in.

.I Ramada.
59 "EN.

#5?me . .Ix
W0 ,asgﬂmﬂ‘Awuu.
x, 3.
a 521.31...
..S . \..L.
(It!)
—‘
.

i V. :5}? kg...
5’... I...-.\ ,
‘whn... i... :9
1‘5!!! Ar.“
v Iii

 

 

{anti

4‘9?!
:ltzv: !.

2.5.3....
.5. A.

 

h ..
:2.
2.3

 

 

Z
fill? LIBRARY
Michioa. State
University

 

 

This is to certify that the
thesis entitled

THE BEECH SCALE (CRYPTOCOCCUS FAG/SUGA) IN
MICHIGAN: DISTRIBUTION, MODELS OF SPREAD AND
RELATION TO FOREST AND WILDLIFE RESOURCES

presented by

NANCY J. SCHWALM

has been accepted towards fulﬁllment
of the requirements for the

Master of degree in
Science FISHERIES AND WILDLIFE

 

 

WM

V Major Professor’s Signature
AUyﬁ/j f wﬂﬁ
7 I

Date

 

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

 

 

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K IProj/AccaPrasICIRCIDateDue indd

 

THE BEECH SCALE (CRYPTOCOCCUS FAGISUGA) TN MICHIGAN:
DISTRIBUTION, MODELS OF SPREAD AND RELATION TO FOREST AND
WILDLIFE RESOURCES.

By

Nancy J. Schwalm

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF SCIENCE
FISHERIES AND WILDLIFE

2009

ABSTRACT

THE BEECH SCALE (CRYPTOCOCC US FAGISUGA) IN MICHIGAN:
DISTRIBUTION, MODELS OF SPREAD AND RELEATION TO FOREST AND
WILDLIFE RESOURCES.

By
Nancy J. Schwalm

The state of Michigan was surveyed from 2004-2006 to locate beech scale infestations
and to collect baseline data of forest and wildlife resources of those study sites. Results
of this survey demonstrated that beech scale was more widely distributed than
previously thought. Beech scale was distributed in the Upper Peninsula in a single
contiguous population, and encompassed nearly all of the distribution of American
beech. In the Lower Peninsula, beech scale was distributed across several disjoint
populations, and was found on several islands within the Great Lakes. The spread of
beech scale was represented using an inverse modeling procedure. Results of these
models showed that spread rates in the Upper Peninsula were higher than in the Lower
Peninsula, and that spread rates depended on land cover types. Spread rates were
modeled as a diffusion-like process, and were substantially lower than previous
estimates based on large-scale jumps in distribution. To date, infestation with beech
scale has shown little evidence of impact on forest wildlife resources, but as Nectria
fungal infestations spread leading to beech bark disease, more widespread impacts are

expected.

ACKNOWLEDGEMENTS

This research would not have been possible without the generous ﬁnancial support by
the Michigan Department of Natural Resources and the United States Forest Service

PTIPS Program, thank you.

A great deal of people assisted me along the way; I would like to recognize a few of
them here. Thank you to Dr. Daniel B. Hayes, my major professor for all that he taught
me and the incredible support and patience that he has had with me from learning
modeling to ﬁnalizing drafts, thank you so very much, you truly are a mentor. Dr.
Deborah G. McCullough has functioned like a co-advisor because of her incredible
knowledge of the insect and forestry world, thank you for all of your time and patience.
Thank you to Dr. Rique Campa and Dr. Michael B. Walters, my graduate committee,
for their time reviewing drafts, listening to ideas and providing guidance and assistance
when needed. Thank you to Amos Desjardins, who collected data for this project the
summer before I started at Michigan State University. For my right-hand man, Daniel
Wieferich who worked in the ﬁeld, entered and checked data and did all of the amazing
GIS functions for my research, thank you for everything. I was blessed with the best
ﬁeld crew that anyone could ask for, thank you to James Wieferich and Will Folland for
your positive attitudes, willingness to learn and work (and play) hard in the ﬁeld.
Finally, I would like to acknowledge the “nonprofessional” help, my husband Ezra, my
Mom and Dad and my dear friends, Malcolm and Henrietta Olson, who all were so
understanding and supportive during the entire process from application to submitting

this thesis, thank you.

iii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... v
LIST OF FIGURES ................................................................................. vii
CHAPTER 1

DISTRIBTUION OF BEECH SCALE INSECTS IN MICHIGAN:
ASSOCIATION WITH FOREST AND WILDLIFE RESOURCES

Introduction ................................................................................... 1
Invasive Species .............................................................................. 3
American beech .............................................................................. 5
Beech bark disease ........................................................................... 7
Advancing front: beech scale ............................................................... 8
Beech scale biology ......................................................................... 9
Killing front: Nectria fungi ............................................................... I4
Nectria taxonomy ........................................................................... 16
Aftermath forest ............................................................................ 19
Wildlife ...................................................................................... 21
Rate of Spread .............................................................................. 23
CHAPTER 2 ......................................................................................... 26

MODELING THE SPATIAL SPREAD OF THE BEECH SCALE
INSECT (CRYPTOCOCCUS FAGISUGA) IN MICHIGAN.

Abstract ....................................................................................... 26
Introduction .................................................................................. 28
American beech and its importance to wildlife ......................................... 30
Beech Scale ................................................................................... 31
Methods
Study Design: Site Selection ..................................................... 32
Study Design: Plot-level measurements ........................................ 33
Study Design: Individual tree measurements .................................. 34
Statistical and spatial analysis methods .................................................. 36
Results ........................................................................................ 38
Distribution of beech scale ....................................................... 39
Forest Resources ................................................................ 4O
Wildlife Resources ................................................................ 42
Discussion
Distribution of beech scale ........................................................ 44
Forest Resources .................................................................. 45
Wildlife Resources ................................................................. 48
Management Implications ................................................................. 54

iv

CHAPTER 3
MODELING THE SPATIAL SPREAD OF THE BEECH SCALE (CRYPTOCOCCUS

FAGISUGA) TN MICHIGAN.
Abstract .................................................................................... 75
Introduction ............................................................................... 77
Methods
Study Area ....................................................................... 82
Model description and structure .............................................. 84
Model parameters ............................................................... 87
Model selection procedures .................................................... 90
Model assumptions and limitations .......................................... 92
Results
Distribution of beech scale infestations ...................................... 93
Model performance: Simple diffusion model ............................... 95
Model performance: Complex model ........................................ 96
Discussion
Distribution of beech scale infestations ..................................... 98
Model performance ........................................................... 102
Management implications ............................................................ 105
APPENDICES
Appendix A Site Coordinates ......................................................... 122
Appendix B Model Parameters ........................................................ 141
Appendix C IF MAP Classiﬁcations ................................................. 142
BIBLIOGRAPHY ................................................................................ 146

LIST OF TABLES
Table 2-1. Satellite populations of beech scale infestations in Michigan ................... 57

Table 2-2. Common name and number of trees by species associated with beech within
study sites. Trees are arranged in descending order according to their abundance within
study sites ............................................................................................. 58

Table 2-3. Results from an ANOVA to compare basal area for American beech and the
seven most abundant other species across beech scale infestation classes. N is the
number of individual trees examined across sites (n=73 7). Basal area is reported in
mZ/hectare ............................................................................................ 59

Table 2—4. Number and percentage of beech trees per diameter class corresponding with

the level of beech scale infestation. Each diameter class is recorded in centimeters and is
represented in the table by the median number in its range of dbh measurements (i.e., dbh
class “5” represents trees that are 1-9 cm dbh) .................................................. 60

Table 2-5. Frequency of occurrence for beech snap, tar spots, crown dieback and cankers
across levels of beech scale infestation ........................................................... 61

Table 2-6. Mean number of beech snags (n=44) and non-beech snags (n=3,886) per site
across levels of beech scale infestation. Basal area is reported in mz/ha .................... 61

Table 2-7. Common name and number of tree species examined within study sites.
Number of cavity trees and percentage of total cavity trees arranged by species and
presented in descending order of abundance .................................................... 62

Table 2.8. Non-beech trees were divided up into 14 diameter at breast height (dbh)
classes. Each dbh-class is represented in the table by the median number in its range of
measurements (i.e., dbh-class “5” represents trees that are 1-9 cm dbh) ..................... 63

Table 2-9. Chi-square table of cavity tree abundance across levels of beech scale
infestation ............................................................................................. 64

vi

Table 2-10. Chi-square table of beech cavity tree abundance across levels of beech scale
infestation ............................................................................................ 64

Table 2-11. Volume of coarse woody debris (i1 SE) and associated level of beech scale
infestation. There were 453 sites in the absent category, 100 in the light and 71 in the
moderate categories respectively .................................................................. 65

Table 2-12. Frequency of occurrence of coarse woody debris pieces in each decay class
and corresponding beech scale infestation level ................................................ 65

Table 3-1. Satellite infestations separated into beech scale infestation class and
approximate size of the area infested as of 2006. Area was calculated using the area
feature in ArcGIS .................................................................................. 108

Table 3-2. Contingency table of model errors for each of the models ..................... 108

vii

LIST OF FIGURES

Figure 1-1. Distribution of the American beech in North America (US. Geological
Survey, 1999) ........................................................................................ 25

Figure 2-1. Adaptive sampling design for designing the advancing front. The star
represents the midpoint between a known infested site and a known uninfested site. . . ...66

Figure 2-2. Site layout with ﬁve plots; center, north, east, south, and west all 100 m
apart. Each site also has two 100 m coarse woody debris transects between the center
and north plot and the center and west plot ...................................................... 66

Figure 2-3. Photos used to standardize levels of beech scale infestation. Photo on the far
left represents beech scale classiﬁcation “trace”, middle photo represents “patchy” and
right photo deﬁnes “whitewashed” (Photos taken by Nancy Schwalm, May 2004) ....... 67

Figure 2-4. Frequency of sites plotted against mean scale to determine beech scale
infestation classes. Mean scale was determined by aggregating all plot-level data across a
site to obtain averages per site ..................................................................... 67

Figure 2-5. Map of Michigan, USA with study sites coded as uninfested (open while
circles) or infested (closed black circles) or no beech sites (triangles) ...................... 68

Figure 2-6. Map of Michigan, USA with beech study sites grouped into eleven distinct
satellite populations ...................................................................................... 69

Figure 2-7a. Map of the Ludington and Silver Lake satellite populations enlarged to
show the detail of sites coded according to their beech scale infestation level. Map
created by Daniel Wieferich on March 30, 2007 ................................................ 70

Figure 2-7b. Map of the Upper Peninsula satellite population enlarged to show the detail
of sites coded according to their beech scale infestation level. Map created by Daniel
Wieferich on March 30, 2007 ...................................................................... 71

viii

Figure 2-8. Mean beech basal area (:t 1 SE) across level of beech scale infestation ...... 72

Figure 2-9. Mean beech diameter at breast height (dbh) (:I: 1 SE) across levels of beech
scale infestation ...................................................................................... 72

Figure 2-10. Percent of beech trees infested with beech scale as a function of tree
diameter at breast height (dbh) .................................................................... 73

Figure 2-11. Frequency of beech trees within each beech scale infestation class across
diameter at breast height (dbh) classes. Diameter at breast height classes represent the
median number in a range of dbh measurements (i.e., dbh class “5” = dbh measurements
1-9 cm, “15” = 10-19 cm...”115” = 110-109 cm). Beech scale infestation classes are
coded as “HV” for heavy infestation, “MD” for moderate infestation, “LT”, for light
infestation and “AB” for uninfested ............................................................... 73

Figure 2-12. Percent of beech trees within each beech scale infestation class across
diameter at breast height (dbh) class. Beech scale infestation classes were coded as “LT”
for lightly infested, “MD” for moderately infested, and “HV” for heavily infested ....... 74

Figure 3-1. Three types of rang-versus-time curves. Range expansion patterns
commonly have an establishment phase (arrow), expansion phase (solid line), and
saturation phase (dashed line), successively. The expansion phase is classiﬁed into three
types. Type 1 shows linear expansion. Type 2 exhibits biphasic expansion, with an
initial slow slope followed by a steep linear slop. In type 3, the rate of expansion
continually increases with time (Shigesada and Kawasaki 1997) ........................... 109

Figure 32. Adaptive sampling design for designing the advancing front. The star
represents the midpoint between a known infested site and a known uninfested site. ...109

Figure 3-3. Photos used to standardize levels of beech scale infestation. Photo on the far
left represents beech scale classiﬁcation “trace”, middle photo represents “patchy” and
right photo deﬁnes “whitewashed” (Photos taken by Nancy Schwalm, May 2004)...110

Figure 3-4. Frequency of sites plotted against mean scale to determine beech scale
infestation classes. Mean scale was determined by aggregating all plot-level data across a
site to obtain averages per site ................................................................... 1 10

ix

Figure 3-5. Map of Michigan, USA with a layer of forest types grouped to locate beech.
Beech was typically found within northern hardwood or deciduous forest cover types.
Typically beech was not abundant in the oak association, coniferous or non-forested
forest types. Data was extracted from IF_MAP data 2001, from MCGI website
(www.mcgistatemius). Created by Daniel Wieferich on October 18th, 2006. Please
note that this image is presented in color ....................................................... 11 l

 

Figure 3-6. Map of Michigan, USA with study sites coded as uninfested (open while
circles) or infested (closed black circles) or no beech sites (small triangles)... ........1 12

Figure 3-7. Map of Michigan, USA with beech study sites grouped into eleven distinct
satellite populations ............................................................................... 1 13

Figure 3-8a. Modeled errors for the Upper Peninsula simple diffusion model mapped to
show the location of model error. White dots in blue background illustrate individual
model errors (SSE = 24) where the model predicted scale infestation in areas that were
absent of infestations. The model accurately predicted absence of infestation in areas
where it was absent (SSE = 0). Created by Daniel Wieferich on 12/ 14/2006. Please note
that this image is presented in color .............................................................. 1 14

Figure 3-8b. Model of Lower Peninsula simple diffusion model: red dots in black
background illustrate individual errors (SSE = 21) where the model predicted scale
infestation in areas that were absent of infestations. White dots on blue background
illustrate a predicted absence of infestation in areas where it was present (SSE = 16).
Created by Daniel Wieferich on 12/14/2006. Please note that this image is presented in
color ................................................................................................. 115

Figure 3-8c. Model of the Upper Peninsula land cover based model: red dots in black
background illustrate individual errors (SSE = 5) where the model predicted scale
infestation in areas that were absent of infestations. White dots on blue background
illustrate a predicted absence of infestation in areas where it was present (SSE = 3).
Created by Daniel Wieferich on 03/21/2007. Please note that this image is presented in
color ................................................................................................. l 16

Figure 3-8d.

Model of the Lower Peninsula land cover based model: red dots in black

background illustrate individual errors (SSE = 14) where the model predicted scale
infestation in areas that were absent of infestations. White dots on blue background
illustrate a predicted absence of infestation in areas where it was present (SSE = 14).
Created by Daniel Wieferich on 03/21/2007. Please note that this image is presented in

color .........

Figure 3-9a.

Figure 3-9b.

Figure 3-9c.
values ........

Figure 3—9d.

Figure 3—9e.

........................................................................................ l 17
Sum of squared errors plotted against spread rate parameter values ....... 118
Sum of squared errors plotted against spread rate parameter values ....... 1 18
Sum of squared errors plotted against spread rate parameter
........................................................................................ 119
Sum of squared errors plotted against spread rate parameter values ....... 119
Sum of squared errors plotted against spread rate parameter values. l 20

xi

Introduction

Non-native forest pests and pathogens have had an increasingly profound impact
on the structure, dynamics, and ecosystem processes of forests in the past century
(Kizlinski et al. 2002; Liebhold et al. 1995). Dominant species in North American forests
such as the American chestnut (Castanea dentata [Marshall] Brokh.), American elm
(Ulmus americana L.), eastern hemlock (Tsuga canadensis [L.] Carriere), and American
beech (Fagus grandifolia Ehrh.) have suffered diebacks from exotic pests (Costello 1995;
Liebhold et a1. 1995; Orwig 2002). This study focuses on the beech scale insect
(Cryptococcusfagisuga Lind.) (Homoptera: Coccidae) which is a non-native invasive
insect and the precursor to beech bark disease (BBD). Beech bark disease has been part
of North America’s forest ecosystems since 1890, (spreading into the northeastern United
States from Canada around 1931 (Ehrlich 1934; Brower 1949; Houston 1975; Houston
and Valentine 1988). Beech scale was ﬁrst documented in Michigan in 1990. Because
of the relatively recent arrival of beech scale to Michigan, and because Michigan is
bounded by water, it provides a unique opportunity to study the spread of beech scale
infestations and to better understand the impacts of this nonindigenous insect pest on our
forests.

To understand rates and patterns of beech scale spread in Michigan, we must
determine the beech scale distribution for the entire state. This is the crucial ﬁrst step in
BBD management because the arrival of scale inevitably leads to BBD, followed by tree
death. There has never been an extensive statewide survey to document beech scale
distribution in Michigan. This project provides the most comprehensive information

about beech scale distribution across the Upper and Lower Peninsula’s of Michigan, in

addition to documenting forest characteristics and species composition for stands
containing beech. The information gathered during this study provides a baseline of the
current state of the surveyed stands so that we may better understand changes initiated by
BBD. Results will help forest health specialists; silviculturists and property owners
prioritize areas for survey, management and public outreach activities. Finally this project
will enhance our general understanding of how beech scale, a nonindigenous forest pest
spreads and increases in density, “knowledge that has become increasingly important as
we grapple with newly discovered exotic forest insect and pathogen pests” (National

Research Council 2002).

Ligasive Species

Invasions by exotic insects and pathogens are one of the most important threats to
the stability and productivity of forest ecosystems around the world (Liebhold et al. 1995;
Vitousak et al. 1996; Pimentel et al. 2000). Invasive species were ranked second,
following habitat degradation, in the list of greatest threats to biodiversity in North
America (Vitousak et al. 1996; Mooney and Hobbs 2000; USFWS 2006). Increasing
international travel and globalization of trade provide pathways for the transport of
nonindigenous species and have negated natural barriers such as oceans, rivers, and
mountain ranges that originally deterred spread of nonindigenous species (Davis 2003).
All regions of the world have been impacted by invasive species (Pimentel et al. 2000)
and huge losses in the agricultural, forestry, livestock and ﬁsheries industries have been
documented. Economic losses and expenditures resulting from the introduction of
invasive species in the United States were estimated at $97 billion in 1991 and estimated
costs in 2006 were $138 billion (USFWS 2006).

Insects and pathogens were viewed historically as two of the most important
damaging agents of forests (Hepting and Jemison 195 8). The invasion of diseases such
as chestnut blight (Cryphonectria parasitica [Murrill] Barr), practically eliminated all the
American chestnut (Castanea dentate) from northeastern forests in the early 20‘h century
(National Research Council 2002). The American chestnut tree comprised more than
one-quarter of the canopy trees in eastern forests. The loss of this species may have
initially appeared to have staggering effects on the ecosystems (Roane et al. 1986);
however, species such as oak (Quercus spp.), hickory (Carya spp.), black cherry (Prunus

serotina) and red maple (A cer rubrum) replaced chestnut in the canopy providing similar

ecosystem function (Yahner 1995; Youngs 2000). The invasive white pine blister rust
(Cronartium ribicola) attacks ﬁve-needled pines including the whitebark pine (Pinus
albicaulis). Whitebark pine, a keystone species of upper subalpine ecosystems, produces
seeds that are an important source of food for a number of birds and mammals including
nutcrackers, squirrels, and bears (Tomback et al. 2001). The above are two of many
examples that illustrate how invasive species can dramatically alter ecosystems, impact
wildlife species, and affect human economies. The ecological changes resulting from
invasive pests, typically set off a cascading chain of events leading to ecosystem changes
(Gibbs and Wainhouse 1986; National Research Council 2002). Consequences of forest
invasive species include cumulative stresses on the host plant and alteration of the
populations of other native species; effects that can extend to other trophic levels
(National Research Council 2002). This study focuses on the distribution and spread of
one of two invasive species that together cause beech bark disease (BBD), a disease
impacting our beech forests in the eastern United States and the communities of wildlife

that depend upon them.

American beech

American beech (hereafter referred to as beech) belongs to the family F agaceae
and is the only native species of the Fagus genus in North America. Beech is a slow-
growing, common, deciduous tree that attains ages of 300 to 400 years. Beech is valued
for its wood and as a source of food and habitat for wildlife. The ﬁne-grained wood is
used for ﬂooring, furniture, and baskets because it is excellent for turning, steam bending,
and veneer, it burns well and is easily treated with preservatives (Tubbs and Houston
1990; Barker et al. 1997). Beech wood is favored for fuel because of its high density and
good burning qualities (Barker et al. 1997). Beech trees are aesthetically pleasing and are
often valued by private property owners for their unique appearance in landscape
settings.

Although beech is now conﬁned to the eastern United States (except for the
Mexican population) it once extended as far west as California and probably ﬂourished
over most of North America before the last glacial period. The current range of beech
extends from Maine to northwest Florida, and west to eastern Wisconsin and Texas,
Michigan is at the northern and western edge of beech range (Figure 1-1).

Beech grows on a variety of soil types, but grows best on deep, rich, well-
drained moist soils and cool, shady, moist locations on fertile bottomlands and uplands
(Rushmore 1961). The largest beech trees are found in the alluvial bottom lands of the
Ohio and Mississippi River valleys and along the western slopes of the southern
Appalachian Mountains. Beech is found at low elevations in the North and relatively
high elevations in the southern Appalachians. Local soil and climatic factors probably

determine whether beech grows at higher elevations (Tubbs and Houston 1990).

American beech trees are a major component of three northern forest cover types
and a minor component of 17 other cover types (Tubbs and Houston 1990). Principle
associates are sugar maple (Acer saccharum), red maple (A cer rubrum), yellow birch
(Betula allegham'ensis), American basswood (Ti/fa americana), black cherry (Prunus
serotina), southern magnolia (Magnolia grandiﬂora), eastern white pine (Pinus strobes),
several hickory species (Carya spp.) and oak species (Quercus spp.) (Halls 1977; Tubbs
and Houston 1990). As a co-dominant species within the maple-beech-birch forest type,
beech inﬂuences many physical and biotic properties of the forest, including maintenance

of canOpy closure and understory light and moisture regimes (Storer et al. 2004).

Beech bark disease

Beech bark disease is caused by an etiological complex consisting of a sap-
feeding beech scale insect, the focus of this study, and one of three fungi that kill phloem
and cambium in the genus Nect‘ria. Beech bark disease has been divided into three major
phases. The leading edge of beech scale infestation is known as the advancing front. The
leading edge of the Nectria fungal invasion following the advancing front is known as the
killing front. The aftermath forest is the ﬁnal result of BBD and is characterized by dead
and declining overstory beech (Shigo 1972; MacKenzie 2004). These phases were
originally proposed by Shigo (1972) and have been widely adopted since then

(MacKenzie 2004).

Advancing front: beech scale

The ﬁrst stage of BBD begins when beech becomes infested, for the ﬁrst time,
with beech scale (Wainhouse and Deeble 1980; Houston and O’Brien 1983). Beech scale
probe the living tissues of the outer bark, extracting protoplasmic materials and causing
the death of punctured cells. Beech scale arrived in North America from Europe.
sometime in the mid-to-late 1800’s on a ship carrying European beech tree (F agus
sylvatica L.) seedlings into the Canadian port of Halifax, Nova Scotia (Ehrlich 1934;
Houston and O’Brien 1983). In 1890, some of the imported trees were found to be
infested with “felted beech coccus,” which was identiﬁed as C. fagi Baer, later renamed
C. fagisuga, the beech scale (Ehrlich 1934). Thirty years later, beech trees in forests
surrounding Halifax began dying and John Ehrlich, then a Ph.D. candidate at Harvard,
began to study, describe, and name the disease (Houston 2004).

By the early 1930’s, beech scale had spread throughout the Maritime Provinces
and into Maine. Most of New England and areas of New York were affected by scale in
the 1960’s, areas of Pennsylvania in the 1970’s, and a major infestation was discovered in
West Virginia in 1981 (Houston and Lonsdale 1979; Houston 1994). Beech scale and
BBD were discovered in the 1990’s in localized areas of Ontario, Virginia, and Ohio
(Houston 1994). Beech scale infestations were ﬁrst documented in northwestern Lower
Peninsula of Michigan in Ludington State Park, Mason County and the east-central
Upper Peninsula in Bass Lake State Forest Campground, Luce County in 2000.
Anecdotal records and data collected from affected stands suggest that beech scale was

present in Ludington State Park as early as 1991 (O’Brien et al. 2001).

Beech scale him

The beech scale is a small insect, 0.5-1.0 mm in length, reddish-brown eyes, a
stylet about 2 mm long, rudimentary antennae and legs, and numerous minute glands
(Shigo I972). The species reproduces parthenogenetically, but is univoltine. Ehrlich
(1934) and Houston and O’Brien (1983) reported an average of 50 eggs per female,
whereas Wainhouse and Gate (1988) state a maximum observed fecundity of 43, with the
average realized fecundity of 4-16 eggs per female. The yellowish-colored eggs are laid
between July and November, depending upon temperature. Eggs hatch 20-25 days later
to become ﬁrst-instar, mobile crawlers (Shigo 1972; Wainhouse 1980; Houston 1994).
Crawlers emerge from the eggs with well-developed legs and antennae (Borror and White
1970; Shigo 1972). Crawlers remain stationary under the females, migrate to cracks in
the bark, establish themselves on other trees after being disseminated by various agents,
or die (Shigo 1972). Mortality of crawlers was estimated to be about 86% but
comparatively few (<1%) were washed off the bark during rainfall, contrary to previous
assumptions (Wainhouse and Gate 1988). The crawler stage is the only mobile stage in
the beech scale life cycle where it can successfully disperse (Wainhouse 1980).

Crawlers settle into cracks or in areas where the bark is rough, usually on the bole
or large branches of the host tree. After settling, the crawler becomes stationary, forces
its stylet into the bark, becomes a second-instar nymph without legs and produces a
white-waxy ﬁlamentous secretion which completely surrounds its body. The waxy
covering protects the scales from environmental hazards such as unfavorable weather

conditions and natural enemies (Speight 1981). The insect hibemates in the second-instar

stage and molts in the spring to become an adult female and remains sessile for the rest of
its life (Borror and White 1970; Shigo 1972).

Environmental and biological controls play a relatively small part in beech scale
population dynamics. Air temperatures of -37° C (-3 5° F) are lethal to scales not
protected by snow (Houston and O’Brien 1983). No parasitoids of beech scale have been
found in Europe or North America despite repeated searches (Wainhouse and Gate 1988).
A number of native predators are known to feed on scale and are effective in reducing
scale populations on individual trees; however, their inﬂuence on the course of the
disease is of little consequence (Houston 2004). A large red velvet mite (Allothrombium
mitchelli Davis) was found to feed on beech scales in the Great Smoky Mountains
National Park (Wiggins 2001). Among Coleoptera, Coccinellidae may be effective at
reducing local populations of scale. The most common enemy is a native coccinellid, the
twice-stabbed ladybird beetle (Chilocorus stigma Say). A cecidomyid ﬂy (Lestidiplosis
sp.) is also common, but generally prefers trees with moderate to heavy scale populations
(Wainhouse and Gate 1988); both adults and larvae feed on scale (Houston 1997). Gall
gnats (Diptera: Cecidomyidae, Lestodiplosis spp.) may also be effective in reducing
beech scale populations, especially at high densities (Storer et al. 2004). The effect of
these predators on the scale population is considered negligible, but they may serve an
important function as long-range vectors for the fungi (Shigo 1962).

Beech scales eventually infest all beech trees with individual stands but certain
trees appear to be resistant (Ehrlich 1934; Shigo 1962; Houston 1983). In the
northeastern forests, this may amount to less that one percent of American beech trees

(Houston and Houston 1994, 2000). Resistant trees can occur as individuals, but often

10

are found in groups (Houston 1983) due to stump sprouting of parent trees. Individual
beech trees vary in their susceptibility to the scale insect based on genetic differences
(Speight 1981). Resistant beech tree bark contains signiﬁcantly less total nitrogen than
that of susceptible trees (Wargo 1988). Low nitrogen concentration is known to limit
establishment and growth of sucking insects (Dadd and Mittler 1965). Tree resistance to
attack may be related to the suitability of the bark for crawler settlement. Scales require
crevices in the bark that may not be present on trees < 25 years old because their bark
may be too smooth (Speight 1981). Other trees may be resistant or partially resistant to
beech scale establishment through physical or chemical attributes or genetic differences.
“Clean” trees are especially evident in some aftermath forests where their smooth, un-
cankered boles stand in sharp contrast to the highly marred stems of their susceptible
neighbors (Houston 2004). These trees were originally thought to be resistant because
the scale had no place to gain a “foothold” and lacked protection from the weather and
enemies. In Pennsylvania, scale traps were used to determine if beech scale will colonize
a tagged resistant tree if given a place to “hide” underneath the trap. These traps consist
of a piece of composite board with foam underneath tied to a resistant tree by a rope.
These traps provide suitable cover for beech scale on an otherwise very smooth-barked
tree. Challenge trials showed the trees were resistant to beech scale even if the scales are
given artiﬁcial protection (Houston 1982, 1983).

Dispersal of beech scale occurs only if crawlers or eggs settle on a suitable host.
Characteristics such as small size, a ﬂat body shape and abundant setae favor passive
dispersal. Eggs and crawlers are transported passively in airstreams, where a small

proportion of the population is wafted upward and dispersed above the canopy

ll

(Wainhouse 1980). However, at least 90-99% of wind-dispersed crawlers are thought to
travel no more than 10 m (Wainhouse 1980; Wainhouse and Gate 1988). There is little
doubt that beech scale is also moved within stands by other insects (e. g., ladybird
beetles), mammals (e. g., squirrels, raccoons) and birds (e.g., nuthatches, creepers,
woodpeckers, titmice) and probably between stands and regions by birds and people
(Houston 2004). The infestations in Michigan, West Virginia, and Ohio, for example,
appear to be centered on campgrounds or scenic areas, suggesting beech scales were
transported by humans perhaps via ﬁrewood.

Long-term monitoring of beech scale establishment and rate of spread has
important implications for public outreach efforts, design of pest surveys and silvicultural
activities. If beech scale spreads primarily by passive dispersal in wind, the rate of
spread should be somewhat predictable. However if spread is primarily by humans,
spread rates may be harder to predict, and control of this dispersal method may involve
public outreach activities designed to educate the public. Historic spread rates of beech
scale in North America have been estimated to be 6-16 km per year (Houston et al. 1979;
Wainhouse 1980; LaChance 1983; Towers 1983; Wainhouse and Gate 1988; Morin et al.
2004). These estimates include both natural and artiﬁcial dispersal. People moving
ﬁrewood, ornamental trees or logs, crawlers on clothing or pets, and vehicles bearing
eggs and crawlers are other potential modes of dispersal that leads to a different pattern
of spread. Such artiﬁcial dispersal can result in establishment of satellite populations and
accelerated spread rates when they eventually coalesce (Shigesada and Kawasaki 1997).
If artiﬁcial dispersal is common, public outreach efforts can be focused on campers or

visitors to recreation sites. Policies designed to restrict infested log or nursery stock

12

movements can be implemented. Such data could help support policies that limit
transportation of ﬁrewood into public parks or campgrounds. By understanding spread
rates, forest managers will be able to focus their management strategies along the leading
edge of the advancing front. Any understanding of how exotics behave as they invade a
new area will provide insights for foresters and wildlife biologists on how to manage
their resources and for preventing similar invasions in the future (Lewin 1987). An
improved understanding of these impacts may be useful in policy decisions relating to

exotic species introductions and to the restoration of beech forests (Storer et al. 2004).

