :3 3.8 58 2888 580:2
5.2 2.2 2.2 82 Eng; 83 8.822-285
8.2 8.2: 3.2 3.2 5533 52» 38.5-“.an
E .3. 8.2 82:82? .8 o o 8?:335

0m x 0m x 0m x
8.1.: 2n:
.33 2un 2.2252 2:252

gag;

 

82.5.83 .29 033

224

..caoE 2: mo .820 8.8938 n. can

53:3 :88 u xa

22.28% 2.280 28 2:5 wanna—.8 no 238:: u :2

 

 

 

 

 

8.2 2.: 8.8 :2 28>
o: 8.2 2 .v 8.2 52:3 380m
32 8.2 8.; :2 02$ 88.78:
2.2 _ 8.2 8. z: 8.8 2282 :32
:2 8.8 2.: 8.2 2555
85 22 Nam 88 5:33 82.2-3585
8.2 8.8 88 8.3. :88». 508:2
2.: S. 2 8. 2 ~08. 8:33 2.3 82822285
22 8.8 8.2 8.8 5.2? =82 8282385
88858 8: 8. a 8.2“ $2 8.8 82.3285
cm x 0m x 0m x
2,": 8:":
33¢ 2&3. 263252 $22.52

 

88:83 .20 23.

225

 

 

 

 

 

 

 

888 888 88 :88 888 82.8 .8388 .520
888 88 888 .88 888 828 5388 8:88 8888.5
:88 :88v 888 2.88 888 88 28% 2A 882 828“.
888 82.8 :88 8:8 £88 88 3% 2-8 883
288 288 888 888 888 88 28» 3 883
88 :88v 828 828 288 88888. .858 8.88
8888 .888 8888 888 28 888E 828 SEE
«0 38.58 05 55? 22.2888 25 ES. meoE<
8.22893..— Bncgo 82> swans 83.83

.83 85.88 888 828-5585
888 z: .8 8888 28 :88 828.8 5388 820
.888 888 S8 2 8 2 .8 888.. 5388 888 8885::
888 :88v 888 88 888 88 3% 2A 882 828

.88 :88v 888 2.8 82 8 88 .58 2-8 883
888 888 888 888 888 88 3% 2 883

_ _8 888 888 88 82 8 88888. 828 8.88
888 _88 :88v 88 888 88883 .8288 SEE

go 8858 85 555, 8582888 25 28. 98:2.
b8> 232.2 .8383 8383 52m 83.83
888$ 8825-885 8888385 8.58385

 

2528 3:32:86 888 $58332 25.82888 25 v.8. mecca 828:3 EB 8.2388 8m cog—823:: m8>o=< E 82? m

.0265. E 8533:. .8222: 9.8 82.83.85. 828 Bang 55?

.39 038,—.

226

gonna 25:8 BE 8 338. 82: 5:803.» N882 S getaway 0.53 83:5.» 823%? =<
.NEoSA 3:» 30:8 .3 69:3: H 87.36..

.waaao 288 .x. u 29526”

.80: 30:03:00 mo 59:5: _88 n 209—b..."

.35: 223 as 48.505 m8: £28 €35 3: .6 282. u 282:.

 

 

: .o NE 8a 8.8 :2 3m :2 SN 3 “use 338
z .o 8: SN 3.2 :._ 5 3o 2 .m 8 “88.: SEE
2 .o 2 .o ..2 3.” m2 3% N2 ”3. a 532m 3% :85: 3mg:
2 .o m2 2 .N a: 02 OS 3.. :.~ 8 538» .588 commas: .85: 3mg:
2 .o 35 :2 one 2: 2A :3 S: E om.fl 2A? B»? v.28 3mg:
2 .o 23 a: 2.2 8.. ”mm 23 8: a own mag NE Emma. .88.. 3mg:
2 .o 33 SN 2.2 2 ._ 3m 3: m. .N 8 can as» 2 Emma :85: 3352
S .o as t: 2.2 8: 8.” 3o 8.: 8 fies: §§£§m
8o 98 3o 2 .3 _2 NS. 23 _ 2 :N fies» 82» 38.5285
0m x 0m x 0m x om x :
Iqqfldul Jam—maul IlzqudHl flafil

 

535:2 .2355: 5%: 05 a 28782

wimsmcoo BE :5 mfiagm 533%? 59a 2: 8.5 48.5.. 30:63: 805.5: mo Saba. N. we :08 mo :5 .2933 :02» 3395
-483 05 :5 8383 §_E=§83 2: .3 3828 85 53220:: :0 oo_Am<O :5 .meOZflu .ZOUHO... .2002: :82 .30 03:...

227

2:008... 03:00 005 .0 00.000. 0.03 :0020w0> 6:82 5 00:05:... 0:03 0030:? 500.03.» =<

.287 09% 30:00 .20 0008:: n 2: A220.
@5530 30:00 o\o n 7530720.

.0000. 000000300 .20 0008:: .80. u 208.0?

.22.... 0.2.» 05 0025.2. 32. 5028 0:28. 2... .... .38.... n 2002...

 

 

 

mm... 20... so... so... 0303 82» 8.85.020
so? 2... 8m... .30 so? as .0 so... 83. .8...: Exam
2. _ .2. .02. 83. 8... .32. mm. ... .0:. as 228 380..
mm. .o 02:. 3...... $2. 2. . .0 no... :2... £2. 2.320 52° .85.. 30052

.N... 2.0.? mm... 82. ~00 .8? So... so... 2.380 0:88 30025. .380 30082
08... 2.2. mm. ... E .o 0... SN... .8? so? 00.. 3.... 2A? 8002 .850 80.52
02.... .8? mm. .o 3.2. 22. 2m... .8? so... om. 28» 2-0 3000. .85.. 8005.:
so... so? 32. 02... NS... «.8... .80... 80.0 cm. as. m.-. 8002 .850 30052

85.5 29.025 200000 2002.. .833 “200025 .200000 .2005: 085088 8:
gggflaal Jgfifiafial 3.5.3.0083... .5532
5020.2

.20 2:05:03 :03: 05 5 002-80— 5 50030005. 0:00.30. 03:00 0:3 :8 .0 3.03:8 0:5 0:0 msafinm 508%?» 800 0:0 800
..0000._ 00030.3: E0500: ..0 30.00 02.03.05 0:: 003003 59% 000.005-0605 05 0:: 003.33 :0_E:€_003 05 .00 0000 303.00

8 _ £20 ...... 202025 .2008: .2003: ... .8550000 8:002:00 .8033. 2.. 05.5.8002 m<>02< a 8...: 0

.25 030,—.

228

0.0.08 0.5.9.. 2...... .35.... .5... n 2002..
.08.. 3800...... ... .35.... ....o. n 2005..
.50.: 05 ..0 00:0 0:00:50 n 00.H

0000.0 .0080: :02: u ....

.0..:: 0:38.030 000:5: u :.

 

8.0 E... 0. 0.: R0 88 ...... ...... 0008.05.02... 8.0.0

and .0. 86 ..NN 9..» find N0 N..N .~.w .N 3M hm... 000300520000. ESE...

0:... 0.. 3w... tam 0w~.w 2.0 ..0. mm. .0.0 h 0:... 0...... 00000005..00:0..00w0:08:00.3530.&000_0

 

 

0S... .0... 8. .N 00... v 00... 00. 0.... .0.0 .0 . 8.0 .0380 .00... 05.....on .00.... 808...).
0 00 x 00 x 0 : 00 x .: M00 .0.
0.0.0.4003 Ififlufil 00.80.3000... I033?
8.8.5.8... 5250...... B...8...-..8.m 5.50385
.28.)... 20000..

 

 

3.0.5.2 ..0 3:05:02 .000: 0.0 :. 002-90. :. 50.0.0005. 0:000:00 0.... 0:0 800.000.,
£0... 0... 0.0a .8000 00.0.. 00030:... 605.0: 30.0.0.0 0:. :. 00.0000 0.... N 05 503.00 2002.. 0:0 2000.0... :. 000500.00 ..0 2.8.0.50
08000.30 0... 0:08.05 $5072 ..0 000.9, 0 0:0 .020. 00030.0: 80.0.0: ..0 0.3.0 .00: 0:0. :. 8:00.030 30090.... 0:0 .00000 00:30:00
.020... 30.0.0.0 :. 003.03 :00.» 00.000.00.00... 05 0:0 .0300». 5.05.8.8... 0.0 .3 00.00.00 0.835.200: ..0 20024 0:0 2000.0... :002

.NNQ 030...

