. .an:
. .axafi. a. anus for.
. . . 4.0:... {4.5
. . 9...“... ‘ s
WE»?
..

Lfid ..
I

to.

.
. x
.L. 3.1
In). 3‘!

I... :3:
It .
p lth}
was}...
(:3.

 

3
.s.
- . .. ;
- 2T a
4

I

 

LI‘B‘RARV”?
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

EFFECT OF LAND COVER CHANGE ON RED-BEADED WOODPECKER
POPULATIONS AT HIGH VS LOW ABUNDANCE RANGE LOCATIONS

presented by

Laura A. Lukomski

has been accepted towards fulfillment
of the requirements for

M. S. degree in GeograRhL

Ram ESL

\

 

Major professor

Date “MSW i 2-003

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

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRCIDateDuosz-pJS

EFFECT OF LAND COVER CHANGE ON RED-HEADED WOODPECKER
POPULATIONS AT HIGH VS. LOW ABUNDANCE RANGE LOCATIONS

By

Laura A. Lukomski

A THESIS
Submitted to
Michigan State University
in the partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Geography

2003

ABSTRACT

EFFECT OF LAND COVER CHANGE ON RED-HEADED WOODPECKER
POPULATIONS AT HIGH VS. LOW ABUNDANCE RANGE LOCATIONS

BY

Laura A. Lukomski

Range collapse studies of endangered species have revealed that location
within a species” geographic range is not a critical factor for survival. However,
wildlife mangers have documented a higher translocation success rate if
transplant populations are relocated within the center of the species geographic
range. The inconsistencies between translocation success and patterns of range
collapse lead to management confusion. The purpose of this study is to help
clarify this confusion by evaluating a case species, that of the red-headed
woodpecker (Melanerpes erythmcephalus), from 1978 to 1993 and 1.) Isolate
land cover changes significantly related to population change; and 2.) Determine
if the red-headed woodpecker is more sensitive to these changes at low versus
high- abundance regions of its range.

The multi-tempcral GIS and statistical analysis revealed the red-headed
woodpecker to be statistically sensitive to changes in agricultural mean patch
dimension, agricultural edge density, and urban area; and abundance change
was not significantly different between low versus high-abundance routes.
Therefore, the red-headed woodpecker was not sensitive to range location in this
study, and, as a result, optimal habitat located in the periphery of the species”

range should not be disregarded for management activities.

ACKNOWLEDGEMENTS

I would like to express my deepest gratitude to my advisory committee, Dr.
David Lusch, Dr. Brian Maurer, and Dr. Catherine Lindell, for their knowledge
and guidance. Thank you.

Thanks to my sister, Kara Lukomski, who will always be my best friend.
Finally, I would like to thank my parents, Andy and Marsha Lukomski, for their

unwavering support and encouragement.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................... vi
LIST OF FIGURES ....................................................................... vii
CHAPTER 1
INTRODUCTION .......................................................................... 1
1.1 Background .................................................................... 1
1.2 Statement of Problem ...................................................... 3
1.3 Overview of Thesis ......................................................... 4
CHAPTER 2
LITERATURE REVIEW .................................................................. 6
2.1 Spatial Distribution of Species ........................................... 6
2.2 Species Reintroduction .................................................... 8
2.3 Spatial Patterns of Decline ................................................ 9
2.4 Breeding Bird Survey ....................................................... 12
2.5 Red-headed Woodpecker ................................................. 14
CHAPTER 3
METHODS ................................................................................... 17
3.1 Overview ....................................................................... 17
3.2 Data Compatibility Issues .................................................. 23
3.3 Data Preparation ............................................................. 26
3.3.1 Bird Data ............................................................ 26
3.3.2 Land Cover Data ................................................. 30
3.3.2.1 1978 Land Cover ..................................... 33
3.3.2.2 1993 Land Cover ...................................... 34
3.4 Analysis ......................................................................... 36
CHAPTER 4
RESULTS .................................................................................... 40
4.1 Red-headed Woodpecker Population Change ....................... 40
4.2 Land Cover Change ......................................................... 43
4.3 Analysis ......................................................................... 48
4.3.1 Multiple Regression .............................................. 48
4.3.2 Hierarchical Clustering Analysis .............................. 49
4.3.3 Analysis of Variance ............................................. 52
CHAPTER 5
DISCUSSION AND CONCLUSIONS ................................................. 56

APPENDICES .............................................................................. 61

Appendix A

Logm Land Cover and Red-headed Woodpecker Change ............ 62

Appendix B

Hierarchical Clustering Analysis Agglomeration Schedule ............. 93
LITERATURE CITED ..................................................................... 96

LIST OF TABLES

Table 1. Land Cover Classification and Justification .............................. 31
Table 2. Land Cover Measures and Description ................................... 32
Table 3. Aggregation of 1978 Land Cover Data .................................... 34
Table 4. Aggregation of 1993 Land Cover Data .................................... 36
Table 5. Red-Headed Woodpecker Change: 1978 -1993 ....................... 41
Table 6. Average Land Cover Change Occurring Along BBS Routes ........ 47
Table 7. Stepwise Regression Output ................................................ 48
Table 8. Analysis of Variance ........................................................... 54

vi

LIST OF FIGURES

Figure 1. Red-Headed Woodpecker Range ......................................... 15
Figure 2. Cartographic Model .......................................................... 19
Figure 2a. Cartographic Model: Bird Data ............................... 19
Figure 2b. Cartographic Model: 1978 Land Cover Data .............. 20
Figure 2c. Cartographic Model: 1993 Land Cover Data ............... 21
Figure 2d. Cartographic Model: Analysis .................................. 22
Figure 3. Majority Filter Example ...................................................... 25
Figure 4. Sample Route Locations .................................................... 28
Figure 5. Sample BBS Routes Quartile Ranking .................................. 29
Figure 6. Change in Red-Headed Woodpecker Abundance:

1978 - 1993 ....................................................................... 42
Figure 7. Box Plot: Red-Headed Woodpecker Change by Quartile ......... 43
Figure 8. 1978 - 1993 Land Cover. BBS Route 52029, Missouri ........... 45
Figure 9. 1978 - 1993 Land Cover. BBS Route 91020, Vlflsconsin ......... 46
Figure 10. Hierarchical Cluster Analysis Results ................................. 50
Figure 11. Hierarchical Cluster Distribution ......................................... 52

vii

CHAPTER 1 . INTRODUCTION

1.1 Background

The structure of land cover on our planet has changed dramatically over
the past century. The human population has increased at an extraordinary rate
and changed the landscape as it continues to grow. Forests have been
converted to agriculture, agricultural lands engulfed by suburban development,
and productive wetlands paved over for shopping centers (Whitney 1994, Rome
2001). As more people demand resources, non-human species are forced to
adapt or compete for diminishing resources (Campbell 1996).

The challenge of our generation is to ensure that these species will have a
place on our planet for the enjoyment of future generations and to preserve
Earth’s biological diversity. The more we can learn about the distribution, habitat
requirements, and biology of an animal, the better prepared we will be to manage
for the species in the future. Current biogeographic research suggests that if we
can understand the dynamics of species’ distributions, then endangered species’
management, fianslomtions, and other conservation management activities
would benefit considerably (Brown 1984, Lomolino and Channell 1995, Wolf et
al. 1996). My thesis research attempts to clarify the distribution enigma of
species response to range location by quantifying the effect of land cover change
at the core and periphery of a species’ range.