I3

MFrgnt: Nectria fungi

The second stage of BBD or killing front refers to stands where both beech scale
populations and infection by Nectria are high, with associated tree mortality (Shi go
1972). Nectria infection begins when groups of dead cells, killed by beech scales, leads
to tearing of the periderm, which enables the Nectria fungi to initiate infection (Ehrlich
1934; Speight 1981). Once past the barrier of phellem (i.e., cork cells that make up the
ﬁrst layer of the periderrn), Nectria is able to advance through the living tissues of bark,
cortex, phloem, cambium, and sapwood. Death of the infected tissues interferes with
normal conduction and storage in the trunk and results in a progressive killing of the tree.
Tree death results when the fungal lesions have coalesced sufﬁciently to block transport
of materials to the crown of the tree. As infestation progresses, the foliage and twigs dry
and die, whole branches cease to leaf out, and large areas of bark on the trunk crack,
usually loosen from the wood, and eventually fall. On younger trees infection is less
abundant because the fungus apparently advances less readily (Ehrlich 1934). It is the
fungal infection and subsequent death of the cambium that leads to growth loss, internal
defect, decay, and tree death (Burns and Houston 1987). Dead bark will crack and ﬁssure
as the tree grows providing additional refuges for the scale and points of entry for
Nectria. Some trees may linger for several years, eventually succumbing to Nectria.

Areas devoid of beech scale or patches of black “wool”, indicative of dead scales,
are evidence of places killed by Nectria. Beech scales cannot live on dead tissue; and as
the tissues die a black fungus grows over them (Shigo 1972). Nectria may infect large
areas on some trees, completely girdling them. The leaves that come out in the spring do

not mature, giving the crowns an open appearance. The leaves turn yellow and usually

14

remain on the tree during the summer season. The chlorotic crowns are typical of trees
dying from water deﬁciency (Shigo 1972).

The killing front, infection by Nectria fungi, typically follows the advancing front
1-4 years following a heavy buildup of scale (Houston, 1996). This estimate is based on
historic records in the northeastern United States and whether this rate is consistent in
newly infested areas such as Michigan is unknown because there has never been a
Nectria distribution study conducted. Modes of transportation and fungal spread rates are
even less understood than those of beech scales. Wind and rain are documented as agents
of transportation for ascomycetes (Twery and Patterson 1984), responsible for infecting
new trees with Nectria fungi. Insect vectors may also aid the spread of Nectria spores.
Ladybird beetles (Chilocorus stigma Say) and Ambrosia beetles (Scolytidae,
Platypodidae) are strong ﬂiers; they go from tree to tree in search of food, often coming
into contact with perithecia and sporodochia of Nectria in their search (Shigo 1962).
Shigo (1964) isolated Nectria species from twice-stabbed ladybird beetles and postulated

that this beetle may serve as a long-range vector for Nectria species (Cotter 1977).

15

Nectria taxonomy

Spaulding et al. (1936) recognized that more than one species of Nectria was
causing cankers on American beech trees in North America following attack by beech
scale. Historical understanding of BBD in North America involved three different
Nectria fungi; two are native and one is introduced. Several taxa of Nectria infect the
bark of beech trees in both North America and Europe and it is not always clear whether
the tree is infected by an introduced or native species (Mahoney et al. 1999). Studies of
Neonectria population genetics revealed the native var. faginata is more closely related to
the Europe variety (Mahoney et al. 1999), leading to the hypothesis that it was
introduced, probably about the same time as beech scale (Houston 1994, 2004). While
the actual origin of N. var. faginata remains unknown, Plante et al. (2002) concurred with
Mahoney et al. (1999) that N. coccinea var. faginata found in the eastern part of North
America may have been introduced.

Nonindigenous N. coccinea var. faginata A. is the main species found in New
England, northern New York, and the Maritime Provinces. The native N. galligena A.
and N. ochreleuca A. are found in western Pennsylvania, West Virginia, and Michigan
(Houston and Mahoney 1987; Wainhouse and Gate 1988; Houston 1994; MacKenzie
2004). Nectria galligena is typically the ﬁrst species of fungus to infect beech trees
because it is already present in the forests on non-beech hosts. The nonindigenous N.
coccinea var. faginata quickly replaces the native N. galligena, as its spreads across the
country following the advancing front (Witter et al. 2004). In Europe, N. coccinea is the
only fungus associated with BBD (Wainhouse and Gate 1988). The fungi causing beech

bark canker (e. g., BBD) have recently been transferred from genus Nectria to the genus

16

Neonectria Wollenw (Castlebury et al. 2006). The genus Nectria (Hypocreales,
Nectriaceae) was described by Wollenweber (1917) based on Nectria ramulariae but was
essentially ignored until Rossman et al. (1999) recognized this genus for species
segregated from Nectria. Fungi associated with BBD will be referred throughout this
thesis as Neonectria.

Our current understanding of BBD pathogens is at least two species of Neonectria
are associated with BBD in North America. The most common is Neonectriafaginata.
The second species is Neonectria ditissima, which was previously referred to as
Neonectria galligena. For many years the fungus causing beech bark disease in North
America was recognized as Nectria coccinea var. faginata (Mahoney et al. 1999).
Castlebury et al. (2006) indicate that Neonectriafaginata should be recognized as a
distinct species from Neonectria coccinea. At present, Neonectriafaginata is known
only on Fagus in North America and Neonectria coccinea sensu stricto is known only on
Fagus in Europe. Castlebury et al. (2006) reported that the isolates from American beech
trees did not reveal any signs of Neonectria coccinea and concluded that it does not occur
in North America. Most studies have indicated that Neonectria ditissima (as Neonectria
galligena) is likely native to North America due to the large amount of genetic variation
present in North American isolates. However, without a similar comparison of the
genetic variation of European populations, it is not possible to draw conclusions
concerning the origin of Neonectria ditissima; therefore it is not clear where Neonectria
ditissima originated (Castlebury et al. 2006).

Though Neonectria may be present in the forest on other hardwood tree species,

experiments have determined that it is only able to enter and infect beech trees on which

17

the insect has been present for at least a year (Ehrlich 1934). In the absence of large
beech scale populations, pathogen spores are unable to penetrate healthy bark (Speight
1981). Once openings in the periderm are created, Neonectria spores enter the sapwood,
and mycelia spread throughout the tree (Ehrlich 1934; Houston 1994). To become
established, Neonectria spores must penetrate the cambium layer (Lortie 1964). Once
inside the tissues of a tree Neonectria grows parasitically, destroying the storage and
vascular systems of the trunk and branches (Ehrlich 1934). Eventually, the vascular
system stops functioning properly, resulting in increased leaf yellowing and eventual
death of the tree (Speight 1981). Secondary factors such as other insect pests and
pathogens cause structure weakening and tree crowns often break off during high winds a

condition referred to as “beech snap” (Houston and O’Brien 1994).

18

Afteﬂh forest

The ﬁnal stage of BBD, the aftermath forest, is characterized by poor quality
surviving trees, resistant trees, beech tree thickets consisting of small beech saplings and
relatively low levels of active disease (Shigo 1972). Declining mature beech trees often
produce dense root-sprouts that are genetically identical to parent trees and equally
susceptible to BBD (Houston 1975). Root sprouts originating from diseased trees are
generally stunted or deformed, contributing to the characteristic aftermath forest structure
that has replaced much of the original beech component of the northeastern United States
(Houston 1994; Houston and Valentine 1987; Ostrofsky and McCormack 1986). Dense
thickets of beech sucker sprouts in the northeastern United States are sometimes referred
to as “beech hell” (M. Ayers, Dartmouth College, pers. comm.) Dense understory
vegetation can limit the regeneration of other species including sugar maple (Houston
1975; Twery and Patterson 1984; Houston and Valentine 1987; Hane 2003), thereby
providing a competitive advantage for beech. Kearney (2006) did not ﬁnd an increase in
the overall abundance of beech seedlings, saplings, or recruits in Michigan forests.

Beech regeneration is also favored when browsing white-tailed deer (Odocoileus
virginianus) severely limit the height growth of more palatable (i.e., non-beech) species
(Kelty and Nyland I981; Marquis and Grisez 1978; Tilghman 1989). In northwestern
Pennsylvania, high deer densities (40-80 deer/sq mi) negatively affected the regeneration
of other tree species, such as red maple and northern red oak (Tilghman 1989). Even if
high populations of deer do not eliminate regeneration of tree seedlings, they may delay
the time period normally required for regeneration (Marquis and Grisez 1978). Michigan

and Pennsylvania are similar in forest composition and deer densities that the same effect

19

on regeneration is likely to occur in Michigan as well. These types of competition,
through crowding and selective browsing may change the species composition of the
aftermath forest to favor beech. Other studies state that as beech decline, other tree
species will replace beech. Following BBD, eastern hemlock (Tsuga canadensis) and
sugar maple (Acer saccharum) in the northeastern United States and red spruce (Picea
rubens) or ﬁr (abies spp.) in the southern Appalachians, eventually become the major
component of the forest (Twery and Patterson 1984; Runkle 1990; Gavin and Peart 1993;
Leak and Smith 1996). Kearney (2006) found that either sugar maple or red maple would
dominate the forest structure following the killing front and subsequent dieback of beech
in her study areas in Michigan.

Trees that are killed by BBD often are invaded by other insects and wood-decay
fungi. Ambrosia beetles and homtails (Hymenoptera: Siricidae) bore into the canker
areas, allowing other fungal agents to enter (Morin et al. 2001). Hypoxylon, a sapwood
decay fungus, often invades a tree. The shoestring root rot fungus, Armillaria mellea,
sometimes invades weakened trees and hastens death. “Beech snap” is an important
management concern in recreational areas, campgrounds and on private property where
property damage or injury to people, pets, or livestock can occur (McCullough et al.

2000)

20

mung:

Beech trees are used by many birds for nesting, roosting, perching, and insect
foraging (Robb and Bookhout 1995). Mammals frequently use cavities in beech trees for
shelter or dens (Tubbs and Houston 1990). Coarse woody debris, produced by mature
beech trees when they lose branches or die, facilitates travel pathways for small mammals
(Graves et al. 1988; Greenberg 2002). Mixed species forests containing beech are critical
habitat for avian species such as the hairy woodpecker (Picoides villosus), brown creeper
(Certhia americana), and solitary vireo (Vireo solitarius) (Thompson and Capen I988).

The loss of a dominant, mast-producing tree species such as beech, and its
replacement by non-mast-producing species such as hemlock, spruce, or ﬁr, not only
affects plant composition of forests, but may also negatively impact the animals that use
these trees for habitat and food (Wiggins et al. 2004). Wildlife communities depend upon
a variety of vegetation types and structures for food, habitat and space requirements.
Trees and shrubs that retain their leaves or needles throughout the winter provide thermal
cover for a variety of wildlife species. Young and immature beech trees characteristically
hold their leaves throughout the winter, providing thermal cover for a variety of wildlife
species. In colder, northern forests dominated by spruce-hardwoods, beech is the sole
hard mast producer (Tubbs and Houston 1990) and one of the few remaining mast-
producing trees at altitudes greater than 4,500 ft (Russell 1953; Whittaker 1956).
Beechnuts can be substantial components of winter diets for a variety of species
including; white-tailed deer, ruffed grouse (Bonasa umbellus), wild turkey (Meleagris
gallopavo), northern bobwhite (Colinus virginianus), and black bear (Ursus americanus)

(Glover 1949; Nixon et al. 1968; Gysel I971; Halls 1977; Beeman and Pelton 1978),

21

especially in northern regions where oaks and hickories are rare (McDonald and Fuller
1994; McLaughlin et al. 1994). BBD may signiﬁcantly reduce beech nut production by
large trees (Costello 1992). Beechnuts are high in fat and are available when other plant
foods such as ﬂeshy fruits and foliage are not (Elowe and Dodge 1989). They are also
high in calcium and moderate in crude protein and phosphorus (Halls 1977). Beechnuts
have a protein content equivalent to com (11% dry mass) and a fat content (17.3% dry
mass) ﬁve times greater than that of corn (Elowe and Dodge 1989). The loss of this mast
resource could impact numerous species of wildlife and potentially have cascading
impacts on our forest ecosystems. While many studies (i.e., Costello 1992; Storer et al.
2004; Kearney et al. 2006) have quantiﬁed wildlife resources in relation to BBD, no
quantitative observational study has concluded a decline in mast production, cavity trees,
coarse woody debris, or wildlife abundance as a result of BBD. While many effects of
BBD on wildlife species have been speculated, none is actually documented. This is a

huge gap in our understanding of BBD and these impacts need to be quantiﬁed.

22

Rate of spread

While historical records document the advance of beech scale and BBD in some
areas of northeastern North America (Houston et al. 1979), there have been few efforts to
quantify the rate and pattern of spread in newly affected areas such as Michigan In the
Allegheny National Forest in Pennsylvania, BBD has been established since at least
1985. Forest health protection specialists conducted roadside surveys, recorded beech
scale presence and beech mortality, then drew contour maps by hand to estimate temporal
progression of the advancing and killing fronts from 1985-1996 (MacKenzie 2004).
These maps provide a limited basis; however, for predicting how rapidly beech scale and
BBD may spread.

Morin et al. (2004) used. existing BBD distribution information and historic
records of invasion years to estimate a spread rate for the entire northeastern region of the
United States. Historical BBD spread rates were estimated from maps depicting the
killing front as contour lines drawn on a map incorporating year’s 1935, 1950, 1960,
1970, and 1975 (Houston 1994). Years 1990, 1999, 2000, 2001, 2002, and 2003, were
compiled into geographic information systems (GIS) to illustrate the advance of the
killing front (Morin et al. 2004). To calculate spread rates, minimum distance from each
infested county back to the area initially infested was calculated in GIS. Average radial
rate of spread was estimated by the slope of the linear regression model of the minimum
distances as a function of the year of initial infestation using. The estimated spread rate
from the regression analysis was then applied to the 2003 BBD distribution to generate a
map representing its predicted spread through 2025 over a 1 km2 raster GIS layer (Morin

et al. 2004). These calculations consider all areas behind the killing front to be infested

23

and also incorporate long-distance (or jump) dispersal into their calculations. Morin et al.
(2004) estimated that BBD spreads at a rate of 14 km/year across all land cover areas but
did not differentiate between beech scale infestation and fungal infection, or various land
cover types. Whether this rate of spread is applicable to Michigan, is not known. One
critical difference between our study and Morin et al (2004) study is that we focus
exclusively on beech scale distributions, and do not incorporate the killing front into our
spread rate calculations. Our spread rates are based on beech scale spread rates which
may not be the same rate as the Neonectria infestations. Stands may be heavily infested
with scale without Neonectria infection for several years; it is unknown how long

between fronts, particularly if a forest is isolated from the killing front.

24

 

 

 

 

 

Figure 1-1. Distribution of the American beech in North America (US. Geological
Survey, 1999).

25

Distribution of beech scale in Michigan:
Association with forest and wildlife resources.
Abstract
A total of 871 sites across Michigan were surveyed from 2004-2006 to document
the presence and level of beech scale (Cryptococcusfagisuga Lind.) infestation, identify
the advancing front and assess forest and wildlife resources. Eleven distinct beech scale
infestations were clustered into populations and were identiﬁed as covering an
approximate area of 15,095 km’. Results showed that beech scale was present in ten
counties not previously known to be infected. Stand characteristics including overstory
composition and basal area, in addition to wildlife resources such as coarse woody debris,
cavity, and snag abundance were quantiﬁed for each site. Thirty-seven other tree species
co-occurred with beech (F agisuga grandifolia Ehrh.). Common associates included:
sugar maple, red maple, northern red oak, ash species, aspen, white oak and eastern
hemlock. Basal area of beech and trees other than beech were not signiﬁcantly related to
levels of beech scale infestation. Beech diameters were positively related to levels of
beech scale infestation. Snag density was signiﬁcantly higher in moderately infested
sites than in other sites. The majority of cavity trees were beech, with non-beech trees
comprising of <1 % of total cavity trees. Beech cavity trees were present in 4% of sites
and their abundance was not signiﬁcantly different among levels of beech scale
infestation. Coarse woody debris abundance and decay class differed signiﬁcantly
among levels of beech scale infestation. Sites not infested with beech scale had the
highest abundance of coarse woody material. Volume of coarse woody debris was not

signiﬁcantly different among levels of beech scale infestation. Presence of beech snap,

26

crown dieback, and tar spots were signiﬁcantly different across levels of beech scale
infestation; however presence of beech cankers was not. Overall, the presence of beech
snap, tar spots, crown dieback and beech cankers were highest in uninfested sites. Within
infested sites, presence of beech snap, tar spots, crown dieback and beech cankers were
highest in moderately infested sites. These data will be useful for long-term monitoring
of beech scale distributions and changes in forest and wildlife resources as a result of

beech scale invasions.

27

Introduction

Beech bark disease (BBD) is caused by an etiological complex consisting of a
nonindigenous sap-feeding beech scale insect (Cryptococcusfagisuga Lind.) and a
parasitic fungus in the genus Neonectria. Beech bark disease has been divided into three
major phases. The leading edge of beech scale infestation is known as the advancing
front. The leading edge of the Neonectria fungal invasion, following the advancing front,
is known as the killing front. The aftermath forest is the ﬁnal result of BBD and is
characterized by dead and declining overstory American beech (F agisuga grandifolia
Ehrh. - hereafter referred to as beech) (Shigo 1972; MacKenzie 2004). This study
focused on the advancing front because, as a precursor for BBD, the advancing front and
areas not yet infested with beech scale provide opportunities to document forest
conditions pre-disease and to monitor effects as BBD progresses.

Beech bark disease has been studied in the United States since the 1930’s,
beginning with John Erhlich’s work in 1934. Research on BBD has addressed an array of
topics including distribution (e.g., Brower 1949; Grifﬁn et al. 2003), spread rate (e. g.,
Houston et a1. 1979; Houston 1994; MacKenzie 2004; Morin et al. 2004), pathology (e. g.,
Wollenweber 1917; Mahoney et al. 1999; Rossman et al. 1999; Castlebury et al. 2006),
effects on wildlife (e. g., Jakubas et al 2004; Storer et a1 2004; Kearney 2006), effects on
stand composition (e.g., Houston 1975 and 2001; Houston and Valentine 1987; Hane
2003; Runkle 2005). This study is unique from other BBD studies in Michigan (e. g.,
O’Brien et al. 2001; McCullough et al. 2002; Storer et al. 2004; Petrillo and Witter 2004;
Kearney 2006) in that it delineated the advancing front distribution in Michigan and

provides baseline information on forest conditions. We conducted an extensive statewide

28

survey, building off of existing BBD study sites in Michigan to document beech scale
distribution.

This project provides comprehensive information about beech scale distribution
across the Upper and Lower Peninsulas of Michigan, in addition to documenting forest
and wildlife resources in stands containing beech. This information provides a baseline
of the current state of the surveyed stands so that we may better understand changes
initiated by BBD. Results will help forest health specialists, silviculturalists and property
owners prioritize areas for survey, management or public outreach activities. Finally, this
project will enhance our general understanding of how beech scale, a nonindigenous
forest pest, spreads and increases in density, “knowledge that has become increasingly
important as we grapple with newly discovered exotic forest insect and pathogen pests”
(National Research Council 2002). The goals of this research were to map the
distribution of beech scale infestation throughout the state of Michigan and to record
stand characteristics such as coarse woody debris, cavities per species, snags, basal area
of all species, and BBD symptoms to provide baseline information on forest conditions
prior to disease. The objectives were to 1) to document the extent of the advancing front
throughout Michigan by surveying sites in all counties containing beech and 2) quantify
stand characteristics including overstory composition, basal area, coarse woody debris,

cavities, snags, beech snap, tar spots, crown condition and cankers.

29

American beech and its importance to wildlife
American beech is a major component of three northern forest cover types and a
minor component of seventeen other cover types throughout North America (Tubbs and
Houston 1990). As a co-dominant tree within the maple-beech forest type, beech
inﬂuences many physical and biotic properties of the forest, including maintenance of
canopy closure and understory light and moisture regimes (Storer et al. 2004). Mammals
frequently use cavities in beech trees for shelter or dens (Tubbs and Houston 1990). Like
the once-prominent American chestnut tree (Castanea dentata (Marshall) Borkhausen),
beech produces hard mast that is an important autumn food source for a large number of
bird and mammal species (Faison 2004). Coarse woody debris produced by mature
beech trees facilitates travel pathways for small mammals (Graves et al. 1988; Greenberg
2 002). The loss of overstory beech could impact numerous species of wildlife and

potentially have cascading impacts in forest ecosystems.

30

Beech scale
The beech scale is a univoltine parthenogenetic insect producing 4-50 yellowish-
colored eggs per adult. Eggs hatch in 20-25 days to become ﬁrst-instar, mobile crawlers
(Shigo 1972; Wainhouse 1980; Houston 1994) that may remain stationary or migrate to
new areas (Borror and White 1970; Shigo 1972). The crawler stage is the only mobile
stage in the beech scale life cycle where it can successfully be dispersed (Wainhouse
1980)
The beech scale was accidentally introduced into North America, from Europe, on
a ship carrying European beech tree (F agus sylvatica L.) seedlings into the Canadian port
of Halifax, Nova Scotia in 1890 (Ehrlich 1934; Houston and O’Brien 1983). By the early
1 930’s, beech scale had spread throughout the Maritime Provinces and into Maine. New
England and areas of New York were infested by the 1960’s. In the 1970’s,
Pennsylvania was infested and by the 1980’s, West Virginia (Houston and Lonsdale
1 979; Houston 1994). Ontario, Virginia, and Ohio reported beech scale infestations in
the 1990’s (Houston 1994). Infestations in Michigan were ﬁrst ofﬁcially documented in
2000, in northwestern Lower Peninsula’s Ludington State Park (Mason County) and the
e«ﬁst-central Upper Peninsula’s Bass Lake State Forest Campground (Luce County),
although anecdotal records indicated that beech scale was present in Ludington State Park

as early as 1991 (O’Brien et al. 2001).

31

Methods

Study Design: Site Selection

In 2004, study sites were located by systematically searching areas beyond the
boundaries of 62 research study sites established in 2002-2003 by Kearney (2006). To
the extent possible, sites were arranged in concentric circles approximately 1 km apart to
locate the advancing front. In 2005, additional sites were surveyed to further deﬁne the
advancing front using an adaptive sampling design (2—2) based on known locations of
uninfested and infested sites. Sites were established by locating beech trees midway
between two established sites where there was a discontinuity in beech scale distribution,
i.e., between a site with no evidence of beech scale and a site with evidence of beech
scale. Sequential bisections were created to deﬁne the advancing front.

Sampling to explore for disjunct populations (here termed satellite populations)
was conducted by systematically dividing a quadrangular map of the state of Michigan
( 1 :150,000) into a north and south hemisphere. Each hemisphere was then further
divided into eight to ten subsections of approximately 104 sq km in which to search for
Stands containing beech. In 2006, sites were surveyed in areas that had not been
previously visited and in cover types predicted to contain beech as a major component
according to the USDA Forest Service Forest Inventory Analysis (F IA) data. All
S€3v‘nuﬁlrches were limited to areas accessible by public or private roads. In 2005 and 2006,

Sel ected sites along the advancing front were revisited to monitor changes in levels of

111f‘estation.

32

Study Design: Plot-level measurements
At each site, I established ﬁve plots where data were collected. Data from each of
the ﬁve plots were pooled and means were calculated to obtain site-level data. The ﬁve
variable radius plots (Held 1983; Pierce and Running 1988) were established using a 10
BAF prism (Panama Angle Gauge) (Figure 2-2). The center plot was initially
established, followed by four additional plots 100 m in each cardinal direction from the
center plot (Figure 2-2). Location of each plot was recorded using a handheld GPS
(Garmin International, Inc., Olathe, Kansas) unit. GPS coordinates were recorded to the
nearest 0.001 degrees, but accuracy depended on canopy coverage and satellites
available. Accuracy ranged from i 3 m to d: 30 m. Plot-level data included basal area of
beech, basal area of all other tree species combined, number of snags, and evidence of

beech snap, porcupine damage, tar spots, crown condition, and cankers.

33

Study Design: Individual-tree measurements

Data recorded for individual trees and snags included species and dbh and number
and size of cavities. Diameter at breast height was measured on each tree at 1.3 m above
ground. To be conservative over the positive identiﬁcation among species of ash, White
ash (Fraxinus americana L), green ash (F raxinus pennsylvanica Marsh.) and black ash
(Fraxinus nigra Marsh were combined into Fraxinus genus rather than recorded as
individual species. Additionally, large-tooth aspen (Populus grandidentata) and quaking
aspen (Populus tremuloides) were also combined into Populus genus rather than recorded
as individual species. Snags were deﬁned as any dead standing tree >8 cm dbh and >1 .8
m tall (Thomas et al. 1979; Kruse 1990). Cavity trees were deﬁned as trees with a nest,

cavity, den or hollow that might shelter a hole-nesting species (Healey et al. 1989) that
were in any live tree >1 m above the ground that provided overhead shelter from
precipitation and did not have cracks or openings other than the entrance (Carey 1983).
Cavities were recorded as small (<6 cm in circumference), large (>6 cm) or multiple (two
or more cavities).

Beech were visually examined, from the ground, for beech scale. Scale
abundance was recorded using a qualitative rating of 0-4 based on visual comparisons
with standardized photographs (Figure 2-3). Beech scale abundance classes were
I”ecorded as: 0) absent, with no detectable scale presence; 1) trace, with only a few
Sczattered scales; 2) patchy infestation, with one or more dense patches of scale; 3)
WHtewashed, with heavy infestation covering the majority of bole and limbs; and 4)

dead/declining trees presumably resulting from BBD, usually covered with dead scales

e1laracteristic of “black wool” (Shigo 1976).

34

Additionally, all beech trees were visually examined from the ground to look for
beech snap, crown dieback, beech cankers and tar spots. Beech snap refers to a beech
crown that has “snapped off”, typically from the wind, after severe weakening of the stem
due to pests or pathogens and only the bole remains upright. Crown dieback was
recorded if >50% of a tree’s crown appeared dead or in severe decline. Beech cankers
were recorded if there was evidence of necrosis on the bark of the stem. Tar spots were
recorded if a black tar-like substance was evident on the stem.

Two coarse woody debris transects, each 100 m long and 2 m wide, were
established between the center and north plot and the center and west plot (Figure 2-2).
Coarse woody debris was deﬁned as dead branches, stems and boles of trees, >10 cm in
diameter, that had fallen and were at <45° anglegto the ground. Diameter at the point of
intersection and length was recorded for each individual piece. Volume of coarse woody

debris was calculated as (length x 7!: (diameter/2f) for each individual piece.

35

Statistical and spatial analysis methods

Although underlying data may not be normally distributed, means based on a
large sample size are assumed to be normally distributed (Stewart-Oaten 1995). As such,
we performed analysis on untransformed data to avoid potential problems with
transformation bias (Hayes et al. 1995). All statistical analyses were performed using
SAS (9.1; SAS Institute, Inc., Cary, North Carolina). Signiﬁcance for all statistical tests
were determined using an u=0.05.
Summary statistics (e.g., total number of trees examined, number of species
examined) are reported as total values for all sites. Most analyses were performed at the
site-level by aggregating all plot-level data across a site to obtain averages per site. Plot-

Ievel comparisons may too easily be affected by local effects or random chance. This

type of aggregation reduces variability within sites.

Sample size for each individual analysis varied and will be presented with each
analysis in the results section. Some sites were examined for beech scale infestation only
and stand-level data were not recorded. This type of sampling occurred in situations
Where we tried to delineate the advancing front and had to concentrate sites in a smaller
area and in places where stand data could not be collected without sampling bias (e. g.,
residential areas, along roadsides, in campgrounds). These sites were used in deﬁning the

advancing front but were excluded from statistical analysis to avoid any sampling bias.
Mean beech scale abundance (i.e., level of infestation) was determined for all

beech sites (n=739). Sites with no beech scale were categorized as beech scale

i1lifestation category “absent" (n=517). Infested sites were divided into three categories;

“ l ight” (n=123) included sites with mean beech scale abundance greater than zero but less

36

than one. “Moderate” sites (n=88) had a mean scale abundance greater than one but less
than or equal to three. “Heavy” sites (n=1 1) had a mean scale abundance greater than
three.

For basal area, beech dbh and snag analyses, simple descriptive statistics were
used to characterize the abundance, mean and variance in the data. General linear models
were used to test for differences among means (Searle 1987) of beech snap, tar spots,
crown dieback and beech cankers. Frequency distributions and chi-square tests were
calculated to asses associations among levels of beech scale infestation and presence of
cavity trees. Fisher’s Exact Test was used to calculate p-values because chi-square

assumptions would be violated due to a small number of expected positive occurrences
(<5). Due to a small sample size of cavity trees in general, I did not examine associations
between cavity size and beech scale infestation level.

All spatial data calculations were performed using ArcView GIS (3.2; ESRI,
Redlands, California) to calculate area of and distances between satellite infestations.
Satellite infestations were visually separated and grouped as distinct infested areas set

apart from other infestations by uninfested beech or unsuitable habitat >10 km apart.
Each satellite population was distinguished by its disjunct location in relation to other
Satellite populations and its distinctive core-to-periphery pattern of infestation. Typically
a satellite infestation had a lighter-to-heavier gradient of infestation from the perimeter to

the core respectively.

37

.R_es_ul_t§

A total of 871 sites were surveyed from 2004-2006. In total, 732 sites with beech
trees and 139 sites devoid of beech were surveyed. In addition, 67 sites along the
advancing front were re-visited to monitor spread in 2005 and 2006. Overall, 26% of all
sites were infested. In the Upper Peninsula, the percentage of infested beech sites was
higher, 47% (68 out of 144) were infested. In the Lower Peninsula, only 21% of beech

sites were infested (125 out of 588).

38

Distribution of beech scale

Beech scale infestations occurred in 15 of the 63 counties (24%) where I
surveyed sites with beech. Infestations were concentrated in the eastern Upper Peninsula
and western Lower Peninsula (Figure 2-5). The distribution in the Upper Peninsula
extends approximately 150 km east-west and approximately 75 km north-south. The
beech scale infestation distribution in the Lower Peninsula extends approximately 250
km north-south and 150 km east-west (Figure 2-5). Beech scale infestations in the Upper
Peninsula appear to be continuous while the Lower Peninsula, ten discontinuous areas of
infestation, or satellite populations were identiﬁed (Figure 2- 6). Combined, these
satellite populations cover approximately 15,100 km2 (Table 2-1). Each satellite
population in the Lower Peninsula had a distinct pattern of infestation in which it
appeared to be more heavily infested at the core and less infested towards its periphery
(Figure 2-7). The Upper Peninsula population did not have a small distinguishable core
area; rather it covers a much larger geographical area than the Lower Peninsula satellite
populations. The Upper Peninsula has a large contiguous population covering more land

area than the two largest Lower Peninsula satellite populations combined (Table 2-1).

39

Forest resources

Thirty-seven tree species co-occurred with beech within our study sites (Table 2-
1). Sugar maple (Acer saccharum Marsh.) was the most abundant (1,741), followed by
red maple (Acer rubrum Linnaeus) (321), northern red oak (Quercus rubra L.) (296), ash
species (Fraxinus spp.) (202), aspen species (Populus spp.) (201 ), white oak (Quercus
alba L.) (158) and eastern hemlock (Tsuga canadensis (L.) Carr.) respectively (130)
(Table 2-2). Mean basal areas were not different for the seven most commonly
encountered species among levels of beech scale infestation (Table 2-3). Mean beech
basal area was not statistically different among level of beech scale infestation (F1707
=1 .50; p=0.2144) however, basal areas exhibited a pattern of increase as beech scale
infestation level also showed a pattern of increase (Figure 2-8).

A total of 4,307 beech trees were examined and grouped into 12 diameter classes
(Table 2-4). The dbh of infested beech trees ranged from 10-117 cm. Mean scale
abundance signiﬁcantly increased as beech dbh increased (F 3, 3633 = 1.79; p = <0.0001)
(Figure 2-9). Within infested sites, approximately 35% of beech trees <65 cm dbh had
some level of beech scale infestation, whereas 35-55% of beech trees >65 cm dbh were
infested with beech scale (Figure 2-10). As beech trees approached 100 cm dbh, the
percentage of infested trees declined sharply (Figure 2-10), but very few trees of this size
were examined (Figure 2-1 1). The majority of infested trees (64%) had a light level of
infestation (924 out of 1447). A smaller proportion (35%) showed moderate levels of
infestation (512 out of 1447) and very few trees (<1%) were heavily infested (11 out of

1447) (Figure 2-12).