229

if
E
m

 

Bod mad :3 S magmas. use 8.3%

 

 

£3 Eco So 8 39835 :28 SEE
as v8.0 9: .o mm 595.. €9on 3% use 8332
m _ ed 83 $3 8 use :EEm .588 use 8332
m3 599 82 S cm“ e8» SA Emma .35. 8mg:
08.0 Eco so? mm own man; «To 832 .85 8mg:

.8? S .0 v2 8 cm“ as» 3 832 52$ 8mg:

<38 mg 35m = 8:2. 8§eame mean

..de gem €83

 

.gwEonz mo 28:8”:— bnaD 2: E 32-nao— E cog—.3525
mEEEwm co_§omu> Ea manage 95 Eat 203 8% 523%? 28 8365“. RE .836 .835. E 35936
3.652. was .votoa cocafaflv 328. n 2: Sm ..mdv 2:06 383 so 2268 5333 3233585 23 Amy—32v
Suva. 585:2 559:: 8333 0:3 3885;.an Mo 2238on 882.8 E 356508 coamflwfl .8 mos—S, m .89 usual

230

Neg—A 820 on: mo 8% :o 82.5: n Ago—Ania.
.22: 820 on: :0 Ram .3 893: u 8:20h
.NEOm mas—o 2% .«o 8% mo 82:3: u o???
.83 2 :2: 32.8 :83: :o o\.. u 550720n

...n 53A 8.8m 383 :o .352 u 269%?
...mdv man 3003 mo 825.: u WOVUGm?»n
.52: 2: .«o .828 23:3» u 2w

 

 

 

 

 

.508 n x.

v.0 0N; 05¢ 55.0 Nvd mvd oo—AQO
05.0 whd :.o and 00.0 0.0 Ago—$0
ONO :.~ :16 N_.— 3.0 5N; Omm<mv
mod oo.wN Rb.— hmfim MK.— —w.@— ZMQOEU
wmé :N San Yom mvN MONN mdAUQmB
me SEQ. ©m.m ohmh ~06 #900 deUQmB

3. x 8 x 3. x

Ru: own: 3n:
<3m0 g amhm

 

.5383
.«o «_smziom 8&3 2: 5 32-32 mats: 828% 8.88 EB _Nm 5 2.8.3895 mamas—8 =0:Sowo> :5 8.388
EB So: “8:350 22$ «:5 Nev—EBB 3A 83 868% BE 3:038: 2: :0 £80: 2: £023 8 28:8,.“ 888 EB
9508 2352. :o 8:388:23 2: 3 8:982 083 860% a 3 .8828 3:32 :o 88.58820 .8333 3223.85
93 2382 528:2 .8383 0:3 8385-285 2: B 8828 2352 5 8.8.53 88. £5: :5 39.8 :o 382

<8 032;

231

...mdv £53 383 go 398:: n Devon—wk,n
.NEoSA mam—o cum .«o 8% mo 398:: n 2:920.
.NEoS 3.20 9% mo 3% mo “38:: n 02.2?
.NEom 9.20 on? mo 8% mo 398:: u emm<0~
.53 fl 3.: 3058 $28 «o .x. n 295720.

 

 

2 80¢ a 8 N: .o <38 2, mg an;

2. 83 a 2 Sod <38 a, 25.8 was

2 53.0 2. 2 :3 g m> 258 Ed.

and 8o... $3 23 83 <>oz< __ego
“novuama ,ooxego “8:20 N035 _zmmozfi 8338 £85

 

53:22 .8 2:25; 3&5 2: 5 a2
-32 5 So toga mEEES 538%? 98 8328 EB 89c 35830 803 3.5 $558 $3.33 33-55.85 v5
Amy—$7; 5360.. c855 ABM—Ha .5333 0:3 vofioEHian 2: .3 @8028 3852 “8.5“ 39.63: Panto: $88.»
8% 30:8 28 .EoEnto—u 53— 257. $553 30:8 8388 2 coxatovcs 33. 3:53 5 m<>OZ< .«o 8:?» m .mNQ 033‘

232

.NEoeA 80m 3008 .00 000800 .I. Ave—Ag?
.3008 0000 o\.. n 7530720.

 

 

80.0 23 cm“ 3% SA 8002 use 5 mg a. 30.5 52300

23 82 cm“ 083 8-0 8002 .880 s <38 0 00:2 .380 08.5

32 #3 cm“ 083 2 8002 0.20 0 mg 0 30.8 80300
mEoSAZc 2000720 80,0538

 

08232 .00 0.85003 0033 05 E 32.82 0_ 000—8000: 3:088 00080wo> 000 83800 003 0.0.0 005800 0003
8.5 80000380 wEwwo. 00008 3000000000 880.. 000000800 0800.“ m 5 003.83 00008050000 000 .0880. 08.00050
005.33 003 080005-803 05 300.0 3% 3008 000 w0_0000 3008 00000000 00 000—80000: m<>OZ< .00 m020> n— .0NG 0303.

233

...00v 8.30 008300 .3030 u 0.90003.

.0050? 0000 3050 00 2:30 u 8.0.00.

N050. 000.0 00.0 0. 0000 3008 .00 000.000 u 00. 3.0.00

.388 8% .x. n 2000200.

.0000. 0.0 .00 00.00 000080 n 00.H

0.00.03 ..0 00.8 0008 n 00

.~00<000.0 00A 003 00.0000 05 .00 3.0000 00.. 00.03 .0 000080 000000 00.0 .00 00800 n 0.

.3008 00A 33 0303 300-0580 20 00 0080

0.0 00.03 00.. .000 0000» 0-. 0030.. ..0 00.000 05 0. 0000000000 003 008 . .000 Ame—00000.0 00A 003 8.0000 030000000 0000 .00 3.0000 00.. 00.03 00
0000000000 00000.. 0000 0.5.3 000080 000000 00.0 00 00.00000 000800000 .000 003000 0.03 000..» 00.0000 0000 00.. 00.00.00.» 8.0008000 .00 000.0.» 00006..

 

 

.m mm o . 0.0 Wm 90.. .0.0. N <3mu

0. .3 00.2. mmd who and and Raw .0. . N ... 5.2.0

5.0. 0.00 0. .0 $0 @010 mm.. 00m 00.0. 0. BmHm
3%

00.2 New .N... 00.0 30 00.0 mw.m 0. .2. m . 0.2/mo

0.0 00.00 m . .o wmd 00.0 mm... and 0~.0~ 0N . was?

0.0. 0.00 v. .o 30.0 v. .0 mm... ...m 0m.- cm 395
3%

0000000000.. .00 0 $060

m. .om 2.00 mm... mm... o o 00.... m0..m 0 was?

N00. Sam 000 00.0 vmd mmd mmd 00.5 0 amhm
3g

00 x 00 x 00 x 00 x 0 80.00 00000..
4% Jamal 400000.. 083.50 .5532

 

00800:). ..0 0.000.003 00003 05 0. 000. -30. 0. 09.800000 00:00.80 00080m0> 000 00000000 00.0 80.0 000.800 0003 809 00.30.
.58. 00.. 080% 00:20 2:0 0 0008028. 320 0880 80200.. .558: s 2300. 0.003 300-0586 08 0002.3 5002
000.0002 ..Bmhmv 00.0003 00.0 00.00.50.020 0.0 30 00000.00 008.000 0. 00va 000 607.20 60.3.5 .meoz<u ..0 0000.2 Nam. 0.00...

234

Table 028. Contribution of different land type associations and forest disturbance periods to
different groups of bird census stations aggregated based on forest cover within 1 mile
(FC lMIL), length of forest edge EDGE, and shrub layer development, WSDC<0.5".
Data were derived from vegetation sampling in 1992-1994 in the Upper Peninsula of
Michigan and GIS coverage of study areas current for 1992.

 

 

Groups #BCS HUM MCC SYL DUK MUN MAN RAP
Forest Cover 1; Edge 1; Shrub l 26 3 4 4 1 11 O 3
Forest Cover 1; Edge 1; Shrub 2 27 2 5 l 3 13 0 3
Fomt Cover 1; Edge 2; Shrub 1 12 3 O 0 4 5 O 0
Forest Cover 1; Edge 2; Shrub 2 9 l l l 2 3 O 2
Forest Cover 2; Edge 1; Shrub l 21 0 0 5 l 0 3 12
Forest Cover 2; Edge 1; Shrub 2 15 0 0 O O 0 6 8
Fomt Cover 2; Edge 2; Shrub l 14 O 0 O 2 3 4 5
Forest Cover 2; Edge 2; Shrub 2 18 O 0 0 0 0 15 3
Forest Cover 3; Edge 1; Shrub l 8 O 0 0 0 0 5 3
Forest Cover 3; Edge 1; snub 2 1 1 o o o o 1 6 4
Forest Cover 3; Edge 2; Shrub l 9 O 0 O O O 5 3
Forest Cover 3; Edge 2; Shrub 2 ll 0 0 0 0 2 6 3

 

235

 

Table D28. (continued)

 

E I; . l E . l
Lg Lg Lg Unl Older
#BCS PF l-5y 6-12y >12y SG G

 

Forest Cover 1; Edge 1; Shrub l 26 12 6 0 0 3 5
Forest Cover 1; Edge 1; Shrub 2 27 11 2 0 9 0 5
Forest Cover 1; Edge 2; Shrub 1 12 3 5 0 2 0 2
Forest Cover 1; Edge 2; Shrub 2 9 5 0 0 2 0 2
Forest Cover 2; Edge 1; Shrub l 21 5 5 3 0 7 0
Forest Cover 2; Edge 1; Shrub 2 15 0 2 8 0 l 4
Forest Cover 2; Edge 2; Shrub 1 14 O 2 3 3 5 1
Forest Cover 2; Edge 2; Shrub 2 l8 0 0 6 9 0 3
Forest Cover 3; Edge 1; Shrub 1 8 0 5 0 0 3 0
Fomt Cover 3; Edge 1; Shrub 2 ll 0 1 1 4 3 2
Forest Cover 3; Edge 2; Shrub l 9 0 5 0 3 O 0
Forest Cover 3; Edge 2; Shrub 2 l l O 1 2 5 3 0

 

‘Forest Cover 1 FC lMIL = >90%. Forest Cover 2 F ClMIL = >80-90%. Forest Cover 3 FC lMIL =
>70-80%.