My research question was developed in response to contradicting
literature about the patterns of species decline in biogeography and ecology

journals (Griffith et al. 1989, Lomolino and Channell 1995, Wolf et al. 1996,

Lomolino and Channell 1997). A species occupies an area or region referred to
as its geographic range, and some general patterns have been observed (Brown
and Lomolino 1998). Species tend to have highest abundance near the center of
their geographic range, and abundance usually decreases gradually toward the
periphery (Hengeveld and Haeck 1982, Brown 1984). The area of high density,
which is theoretically located in the center of a species’ geographic range, is
referred to as the core. The outer limit of a species range, where population
density is the lowest, is called the periphery. For the purpose of my study, the
core is defined as the geographic region with the highest density (i.e., the upper
quartile of my sample routes). The terms high-abundance and core will be used
interchangeably. The periphery in this study is the geographic region with the
lowest density (i.e., the bottom quartile of my sample routes). The terms low-
abundance and periphery will be used interchangeably.

Based on general theories of species distribution, species decline should
be initiated in the outer extent of its range where populations are low and suitable
habitat is minimal (Lomolino and Channell 1995). Yet, this pattern is not routinely
exhibited.

Many species have a present-day range found in the peripheral extent of
their historic range (i.e., the geographical range exhibited by a species in the
past; Lomolino and Channell 1995). On the other hand, translocations and re-
introduction sites of species are found to be more successful if they are located
in the core of the species’ historic range (Griffith et al. 1989, Wolf et al. 1996).

The inconsistencies between translocation success and patterns of range

collapse could lead to management confusion, which would be reduced if we
could identify areas where species are more susceptible to decline. Research
conducted at the landscape-scale has demonstrated that land cover change and
degradation of habitat are influential in promoting species decline (Wilcox and
Murphy 1985, Flather and Sauer1996, Bohning-Gaese 1997). However, the
effect of land cover change on populations at the periphery versus the core of

their range has not, to my knowledge, been studied.

1 .2 Statement of Problem

I will evaluate the effect of land cover change on the abundance of the
red-headed woodpecker (Melanerpes erythrocephalus) at the periphery and core
of its range. The red-headed woodpecker was selected for my study because its
population has been declining, population data are readily available, multi-
temporal land cover data are available throughout its range, and its range
boundaries remain relatively constant year round (Short 1982, Page 1996, Sauer

et al. 2001).

My study attempts to answer the following two questions:

1) What land cover changes are significantly related to red-headed woodpecker
population change?

2) Are red-headed woodpecker populations more sensitive to land cover
changes in low-abundance regions of their range than in high abundance
regions?

 

Based on the central patterns of species distribution, I hypothesize that
peripheral red-headed woodpecker populations will be more sensitive to land
cover change than core populations. Therefore, for routes that experience
similar land cover changes, I expect that the mean population change in the

periphery will be significantly different than the mean change in the core. The

null hypothesis (Ho) that will be tested states:

Ho: (mean change along low-abundance routes) = (mean change along

high-abundance routes)

The alternative hypothesis (Ha) that will be tested states:

Ha; (mean change along low-abundance routes) <> (mean change along

high-abundance routes)

If a species is more vulnerable to extinction at the furthest extent of its range,
then habitat disturbances, competition, or human activity may cause anomalous
patterns of species decline, and restoration efforts should be concentrated in the
core of a species range. However, if species decline is not related to geographic
range location, then managers should consider all areas of the range when

selecting restoration sites.

1.3 Overview of the Thesis
Chapter two gives a detailed account of the relevant literature in this topic
area. The discussion will begin broadly, focusing on current paradigms relating

to species distribution and the relationship of these paradigms to species

reintroduction. Next, I will discuss patterns of species decline, as they relate to
range collapse and reintroduction strategies. Chapter two will conclude with a
discussion of a key data source used in this study, the Breeding Bird Survey and
the study species, the red-headed woodpecker.

Chapter three presents the methods used to answer the research
question. The chapter is organized around the bird data, land cover data, and
statistical analysis. The bird data section describes population change
determination and route selection. The land cover data subchapter includes
details about data processing, land cover classification, the geographic
information system layers used, processes that yield the final land cover output,
and calculating the change in landscape. The statistical analysis section
provides details about statistical methods used to determine which land cover
changes were significantly related to bird population change, determination of
similar routes in the core and periphery, and analysis of core/peripheral
population response to land cover change.

Chapter four presents the results of the analyses. A detailed discussion
and interpretation of the results follows in chapter five, which concludes with a
summary of significant findings and implications of this study as well as some

suggestions for future research.

CHAPTER 2. LITERATURE REVIEW

2.1 Spatial Distribution of Species

Species exhin unique geographic distributions, which are dynamic over
space and time. The spatial distributions of species demonstrate some general
patterns. First, most species are not evenly distributed throughout their range.
Hengeveld and Haeck (1982) examined the spatial distribution of abundance
within the geographic ranges of selected species in northwest Europe. Their
goal was to determine whether species density was homogenous or
heterogeneous across a range. Through the use of grid sampling, data were
collected throughout the selected species’ ranges and processed in an indirect
interpolation model. All populations were found to be heterogeneous across their
range, and a tendency for the highest population densities to be located in the
center and the lowest densities to be at the limits of the geographic range was
revealed (Hengeveld and Haeck 1982).

Brown’s (1984) research findings paralleled those of Hengeveld and
Haeck (1982): density was greatest near the center of the range, and population
abundance decreased gradually toward the peripheral edge. Brown synthesized
studies conducted on the relationship between abundance and distribution, and
developed a theory to explain the general patterns consistent with the research.

In addition to the general patterns of distribution within a species’
geographic range, it was found that the spatial distribution of species is positively

correlated with abundance (Brown 1984). Thus, species with high average

abundance will have a larger geographic range compared to those with low
average abundance. These patterns of distribution have some exceptions:
anomalous distributions occur at sharp environmental contrasts and sporadic
distributions are caused by environmental patchiness (Brown 1984). However, in
landscapes with little environmental patchiness or abrupt environmental
obstacles, species abundance should be greatest at the core and gradually
decline to the outer limits of the geographic range.

Environmental limiting factors affect geographic range boundaries and
population density. In locales with good habitat, so-called source populations
occur (i.e., populations in which the birth rate exceeds the death rate; Brown and
Lomolino 1998). The habitat at the outer limits of a species’ range is of marginal
quality. Thus, it would seem that sink populations (i.e., the death rate exceeds
the birth rate) would be more common at the peripheral limits of a geographic
range.

Hengeveld and Haeck (1982) expected the amount of favorable habitat to
decrease at the margins of the range, but anticipated that the density of species
would remain the same at optimal habitat locations in the margin and central
range locations. This outcome was not upheld; independent of scale and habitat
quality, species became more rare as they approached the range margins
(Hengeveld and Haeck 1982). Thus, species responded differently to habitat
conditions throughout their range.

Based on the central patterns within a single species’ distribution,

population decline should first appear at the edge of the range, where sink

populations are prevalent, population density is low, and habitat is marginal
(Lomolino and Channell 1995). However, this pattern of range collapse is not

exhibited in relic species.

2.2 Species Reintroduction

Lomolino and Channell (1995) reviewed terrestrial mammals whose
present distributions were reduced to less than 25 percent of their historic range.
They argued that if patterns of range collapse are related to the occurrences of
sink populations, then populations of endangered species should be located in
the center of their historic range. They used a geographic information system to
divide the historic range into core and peripheral regions. Peripheral regions
were determined by buffering the edges of the range until the interior area
equaled that of the outer. Analysis was completed through the use of an index to
compare present range locations to their historic locations. The authors found
that 23 of the 31 species analyzed were located in the periphery of their historic
range (Lomolino and Channell 1995).

Based on this discovery, the consideration of peripheral sites as potential
reintroduction locations was recommended (Lomolino and Channell 1995). Yet,
analyses of successful wildlife translocations have revealed the opposite.

Griffith et al. (1989) surveyed professionals involved in wildlife
translocations to uncover patterns of success within these programs. Seven
variables were correlated with translocation success, one of which was the

location of the release site in relation to the historic range of the species.