40

Beech trees were evaluated for porcupine feeding, beech snap, tar spots, crown
dieback, and cankers. No evidence of porcupine feeding on beech trees was found.
Occurrence of beech snap, tar spots, crown dieback and cankers were low, averaging less
than 3.5% of sites. Occurrence of beech snap, tar spots and crown dieback were
statistically different among levels of beech scale infestation; beech snap (p=0.0073), tar
spots (p=0.0099) and crown dieback (p=<0.0001). Occurrence of beech cankers was not
statistically different among levels of beech scale infestation (=0.0619). Overall,
presence of beech snap, tar spots, crown dieback and beech cankers were highest in
uninfested sites. Within infested sites, presence of beech snap, tar spots, crown dieback

and beech cankers were highest in moderately infested sites (Tables 5-8).

41

Wildlife resources

A total of 291 snags representing 22 different tree species were observed in 148 of
730 sites (Table 2-2). Snag density for all species other than beech was signiﬁcantly
related to levels of beech scale infestation (F 3~ 727 = 3.91; p = 0.0087; Table 2-6) and was
highest in moderately infested sites (Table 2-6). The total number of beech snags was
positively related to increasing levels of beech scale infestation (F 1, 727 = 3.07; p =
0.0272; Table 2-6). Beech snag density was higher in moderately infested sites than in
uninfested and lightly infested sites followed (Table 2-6).

Cavity trees, other than beech, totaled 33 trees out of 3,491 trees examined
(0.95%). Only 11 species of trees other than beech had cavities. Sugar maple, which was
very abundant in transects, had the most cavities (n=1 7) but only <1% of sugar maples
examined had a cavity. Other tree species generally provided a small number of cavities,
but white pine, black oak, aspen, yellow birch and red maple all had a higher percentage
of trees with cavities than sugar maple (Table 2-7). Even though the proportion of trees
with cavities increased with dbh, the number of cavity trees peaked in the 45 to 65 cm
dbh size classes because of their greater abundance (Table 2-8). Cavity tree abundance
was not different among levels of beech scale infestation, X2 (3, n= 3,524) = 3.3, p= 0.35
(Table 2-9). Beech cavity tree abundance was also not different among levels of beech
scale infestation, X2 (3, n= 5,131) = 6.63, p= 0.085 (Table 2-10).

In total, 1,230 pieces of coarse woody debris were recorded along 119 transects.
Abundance of coarse woody debris was signiﬁcantly different among levels of beech
scale infestation (F2, m. = 7.79; p=0.0004). Sites without beech scale had the greatest

amount of coarse woody material, followed by moderately infested sites and lightly

42

infested sites (Table 2-12). Volume of coarse woody debris was not signiﬁcantly

different among levels of beech scale infestation (F2, “9 = 0.24; p=0.7840; Table 2-11).

43

Discussion
Distribution of beech scale

Beech scale is more widely distributed in Michigan than previous surveys
revealed. Kearney (2006) reported that beech scale infestations were limited to a ﬁve-
county area in 2002-2003 (Chippewa, Manistee, Mason, Luce, and Oceana counties)
based upon surveys by Witter and Petrillo (2005). Michigan’s advancing front is
spreading into new areas, creating smaller satellite infestations outside the original ﬁve-
county front. Satellite populations were not evenly distributed between the Peninsulas,

with a more fragmented distribution in the Lower Peninsula.

44

Forest resources

Although many species occur in northern hardwood stands, the forest nearly
always include sugar maple, white ash, red maple, beech and eastern hemlock and
occasionally aspen and northern red oak (Eyre 1980; Tubbs and Houston 1990; Dickman
and Leefers 2003). Primary associates with beech in our study sites were similar to other
studies involving BBD in the United States (e.g., Forrester and Runkle 2000; Grifﬁn et al
2003; Latty et a1. 2003; Kearney 2006).

Few studies have reported forest conditions in relationship to beech scale
infestation; most report their results in comparison to BBD. This study focused
exclusively on the distribution of beech scale because trees infested with beech scale are
likely to eventually become infected with BBD (Ehrlich 1934; Speight 1981; Grifﬁn et al.
2003). Beech bark disease studies conducted in northeastern United States reported 80-
90% mortality of mature beech as a result of BBD (Houston 1984; Krasny I992; Leak
2006). Results from this study showed that in sites with beech scale, only 35-60% of
beech trees were infested. Lacking comparable studies, we are unable to conclude how
Michigan’s level of infestation compares to that of other areas. This study reveals only a
snapshot in time of beech infested with beech scale. The advancing front is likely too
recent in Michigan to have infested all susceptible trees within our sites and the number
of infested trees will increase.

Factors affecting population dynamics of beech scale are poorly documented.
Past studies have suggested speciﬁc geographic, climatic or biological conditions such as
extreme winter temperatures and heavy autumn rainfalls that can temporarily reduce

beech scale populations (Houston and Valentine 1988). Erlich (1934) noted that climatic

45

limitations are undoubtedly important in restricting beech scale range where it has been
present long enough to allow wide distribution as in Europe. At the time of this study, it
appeared that Michigan forests were less infested than forests in the northeastern United
States. Future studies are needed to determine if infestation rates remain the same as
infection rates.

Results from this study did not reveal a signiﬁcant relationship between beech
basal area and basal area of the other seven most abundant tree species. I also found that
mean beech basal area was not statistically different among stands with varying levels of
beech scale infestation. This ﬁnding coincides with Grifﬁn et al. (2003) and Kearney
(2006) whom did not ﬁnd a signiﬁcant difference among beech scale infestation and
density of beech in New York and Michigan, respectively. I did ﬁnd that beech basal
area was highest in moderately infested sites.

Factors that predispose a stand to beech scale infestation, and subsequently BBD,
are uncertain. Ehrlich (1934) stated that the density of beech would “inﬂuence
infestation only as they affect retention of moisture and protection against driving rains,
hot sun and strong winds.” Similar to Erhlich’s (1934) idea about beech basal area
inﬂuencing moisture retention, Twery and Patterson (1984) hypothesized that the
presence of eastern hemlock would enhance shading and moisture retention, conditions
which have been correlated with beech scale colonization and survival. Studies have
found an increase in eastern hemlock in response to the loss of beech due to BBD (Twery
and Patterson 1984; Runkle 1990; Le Guerrier et al. 2003). Eastern hemlock was one of
the most commonly occurring tree species in our sites, but there was no signiﬁcant

relationship between basal area of eastern hemlock and beech scale infestation.

46

Similarly, Grifﬁn et al. (2003) did not ﬁnd any signiﬁcant relationship between hemlock
basal area and BBD severity in New York.

We surveyed beech trees with diameters ranging 10-117 cm dbh and the larger
trees consistently were more highly infested than the smaller diameter trees. This ﬁnding
is consistent with the literature (e. g., Ehrlich I934; Shigo 1963, 1964; Houston et al.
1979; Fernandez and Boyer 1988; Runkle 1990; Grifﬁn et a1. 2003) and is probably a
result of more suitable habitat for scale on the bark of older beech. Small-diameter trees
can still be infested (Ehrlich 1934), however, as was observed during this study.

Sites were examined for evidence of porcupine feeding on beech trees because
studies in the Allegheny National Forest in Pennsylvania observed that porcupines fed on
beech trees without scale that were surrounded by infested beech trees. The author
theorized that these trees exhibited a resistance to beech scale (R. White, USDA Forest
Service Allegheny National Forest, personal communication, April 17, 2005). We did
not ﬁnd any evidence of porcupine feeding on any beech trees, regardless of beech scale
infestation.

There was no clear progression of increasing abundance of beech snap, tar spots,
crown dieback and beech cankers from lightly infested to heavily infested sites. Results
from this analysis showed that moderately infested sites had the most occurrences of
beech snap, tar spots, crown dieback and beech cankers, but there were not enough trees

in the heavily infested sites to show any strong relationship.

47

Wildlife resources

Snags are an important wildlife resource, used for a variety of taxa. They provide
perches for singing, hunting, foraging, resting and roosting, as well as foraging sites for
insect-eating birds, mammals, reptiles and amphibians (Miller 1994). In the northern
hardwood forests where most of our study sites were located, over 40 species of birds and
mammals use snags and dead portions of live trees for nest sites, dens, escape cover and
winter shelters (Evans and Conner 1979; DeGraaf and Shigo 1985). Each forest
community has different requirements in terms of the number, species and size of snags
necessary to support all the cavity users associated with that community. Height, dbh,
condition, tree species, location and abundance of snags have a direct impact on the
wildlife species that utilize a stand (DeGraaf and Shigo 1985). Bunnell et al. (2002)
suggested maintaining a target density of 2-3 large snags (30 cm dbh) and 10-20 smaller
snags per hectare throughout the stand. During this study, density of snags was highest
within moderately infested sites for both beech and non-beech species. This is likely
explained because sites there had larger-diameter trees than uninfested or lightly infested
sites. Beech scale is a relatively recent (< 20 years) invader to Michigan’s forests and as
the advancing front progresses into the killing front, changes in forest ecosystems will
likely become more evident. Results from our study showed that increasing levels of
scale were positively related to the number of snags. Following beech scale infestation,
the killing front moves through a forest stand and will result in declining health and death
of overstory beech, producing an increasing number of beech snags in the infested forest.
This study provides a baseline data for determining how snag and snag—using wildlife

may change as BBD progresses.

48

Cavity trees are trees that are living or partially living and possess a cavity large
enough to serve as shelter for birds and mammals. Cavities are created by injury, disease,
woodpeckers or loss of large limbs. The best cavity trees have healthy crowns that
protect a cavity from the elements and provide multiple beneﬁts to the occupant such as
protection from predators and foraging opportunities including mast production (Miller
1994). Smaller cavities are utilized by species such as chickadees (Poecile spp.),
nuthatches (Sitta spp.) and northern pygmy owls (Glaucidium californicum), while larger
cavities are used by species such as pileated wood peckers (Dryocopus pileatus), wood
ducks (A ix sponsa) and northern ﬂickers (Colaptes auratus). Cavities were unevenly
distributed among trees species (Table 2-7). Results from this study showed that of the
non-beech trees, sugar maple had the most cavities, but was probably more a product of
abundance than anything else. This result concurred with Keneﬁc and Nyland (2007)
who also found that sugar maple accounted for about half of observed cavity trees.

Less than 1% of all trees within study sites were observed to have cavities. Only
3.8% of cavity trees were beech, which is much lower that similar studies in the
Monongahela National Forest in West Virginia. Beech comprised of 36.7% of cavity
trees in the Kahler and Anderson (2006) study in New York whereas Carey (1983) found
29% of cavity trees were beech. Initially, we thought that the low number of cavity trees
could be explained by reduced visibility due to heavy crown cover because this study was
conducted in May—August when crowns are fully developed and can block views of upper
canopy cavities. We suspected that our number would increase in the fall after leaf

senescence. Healy et al. (1989) and Kahler and Anderson (2006) however, estimated that

49

80% of hardwood cavities were detected from the ground using binoculars. I did not use
binoculars, which may potentially explain the relatively few number of cavities recorded.

Beech accounted for the majority of cavity trees but, they represented slightly
more than 1% of all beech trees surveyed. Other studies (e.g., Kearney 2006; Gysel
1961; Robb and Bookhout 1995) also found that beech trees had more cavities than any
other species. Kahler and Anderson (2006) found that black locust (Robinia
pseudoacacia) followed by beech, were signiﬁcantly more likely to have cavities than all
other species in their studies in the Monongahela National Forest. In my study stands, we
only encountered one black locust tree (Table 2-2). Fan et al. (2003) also found beech to
be highly susceptible to cavity formation during studies conducted in Illinois, Indiana and
Missouri. They noted that beech were highly prone to damage and/or infection from a
number of sources, in part because of the thin bark and high susceptibility to ﬁre, logging
damage and decay-causing fungi. Older trees are almost invariably hollow as a result of
the presence of various heart rot fungi (Hicks 1998).

In the analysis, abundance of beech cavity trees was not signiﬁcantly related to
beech scale infestation class but, cavity trees were less common in uninfested sites than
moderately infested sites. Heavily infested sites did not have any cavities but, there were
only 11 of these sites. Kearney (2006) found that tree diameter was signiﬁcantly related
to number of cavities. If consistent, declines in mature beech from BBD will likely
reduce the abundance of cavities available for wildlife. Our data suggest that middle-
sized trees were observed to have the most number of cavities but few larger-diameter

trees were examined.

50

Kahler and Anderson (2006) cautioned against assessing the value of the beech as
a cavity tree resource because of BBD. When beech bark disease infects a forest for the
ﬁrst time, a high proportion of large, mature trees are killed (Tubbs and Houston 1990)
and replaced by trees that are too small for cavity formation (Houston 1994). Perhaps
initially BBD will increase cavity abundance; in the long-term beech may not be viable
cavity resource. What will replace it as a cavity resource in the aftermath forests is
unknown.

Dead wood lying on the forest ﬂoor is commonly referred to as coarse woody
debris. It can take the form of fallen logs, broken branches or downed treetops. Coarse
woody debris provides habitat elements useﬁrl for many species of amphibians, reptiles,
birds and mammals that may be important to their survival and migration (Harmon et al.,
1986). In an old-growth maple-beech forest, 89% of the bird species that were permanent
residents and fall/winter visitors used coarse woody debris (Williams 1936). Twenty-
eight birds, 18 mammals, 23 reptiles and amphibians and hundreds of invertebrates and
fungi use coarse woody debris in temperate deciduous forests (e. g., New England forests)
(DeGraaf and Rudis 1986; Keddy and Drummond 1996).

In the short-term, BBD has the potential to increase coarse woody debris which
will positively inﬂuence wildlife habitat but, in the long-term, it may negatively inﬂuence
wildlife habitat. Extensive volumes of literature describe the relationship between coarse
woody debris and animals (e.g., Menzel et al. 1999; Stone et al. 1999; Butts and
McComb 2000; Greenburg 2002; Bate et al. 2004). Each species of wildlife in each
region requires a different volume and size of CWD. Of all habitat variables assessed,

downed wood is the least consistently measured, and it is impossible to equate number of

51

pieces, volume, and percent cover to extract broad patterns (Bunnell and Huggard 1999).
In a study of 12 forest stands in Virginia, CWD volume ranged from 4-24 m3 /ha
(Fuhrrnan 2004). Kearney (2006) found CWD volumes in Michigan beech stands range
from 25-235 m3 /ha. I found CWD volumes considerably higher (27-36 m3 /ha) than the
Virginia study, but within the broad range of Michigan study. Many factors inﬂuence the
distribution and abundance of CWD including wind throw, topography insects and
diseases which can affect stands of trees and highly exaggerate patterns for an area
(Harmon et al. 1986; Rubino and McCarthy 2003). Much research has been conducted
on CWD in relationship to wildlife. Several sources suggest that greater mean CWD
volumes are associated with more wildlife and that low CWD volumes can be limiting to
wildlife (Harmon et al. 1986, Newton 1994, Carey and Johnson 1995), however, I was
not able to ﬁnd a quantitative estimate of CWD volumes for wildlife. Instead, wildlife
studies in relationship to CWD focused on volume, abundance and decay class as related
to wildlife populations. Hagan and Grove (1999) stated that “if forest ecologists don’t
know how much CWD is needed to maintain biodiversity, how are foresters supposed to
know?” Further study is need on this topic, especially in relation to BBD. Traditionally,
BBD has led to short-term regional increases in coarse woody debris, thus the disease
may play an important role that inﬂuences landscape scale wildlife habitat characteristics
(McGee 2000; Morin et al.). As BBD progresses in Michigan, there should be an
increasing amount of coarse woody debris as more beech die and fall to the forest ﬂoor.
Dead tree crowns will snap and fall to the forest ﬂoor, increasing downed coarse woody
debris and creating snags and openings in the canopy. Changes as a result of BBD will

likely increase wildlife habitat initially by increasing the number of snags, volume of

52

coarse woody debris and number of cavities, but the long-term effects on wildlife
populations are unknown. Further studies regarding these changes in relationship to
various stages will be extremely important as we try to understand changes initiated by

the BBD complex.

53

Management implications

New scale infestations can be difﬁcult to detect and new satellite populations are
established through many different means; humans, small mammals, birds, and wind
currents. People moving ﬁrewood or other materials bearing viable crawlers is another
potential mode of dispersal into new areas. While reviewing invasive species literature,
the following examples provide insight into potential management considerations.

Andow et al. (1990) and Muirhead et al. (2003) found that long-distance dispersal
accelerated spread rates of cereal leaf beetles (Oulema melanopus) and emerald ash borer
(A grilus planipennis) respectively, by providing opportunities for ‘nascent foci’ to
develop, from which new populations or coalescing nodes can be founded. The cereal
leaf beetle was observed to spread much faster than microscale data suggested, likely due
to macroscale movements such as through air currents or human transport (Andow et al.
1990). Likewise, the emerald ash borer diffusive spread models were “unable to account
for 17 of the 48 new p0pulations in the Great Lakes during 2004” due to long-distance
dispersal (Muirhead et al. 2003). This resulted in an artiﬁcially higher dispersal rate
when establishing new satellite infestations. Artiﬁcial dispersal can result in
establishment of satellite populations and accelerated spread rates when they eventually
coalesce (Shigesada and Kawasaki 1997). By understanding likely spread rates, forest
managers would have time to focus their management strategies along the advancing
front and to adapt their management plans to incorporate impacts from the disease and to
target property owners in the vicinity. Moody and Mack (1988) found the spread of

exotic plants to be greatly accelerated through the growth of satellite foci which

54

“eventually exceed the range occupied by the spread of a main focus” and they stressed
the importance of focusing on satellite populations in managing spread.

Focusing management efforts such as scale control, on satellite infestations would
likely be the best strategy for controlling spread and thereby reducing the forest impacts.
Taylor and Hastings (2004) suggested “eradication prioritization for isolate, low-density
smooth cordgrass (Spartina alterniﬂora) colonies as opposed to high-density core
populations owing to faster spread capabilities of the former.” Sharov et al. (2002) stated
that eradication efforts “targeted at isolated satellite colonies along the invasion front
dramatically reduced the overall rate of spread by the gypsy moth (Lymantria dispar) in
North America. While controlling isolated satellite infestations appears to be the best
management strategy for reducing the spread rate, it also is a major challenge. Locating
beech scale is easy. Theoretically, new satellite infestations can be established from a
variety of means including people moving ﬁrewood, bird or mammal migrations, wind or
water currents, therefore making it difficult to not only detect, but to manage.

Despite our best efforts to control the spread of beech scale, this will likely not
lead to a total eradication from our forests even if it was deemed a worthy endeavor and
all the funding and personnel were in place. Federal and or state quarantines are a means
to limit the transportation of infested material out of the quarantined area during the
critical period of scale development i.e., the crawler stage but like all regulations,
quarantines are not totally effective as they rely upon cooperation, enforcement and
catching every single violation of the law. Additionally, quarantines do not regulate
unintentional movement of crawlers or bird, mammal, wind or water movements. Rather,

quarantines are a people—management tool designed to slow the transport of crawlers out

55

of a known infested area. These measures do not safeguard against total compliance by
people, bird, animal or wind movements nor do they protect from transportation of
infested material out of non-quarantined areas. In short, it is a management tool only as
effective as the compliance that it receives and the data of known infestations.

The advancing front is the prelude to BBD and likely the best place to employ
management activities. We have the knowledge of the beech scale distribution in
Michigan therefore; we know where to expect BBD in the future. Additionally, we are
identifying new areas of infestation as this long-studied disease enters into new areas
such as Michigan. This knowledge can help us to manage this exotic forest pest in

forested ecosystems.

56

Table 2-1. Satellite populations of beech scale infestations in Michigan.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Number Light Moderate Heavy . Area
Satellite name .Of Scale Scale Scale InfeStfd
srtes (Km )
Beaver Island 8 2 1 5 91
Bois Blanc Island 4 2 1 1 40
Benzie County 3 2 1 0 10
Cadillac 16 1 1 3 2 595
Emmet County 7 5 1 1 506
Fisherman’s Point 1 0 0 1 60
Leelanau 2 2 0 0 I74
Ludington 59 22 15 22 2,533
Mackinaw Island 37 24 9 4 6
Silver Lake 68 17 16 35 1,267
Upper Peninsula 3 2 1 0 9,823

 

 

 

 

 

57

 

Table 2-2. Common name and number of trees by species associated with beech within

study sites. Trees are arranged in descending order according to their abundance within
study sites.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Species Number of trees examined Number of Slag
American beech 3,445 57
Sugar maple 1,741 47
Red maple 321 17
Red oak 296 8
Ash species 202 11
Aspen species 201 23
White oak 158 13
Eastern hemlock 130 12
Black cherry 1 14 7
Black oak 99 0
Unknown 91 26
White pine 76 17
Paper birch 70 19
Red pine 70 5
Yellow birch 70 9
American h0phornbeam 60 2
Northern white cedar 36 4
Jack pine 25 1
Elm species 16 3
Fir species 16 2
Spruce species 15 0
Basswood 14 0
Eastern cottonwood 12 4
Balsam fir 9 0
Black walnut 8 1
Unknown oak species 5 0
River birch 5 l
Tulip poplar 5 0
Stripped maple 4 0
Ironwood 3 0
Sycamore 3 0
Black birch 2 0
Box elder 2 0
Hickory species 2 0
Sassafras 2 0
Apple 1 0
Black locust 1 0
Choke cherry l 0
Total 7,331 289

 

 

 

 

LA
00

Table 2-3. Results from an ANOVA to compare basal area for American beech and the

seven most abundant other species across beech scale infestation classes. N is the

number of individual trees examined across sites (n=737). Basal area is reported in

 

 

 

 

 

 

 

 

 

mz/hectare.
Absent Light Moderate Heavy
n=517 sites n=123 sites n=88 sites “=11 srtes
Tree
species N Basal area Basal area Basal area Basal area pl-
Mean 4 SE Mean :1: SE Mean 3: SE Mean i SE V” "9
“3:3?“ 2,435 6.01 :h 0.06 6.22 i 0.51 7.42 :I: 0.64 7-64 i 3-26 0.2144
3"3’" 2,069 6.36 :I: 0.25 5.69 2: 0.67 5.83 i 0.06 1-15 i 4'13 0-4167
maple
Red 393 1.08:1:0.l4 1.124:0.34 l.l7:h0.32 198*0-14 ”-1990
maple
Norther“ 316 0.92 i 0.14 1.06 :1: 0.32 0.69 3: 0.30 0 090%
red oak
A39?“ 236 0.55 i 0.11 0.69 i 0.28 0.37 d: 0.25 0 0-08'1
specres
Asl‘ 213 0.71 i 0.09 0.83 :l: 0.23 ~ 0.14 :i: 0.23 0 0-8346
specles
“(:23 172 0.44 a 0.07 0.44 a 0.07 0.05 .4: 0.18 0 0209
Easter" 154 0.46 a 0.09 0.46 a 0.09 0.83 a 0.21 0 0-1703
hemlock

 

 

 

 

 

 

 

 

* Denotes signiﬁcance between uninfested and infested sites at 01 = 0.05.

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Eris ~A = 2 NE 3 3m 3 83 £33.
ﬂ o o o o o o o2 a m2
e o o mm N o o no v mm a
3 o O on m cm m om m m3
5 o o 3 v w m w me a me
am o o em 2 mm a # em mm mm
3 a LA _ 3 M: “N mm me me me
«E ~A H 2 am 3 cm on >3 mm
mmv ~A m S Vm om ow so am mm
2:. o c 3 mm mm of we owe me
mom EA v E 12 mm 3m mo Ev mm
mma _A a 2 mg g com me 0mm 3
mme EA m 2 we mm N: we 3% 3
¢m~ o o 3 em 3 me K m3 m
:33. “MW 23m SMHWME 2.8m “oh—MM 23m 2.8m eZ 28m 2 mas—O
Eoocom .053: “59.3 3.23:2 2.3on EH..— 2523 .588“:—

 

 

60

did So a; 0.8 35 moon. $5352 am: 330 new
rod 85893308 new we emcee m: E 398:: 50.62: of E 038 oﬁ E woeeomoaoc E 28 $808350 5 3288 mm 320 88:86

zoom douﬁmoeﬁ 28m cocoa mo _o>o_ 2.: 53» memecoamoboo mmﬂo 53:86 com moo: scoop mo owﬁcoeoa ES 598: Z .Ym oEah

Table 2-5. Frequency of occurrence for beech snap, tar spots, crown dieback and

cankers across levels of beech scale infestation.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No Light Moderate Heavy
Scale Scale Scale Scale Total
Infestation Infestation Infestation

Sites with beech snap 27 2 11 0 40
Sites without beech snap 426 98 60 3 587
Sites with tar spots 11 3 8 0 22
Sites without tar spots 442 97 63 3 605
Sites with crown dieback 20 2 12 2 36
Sites without crown dieback 433 98 59 1 591

Sites with cankers l 0 2 0 3
Sites without cankers 452 100 69 3 624
Total sites examined 453 100 71 3 627

 

Table 2-6. Mean nLunber of beech snags (n=44) and non—beech snags (n=3,886) per site
across levels of beech scale infestation. Basal area is reported in mZ/ha.

 

 

 

 

 

 

 

 

Beech Non-beech

Basal area Basal area

Mean :1: SE Mean :1: SE

Absent 0.16 :I: 0.05 0.87 i 0.09

Light 0.11 i 0.07 0.23 d: 0.18

Moderate 0.41 d: 0.09 1.12 :t 1.12
Heavy 0 0

 

 

6'1

 

Table 2-7. Common name and number of tree species examined within study sites.
Number of cavity trees and percentage of total cavity trees arranged by species and
presented in descending order of abundance.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Number of Number of Percent of

Species trees cavity cavity
examined trees trees

American beech 4,945 186 3.76
Sugar maple 1,741 17 0.98
Red maple 321 3 0.93
Red oak 296 2 0.68
Ash species 202 l 0.50
Aspen species 201 3 1.49
White oak 158 0 0.00
Eastern hemlock 130 0 0.00
Black cherry l 14 1 0.88
Black oak 99 1 1.01
Unknown 91 2 2.20
White pine 76 2 2.63
Paper birch 70 0 0.00
Red pine 70 0 0.00
Yellow birch 70 I 1.43
American h0phornbeam 60 0 0.00
Northern white cedar 36 0 0.00
Jack pine 25 0 0.00
Elm species 16 0 0.00
Fir species 16 0 0.00
Spruce species 15 0 0.00
Basswood 14 0 0.00
Eastern cottonwood 12 0 0.00
Balsam fir 9 0 0.00
Black walnut 8 0 0.00
Unknown oak species 5 0 0.00
River birch 5 0 0.00
Tulip poplar 5 0 0.00
Stripped maple 4 0 0.00
Ironwood 3 0 0.00
Sycamore 3 0 0.00
Black birch 2 0 0.00
Box elder 2 0 0.00
Hickory species 2 0 0.00
Sassafras 2 0 0.00
Apple 1 0 0.00
Black locust 1 0 0.00
Choke cherry 1 0 0.00

 

 

 

 

 

62

Table 2-8. Non-beech trees were divided up into 14 diameter at breast height (dbh)
classes. Each dbh-class is represented in the table by the median number in its range of
measurements (i.e., dbh-class “5” represents trees that are 1-9 cm dbh).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Diameter Number Number of Percent of
class (cm) of trees cavity trees cavity trees
5 73 0 0
15 443 2 0.45
25 1,034 3 0.29
35 854 2 0.23
45 539 6 1.11
55 288 8 2.78
65 154 5 3.25
75 56 1 1.79
85 28 2 7.14
95 10 l 10.00
105 5 1 20.00
115 5 2 40.00
135 1 0 0
145 l 0 0

 

 

 

 

 

63

Table 2-9. Chi-square table of cavity tree abundance across levels of beech scale

 

 

 

 

 

 

 

 

 

 

infestation.
Absent Light Moderate Heavy Total
Cavity trees 27 l 5 0 33
Non-cavity trees 2,598 474 413 6 3,491
Total trees examined 2,625 475 418 6 3,524

 

Table 2-10. Chi-square table of beech cavity tree abundance across levels of beech scale

infestation.

 

 

 

 

 

 

 

 

 

 

Absent Light Moderate Heavy Total

Cavity trees 132 30 24 0 I86
Non-cavity trees 3,099 1,162 673 11 4,945
Total 3,231 1,192 697 11 5,131

 

64

 

 

Table 2-11. Volume of coarse woody debris (i1 SE) and associated level of beech scale
infestation. There were 453 sites in the absent category, 100 in the light and 71 in the

moderate categories respectively.

 

 

 

 

 

 

 

 

 

 

Absent Light _ Moderate
Volume (111’) per hectare 36.25 i 4.25 33 a: 10.5 27.25 e 12.25
Mean number of pieces per hectare 10.18 d: 0.63 6.09 d: 1.85 13.73 i 1.58
Mean diameter 9.65 :I: 0.28 13.97 i 1.15 9.05 :1: 0.56
Mean length 13.67 i 0.26 10.07 :t 1.03 13.54 :t 0.64
Total number of pieces per categog 957 67 206

 

Table 2-12. Frequency of occurrence of coarse woody debris pieces in each decay class
and corresponding beech scale infestation level.

 

 

 

 

 

 

 

 

Total number
Decay Class Absent Light Moderate of pieces
1 1 1 0 7 3 8 1 55
2 343 23 73 439
3 294 26 52 372
4 1 59 5 30 194
5 5 1 6 13 70

 

 

 

 

 

 

65

  

Infested with beech scale

0

Uninfested with beech scale

Figure 2-1. Adaptive sampling design for designing the advancing front. The star
represents the midpoint between a known infested site and a known uninfested site.

 

Figure 2-2. Site layout with ﬁve plots; center, north, east, south, and west each 100 m
apart. Each site also has two 100 m coarse woody debris transects between the center
and north plot and the center and west plot.

66

 

Figure 2-3. Photos used to standardize levels of beech scale infestation. Photo on the far
left represents beech scale classiﬁcation “trace”, middle photo represents “patchy” and
right photo deﬁnes “whitewashed” (Photos taken by Nancy Schwalm, May 2004).

25 1

N
O

15

Frequency of sites

 

0 I 7 : 1 . 1
0 0.5 1 1.5 2 2.5 3 3.5 4 I
Mean Scale

 

Figure 2-4. Frequency of sites plotted against mean scale to determine beech scale
infestation classes. Mean scale was determined by aggregating all plot-level data across a
site to obtain averages per site.

67

 

 

 

 

 

 

 

 

 

 

Legend

0 Presence

 

 

 

 

0 Absence

 

No Beech Trees

 

 

 

 

Created By: Daniel Wieferich
Sept. 5. 2006

 

mgncAN STATE
UNIVERSITY

     

 

 

 

 

Miles
0 15 30 60 90 120

\

Figure 2-5. Map of Michigan, USA with study sites coded as uninfested (open white
circles) or infested (closed black circles) or no beech sites (triangles).

 

 

 

 

68

 

 

Legend

A Scale Presence

0 Scale Absence

 

Created By: Daniel Vlﬁeferich
Sept. 5, 2006

 

Miles
1 0

n=—
Q5 30 60 90 2

 

 

 

 

 

 

 

 

 

o
, ”:9
~ee~a°oooo
o o°°oo.°° °
$00.00 0 0
Q ' (0% n
6 03° 3°, 0
.1 o n
o o o
0
00° @909 0° 0 o
090 I!) <9 00 o
o 0°00 0°
° 9
o
0
V
00° 0°o
o
0 yo 0
o
0
00000 P
o 4°77”
0 o 3
o

 

 

 

 

 

 

 

 

 

Figure 2-6. Map of Michigan, USA with beech study sites grouped into eleven distinct

satellite populations designated as follows. 1-

Upper Peninsula; 2-Mackinac Island; 3-

Bois Blane; 4-Beaver Island; 5-Emmet County; 6-Fishermen's Point; 7-Leelanau County;

8-Benzie County; 9-Cadillac; 10-Ludington; 1

69

l-Silver Lake.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

° 0
O
O
o O
O o O o
o
o o 0 o
O o 0 O . 0
0 o
9 o
o 0 000
o
”:0 o
o
O 00 0 O
o
o
o
o
O
0
0 4 8 16 24
-=—_——_—_—_—Miles
f
Reference
r 1 . ,__
Scale lntensi i" g.
0 No Scale I " «A. |
3"8 . o: "
0 Trace F
I If.
O Patchy
‘1: I~i~‘,’:'._."
0 Heavy 1 Lil J
L J k I I J

 

Figure 2~7a. Map of the Ludington and Silver Lake satellite populations enlarged to
show the detail of sites coded according to their beech scale infestation level. Map
created by Daniel Wieferich on March 30, 2007.