2EDGE 1 EDGE <2000m. EDGE 2 EDGE >2000m.

3SHRUB l WSDC<0.05 = <50. SHRUB l WSDC<0.5 = >50.
‘HUM = Huron Mountains.

5MCC = McCormick Wilderness Area.

6SYL = Sylvania Recreational Area.

7DUK = Dukes Experimental Fomt.

8MUN = Munising.

9MAN = Manistique.

loRAP = Rapid River.

“PF = primary fomt.

12Lg l-Sy = managed format logged 1-5 years ago.

l3Lg 6-12y = managed forest logged 6- 12 years ago.

l‘Lg >12y = managed for0st logged >12 years ago.

15Unl SG = managed forest unlogged second growth.

l“Oldver G = managed fomt older growth.

236

 

Table D29. P values in a 3-way ANOVA, indicating the statistical significance of effects
of F ClMlL‘, EDGE2 and WSDC<0.53 on densities of black-throated blue
warbler (BTBW), American redstart (AMRE) and the variable BTARCS‘.
Bird densities and vegetation were sampled in 1992-1994 in the Upper

 

 

Peninsula of Michigan.
Effect BTBW AMRE BTARCS

FClMIL 0.028 O 0.001
EDGE 0.101 O 0 g

WSDC<0.5 0.003 0.001 0.98 E

FClMIL X EDGE 0.056 0.001 0.00] i.
FClMIL X WSDC<0.5 0.025 0.97 0.237 !
EDGE X WSDC<0.5 0.627 0.012 0.089

FClMIL X EDGE X WSDC<0.5 0.022 0.64 0.023

 

1FClMIL = group based on forest cover within 1 mile: >90°/o; <90->80%; >70-<80°/o.
2EDGE = effect of group based on length of edge within a 500m radius: >2000m; <2000m.
3WSDC<0.5 = effect of group based on number of deciduous woody stems: <50; >50.
‘BTARCS = difi‘erence between the density of the black-throated blue warbler and the
combined densities of the American redstart and chestnut-sided warbler.

237

bill!

 

 

3m :3- S: 32 2: 8.2. S a? novooma €53? u .546»
28 q. 8. <2 “.8 68 Sm : 8v Woven; £85? n ,Efiom
man a? 0.: Sn 8.: 2: 2 8A «.3083 £52? u $sz
8.2 mg v. : 2: S; >2. mm 8v novooma €085? n dEB
man «.8 ”.2 a? v.2 ”.2 on SA novuoma mgo? n ,EEB
N8 o: o. : he 8.: 8.2“ mm cmv novooma ”.80? u 152::
Q8 Ed. 8.: 32 «.2 08 on :58? monm @988? u Egon
”.8 SM «.2 08 ”.2 m2 2 888v meow @837 u diam
8.8 v. _- w. : can “.2 «.8 NM :58? meow $8-0? u 49:8
we 3” a: a: «.1 32 on 888v moam $8-0? u 5:8
08 SN 0.: Zn 3: as 8 888A meow deco? u 526;
0.2 0.2 3 8.3 5.: we. 8 588v moo”,— §8A n 4928
8.2 ow E 0.8 8.8 on” a 0? n 090nt
S: 2 3. o.: M; 8.3. 8 8v u novuoma
8.: N3- 3 o: 2: Sn 2 :58? n momm—
md 03. 3 EN 3 N2 we 883v u meow
S: of: 2: in a: n: on 9888?"ng
N: w: 2 3m 2: 3w 8 $8-0? n 49:8
0.2 3: 3 ma _.: a? E ego? u 43::
8 x 8 x 8 x mum» 3:90

 

Eudobfi m .3 $2? :88 28 .Amgv tau—x: 523:2 28 555.5 3333 0:3 BEoEfixoa—n mo 838% :82

.cnwiog mo 23.—Eon 5&5 05 E woo—-32 E 3383 203 538%? v5 828% gm .mdvUDmB
mo m_o>o_ N 28 .mOOm .«o £26. N {Ea—Um mo 232 n so woman cogocsabm a E @258 2.23% 328 BB «0 8:on

.oMD oBuH

238

 

.52: on: .6 :88 23:8: n 8:

56:0: 52: n x:

8:033: 25:8 “:3 .3 33:5: u mom—mm

:58: 36:3 25:28: :0 :83 anew n m.ovUDm3..

:38: 25:8 23 EB: 338-80% : £523 emu: :0 EMS— :o :83 ESE H ”759mm

.0::— _ :32? :98 .86.: :o :88 98% n. AME—0%

.8323 32855823

:5 2830.. 585:2 05:: 82:5: 8:388 on: :5 3333 0:3 3885-383 2: :0 56:0: 2: 803:3 85.5%: n mom‘rm.

 

 

:8 $2- :8 :2 :8 :8 : 8A 3va J:88A man: 885? u :26:
:8 :8. :8 :8 :8 v.8 : 8v 3%? mecca? mom: m88.0? n :26:
:8 : 8 :8 :8 :8 :8 : 8A :32: €88v m8: m88.0? u :26:
:9. :8 :8 :8 :8 :8 w 8v 3va €88v mom: €88? u :26:
:8 :8- :2 :2 :8 :8 2 8A 3va €88A moo: m92.228 u :26:
:8 : _ m :2 :8 :8 :2 2 8v 3%? €88A mom: M88.28 u. :26:
:8 :8. :2 :2 :8 :8. 2 8A 3va m:52an mom: m88.98 u :26:
:8 w. 8 :2 :2 :2 :8 8 8v :0va maoo8v :09: 885% u :26:
:8 :9 :8 :8 :8 :8 a 8A :0va M:58? :09: 88A u :26:
:8 :2 :2 ...: :8 :8 2 8v :32,» mecca? :09: 88A u :26:
:8 :2 :2 :8 :2 :8 8 8A 2va maoo8v mom: J88x u :26:
:8 :o :2 :2 :2 :8 8 8v :OVmB J588v men: 88 u :26:
:2 :8- :: :2: :2 :8 8 8A 3683 :58? men:
:2 :2- ...: :2. :2 :8 8 8v :ovooma €88A man:
:2 :8 :8 :8 :2 :8 2 8A :36nt m888v man:
:2 :8 :2 :2 :2 :8: mm 8v :ovoama meoo8v mom:
3 x cm x on x mum: 3:80

mUMQHm rail amid

 

685:2: .8: 2%:

239

 

 

:0353 32:23:80.3

:5 58:0: 50:05 05 mo :0E::0: 5:388 05 :5 :0353 0:3 35858.53 0.: .3 32:5: 05 503:0: 350%: n mug???

 

 

mmmo Nmoo mmNo moVUQmB x 2.9—OED x mOOm x §_Q>>EZ
mood whoo mmmo moVUDmB x 2.9—0720 x mOQm
:oo mhmo mood movUDm>> x meOZ<U x HER—332
omoo wovo mNNo movonma x mODm x :52:ng
mood on— .o omoo meOZ<U x NOON x E2;
mmvo meo moo movUDmB x meOZ<U
woo m. ..o ciao moVOQmB x MOON
hhvo Rod :5 meOZ<U x NOAH-m
wood as .o goo movonma X AME—DE
cmoo v:.o :N. .o meOZ<U x 52—ng2
of .o whoo _Noo mOQm x AHSZDBEZ
Nooo w_o.o _ooov moVODm>>
ommo v. _ .o oooo meOZ<U
wmvo :2 .o o—wo mOQm
:mo mm. .o ooo anal—Q;
moghm MKS? BmHm :95..—

 

.NoE .8: 808:0
5:222 :o 2:88: .08: 2: s :83 :8: :o :2: =5: :5 82-82 a 8:588: :25: 8888., :8 5:38
E3 89: :0883: 003 San Amovonmav :83: 30:3 :0 538:: :5 .meozlxo 880:: 3:50 8. donm 0w:0
88: 2:2 2 :32: 3:2 2:. 2 :52 :20: 82,52 505.8 :o 8 8 83: 8:2: :85 3888 283 858
035:? 0.: :5 58:0: 50:05 :5 c0353 0:3 8585-053 05.3 83:5: 0.: 3033 :_ $5072 53 v :_ 83:3 : 2mm 0.5:.