Translocations are considered successful if they result in self-sustaining
populations (Griffith et al. 1989). Releases in the core of the species’ historic
range were 76 percent successful. Conversely, releases in peripheral historic
range locations were only 48 percent successful. Stepwise regression of the
survey variables revealed that location of release was statistically significant in
predicting translocation success (Griffith et al. 1989).

This study was revisited in 1996 to verify the variables that Griffith et al.
(1989) attributed to translocation success (Wolf et al. 1996). A series of
questionnaires were issued to people involved with translocation programs, and
multiple regression analysis was used to determine the variables that were
significant predictors of translocation success. The results revealed that both
mammals and birds were sensitive to release location within the range. Animals
relocated in the core of their historic range had a significantly higher success
rate. Wolf et al. (1996) claimed that their findings contradict the conclusions of
Lomolino and Channell (1995), i.e., relocating a species in the periphery of their
historic range is significantly related to translocation failure. They also pointed
out that patterns of decline are more complex than a simple distinction between

core and periphery.

2.3 Spatial Patterns of Decline
Complex patterns of decline were revealed in a study of two species of
warblers that are declining over their breeding range, and temporal and spatial

abundance maps were created to explore patterns of population change (Villard

and Maurer 1996). Areas of decline were scattered throughout the geographic
range and were inconsistent in temporal analyses. The inconsistencies in the
patterns of decline demonstrate the extreme difficulty in applying general
patterns to spatially and temporally dynamic processes. Species abundance
fluctuates in response to external environmental stimuli. Perhaps looking at
within-habitat interactions between abundance and environmental conditions can
clarify the conflicting observations of species decline. While many researchers
have looked at general patterns of species decline, very few have linked these
patterns to the underlying processes causing decline.

ln landscape-level studies, environmental conditions such as temperature,
vegetation, forest size, and habitat heterogeneity are often described as factors
limiting species abundance (Smith 1974, Williams 1975, Bohning-Gaese 1997).
For example, temperature was isolated as a limiting factor for pika abundance
(Smith 1974). High-altitude pikas are more likely to have a greater home range
and abundance than low-altitude populations because average daily
temperatures were not limiting activity and mobility (Smith 1974).

Most species, however, have more than one environmental variable
affecting their abundance (Hutchinson 1957). For example, a study evaluated
habitat utilization of woodpeckers in central Illinois, and each species of
woodpecker was found to occupy a specific niche (Williams 1975). Foraging
techniques, nesting strategies, and utilization of certain height classes of

vegetation were some of the many strategies used to reduce competition and

10

permit co-existence between species. Thus, the woodpeckers utilized these
strategies to fulfill their multiple life requirements.

If one of the variables significantly related to a species’ survival is no
longer available, then the species must adapt, move, or go extinct (Brown and
Lomolino 1998). Therefore, a decline in species abundance is a direct result of
one or many changing environmental conditions. Our understanding of the
general patterns of decline would be enhanced if we could pinpoint the
underlying reasons for the decline. General patterns of decline should be linked
back to their causes whenever possible.

For example, ecologists have linked habitat loss to the decline in
Neotropical migratory bird abundance. F lather and Sauer (1996) analyzed the
relationship between land use and Neotropical bird abundance at a small-scale
and found Neotropical migrant abundance was correlated with many aspects of
landscape structure. Birds were more abundant in areas of forests and wetlands,
especially areas with larger forest patch size and fewer edges and where forests
were dispersed throughout the environment (Flather and Saurer 1996).
However, the land cover data were for one time period; two time periods of land
cover are necessary for a true temporal change analysis. Nevertheless, this
study attempted to connect patterns of decline to the processes underlying the

pattern.

11

2.4 Breeding Bird Survey

Since 1966, bird populations have been monitored annually nation wide
through the Breeding Bird Survey (Fallon et al. 2000). The Breeding Bird Survey
(BBS) has unveiled that the population of many species of birds, including the
red-headed woodpecker, have declined since the start of the survey
(Cunningham and Saigo 1995, Sauer et al. 2001). The survey is conducted in
early June during the peak breeding season throughout North America (Fallon et
al. 2000). Birds call more often during the breeding season, thus there is more
likelihood of observing (hearing) a bird during this time period. BBS routes
contain 50 stop locations at approximately half-mile (800 m) intervals; the same
route and corresponding stop locations are often surveyed yearly. The exception
to the half-mile stop interval occurs if there are hazards or excessive noise at a
location (Adams 2000). Birds within a quarter-mile radius of the stop location are
recorded (Fallon et al. 2000). Birds are recorded by sight and sound, with each
individual bird recorded only once. Volunteers follow BBS procedures
consistently every year.

Breeding Bird Survey data are collected by experienced, volunteer birders.
Volunteers are required to know the song, calls, and visual identification of all
avian species that may be encountered (Sauer et al. 2001). The observers
submit to BBS headquarters their annual roadside counts and document the
weather conditions, number of vehicles encountered, wind speeds, and

excessive noise sources (Sauer et al. 2001).

12

Some concerns were initially raised regarding data obtained from the BBS
including the appropriateness of conducting a survey along a road, excessive
noise impeding an observers’ surveying, and the skill differences among
observers. The rationale of conducting bird surveys along roads, as opposed to
off road, has been explored (Keller and Scallan 1999). Traditionally, human-
induced changes in habitat occur near roadways. The worry is that these habitat
changes are not representative of what is occurring away from transportation
networks. However, the study discovered that habitat changes along roadsides
matched changes occurring in the landscape within 1600 meters of BBS routes,
but changes in urbanized areas were more likely near the road (Keller and
Scallan 1999). BBS route locations avoid urban areas, and road surveys are
easier to collect than off-road surveys, thus the use of roadside surveys seems
appropriate.

Noises associated with traffic, weather, and residential living can hinder an
observer’s ability to hear and accurately record birds. Vehicles that pass during
each three-minute stop are recorded on BBS data sheets, and the survey is
conducted under specific weather guidelines (Fallon et al. 2000). Counts are
conducted on days with little wind, no precipitation, and good visibility, with
weather conditions recorded on the field sheets. Since factors impacting bird
observations are recorded, survey sites with excessive noise can be eliminated
from subsequent analyses.

Changes in observers on BBS routes can greatly influence the perception

of bird populations (Link et al. 1994). BBS routes are sampled by thousands of

13

volunteers across North America. The observers may vary in skills and abilities,
thus resulting in inaccurate and inconsistent counts between routes and survey
years. It was found that changes in bird counts were significantly greater when
observers differed between years (Link et al. 1994). The impacts of this change
can result in inaccurate interpretations of changes in individual species along a
route. To combat this problem, the BBS includes unique observer numbers
associated with each route and year. A change in observer can be accounted for
in an analysis. For example, the year immediately following an observer change
may not be included in the calculation of BBS trends on the BBS web site (Sauer
et al. 2001). So, researchers are made aware of changes in observers and can
make adjustments if they think it is necessary. Even with all the associated and
potential problems, the BBS is still one of the most comprehensive species

monitoring programs in North America (Sauer et al. 2001).

2.5 Red-headed Woodpecker

The red-headed woodpecker was selected for this study because
populations: have been declining, population data are readily available, its range
is located in North America where land cover data are available, and its range
limits remain relatively stable year round (Short 1982, Page 1996, Sauer et al.
2001).

The red-headed woodpecker is an eastern North America species. It’s

range limits (Figure 1) extend from southeast Canada to Florida, and the Atlantic

14

coast to Colorado (Brown et al. 1999). The woodpecker is most abundant in the

central mid-west states of Iowa, Missouri, and Illinois (Sauer et al. 2001).

 

Red-headed Woodpecker Range,

 

,, 1 ,, North America
* .3. K 7.0"
l’ '5; I u i Y {Elia is: a}._ $77...
If V. we. "I - $993 W -\T§E‘:’
4“” 4 I . .3»: 41.. “'71: EN; V's; :§: at $0,?
g E. ~".'-‘r‘ ‘ 'g a”...
“as was, i »; ‘s .2. s fi”% 4,;
',‘ {’39— ‘9' 7% 7 ~. E:
a: :_ , >4" . .g E“

 

mm
www.mnjzoor

 

 

Figure 1. Red-Headed Woodpecker Range.