70

 

/:/.’1”-‘:““‘-/:fﬁ 7
,//,\. - . - ° .

 

 

 

 

.
O
O
L_.,.._..-_ -..._2 _.___.._.._ .
O 0
_. O

o o 0 JR 0 0 1gb
f‘ﬁr o I .N. W . O
‘ o | F Reference ‘
L Scale Intensity .
O No Scale . *-

 

 

 

 

 

 

 

 

0 4.5 9 18
Miles 0 Trace
O Patchy
0 Heavy k

 

 

 

 

Figure 2-7b. Map of the Upper Peninsula satellite population enlarged to show the detail
of sites coded according to their beech scale infestation level. Map created by Daniel
Wieferich on March 30, 2007.

71

 

 

 

 

50 . n= 11 .

" I

l 45 ,

_ 40 I

g 35 n= 88 -

.c n= 123 i 9 1

g 30, n= 517 I

= 251 5 I 1

2 20 I _ I
I

15 . , - , . I

Absent Light Moderate Heavy I

Beech scale infestation class I

I

Figure 2-8. Mean beech basal area (i 1. SE) across level of beech scale infestation.

72

0|
O

1 ' n=ll

 

 

I s:
.D
I 1: 45
1 ‘5
I 3 n= 512
.0 4° ‘ _
4 c n— 2860
. a, I n= 924
I g 35 -
I 30 . v .
I Absent Light Moderate Heavy

Scale abundance classes

Figure 2-9. Mean beech diameter at breast height (dbh) (d: 1 SE) across levels of beech
scale infestation.

70

SDI
.I

40I

I

I

I

I

I

I

I I
101

OI

I

01

0)
0

Percent infested

N
O

15 30 45 60 75 90 105 120 I
dbh (cm) I

Figure 2-10. Percent of beech trees infested with beech scale as a function of tree
diameter at breast height (dbh).

73

1 000

I 900 ;

, 800

I . H.
E‘ 600 MD

I 3 500 I i
g 400 1:1 LT .

I "' 300 DAB

1 200

100 1

   

1158

I

1r

, I E = _.
55 65

75 85 95 105 115 125

dbh (cm) class

0
01m

Figure 2-11. Frequency of beech trees within each beech scale infestation class across
diameter at breast height (dbh) classes. Diameter at breast height classes represent the
median number in a range of dbh measurements (i.e., dbh class “5” = dbh measurements
1-9 cm, “15” = 10-19 cm...”115” = 110—109 cm). Beech scale infestation classes are
coded as “HV” for heavy infestation, “MD” for moderate infestation, “LT”, for light
infestation and “AB” for uninfested.

 

Percent infested

25 35 45 55 65 75 85 95105115125
dbh (cm) class

I

I 515

Figure 2-12. Percent of beech trees within each beech scale infestation class across
diameter at breast height (dbh) class. Beech scale infestation classes were coded as “LT”
for lightly infested, “MD” for moderately infested, and “HV” for heavily infested.

74

Modeling the spatial spread of the beech scale
(Cryptococcusfagisuga) in Michigan.

Abstract

The spread of invasive species is a growing concern for the ecological well being
of forest ecosystems worldwide. Effectively managing invasive species includes the
ability to predict how rapidly they will spread into new areas. Attempts have been made
to model the spread of beech scale (Cryptococcus fagisuga) however; few efforts have
been made to quantify the rate and pattern of spread in newly affected areas. Here I
present an approach for modeling the spread of beech scale. I surveyed all counties in
Michigan where the American beech (F agus grandifolia Ehrh.) exists, to document the
distribution of beech scale. I then utilized an inverse modeling approach to design a suite
of dynamic models, entitled SCALESPREAD, to estimate spread rates for beech scale
populations throughout Michigan. My initial model was based on a simple diffusion
model with one parameter, spread rate. I also developed a model based on land cover
type which had four parameters, spread rate for forest containing beech, deciduous non-
beech forest, coniferous forest and other land cover types. I then used the observed
distribution of beech scale to develop parameter estimates for spread rates in the Lower
and Upper Peninsulas of Michigan. Results of the simple difﬁlsion model suggested that
beech scale spreads at a rate of 1.5 km/year in the Lower Peninsula and 4 km/year in the
Upper Peninsula. The land cover based model for the Lower Peninsula indicated that
beech scale spreads at a rate of 1.5 km/year in forests containing beech, 1 km/year in
deciduous non-beech forest, 0.75 km/year in coniferous-dominated forest, and 1.5

km/year in other land cover types. The land cover based model for the Upper Peninsula

75

indicated that beech scale spreads at a rate of 5 km/year in beech forest, 2.5 km/year in
deciduous non-beech forest, 2.5 km/year in coniferous forest, and 1 km/year in other land

cover types. Although the simple diffusion model provided a reasonable ﬁt to the data,

the comparison of AICc values indicated that the land cover based model was

signiﬁcantly better than the simple diffusion model with an AICc value improvement of

81 and 38 for the Upper Peninsula and Lower Peninsula, respectively. These models

provide the ﬁrst estimate of dispersion rates for beech scale in Michigan.

76

Introduction

The spread of invasive species has a growing impact on the economic value and
ecological well being of ecosystems worldwide. By one estimate, invasive species have
staggering economic and environmental costs, approximately $137 billion per year in the
United States (Pimentel et a1 2000). Invasions by exotic insects and diseases are one of
the most important threats to the stability and productivity of forest ecosystems around
the world (Liebhold et a1. 1995; Vitousak et al. 1996; Pimentel et al. 2000). Over the last
century, forests of eastern North America have suffered devastating effects by well-
known invasive species and diseases such as the chestnut blight, gypsy moth (Lymantria
dispar Linnaeus), hemlock wooly adelgid (Adelges tsugae), and beech bark disease
(BBD) (Mattson 1997). Invasive species are also known to result in a multitude of
community level effects, including changes in plant species richness, community
structure, vegetation dynamics, and plant—animal interactions. Understanding insect
community and population dynamics are crucial to understanding invasions, and there
remains a great deal to know (National Research Council 2002).

This study focuses on the distribution and spread of beech scale (Cryptococcus
fagisuga Lind), one of two nonindigenous organisms that together cause BBD. Beech
bark disease is currently spreading across North America, endangering American beech
(Fagus grandifolia Ehrh.) resources and the communities of wildlife that depend upon
them. Beech bark disease is caused by an etiological complex consisting of the sap-
feeding beech scale and a parasitic fungus in the genus Neonectria. Beech scale ﬁrst
arrived in North America, from Europe, sometime in the mid-to-late 1800’s on a ship

carrying European beech (Fagus sylvatica L.) seedlings into the Canadian port of

77

Halifax, Nova Scotia (Ehrlich 1934; Houston and O’Brien 1983). Since that time, the
distribution of beech scale has expanded to include much of the distribution of American
beech.

Beech bark disease has been divided into three major phases: the leading edge of
beech scale infestation, known as the advancing front; the leading edge of the fungal
invasion, known as the killing front; and ﬁnally, the loss of overstory beech from the
forests, known as the aftermath forest. For early detection, locating the advancing front is
particularly important, which is why it was the focus of my research.

While historical records document the advance of beech scale and BBD in some
areas of northeastern North America (Houston et al. 1979), there have been few efforts to
quantify the rate and pattern of spread in newly affected areas such as Michigan In the
Allegheny National Forest in Pennsylvania, BBD has been established since at least
1985. Forest health protection specialists conducted roadside surveys, recorded beech
scale presence and beech mortality, then drew contour maps by hand to estimate temporal
progression of the advancing and killing fronts from 1985-1996, where they predicted
beech scale moved at 10 km/year (MacKenzie 2004). These maps provide a limited
basis; however, for predicting how rapidly beech scale and BBD may spread.

Morin et al. (2004) used existing BBD distribution information and historic
records of invasion years to estimate a spread rate for the entire northeastern region of the
United States. Historical BBD spread rates were estimated from maps depicting the
killing front as contour lines drawn on a map incorporating year’s 1935, 1950, 1960,
1970, and 1975 (Houston 1994). Years 1990, 1999, 2000, 2001, 2002, and 2003, were

compiled into geographic information systems (GIS) to illustrate the advance of the

78

killing front (Morin et al. 2004). To calculate spread rates, minimum distance from each
infested county back to the area initially infested was calculated in GIS. Average radial
rate of spread was estimated by the slope of the linear regression model of the minimum
distances as a function of the year of initial infestation using. The estimated spread rate
from the regression analysis was then applied to the 2003 BBD distribution to generate a
map representing its predicted spread through 2025 over a 1 km2 raster GIS layer (Morin
et al. 2004). These calculations consider all areas behind the killing front to be infested
and also incorporate long-distance (or jump) dispersal into their calculations. Morin et al.
(2004) estimated that BBD spreads at a rate of 14 km/year across all land cover areas but
did not differentiate between beech scale infestation and fungal infection, or various land
cover types. Whether this rate of spread is applicable to Michigan, is not known. One
critical difference between our study and Morin et a1 (2004) study is that we focused
exclusively on beech scale distributions, and do not incorporate the killing front into our
spread rate calculations. Our spread rates are based on beech scale spread rates which
may not be the same rate as the Neonectria infestations. Stands may be heavily infested
with scale without Neonectria infection for several years, particularly ifa forest is
isolated from the killing front.

People have long been interested in understanding and predicting spread rates of
invasive species because of their potential impacts on humans, the environment, and
global biodiversity (Muirhead et al. 2005). “Few events are as important in predicting the
future role of a nonindigenous plant, arthropod, or pathogen as its attainment of the
population size at which it rapidly adds members and spreads simultaneously into a new

range” (Elton 1958). Life history, morphology, and behavioral traits related to dispersal

79

of newly established species obviously play important roles in determining the rate of
range expansion, and knowledge of such characteristics would be useful in predicting the
likelihood of an invasion (Hastings 1996). Once an immigrant population has arrived, it
will become a successﬁll invader only if the population is able to increase in abundance
and spread from its point of entry. Population expansion typically consists of three steps:
an initial establishment phase with little or no expansion, an expansion phase, during
which the territory it inhabits is ﬁlled, and a saturation phase if there is a geographical
limit to the available space (Shigesada and Kawasaki 1997) (Figure 3-1). Focusing on
the expansion phase, the patterns are further divided into three categories. Type I, the
range always expands linearly with time. Type II expansion phase involves a slow initial
spread followed by linear expansion at a higher rate. Type III expansion phase occurs
when spread rate is continually increasing with time, resulting in a convex curve
(Shigesada and Kawasaki 1997). Beech scale follows a Type III expansion phase
because it has a brief establishment phase where the population is not spreading into new
areas, rather it is establishing itself. This is followed by exponential population growth
where the reproduction is high and offspring begin to disperse. The ﬁnal stage, the
saturation phase, occurs when the expansion phase levels to an asymptote, presumably
when the trees in a given area are no longer able to provide enough sustenance for the
scale and the offspring are forced to diffuse into new areas or die.

The simplest mathematical representation of range expansion by an alien invader
is a reaction-diffusion model, which combines exponential population growth with
random (diffusive) spread (Skellam 1951). This model predicts a radial rate of spread

that initially increases following establishment of the founding p0pulation, followed by a

80

period of constant radial range expansion until spread decelerates as the species saturates
its potential range (Shigesada and Kawasaki 1997). Despite its simplicity, there has been
remarkable congruence between this model’s predictions and actual spread data from a
variety of organisms (Levin 1989). It is from these basic principles that my model,
SCALESPREAD, is derived.

Ideally, the model will be simple enough to provide a general framework for
representing the spatial dynamics of scales in Michigan, yet be realistic enough to allow
predictions of their rate of spatial spread and the result will provide a tool for forest
managers to us to predict spread. There has not been an intensive statewide survey for
documenting beech scale distribution in Michigan, nor has an analysis of spread rate been
conducted This project is the ﬁrst to provide beech scale distribution information for the
entire state of Michigan while also documenting forest characteristics and species
composition. This information provides baseline data on the current state of the forests
so that changes initiated by scale infestation may be better understood and appropriate
management strategies can be developed. This project will enhance the general
understanding of how this nonindigenous forest pest becomes established and spreads,
“knowledge that has become increasingly important as we grapple with newly discovered
exotic forest insect and pathogen pests” (National Research Council 2002). In this
research, my objectives were: 1) to document beech scale distribution throughout
Michigan by surveying sites in all counties containing beech; 2) to construct a hierarchy
of dynamic models representing potential factors affecting spread rate; and 3) to compare
the modeled spread rates with estimates from other areas such as the Allegheny National

Forest in Pennsylvania.

81

Methods
Study area

From 2004-2006 the state of Michigan was intensely surveyed to locate beech
scale infestations, covering all counties with beech, based upon Michigan’s Department
of Natural Resources Integrated Forest Monitoring, Assessment, and Prescription
(IFMAP) project. In 2004, sites were located by systematical searching areas from the
edge of 62 research sites established in 2002-2003 by Kearney (2006). To the extent
possible, sites were arranged in concentric circles approximately 1 km apart to locate the
advancing front. In 2005, sites were surveyed to further deﬁne the advancing front using
an adaptive sampling design (Figure 3-2) based on known locations of uninfested and
infested sites. Sites were established by locating beech stands approximately midway
between two established sites where there was a discontinuity in beech scale distribution
(i.e., between a site with no evidence of beech scale and a site with evidence of beech
scale). Sequential bisections were created to identify the advancing front. Sampling to
explore for disjunct populations (here termed satellite populations) was conducted by
systematically dividing quadrangular maps (1:150,000) into a north and south
hemisphere. Each hemisphere was then further divided into eight to ten subsections of
approximately 104 square kilometers in which to search for stands containing beech. In
2006, sites were surveyed in areas that had not been previously surveyed and in cover
types known to contain beech as a major component according to the USDA Forest
Service Forest Inventory Analysis data. All searches were limited to areas accessible by
public or private roads. In 2005 and 2006, selected sites along the advancing front were

revisited to monitor changes in levels of infestation.

82

Within each site, beech trees were examined for scale, and scale abundance was
recorded using a qualitative rating of 0-4 based on visual comparisons using standard
photographs (Figure 3-3). Beech scale abundance was classiﬁed as: 0) absent, with no
detectable scale presence; 1) trace, with a few scattered scales; 2) patchy infestation, with
one or more patches of scale; 3) whitewashed, with heavy infestation covering the bole
and limbs; and 4) dead/declining trees, likely as a result of BBD. Mean beech scale
abundance (i.e., level of infestation) was determined for each site (n=739 sites). There
were 517 sites with no beech scale. The distribution of sites with mean beech scale
abundance >0 (i.e., infested sites) were plotted to determine infestation categories (Figure
3-4). Infested sites were divided into three categories, light (n=123) were sites with mean
beech scale abundance between 0 and 1.0. Moderately infested sites (n=88) had a mean
beech scale abundance between 1.0 and 3.0. Heavily infested sites (n=11) had a mean

beech scale abundance >3.

83

Model description and structure

I used an inverse modeling approach (e.g., Nibbelink and Carpenter 1998) that
combines the power of dynamic modeling with the need to recognize stochasticity in the
processes of pest dynamics. In this method, combinations of parameters spanning a
range of plausible values were simulated and then parameter values that best ﬁt the data
were selected (Swartzman and Kaluzny 1987). This method is described as inverse
modeling because the model itself is used to estimate the unknown parameters by ﬁtting
them to known output (Parker 1977). When model simulations were completed, the
number of statistical errors were used to rate the ﬁt of the parameter values set. The
parameter values set that produced the best ﬁt and had the least discrepancy between
predicted and observed beech scale distribution was the ﬁnal set retained.

Fundamental to this modeling approach is the idea that movement of the pest
across the landscape is analogous to a diffusion process. The rate of movement
(diffusion), however, may not be a ﬁxed number but potentially could vary depending
upon factors such as prevailing wind direction, density of beech trees, and other factors
(Shigesada and Kawasaki 1997). In this particular application, the inverse modeling
approach takes the philosophy that the rate of movement is not known, but the current
distribution is known. Thus, a mathematical search is conducted across a range of
movement rates to make predictions of distribution. The movement rate that produced
the best statistical ﬁt between the predicted distribution and the known distribution was
then selected as the best parameter set for the model. The challenge of how best to
incorporate landscape heterogeneity into the model was addressed in the land cover based

model where each land cover type had an associated permeability parameter. I deﬁned

84

the permeability parameter as the value assigned to how resistant the land cover type was
to beech scale dispersion. Smaller values of permeability were indicative of land cover
types where it was more conducive for beech scale to establish and spread (e.g., beech
forest). Higher values of permeability indicated land cover types that were unlikely
places for beech scale to establish and spread (e. g., coniferous forest).

I created a suite of models in Microsoft Excel to incorporate spatially explicit
information about the location of satellite populations and the edge of the advancing
fronts. Each model was laid out in a 100x100 km structure, with 0.5 km grid cells. This
was large enough to incorporate several of our larger satellite populations in each of the
Upper and Lower Peninsulas. Three of our smallest infestations, located in Leelanau,
Benzie and Emmet County were not included in the model because they did not ﬁt within
the modeling space. Initially, I divided the modeling landscape into 1 km grid cells but
found this level of aggregation to be too coarse for the detailed infestation data.

The models are grouped into two general categories, simple diffusion models and
land cover based models. Both categories of models simulated the spread of beech scale
for the Lower Peninsula and the Upper Peninsula. In the simple diffusion model, spread
was represented as a ﬁxed number of cells each year from infested cells in all directions
regardless of the land cover type. In the land cover based model, spread varied according
to a parameter speciﬁc to each land cover type. Land cover types were aggregated into
four categories to avoid an overly complex land cover layer and to maintain a relatively
small number of potential parameters. The four land cover types used in
SCALESPREAD were; beech (i.e., northern hardwood forest which contains the maple-

beech forest type), deciduous non-beech, coniferous-dominated forest, and other (e.g.,

85

 

urban, agriculture, water). Cover types were determined using the IF MAP classiﬁcation

system (Appendix 2).

86

Model Parameters
The simple diffusion model included only one parameter to adjust, spread rate.

This parameter allowed an infested cell to spread across a ﬁxed number of adjacent cells,
equally in all directions, into uninfested cells. The simple diffusion model included three
parameters; spread rate, year of infestation and initial infestation location (i.e., start point
cell). These parameters were adjusted to best match predicted distribution of each
satellite population with the observed distribution. At each step of the modeling process,
the parameter values were adjusted to reduce the number of modeled errors between the
predicted distribution (i.e., modeled distribution) and the actual distribution (based on
empirical data). The model ﬁtting process was iterative to adjust the model to best
represent the ﬁeld observations. Parameters were systematically adjusted one at a time
until the smallest error value was obtained. Thus, I attempted to recover all three
unknown parameters simultaneously by adjusting the year of initial infestation, start point
cell and spread rate.

The land cover based model had one parameter to control maximum spread rate
for habitats suited to dispersal of beech scale, and had four additional parameters (here I
refer to these as permeability parameters) that limit movement through cells relative to
the‘beech scale rate of dispersal based on cover type. In this model, I assumed that beech
scale would spread faster in beech cover types than other cover types, such as urban or
agricultural areas. Therefore, beech forests would have a lower permeability (i.e., easier
to penetrate) value than urban or agricultural areas. In the land cover based model,
parameter values were selected for beech forest, deciduous non-beech forest, coniferous

non-beech forest and other including urban, water, and agriculture. Using the principle of

87

parsimony in modeling, we chose to categorize wetlands, sparsely vegetated areas,
agriculture areas, lowland forests, and open land into the category “other” rather than
separate them into their own categories. Deciduous non-beech forests included oak
association, aspen association, and other upland/mixed deciduous forests that did not list
beech as an overstory species. Beech forests included the northern hardwood association.
Coniferous forests were based upon the upland coniferous forest categories. These
categories were based upon IF MAP description of classes according to IFMAP decision
rules for forest classes used in statewide maps of Michigan (Appendix 2).

Year of infestation and initial start point were constrained by three factors. The
ﬁrst constraint was documentation of a year of known infestation. A letter written from
the Michigan Department of Natural Resources documented beech scale on beech trees in
Ludington State Park Campground in 1991. This constraint was incorporated into the
modeling by not allowing the start year for the Ludington satellite population to be any
later than 1990, thereby allowing at least one establishment year prior to discovery.
Anecdotal records indicate that beech scale was reported in Bass Lake State Park
Campground in the Upper Peninsula around this same time, although no written
documentation lists a speciﬁc year. A second constraint was the empirical evidence of
beech scale spread rates area limited by the actual distribution. Preliminary evaluations
of beech scale spread using a literature value of 14 km/year (Morin et al. 2005), produced
results inconsistent with observed distribution. In fact, using a start year of 1990 for
Ludington State Park and a spread rate of 10 km/year would infest an area well beyond

the current distribution. Based on this, spread rates were limited to 10 km/year or less.

88

The ﬁnal constraint was that the initial infestation cell (and subsequent advancing front)

had to fall within the boundaries of the observed distribution for each population.

89

Model selection procedure

This modeling approach assumes that while several model formulations may
make biological sense, a single model formulation cannot be chosen a priori. Burnham
and Anderson (2002) stressed three principles that regulate our ability to make inferences
in the sciences 1) parsimony, 2) having several working hypotheses, and 3) strength of
evidence. I strove to provide the most parsimonious model that would accurately
describe beech scale spread throughout Michigan. SCALESPREAD was created as a
series of models with different cell sizes and numbers of parameters to ﬁnd the best
model ﬁt with the empirical data. I built a hierarchy of models with increasing
complexity which allowed goodness of ﬁt between model predictions and the data to
guide us in the model choice (Burnham and Anderson 2002). The principle of several
working hypothesis consists in testing a hypothesis from one experiment (e. g., start point,
year of infestation, spread rate) then according to the results, formulating a new
hypothesis to test with a new experiment (Chamberlin 1965).

Model ﬁt was measured through comparison of errors comparing the modeled
distribution to the known distribution. In this approach, errors could occur by having
predicted infestation where none was observed or having a predicted absence of
infestation where infestation was observed. These errors were recorded in 2x2
contingency tables for each of the four models.

For model selection, I computed the log-likelihood from the contingency table,
and used Akaike’s information criterion to determine which model was the most

representative and parsimonious. I used the small sample Akaike’s information criterion

(AICC) because the overall sample size was relatively small for this type of modeling

9O

(generally <40 sites per satellite population) (Burnham and Anderson 2002). As sample
size increases, the last term of the AICc approaches zero, and the AICc tends to yield the
same conclusions as the AIC (Burnham and Anderson 2002).

The AICc formula is:

2K(K+1)
n-K-l

 

AICc = AIC +

where K is the number of estimated parameters and n is the sample size. The AICc takes

both the residual variance and the number of parameters in the model into account and

balances the errors of “underﬁtting” and “overﬁtting” the models. The lower the AICc

value, the better ﬁt the model. From the output of each model, maximum likelihood
estimates of the variance were calculated as the reSidual sum of squares divided by the
number of data (i.e., number of errors).

The log—likelihood of the model given the data reﬂects the overall ﬁt of the
model. For binomial data with correct and incorrect model prediction arranged in a
contingency table, the log-likelihood can be estimated via the formula for the G-statistic

as follows:

G = 22 Oi - ln(Oz'/ Ei)

Where 0i is the frequency observed in a cell, 2 is sigma (i.e., sum), In denotes the natural

logarithm (loge to the base of e) and E,- is the frequency expected on the null hypothesis.

9]

Model Assumptions and Limitations

For many population models it is natural to assume that individuals disperse
within a continuous spatial habitat (Hardin et al. 1990). In the simple diffusion models,
spread depends upon the adjacent cells all being homogenously equally able to accept
scale infestations and the scale has an equal probability of moving to adjacent cells. The
models treat the landscape as homogenous, with respect to scale dispersion and
movement is equally probable in all directions.

For many models of biological populations it is natural to assume that time
advances discretely. This assumption corresponds to populations with seasonal life
cycles or synchronized generations (Hardin et al. 1990). Beech scales are univoltine,
which allows distinct periods of population growth on an annual basis.

Long-distance or artiﬁcial spread was not represented in SCALESPREAD;
instead I focused on natural diffusive dispersal in a contiguous area. SCALESPREAD
also did not include Michigan’s islands to determine spread rates, but these data may be
useful in the future. Additionally, the land cover based model is limited to a maximum
move of seven cells (3.5 km/year) within a function because of the limitations to the
programming environment; Microsoft Excel has a limit of seven embedded functions
within a formula. Finally, the models are based on discrete grid cells whereas the actual
distribution of beech is continuous. The edges of the discrete modeled distribution and
the observed point data did not always match up well because nature is not hard-lined,
however, this discrepancy is unavoidable in a computer environment but at a landscape

scale, the errors were negligible.

92

BELIEF:
Distribution of beech scale infestations

A total of 871 sites were surveyed from 2004-2006. In total, 732 beech sites and
139 sites devoid of beech were surveyed. In addition, 67 sites along the advancing front
were re-visited to monitor spread between study years. Overall, 26% of sites surveyed
were infested. In the Upper Peninsula the percentage of infested sites was higher, 47%
(68 out of 144) of beech sites were infested. In the Lower Peninsula, the percentage of
infested sites was lower, 21% (125 out of 588).

Beech occurs in 63 counties in Michigan. Beech scale occurred in 15 of those
counties congregated in the eastern Upper Peninsula and the western Lower Peninsula
(Figure 3-6). The distribution across the Upper Peninsula was more lateral than vertical
with the distribution extending approximately 150‘km east-west and approximately 75
km north-south. The distribution across the Lower Peninsula was more vertical than
lateral, with the distributions approximately 250 km north-south and approximately 150
km east-west (Figure 3-6). Infestation in the Upper Peninsula was more continuous than
the Lower Peninsula, where the distribution was divided into eleven discontinuous areas
of infestation or satellite p0pulations (Figure 3-7). Satellite populations were
distinguished by their disjoint location in relation to other populations of beech scale
infestation and their distinctive core-to-periphery pattern of infestation. Distance
between satellite populations ranged from 11 to 38 km apart. Michigan’s total area of
infestation covers approximately 15,095 km2 (Table 3-1).

In 2005, nine uninfested sites (selected from 2004) along the advancing front

were re-visited, of which only two became infested suggesting that beech scale spread

93

was <1 km during the 2004-2005 year. In 2006, 58 uninfested sites (selected from 2004-
2005 sites) were re-visited along the advancing front. Of these re-visited sites, 18
became infested during the 2005-2006 year, demonstrating that scale is able to disperse 1.

to 3 km annually.

94

Model performance: Simple di[fusion model

The best ﬁtting set of parameters in the simple diffusion model indicated that the
overall spread rate averaged 4 km/year in the Upper Peninsula and 1.5 km/year for the
Lower Peninsula. The simple diffusion model for the Upper Peninsula had a total of 24
errors; all of which were model over-predictions (i.e., the model projected beech scale
infestations in uninfested areas) (Figure 3-8a). The simple diffusion model for the Lower
Peninsula had a total of 37 errors; 21 of which were model over-predictions and 16 were
model under-predictions (Figure 3-8b).

A contingency table was created for each of the models (Table 3-2). The log-
1ikelihood associated with the Upper Peninsula was 12. This particular approximation of
the likelihood was conservative because the model had an under-prediction value of zero
and in the likelihood function you cannot take the log of zero. In place of the zero, I used
0.001 as the under-predicted value, which was near the asymptotic value for zero that

could be used. The log-likelihood associated. with the Lower Peninsula simple diffusion

model was 44. AICC values were calculated for the each of the simple diffusion models,

the Upper Peninsula model had an AICc value of -21 and the Lower Peninsula had a

value of -86.

95

Model Performance: Complex model

The best ﬁtting complex model parameters indicated that the overall spread rate
averaged 5 km/year in the Upper Peninsula and 1.5 km/year for the Lower Peninsula.
Spread rate was calculated for each land cover type. In the Upper Peninsula, the most
permeable land cover type was “beech” which the model predicted allows a spread rate of
5 km/year. The second most permeable land cover types were “deciduous non-beech”,
and “coniferous,” each with a spread rate of 2.5 km/year. The “other” cover type had a
much lower permeability with a spread rate of 1 km/year. In the Lower Peninsula, the
most permeable cover types were “beech” and “other”, each with a spread rate of 1.5
km/year. The second most permeable land cover types were “deciduous non-beech” and
“coniferous” with spread rates of 1 km/year and .75 km/year respectively (Appendix 1).
Sensitivity analysis was conducted by adjusting one parameter at a time and recording the
error values to determine the least number of errors. Sum of squared error values were
plotted against spread rates to visualize the sensitivity of error rates to changes in
parameter estimates and to determine the parameter set that produced the least number of
errors (i.e., the lowest point on the graph) (Figures 3-9a—e).

The complex model for the Upper Peninsula had a total of eight errors; ﬁve of
which were model over-predictions (i.e., the model projects beech scale infestations in
uninfested areas) and three were model under-predictions (Figure 3-8c). The complex
model for the Lower Peninsula had a total of 28 errors; 14 of which were model over-
predictions and 14 were model under-predictions (Figure 3-8d). Contingency tables were
created for each of the simple diffusion models (Table 3-2). The log-likelihood

associated with the Upper Peninsula was 51. The log-likelihood associated with the

96

Lower Peninsula simple diffusion model was 67. The Upper Peninsula model had an

AICc value of -93 and the Lower Peninsula had a value of -125.

97

Discussion
Distribution of beech scale infestations

The distribution of beech scale infestations in Michigan was divided into 11
satellite populations. Satellite populations were not evenly distributed between the Upper
Peninsula and the Lower Peninsula. The Upper Peninsula had one only satellite
population however it is the largest, covering more land area than the two largest Lower
Peninsula satellite populations combined. The Lower Peninsula had ten satellite
populations of varying sizes. Each satellite population appeared to be more heavily
infested at the core and become progressively less infested towards its periphery,
suggesting an advancing front that is moving away from the center of infestation.

To accurately map the advancing front, more sites were required in the Lower
Peninsula to delineate the many disjointed satellite populations. Fewer sites were
required to accurately map the advancing front in the Upper Peninsula because of the
continuous nature of the single large infestation. The Upper Peninsula has a larger
percent infested site than the Lower Peninsula which may be the result of more
continuous stands of maple-beech forest. Forest stands with a high density (e. g., 160 sq.
ft. of basal area/acre) of mature beech are considered to be a major source of dispersion
(Le Guerrier et al. 2003). In contrast, the beech scale distribution in the Lower Peninsula
is not as extensive. Substantial areas of beech in the northern Lower Peninsula remain
uninfested. In the Lower Peninsula, beech scale infestations extended from Oceana
County in the south, to Emmet County in the north. Beech scale infestations are bounded

on the west by Lake Michigan and extend eastward to Wexford County covering a

98

considerably smaller extent (<50%) of the available beech range than in the Upper
Peninsula.

There are widely-varying annual ﬂuctuations in the distance that beech scale may
spread annually. There were no observed changes in infestation of sites along the
advancing front from 2003-2004. Twenty-ﬁve percent of uninfested sites along the
advancing front became infested 2004-2005, whereas 41% of uninfested sites became
infested from 2005-2006. This suggests that beech scale spread is variable from year-to-
year and small-timescale snapshots may not accurately portray scale spread rates.
Further study on annual beech scale spread will be needed to quantify these ﬂuctuations
in spread rates to help our understanding of beech scale spread throughout Michigan.
Contributing factors such as precipitation, mast cycles, or other factors may need to be
studied in relationship to beech scale movement. Studies are needed to quantify the role
of various factors-such as wind, birds, natural migrations, human-mediated actions-that
contribute to long-distance transport of individual species” (National Research Council
2002)

Migrating birds, wildlife or humans are also likely causes of long-distance
dispersal. In studying biological invasions it was found that human-mediated dispersal
transports a signiﬁcant number of individuals to distances farther from the source than
they could disperse naturally and that many disjoint populations could not be explained
by diffusive spread (Muirhead et al. 2006; Herbert and Cristescu 2002). In this study, we
found a number of disjoint populations in the Lower Peninsula, located in areas where
humans often transport ﬁrewood, albeit at a low frequency in some areas. Analogous to a

wildﬁre, beech scale spreads from the main infestation and produces smaller “spot ﬁres”

99

(smaller disjunct infestations) that appear outside the perimeter of the main infestation.
This type of spread is likely not passive dispersal because the areas between infestations
are often uninfested beech forests, suggesting that beech scale jumped those areas.
Passive beech scale dispersal is limited to a very small area (<1 m) surrounding the
infested trees. Few larvae (<1%) become trapped in airstreams above the canopy and
potentially disperse (Houston et al. 1979). The main infestation eventually engulfs these
“spot ﬁres” and the infestation grows geometrically larger into one major infestation.
The process is repeated, with more “spot ﬁres” preceding the main front of infestation
thereby leading to a larger and faster-moving advancing front.