240

 

 

 

::.: ::.:: ::.: ::.:: ::.:_ ::.:: :: o:A :.ovoo:3 mo:A.::,::322
::.: ::.2 ::.: ::.2 ::.2 :3: :: o:v :.ovoo:3 5:335:35
::.: 2:: ::.: 2:: ::.:: ::.:: :_ o:A:.ovoo:3 m9952535
::.:. 8.2 ::.:. 8.2 ::.:: ::.: :: o:v :.ovo::3 mova—253:2
::H: _:..:: 8H: 3:: ::.2 3:: S o:A:.ovoo:3
2 2 R: ::: ::2 ::.2 ::.:: 2 av 933:3
:3: :8: ::.: ::Ho: ::.: ::.:2 :: $229525
:2 ::2 2w: 3:: ::.:: ::.:: :: 22325025
::.:: 8.2 :..: 3:: ::.2 ::.:: :: 8:8:Amoam
:: 2 :: :: :5 :: :: ::.: :::.2 :: accovmmoom
::2 ::.t ::.: ::Ho: ::.: ::.:: :2 835535
3:: 8:: $2 2:: :2 Z: R 5.9335322
0: x 0: x 0: x mum: 3:90

3%

 

5:222 ..o 2355: 5&2 2: s ::2 :82 :55: 3.98:
553%?» :5 m8:::50 ::3 82o 05 35:03 :50 .2358: :::50 E3 2 55 038 5:388 55 :5 0052.333: .3 35.
..o 05 5 83:3»: 053 55 :355 988: 2350828 80:030. :5 ::.:2w .3 :508 3:0 33 555:8: 05 5905 .3 $00.33 58
.5852 :5 80:3: 32.8 E: :o 2:2: __a ..o 285 85:: :33 .:.ovuo:3 :5 2:325 .moom J22 5322:: 38:: 2:

:2 2 :8: 23:3: 33-: 5 .3558 ..o :5 .855: 5:3. 5:85 :5 .22? 33 385-435.: 88:. was: .83 :8 03::

241

 

 

 

 

3.3 35 no.3 I .3 2.2 3.3 mm omA WoVUQmB gems—«29322
2 .2 MN: vms 3.2 3.2 82m 2‘ omv deUQmB mgomfi_2_n>»zz
2 .Nm 250 8.3 2 .3 «ER 5.2; 2 3A ndVUQmB §omva=2_§mz
002 2.3. 3.2 3.2 Sfim 3.x mm owv Devon—ma 2x69225332
2.: «me was 3.3 3.: 3.3 we e\..2AZ”..EOZ<U gems—253:2
2 .2 ~93 cow 2.: on: 3.6m mo 3m _vamOZ<U m$om§_25322
when 3.: 2.2 2 .3 co; m $62 2 $n_AmeOZ<U .bed—zaamz
MNNN cod- 3.2 _ fem 21K «mam mm $2VmeOZ<U gomvdzaamz
mm.wm 2 .Nm 3.2 3.3 3.3 v9.2 2 omA Wovooma mooomA mgm
2.2“ end. 2.: 3.2 3.2 Sam om env 090nt mooomA mOOm
3.3 3.3 mmd a2” 82 undo— ? omA Devon—25 mooomv mOOm
3.2 B.» on.» 3.2 3.2 modm S. omv Devon—m2, Mooomv mgm
2 .3 3.2. No... 2.8 ~32 mg: on owA deUDmB m$295.5720
35m $6 3.3 3.3 whom omwm em omv Wovuama MG\e2A7m50720
909 5.2.. 2.2 we; m vmdm 90% mm cmA OoVUQmB M.x.2vaEOZ<U
3.2 2.7 on.» E .2 a2 34% av cmv Wovuama ”gm—vamOZ<U
cm x 3 x 8 x mum.“ 3:95
Goa—5:03 .Nma 052.

242

2.03 3003 303200: :0 :08: anew n Wovoflmzr
m:_:0:0 20:00 :0 :08: anew n meoz<u .

.:0m§m 25:00 EB 80¢ 3:08:58 0 £53» 0200 m0 522 :0 :83 anew u mamm—M

mavm: 0:8 2 :23? 60.0.2 00033:: 505.5: .20 ..\o :0 :83 anew" 22532?
A8533 00233585 + tau—00: 505:2 .«0 33:03 - $3.33 02: 0800:5283 m0 26:0: u mumEb.

 

 

FNNN omdm- 2 .2 coda o~.2 2mg 2 SA 20va xxi _AmeOZ<U Moms—23522
23m 320 8.: 00.3. 2km and: 2 omv 0°va gm—AZm—moz’xu moms—.222
2.2 ”0.2” hm.» m~.mm 3.2 «NS. mm omA new? M..\..m_vszOvZU STE—25322
$2 3.3 2..” 3.x 8.2 32m _m cmv ndeB m..\..2V7m50720 moms—.2922
2:3 2.3- 3.3 3.2 mmdm 2.2 2 omv demB xx.2va.—OZ<U memq_2_9$mz
343 8.3. 3.2 $6» 2 .vm v0.0: 1 $2AZmA—OZ<U mooo~A moom gems—2—QBIZ
$2 3.3 2..” N: m 3.2 5.8 _m {on _vaEOZ<U mooomA momm— gems—253:2
2 .cm «can 3 .o Nadm «N2 8.3 mm {on 7me0720 mooomv mOQm Mgown—.253:2
3.2 2 . _ m nmw oo._ _ and _ ~05. mm ..\om VmeOZ<U 68V mOQm gems—25312
Nwém omfim o~2 2.2 2 .3 «man om o\om _vamOZ<U Mooomv moam m$omVA_E_D\E.—Z
0m x 0m x 0m x mum», 3:05
lama 32¢ Emma

 

ABEESV an: 23

243

LITERATURE CITED

Agee, J .K., and DR. Johnson. 1988. Ecosystem management for parks and wilderness. .
University of Washington Press, Seattle. ‘4

Ahlgren OE, and LP. Ahlgren. 1983. The human impact on northern forest ecosystems.
Pages 33-51 in S.L. Flader (editor): the Great Lakes forest an environmental and social
history. 336 pp.

 

Albert, D.A., S.R. Denton, and B.V. Barnes. 1986. Regional landscape ecosystems of
Michigan. School of Natural Resources. University of Michigan 48109-1115.

Allen, T.F.H., and TB. Starr. 1982. Hierarchy perspectives for ecological complexity.
The University of Chicago Press. Chicago.

Allen, T.F.H. and T.W. Hoekstra. 1990. The confusion between scale-defined levels and
conventional levels of organization in ecology. Journal of Vegetation Science 1: 5-12.

Allison, TD. 1990. The influence of deer browsing on the reproductive biology of
Canada yew. Direct effect on pollen, ovules, and seed production. Oecologia 83: 523-529.

Alverson, W., D. Waller, and S. Solheim. 1988. Forests too deer: edge effects in northern
Wisconsin. Conservation Biology 2: 348-358.

Ambuel, B., and SA. Temple. 1982. Songbird populations in southern Wisconsin forests:
1954 and 1979. J. Field Omith. 53: 149-158.

Ambuel, B., and SA. Temple. 1983. Area-dependant changes in the bird communities
and vegetation of southern Wisconsin forests. Ecology 64: 1057-1068.

Anderson, SH. 1981. Correlating habitat variables and birds. Pages 538-542 in C].
Ralph and 1M. Scott (editors). Estimating numbers of terrestrial birds. Studies in Avian
Biology No 6. Allen Press. Lawrence, Kansas.

Anderson, SH, and H.H. Shugart, Jr. 1974. Habitat selection of breeding birds in East
Tennessee deciduous forest. Ecology 55: 828-837.

244

Anderson, RC, and CL. Loucks. 1979. White-tailed deer (Odocoileus virginianus)
influence on the structure and composition of T suga canadensis forests. Journal of
Applied Ecology 16: 855-861.

Arcese, P. and A.R.E. Sinclair. 1997. The role of protected areas as ecological baselines.
J. Wildlife Management 61(3): 587-602.

Askins, R.A., M.J. Philbrick, and D.S. Sugeno. 1987. Relationship between the regional
forest abundance and the decline and recovery of forest bird communities. Biological
Conservation 39: 129-152.

Askins, R. A., and M.J. Philbrick. 1987. Effect of changes in regional forest abundance on
the decline and recovery of a forest bird community. Wilson Bulletin 99: 7-2]

Askins R.A., J .F. Lynch, and R. Greenberg. 1990. Population declines in migratory birds
in eastern North America. Pages 1-57 in D.M. Power, editor. Current omithology.
Plenum Press. New York.

Baichich, P]. 1997. A guide to the nests, eggs, and nestlings of North American Birds.
Academic Press. New York. 347pp.

Bailey, R. G., R.D. Fister, and J .A. Henderson. 1978. Nature of land and resource
classification-a review. J. For.76:650-655.

Bailey, KG. 1985. The factor of scale in ecosystem mapping. Env. Management 9: 271-
276.

Barnes, B.V., K.S. Pregitzer, T.A. Spies, and V.H. Spooner. 1982. Ecological forest site
classification. Journal of Forestry 80 (8): 493-498.

Bent, A. C. 1963. Life histories of North American wood warblers. Part 1. Dover
Publications. New York. 367pp.

Binford, LC. 1991. Black-throated blue warbler (Dendroica caerulescens). Pp 402-403
in R. Brewer, G.A. McPeak, and R.J. Adams: The atlas of breeding birds of Michigan.
Michigan State University Press. East Lansing, Michigan. 594pp.

Blake, J .G., and J .R. Karr 1984. Species composition of bird communities and the
conservation benefit of large vs small forests. Biological Conservation 30: 173-187.

Block, W.M., D.B. Finch, and LA. Brennan. 1995. Single species versus multiple species
for management. Pages 461-476 in TE. Martin, and DB. Finch (editors). Ecology and
management of neotropical migratory birds: a synthesis and review of critical issues.
Oxford University Press. New York.

245

u .1!- -hdr’fl

 

at“ -—-—
. . .1.
_ ‘ . .'. . -_-
___. -

Bond, RR. 1957. Ecological distribution of breeding birds in the upland forests of
southern Wisconsin. Ecological Monograghs 27:351-384.

Bormann, F.H., and GE. Likens. 1979a. Patterns and processes in a forest ecosystem.
Springer-Verlag, New York. 253pp.

Bormann, F.H., and GE. Likens. 1979b. Catastrophic disturbance and the steady state in
northern hardwood forests. American Scientist 67: 660-669.