Throughout its range, the red-headed woodpecker inhabits open
deciduous woodlots, wooded field edges, riparian forests, and urban parks (Short
1982, Robert 1989, Winkler et al. 1995, Page 1996, Brown et al. 1999). It
resides in habitats containing dead trees for nesting and open areas for foraging
insects. The red-headed woodpecker is an omnivorous species, eating insects,
berries, nuts, corn, sap, seeds, and other bird's eggs (Short 1982, Winkler et al.

1995, Page 1996). It occupies a home range size of 0.04 to 2.0 hectares

 

(Winkler et al. 1995). The woodpecker is monogamous and generally has a
clutch size of four to seven eggs and produces up to two broods per year; both
male and female excavate the cavity and tend to the young (Short 1982, Ehrlich
et al. 1992, Page 1996). The red-headed woodpecker can be migratory, mostly
in northern range locations, but range limits remain fairly constant year round
(Short 1982, Page 1996).

BBS data have revealed that red-headed woodpecker populations have
declined at a rate of 2.4% from 1966 to 1998 in the United States (Sauer et al.
2001). During the same time period, the red-headed woodpecker has
experienced declines in 23 of the 35 States within its range (Sauer et al. 2001).
The woodpecker is considered imperiled in six states located on the outer limits
of its breeding range: Connecticut, Delaware, New Hampshire, New Mexico,
Rhode Island, and Utah (Page 1996).

Current literature suggests that declines in red-headed woodpecker
populations may be attributed to a loss in nesting habitat due to clear cutting,
removal of dead trees in urban and residential areas, and cleaner agricultural
practices such as monoculture fields and removal of hedgerows (Kilham 1958,
Page 1966, Robert 1989, Brown et al. 1999). In addition, competition for nest
cavities with the exotic European starling (Stumus vulgan's), which often
overtakes newly excavated nest cavities for their own broods, has contributed to

the woodpeckers decline (lngold 1989, Page 1996, Brown et al. 1999).

16

 

CHAPTER 3. METHODS

3.1 Overview

Based on the habitat requirements and suggested factors leading to red-
headed woodpecker decline, I calculated land cover characteristics that might
impact woodpecker abundance in order to determine the land cover variables
that are statistically significant to red-headed woodpecker abundance.

Red-headed woodpecker data were obtained from the Breeding Bird
Survey, and land cover data were obtained from the United States Geological
Survey. Figure 2 is a cartographic model of the methods used in my research
(Figures in this thesis, with the exception of figure 2 and 3, are presented in
color). Figure 2a shows the steps used to analyze red-headed woodpecker data
obtained from the BBS. The number of sample routes was determined through a
power analysis, and a random number table was used to select individual routes.
Core and peripheral routes were distinguished by ranking routes based on
average populations. Routes were ranked into quartiles: definite core (i.e., with
largest numbers of red-headed woodpeckers detected), intermediate core,
intermediate periphery, and definite periphery (i.e., with the fewest red-headed
woodpeckers detected).

Land cover data (Figure 2b and Figure 2c) were resampled to obtain the
same resolution and land cover attributes (based on red-headed woodpecker
biological requirements) and were analyzed in a geographic information system
(GIS). Trends in red-headed woodpecker populations from 1978-1993 were

calculated, and a multiple regression analysis was used to determine which land

17

cover attributes were significantly related to red-headed woodpecker population
change. Figure 2d shows the steps used to analyze the data. A hierarchical
clustering analysis was used to isolate routes that were similar in the land cover
attributes related to woodpecker abundance. Finally, an analysis of variance test
was used on clustered routes to demonstrate whether or not the red-headed
woodpecker is more sensitive to land cover change at high-abundance locations

or low-abundance range locations.

18

 

8.5353580

 

 

 

 

 

 

 
 

89.32.

«87%..
31¢

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

388 385 Al
\ 5:30 $82
I 3530 / gases... as.
2.2.332 8.5. 3>e<
£338.. 5833. a... 8a m. 6 I

 

 

 

 

 

 

 

 

 

 

8535mm

 

 

 

 

 

 

 

38 Em "sec: 2.528an .8 2:2“.
.38: 96283.8 a 26ml

19

 

 

 

.8 23m 5 RES

 

 

 

 

 

gnaw—.0
360.33 on?

 

 

   

.555 ES .
Saw 35$

25
o. 5250

    
 
   

    

 
 

2.55
528w E35 can“.

 
   

833.0330
23:32 . .

  
  

 

T

 

22
Buses”.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T

823.5 5.:

.235 so... 2

    

69885.

26

 
    

Gamma—o
ages!

Ouz_ .

2 52.8

   
  

a. 2:9". sea

 

 

 

 

   
 

 

55.82

$8~
i4 3.8238
5392332

 

 

 

 

2».
Eu a8:

 

 

580.93.. ”SF

 

Sun .33 En.— 22 gave: 2.395050 .nN 2:9“.

.G.E8V N 2:9“.

 

 

20

 

 

32.85338 .8.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

go
.9ng 8a..
333
8.886 1.92 fifim
8386220
.2823 8:980
088.. mom §>e< ”8.983
8.98.23 Al Sum 80 .353.
.528 82 $3 58 02 £88
SEES. .6592 88.82
gm E 8w
5,8; Egg 2 All a £88.. 5.8
86983 «.8
8a 3.880 25:38,.
9.252 2».
88: 2.8.... 39m
8308 $8~ Alli So 8252 8955358 2.. 95:8 mom:
.RuEEEaB .98.!!88. All e Al “8:23.88 Al.
€212 :83 83923 n8.

 

 

 

 

 

 

 

 

 

Sun 330 Ea.— 89 ".30: 02352.30 .oN 2:2“—
.AEc8V ~ 23:

 

21

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.3 A1 385......
an: .930 £839..
.533 Sam 8 go
8mm .3.
85953 31¢
3% A1 38.28220 All 9132 All 8553.».
Sea a 8.3... .9823 385:5 8.8281 £332 a .98.?! 5.;
685:3 838. 3:8 195m mom
a 2...... so...
.8.
83.83.20
3.60.93 89
8...“... .096 w:
5:86.03 8.60.9.3
3.3.2... £2.82 a ea... 22..

._. 6255 _

/ 3mm... \ .e.

855 .2699! 8.3 .a 8 855
6 m .388. £58.. m ? 8.2 33.9 .2865. o
.83. 22 8.60.9.3 «8.

 

 

 

 

 

 

 

 

 

 

 

«7.222 ".23: 2.35950 .3 2:2“.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

$2.8. u 8.5....

 

 

22

_._! -flL— _~_____._._.. __._..‘ --

3.2 Data Compatibility Issues

Three main GIS sources were used in this research. The BBS routes
were in vector format and were downloaded from the USGS (2001); 1978 land
cover data were also in vector format and were obtained from the USGS/EPA
(1998); and the 1993 land cover data were in raster format and were available
from the USGS (2001). Compatibility issues are almost certain when using data
from varying sources and in different formats. The BBS routes and the 1978 land
cover data were unprojected in decimal degree latitude and longitude. However,
1993 land cover data were projected in Albers Conical Equal Area, based in
meters. These two geographic formats are not compatible. To overcome this, I
projected all my GIS data into the Albers Conical Equal Area projection because
it does not distort area. This was important because many area measurements
were critical and necessary for my research. Additionally, the Albers projection is
based in meters, making the calculation of the land cover metrics less
complicated (Snyder 1993).