I examined ﬁve islands in Michigan; three of which, Bois Blane, Mackinac and
Beaver Islands, were infested and two, North and South Manitou Islands, were not. As
these islands are separated from mainland Michigan by water, short distance dispersal
consistent with what I observed on the mainland can be ruled out. Beech scale would
have to utilize long-distance dispersal mechanisms to pass over such vast areas of
unsuitable habitat. Some of the potential long-distance dispersal agents might include
birds and humans. Interestingly, the three infested islands have year-around human
residents and open access via ferry for human goods. The two uninfested islands do not
have year around human residents and have a limited number of visitors and restricted
human goods. North and South Manitou Islands are managed by the National Park
Service and have a restricted number of “low impact” campers per day. This dichotomy
suggests that humans, and not birds, are the predominant agent in long-distance dispersal

to these islands. This also suggests that birds may be a less important agent in long-

100

distance dispersal on the mainland as well. F ollow-up research would be important to

monitor these islands through time.

101

Model Performance

The simple diffusion model performed very well despite having only a single
spread rate parameter. Both the simple and complex models predicted the same spread
rate for the Lower Peninsula. The only difference between the two model’s predictions
was regarding the spread rate for the Upper Peninsula. The simple diffusion model
predicted a spread rate of 4 km/year and the complex model predicted a spread rate of 5
km/year.

The land cover based models reduced the total number of sum of squared errors

from 61 to 36 which lead to a substantial reduction in the AICc value, despite the increase

number of parameters. The simple diffusion model provided a reasonable ﬁt to the data
and if a parsimonious model is most desirable, would be an acceptable substitution for
the more complex land cover based models. These models provide the ﬁrst estimate of
dispersion rates for beech scale in Michigan, and the ﬁrst estimates of differences in
dispersal rate as a function of land cover.

The Lower Peninsula has a more heterogeneous landscape than the Upper
Peninsula (Figure 3-5). This heterogeneity may account for the presence of satellite
populations in the Lower Peninsula as compared to the larger population in the Upper
Peninsula where the forested landscape is not as fragmented. This is consistent with
conclusions by the National Research Council that “Environmental heterogeneity, the
patchiness of the environment, can also inﬂuence the rate of spread” (National Research
Council 2002). This may also account for the differences in spread rates between the
Upper and Lower Peninsulas. Overall spread rate was four times faster in the Upper

Peninsula than in the Lower Peninsula. The continuous beech forest may allow beech

102

scales to move more easily throughout the landscape, thereby increasing the spatial
spread at a more accelerated rate than in areas, such as the Lower Peninsula, where beech
forest patches are surrounded by non-suitable habitat for beech scales to disperse through.

The pattern of scale infestation in the Upper Peninsula satellite population
suggests that there may have been more than one point of infestation. Though I lack
direct evidence of multiple initial points of infestation, l theorize that sometime after the
original 1990 infestation, a second infestation started in the Upper Peninsula and has
already coalesced with the original satellite population. My theory of more than one start
point may explain some of the discrepancies in the modeling. Without supporting data
(e. g., historical observations), I did not incorporate a second start point in either of the
Upper Peninsula’s models. This may also explain the higher spread rate estimated for the
Upper Peninsula. Perhaps beech scale is not spreading faster in the Upper Peninsula, but
the larger extent represents the impact of a several smaller satellite infestations coalescing
into a larger population. It is interesting to note that my estimates for spread are
considerably smaller than prior BBD spread estimates (6-16 km/year). This discrepancy
may be due in part to what is actually being modeled. One signiﬁcant difference in this
study versus Morin et al. (2004) is that we did not consider the spread rate of the killing
front which may be faster than the advancing front. Additionally, Michigan’s beech scale
infestation is relatively new and the satellite infestations were more easily recognized as
individual infestations. Our modeling efforts calculated spread rate based upon these
individual infestations rather than a conglomerate of all infestations. This distinction can
have an effect on spread rate that gives the false impression of a faster spread rate.

Historic BBD spread rates may have unknowingly incorporated several disjoint

103

infestations into one larger infestation and therefore appears that BBD has engulfed more

landscape than occurred via local dispersive processes.

104

Management implications

Nonindigenous species are often rare during the establishment phase of their
colonization and are typically not detected for several years; therefore it is difﬁcult to
know how long they persisted in a certain area (Carey 1996; National Research Council
2002). Cryptic behavior such as “hide and survive” techniques, small bodies,
concealment within small crevices in bark and their ability to remain in a dormant stage
increase an invasive insects chance of survival during transport to new areas and reduce
the likelihood that it will be detected in its new environment (National Research Council
2002). Low population densities might be exacerbated in species, such as the beech
scale, that can reproduce parthenogenetically and lack the limitations of a sexually-
producing organism. Additionally, nonindigenous insects initially lack natural enemies
which aides in their survivability. New populations must exceed a threshold density
before it can easily be detected and this threshold will depend on traits or behavior of the
organism, including the extent of the damage it causes (National Research Council 2002).

Inadvertent transportation of a nonindigenous organism by humans can establish
new foci at substantially greater distances than would occur by natural dispersal
mechanisms of the species. Such transportation has been shown to have substantially
increased the spread rate of such species as the gypsy moth. Bird dispersal is another
common mechanism for invasive spread in forest ecosystems (National Research Council
2002). Moody and Mack (1988) stressed the importance of focusing on satellite
populations in controlling spread of invading plants because when two smaller
populations eventually coalesce, they produce a single faster-spreading advancing front.

In an invasive plant study, Taylor and Hastings (2004) suggested eradication efforts to

105

target isolated, low-density colonies as opposed to high-density core populations owing
to faster spread capabilities of the former. Since the onset of this project, two of the
major satellite populations in the Lower Peninsula, Ludington and Silver Lake, have
coalesced into one larger population. Within even a few years time it may be
increasingly difﬁcult to distinguish one satellite population from another, especially in
the Lower Peninsula.

If beech scale spreads primarily through passive dispersal (e. g., wind) the rate of
spread should remain fairly predictable. However, if artiﬁcial spread, human-induced
(e. g., movement of ﬁrewood) may be harder to predict, and control of this dispersal
method may involve public outreach activities designed to educate the public and limit
artiﬁcial spread. Such artiﬁcial dispersal can result in establishment of satellite
populations and accelerated spread rates when they eventually coalesce (Shigesada and
Kawasaki 1997). Long-term monitoring of the establishment and rate of beech scale
spread, the precursor to BBD, has important implications for public outreach efforts,
design of pest surveys and silvicultural activities. Public outreach efforts could focus on
education or policies that limit the transportation of ﬁrewood out of infested areas. By
understanding spread rates, forest managers and property owners have time to incorporate
impacts of BBD into their management plans to mitigate losses of beech resources along
the advancing front. General patterns can be recognized and may be useful for predicting
invasions (National Research Council 2002).

Managers of natural landscapes are faced with decisions about how to control
these species and or minimize their impacts on the natural resource on a limited budget.

Very often the problem with invasive species is that they are so extensive in their range

106

that management actions are generally taken after a nonindigenous species has invaded,
rather than preventative action (N eubert 2004). Decisions on how to best allocate
resources to nonnative species management should be based on a risk analysis that
evaluates the potential for long-term, negative effects on natural ecosystems, including
populations of native species (N eubert 2004). I would suggest focusing management
strategies on controlling the small satellite populations before they have the opportunity
to coalesce into larger faster-moving fronts. Smaller populations are logistically easier to
contain and treat before an area becomes too large. In a study on invasive plants, Mack
(1985) concluded that the area occupied through the growth of satellite foci eventually
exceeds the range occupied by the spread of a main focus. If the initial area of a single
large focus and the initial collective area of many small foci are equal and all grow at the
same constant rate, the small foci will collectively occupy space much faster than the

single large focus.

107

Table 3-1. Satellite infestations separated into beech scale infestation class and
approximate size of the area infested as of 2006. Area was calculated using the area
feature in ArcGIS.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Estimated Approximate
Start area of
Satellite name Year infestaztion

(km )
Beaver Island * 91.23
Bois Blane Island * 40.25
Benzie County * 448.67
Cadillac 2004 595.09
Emmet County * 506.14
Fisherman’s Point 2004 59.82
Leelanau * 173.82

Ludington 1989 2,386.06

Mackinaw Island * 6.17

Silver Lake 1997 964.84

Upper Peninsula 1989 9,823.06

 

* Indicates satellite populations that were not included in modeling.

Table 3-2. Contingency table of model errors for each of the models.

 

 

 

 

 

 

 

 

 

 

Modeled Scale
Observed Scale Occurrence Occurrence
Present Absent

Upper Peninsula Present 44 0
simple diffusion Absent 24 6
Lower Peninsula Present 50 21
simple diffusion Absent 16 69
Upper Peninsula Present 39 3
complex land cover Absent 5 27
Lower Peninsula Present 52 14
complex land cover Absent 14 76

 

 

 

 

 

 

108

 

 

m ” --------- I’ --------- ’ ------
.2 I /
E Type 3 Type 2 Type 1
8,
c
as
C
O
establishment Tlme

phase

Figure 3-1. Three types of rang-versus-time curves. Range expansion patterns
commonly have an establishment phase (arrow), expansion phase (solid line), and
saturation phase (dashed line), successively. The expansion phase is classiﬁed into three
types. Type 1 shows linear expansion. Type 2 exhibits biphasic expansion, with an
initial slow slope followed by a steep linear slop. In type 3, the rate of expansion
continually increases with time (Shigesada and Kawasaki 1997).

  

Infested with beech scale

0

Uninfested with beech scale

Figure 3-2. Adaptive sampling design for designing the advancing front. The star
represents the midpoint between a known infested site and a known uninfested site.

109

 

Figure 3-3. Photos used to standardize levels of beech scale infestation. Photo on the far
left represents beech scale classiﬁcation “trace”, middle photo represents “patchy” and
right photo deﬁnes “whitewashed” (Photos taken by Nancy Schwalm, May 2004).

NNQJO)
001001

Frequency of site:
E3 5‘»

 

Mean Scale

 

 

 

Figure 3-4. Frequency of sites plotted against mean scale to determine beech scale
infestation classes. Mean scale was determined by aggregating all plot-level data across a
site to obtain averages per site.

110

 

   

  
 
     
 
 
    
  

5299—“
El Non Forested
Northern Hardwoods
- Oak Association
l:l Decidous
- Coniferous

Created By: Daniel Wieferich
October 18. 2006

MICHIGAN STATE

 

 

UNIVERSITY

O 15 30 60

 

90
Miles

 

 

Qata extracted from IF_MAP data 2001, from MCGI website (www.mcgi.state.mi.us).

 

Figure 3-5. Map of Michigan, USA with a layer of forest types grouped to locate beech.
Beech was typically found within northern hardwood or deciduous forest cover types.
Typically beech was not abundant in the oak association, coniferous or non-forested
forest types. Data was extracted from IF_MAP data 2001, from MCGI website
(www.mcgi.state.mi.us). Map was created by Daniel Wieferich on October 18", 2006.
Please note that this image is presented in color.

111

 

Legend

0 Presence

0 Absence

No Beech Trees

Created By: Daniel Wieferich
Sept. 5, 2006

MICHIGAN S
—-———~———————————— ..

UNIVER

 

 

 

Miles
120

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3-6. Map of Michigan, USA with study sites coded as uninfested (open white
circles) or infested (closed black circles) or no beech sites (small triangles).

112

 

 

 

Legend
A Scale Presence

0 Scale Absence

 

Created By: Daniel Viﬁeferich
Sept. 5, 2006

Miles
120

ct—Z—
{5 30 60 90

 

o o .‘ “f A
o ‘ ‘
o ‘3: ‘
O ‘A A
o o A A0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-6. Map of Michigan, USA with beech study sites grouped into eleven distinct
satellite populations designated as follows. l-Upper Peninsula; 2-Mackinac Island; 3-
Bois Blane; 4-Beaver Island; 5-Emmet County; 6-Fisherrnen's Point; 7-Leelanau County;

8-Benzie County; 9-Cadillac; lO-Ludington; ll-Silver Lake.

113

 

Legend
2006 Scale Presence _ _ ' "
1:] Absence o 5 10 20 30 4o
- Presence

Basic Model Prediction
- Absence

- Presence

 

 

 

Figure 3-8a. Modeled errors for the Upper Peninsula simple diffusion model mapped to
show the location of model error. White dots in blue background illustrate individual
model errors (SSE = 24) where the model predicted scale infestation in areas that were
absent of infestations. The model accurately predicted absence of infestation in areas
where it was absent (SSE = 0). Created by Daniel Wieferich on 12/14/2006. Please note
that this image is presented in color.

114

 

 

Legend
_ . - — '
Basic Model Predrctron 5 10 20 30 40
- Absence
_ Presence
2006 Scale Presence
[:3 Absence
- Presence

 

 

 

Figure 3-8b. Model of Lower Peninsula simple diffusion model: red dots in black
background illustrate individual errors (SSE = 21) where the model predicted scale
infestation in areas that were absent of infestations. White dots on blue background
illustrate a predicted absence of infestation in areas where it was present (SSE = 16).
Created by Daniel Wieferich on 12/14/2006. Please note that this image is presented in
color.

115

 

 

 

Lem ..
2006 Scale Presence 5 10 20 30 4o
[:3 Absence

_ Presence

LU Model Prediction
- Absence

- Presence

 

 

 

Figure 3-8c. Model of the Upper Peninsula land cover based model: red dots in black
background illustrate individual errors (SSE = 5) where the model predicted scale
infestation in areas that were absent of infestations. White dots on blue background
illustrate a predicted absence of infestation in areas where it was present (SSE = 3).
Created by Daniel Wieferich on 03/21/2007. Please note that this image is presented in
color.

116

 

 

ngend .
LU Model Prediction 0 5 10 20 30 40
- Absence
- Presence
2006 Scale Presence
[:1 Absence
- Presence

 

 

 

Figure 3-8d. Model of the Lower Peninsula land cover based model: red dots in black
background illustrate individual errors (SSE = 14) where the model predicted scale
infestation in areas that were absent of infestations. White dots on blue background
illustrate a predicted absence of infestation in areas where it was present (SSE = 14).

117

 

so -
7o ..
60 l ’
50 , l
40
30
20
10

O l l I ' l 7 l ' I i
0 1 2 3 4 5

Spread Rate

0. O.-. O

0%..

“O...

”.00 0

Sum of Squared Error

 

 

O 0.0... O

 

Figure 3-9a. Sum of squared errors plotted against spread rate parameter values.

 

80
70 s
60 4
50

40 r
30 e
20 w
10

0 ﬂ 7 ‘ 'T T‘ T" 7 7' 7 I ' 7' I T I
O 0.5 1 1.5 2

Beech Forest

0“

O. “O
.0“.
00.00

Sum of Squared Error
0

 

 

2.5

 

Figure 3-9b. Sum of squared errors plotted against spread rate parameter values.

118

 

 

 

Sum of Squared Error

 

 

20 i
10

 

0.5

O

000.

O

1.5

Deciduous forest

.0 009

O

2.5

0.0

O.

 

Figure 3-9c. Sum of squared errors plotted against spread rate parameter values.

 

Sum of Squared Error

 

80
70 *

60 i
50 .

40
30

10

 

0.5

O...

O

A

“000

1.5

Coniferous Forest

W00 .0

0.009

2.5

.00.

 

119

 

 

Figure 3-9d. Sum of squared errors plotted against spread rate parameter values.

 

O”

70
60

.00

40

O
OO
00

20
10

Sum of Squared Error

 

Other

 

O.

 

Figure 3-9e. Sum of squared errors plotted against Spread rate parameter values.

120

 

APPENDICIES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Conny Site Latitude Longitude Date Scale
ALCONA 1082 44.846483 83.445100 5/30/2006 No Scale
ALCONA 1083 44.779750 -83.403950 5/30/2006 No Scale
ALCONA 1084 44.683800 -83.403650 5/30/2006 No Scale
ALCONA 1085 44.569750 -83.368150 5/30/2006 No Beech
ALCONA 1086 44.566800 -83.592017 5/30/2006 No Beech
ALCONA 1087 44.692800 83.666033 5/30/2006 No Scale
ALCONA 1093 44.784500 83.832050 5/30/2006 No Scale
ALGER 200 46.665983 86.009600 8/10/2004 Patchy
ALGER 201 46.633900 86.1 19517 8/10/2004 Trace
ALGER 202 46.606183 86.215600 8/10/2004 No Beech
ALGER 203 46.588650 86.222817 8/10/2004 No Scale
ALGER 204 46.526350 86.232933 8/10/2004 Trace
ALGER 205 46.516050 86.327633 8/11/2004 No Scale
ALGER 206 46.485133 86.394683 8/11/2004 No Scale
ALGER 207 46.428433 86.448917 8/1 1/2004 No Beech
ALGER 208 46.419300 86.546400 8/11/2004 No Scale
ALGER 209 46.489367 86.954217 8/12/2004 No Scale
ALGER 210 46.460750 86.926150 8/12/2004 No Beech
ALGER 211 46.376800 86.835800 8/12/2004 No Scale
ALGER 212 46.393467 86.766133 8/12/2004 No Scale
ALGER 213 46.375550 86.700533 8/12/2004 No Scale
ALGER 214 46.307833 86.716617 8/12/2004 No Scale
ALGER 215 46.264133 86.627550 8/12/2004 No Scale
ALGER 219 46.358050 86.469700 8/13/2004 No Beech
ALGER 732 46.655183 85.929017 7/12/2005 Whitewashed
ALGER 743 46.467267 86.552500 7/6/2005 No Scale
ALGER 745 46.407150 86.574450 7/6/2005 No Scale
ALGER 747 46.319500 86.615700 7/6/2005 No Scale
ALGER 750 46.557283 86.363767 7/6/2005 No Scale
ALGER 752 46.564083 85.927800 7/12/2005 Whitewashed
ALGER 760 46.333350 86.799167 7/25/2005 No Scale
ALGER 761 46.506917 86.363483 7/7/2005 No Scale
ALGER 763 46.520783 86.268700 7/7/2005 Patchy
ALGER 765 46.535583 86.252533 7/7/2005 Whitewashed
ALGER 767 46.5523 83 86.150400 7/7/2005 No Scale
ALGER 769 46.547500 86.062600 7/7/2005 Trace
ALGER 1170 46.205850 86.754167 6/27/2006 No Scale
ALGER 1172 46.436767 86.658250 6/28/2006 No Scale
ALGER 1173 46.396000 86.928233 6/28/2006 No Scale
ALGER 1174 46.381317 87.086033 6/28/2006 No Scale
ALGER 1253 46.557233 86.3641 17 7/24/2006 Trace
ALGER 1254 46.552283 86.150383 7/25/2006 Whitewashed
ALGER 1255 46.520517 86.303133 7/25/2006 No Scale
ALGBR 1256 46.506783 -86.363483 7/25/2006 Trace
ALGER 1257 46.539867 86.446233 7/25/2006 Trace
ALGER 1258 46.467233 86.552517 7/25/2006 No Scale
ALGER 1259 46.474533 86.428000 7/25/2006 No Scale

 

 

 

 

 

 

122

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ALGER 748 46.649020 -86.027480 7/14/2005 Patchy
ALLEGAN 838 42.703050 -86. 197217 8/8/2005 N0 Scale
ALLEGAN 1063 42.647167 -85.898567 5/25/2006 No Beech
ALLEGAN 1064 42.636000 -85.898767 5/25/2006 No Scale
ALLEGAN 1065 42.485067 -86.010467 5/26/2006 No Scale
ALLEGAN 1066 42.553000 -85.995700 5/26/2006 No Scale
ALPENA 1078 45.176150 -83.731100 5/29/2006 No Scale
ALPENA 1079 45.0421 17 -83.563283 5/29/2006 No Beech
ALPENA 1080 44.982733 -83.616967 5/29/2006 No Scale
ALPENA 1081 44.898583 -83.668150 5/29/2006 No Beech
ANTRIM 585 44.881833 -85.423033 6/15/2005 No Scale
ANTRIM 587 44.863150 -85.353250 6/15/2005 No Scale
ANTRIM 591 44.989833 -85.l34683 6/15/2005 No Scale
ANTRIM 599 44.961200 -85.134650 6/15/2005 No Scale
ANTRIM 618 45.064567 -84.907433 6/20/2005 No Scale
ANTRIM 663 45.168500 -85.226467 6/22/2005 No Scale
ANTRIM 666 45.049900 -85.153950 6/22/2005 No Scale
ANTRIM 668 45.090683 -85.138650 6/22/2005 No Scale
ANTRIM 671 45.182200 -85.376550 6/22/2005 No Scale
ANTRIM 673 45.075667 -85.361200 6/22/2005 No Scale
ANTRIM 675 45.159350 -85.299850 6/22/2005 No Scale
ANTRIM 677 45.048683 -85.299383 6/22/2005 No Scale
ANTRIM 679 45.029750 -85.217217 6/22/2005 No Scale
ANTRIM 681 45.103017 -85.144617 6/22/2005 No Scale
BARRY 1067 42.594717 -85.465600 5/26/2006 No Scale
BAY 1 194 43.668550 -83.908917 7/6/2006 No Beech
BENZIE 535 44.535767 -86.105417 6/8/2005 No Scale
BENZIE 536 44.590067 -86. 1 023 67 6/8/2005 Trace
BENZIE 537 44.683267 -86.1 13633 6/8/2005 No Scale
BENZIE 538 44.720683 -86.062250 6/8/2005 No Scale
BENZIE 539 44.764000 -86.074600 6/8/2005 N0 Scale
BENZIE 545 44.757517 -8S.999050 6/9/2005 No Scale
BENZIE 547 44.718183 -85.885667 6/9/2005 No Scale
BENZIE 548 44.644917 -85.979617 6/9/2005 N0 Scale
BENZIE 549 44.680067 -85.957933 6/9/2005 No Scale
BENZIE 550 44.617150 -85.909067 6/9/2005 No Scale
BENZIE 551 44.617533 -86.046283 6/9/2005 No Scale
BENZIE 552 44.695850 -86.230717 6/9/2005 No Scale
BENZIE 553 44.531017 -86.129967 6/9/2005 Trace
BENZIE 554 44.525450 -85.958850 6/9/2005 Patchy
BENZIE 569 44.547333 —85.818517 6/14/2005 No Scale
BENZIE 573 44.6781 17 -85.841933 6/14/2005 No Scale
BENZIE 575 44.739933 -85.858383 6/14/2005 No Scale
BERRIEN 832 41.839433 -86.625583 8/8/2005 N0 Scale
BERRIEN 834 41.904450 -86.601800 8/8/2005 No Scale
CASS 1 190 41.948700 -85.769467 7/5/2006 No Scale
CHARLEVOIX 614 45.186717 -84.751350 6/20/2005 No Scale
CHARLEVOIX 616 45.146717 -84.935217 6/20/2005 No Scale

 

 

 

 

 

 

 

123

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CHARLEVOIX 623 45.277800 -84.935633 6/20/2005 No Scale
CHARLEVOIX 652 45.349517 -85.l66383 6/21/2005 No Scale
CHARLEVOIX 654 45.310900 -85.178083 6/21/2005 No Scale
CHARLEVOIX 656 45.312250 -85.058583 6/21/2005 No Scale
CHARLEVOIX 658 45.241900 -85.042883 6/21/2005 No Scale
CHARLEVOIX 660 45.237300 -85.093317 6/21/2005 No Scale
CHARLEVOIX 662 45.171817 -85.130967 6/21/2005 No Scale
CHARLEVOIX 664 45.128200 -85.040050 6/22/2005 No Scale
CHARLEVOIX 665 45.215550 -85.201717 6/22/2005 No Scale
CHARLEVOIX 665 45.214933 -85.201867 6/22/2005 No Scale
CHARLEVOIX 667 45.279283 -85.254583 6/22/2005 No Scale
CHARLEVOIX 669 45.255567 -85.32 1 833 6/22/2005 No Scale
CHARLEVOIX 762 45.747867 -85.539133 7/26/2005 Dead/Declining
CHARLEVOIX 764 45.734033 -85.556917 7/26/2005 No Scale
CHARLEVOIX 766 45.725367 -85.564267 7/26/2005 Whitewashed
CHARLEVOIX 768 45.687983 -85.5591 17 7/26/2005 No Scale
CHARLEVOIX 770 45.646733 -85.491433 7/27/2005 Patchy
CHARLEVOIX 772 45.606000 -85.4963 17 7/27/2005 No Scale
CHARLEVOIX 774 45.575517 -85.570633 7/27/2005 No Scale
CHARLEVOIX 776 45.659700 -85.553750 7/27/2005 No Scale
CHARLEVOIX 778 45.659283 -85.579433 7/27/2005 No Scale
CHARLEVOIX 780 45.647350 -85.583083 7/27/2005 No Scale
CHARLEVOIX 782 45.608700 -85.592383 7/27/2005 No Scale
CHARLEVOIX 784 45.584650 -85.596467 7/27/2005 No Scale
CHARLEVOIX 1096 45.273 820 ~84.737970 5/31/2006 No Scale
CHARLEVOIX 1236 45.307667 -85.306633 7/17/2006 No Scale
CHARLEVOIX 1237 45.307367 -85.31 1067 7/17/2006 Whitewashed
CHARLEVOIX 1238 45.646617 -85 .49 l 300 7/18/2006 Whitewashed
CHARLEVOIX 1239 45.614517 -85 .49 l 367 7/18/2006 Whitewashed
CHARLEVOIX 1240 45.606300 -85.496500 7/18/2006 No Scale
CHARLEVOIX 1241 45.575433 -85.570600 7/18/2006 No Scale
CHARLEVOIX 1242 45.584017 -85.596567 7/18/2006 No Scale
CHARLEVOIX 1243 45.608233 -85.592250 7/18/2006 No Scale
CHARLEVOIX 1244 45.647767 -85.582933 7/18/2006 No Scale
CHARLEVOIX 1245 45.659483 -85.579267 7/18/2006 No Scale
CHARLEVOIX 1246 45.687783 -85.558800 7/18/2006 Trace
CHARLEVOIX 1247 45.749783 -85.539033 7/18/2006 Dead/Declining
CHARLEVOIX 1248 45.734183 -85.557067 7/18/2006 Trace
CHARLEVOIX 1249 45.694150 -85.502583 7/19/2006 No Beech
CHARLEVOIX 1250 45.609317 -85.513767 7/19/2006 No Beech
CHARLEVOIX 1251 45.633933 -85.569633 7/19/2006 No Scale
CHARLEVOIX 1252 45.655850 -85.527850 7/19/2006 No Beech
CHEBOYGAN 619 45.205200 -84.591 100 6/20/2005 No Scale
CHEBOYGAN 621 45.244350 -84.666867 6/20/2005 No Scale
CHEBOYGAN 628 45.504983 -84.576183 6/21/2005 No Scale
CHEBOYGAN 630 45.550933 -84.715917 6/21/2005 No Scale
CHEBOYGAN 632 45.574917 -84.631650 6/21/2005 No Scale
CHEBOYGAN 634 45.691633 -84.728233 6/21/2005 No Scale

 

 

 

 

 

 

124

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CHEBOYGAN 638 45.683233 -84.650650 6/21/2005 No Scale
CHEBOYGAN 640 45.664617 -84.638567 6/21/2005 No Scale
CHEBOYGAN 642 45.579550 -84.328167 6/21/2005 No Scale
CHEBOYGAN 644 45.538733 -84.396283 6/21/2005 No Scale
CHEBOYGAN 646 45.569167 -84.454183 6/21/2005 No Scale
CHEBOYGAN 648 45.318967 -84.508033 6/21/2005 No Scale
CHEBOYGAN 650 45.254217 -84.477317 6/21/2005 No Scale
CHEBOYGAN 683 45.435617 -84.470750 6/23/2005 No Scale
CHEBOYGAN 685 45.343850 -84.431900 6/23/2005 No Scale
CHEBOYGAN 691 45.274017 -84.307667 6/23/2005 No Scale
CHEBOYGAN 1273 45.744217 -84.670433 7/27/2006 No Scale
CHIPPEWA 775 46.431417 -85.076000 7/1 1/2005 Whitewashed
CHIPPEWA 777 46.634417 -85.1 15333 7/1 1/2005 Whitewashed
CHIPPEWA 793 46.349600 -85. 144733 7/12/2005 Whitewashed
CHIPPEWA 797 46.261400 -85.012300 7/12/2005 Dead/Declinilg
C HIPPEWA 805 46.462167 -84.668767 7/12/2005 Trace
CHIPPEWA 1 105 46.003570 -84.185810 6/5/2006 No Scale
CHIPPEWA 1 106 45.961810 ~84.021250 6/5/2006 No Scale
CHIPPEWA 1 107 46.089440 -83.724680 6/6/2006 No Beech
CHIPPEWA 1 108 46.072390 -83.618990 6/6/2006 No Beech
CHIPPEWA 1 109 46.031 190 -83.673410 6/6/2006 No Scale
CHIPPEWA 1 1 10 46.013050 -83.703080 6/6/2006 No Scale
CHIPPEWA 1111 45.998700 -83.667490 6/6/2006 N0 Scale
CHIPPEWA 1 112 45.943410 -83.545520 6/6/2006 No Scale
CHIPPEWA 1 1 13 46.002410 -83.536660 6/6/2006 No Scale
CHIPPEWA l 1 14 45.959890 -83.61 1910 6/6/2006 No Scale
CHIPPEWA 1 1 15 45.975230 -83.697280 6/6/2006 No Scale
CHIPPEWA 1 1 16 45.984440 -83.787600 6/6/2006 No Scale
CHIPPEWA 1 1 17 45.994590 -84.072580 6/7/2006 No Scale
CHIPPEWA 1 1 18 46.158980 -84.175060 6/7/2006 N0 Beech
CHIPPEWA 1 119 46.212850 -84.271260 6/7/2006 N0 Beech
CHIPPEWA 1 120 46.375767 -84.353450 6/7/2006 N0 Beech
CHIPPEWA 1 121 46.409633 -84.7391 17 6/7/2006 No Scale
CHIPPEWA 1 122 46.431783 -84.6975 83 6/7/2006 Patch

CHIPPEWA l 123 46.450733 -84.677290 6/7/2006 Trace
CHIPPEWA 1 124 46.403683 -84.93 8933 6/7/2006 Trace
CHIPPEWA 1 125 46.422850 -84.908770 6/7/2006 Patchy
CHIPPEWA 1 126 46.413483 -84.835050 6/7/2006 Trace
CHIPPEWA 1127 46.302783 -84.591733 6/7/2006 No Beech
CHIPPEWA 1 128 46.186783 -84.787083 6/7/2006 No Beech
CHIPPEWA 1186 46.331983 -84.974350 6/29/2006 No Beech
CLARE 697 44.069317 -84.869483 6/27/2005 No Scale
CLARE 699 44.106717 -84.692950 6/27/2005 No Scale
CLARE 1032 43.842717 -85.009867 5/23/2006 No Scale
CLINTON 140 42.812450 -84.389517 5/23/2005 No Beech
CLINTON 1000 42.914417 -84.590567 5/8/2006 No Beech
CRAWFORD 604 44.801367 -84.645800 6/ 16/2005 N0 Scale
CRAWFORD 605 44.680533 -84.646733 6/ 16/2005 N0 Beech

 

 

 

 

 

 