Bourdo, EA. 1983. The forest the settlers saw. Pp 3-16. in S.L. Flader editor. The Great
Lakes forest: an environmental and social history. University of Minnesota Press.
Minneapolis. 336pp

Braun, EL. 1950. Deciduous forests of Eastern North America. The Blakiston Co.,
Philadelphia. 594pp.

Brewer, R. I991. Original fauna and postsettlement changes. Pages 33-58 in R. Brewer,
G.A. McPeak, and R.J. Adams. The atlas of breeding birds of Michigan. Michigan State
University Press. East Lansing, Michigan. 594pp.

Brewer, R., G.A. McPeak, and R.J. Adams. 1991. The atlas of breeding birds of
Michigan. Michigan State University Press. East Lansing, Michigan. 594pp.

Brittingham, MC, and SA. Temple. 1983. Have cowbirds caused forest songbirds to
decline? Bioscience 33: 31-35.

Brubaker, LB. 1975. Postglacial forest patterns associated with till and outwash in
Northcentral Upper Michigan. Quaternary Res.5: 499-527.

Buckland, ST, D.R. Anderson, KP. Bumham and J .L. Lake. 1993. Distance sampling.
Chapman and Hall. New York. 446 pp.

Canfield, RH. 1941. Application of the line interception method in sampling range
vegetation. J. For. 39 (4): 388-394.

Canham, C .D., and O. L. Loucks. 1984. Catastrophic windthrow in the presettlement
forests of Wisconsin. Ecology 65 (3):793-809.

Cody, ML. 198 1. Habitat selection in birds: the roles of vegetation structure, competitors,
and productivity. BioScience 31(2): 107-113.

Cody, ML. 1978. Habitat selection and interspecific territoriality among the sylviid
warblers of England and Sweden. Ecological Monographs 48: 351-396.

Cody, ML. 1974. Competition and the structure of bird communities. Princeton
University Press.

246

Collar N.J., and P. Andrew. 1988. Birds to watch: the ICBP world checklist of threatened
birds. ICBP Technical Publication No.8 Smithsonian Institution Press. Washington, DC.
288pp.

Cook, W.W. 1904. Distribution and migration of North American warblers. US. Dept. Of
Agric. Division of Biological Survey. Bulletin N0 18. l39pp.

Crow, T. R., A. Handy and D.M. Waller. 1994. Report on the scientific round table on
biological diversity convened by the Chequamegon and Nicolet National Forests. USDA-
Forest Service, NC-166. 55pp.

Crow, TR. 1995. The social, economic and ecological significance of hemlock in the lake
states. Pages 11-17 in G.Mroz and J. Martin (editors). Hemlock ecology and management.
Proceedings of a regional conference on ecology and management of eastern hemlock held
in Iron Mountain, Michigan Sept 27-28, 1995. Published by the Department of Forestry,
University of Wisconsin, Madison. 200pp.

Cunningham, R.W. and HG. White. 1941. Forest resources of the Upper Peninsula of
Michigan. US. Dept. Age. M.C. Pub. 429.

Curtis, J .T. 1959. The vegetation of Wisconsin. Univ. Wise. Press. Madison. 659pp.

Dawson, D.K. 1979. Bird communities associated with succession and management of
lowland conifer forests. Pages 120-127 in R.M. DeGraaf, and K.E. Evans, compilers.
1979. Workshop Proceedings. Management of north-central and north-eastern forests for
nongame birds. U.S. Dep. Agric., For. Serv. GTR NC-5].

Davis, ME. 1981. Quaternary history and the stability of forest communities. Pages 132-
153 in DC. West, H.H. Shugart, and DB. Botkin (editors). Forest succession, concepts
and application. Springer-Verlag. New York.

Davis, M.B., S. Sugita, R.R. Calcote, T.E. Parshall, and J .3. Ferrari. 1995. 3000 years of
abundant hemlock in Upper Michigan. Pages 19-27 in G. Mroz and J. Martin (editors).
Hemlock ecology and management. Proceedings of a regional conference on ecology and
management of eastern hemlock. Department of Forestry, University of Wisconsin.
Madison.

DeGraaf, R.M., and K.E. Evans, compilers. 1979. Workshop Proceedings. Management
of north-central and north-eastern forests for nongame birds. U.S. Dep. Agric., For. Serv.
GTR NOS 1.

DeGraaf, R.M., M. Yarnasaki, and W. B. Leak. 1992. Management of New England
northern hardwoods, spruce-fir, and eastern white pine for neotropical migratory birds.
Pp. 363-373, in D. Finch and P.W.A. Stangel (editors). Status and management of
neotropical migratory birds. USDA-Forest Service General Technical Report RM-229.

247

Delcourt, H.R., P.A. Delcourt, and T. Webb III. 1983. Dynamic plant ecology: the
spectrum of vegetational change in space and time. Quaternary Science Revue. 1: 153-
175.

Donovan, T.M., F.R. Thompson, J. Faaborg and J. Probst. 1995. Reproductive success of
migratory birds in habitat sources and sinks. Conservation Biology 9 (6): 1380-1395.

Doepker, R.V., D.E. Byer Jr., and M. Donovan. 1995. Deer population trends in
Michigan’s Upper Peninsula. Pages 99-108 in G. Mroz, and J. Martin (editors). Hemlock
ecology and management. Proceedings of a regional conference on ecology and
management of eastern hemlock. University of “Wisconsin-Madison.

Dunning, J .B., B]. Danielson, and HR. Pulliarn. 1992. Ecological processes that affect
populations in complex landscapes. Oikos 65: 169-175.

Erlich, PR, and AH. Erlich. 1981. Extinction: the causes and consequences of the
disappearance of species. Random House. New York. 305p.

Erlich, PR. and E0. Wilson. 1991. Biodiversity studies: science and policy. Science 253:
758-762.

Farrand, W.R., and D.L. Bell. 1982. Quaternary geology of northern Michigan. Map.
Michigan Department of Natural Resources, Geological Survey Division.

Faaborg J ., M. Brittingham, T. Donovan, and J. Blake. 1995. Habitat fragmentation in the
temperate zone. Pages 357-380 in D. M. Finch, and P.W.A. Stangel, editors. Status and
management of neotropical migratory birds. USDA Forest Service Gen.Tech. Rep. RM-
229.

Finch, D.M. 1991. Population ecology, habitat requirements, and conservation of
neotropical migratory birds. USDA Forest Serv. Gen.Tech. Rep. RM-205, Fort Collins,
CO.

Fahrig, L. 1997. Relative effects of habitat loss and fragmentation on population
extinction. J. Wildlife Management 61(3): 603-610.

Fonnan, R.T.T., and M. Godron. 1986. Landscape ecology. Wiley and Sons. New
York.

Franklin, J .F . 1993. Preserving biodiversity: species, ecosystems, or landscapes.
Ecological Applications 3 (2): 202-205.

Franzreb, K.E. 1977. Bird population changes after timber harvesting of a mixed conifer
forest in Arizona. USDA Forest Service. Research Paper RM-184. 26pp.

248

 

F reemark, K.E., and HG. Merriam. 1986. Importance of area and habitat heterogeneity
to bird assemblages in temperate forest fragments. Biological Conservation 36: 115-141.

Freemark, K.E., J .R. Probst, J .B. Dunning, and SJ. Hejl. 1992. Adding a landscape
perspective to conservation and management planning. In D.M. Finch and P.W.A. Stangel
editors: Status and management of neotropical migratory birds. USDA Forest Service,
General Technical Report- RM 229.

Freemark, K.E., J .R. Probst, J .B. Dunning, and SJ. Hejl. 1995. A landscape ecology
perspective for research, conservation, and management. In TE. Martin and D.M. Finch
(editors). Ecology and management of neotropical migratory birds. Oxford University
Press, New York.

F relich, LE, and C .G. Lorimer. 1991. Natural disturbance regimes in hemlock-hardwood
forests of the Upper Great Lakes region. Ecological Monographs 61 (2): 145-164.

Gates, J .E., and L.W. Gysel. 1978. Avian nest dispersion and fledgling success in field-
forest ecotones. Ecology 59: 871-883.

Gentry, AH. 1996. Species expirations and current extinction rates: a review of the
evidence. Pages 17-26 in RC. Szaro, and D.W. Johnson (editors). Biodiversity in
managed landscapes. Oxford University Press. New York. 778pp.

Godfrey, W.E. 1986. The birds of Canada. Natural Museum of Natural Sciences. National
Museums of Canada

Godron, M., and R.T.T. F orrnan. 1983. Landscape modification and changing ecological
characteristics. pages 12-28 in Mooney HA, and M. Godron (editors). Disturbance and
ecosystems components of response. Ecological Studies 44. Springer Verlag. New York.

Golley, F .B. 1989. A proper scale; editor’s comment. Landscape Ecology 2 (2): 71-72.

Graham, SA. 1941. Climax forests of the Upper Peninsula of Michigan. Ecology
22(4):355-362.

Grumbine, RE. 1990. Protecting biological diversity through the greater ecosystem
concept. Natural Areas J. 10:114-120.

Grumbine, RE. 1994. What is ecosystem management? Conserv. Biol. 8:27-38.

Gutzwiller, K.J., and SH. Anderson. 1987. Multiscale associations between cavity-
nesting birds and features of Wyoming strearnside woodlands. The Condor 89:534-548.