I chose to keep the land cover data in raster format and to retain the BBS
buffers as vector data. The benefits of raster data in change analysis far
outweigh vector data. Spatial analysis, filtering, and mathematical modeling are
easy due to the simple shape of the pixels in raster analysis (Burrough and
McDonnell 1998). The raster grids were clipped to the polygon- represented
route buffers so that the 0.25-mile survey radius was upheld. If the center of the
grid cell fell within the survey radius, then it was included in the output; this

prevented sub-pixel fragments (ERSI 2000).

23

An additional data incompatibility was the resolution of the 1978 and 1993
land cover data. The 1978 data were determined from aerial photographs using
a 200 x 200 meter minimum mapping unit. The 1993 data, in contrast, were
classified from Landsat TM satellite imagery, which has a 30-meter resolution.
Since my research attempts to identify changes in the landscape, it is imperative
that the land cover data sets be at comparable resolutions. Using the 30-meter
data with the ZOO-meter data will result in inaccurate representations of
occurrences in the landscape, such as drastic increases in fragmentation in the
1993 data. To combat this problem, I degraded the 1993 data to a resolution of
approximately 200 meters. This was accomplished by using a non-overlapping,
majority filter. The majority filter uses a specified window size and reassigns all
the pixels within that window with the modal pixel value. This preprocessing
assured that the 1978 and 1993 data were at comparable resolutions for change
analysis. However, degrading the resolution of the 1993 data created additional
problems.

The non-overlapping window scrolls through a scene and reclassifies pixel
values; reassigned values depend upon the placement of the window. For
example, in the edges of the homogenous landscape of Figure 3, representing a
coastline, window 1 would classify the aggregated pixel as water, but window
placement one row over (window 2), would reclassify as sand. However, the
land cover data are nominal, which are not suitable for an averaging filter, and
the majority filter finds the modal value in the window, which is appropriate when

aggregating to a coarser resolution. Thus, a non-overlapping majority filter was

24

an efficient and feasible solution to my resolution issue, and determined to be

adequate for the scope of this research.

 

 

 

 

 

 

 

Moving Window 1 Moving Window 2
2 2 1 2 2 1
2 2 1 2 2 1
2 1 1 2 1 1
5 water pixels 2 water pixels
1 sand pixel 4 sand pixels
Classified as Water Classified as Sand
1 = Sand 2 = Water

 

Figure 3. Majority Filter Example.

Finally, the 1978 land cover data were obtained from aerial photographs
and 1993 land cover data were obtained from satellite imagery. The differences
in source-document format may pose classification errors or discrepancies. For
example, reflectance from satellite imagery may be difficult to interpret and pixels
classified as residential from aerial photography may be classified as forest from
satellite image interpretation. The 1978 data were classified into 37 different
classes; the 1993 data were classified into 21 classes. I regrouped each year’s
classification scheme into ten classes. The classes I chose are easy to interpret
from both aerial photography and satellite images. Thus, the chances for
inaccurate representation of land cover due to the origins of the two land cover

data sets were minimized.

25

3.3 Data Preparation

3.3.1 Bird Data

I attempted to minimize the potential biases in the BBS by 1.) sampling
from continuously surveyed routes during my study years; 2.) assessing selected
routes for excessive noise or traffic; and 3.) accounting for observer changes
during my study period.

A power analysis was used to determine the number of BBS routes
necessary to provide statistically significant results. The power analysis yielded
a value of 45 routes to determine variation in bird abundance between core and
periphery, and a maximum value of 200 routes to determine variation between
the study years, 1978-1993, with a 0.9 alpha and beta level.

I decided to use 80 sample routes due to the practicality of assessing
landscape attributes. A total number of 80 routes is in between the minimum and
maximum sample values obtained from the power analysis, and l determined that
80 routes were the most data that could be processed within the time and
computing constraints I had. Also, 80 routes should have been enough to
determine the major changes in both landscape attributes and bird abundance.

The BBS data for red—headed woodpecker from 1977 to 1994 were
downloaded from the BBS website (USGS and Patuxent Wildlife Research
Center 2001). These data were brought into Microsoft Access (2000), and routes
that were not surveyed continuously during my study period were deleted from
the candidate pool. The remaining BBS data were brought into Microsoft Excel

(2000) and average red-headed woodpecker abundance was calculated for the

26

years between 1977 and 1994 for each potential sample route. Routes were
numbered based on average abundance; the route with the highest average
abundance was given a value of one.

The Excel file was saved as a database file (.dbf) and imported into ESRI
ArcVIew 3.2 (2000) where it was joined to the BBS route shapefile that had been
downloaded from the USGS National Atlas (2001). In an attempt to homogenize
the land cover types occurring on the BBS routes, I only included routes
occurring on glaciated landscapes. Glaciated landscapes provided some degree
of normalization for topography, soils, and climate. Land cover changes would
be more similar in Minnesota and Michigan, for example, than comparing
changes occurring in Minnesota and Florida (Harpstead et al. 1980). The Major
Land Resource Areas (MLRA) from the Natural Resources Conservation Service
(NRCS) were downloaded from the NRCS website (2001) as an Arclnfo
exchange file (.e00) and converted into a shapefile using the ArcVIew Import 71
program (ESRI 2000). The MLRA shapefile was queried for land areas
containing glaciated landscapes using the landscape name field. The selected
landscapes were converted into a shapefile. The BBS route shapefile was then
clipped to the glaciated landscape shapefile using the Geoprocessing Wizard
extension in Archew (ESRI 2000). The remaining BBS routes were those that
were surveyed every year between 1978 and 1993 and located in glaciated
landscapes.

The ranked routes on glaciated landscapes were divided into four equal

groups (i.e., quartiles), based on average abundance from 1977 to 1994. These

27

four groups represent definite core (i.e., the uppermost quartile), transition core,

transition periphery, and definite periphery (i.e., the lowermost quartile), and were

labeled 1 through 4 correspondingly.

Twenty routes were selected from each quartile through a random

numbers table and converted to a new shapefile. Figure 4 shows the BBS

sample routes selected for this study.

 

Sample Route Locations

0 Routes

 

El

MapExtent
. a?
O . .0

[:1

 

 

Figure 4. Sample Route Locations.

Figure 5 is the sample routes cartographically represented by their quartile

ranking. Most of the routes categorized as a one, or core, are located in the

center of the bird's range. The peripheral routes, or low-abundance, categorized

as a four, are located on the perimeter of the red-headed woodpecker’s range.

28

This pattern was expected based on the research findings of Hengeveld and

Haeck (1982) and Brown (1984).

 

Sample BBS Routes: RW‘“

Quartile Ranking ' 1iCore
“ 2: Transition Core

0 3: Transition Periphery

 

06) OCH“ 1 /
O . e ' Q

O 4: Periphery
O
. . O .
‘ é) :13
g Q - (f) (T: . t
@ 9 8 .0050 O ' O ” e
e 0 0 g?) e
. ,
i e . o. z 0 .7
T .0 .63 o O ’L‘
i . 8 Q
l
l
E
l

 

L-- f” -- -----_, FJ

Figure 5. Sample BBS Routes Quartile Ranking.

I created a 0.25-mile buffer around the linear routes to represent the area
surveyed by the BBS. BBS observers record birds seen and heard within 0.25-
miles of survey stops (Sauer et al. 2001).

To calculate change in red-headed woodpecker abundance from 1978 to
1993, the sample routes table was exported and opened in MS Excel (2000).
Dividing the 1993 abundance by 1978 abundance and taking the logarithm (to

normalize data) of this value and adding a correction factor of 1.285, to ensure

29

positive values, calculated a change ratio. However, I had to correct my data for
BBS observer variation, due to problems with BBS data due to observer
differences (Link et al. 1994). l accomplished this by calculating a mean count
for each route over the entire study period. I, then, calculated a mean count for
each individual observer on the route. A correction factor was obtained by
subtracting the observer mean from the overall mean, and this correction factor
was applied for every year of observation by that individual. The average count
of three years around my land cover time periods was calculated to obtain a
more representative measure of bird abundance surrounding the land cover year.
For example, to obtain abundance for 1978, l averaged the observer-corrected,
red-headed woodpecker counts from 1977, 1978, and 1979. Upon calculating an
observer corrected average surrounding my two time periods, I calculated the
logarithm ratio of change from 1978 to 1993. The bird abundance change was
exported as a .dbf file, brought into ArcVIew (ESRI 2000), and joined to the

corresponding route shapefile to be used in the final analysis.