125

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CRAWFORD 607 44.753950 -84.777067 6/16/2005 No Scale
CRAWFORD 612 44.776867 -84.781083 6/16/2005 No Scale
CRAWFORD 617 44.598067 -84.765667 6/16/2005 No Scale
DELTA 714 45.767200 -86.599717 7/6/2005 No Scale
DELTA 716 45.700633 -86.663900 7/6/2005 No Scale
DELTA 718 45.785700 -86.469867 7/6/2005 No Scale
DELTA 749 46.157417 -86.635817 7/6/2005 No Scale
DELTA 751 46.073883 -86.557167 7/6/2005 No Scale
DELTA 753 45.788317 -86.863250 7/6/2005 No Scale
DELTA 755 46.036000 ~86.857100 7/6/2005 No Scale
DELTA 756 45.625767 -87.303833 7/20/2005 No Scale
DELTA 757 46.134650 -86.837600 7/6/2005 No Scale
DELTA 758 45.840917 -87.1 1 1 133 7/20/2005 No Scale
EATON 133 42.577867 -84.758250 5/19/2005 No Scale
EATON 134 42.759433 -84.759833 5/19/2005 No Scale
EMMET 625 45.352433 -84.816567 6/20/2005 No Scale
EMMET 627 45.326267 -84.741800 6/20/2005 No Scale
EMMET 629 45.441717 -84.76661 7 6/21/2005 No Scale
EMMET 631 45.516083 -84.762033 6/21/2005 No Scale
EMMET 633 45.628750 -84.792533 6/21/2005 No Scale
EMMET 635 45.719233 -84.772850 6/21/2005 No Scale
EMMET 637 45.766750 -84.769517 6/21/2005 Whitewashed
EMMET 641 45.650717 -85.014617 6/21/2005 No Scale
EMMET 643 45.606650 -85.085083 6/21/2005 No Scale
EMMET 645 45.551 100 -85.0161 17 6/21/2005 No Scale
EMMET 647 45.550283 -84.935817 6/21/2005 No Scale
EMMET 649 45.613783 -84.9296 1 7 6/21/2005 No Scale
EMMET 651 45.636700 -84.849550 6/21/2005 Trace
EMMET 653 45.551250 -84.845683 6/22/2005 No Scale
EMMET 655 45.471 133 -84.852250 6/22/2005 No Scale
EMMET 657 45.456783 -84.926967 6/22/2005 No Scale
EMMET 659 45.377000 -84.799567 6/22/2005 No Scale
EMMET 661 45.407783 -84.906317 6/22/2005 No Scale
EMMET 804 45.508017 -85.069417 8/4/2005 No Scale
EMMET 806 45.589067 -85.015433 8/4/2005 Trace
EMMET 826 45.458067 -85.066983 8/4/2005 No Scale
EMMET 828 45.680267 -84.892900 8/4/2005 No Scale
EMMET 830 45.655683 -84.774100 8/4/2005 No Scale
EMMET 1097 45.313160 —84.746820 5/31/2006 No Scale
EMMET 1098 45.310333 -84.882050 5/31/2006 No Scale
EMMET 1099 45.492980 -84.876680 5/31/2006 No Scale
EMMET 1 100 45.547090 -84.855870 5/31/2006 No Scale
EMMET 1 164 45.576433 -84.984733 6/22/2006 No Scale
EMMET 1 165 45.575717 -85.057183 6/22/2006 No Scale
EMMET 1 181 45.680667 -84.841267 6/29/2006 No Scale
EMM ET 1 182 45.707900 -84.910067 6/29/2006 Trace
EMMET 1 183 45.716867 -84.868267 6/29/2006 Patchy
EMMET 1274 45.719550 -84.772133 7/27/2006 No Scale

 

 

 

 

 

 

126

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EMMET 1275 45.694467 -84.887000 7/27/2006 Trace
EMMET 1276 45.680150 -84.893200 7/27/2006 No Scale
EMMET 1277 45.608840 -85.082960 7/27/2006 No Scale
EMMET 1278 45.549833 -85.015500 7/27/2006 Trace
GLADWIN 741 44.016500 -84.520550 6/30/2005 No Scale
GLADWIN 1135 43.979067 -84.209850 6/12/2006 No Beech
GRAND

TRAVERSE 563 44.512217 -85.519717 6/14/2005 No Scale
GRAND

TRAVERSE 567 44.546733 -85.677517 6/14/2005 No Scale
GRAND

TRAVERSE 571 44.604517 -85.797917 6/14/2005 No Scale
GRAND

TRAVERSE 577 44.668067 -85.675950 6/14/2005 No Scale
GRAND

TRAVERSE 579 44.635517 -85.599000 6/14/2005 No Scale
GRAND

TRAVERSE 581 44.627783 -85.556700 6/14/2005 No Scale
GRAND

TRAVERSE 583 44.712433 -85.494300 6/14/2005 No Scale
GRAND

TRAVERSE 586 44.763033 -85.403083 6/15/2005 No Scale
GRAND

TRAVERSE 588 44.587800 -85.345683 6/15/2005 No Scale
GRATIOT 1134 43.251550 -84.409233 6/12/2006 No Beech
HILLSDALE 1192 41.833333 -84.473583 7/5/2006 No Beech
lNGHAM 130 42.717733 -84.4771 17 5/19/2005 No Scale
lNGHAM 131 42.689233 -84.51 1217 5/19/2005 No Scale
INGHAM 132 42.609200 -84.591217 5/19/2005 No Scale
lNGHAM 135 42.530750 -84.471433 5/23/2005 No Beech
lNGHAM 136 42.526700 -84.363667 5/23/2005 No Beech
lNGHAM 137 42.596400 -84.279033 5/23/2005 No Beech
lNGHAM 138 42.705367 -84.369233 5/23/2005 No Scale
lNGHAM 139 42.755183 -84.407617 5/23/2005 No Beech
lNGHAM 754 42.577817 -84.253950 7/12/2005 No Scale
IONIA 1 187 42.820433 -84.927533 7/5/2006 N0 Beech
IONIA 1 188 42.936250 -85.129617 7/5/2006 No Scale
IOSCO 1088 44.456300 -83.766350 5/30/2006 No Beech
ISABELLA 1030 43.481633 -84.909283 5/23/2006 No Scale
ISABELLA 1031 43.524583 -85.047200 5/23/2006 No Scale
KALAMAZOO 1 191 42.086083 -85.327517 7/5/2006 No Beech
KALKASKA 589 44.857717 -85. I 89300 6/15/2005 No Scale
KALKASKA 590 44.526667 -85.316750 6/15/2005 No Scale
KALKASKA 592 44.583950 -85.177983 6/15/2005 No Scale
KALKASKA 593 44.839367 -85.050600 6/15/2005 No Scale
KALKASKA 594 44.641033 -85.143433 6/15/2005 No Scale
KALKASKA 595 44.792533 -85.193333 6/15/2005 No Scale
KALKASKA 596 44.654900 —85.196450 6/15/2005 No Scale
KALKASKA S97 44.782900 -85.268300 6/15/2005 No Scale

 

127

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

KALKASKA 598 44.670000 -85.154500 6/15/2005 No Scale
KALKASKA 600 44.698683 -85.073783 6/15/2005 No Scale
KALKASKA 601 44.728217 -85.071367 6/15/2005 No Scale
KALKASKA 602 44.572567 -85.21 1800 6/15/2005 No Scale
KALKASKA 603 44.728600 -85.072733 6/15/2005 No Scale
KALKASKA 609 44.770933 -84.884450 6/16/2005 No Scale
KALKASKA 61 1 44.713383 -84.976050 6/16/2005 No Scale
KALKASKA 613 44.635333 -84.914350 6/16/2005 No Scale
KALKASKA 615 44.539900 -84.955833 6/16/2005 No Scale
KENT 1054 43.252883 -85.697300 5/25/2006 No Scale
KENT 1068 42.785467 -85.408700 5/26/2006 No Beech
KENT 1069 42.948983 -85.31 1967 5/26/2006 No Beech
KENT 1070 43.043567 -85.484900 5/26/2006 No Beech
LAKE 94 43.948033 -85.997367 8/2/2004 No Beech
LAKE 95 43.944833 -85.901433 8/2/2004 N0 Beech
LAKE 96 43.944550 -85.91 1 183 8/2/2004 No Scale
LAKE 1 13 44.147817 -85.987633 8/4/2004 No Beech
LAKE 122 44.081550 -86.014333 8/4/2004 No Scale
LAKE 124 43.918317 -86.033500 8/5/2004 No Scale
LAKE 690 44.132250 -85.594533 6/29/2005 No Scale
LAKE 1024 43.849867 -85.824667 5/10/2006 No Beech
LAKE 1025 43.923033 -85.764050 5/10/2006 No Beech
LAKE 1034 44.140500 -85.870500 5/24/2006 No Beech
LAKE 1036 44.032233 —86.015600 5/24/2006 No Scale
LAKE 1037 43.887683 -85.985167 5/24/2006 No Scale
LAKE 1038 43.823333 -85.970533 5/24/2006 No Beech
LAPEER 1 196 43.166783 -83.378933 7/6/2006 No Scale
LAPEER 1 197 42.942617 -83.346200 7/6/2006 No Scale
LEELANAU 540 44.845950 -86.035900 6/8/2005 No Scale
LEELANAU 541 44.897033 -86.020983 6/8/2005 Trace
LEELANAU 542 44.935383 -85.925250 6/8/2005 No Scale
LEELANAU 543 44.877067 -85.914717 6/8/2005 Trace
LEELANAU 544 44.842800 -85.969817 6/8/2005 No Scale
LEELANAU 546 44.807283 -85.957233 6/9/2005 No Scale
LEELANAU 562 44.810400 -85.855250 6/14/2005 No Scale
LEELANAU 564 44.878800 -85.853750 6/14/2005 No Scale
LEELANAU 566 44.922017 -85.864367 6/14/2005 No Scale
LEELANAU 568 44.891550 -85.744700 6/14/2005 No Scale
LEELANAU 570 44.982217 -85.726950 6/14/2005 No Scale
LEELANAU 572 44.994433 -85.759900 6/14/2005 No Scale
LEELANAU 574 44.980483 -85.776433 6/14/2005 No Scale
LEELANAU 576 44.958967 -85.796383 6/14/2005 No Scale
LEELANAU 578 45.006317 -85.633150 6/14/2005 No Scale
LEELANAU 580 45.102800 -85.643583 6/14/2005 No Scale
LEELANAU 582 44.887300 -85.674500 6/14/2005 No Scale
LEELANAU 584 44.802483 -85.652217 6/14/2005 No Scale
LEELANAU 786 45.003483 -86.134283 8/1/2005 No Scale
LEELANAU 788 45.108233 -85.985950 8/2/2005 No Scale

 

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LEELANAU 790 45.1 19917 -86.053850 8/3/2005 No Scale
LEELANAU 792 45.138150 -86.047450 8/3/2005 N0 Scale
LEELANAU 794 45.142933 -86.019717 8/3/2005 No Scale
LEELANAU 796 45.129400 -86.014417 8/3/2005 No Scale
LEELANAU 798 45.122850 ~85.989567 8/3/2005 No Scale
LEELANAU 800 45.002750 -86.139500 8/1/2005 No Scale
LEELANAU 802 45.005683 -86.1 12983 8/1/2005 No Scale
LEELANAU 808 45.1 16717 -85.980483 8/2/2005 No Scale
LEELANAU 810 45.097950 -86.006183 8/2/2005 No Scale
LEELANAU 812 45.1 1 1017 -86.059217 8/3/2005 N0 Scale
LEELANAU 814 45.098900 -86.047400 8/3/2005 No Scale
LEELANAU 816 45.080900 -86.029600 8/3/2005 No Scale
LEELANAU 818 45.073433 -86.007267 8/3/2005 No Scale
LEELANAU 820 45.080033 -85.989600 8/3/2005 No Scale
LEELANAU 822 45.090917 -85.989800 8/3/2005 No Scale
LEELANAU 824 45.103433 -85.985067 8/3/2005 No Scale
LEELANAU 1 158 45.01 1700 -86.1 13883 6/20/2006 No Scale
LEELANAU 1 159 45.019050 -86.1 13817 6/20/2006 No Scale
LEELANAU 1 160 45.032717 -86.1 12600 6/20/2006 No Scale
LEELANAU 1 161 45.034233 -86.1 12850 6/20/2006 No Beech
LEELANAU 1 162 45.025533 —86.120717 6/20/2006 No Scale
LEELANAU l 163 44.998933 -86.139850 6/21/2006 No Scale
LEELANAU 1232 44.845650 -86.035883 7/17/2006 No Scale
LEELANAU 1233 44.842833 -85.969767 7/17/2006 No Scale
LEELANAU 1234 44.917550 -85.874900 7/17/2006 No Scale
LEELANAU 1235 44.922017 -85.864367 7/17/2006 No Scale
LUCE 241 46.3 10200 -85.694900 8/ 1 7/2004 Patchy
LUCE 726 46.589300 -85.600417 7/12/2005 Dead/Declinin

LUCE 728 46.651683 -85.745033 7/12/2005 Dead/Declining
LUCE 730 46.669700 -85.831650 7/12/2005 Dead/Declining
LUCE 742 46.335917 -85.783100 7/12/2005 Dead/Declining
LUCE 744 46.462317 -85.707250 7/12/2005 Dead/ Declining
LUCE 746 46.450583 -85.801350 7/13/2005 Whitewashed
LUCE 779 46.577883 -85.252400 7/1 1/2005 Dead/Declining
LUCE 781 46.553667 -85.370083 7/1 1/2005 Dead/Declining
LUCE 783 46.665633 —85.307633 7/1 1/2005 Whitewashed
LUCE 785 46.657200 -85.533733 7/1 1/2005 Patchy
LUCE 787 46.453967 -85.597267 7/1 1/2005 Whitewashed
LUCE 789 46.494767 -85.426883 7/12/2005 Dead/ Declining
LUCE 791 46.414383 -85.634167 7/12/2005 Whitewashed
LUCE 795 46.477017 -85 .239400 7/12/2005 Whitewashed
MACKINAC 242 46.194917 -85.8181 17 8/17/2004 No Scale
MACKJNAC 243 46.205767 -85.754100 8/ 17/2004 Patchy
MACKINAC 244 46.203000 -85.697767 8/17/2004 Trace
MAC KINAC 736 46.100767 -85.782933 7/12/2005 Whitewashed
MACKINAC 738 46.035550 -85.696767 7/12/2005 No Scale
MAC KINAC 740 46.222383 -85 .572000 7/12/2005 Whitewashed
MAC KINAC 799 46.174450 -85.1 84867 7/ 12/2005 Whitewashed

 

 

 

 

129

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MACKINAC 801 46.1 10483 -85.4465 50 7/12/2005 Whitewashed
MACKIN AC 803 46.0393 17 -85.1 12283 7/12/2005 Patchy
MACKINAC 807 45.988333 84.925633 7/14/2005 Dead/Declining
MACKINAC 809 45.892833 -84.806950 7/14/2005 No Scale
MACKINAC 81 1 45.966483 -84.76l417 7/14/2005 No Scale
MAC KINAC 1 101 45.928920 -84.913390 6/5/2006 Whitewashed
MACKINAC 1 102 45.928610 -84.912800 6/5/2006 No Scale
MACKINAC 1 103 45.961900 -84.898250 6/5/2006 Patchy
MACKINAC 1 104 46.085960 -84.369680 6/5/2006 No Scale
MACKINAC 1 129 46.063067 -85.146600 6/8/2006 Dead/Declining
MACKINAC 1 130 46.065950 -85.026667 6/8/2006 Trace
MACKINAC 1 131 46.028433 -84.918850 6/8/2006 No Scale
MACKINAC 1 132 46.102300 -84.879950 6/8/2006 No Scale
MACKINAC 1 133 46.073600 -84.765533 6/8/2006 No Scale
MACKINAC 1 136 45.855933 -84.605317 6/13/2006 No Beech
MACKINAC 1 137 45.877600 -84.624067 6/13/2006 No Beech
MACKINAC 1 138 45.879500 -84.629000 6/13/2006 No Scale
MACKINAC 1 139 45.871 100 84624183 6/13/2006 Whitewashed
MACKINAC 1 140 45.874000 -84.635467 6/13/2006 No Scale
MACKINAC 1 141 45.872917 -84.643650 6/13/2006 Trace
MACKINAC 1 142 45.871467 -84.645267 6/13/2006 Patchy
MACKINAC 1 143 45.870767 -84.646217 6/13/2006 Whitewashed
MACKINAC 1 144 45.865550 -84.645033 6/13/2006 Whitewashed
MACKINAC 1 145 45.858633 -84.637617 6/13/2006 No Scale
MACKINAC 1146 45.862783 -84.633467 6/13/2006 No Scale
MACKINAC 1 147 45.861783 -84.625417 6/13/2006 Patchy
MACKINAC 1 148 45.857733 -84.609950 6/13/2006 Trace
MACKINAC 1 149 45.809000 -84.571017 6/14/2006 Dead/Declining
MAC KINAC 1 150 45.794467 -84.536400 6/14/2006 Trace
MACKINAC 1 151 45.790550 -84.520783 6/14/2006 Patch

MACKINAC 1 152 45.770733 -84.513100 6/14/2006 Trace
MACKINAC 1 153 45.752350 -84.4931 17 6/14/2006 No Scale
MACKINAC 1 154 45.779017 -84.384667 6/14/2006 No Beech
MACKINAC 1 155 45.743217 -84.385900 6/14/2006 No Beech
MACKINAC 1 156 45.761883 -84.424383 6/14/2006 No Scale
MACKINAC 1 157 45.772750 -84.451883 6/14/2006 No Scale
MACKINAC 1 166 46.081817 -85.594483 6/26/2006 Patchy
MACKINAC 1 167 45.983700 -85.7001 17 6/26/2006 No Scale
MACKINAC 1 171 46.177167 -85.795233 6/27/2006 Whitewashed
MACKINAC 1 180 45.945883 -84.859750 6/29/2006 Whitewashed
MACKINAC 1270 46.035033 -85.696800 7/26/2006 Patchy
MAC KINAC 1271 45.987900 -85 .749233 7/26/2006 Patchy
MACKINAC 1272 45.966717 -84.762600 7/27/2006 No Scale
MACKINAC 1279 45.892750 -84.806767 7/24/2006 Patchy
MANISTEE 67 44.184217 -86.1 17000 7/27/2004 No Beech
MANISTEE 68 44.172933 -86.207867 7/27/2004 No Beech
MANISTEE 69 44.175000 -86.256517 7/27/2004 No Beech
MANISTEE 70 44.284333 -86.3 10600 7/28/2004 Whitewashed

 

 

 

 

 

 

130

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MANISTEE 71 44.274133 -86.202717 7/28/2004 No Scale
MANISTEE 72 44.268733 -86.077650 7/28/2004 No Scale
MANISTEE 73 44.270800 -85.944433 7/28/2004 No Beech
MANISTEE 74 44.321583 -85.961000 7/27/2004 No Beech
MANISTEE 75 44.314033 -86.088500 7/27/2004 No Scale
MANISTEE 98 44.316733 -86.196017 8/3/2004 Trace
MANISTEE 99 44.373633 -86.184883 8/3/2004 Trace
MANISTEE 100 44.371650 -86.071833 8/3/2004 No Beech
MANISTEE 101 44.375917 -85.975800 8/3/2004 No Beech
MANISTEE 102 44.367667 -85.933467 8/3/2004 No Scale
MANISTEE 103 44.373450 -85.831067 8/3/2004 No Scale
MANISTEE 104 44.419750 -85.859383 8/3/2004 No Scale
MANISTEE 105 44.426150 -85.919717 8/3/2004 No Beech
MANISTEE 106 44.426550 -85.927500 8/3/2004 No Beech
MANISTEE 107 44.483133 -85.999600 8/3/2004 No Scale
MANISTEE 108 44.493500 -85.943650 8/3/2004 N0 Scale
MANISTEE 109 44.4841 17 -86.038750 8/3/2004 N0 Scale
MANISTEE 1 10 44.487217 -86.092900 8/3/2004 N0 Scale
MANISTEE 1 1 1 44.476050 -86.169917 8/3/2004 N0 Scale
MANISTEE 1 12 44.432250 -86.181900 8/3/2004 No Scale
MANISTEE 1 14 44.168917 -85.905950 8/4/2004 No Beech
MANISTEE 1 16 44.252650 -85.860400 8/4/2004 No Beech
MANISTEE 1 l8 44.345633 -85.839467 8/4/2004 No Scale
MANISTEE 1 19 44.223500 -85.910183 8/4/2004 No Beech
MANISTEE 120 44.228900 -86.013300 8/4/2004 N0 Beech
MANISTEE 121 44.222450 -86.216883 8/4/2004 No Beech
MANISTEE 123 44.269733 -86.120217 8/4/2004 No Scale
MANISTEE 182 44.347317 -85.894467 6/1/2005 No Scale
MANISTEE 183 44.438733 -85.997333 6/1/2005 N0 Scale
MANISTEE 184 44.479150 -86.243383 6/1/2005 N0 Scale
MANISTEE 185 44.430600 -86.230367 6/1/2005 N0 Scale
MANISTEE 186 44.403 700 -86.225900 6/ 1/2005 Trace
MANISTEE 187 44.400950 -86.227433 6/1/2005 Whitewashed
MANISTEE 188 44.388983 -86. 196467 6/1/2005 Patchy
MANISTEE 189 44.388550 -86.164800 6/1/2005 Trace
MANISTEE 190 44.403133 -86.167300 6/1/2005 No Scale
MANISTEE 191 44.406400 -86.194417 6/1/2005 Trace
MANISTEE 192 44.358933 -86.1451 17 6/1/2005 No Scale
MANISTEE 193 44.387633 -86.125017 6/1/2005 No Scale
MANISTEE 194 44.371417 -86.122367 6/1/2005 No Scale
MANISTEE 195 44.329767 -86. 147667 6/1/2005 No Scale
MANISTEE 196 44.3091 17 -86.172700 6/1/2005 Whitewashed
MANISTEE 197 44.315767 -86. 167733 6/2/2005 Patchy
MANISTEE 198 44.340083 -86.162250 6/2/2005 No Scale
MANISTEE 199 44.381283 -86.155850 6/2/2005 N0 Scale
MANISTEE 500 44.358517 -86.171950 6/2/2005 N0 Scale
MANISTEE 501 44.349650 -86.202467 6/2/2005 No Scale
MANISTEE 502 44.331483 -86.197933 6/2/2005 Patchy

 

131

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MANISTEE 503 44.331267 -86.187900 6/2/2005 No Scale
MANISTEE 504 44.3 16833 -86. 1921 17 6/2/2005 Whitewashed
MANISTEE 505 44.262683 -86.181617 6/2/2005 No Scale
MANISTEE 506 44.267067 -86.1773 50 6/2/2005 Trace
MANISTEE 530 44.250883 -86.200683 6/7/2005 Patchy
MANISTEE 531 44.170733 -86.103333 6/7/2005 No Scale
MANISTEE 532 44.312883 -86.141 133 6/8/2005 Trace
MANISTEE 533 44.304967 -86.102350 6/8/2005 Trace
MANISTEE 534 44.344967 -86.268333 6/8/2005 Patchy
MANISTEE 847 44.396233 -86.221350 6/1/2005 N0 Scale
MANISTEE 1008 44.194767 -86.037100 5/9/2006 N0 Scale
MANISTEE 1009 44.267290 -86. 1 783 00 5/9/2006 Whitewashed
MANISTEE 1226 44.284070 ~85.863730 7/13/2006 No Scale
MASON 61 43.916650 -86.442783 7/26/2004 Whitewashed
MASON 62 44.1 12050 -86.419450 7/27/2004 Patchy
MASON 63 44.131367 -86.333800 7/27/2004 No Beech
MASON 64 44.126567 -86.287650 7/27/2004 Patchy
MASON 65 44.146783 -86.2005 00 7/2 7/2004 Trace
MASON 66 44.144017 -86.102983 7/27/2004 No Beech
MASON 84 43.833017 -86.428300 8/1/2004 No Beech
MASON 85 43.871850 -86.43 6967 8/1/2004 Patchy
MASON 86 43.900933 -86.398483 8/ 1/2004 Whitewashed
MASON 88 43.843333 -86.399150 8/2/2004 Trace
MASON 89 43.847017 -86.319683 8/2/2004 Trace
MASON 90 43.825133 -86.259333 8/2/2004 No Beech
MASON 91 43.959383 -86.339200 8/2/2004 Patchy
MASON 92 43.955200 -86.207050 8/2/2004 Trace
MASON 93 43.945950 -86.077133 8/2/2004 N0 Beech
MASON 97 44.018867 -86.145083 8/2/2004 N0 Beech
MASON 127 44.1 16900 -86.374267 7/27/2004 No Beech
MASON 158 44.040050 -86.4963 1 7 5/26/2005 Whitewashed
MASON 159 43.993067 -86.463183 5/26/2005 Whitewashed
MASON 160 43.971350 -86.4583 50 5/26/2005 Whitewashed
MASON 161 43.944700 -86.398467 5/26/2005 'No Beech
MASON 162 43.890650 -86.284883 5/26/2005 No Beech
MASON 163 43.876533 -86.331667 5/26/2005 Whitewashed
MASON 164 43.876667 -86.368283 5/26/2005 Patchy
MASON 165 43.868800 -86.399983 5/26/2005 Whitewashed
MASON 171 43.848150 -86.077333 5/31/2005 No Beech
MASON 172 43.871900 -86.102500 5/31/2005 No Scale
MASON 173 43.874083 -86.1 10733 5/31/2005 No Scale
MASON 174 43.873750 -86.135050 5/31/2005 No Scale
MASON 175 43.874683 —86.190933 5/31/2005 No Scale
MASON 176 43.875383 -86.231483 5/31/2005 Patchy
MASON 177 43.874333 -86.227717 5/31/2005 Patchy
MASON 178 43.889867 -86.229783 5/31/2005 No Scale
MASON 179 43.890050 -86.227817 5/31/2005 Patchy
MASON 180 43.904550 -86.218483 5/31/2005 Patchy

 

132

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MASON 181 43.975367 ~86.103317 5/31/2005 Patchy
MASON 511 43.872067 ~86.248200 6/6/2005 No Scale
MASON 512 43.838850 -86.201450 6/6/2005 No Scale
MASON 513 43.835733 ~86.175550 6/6/2005 No Scale
MASON 5 14 43 .83 7000 ~86.]61683 6/6/2005 Whitewashed
MASON 517 43.911333 ~86.153117 6/7/2005 No Beech
MASON 518 43.914067 ~86. 163433 6/7/2005 Trace
MASON 5 19 43.903767 ~86.190683 6/7/2005 Trace
MASON 520 43 .9021 17 ~86.175050 6/7/2005 Trace
MASON 521 43.937650 -86.050033 6/7/2005 No Scale
MASON 522 43.984800 ~86.054617 6/7/2005 No Scale
MASON 523 44.017817 ~86.080133 6/7/2005 Trace
MASON 524 44.016317 ~86.] 15900 6/7/2005 Trace
MASON 525 44.032250 ~86.077217 6/7/2005 No Scale
MASON 526 44.041767 ~86. 120400 6/7/2005 Wh itewashed
MASON 527 44.066317 ~86.132400 6/7/2005 Trace
MASON 528 44.104933 -86.1253 83 6/7/2005 Trace
MASON 529 44.145717 -86.221 133 6/7/2005 Trace
MASON 1010 44.1 1 1760 ~86.41 8540 5/9/2006 Whitewashed
MASON 101 l 44.085500 ~86.435460 5/9/2006 Dead/Declinirl

MASON 1012 44.085240 ~86.3 87150 5/9/2006 Whitewashed
MASON 1013 44.092120 ~86.36683O 5/9/2006 Wh itewashed
MASON 1 014 44.042450 ~86.496280 5/ 10/2006 Wh itewashed
MASON 1015 44.044730 ~86.49657O 5/10/2006 Dead/ Declining
MASON 1016 44.037810 ~86.504890 5/10/2006 Dead/Declining
MASON 1035 44.080400 ~86.081233 5/24/2006 No Scale
MASON 121 1 43.825300 ~86.152590 7/10/2006 Dead/Declining
MASON 1229 43.873733 ~86.] 1 1933 7/13/2006 No Scale
MASON 1230 43.835917 ~86.175417 7/13/2006 Trace
MASON 1231 43 .83 8883 ~86.201517 7/13/2006 Trace
MECOSTA 1003 43.604767 -85.443333 5/8/2006 No Scale
MECOSTA 1004 43.703533 ~85.200650 5/8/2006 No Scale
MECOSTA 1049 43.558017 ~85.522833 5/25/2006 No Scale
MECOSTA 1050 43.515617 ~85.402700 5/25/2006 No Scale
MENOMINEE 759 45.771767 -87.366917 7/6/2005 No Scale
MENOMINEE 1168 45.433300 -87.375550 6/27/2006 No Scale
MENOMINEE 1169 45.515317 ~87.756817 6/27/2006 No Beech
MIDLAND 1193 43.655217 ~84.586717 7/6/2006 No Beech
MISSAUKEE 693 44.463367 -84.972467 6/27/2005 No Scale
MISSAUKEE 695 44.305817 ~84.994050 6/27/2005 No Scale
MISSAUKEE 707 44.454200 ~85.240100 6/28/2005 No Scale
MISSAUKEE 709 44.359867 ~85.31 1217 6/28/2005 No Scale
MISSAUKEE 713 44.294983 -85.135017 6/28/2005 No Scale
MISSAUKEE 715 44.445000 -85.074883 6/28/2005 No Scale
MISSAUKEE 717 44.451800 ~85.135450 6/28/2005 No Scale
MISSAUKEE 719 44.367417 ~85.203433 6/28/2005 No Scale
MISSAUKEE 721 44.307633 ~85.034950 6/28/2005 No Scale
MISSAUKEE 1 184 44.291733 -84.890500 6/30/2006 No Beech

 

133

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MISSAUKEE 1 185 44.205033 ~84.872183 6/30/2006 No Beech
MONROE 1201 41.878633 -83.695767 7/7/2006 N0 Beech
MONTCALM 1001 43.290167 ~85.013983 5/8/2006 N0 Scale
MONTCALM 1002 43.427467 -85.084433 5/8/2006 No Scale
MONTCALM 1028 43.408017 ~85.013883 5/23/2006 No Beech
MONTCALM 1029 43.460652 -84.886467 5/23/2006 No Scale
MONTCALM 1051 43.432283 ~85.343550 5/25/2006 No Scale
MONTCALM 1053 43.329667 ~85.535333 5/25/2006 No Scale
MONTCALM 1071 43.127500 ~85.283167 5/26/2006 No Beech
MONTCALM 1072 43.285417 ~85.255750 5/26/2006 No Scale
MONTMORENCY 559 44.960200 -84.371333 6/13/2005 No Scale
MONTMORENCY 561 44.868183 ~84.198717 6/13/2005 No Scale
MONTMORENCY 1077 45.135400 ~84.121067 5/29/2006 No Scale
MONTMORENCY 1095 44.957200 ~84.006267 5/30/2006 No Beech
MUSKEGON 141 43.131483 ~86.266750 5/24/2005 No Scale
MUSKEGON 142 43.263333 ~86.358650 5/24/2005 No Scale
MUSKEGON 143 43.344733 ~86.396050 5/24/2005 No Beech
MUSKEGON 144 43.454600 ~86.266183 5/25/2005 No Scale
MUSKEGON 1056 43.431750 ~86.099383 5/25/2006 No Scale
MUSKEGON 1057 43.337150 ~86.082250 5/25/2006 No Scale
MUSKEGON 1058 43.364750 ~86.189017 5/25/2006 No Scale
MUSKEGON 1059 43.139983 -86.087383 5/25/2006 No Scale
NEWAYGO 1023 43.779267 ~85.91 1217 5/10/2006 No Beech
NEWAYGO 1026 43.815267 ~85.643100 5/10/2006 No Scale
NEWAYGO 1027 43.663767 ~85.627483 5/10/2006 No Scale
NEWAYGO 1039 43.742100 ~86.018817 5/24/2006 No Scale
NEWAYGO 1044 43.588333 ~85.900033 5/24/2006 No Scale
NEWAYGO 1046 43.699150 ~85.812767 5/25/2006 No Beech
NEWAYGO 1047 43.610650 -85.761 1 17 5/25/2006 No Scale
NEWAYGO 1048 43.558800 ~85.662067 5/25/2006 No Scale
NEWAYGO 1052 43.458867 ~85.582133 5/25/2006 No Beech
NEWAYGO 1055 43.480200 -85.788150 5/25/2006 No Scale
OAKLAND 1 198 42.794333 -83.509850 7/6/2006 No Scale
OAKLAND 1 199 42.644650 ~83.549717 7/6/2006 No Scale
OCEANA 1 43.617950 ~86.499583 7/16/2004 No Beech
OCEANA 2 43.617350 ~86.479050 7/16/2004 No Beech
OCEANA 3 43.621 150 ~86.439100 7/16/2004 No Beech
OCEANA 4 43.667700 ~86.488650 7/17/2004 Trace
OCEANA 5 43.631800 ~86.488050 7/17/2004 No Beech
OCEANA 6 43.632000 ~86.473583 7/17/2004 Patchy
OCEANA 7 43.6321 17 ~86.457383 7/17/2004 No Beech
OCEANA 8 43.624500 ~86.517933 7/17/2004 No Beech
OCEANA 9 43.624717 ~86.509600 7/1 7/2004 Trace
OCEANA 10 43.623717 ~86.498800 7/17/2004 No Beech
OCEANA 1 1 43.638967 -86.51 1833 7/17/2004 No Beech
OCEANA 12 43.639167 -86.497017 7/17/2004 Trace
OCEANA 13 43.638933 ~86.467983 7/17/2004 No Beech
OCEANA 14 43.639100 ~86.459150 7/17/2004 No Scale