Haila, Y. 1986. North European landbirds in forest fragments: evidence for area effects.
Pp 315-319 in J. Vemer, M.L. Morrison and 0]. Ralph (editors). Wildlife 2000;

249

 

».~' 4"\2‘ _ 5

modeling habitat relationships of terrestrial vertebrates. The University of Wisconsin
Press. Madison, Wisconsin.

Hamady, M. 1983. Bird communities along a forest maturity gradient of northern
hardwoods in Upper Michigan. Unpublished M.S. thesis. Colorado State University.
144pp.

Hansen, Al, and D.L. Urban. 1992. Avian response to landscape pattern: The role of
species' life histories. Landscape Ecology 7(3): 163-180.

Harris, L. D., and G. Silva-Lopez. 1992. Forest fragmentation and the conservation of
biological diversity. Pages 197-237 in P.L. Fiedler, and SK. Jain (editors). Conservation
biology the theory and practice of nature conservation and management. Chapman and
Hall, New York. 507pp.

Harrison, H.H. 1975. A field guide to bird’s nests in the United States east of the
Mississippi River. Houghton Mifflin Company. Boston. 257pp.

Harper, J. 1992. Foreword. Pages xii to xviii in P.L. F iedler and SK. Jain (editors).
Conservation biology the theory and practice of nature conservation preservation and
management. Chapman and Hall. New York. 507pp.

Hart, J .H. 1990. Nothing is permanent except change. Pages 2-17 in J .M. Sweeney
(editor). Management of dynamic ecosystems. Proc. Slst Midwest Fish and Wildlife
Conference. Springfield, Illinois. 180pp.

Haufler, J .B. 1990. Static management of forest ecosystems. Pages 124-130 in J .M.
Sweeney (editor). Management of dynamic ecosystems. Proc. Slst Midwest Fish and
Wildlife Conference. Springfield, Illinois. 180pp.

Heinselman, ML. 1973. Fire in the virgin forests of the Boundary Water Canoe Area,
Minnesota. Quaternary Research 3: 329-382.

Hilden, O. 1965. Habitat selection in birds. Ann. Zool. Fennica 2:53-75.

Hill NP, and J .M. Hagan. 1991. Population trends of some northeastern north American
landbirds: a half century of data. Wilson Bulletin 103 (2): 165-338.

Holmes, RT. and S. Robinson. 1981. Tree species preferences of foraging insectivorous
birds in a northern hardwoods forest. Oecologia 48:31-35.

Holmes, RT, and T.W. Sheny. 1985. Assessing population trends of New Hampshire
forest birds: local vs regional patterns. Auk 105: 786-768.

Hutto, R.L., S. Reel, and PB. Landers. 1987. A critical evaluation of the species
approach to biological conservation. End. Sp. Update 4 (12): 1-4. 1

250

 

 

 

Irland, LC. 1982. Wildlands and woodlots: the story of New England’s forests.
University Press of New England. Hanover, New Hampshire. 217 pp.

James, EC. 1971. Ordination of habitat relationships among breeding birds. Wilson
Bulletin 83: 215-236.

James PC, and CE. McCulloch. 1995. The strength of inferences about causes of trends
in populations. Pages 40-51 in TE. Martin, and DB. Finch (editors). Ecology and
management of neotropical migratory birds: a synthesis and review of critical issues.
Oxford University Press. New York.

Karr, J .R. 1982. Avian extinction on Barro Colorado Island, Panama: a reassessment.
Amer. Natur. 119:220-239.

Kepler, C .B., and J .M. Scott. 1981. Reducing bird count variability by training observers.
Pages 366-371 in C]. Ralph and J .M. Scott (editors). Estimating numbers of terrestrial
birds; Studies in Avian Biology No 6. Allen Press. Lawrence. Kansas. 630pp.

Koford, R.R., J. B. Dunning JR., C .A. Ribic, and D.M. Finch. 1994. A glossary for avian
conservation biology. Wilson Bull. 106(1): 121-137.

Kotliar, N.B., and J .A. Wiens. 1990. Multiple scales of patchiness and patch structure: a
hierarchical framework for the study of heterogeneity. Oikos 59:253-260.

Krummel, J.R., R.H. Gerdner, G. Sugihara, R.V. O'Neill, and PR. Coleman. 1987.
Landscape patterns in a disturbed environment. Oikos 48: 321-324.

Laake, J. L., S.T. Buckland, D.R. Anderson, and KP. Bumham. 1994. DISTANCE
User’s Guide. Version 2.1. Colorado Cooperative Fish and Wildlife Research Unit
Colorado State University, Fort Collins, CO.84 pp.

Large, ET 1993. Implementation of an ecosystem approach to endangered species
conservation. U.S. Endangered species Tech Rep. Bull. 10: 3-6.

Lord, J .M., and DA. Norton. 1990. Scale and the spatial concept of fragmentation.
Conserv. Biol. 4: 197-202.

Lorimer, CG. 1977. The presettlement forest and natural disturbance cycle of
northeastern Maine. Ecology 58: 139-148.

Lavage, TE. 1979. Foreword. Pages ix-x in ME. Soulé and B. A. Wilcox (editors).
Conservation biology-an evolutionary-ecological perspective. Sinauer Associates, Inc.
Sunderland, Massachusetts. 395pp.

Lynch, J .F ., and RF. Whitcomb. 1978. Effects of the insularization of the eastern
deciduous forest on avifaunal diversity and turnover. Pages 461-490 in A. Marmelstein

251

 

(editor). Classification, inventory, and analysis of fish and wildlife habitat. FWS/OBS-78-
76. US. Department of the Interior, Fish and Wildlife Service, Office of Biological
Services, Washington DC.

Lynch J .F ., and DP. Whigham. 1984. Effect of forest fi'agmentation on breeding bird
communities in Maryland, USA. Biological Conservation 28: 287-324.

MacArthur, RH. and E0. Wilson. 1967. The theory of island biogeography. Princeton
Univ. Press. Princeton NJ.

MacArthur, RH. 1958. Population ecology of some warblers of northeastern coniferous
forests. Ecology 39: 599-619.

Marcot, B.G., and DD. Murphy. 1996. Population viability analysis and management.
Pages 58-76 in R. Szaro, editor. Biodiversity in managed landscapes: Theory and
Practice. Proceedings of a conference 13-17 July. Sacramento, CA. Oxford University
Press. New York 778pp.

Martin TE, and D.M. Finch. 1995. Introduction: importance of knowledge and its
application in neotropical migratory birds. Pages XIII - XV in TE. Martin and DB. Finch,
(editors). Ecology and management of neotropical migratory birds: a synthesis and review
of critical issues. Oxford University Press. New York.

McCann, MT 1991. Land, climate and vegetation of Michigan. Pages 15-31 in R.
Brewer, G.A. McPeek and R.J. Adams (editors). The atlas of breeding birds of Michigan.
Michigan State University Press. East Lansing. 594pp.

Meentemeyer V. and ED. Box. 1987. Scale effects in landscape studies. Pages 15-34 in
MG. Turner (editor). Landscape heterogeneity and disturbance. Ecological Studies 64.
Springer-Verlag. New York. 239pp.

Meffe, GK. 1996. Conserving genetic diversity in natural systems. Pages 41-57 in RC.
Szaro and D. W. Johnston (editors). Biodiversity in managed landscapes theory and
practice. Oxford University Press. New York. 778pp

Milne, ET. 1991. Heterogeneity as a multiscale characteristic of landscapes. Pages 69-84
in J. Kolasa, and S.T.A. Pickett (editors). Ecological Heterogeneity. Springer-Verlag.
New York.

Miller, S.G., S.P. Bratton, and J. Hadidian. 1992. Impacts of white-tailed deer on
endangered and threatened vascular plants. Natural Areas Journal 12 (2): 67-74.

Mladenofl‘, DJ, and F. Steams. 1993. Eastern hemlock regeneration and deer browsing

in the northern Great Lakes Region: A re-examination and model simulation.
Conservation Biology 7: 889-900.

252

 

Nfladenoff, D.J., M.A. White, J. Pastor, and T. R. Crow. 1993. Comparing spatial
pattern in unaltered old-growth and disturbed forest landscapes. Ecological Applications
3 (2): 294-306.

Mladenofi’, D. J. 1995. The role of eastern hemlock across scales in the northern lake
states. Pages 29-42 in G. Mroz, and J. Martin (editors). Hemlock ecology and
management. Proceedings of a regional conference on ecology and management of
eastern hemlock held in Iron Mountain, Michigan Sept 27-28, 1995. Published by the
Department of Forestry, University of Wisconsin, Madison.

Mooney, HA. and M. Godron (editors). 1983. Disturbance and ecosystems; components
of response. Ecological Studies 44. Springer Verlag. New York. 292pp.

Morrison, M.L., R.W. Mannan, and G.L. Dorsey. 1981. Effects of number of circular
plots on estimates of avian density and species richness. Pages 405-408 in C.J. Ralph, and
J .M. Scott, editors. Estimating numbers of terrestrial birds. Studies in Avian Biology No
6. Allen Press. Lawrence, Kansas. 630pp.

Morse, DH. 1976. Variables affecting the density and territory size of breeding spruce-
woods warblers. Ecology 57: 290-301.

National Fish and Wildlife Foundation 1993. 1993 Annual Report. NFWF 4 (1). 87pp.

National Fish and Wildlife Foundation 1994. Neotropical migratory bird conservation
initiative. Washington DC.