3.3.2 Land Cover Data

Land cover data from 1978 and 1993 were reclassified, through
aggregation, into ten broad classes based, primarily, on the red-headed
woodpeckers habitat requirements and variables that the literature has
suggested leads to a change in abundance: deciduous forest, agriculture,

residential, urban (commercial and industrial areas), transitional, wetland, water,

30

grassland, evergreen forest, and mixed forest. Table 1 outlines the land cover

classes I regrouped the existing classes into and the corresponding justifications.

Table 1. Land Cover Classification and Justification.

 

Land Cover Justification
Agriculture Food source (Short 1982, Page 1996)
Deciduous Forest Prefers deciduous forest (Page 1996,

Williams 1975, Short 1982, Winkler et al. 1995)
Evergreen Forest Not preferred (Page 1996)

 

 

Grassland Open area for foraging (Page 1996)

Mixed Forest Can't be further classified

Residential Can thrive in parks and old residential
neighborhoods (Winkler et al. 1995, Page 1996)

Transitional Can't be further classified.

Urban Loss of birds due to urban clean-up

(removal of dead trees), competition, and destruction
of habitat (Page 1996)

Water Water source

Wetland Utilizes lowland areas for water and food
sources (Short 1982, Winkler et al. 1995, Pa e 1996

 

After reclassifying the land cover classes, six landscape metrics were
calculated based on these classes. Based on the literature, the following metrics
are thought to impact red-headed woodpecker abundance: area, mean patch
size, total edge, edge density, mean patch fractal dimension, and interspersion
and juxtaposition measures. Table 2 outlines the calculations used in my

research and a summary of what each attribute calculates.

31

Table 2. Land Cover Measures and Description.

 

 

Land Cover Measure What it Measures

Area Sum of area for each land cover class
Mean patch size Mean area for each land cover class

Total edge Sum of perimeter of each land cover class
Edge density Sum of perimeterl sum of area

 

Mean patch fractal dimension Measures shape or complexity; fragmentation
lnterspersion and Juxtaposition The interspersion of the land class compared
To other land classes on the BBS route

 

 

I will use area and mean patch size measurements to determine the
impact that a loss or gain in a land cover may have on the species. Edge
measurements, such as total edge and edge density, will isolate fragmentation
effects at the patch level on red-headed woodpecker abundance. The mean
patch fractal dimension measurement returns an index value between 1 and 2,
where a measure of one indicates a simple shape and values approaching 2
reveal a more complex shape (Elkie et al. 1999). Mean patch fractal dimension
evaluates shape complexity and reveals the type of fragmentation that is
occurring providing more insight than a simple edge density or total edge
measurement. Finally, the interspersion and juxtaposition measurement is the
percentage of interspersion of the land cover types with other land cover types
found along the BBS route. A value of 0 percent indicates that the land cover
type is adjacent to only one other patch type and a value of 100 percent occurs
when the land cover type is equally adjacent to all patch types found on the BBS
route (Elkie et al. 1999). The interspersion and juxtaposition measurement is a
good indicator of the fragmentation occurring among the land cover type as a

group, and not just individual patches.

32

3.3.2.1 1978 Land Cover

Land cover data from 1978 were obtained from the United States
Environmental Protection Agency (USEPA and USGS 1998). These data were
originally collected by the USGS but were subsequently converted into digital
format by the EPA. The USGS classified the landscape from aerial photography
using a 10-acre minimum mapping unit (200—meter resolution).

Downloaded quadrangles from corresponding BBS sample routes were
brought into Arc/INFO (ESRI 2000). The 1978 land cover coverage was
reprojected into the Albers Equal Area Conic projection and converted into a grid
file format at a 30—meter grid cell. The original data were classified using the
Anderson et al. (1976) system at level 2. For the purposes of my research, I
recoded the classification down to ten broad classes. Table 3 shows the original
1978 land classes and the corresponding new classification used for this
research. All reclassification was completed in ArcView (ESRI 2000). The land
cover data were clipped to the 80 sample BBS routes each with a 0.25-mile
buffer using the MILA Grid Utilities Extension (Guissard 1999). The result is the
land cover in 1978 surveyed by the BBS for the presence of songbirds. Finally,
the Patch Analyst Extension (Rempel 2000) for ArcView (ESRI 2000) was used
to calculate the selected landscape metrics for each of the ten land classes.

The results were exported and brought into MS Excel (2000).

33

Table 3. Aggrggation of 1978 Land Cover Data.

 

 

flinal Classification New Classificatio
Residential Residential
Commercial and services Urban
Industrial Urban
Mixed urban Urban
Other urban Urban
Cropland and pasture Agriculture
Orchards, groves, vineyards Agriculture
Confined feeding Operations Urban
Other agricultural Agriculture
Herbaceous rangeland Grassland
Mixed rangeland Grassland
Deciduous forest Deciduous Forest
Evergreen forest Evergreen Forest
Mixed forest Mixed Forest
Streams and canals Water
Lakes Water
Reservoirs Water
Bays and estuaries Water
Forested wetland Wetland
Nonforested wetland Wetland
Strip mines, quarries, gravel pits Urban
Transitional Areas Transitional

 

3.3.2.2 1993 Land Cover

 

Land cover from 1993 was obtained from the United States Geological
Survey (2001). The National Land Cover Data (NLCD) were classified from
Landsat TM imagery, which has a 30-meter resolution.

The 1993 NLCD can be downloaded by state (USGS 2001). Each state
from the corresponding survey routes were downloaded, extracted, brought into
ERDAS Imagine (1999), and converted to an image file. The image file was

opened in ArcView (ESRI 2000) and converted to grid format using the Spatial

Analyst extension (ESRI 2000). For the purposes of my research, l recoded the
classification (using ArcView) down to ten broad classes. Table 4 shows the
original 1993 land cover classes and the corresponding new classification for this
research. The grid for each state was processed in Arc/INFO Grid to aggregate
the 30-meter pixels into approximately ZOO-meter blocks for compatibility with the
- 1978 data. I completed this task using the block majority function with a 7 by 7
non-overlapping window. Block majority takes the most frequently occurring
pixel value in the specified window and converts all pixels within the window to
that value. If there is a tie, and two values are equally abundant, then the filter
classifies the pixels as “no data.” A 7 by 7 pixel window for aggregation was
selected to create a coarse resolution to compare with the 1978 data (30-meter *
7 =210—meter). The 200-meter, 1993 land cover data were brought into ArcView
and clipped to the sample routes with a 0.25 mile buffer using the MILA Grid
Utilities extension (Guissard 1999). The Manual Grid Editor extension (Luijten
2001) was used to add land cover values to any “no data pixels.” The non-
aggregated NLCD data and 1-km AVHRR imagery (EROS and USGS 2001)
were used as reference, and no data pixels were, ultimately, assigned a value

based on AVHRR classification.

35

Table 4. Aggregation of 1993 Land Cover
Data.