 

 

 

 

 

 

134

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OCEANA 15 43.642550 ~86.460000 7/17/2004 No Beech
OCEANA 16 43.633850 -86.534333 7/18/2004 Whitewashed
OCEANA l7 43.667750 -86.458217 7/18/2004 Trace
OCEANA 18 43.660767 ~86.451900 7/18/2004 No Scale
OCEANA 19 43 .661 150 ~86.438517 7/18/2004 Trace
OCEANA 20 43.663533 ~86.432183 7/18/2004 No Beech
OCEANA 21 43.656633 -86.480500 7/19/2004 No Beech
OCEANA 22 43.674733 ~86.442733 7/19/2004 No Beech
OCEANA 23 43.672100 ~86.418267 7/19/2004 No Beech
OCEANA 24 43.646267 -86.427267 7/19/2004 No Beech
OCEANA 25 43.616983 ~86.408317 7/19/2004 No Scale
OCEANA 26 43.617167 ~86.438150 7/19/2004 No Beech
OCEANA 27 43.610050 ~86.475133 7/19/2004 No Beech
OCEANA 28 43.607233 ~86.518400 7/19/2004 Trace
OCEANA 29 43.681883 -86.444467 7/19/2004 No Beech
OCEANA 30 43.651600 ~86.398100 7/19/2004 No Scale
OCEANA 31 43.645283 -86.382617 7/22/2004 No Scale
OCEANA 32 43.633567 -86.405783 7/22/2004 No Beech
OCEANA 33 43.630300 ~86.451217 7/22/2004 Patchy
OCEANA 34 43.620283 ~86.468450 7/22/2004 No Scale
OCEANA 35 43.588167 ~86.513533 7/22/2004 Patchy
OCEANA 36 43.569250 -86.508483 7/22/2004 Patchy
OCEANA 37 43.541350 ~86.491967 7/22/2004 No Scale
OCEANA 38 43.551967 ~86.478417 7/22/2004 Trace
OCEANA 39 43.556567 -86.416717 7/23/2004 No Scale
OCEANA 40 43.544783 -86.415767 7/23/2004 No Scale
OCEANA 41 43.527950 ~86.418033 7/23/2004 No Beech
OCEANA 42 43.517700 ~86.440167 7/23/2004 No Scale
OCEANA 43 43.531200 ~86.461733 7/23/2004 No Scale
OCEANA 44 43.544950 -86.452333 7/23/2004 Trace
OCEANA 45 43.558883 -86.447400 7/23/2004 No Beech
OCEANA 46 43.566267 ~86.4228OO 7/23/2004 Patchy
OCEANA 47 43.569700 ~86.417850 7/23/2004 No Beech
OCEANA 48 43.591400 ~86.406750 7/23/2004 Trace
OCEANA 49 43.582950 -86.377717 7/23/2004 No Scale
OCEANA 50 43.563400 ~86.387250 7/25/2004 Trace
OCEANA 51 43.561633 ~86.365567 7/25/2004 Trace
OCEANA 52 43.576867 -86.300683 7/25/2004 No Beech
OCEANA 53 43.675783 ~86.457717 7/26/2004 Trace
OCEANA 54 43.714583 ~86.474833 7/26/2004 Trace
OCEANA 55 43.756367 ~86.4341 17 7/26/2004 No Beech
OCEANA 56 43.740683 -86.430283 7/26/2004 Trace
OCEANA 57 43.726500 ~86.409233 7/26/2004 No Scale
OCEANA 58 43.775933 -86.41 7967 7/26/2004 Trace
OCEANA 59 43.761000 ~86.346867 7/26/2004 No Beech
OCEANA 60 43.768517 ~86.374533 7/26/2004 No Scale
OCEANA 76 43.510150 ~86.357050 8/1/2004 No Scale
OCEANA 77 43.556383 ~86.325133 8/1/2004 No Scale

 

135

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OCEANA 78 43.582533 -86.338017 8/1/2004 No Beech
OCEANA 79 43.609783 -86.338317 8/1/2004 No Scale
OCEANA 80 43.659217 -86.326200 8/1/2004 No Scale
OCEANA 81 43.702417 ~86.3373 50 8/1/2004 N0 Beech
OCEANA 82 43.698367 ~86.416683 8/1/2004 No Beech
OCEANA 83 43.797900 ~86.338183 8/1/2004 No Scale
OCEANA 87 43.802183 -86.393483 8/2/2004 No Beech
OCEANA 126 43.737517 ~86.186550 8/5/2004 ' No Scale
OCEANA 128 43.679733 ~86.412667 5/17/2005 No Scale
OCEANA 129 43.659083 ~86.369417 5/17/2005 No Scale
OCEANA 145 43.485400 -86.359483 5/25/2005 No Beech
OCEANA 146 43.530633 ~86.439467 5/25/2005 No Beech
OCEANA 147 43.534833 ~86.4377l7 5/25/2005 No Scale
OCEANA 148 43.543750 -86.358867 5/25/2005 No Scale
OCEANA 149 43.614967 ~86.431067 5/25/2005 No Scale
OCEANA 150 43.614433 -86.468333 5/25/2005 No Scale
OCEANA 15 l 43.660750 ~86.497083 5/25/2005 Trace
OCEANA 152 43.648967 ~86.498067 5/25/2005 No Scale
OCEANA 153 43.664067 ~86.482983 5/25/2005 Patchy
OCEANA 1 54 43.667633 ~86.468100 5/25/2005 Whitewashed
OCEANA 155 43.696300 ~86.454400 5/25/2005 No Scale
OCEANA 156 43.691500 -86.487450 5/25/2005 No Scale
OCEANA 157 43.733500 -86.471333 5/25/2005 No Beech
OCEANA 166 43.697233 ~86.392767 5/26/2005 Patchy
OCEANA 167 43.495750 ~86.371500 5/26/2005 No Scale
OCEANA 168 43.758817 ~86.214267 5/31/2005 No Scale
OCEANA 169 43.772133 —86.128567 5/31/2005 No Scale
OCEANA 170 43.802700 -86.087133 5/31/2005 No Scale
OC EANA 507 43.816483 -86.358533 6/6/2005 Trace
OCEANA 508 43.798433 ~86.298167 6/6/2005 N0 Beech
OCEANA 509 43.787700 -86.292217 6/6/2005 Trace
OCEANA 510 43.817533 ~86.268883 6/6/2005 No Scale
OCEANA 515 43.790667 ~86.174950 6/6/2005 No Scale
OC EANA 516 43.793433 ~86.244883 6/6/2005 Trace
OCEANA 1017 43.648180 -86.519170 5/10/2006 Patchy
OCEANA 1018 43.630133 ~86.331450 5/10/2006 Trace
OCEANA 1019 43.614733 -86.278200 5/10/2006 Patch

OCEANA 1020 43.599817 ~86.266383 5/10/2006 No Scale
OCEANA 1021 43.620633 ~86.178800 5/10/2006 Whitewashed
OCEANA 1022 43.672817 ~86.139283 5/10/2006 No Scale
OCEANA 1040 43.731767 ~86.167150 5/24/2006 No Scale
OCEANA 1041 43.625450 ~86.126317 5/24/2006 No Scale
OCEANA 1042 43.587717 -86.153000 5/24/2006 No Scale
OCEANA 1043 43.532200 ~86.1 13550 5/24/2006 No Scale
OCEANA 1202 43.545440 -86.4523 20 7/10/2006 Whitewashed
OCEANA 1203 43 .543 840 ~86.3 58630 7/10/2006 Trace
OCEANA 1204 43.509870 -86.357320 7/10/2006 No Scale
OCEANA 1205 43.614470 ~86.4683 80 7/10/2006 No Scale

 

136

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

OCEANA 1206 43.648730 ~86.497970 7/10/2006 Trace
OCEANA 1207 43.768710 ~86.3 74470 7/10/2006 Trace
OCEANA 1208 43.797430 ~86.33 8240 7/10/2006 No Scale
OCEANA 1209 43.790370 ~86.175200 7/10/2006 No Scale
OCEANA 1210 43.790920 ~86.178850 7/10/2006 Trace
OCEANA 839 43.617150 ~86.408417 8/8/2005 N0 Scale
OCEANA 840 43.645667 -86.382683 8/8/2005 No Scale
OCEANA 845 43.541583 ~86.491550 8/8/2005 No Scale
OCEANA 844 43.548033 ~86.427650 8/8/2005 No Scale
OCEANA 846 43.527300 -86.462733 8/8/2005 N0 Scale
OCEANA 843 43.584617 ~86.3 77933 8/8/2005 No Scale
OCEANA 842 43.507517 ~86.353517 8/9/2005 No Scale
OCEANA 841 43.797483 ~86.338200 8/9/2005 No Scale
OGEMAW 704 44.445650 ~84.146967 6/30/2005 No Scale
OGEMAW 706 44.272183 -84.307200 6/30/2005 No Scale
OGEMAW 1089 44.470100 ~83.944933 5/30/2006 No Scale.
OSCEOLA 676 44.148733 ~85.324500 6/28/2005 No Scale
OSCEOLA 678 44.073417 ~85.382483 6/28/2005 No Scale
OSCEOLA 680 44.110100 ~85.187083 6/28/2005 No Scale
OSCEOLA 692 44.055817 ~85.503083 6/29/2005 No Scale
OSCEOLA 1005 43.970617 ~85.266l83 5/8/2006 No Scale
OSCEOLA 1033 43.894750 ~85.433833 5/23/2006 No Scale
OSCODA 555 44.536217 ~84.356017 6/13/2005 No Beech
OSCODA 1090 44.706967 ~84.217017 5/30/2006 No Scale
OSCODA 1091 44.616033 ~84.129600 5/30/2006 No Scale
OSCODA 1092 44.632317 ~83.941 100 5/30/2006 No Beech
OSCODA 1094 44.754033 ~84.060633 5/30/2006 No Scale
OTSEGO 556 44.939133 ~84.599283 6/13/2005 No Scale
OTSEGO 557 44.971200 -84.447033 6/13/2005 No Scale
OTSEGO 558 45.015750 ~84.590200 6/13/2005 No Scale
OTSEGO 560 44.880183 -84.677950 6/13/2005 No Scale
OTSEGO 606 44.958217 ~84.706717 6/16/2005 No Scale
OTSEGO 608 44.921 117 ~84.776683 6/16/2005 No Scale
OTSEGO 610 44.888017 -84.787417 6/16/2005 No Scale
OTSEGO 620 45.1 13000 ~84.816450 6/20/2005 No Scale
OTSEGO 622 45.027600 ~84.755550 6/20/2005 No Scale
OTSEGO 624 45.099617 ~84.699017 6/20/2005 No Scale
OTSEGO 626 45.061767 ~84.623483 6/20/2005 No Scale
OTSEGO 670 45.159083 ~84.512067 6/22/2005 No Scale
OTSEGO 672 45.156167 ~84.417450 6/23/2005 No Scale
OTTAWA 1045 43.1691 17 ~85.890017 5/25/2006 No Beech
OTTAWA 1060 43.001900 ~86.184883 5/25/2006 No Scale
OTTAWA 1061 42.935450 ~86.126383 5/25/2006 No Scale
OTTAWA 1062 42.875150 ~86.178867 5/25/2006 No Scale
PRESQUE ISLE 687 45.435767 ~84.224883 6/23/2005 No Scale
PRESQUE ISLE 689 45.356683 ~84.152450 6/23/2005 No Scale
PRESQUE ISLE 1073 45.396183 ~84.055283 5/29/2006 No Beech
PRESQUE ISLE 1074 45.4131 17 -83.886333 5/29/2006 No Scale

 

 

 

 

 

 

137

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PRESQUE ISLE 1075 45.222533 -83.621933 5/29/2006 No Beech
PRESQUE ISLE 1076 45.260683 ~83.483500 5/29/2006 No Scale
SCHOOLCRAFT 216 46.289850 -86.570883 8/12/2004 No Beech
SCHOOLCRAFT 217 46.279617 ~86.542617 8/13/2004 No Scale
SCHOOLCRAFT 218 46.264900 -86.490883 8/13/2004 No Scale
SCHOOLCRAFT 220 46.346650 ~86.353550 8/13/2004 No Beech
SCHOOLCRAFT 221 46.348167 ~86.278450 8/15/2004 No Beech
SCHOOLCRAF T 222 46.502517 -86.271767 8/15/2004 Trace
SCHOOLCRAFT 223 46.428200 ~86.361800 8/15/2004 No Scale
SCHOOLCRAFT 224 46.371683 ~86.291417 8/15/2004 No Scale
SCHOOLCRAFT 225 46.345400 ~86.104800 8/15/2004 No Beech
SCHOOLCRAFT 226 46.269600 ~85.928233 8/15/2004 No Beech
SCHOOLCRAFT 228 46.277667 ~86.259500 8/16/2004 No Scale
SCHOOLCRAFT 229 46.220583 ~86.229933 8/16/2004 No Scale
SCHOOLCRAFT 230 46.154450 ~86.198950 8/16/2004 No Beech
SCHOOLCRAFT 231 46.125150 ~86.227983 8/16/2004 No Scale
SCHOOLCRAFT 232 46.070900 ~86.262150 8/16/2004 No Beech
SCHOOLCRAFT 233 46.000583 -86.272083 8/16/2004 No Beech
SCHOOLCRAFT 234 45.979050 ~86.188950 8/16/2004 No Scale
SCHOOLCRAFT 235 46.021483 -86.1 16183 8/16/2004 No Beech
SCHOOLCRAFT 236 46.043067 -86.082600 8/16/2004 No Scale
SCHOOLCRAFT 237 46.099800 ~86.026350 8/16/2004 No Beech
SCHOOLCRAFT 238 46.160500 ~86.001083 8/16/2004 No Beech
SCHOOLCRAFT 239 46.216017 -85.970667 8/17/2004 No Beech
SCHOOLCRAFT 240 46.241750 ~85 .9441 17 8/17/2004 Trace
SCHOOLCRAFT 245 46.185783 ~85.928283 8/17/2004 No Scale
SCHOOLCRAFT 246 46.159350 -85.928250 8/17/2004 Trace
SCHOOLCRAFT 708 46.291750 ~86.447750 7/5/2005 No Scale
SCHOOLCRAFT 710 46.068083 ~86.468200 7/6/2005 No Scale
SCHOOLCRAFT 712 45.967133 ~86.364000 7/6/2005 N0 Scale
SCHOOLCRAFT 720 45.840167 ~86.368867 7/6/2005 No Scale
SCHOOLCRAFT 722 45.981267 ~86.136433 7/6/2005 No Scale
SCHOOLCRAFT 724 46.071750 ~86.058467 7/6/2005 N0 Scale
SCHOOLC RAF T 734 46.164983 ~85 .927850 7/12/2005 Patchy
SCHOOLCRAF T 771 46.4591 17 ~86.170917 7/7/2005 Trace
SCHOOLCRAFT 773 46.419367 -86.157517 7/7/2005 No Scale
SCHOOLCRAFT 1 175 46.428200 ~86.361800 6/28/2006 No Scale
SCHOOLCRAF T 1 176 46.502583 -86.271917 6/28/2006 Patchy
SCHOOLCRAFT l 177 46.502150 -86.272000 6/28/2006 Whitewashed
SCHOOLCRAFT 1 178 46.461 100 ~86.261550 6/28/2006 Whitewashed
SCHOOLCRAFT 1 179 46.426650 ~86.083033 6/28/2006 No Beech
SCHOOLCRAFT 1260 46.413700 ~86. 146800 7/25/2006 Whitewashed
SCHOOLCRAFT 1261 46.419483 ~86.157633 7/25/2006 Dead/ Declining
SCHOOLCRAFT 1262 46.371600 ~86.291317 7/26/2006 No Scale
SCHOOLCRAF T 1263 46.371817 ~86.291400 7/26/2006 Trace
SCHOOLCRAFT 1264 46.277500 ~86.260050 7/26/2006 No Scale
SCHOOLCRAFT 1265 46.220633 ~86.229783 7/26/2006 No Scale
SCHOOLCRAFT 1266 46.125200 ~86.227967 7/26/2006 No Scale

 

138

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SCHOOLCRAFT 1267 45.966150 ~86.363650 7/26/2006 No Scale
SCHOOLCRAF T 1268 45.981917 -86.136200 7/26/2006 Dead/ Declining
SCHOOLCRAFT 1269 46.071700 ~86.058217 7/26/2006 No Scale
TUSCOLA 1 195 43.459900 ~83.365883 7/6/2006 No Beech
VAN BUREN 836 42.330550 ~86.304533 8/8/2005 No Scale
VAN BUREN 837 42.337033 ~86.306933 8/8/2005 No Scale
VAN BUREN 1 189 42.298383 ~85.790067 7/5/2006 No Scale
WAYNE 1200 42.431667 ~83.519750 7/7/2006 No Scale
WEXFORD 1 15 44.201217 ~85.799050 8/4/2004 No Scale
WEXFORD 1 17 44.325033 ~85.820250 8/4/2004 No Beech
WEXFORD 565 44.497133 ~85.608100 6/14/2005 No Scale
WEXFORD 674 44.251800 -85.359633 6/28/2005 No Scale
WEXFORD 682 44.217867 ~85.500717 6/29/2005 No Scale
WEXFORD 684 44.222467 -85.603867 6/29/2005 No Scale
WEXFORD 686 44.246383 -85.703717 6/29/2005 No Scale
WEXFORD 688 44.185600 ~85.708067 6/29/2005 No Scale
WEXFORD 694 44.265800 ~85.608917 6/29/2005 No Scale
WEXFORD 696 44.275883 ~85.619667 6/29/2005 No Scale
WEXFORD 698 44.280733 ~85.640267 6/29/2005 Trace
WEX FORD 700 44.284083 -85.641467 6/29/2005 Trace
WEXFORD 701 44.330483 ~85.407133 6/28/2005 No Scale
WEX FORD 702 44.329100 -85 .609867 6/29/2005 Trace
WEXFORD 703 44.453983 ~85.412083 6/28/2005 No Scale
WEXFORD 705 44.425150 ~85.373317 6/28/2005 No Scale
WEXFORD 71 1 44.260150 -85.336167 6/28/2005 No Scale
WEXFORD 723 44.303517 -85.489400 6/29/2005 No Scale
WEXFORD 725 44.378150 -85.536033 6/29/2005 No Scale
WEXFORD 727 44.406550 -85.6191 17 6/29/2005 No Scale
WEXFORD 729 44.443583 -85.697483 6/29/2005 No Scale
WEXFORD 731 44.469067 ~85.771000 6/29/2005 No Scale
WEXFORD 733 44.352517 ~85.73 8767 6/29/2005 No Scale
WEXFORD 735 44.251467 ~85.660950 6/29/2005 Patchy
WEXFORD 737 44.251 150 ~85.699667 6/29/2005 No Scale
WEXFORD 739 44.251667 -85.644567 6/29/2005 No Scale
WEXFORD 813 44.208817 ~85.604567 7/25/2005 No Scale
WEXFORD 815 44.202800 ~85.657533 7/25/2005 No Scale
WEXFORD 817 44.206000 ~85.640500 7/25/2005 No Scale
WEXFORD 819 44.237000 -85.623267 7/25/2005 No Scale
WEXFORD 821 44.240417 ~85 .647800 7/25/2005 Whitewashed
WEXFORD 823 44.277217 -85.707467 7/25/2005 No Scale
WEXFORD 825 44.343967 ~85.675917 7/25/2005 No Scale
WEXFORD 827 44.344133 ~85.675583 7/25/2005 Trace
WEXFORD 829 44.381333 ~85.616433 7/25/2005 No Scale
WEXFORD 831 44.353367 ~85.567017 7/25/2005 No Scale
WEXFORD 833 44.303500 ~85.559600 7/25/2005 No Scale
WEXFORD 835 44.222667 ~85.609667 8/25/2005 Trace
WEXFORD 1006 44.246820 ~85.703310 5/9/2006 N0 Scale
WEXFORD 1007 44.248100 ~85 .674600 5/9/2006 Trace

 

 

 

 

 

 

139

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WEXFORD 1212 44.435813 -85.738550 7/1 1/2006 Trace
WEXFORD 1213 44.343350 -85.675340 7/1 1/2006 Whitewashed
WEXFORD 1214 44.309110 -85.559510 7/12/2006 No Scale
WEXFORD 1215 44.276000 ~85.619530 7/12/2006 Trace
WEXFORD 1216 44.237180 -85.617080 7/12/2006 Trace
WEXFORD 1217 44.251660 ~85.6445 80 7/12/2006 Patchy
WEXFORD 1218 44.254340 ~85 .622660 7/12/2006 Patchy
WEXFORD 1219 44.222590 -85.603540 7/12/2006 No Scale
WEXFORD 1220 44.208710 -85.604630 7/12/2006 No Scale
WEXFORD 1221 44.206550 ~85.64052O 7/12/2006 No Scale
WEXFORD 1222 44.202800 -85.657532 7/12/2006 No Scale
WEXFORD 1223 44.246270 ~85.704010 7/12/2006 Trace
WEXFORD 1224 44.276990 ~85.707470 7/12/2006 No Scale
WEXFORD 1225 44.248550 -85.687820 7/12/2006 Trace
WEXFORD 1227 44.351883 -85.706467 7/13/2006 No Scale
WEXFORD 1228 44.380867 ~85.617l83 7/13/2006 No Scale

 

140

 

Appendix B.
Parameter values and start points to obtain the lowest SSE for the four main models

LP SIMPLE satellite name “model coordinate” (year)
Cadillac FT4775 (2004)
Benzie CP4768 (2004)
Silver BC3159 (1997)
Ludington A01 103 (1989)
Spread Rate = 3 (1.5 km/year)

LP COMPLEX satellite name “model coordinate” (year)
Cadillac FT4775 (2004)
Benzie CP4769 (2004)
Silver BC3151 (1997)
Ludington A01 103 (1989)

Complex model parameter values for LP
Max move = 3 (1.5 km/year)
B = 1 (1.5 km/year)
D = 1.5 (1 kin/year)
C = 2 (.75 km/year)
O = l (1.5 km/year)

Number of observations LP = 156

SSE LP simple diffusion model = 37

SSE LP land cover based model = 24

UP SIMPLE satellite name “model coordinate” (year)
Bass Lake GP1359 (1990)

Spread Rate = 8 (4 km/year)

UP COMPLEX satellite name “model coordinate” (year)
Bass Lake FW1338 (1990)
Complex model parameter values for UP
Spread rate = 10 (5 km/year)
B = l (5 km/year)
D = 2 (2.5 km/year)
C = 2 (2.5 km/year)
O = 6 (1 km/year)

Number of observations UP = 74
SSE UP simple diffusion model = 16
SSE LP land cover based model = 8

141

Appendix C. Description of Classes Used in the Michigan Statewide Map

This is an explanation of the values present in the Michigan statewide raster map,
with the associated rules used to arrive at the class labels. Arabic numbers in bold type
are those included in the map. Classiﬁcation scheme should be viewed as a series of
sequential if-then statements. Order counts. For example, consider a forest stand where
50% of the canOpy is Aspen, 20% Maple, and 30% Pine. Because Aspen precedes
Upland Mix in the decision rules, the forest types out as Aspen (413) rather than Mixed
Deciduous (419).

Class numbers were chosen in part to be Similar to existing MIRIS Land Cover labels and
their decision rule sequence does not necessarily match the numeric order (for example
class 110 follows class 122 in the decision rules).

Number in parentheses following classiﬁcation name is the grid value in the raster map.

1 Urban
Land areas greater than 10% man-made structures including paved and gravel
roads and parking lots.

121 Airports (3)
Impervious land within airport grounds, including runways.

122 Road/Parking Lot (4)
Roads or parking lots.

 

123 High Intensity Urban (2)
Land area greater than 25% solid impervious cover made from man-
made materials, other than airports, roads, or parking lots.

11 Low Intensity Urban (1)
Land area is greater than 10% and less than 25% man-made structures
including paved and gravel roads and parking lots.

11 Agricultural
Land intensely managed for vegetation production excluding forestry.

2111 Non-vegetated Farmland (5)
Land area tilled for crop production with less than 25% currently
vegetated.

2112 Row crops (6)
Vegetation consists of annual crops planted in rows (e. g. corn,
soybeans).

2113/212 Forage Crops/ Non-tilled herbaceous agriculture (7)

142

Vegetation used for fodder production (e. g. alfalfa, hay). Also includes
land used for pasture, or non-tilled herbaceous agriculture.

222 Orchards/Vineyards/Nursery (9)
Woody trees not grown for Christmas trees.

UPLAN D
Land not periodically ﬂooded nor on hydric soils.

III Upland Openland
Less than 25% of land area is covered by tree canopy, and greater than 25% of
land area is vegetated.

350 Parks/Golf Courses (13)
Maintained for recreational purposes.

320/330 [ﬁgland Shrub/Low Density Trees (12)
The combination of woody shrubs and tree canopy (woody cover)
covers more than 25% of the land area.

310 Herbaceous Openland (10) ,
Less than 25% of land area consists of woody cover.

IV Upland Forest
Proportion of trees exceeds 25% of land area.

A. Upland Deciduous Forest
Proportion of deciduous trees exceeds 60% of the canopy.

411 Northern Hardwood Association (14) ,
Combination of Maples, Beech, Basswood, White Ash, Cherry, Yellow
Birch exceeds 60% of the canopy.

412 Oak Association (15)
Proportion of Oaks exceeds 60% of the canopy.

413 Aspen Association (16)
Proportion of Aspen exceeds 40% of the canopy.

 

 

414 Other Upland Deciduous (17)
Proportion of any other single species exceeds 60% of the canopy.

419 Mixed Upland Deciduous (l 8)
Proportion of deciduous trees exceeds 60% of the canopy.

143

VI.

VII.

Upland Coniferous Forest

421/422 Pines (19)
Proportion of pines exceeds 60% of the canopy.

423 Other Upland Conifers (20)
Proportion of non-pine upland conifers exceeds 60% of the canopy.

 

429 Mixed Upland Conifers (21)
Proportion of coniferous trees exceeds 60% of the canopy.

 

 

43 @land Mixed Forest (22)
Mixed forest not falling into any other category. Proportion of conifers
to deciduous ranges from 40%:60% to 60%:40%.

Water
50 Water (23)
Proportion of open water exceeds 75% of land area.
LOWLAND
Land is periodically ﬂooded and/or on hydric soils.

Lowland Forest
Proportion of trees exceeds 25% of land area.

611 Lowland Deciduous Forest (24)
Proportion of deciduous trees exceeds 60% of the canopy.

 

 

612 Lowland Coniferous Forest (25)
Proportion of coniferous trees exceeds 60% of the canopy.

613 Lowland Mixed Forest (26)

Mixed forest not falling into any other category. Proportion of conifers
to deciduous ranges from 40%:60% to 60%:40%.

Non-forested Wetlands

Proportion of trees is less than or equal to 25% of land area.

621 Floating Aquatic (27)

144

 

VIII

Proportion of ﬂoating aquatic vegetation exceeds 60% of non-water
coven

 

 

 

 

622 Lowland Shrub (28)
Proportion of lowland shrub exceeds 60% of non-water cover.
623 Emergent Wetland (29)
Proportion of emergent vegetation exceeds 60% of non-water cover.
629 Mixed Non-forest Wetland (30)
Non-forested wetlands not falling into any other category.
Bare/Sparsely Vegetated
Land is less than 25% vegetated.
710 Sand/Soil (31)
Land cover is formed primarily of sand or bare soil.
720 Exposed Rock (32)
Land cover is formed of solid rock.
730 Mud Flats (33)
If periodically ﬂooded.
790 Other Bare/Sparselv Vegptated (35)

 

145

BIBLIOGRAPHY

146

Andow, D.A., P.M. Kareiva, S.A. Levin and A. Okubo. 1990. Spread of invading
organisms: Patterns of spread. Landscape Ecology 4: 177-188.

Ayers, M. 2004, November 16. Personal communication. Phone interview.

Bate, L.J., T.R. Torgersen, M.J. Wisdom, and E.O. Garton. 2004. Performance of
sampling methods to estimate log characteristics for wildlife. Forest Ecology and
Management 199: 83-102.

Beeman, LE. and MR. Pelton. 1978. Seasonal foods and feeding ecology of black
bears in the Smoky Mountains. pp. 141-147 in Bears: their biology and management
(C.J. Martinka and L.J. McArthur, Eds.). Bear Biological Association Conference
Services Pub. 3: 375 pp.

Borror, DJ. and RE. White. 1970. The Peterson Field Guide Series: A Field Guide to
Insects of America north of Mexico. Houghton Mifﬂin Company, New York. 137-
138.

Brower, A.E. 1949. The Beech Scale and Beech Bark Disease in Acadia National Park.
Journal of Economic Entomology 42: 226-229.

Bunnell, Fred L., Boyland, Mark Elke Wind. 2002. How should we spatially distribute
dead and dying wood? USDA Forest Service Gen. Tech. Rep. PSW-GTR-181.

Bunnell, Fredrick and David J. Huggard. 1999. Biodiversity across Spatial and
temporal scales: problems and opportunities. Forest Ecology and Management 115:
113-126.

Burnham, K.P. and DR. Anderson. 2002. Model selection and multimodal inference:
A practical information-theoretical approach. 2'”. Ed. New York: Springer-Verlag.

Burns, BS. and DR. Houston. 1987. Managing Beech Bark Disease: Evaluating
Defects and Reducing Losses. Northern Journal of Applied Forestry 4: 28-33.

Butts, SR. and WC. McComb. 2000. Associations of forest-ﬂoor vertebrates with
coarse woody debris in managed forests of Western Oregon. Journal of Wildlife
Management 64: 95-104.

Canham, CD. 1988. Growth and canopy architecture of shade-tolerant trees:
responses to canopy gaps. Ecology 69: 786-795.

Carey, AB. 1983. Cavities in trees in hardwood forests. In Proceedings, Symposium
on snag habitat management, 7-9 June 1983, Flagstaff, Arizona Technical
coordinators: J .W. Davis, G.A. Goodwin, and RA. Ockenfels. USDA Forest Service
Rocky Mountain Forest Range Experimental Station Technical Report. RM-99: 167-
184.

147

Carey, J .R. 1996. The incipient Mediterranean fruit ﬂy population in California:
Implications for invasion biology. Ecology 77: 1690-1697.

Castello, J .M., Leopold, DJ. and P.J. Smallidge. 1995. Pathogens, patterns, and
processes in forest ecosystems. Bioscience 45: 16-24.

Castlebury, L.A., Rossman, A.Y., and A. S. Hyten. 2006. Phylogenetic relationships
of Neonectria/Cylindrocarpon on F agus in North America. Canadian Journal of
Botany 84: 1417-1433.

Chamberlin, TC. 1965. The method of multiple working hypotheses. Science 148:
754-759. (Reprint of 1890 paper in Science 15: 92).

Costello, CM. 1992. Black bear habitat ecology in the central Adirondacks as related
to food abundance and forest management. Syracuse, NY: State University of New
York, College of Environmental Science and Forestry, M.S. thesis.

Cotter, H. Van T. 1977. Beech Bark Disease: Fungi and Other Associated Organisms.
New Hampshire: University of New Hampshire Department of Botany and Plant
Pathology, M.S. thesis.