Naveh, Z. and A. Lieberman. 1994. Landscape ecology, theory and application. Second
edition. Springer Verlag. New York

Nichols, GE. 1935. The hemlock-white pine-northern hardwood region of North
America. Ecology 16(3): 403-422.

Niemi, G.J., and J .R. Probst. 1990. Wildlife and fire in the Upper Midwest. Pages 32-46
in J.M. Sweeney (editor). Management of dynamic ecosystems. Proceedings of a
symposium held at the 5 1 st Midwest Fish and Wildlife Conference, Springfield, Illinois.
180pp.

Niles, L]. 1992. Protection of neotropical migrants as major focus of wildlife
management. Pages 392-395 in D. M. Finch and P. W. Stangel (editors). Status and
management of neotropical migratory birds. USDA- Forest Service, General Technical
Report RM-229.

Noon, B.R., V. P. Bingman, and J .P. Noon. 1979. The effects of changes in habitat on
northern hardwood forest bird communities. Pages 33-48 in R.M. DeGraaf, and K. E.

Evans (editors). Management of North Central and Northeastern forests for nongame
birds. USDA Forest Service. GTR NC-51. Minneapolis, MN. 268pp.

253

 

Noon, B.R., D.K. Dawson, D.B. Inkley, C.S. Robbins, and SH. Anderson. 1980 .
Consistency in habitat preference of forest bird species. Transactions of the North
American Wildlife Conference 45: 226-244.

Noss, RF. 1983. A regional landscape approach to maintain diversity. BioScience 33:
700-706.

Noss, RF. 1987. From plant communities to landscapes in conservation inventories: a
look at The Nature Conservancy (USA). Biol. Conserv. 41:1 1-3 7.

Noss, RF. 1992. Issues of scale in conservation biology. Pages 240-250 in P.L. Fiedler,
and SK. Jain (editors) . Conservation biology the theory and practice of nature
conservation and management. Chapman and Hall. New York.

 

Ohmann, L.F. 1979. Northeastern and North Central forest types and their management.
Pages 22-31 in R.M. DeGraaf and K. E. Evans (editors). Management of North Central ‘1:
and Northeastern forests for nongame birds. USDA Forest Service. GTR NC-51.

Minneapolis, MN. 268pp.

Opdarn P. 1991. Metapopulation theory and habitat fragmentation: a review of holarctic
breeding bird studies. Landscape Ecology 5 (2): 93-106.

Pastor, J ., and M. Brochart. 1990. The spatial pattern of a northern hardwood -conifer
landscape. Landscape Ecology 4: 55-68.

Pastor, J. and DJ. Mladenoff. 1992. The southern boreal-northern hardwood forest. Pages
216-240 in Shugart H.H., R. Leemans, and G. B. Bonan (editors). A systems analysis of
the global boreal forest. Cambridge University Press. 565p.

Pearson, TO. 1936. Birds of America. Doubleday and Company, Inc.Garden City, New
York. 289pp.

Pearson, S.M., M.G. Turner, R.H. Gardner, and RV. O'Neill. 1996. An organism-based
perspective of habitat fragmentation. Pages 76- 93 in RC. Szaro and D.W. Johnson
(editors). Biodiversity in managed landscapes. Oxford University Press. New York.
778pp.

Peterjohn, B.C., J. Sauer, and CS. Robbins. 1995. Population trends from the North
American Bird Survey. Pages 3-39 in TE. Martin and DB. Finch (editors). Ecology and
management of neotropical migratory birds: a synthesis and review of critical issues.
Oxford University Press. New York.

Peterson, RT. 1980. A field guide to the birds of Eastern and Central North America.
Houghton Mifllin Company. Boston.

254

 

Petit, L.J., D.R. Petit, and TE. Martin. 1995. Landscape-level management of migratory
birds: looking past the trees to see the forest. Wildlife Society Bulletin 23 (4): 420-429.

Pickett, ST. A., and P. S. White. 1985. Patch dynamics: a synthesis, pages 371-384 in
Pickett, ST. A., and P. S. White (editors). The ecology of natural disturbance and patch
dynamics. Academic Press, New York.

Pickett, ST. A. , V. T. Parker and PL. Fiedler. 1992. The new paradigm in ecology:

implications for conservation biology above the species level. Pages 66-88 in P.L. Fiedler
and SK. Jain (editors). Conservation biology the theory and practice of nature ‘
conservation and management. Chapman and Hall. New York. r

Pielou, EC. 1991. After the ice age, the return of life to glaciated North America. The r.
University of Chicago Press. Chicago. 366pp. :15

 

Pregitzer, KS. 1981. Relationships among physiography, soils and vegetation in the “h
McComtick Experimental Forest, Upper Michigan. Ph.D. Dissertation, University of '1
Michigan, Ann Arbor, MI.

Pregitzer, KS, and B.V. Barnes. 1982. The use of ground flora to indicate edaphic
factors in upland ecosystems of the McCormick Experiment Forest, Upper Michigan. Can.
J. For. Res. 12 (3): 661-672.

Pregitzer, KS, and B.V. Barnes. 1984. Classification and comparison of upland
hardwood and conifer ecosystems of the Cyrus H. McCormick Experimental Forest,
Upper Michigan. Canadian Journal of Forest Research 14 (3): 362-375.

Probst, J .R., and TR. Crow. 1991. Integrating biological diversity and resource
management. Journal of forestry 89: 12-17.

Probst, J .R., D.S. Rakstad, and D. J. Rug. 1992. Breeding bird communities in
regenerating and mature broadleaf forests in the USA Lake States. Forest ecology and
management 49: 43-60.

Probst, J .R., and F .R. Thompson. 1994. A comprehensive, regional strategy for
integrating wildlife and timber management into ecosystem management. 56th Midwest
Fish and Wildlife Conference. Indianapolis, IN.

Pulliam, HR. 1988. Sources, sinks, and population regulation. Am. Nat. 132(5): 652-
661.

Ralph, C. J ., and J. M. Scott (editors). 1981. Estimating numbers of terrestrial birds.

Studies in Avian Biology No. 6. Cooper Ornithological Society. Allen Press. Lawrence,
Kansas. 630pp.

255

Ralph, C.J., G.R. Geupel, and P. Pyle. 1991. Handbook for population monitoring of
landbirds (Draft). USDA. Forest Service, Redwood Sciences laboratory. Arcata, CA.

Reiners, WA. 1983. Disturbance and basic properties of ecosystem energetics. Pages 83-
98 in HA. Mooney and M.Godron (editors). Disturbance and ecosystems, components of
response. Springer-Verlag, New York. 292 pp.

Reinoehl, J. 1991. American redstart (Setophaga uticilla) pp 428 in R. Brewer, and GA.
McPeak, and R.J. Adams: the atlas of breeding birds of Michigan. Michigan State
University Press. East Lansing, Michigan.

Reynolds, R.T., J .M. Scott, and RA. Nussbaum. 1980. A variable- circular plot method ,
for estimating bird numbers. Condor 82: 309-313. ;:

Riddle, BR, and KB. Jones. 1996. Relating phylogeographic population structure to
issues in conservation biology. Pages 108-109 in RC. Szaro, and D.W. Johnston in,
(editors). Biodiversity in managed landscapes; theory and practice. Oxford University

Press. New York.

 

Risser, PG. 1985. Toward a holistic management perspective. BioSci. 35: 414-418.

Robbins, CS. 1979. Effect of forest fi'agmentation on bird populations. In Degraaf and

Evans, editors: Management of northcentral and northeastern forests for nongame species.
USDA. Forest Service. GTR NC-51.

Robbins, C .S., D. Bystrak, and PH. Geissler. 1986. The Breeding Bird Survey: its first
fifteen years, 1965-1979. US. Fish Wildl. Serv., Resour. Pub.. 157. 196 pp.

Robbins, 05., J .R. Sauer, R. Greenberg, and S. Droege. 1989a. Population declines in
North American birds that migrate to the Neotropics. Proc. Nat. Acad. of Sci. 86: 7658-
7662.

Robbins, C .S., Dawson D.K., and BA. Dowell. l989b. Habitat-area requirements of
breeding forest birds of the Middle Atlantic States. Wildlife Monograph 103. Washington
DC. The Wildlife Society.

Robinson, SK. 1988. Reappraisal of the costs and benefits of habitat heterogeneity for
nongame wildlife. Trans. 53rd N.A. Wildl. and Nat. Res. Conf.

Robinson, SK. 1992. Population dynamics of breeding Neotropical migrants in a
fragmented Illinois landscape. Pages 408-418 in J.M. Hagan III and D.W. Johnston
(editors). Ecology and conservation of Neotropical migrant landbirds. Smithsonian
Institution Press, Washington, DC.

Rosenberg, K.V., and MG. Raphael. 1986. Effects of forest fragmentation on vertebrates
in Douglas-fir forests. Pages 263-272 in J .Vemer, M.L. Morrison and C]. Ralph.

256

 

(editors). Wildlife 2000, modeling habitat relationships of terrestrial vertebrates. The
University of Wisconsin Press. Madison, Wisconsin.

Roth, RR. 1976. Spatial heterogeneity and bird species diversity. Ecology 57: 773-782.

Rotenberry, J. T., and J .A. Wiens. 1980. Habitat structure, patchiness, and avian
communities in North American steppe vegetation: a multivariate analysis. Ecology
61:1228-1250.

Runkle, J .R. 1982. Patterns of disturbance in some old-growth forests of eastern North
America. Ecology 63(5): 1533-1546.