Original Classification New Classification

 

 

Open Water Water

Low Residential Residential

High Residential Residential
Commercial Urban

Strip Mines, gravel pits Urban
Transitional Transitional
Deciduous Forest Deciduous Forest
Evergreen Forest Evergreen Forest
Mixed Forest Mixed Forest
Orchards, vineyards Agriculture
Grassland Grassland
Pasture, hay Agriculture

Row crops Agriculture

Small grain Agriculture
Fallow Agriculture
Urban grasses Urban

Woody wetlands Wetland
Herbaceous wetlands Wetland

 

 

Upon identification and classification of all “no data pixels”, the Patch
Analyst Extension (Rempel 2000) in ArcView (ESRI 2000) was used to calculate
the selected landscape metrics for each of the ten land classes. The results were

exported and brought into MS Excel (2000).

3.4 Analysis

The 1978 and 1993 land cover characteristics table for each route were
brought into MS Excel (2000) as separate tables. I subtracted each 1993
variable from the 1978 variable on each route. This provided a raw change

number associated with the land cover measure, retaining the same

36

measurement units as the original data. Change data for all routes were
contained in one table. Since land cover change may be negative, the absolute
value of the lowest negative value for each variable plus one was added to every
entry. This ensured that the values were scaled, yet positive, with the lowest
value in each variable expressed as one. This was necessary in an attempt to
normalize the data for further analyses by taking the base ten logarithm of all
values; the log", transform is only possible on positive integers. A normal
distribution of data is necessary for many statistical techniques, and taking the
logarithm can rescale the original values to obtain a normal distribution (Diekhoff
1992). I copied the new change values for each route and corresponding land
cover characteristics into a new table and exported them as a .dbf file. This
database file was added into ArcView (ESRI 2000) and joined with the red-
headed woodpecker change and sample route shapefile. The complete table of
red-headed woodpecker change and land cover attribute changes for each
sample route was brought into SPSS 11.0 (2001).

I used stepwise multiple regression, with entry F probability less than 0.5
and removal at 0.1, to decipher which variables were significantly related to red-
headed woodpecker abundance change. Stepwise multiple regression was
selected because it minimizes the impact of multicollinearity and removes or
includes variables based on the effect this has on the R square value (Diekhoff
1992). Multicollinean’ty occurs when two or more variables are correlated, thus
inclusion of such variables in the regression equation is redundant, does not add

to the predictive power of the equation, and reduces the statistical significance of

37

R (Diekhoff 1992). Stepwise regression avoids these problems by not selecting
strongly correlated variables for inclusion in the equation (Diekhoff 1992). This
method evaluates each variable and includes only those that add to R-square.
Variables can also be removed if their absence does not detract much from the
equation’s predictive power (Diekhoff 1992, SPSS 2001).

The variables found to be significantly related to the change in red-headed
woodpecker abundance were isolated into a separate table. I used a hierarchical
cluster analysis to isolate routes with similar land cover changes, which are
significantly related to red-headed woodpecker change. The hierarchical
clustering routine isolates relatively homogenous cases based on distance
measures (SPSS 2001). Several clustering methods were attempted and the
furthest neighbor method was selected as the clustering strategy because it
provided distinct clusters with adequate numbers of routes for further analysis.
Furthest neighbor uses the span between the least similar cases as the distance
between the element and the cluster (Diekhoff 1992, SPSS 2001). I used the
squared Euclidean distance to measure proximity of cases. This is computed by
taking the sum of the squared differences between the case’s variables. In my
research, the cases are the BBS routes and the variables are the land cover
characteristics significantly related to red-headed woodpecker abundance
change. The result and interpretation of the hierarchical clustering analyses
allows me to group routes with similar land cover changes, which will later be
used in the analysis of differences in abundance changes occurring along high-

populated and low-populated routes.

38

The groups of similar routes were isolated into their own table. The
rankings of routes based on average abundance were used to analyze high-
abundance routes (value of one) and low-abundance routes (value of four) with
the analysis of variance method (ANOVA). ANOVA allows two or more samples
to be compared with a single independent variable, red-headed woodpecker
abundance change, for difference (Sprinthall 1990, Diekhoff 1992). The null
hypothesis states that the mean change in abundance of group one is equal to
the mean of group two. If the two means are significantly unequal, then I can
conclude that woodpecker populations on high- abundance routes are changing
at different rates than low-abundance routes. The ANOVA analysis was
completed with routes classified in groups a.) one and two verses three and four
b.) and one versus four. The results of the ANOVA analysis will detect whether,
with external factors such as land cover change and broad landscape classes

equivalent, core and peripheral responses to land cover changes are similar.

39

CHAPTER 4. RESULTS

4.1 Red-headed Woodpecker Population Change
The change analysis, with corrections for observer effects, is displayed in
Table 5. The following formulas were used to calculate the red-headed
woodpecker population change values found in Table 5:
Percent = 1999 -1978 x 100
1 978

Ratio = 1993
1 978

Correction = Ratio + ( |Lowest Negative Valuel + 1 )

Log (bird) = log (Correction)

Negative change values were calculated in instances where the observer
correction indicated that the observer overestimated red-headed woodpecker
abundance. Therefore, some adjusted change values were negative. A
correction factor of the absolute value of the lowest negative value plus one was

added to ratios to provide positive change values.

40

 

Table 5. Red-headed Woodpecker Change: 1978-1993.

 

Route Percent Ratio Correction LogQird)

Route

Percent Ratio Correction Loggblrd)

 

18014
34002
34009
3401 0
3401 5
34022
34024
34026
34037
34039
34040
34045
34046
34047
34048
34059
34061
34062
34064
35002
3501 0
35029
36005
36006
36008
3601 0
36022
36024
36027
38027
49020
49022
49032
49035
49039
50002
50003
50007
50008
5001 7

0.0%
-33.4%
-53.8%
-31 .2%

-100.0%
21 .1 %
-39.0%
-83.3%
36.4%

-2.3%
-72.7%
21 .1 %
-31 .7%
-22.4%
26.6%
109.1%
50.0%

15.8%
-12.4%
-25.6%

-7.1%

18.3%
44.7%
-38.2%
-22.5%
-52.5%
-39.0%

-8.7%
-24.1%
43.2%
-81 .3%
366.7%
-41 .2%
-77.8%
-36.2%
-73.4%
-37.1%
-68.9%
-60.0%

10.6%

0.00
0.67
0.46
0.69
0.00
1.21
0.61
0.17
1.36
0.98
0.27
1.21
0.68
0.78
1.27
2.09
1.50
1.16
0.88
0.74
0.93
1.18
0.55
0.62
0.78
0.47
0.61
0.91
0.76
0.57
0.19
4.67
0.59
0.22
0.64
0.27
0.63
0.31
0.40
1.11

1.29
1.95
1.75
1.97
1.29
2.50
1.90
1.45
2.65
2.26
1.56
2.50
1.97
2.06
2.55
3.38
2.79
2.44
2.16
2.03
2.21
2.47
1.84
1.90
2.06
1.76
1.90
2.20
2.04
1.85
1.47
5.95
1.87
1.51
1.92
1.55
1.91
1.60
1.69
2.39

0.11
0.29
0.24
0.30
0.11
0.40
0.28
0.16
0.42
0.35
0.19
0.40
0.29
0.31
0.41
0.53
0.44
0.39
0.33
0.31
0.35
0.39
0.26
0.28
0.31
0.25
0.28
0.34
0.31
0.27
0.17
0.77
0.27
0.18
0.28
0.19
0.28
0.20
0.23
0.38

 

50018
50025
50026
52028
52029
54012
54013
61043
61047
61052
61061
61063
61065
61070
61072
66002
66033
66060
66061
72022
81010
91010
91011
91012
91015
91020
91034
91036
91042
91045
91046
91047
91050
91051
91054
91055
91056
91057
91063
91070

-58.2%
-47.9%
-47.7%
16.5%
-64.8%
-58.6%
-64.0%
-78.9%
-100.0%
-128.6%
0.0%
-49.2%
-80.0%
-69.2%
0.0%
100.0%
2.9%
8.6%
19.8%
0.0%
52.7%
61.3%
-1 18.8%
-20.3%
30.1%
-22.2%
-44.4%
6.0%
-47.9%
33.3%
-60.0%
-87.4%
13.7%
47.3%
-35.7%
-42.5%
33.7%
-54.8%
-83.3%
-52.6%