Dadd, RH. and TE. Mittler. 1965. Studies on artiﬁcial feeding of the aphid Myzus
persicae (Sulzer) 111. Some major nutritional requirements. Journal of Insect
Physiology 11: 717-743.

Davis, MA. 2003. Biotic Globalization: Does Competition from Introduced Species
Threaten Biodiversity? Bioscience 53: 481-489.

DeGraaf, RD. and DD. Rudis. 1986. New England wildlife; Habitat, natural history,
and distribution. General Technical Report NE-lO8. Radnor, PA: USDA Forest
Service.

DeGraaf, RM. and AL. Shigo. 1985. Managing cavity trees for wildlife in the
Northeast. USDA Forest Service General Technical Report NE-lOl.

Dickman, D. and L. A. Leefers. 2003. The Forests of Michigan. The University of
Michigan Press. 297 p.

Dwyer, G. 1992. On the Spatial spread of insect pathogens: Theory and experiment.
Ecology 73: 479-494.

Eiler, J .H., W.G. Wathen, and MR. Pelton. 1989. Reproduction in black bears in the
southern Appalachian Mountains. Journal of Wildlife Management 53: 353-360.

Elowe, K.D. and WE. Dodge. 1989. Factors affecting black bear reproductive success
and cub survival. Journal of Wildlife Management 53: 962-968.

148

Elton, CS. 1958. The ecology of invasions by animals and plants. London Methuen
and Company LTD New York: John Wiley ad Sons Inc.

Erlich, J. 1934. The beech bark disease: A Nectria disease of the Fagus, following
Cryptococcus Fagi (Baer). Canadian Journal of Forest Research 10: 593-701.

Evans, K.E. and RN. Conner. 1979. Snag management. In Proceedings, workshop on
the management of north central and northeastern forests for nongame birds, 23-25
January, 1979, Minneapolis, Minnesota. Edited by RM. Degraff and K.E. Evans.
USDA Forest Service General Technical Report NC-S 1. pp. 214-225.

Eyre, F.H., Ed. 1980. Forest cover types of the United States and Canada. Society of
American Foresters, Washington, DC. 148 p.

F aison, E.K. and DR. Houston. 2004. Black bear foraging response to beech bark
disease in northern Vermont. Northeastern Naturalist 11: 387-394.

Fan, 2., SR, Shirley, M.A. Spetich, F.R. Thompson and DR. Larsen. 2003.
Distribution of cavity trees in Midwestern old-growth and second-growth forests.
Canadian Journal of Forest Resources 33: 1481-1494.

Fernandez, MR. and MG. Boyer. 1988. Beech bark disease —- A survey of the
Toronto area. Canadian Plant Disease Survey 68: 157-159.

Forrester, J .A. and J .R. Runkle. 2000. Mortality and replacement patterns of old-
growth Acer-Fagus woods in the Holden Arboretum, Northeastern Ohio. American
Midland Naturalist. 144: 227-242.

Fuhrman, Nicholas, A. 2004. An analysis of the ecology and public perception of
coarse woody debris in Virginia. Virginia Polytechnic Institute and State University.
Blacksburg, Virginia.

Gavin, D.G. and DR. Peart. 1993. Effects of beech bark disease on the growth of
American beech (F agus grandifolia). Canadian Journal of Forest Research 23: 1566-
1575.

Gibbs, J .N., and D. Wainhouse. 1986. Spread of forest pests and pathogens in the
northern hemisphere. Forestry 69: 141-153.

Glover, FA. 1949. Fox foods on West Virginia wild turkey range. Journal of
Mammology 30: 78-79.

Goodburn, J .M. and CG. Lorimer. 1998. Cavity trees and coarse woody debris in old-
growth and managed northern hardwood forests in Wisconsin and Michigan. Canadian
Journal of Forest Research 28: 427-438.

149

Graves, S., J. Maldonado, and J .O. Wolff. 1988. Use of ground and arboreal
microhabitats by Peromyscus leucopus and Peromyscus maniculatis. Canadian J oumal
of Zoology 66: 277-278.

Greenberg, CH. 2002. Response of white—footed mice (Peromyscus leucopus) to coarse
woody debris and micro Site use in southern Appalachian tree fall gaps. Forest Ecology
and Management 164: 57-66.

Grifﬁn, J .M., G.M Lovett, M.A. Arthur, and KC. Weathers. 2003. The distribution
and severity of beech bark disease in the Catskills Mountains, N.Y. Canadian Journal
of Forest Research 33: 1754-1760.

Gysel, L.W. 1961. An ecological study of tree cavities and ground burrows in forest
stands. Journal of Wildlife Management 35: 516-519.

Gysel, L.W. 1971. A 10-year analysis of beechnut production and use in Michigan.
Journal of Wildlife Management 35: 516-519.

Halls, L. K., ed. 1977. Southern Fruit-producing Woody Plants Used by Wildlife.
USDA Forest Service General Technical Report SO-16. Southern Forest Experiment
Station.

Hanaburgh, Christine. 2001. Modeling the effects of management approaches on
forest and wildlife resources in northern hardwood forests. Michigan State University.
East Lansing, Michigan 48224.

Hane, EN. 2003. Indirect effects of beech bark disease on sugar maple seedling
survival. Canadian Journal of Forest Research 33: 807-813.

Hardin, D.P., Takac, P., and G.F. Webb. 1990. Dispersion population models discrete
in time and continuous in space. J. Mathematical Biology 28: 1-20.

Harlow, H.J., T. Lohuis, R.G. Grogan, T.D.I. Beck. 2002. Body mass and lipid
changes by hibernating reproductive and non-reproductive black bears (Ursus
americanus). Journal of Mammology 83: 1020-1025.

Harmon, M.E., J .F . Franklin, F.J. Swanson, P. Sollins, S.V Gregory, J .D. Lattin, N.H.
Anderson, S.P. Cline, N.G. Aumen, J .R. Sedell, G.W. Lienkaemper, K. Cromack, and

K.W. Cummins. 1986. Ecology of coarse woody debris in temperate ecosystems.
Advances in Ecological Research 15: 133-302.

150

 

Hastings, A. 1996. Models of spatial spread: Is the theory complete? Ecology 78:
2145-2152.

Hastings, A., Cuddington, K., Davies, K.F., Dugaw, C.J., Elmendorf, S., Freestone, A.,
Harrison, 3., Holland, M., Lambrinos, J ., Malvadkar, U., Melbourne B.A, Moore K.,
Taylor, C., and D. Thomson. 2005. The spatial spread of invasions: new developments
in theory and evidence. Ecology letters 8: 91-101 .

Hayes, D. B., J. K.T., Brodziak, and J .B. O’Gorman. 1995. Efﬁciency and bias of
estimators and sampling designs for determining length-weight relationships of ﬁsh.
Canadian Journal of Fisheries and Aquatic Sciences 52: 84-92.

Healy, W.M., R.T. Brooks and R.M. Degraaf. 1989. Cavity trees in sawtimber-size
oak stands in central Massachusetts. Northern Journal of Applied Forestry 6: 61-65.

Held, Michael E. 1983. Pattern of beech regeneration in the east-central United States.
Bulletin of the Torrey Botanical Club 110: 55-62.

Hepting, G.H., and GM. Jemison. 1958. Forest protection. Pages 185-220 in Timber
Resources for America’s Future. Forest Research Report Number 14. USDA Forest
Service. »

Herbert, P.D.N. and M.E.A. Cristescu. 2002. Genetic perspectives on invasions: the
case of the Cladocera. Canadian Journal of Fisheries and Aquatic Science 59: 1229-
1234.

Heyd, R.L. 2004. Managing beech bark disease in Michigan in: Beech bark disease:
Proceedings of the beech bark disease symposium, June 16-18, Saranac Lake, New
York. USDA Forest Service General Technical Report NE-33 1. pp. 128-132.

Hicks, R.R., JR. 1998. Ecology and management of central hardwood forests. John
Wiley and Sons, Inc., New York, New York. John Wiley and Sons, Inc., New York,
New York.

Holmes, E. E. 1994. Partial differential equations in ecology: Spatial interactions and
population dynamics. Ecology 75: 17-29.

Houston DR. 1994. Major new tree disease epidemics: Beech bark disease. Annual
Review Phytopathology 32: 75-87.

Houston, DR. 1975. Beech bark disease: The aﬁermath forests are structured for a
new outbreak. Journal of Forestry 73: 660-663.

Houston, DR. 1980. Beech bark disease: what we do and do not know. Annales
Sciences Forestieres 37: 269-274.

151

Houston, DR. 1982. A technique to artiﬁcially infest beech bark with the beech scale,
Cryptococcusfagisuga (Lindinger). USDA Forest Service Research Paper NE-507. 8 p.

Houston, DR. 1983. American beech resistance to Cryptococcusfagisuga.
Proceedings IUFRO beech bark disease work party conference, pp. 38-42. USDA
Forest Service General Technical Report WO-37.

Houston, DR. 1984. What is happening to the American beech? Conservationist, 38:
22-25.

Houston, DR. 1987. Forest tree declines of past and present: Current understanding.
Canadian Journal of Plant Pathology 9: 349-360.

Houston, DR. 1996. Potential for Biologically Based Control of Beech Bark Disease
in the Southern Appalachians. In: Proceedings from the: Southern Appalachian
biological initiative workshop proceedings. The North Carolina Arboretum, Asheville,
North Carolina, September 26-27.

Houston, DR. 1997. Beech Bark Disease. In: Exotic Pests of Eastern Forests,
Conference Proceedings (K. O’Britton, ed.), April 8-10, Nashville, Tennessee. USDA
Forest Service and TN Exotic Pest Plant Council.

Houston, DR. 2004. Beech bark disease: 1934-2004: What’s new since Ehrlich? In:
Beech bark disease: Proceedings of the beech bark disease symposium, June 16-18,
Saranac Lake, New York. USDA Forest Service General Technical Report NE—33 1.
pp. 2-13.

Houston, DR. and EM. Mahoney. 1987. Beech bark disease: Association of Nectria
ochroleuca in West Virginia, Pennsylvania, and Ontario. Phytopathology. 77(11):
1615.

Houston, DR. and J. T. O’Brien. 1983. Beech bark disease: Forest Insect and Disease
leaﬂet 75. US. Department of Agriculture: Forest Service 1983. Accessed on October
13, 2004 http://www.na.fs.ged.us/spfo/pubs/ﬁdls/beechbark/ﬁdl-beech.htrn.

Houston, D.R., E.J. Parker, and D. Lonsdale. 1979. Beech bark disease: patterns of
spread and development of the initiating agent Cryptococcusfagisuga. Canadian
Journal of Forest Research 9: 336-343.

Hugie, RD. 1982. Black bear ecology and management in the northern conifer-
deciduous forests of Maine. PhD Dissertation. University of Montana, Missoula. 201

PP-

Kahler, Harry A. and James T. Anderson. 2006. Tree cavity resources for dependent
cavity-using wildlife in West Virginia Forests. Northern Journal of Applied Forestry
23(2): 114-201.

152

 

Kearney, A. 2006. Impacts of beech bark disease on stand composition and wildlife
resources in Michigan. Michigan State University. East Lansing, Michigan 48224.

Kearney, A., McCullough, D.G., and M. Walters. 2004. Impact and spread of beech
bark disease in Michigan: A progress report. Unpublished data for the Michigan
Department of Natural Resources.

Keddy, RA. and CG. Drummond. 1996. Ecological properties for the evaluation,
management and restoration of temperate deciduous forest ecosystems. Ecological
Applications 6: 748-762.

Kelty, M.J. and R.K Nyland. 1981. Regenerating Adirondack northern hardwoods by
shelterwood cutting and control of deer density. Journal of Forestry 79: 22-26.

Keneﬁc, Laura S. and Ralph D. Nyland. 2007. Cavity trees, snags and selection
cutting: A northern hardwood case study. Northern Journal of Applied Forestry 42:
192-196. .

Kizlinski, M.L., D.A. Orwig, R.C. Cobb, and DR. Foster. 2002. Direct and indirect
ecosystem consequences of an invasive forest pest on forests dominated by eastern
hemlock. Journal of Biogeography 29: 1489-1503.

Kot, M. and W.M. Schaffer. 1986. Discrete-time growth-dispersal models.
Mathematical Biosciences 80: 109-136.

Kot, M., Lewis, M. A., and P. Van Den Driessche. 1996. Dispersal data and the spread
of invading organisms. Ecology 77: 2027-2042.

Krasny, ME. and M.C. Whitmore. 1992. Gradual and sudden forest canopy gaps in
Allegheny northern hardwood forests. Canadian Journal of Forest Research 22: 139-
143.

Kruse, R.L. 1990. The dynamics of wildlife habitat in northern hardwood ecosystems
in New York’s Adirondack region. Ph.D. dissertation. State University of New York
College of Environmental Science and Forestry. Syracuse, NY. 220 p.

LaChance, D. 1983. Status of beech bark disease in the Province of Quebec. In:
Proceedings IUFRO beech bark disease work party conference, pp. 38-42. USDA
Forest Service General Technical Report WO-3 7.

Latty, E.F., C.D. Canham, and PL. Marks. 2003. Beech bark disease in northern

hardwood forests: the importance of nitrogen dynamics and forest history for disease
severity. Canadian Journal of Forest Research 33: 257-268.

153

 

Le Guerrier, C., D.J. Marceau, A. Bouchard, and J. Brisson. 2003. A modeling
approach to assess the long-term impact of beech bark disease in northern hardwood
forest. Canadian Journal of Forest Research 33: 2416-2425.

Leak, W.B. 2006. Fifty-year impacts of the beech bark disease in the Bartlett
Experimental Forest, New Hampshire. Northern Journal of Applied Forestry 23: 141-
143.

Leak, W.B. and ML. Smith. 1996. Sixty years of management and natural disturbance
in a New England forested landscape. Forest Ecology and Management 81: 63-73.

Lee, S.D., Park, S., Park Y.S., Chung, Y.J., Lee, B.Y. and TS. Chon. 2007. Range
expansion of forest pest populations using the lattice model. Ecological Modeling 203:
157-166.

Levin, SA. 1989. Analysis of risk for invasions and control programs. In Biological
invasions: a global perspective. Wiley, Chichester, U, K. pp. 425-435.

Lewin, R. 1987. Ecological invasions offer opportunities. Science 238: 752-753.

Lewis, Mark. 1997. Variability, patchiness and jump dispersal in the spread of an
invading population. In Spatial Ecology: The Role of Space in Population Dynamics
and Interspeciﬁc Interactions. Princeton University Press. pp. 46-69.

Liebhold, A.M., Halverson, J .A. and G.A. Elmes. 1992. Gypsy moth invasion in
North America: A quantitative analysis. Journal of Biogeography 19: 513-520.

Liebhold, A.M., W.L MacDonald, D. Bergdahl and V.V. Mastro. 1995. Invasion by
exotic forest pests: A threat to forest ecosystems. Forest Science 41: 1-49.

Lonsdale, D. 1983. Fungal associations in the build-up and decline of Cryptococcus
fagisuga populations. Proceedings IUFRO Beech bark disease work party conference
pp. 99-104.

Lortie, M. 1964. Pathogenesis in Cankers Caused by Neonectria galligena.
Phytopathology 54: 261 -262.

Ludwig, D., Aronson, D.G., and HF. Weinberger. 1979. Spatial patterning of the
spruce budworrn. Journal of Mathematical Biology 8: 217-258.

Mack, R. N. 1981. lnvading plants: their potential contribution to population biology.

Studies on plant demography: A F estschriﬁ for John L. Harper. (Ed. By J. White), pp.
127-142.

154

MacKenzie, M. 2004. The picture of beech bark disease that we perceive is dependant
upon the scale of the maps on which we base the image. Unpublished data.

Mahoney, E.M., Milgroom, M.G., Sinclair, W.A., and DR. Houston. 1999. Origin,
genetic diversity, and population structure of Neonectria coccinea var. faginata in
North America. Mycologia 91: 583-592.

Marquis, DA. and T.J. Grisez. 1978. The effect of deer exclosures on the recovery of
vegetation in failed clear-cuts on the Allegheny Plateau. USDA Forest Service
Research Note NE-270 Broomall, Pennsylvania.

Mattson, W.J. 1997. Exotic insects in North American forests — ecological systems
forever altered. In Proceedings of Exotic Pests of Eastern Forests, 8-10 April 1997,
Nashville, TN. Edited by Kerry 0. Britton. USDA Forest Service and Tennessee

Exotic Pest Plant Council, Nashville, Tennessee. pp. 187-193.

McCullough D.G., Heyd, R.L., and J .G. O’Brien. 2001. Biology and management of
beech bark disease: Michigan’s newest exotic forest pest. Michigan State University
Extension Bulletin: E-2746. pp. 12. Reprinted October 2002.

McDonald, J .E., J r., T.K Fuller. 1994. Black bear food habits: Beyond the same old
scats. In: Thompson, I.D. (Ed.), Proceedings of the International Union of Game
Biologists, XXI Congress, Forest and wildlife towards the 21St century, August 15-20,
1993, Halifax, Nova Scotia, pp. 293-298.

McGee, G.G. 2000. The contribution of beech bark disease-induced mortality to
coarse woody debris loads in northern hardwood stands of Adirondack Park, New
York, USA. Canadian Journal of Forest Research 30: 1453-1462.

McLaughlin, C.R., Matura, G.J., Jr., and R.J. O’Connor. 1994. Synchronous
reproduction by Maine black bears. International Conference on Bear Research and
Management 9: 471-479.

Menzel, M.A., W.M. Ford, J. Laerrn, and D. Krishon. 1999. Forest to wildlife
opening: habitat gradient analysis among small mammals in the southern Appalachians.
Forest Ecology and Management 114: 227-232.

Michigan Department of Natural Resources Forest, Mineral & Fire Management
Division. 2004. Michigan 2004 Forest Health Highlights. Michigan Department of
Natural Resources '
http://www.michi£an.gov/documents/2004ForestHealthHighlights3_l 16430_7.pdf

Miller, B.K. 1994. Woodland wildlife management. Woodland Management

Cooperative Extension Service Bulletin Purdue University: FNR-102. pp. 14.
Reprinted October 1994.

155

Miller-Weeks, M. 1982, Current status of beech bark disease in New England and New
York. Po 21-23. In Proceedings, IUFRO Beech Bark Disease Working Party
Conference, Hamden, CT. US Forest Service General Technical Report WO-37.

Moody, ME. and RN. Mack 1988. Controlling the spread of plant invasions: the
importance of nascent foci. The Journal of Applied Ecology 25: 1009-1021.

Mooney, H.A. and J. Hobbs. 2000. Introduction. In H.A. Mooney and J. Hobbs (Eds).
Invasive species in a changing world. Island Press: pp. xiii-xv.

Morin, R.S., A. M. Liebhold, E.R. Loader, A.J. Lister, K.W. Gottschalk, and DB.
Towards. 2001. Mapping host-species abundance of three major exotic forest pests.
USDA Forest Service Northeastern Research Station, Research Paper NE-726.

Muirhead, J .R., Leung, B., Overdijk, C.V., Kelly, D.W., Nandakumar, K., Marchant,
K.R. and H.J. Maclsaac. 2006. Modeling local and long-distance dispersal of invasive
emerald ash borer Agrilus planipennis (Coleoptera) in North America. Diversity and

Distributions 12: 71-79.

National Research Council. 2002. Predicting invasions of nonindigenous plants and
plant pests. Washington, DC: National Academy Press.

Negri, Stephen J. 1995. Analysis of Habitat Suitability Models for Primary Cavity-
nesting Birds in Michigan’s Upper Peninsula. Michigan State University. East
Lansing, Michigan 48224.

Nelson, R.A., Folk, G.E., Jr., Pfeiffer, E.W., Craighead, J .J ., Jonquil, C.J., and D.L.
Steger. 1983. Behavior, biochemistry, and hibernation in black, grizzly, and polar
bears. International Conference on Bear Research and Management 5: 284-290.

Neubert, MG. and H. Caswell. 2000. Demography and dispersal: Calculation and
sensitivity analysis of invasion speed for structured populations. Ecology 81: 1613-
1628.

Neubert, M.G., I.M. Parker. 2004. Projecting rates of spread for invasive species. Risk
Analysis 24: 817-831.

Nibbelink, N.P., S.R. Carpenter. 1998. Interlake variation in growth and size structure
of bluegill (Lepomis macrochirus): Inverse analysis of an individual-based model.

Canadian Journal of Fishery Aquatic Science. 55: 387-396.

Nixon, C.M, Worley, D.M., M.W. McClain. 1968. Food habits of squirrels in southeast
Ohio. Journal of Wildlife Management 32: 294-305.

O’Brien, J .G., M.E. Ostry, and ME. Mielke. 2001. First report of beech bark disease
in Michigan. Plant Disease 69: 905.

156

Orwig, D.A. 2002. Ecosystem to regional impacts of introduced pests and pathogens:
Historical context, questions and issues. Journal of Biogeography 29: 1471-1474.

Ostrofsky, W.D. and ML. McCormack. 1986. Silvicultural management of beech and
the beech bark disease. Northern Journal of Applied Forestry 3: 89-91.

Parker, R.L. 1977. Understanding inverse theory. Annual Review of Earth and
Planetary Science 5: 35-64.

Pierce, Lars L. and Steven W. Running. 1988. Rapid Estimation of coniferous forest
leaf Area index using a portable integrating radiometer. Ecology 69: 1762-1767.

Pimentel, D., L. Latch, R. Zuniga, and D. Morrison. 2000. Environmental and

economic costs associated with non-indigenous species in the United States.
BioScience 50: 53-65.

Plante, F., C. H. Hame. 1995. Factors inﬂuencing wood duck use of natural cavities.
Journal of Wildlife Management

Plante, F., Hamelin, RC. and L. Bemier. 2002. A comparative study of genetic
diversity of populations of Neonectria galligena and N. Coccinea var. faginata in North
America. Mycological Research 106: 183-193.

Poland, T.M., McCullough, D.G., Petrice, TR, and NW. Siegert. 2001. Overview on
the pest status and research plans of beech bark disease: a new exotic in Michigan.
Newsletter of the Michigan Entomological Society. 46(3): 10.

Pyle, C. and M.M Brown. 1998. A rapid system of decay classiﬁcation for hardwood
logs of the eastern deciduous forest ﬂoor. Journal of the Torrey Botanical Society 125:
237-245.

Rafferty, Dan, Masters, Ron, Dr. and Green Champe. 2008. Snags, cavity trees and
downed logs. Oklahoma State University Cooperative Extension Service Wildlife
Management Notes 4.

Roane, M.K., G.K. Grifﬁn, and J .R. Elkins. 1986. Chestnut blight, and other Endothia
disease, and the genus Endothia. St. Paul, MN: American Phytopathological Society.

Robb, J .R. and TA. Bookhout. 1995. Factors Inﬂuencing Wood Duck Use of Natural
Cavities. Journal of Wildlife Management 59: 372-3 83.

Rossman A.Y., Samuels, G.J., Rogerson, C.T., and R. Lowen. 1999. Genera of

Bionectriaceae, Hypocreaceae, and Nectriacceae (Hypocreales, Ascomycetes). Study.
Mycologia. 42: 1-248.

157

Runkle, J .R. 1990. Eight years change in an old Tsuga canadensis woods affected by
beech bark disease. Bulletin of the Torrey Botanical Club 117: 409-419.

Rushmore, F.J. 1961. Silvical characteristics of beech (Fagus grandifolia). USDA
Forest Service Northeast Experimental Station Research Paper 161 p.

Russell, NH. 1953. The beech gaps of the Great Smoky Mountains. Ecology 34: 366-
374.

Searle, S. R. 1987. Linear models for unbalanced data. John Wiley and Sons, New
York. 536 p.

Shaffer, ML. 1981. Minimum population sizes for species conservation. Bioscience
31: 131-134.

Sharov, A. A., and A. M. Liebhold. 1998. Model of slowing the spread of the gypsy
moth (Lepidoptera: Lymantriidae) with a barrier zone. Ecological Applications 8:
1170-1179.

Sharov, AA. 2004. Bioeconomics of managing the spread of exotic pest species with
barrier zones. Risk Analysis 24: 879-892.

Sharov, A.A., Leonard, D., Liebhold, A.M. Roberts, EA. and W. Dickerson. 2002.
Slow the Spread: A national program to contain the gypsy moth. Journal of Forestry
100: 30-35.

Shigesada, N., and K. Kawasaki. 1995. Modeling stratiﬁed diffusion in biological
systems. The American Naturalist 146: 229-251.

Shigo, AL. 1962. Another scale insect on beech. USDA Forest Service Station Paper
168. Northeast Forest Experiment Station pp. 13.

Shigo, AL, 1972. The beech bark disease in the northeastern United States. Journal of
Forestry 70: 286-289.

Shigo, AL. 1976. The beech bark disease. Journal of Arboriculture 2: 21-25.

Skellam, J .G. 1951. Random dispersal in theoretical populations. Biometricka 38:
196-218.

Sokal, RR. and F.J. Rohlf. 1994. Biometry: the principles and practice of statistics in
biological research, 3lrd edition. Freeman Publishing Company, New York.

Spaulding, P., Grant, T.J., and T.T. Ayers. 1936. Investigations of Neonectria
diseases in hardwoods of New England. Journal of Forestry. 34: 169-179.

158

Speight, MR. 1981. Tree Pests-5 Beech Scale (Cryptococcus fagisuga Lind) and
Ambrosia beetle (Xyloterus domesticum (L)). Arboricultural Journal 5: 143-146.

Stewart-Oaten, Allan. 1995. Rules and judgments in statistics: three examples.
Ecology 76: 2001-2009.

Stone, J ., J. Parminter, A. Arsenault, T. Manning, N. Densmore, G. Davis and A.
MacKinnon. Dead tree management in British Columbia. 2002. USDA Forest Service
Technical Report PSW-GTR-l 81.

Storer, A.J., J .N. Rosemier, B.L. Beachy and DJ. F lashpohler. 2004. Beech bark
disease, In Proceedings of the beech bark disease symposium. USDA Forest Service
General Technical Report NE-331. pp. 72-78.

Suarez, A. V. D. A Holway, and T.J. Case. 2001. Patterns of spread in biological
invasions dominated by long-distance jump dispersal: Insights from Argentine ants.
Proceedings of the National Academy of Sciences of the United States 98: 1095-1100.

Swartzman, G.L., and SP. Kaluzny. 1987. Ecological simulation primer. Macmillan
Publishing Company, New York. '

Taylor, CM and A. Hastings. 2004. Finding optimal control strategies for invasive
species: a density-structured model for Spartina alterniflora. Journal of Applied
Ecology 41: 1049-1057.

Thomas, J .W. Anderson, R.G., Master, C. and EL. Bull. 1979. Snags. In Wildlife
habitats in managed forests: The blue Mountains of Oregon and Washington. USDA
Forest Service Agriculture Handbook 553: 60-76.

Thompson, F.R. III, and DE. Capen. 1988. Avian assemblages in seral stages of a
Vermont forest. Journal of Wildlife Management 52: 771-777.

Tilghman, Nancy G. 1989. Impacts of white-tailed deer on forest regeneration in
northwestern Pennsylvania. Journal of Wildlife Management 53: 524-532.

Tillman, David and Peter M. Kareiva. 1997. Spatial Ecology: The Role of Space in
Population Dynamics and Interspeciﬁc Interactions. Princeton University Press.

Tomback, DE, SE Arno, and RE. Keane. 2001. Whitebark Pine Communities:
Ecology and Restoration. Island Press Washington DC.

Towers, B. 1983. Status of beech bark disease in Pennsylvania. Proceedings IUFRO

beech bark disease work party conference, pp. 38-42. USDA Forest Service General
Technical Report WO-37.

159

Tubbs, CH. and DR. Houston. 1990. Fagus grandifolia Ehrh. American Beech.
USDA Forest Service northeastern area state and private forestry. Accessed on
November 11, 2004.
http://www.na.fs.fed.us/spfo/pubs/silvics_manual/Volume2/fagus/grandifolia.htm.

Tubbs, C.H., R.M DeGraaf, M. Yamasaki, and W.M. Healy. 1987. Guide to wildlife
tree management in New England Northern Hardwoods. USDA Gen Technical Report
NE-118.

US. Congress Ofﬁce of Technology Assessment. 1993. Harmful Non-Indigenous
Species in the United States. OTA-F-565. Washington, DC. Accessed on April 25‘“.
http://www.wws.princeton.edu/~ota/disk1/1 993/9325_n.htm1

 

US. Fish and Wildlife Service. Invasive Species. Accessed on August 27, 2006
http://www.fws.Lov/midwest/EcosystemConservation/exotic.htrnl

 

US. Geological Survey. 1999. Digital representation of Atlas of United States Trees
by Elbert L. Little, Jr. Accessed on August 23, 2006.
http://esm:r.usgs.gov/data/atlas/little/fagpgran.pdf

 

Veit, RR. and M.A. Lewis. 1996. Dispersal, population growth, and the allee effect:

Dynamics of the house ﬁnch invasion of eastern North America. American Naturalist
148: 255-274.

Vitousak, P.M., C.M., D’Antonio, L.L. Loope, and R. Westbrooks. 1996. Biological
invasions as global environmental change. American Scientist 84: 468-478.

Wainhouse, D and I.M. Gates. 1988. The beech scale pp. 66-85, In: Dynamics of forest
insect populations: Patterns, causes and implications. A.A. Berryman, ed. Plenum
Press, NY.

Wainhouse, D. 1980. Dispersal of ﬁrst instar larvae of the Felted Beech Scale,
Cryptococcus fagisuga. The Journal of Applied Ecology 17: 523-532.

Wainhouse, D. and R. Deeble. 1980. Variation in susceptibility of beech (Fagus spp.)
to beech scale (Cryptococcus fagisuga). Annales Sciences F orestieres 37 : 279-289.

Wainhouse, D. and RS. Howell. 1983. Intraspeciﬁc variation in beech scale
populations and in susceptibility of their host F agus sylvatica. Ecological Entomology
8: 351-359.

Wargo, PM. 1988. Amino nitrogen and phenolics constituents of bark of American

beech, F agus grandifolia, and infestation by beech scale, Cryptococcusfagisuga.
European Journal Forest Pathology 18: 279-290.

160

 

Weber, E. 1998. The dynamics of plant invasions: a case study of three exotic
goldenrod species (Solidago L.) in Europe. Journal of Biogeography 25: 147-154.

White, B. 2005 April 17. Personal communication.

Whittaker, RH. 1956. Vegetation of the Great Smoky Mountains. Ecological
Monographs 26: 1-80.

Wiggins, G.J., J .F . Grant, M.T. Windham, R.A. Vance, B. Rutherford, R. Klein, K.
Johnson, and G. Taylor. 2004. Associations between causal agents of the beech bark
disease complex [Cryptococcusfagisuga (Homoptera: Cryptococcidae) and Neonectria
spp.] in the Great Smoky Mountains National Park. Entomological Society of America

33: 1274-1281.

Willgins, G.J., J .F. Grant, and W. Cal Welbourn. 2001. Allothrombium mitchelli
(Acari: Trombidiidae) in the Great Smoky Mountains National Park: Incidence,
Seasonality, and Predation on Beech Scale (Homoptera: Eriococcidae). Entomological
Society of America 94: 896-901.

Williams, AB. 1936. The composition and dynamics of a beech-maple climax
community. Ecological Monographs 6: 318-408.

Witter, J .A., J .L. Stoyenoff, H.A. Petrillo, J .L. Yocum, and J .1. Cohen. 2004. Effects
of Beech Bark Disease on Trees and Ecosystems. In: Beech bark disease: Proceedings
of the beech bark disease symposium, June 16-18, Saranac Lake, New York. USDA
Forest Service General Technical Report NE-331. pp. 128-132.

Wollenweber, H.W. 1917. F usarium autographica delineatum. Annals of Mycologia.
15: 1-96.

Workman, R.D., Hayes DB, and T.G. Coon. 2002. A model of steelhead movement
in relation to water temperature in two Lake Michigan tributaries. Transactions of the

American Fisheries Society 131: 463-475.

Yahner, Richard H. 1995. Eastern Deciduous Forest: Ecology and Conservation.
Minneapolis, Minnesota: University of Minnesota Press.

Youngs, R. L. 2000. “A right smart little jolt,” Loss of the American chestnut and a
way of life. Journal of Forestry 98: 17-21.

161