Ruggiero, L.F., G.D. Hayward, and J. R. Squires. 1994. Viability analysis in biological
evaluations: concepts of population viability analysis, biological population, and ecological
scale. Conservation Biology 8 (2): 364-372.

Samson, EB. 1992. Conserving biological diversity in sustainable ecological systems.
Trans. 57th N.A. Wildl. & Nat. Res. Conf. (308-320).

Sauer, J .R., and S. Droege. 1990. Survey designs and statistical methods for the
estimation of avian population trends. US. Fish and Wildlife Service Biological Report 90
(1). Patuxent Wildlife Research Center and Office of Migratory Bird Management. Laurel,
Maryland. 166pp.

Schoener, T.W. 1966. Sizes of feeding territories among birds. Ecology 49 (1): 123-141.

Scott, J .M., F. Davis, B. Csuti, R. Noss, B. Butterfield, C. Groves, H. Anderson, S.
Caicco, F. D'erchia, T.C. Edwards, Jr., J. Ulliman, and R.G. Wright. 1993. Gap analysis:

a geographic approach to protection of biological diversity. Wildlife Monographs. No 123.

41pp.

Senner, SE. 1988. Saving birds while they are still common. U.S. Endangered species
Tech Rep. Bull. 5(6): 1-4.

Small, M.F., and ML. Hunter. 1988. Forest fragmentation and avian nest predation in
forested landscapes. Oecologia 76: 62-64.

Smith, KG. 1977. Distribution of summer birds along a forest moisture gradient in an
Ozark watershed. Ecology 58; 810-819.

Simpson, T.B., P.E. Stuart, and B.V. Barnes. 1989. Cover types at the Huron Mountain

Club, Marquette County, Michigan. Huron Mountain Wildlife Foundation, Big Bay,
Michigan.

257

 

Simpson, T.B., P.E. Stuart, and B.V. Barnes. 1990. Landscape ecosystems and cover
types of the reserve area and adjacent lands of the Huron Mountain Club. Occasional
Papers of the Huron Mountain Wildlife Foundation No 4.

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

Soulé. ME. 1991. Conservation: tactics for a constant crisis. Science 253: 744-50.

Stanley, T. R. Jr. 1994. Ecosystem management and the arrogance of humanism.
Conservation biology 9(2): 255-262.

Steams, F. 1990. Forest history and management in the northern Midwest. Pages 109-
122 in J .M. Sweeney. Management of dynamic ecosystems. Proc. 51st Midwest Fish and
Wildlife Conference. Springfield, Illinois. 180pp.

Steele, B]. 1992. Habitat selection by breeding black-throated blue warblers at two spatial
scales. Omis Scandinavica 23:33-42.

Swanson, F .J ., T.K. Kratz, N. Cain, and RG. Woodman see. 1988. Landform efl‘ects on
ecosystem patterns and processes. Bioscience 38 (2): 92-98.

Tansley, AG. 1935. The use and abuse of vegetational concepts and terms. Ecology, 16,
284-307.

Temple, S.A., M.J. Mossman, and B. Ambuel. 1979. The ecology and management of
avian communities in mixed hardwood-coniferous forest. Pages 132-151 in R.M. DeGraaf
and K. E. Evans (editors). Management of North Central and Northeastern forests for
nongame birds. USDA Forest Service. GTR NC-51. Minneapolis, MN. 268pp.

Terborgh, J.W. 1992. Why American songbirds have vanished. Scientific American. May
1992.

Terborgh, J.W. 1989. Where have all the birds gone ? Princeton University Press.
Princeton, NJ.

Tubbs, CH. 1977. Manager‘s handbook for northern hardwoods in the north-central
States. USDA- Forest Service. GTR NC-39. St. Paul. MN.

Tubbs, CH, and RM. Goodman. 1983. North Central types. Pages 55-66 in USDA-
Forest Service: Silvicultural systems for the major forest types of the United States.
USDA, Agricultural Handbook N° 445.

Turner, MG. 1989. Landscape ecology: the effect of pattern on process. Annu.
Rev.Ecol. Syst. 20: 171-197.

258

 

 

Turner, M.G. (editor). 1987. Landscape heterogeneity and disturbance. Ecological studies
64. Springer-Verlag. New York. 239pp.

Tyrrell, LE. and TR. Crow. 1993. Analysis of structural characteristics of old-growth
hemlock-hardwood forests along a temporal gradient. Pages 23 7-246 in J .F . F arlich, R.P.
McIntosh, O.L. Loucks (editors). J .T. Curtis- 50 years of Wisconsin plants. Wisconsin
Academy of Sciences, Arts and Letters. 339pp.

US. Fish and Wildlife Service. 1994. Defining our firture: an ecosystem approach to fish
and wildlife conservation. Fish and Wildlife News. April, 2-3.

 

USDA. Forest Service. 1993. National Hierarchical Framework of Ecological Units.
Ecomap, USDA Forest Service, Washington, DC. 20pp.

USDA. Forest Service. 1986. Hiawatha National Forest Plan. Escanaba, MI.

‘ IL

Webb, W.L., R.T. King, and BF. Patrick. 1956. Effect of white-tailed deer on a mature
northern hardwood forest. Journal of forestry 54: 391-398.

Whitcomb R.F., C.S. Robbins, J.F. Lynch, B.L. Whitcomb, M.K. Klimkiewicz, and D.
Bystrak. 1981. Effects of forest fragmentation on avifauna of the eastern deciduous
forest. Pages 125-205 in R.L. Burgess, and D.M. Sharpe (editors). Forest island dynamics
in man-dominated landscapes. Springer-Verlag. New York.

Whitcomb, B.L., R.F. Whitcomb, and D. Bystrak. 1977. Island biogeography and habitat
islands of eastern forest. III. Long-term turnover and effects of selective logging on the
avifauna of forest fragments. American Birds 31: 17-23.

White, MA, and DJ. Mladenoff. 1994. Old-grth forest landscape transitions from
pre-European settlement to present. Landscape Ecology 9(3): 191-205.

Whitney, G.G., 1987. An ecological history of the Great Lakes forest of Michigan. J.
Ecol. 75:667-684.

Whitney, G.G., and W.J., Somerlot. 1985. A case study of woodland continuity and
change in the American Midwest. Ecol. Conserv. 31:265-287.

Whitmore, RC. 1977. Habitat partitioning in a community of passerine birds. Wilson Bull.
89 (2):253-265.

Wiens, J .A. 1989. Spatial scaling in ecology. Functional Ecology 3: 385-397.
Wiens, J.A., J.F. Addicott, T.J. Case, and J. Diamond. 1986a. Overview: the importance

of spatial and temporal scale in ecological investigations. Pages 145-153 in J. Diamond
and T.J. Case (Editors). Community ecology. Harper and Row. New York.

259

 

Wiens, J .A., J.T. Rotenberry, and B. Van Horne. 1986b. Habitat occupancy patterns of
North American shrubsteppe birds: the effects of spatial scale-Oikos 48:132-147.

Wiens, J .A. 1981. Scale problems in avian censusing. Pages 513-521 in C.J. Ralph, and J.
Michael Scott (editors): Estimating numbers of terrestrial birds. Studies in Avian Biology
No. 6. Cooper Ornithological Society

Wiens, J .A., and J .T. Rotenberry. 1981. Avian habitat occupancy. Pages 522-532 in CI.
Ralph, and J. Michael Scott (editors): Estimating numbers of terrestrial birds. Studies in
Avian Biology No. 6. Allen Press. Lawrence, Kansas.

Wiens, J.A. 1976. Population responses to patchy environments. Ann. Rev. Ecol. Syst. 7:
81-120.

 

Wienstein, DA, and H.H. Shugart.1983. Ecological modeling of landscape dynamics.
Pages 29-45 in HA. Mooney, and M. Godron (editors): disturbance and ecosystems.
Ecological Studies 44. Springer- Verlag. New York. 292pp.

‘4‘.

Wilcove, D.S., and RF. Whitcomb. 1983. Gone with the trees. Natural History 92: 82-
91.

Wilcove, D. 1985. Nest predation in forest tracts and the decline of migratory songbirds.
Ecology 66: 1211-1214.

Wilcove, D. 1987. From fragmentation to extinction. Natural Areas Journal 7 (1): 29
Wilcove, D.S. 1988. Forest fragmentation as a wildlife issue in the eastern United States.
Pages 1-5 in RM. DeGraaf and W.M. Healy (editors). Is Fragmentation a management

issue in the Northeast?. USDA-Forest Service, General Technical Report NE-140.

Wilcox, BA. and DD. Murphy. 1985. Conservation strategy: the effects of
fragmentation on extinction. American Naturalist 125: 879-887.

Wilson, E0. 1988. The current state of biological diversity. Pages 3-18 in E0. Wilson
(editor). Biodiversity. National Academy Press. Washington DC.

Wilson, E0. 1992. The diversity of life. Belknap Press. Harvard University. Cambridge,
MA. 429pp.

Van Deelen, T. R. 1995. Seasonal migrations and mortality of white-tailed deer in
Michigan’s Upper Peninsula. Unpublished Ph.D. dissertation. Michigan State University.
East Lansing.

Van Horn, B. 1983. Density as a misleading indicator of habitat quality. J. Wildl. Manag.
47: 893-901.

260

Yahner, RH, and DP. Scott. 1988. Effects of forest fragmentation on depredation of
artificial nests. Journal of Wildlife Management 52: 158-161.

 

261