0.42
0.52
0.52
1.17
0.35
0.36
0.41

0.21

0.00
-0.29
1.00
0.51

0.20
0.31

1.00
2.00
1.03
1.09
1.20
1.00
1.53
1.61

-0.19
0.80
1.30
0.78
0.56
1.06
0.52

1.33
0.40
0.13
1.14
0.53
0.64
0.58
1.34
0.45
0.17
0.47

1.70
1.81
1.81
2.45
1.64
1.65
1.70
1.50
1.29
1.00
2.29
1.79
1.49
1.59
2.29
3.29
2.31
2.37
2.48
2.29
2.81
2.90
1.10
2.08
2.59
2.06
1.84
2.35
1.81
2.62
1.69
1.41
2.42
1.81
1.93
1.86
2.62
1.74
1.45
1.76

0.23
0.26
0.26
0.39
0.21
0.22
0.23
0.18
0.11
0.00
0.36
0.25
0.17
0.20
0.36
0.52
0.36
0.38
0.40
0.36
0.45
0.46
0.04
0.32
0.41
0.31
0.27
0.37
0.26
0.42
0.23
0.15
0.38
0.26
0.29
0.27
0.42
0.24
0.16
0.25

 

41

 

Figure 6 shows the distribution of change across the sample routes.
Overall, the red-headed woodpecker is experiencing declines from 1978 to 1993.
Woodpecker abundance has decreased 22.2 percent across the sample routes

taken as a whole.

 

Change In Red-headed Woodpecker Chance
Abundance 1978-1993. e -130 ‘36- -5o %
o 4096- 0%
0 +10%- +370%
Q
. O
c? 0° ° ‘
0 0 Of . t
O“ . ‘ .0...O . . . u .
. 1 o . r; 0
"1° 0 e ' 0°
. 1 . . O
s ,. s - o o _
t ‘9. ° ‘
_ L_ O
0‘ e o o) ' <99 0 O
- coo - ' . it

 

 

Figure 6. Change in Red-Headed Woodpecker Abundance:
1978-1993.

The distribution of declining and increasing abundance routes appears
fairly random throughout the sample region. To test this general observation,
route changes of quartile rankings were evaluated in an ANOVA. The ANOVA
yielded an F ratio of 0.485, which is not significant at the 95 percent confidence
level. Figure 7 is a box plot of change in each quartile. Woodpecker population
change in quartile 1, core routes, is the least variable and quartile 4, peripheral

routes, is the most variable. Yet, the ANOVA and visual inspection of median

42

change values in the box plot indicates that mean change is similar across the
sample region. Therefore, there are no significant differences in red-headed
woodpecker population change based on range location; this upholds the general
observation in Figure 6 that the distribution of declining and increasing

abundance routes is random across the sample region.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Box Plot: Red-headed Woodpecker Change by Quartile Ranking
0.50 G
30 so
. .1
E __
o _i_
i 0 40 _—
E ‘ " . I ~ . , _ _
8 ——-,-L- -.__J
£020 _J_ '
2
0m ——
1 2 3 4
Quartile Ranking

 

 

 

Figure 7. Box Plot: Red-headed Woodpecker Change by
Quartile Ranking.

4.2 Land Cover Change

A complete table of log land cover variables with a correction factor to
obtain positive values, and with the corresponding log red-headed woodpecker
change by route can be found in Appendix A. Figure 8 is the land cover change

occurring along BBS route 52029, which is considered a core route in my

43

research. Figure 9 is the land cover change occurring along BBS route 91020,
which is considered a peripheral route in my study. These figures provide a
representative depiction of land cover change occurring across BBS routes.

Overall, the most prominent changes in land cover are occurring in
agriculture, mixed forest, and deciduous cover types. Deciduous forest and
agriculture types are increasing in edge and fragmentation, while the total
acreage of deciduous forest is increasing and the total acreage of agriculture is
decreasing. The amount of mixed forest is decreasing along BBS routes. Table
6 lists the land cover variables that were measured and the average change

occurring along the BBS sample routes.

 

.5822 .083 28a wmm .550 2.3 82.32 .w sac

 

 

 

 

.28 .8... 2.2
n .- u
. H
u .1.
9.2820 238..qu ‘ .- . f . i u .l 325$ 8
2.2820 2 23:33.3. 8 . - .I u . ‘ —.- 25820 a. .. UV
Lil.
33.322238? ‘ .— .35... «83.58 A... - J.
a r
2.9.6on 2 2239.3 ‘ H 2333.? m/Mxtiv
22.5.? Susanne-o ”Wu H 50:00 acui— ._0>OO U54 98w
35:0 02 8 ..-. a
sac-:0 u ..
8a.. 8on9. r r
093:0 .260 2.3..
80% . . .. ..
b I _ a... .
“lit. wk 4:).
.- WUMH.)J L1 .-l J —
A“ i-. l. w-i i. . n53 hf
. My“ 1.. +I|.- - —

 

.....-2i .32 .5532 58% 83¢ wmm .-
. .260 2:3 mam T 22. .

 

 

45

 

1978 -1 993 Land Cover 73%;» ,
BBS Route 91020, Wisconsin 7 i 31%

1993 Land Cover

  

Lmd Cover -.
C3 Agriculture “ 5:33:99” C3 Residential “ Water

CL3 2:15:11 ous “ Mixed Forest 3 Urban C3 Wetland
1978 Land Cover

 

 

 

Land Cover Change
1978 to 1993

'_ nor - ii - "" ’5"
Change
c3 No Change 15;: To Vlblland (as Mixed to Urban (:3 Mix,“ 1°
Agriculture
K Mixed to Evergeenl “ Water to G3 Agriculture to C3 Agfiwjmr.
Deciduous Deciduous Residential to Urban
(:3 Agriculture to Forest if"; Vibtland Loss “ Residential to Urban t j

 

Fugue 9. 1978-1993 Land Cover. BBS Route 91020, Wsconsin.

46

 

 

Table 8. Average Land Cover Change Occurring Alogg BBS Routes.

 

 

 

 

Land Cover Area Mean Patch Size Edge
(Hectares) (Hectares) (Meters)
Agriculture -103.81 -69.41 10530.38
Deciduous Forest 204.02 -11.78 19671.89
Evergreen Forest -2.75 -19.16 3746.79
Grassland 54.25 6.07 6590.45
Mixed Forest -264.05 -41.13 -3310.00
Residential -5.24 2.16 -1335.50
Transitional -8.32 -10.32 216.00
Urban -17.90 -4.01 -1 781.74
Water 1 1.72 0.62 845.1 1
Wetland 27.34 -5.45 5225.89
Land Cover Edge Density Mean Patch Dimension lnterspersion
(Meters! Hectare) (Index: 1 - 2) (Percent)
Agriculture 3.78 -0.01 -14.71
Deciduous Forest 7.08 0.07 13.18
Evergreen Forest 1.32 0.38 25.41
Grassland 2.29 1.02 21.50
Mixed Forest -1.13 0.00 7.43
Residential —0.46 -0.12 -4.34
Transitional 0.14 -0.50 -7.21
Urban -0.65 -0.24 -0.33
Water 0.30 0.07 12.58
Wetland 1.81 0.24 22.54

 

47

 

4.3 Analysis

4.3.1 Multiple Regression

Stepwise regression yielded three land cover variables that were
significantly related to red-headed woodpecker abundance. Mean patch
dimension of agriculture, agriculture edge density, and urban area were all
significant predictors of red-headed woodpecker change at the 95 percent
confidence level with an R2 of 0.161. Table 7 summarizes the multiple

regression statistics

Table 7. Stepwise Regression Output.

Dependent Red-headed woodpecker change

independent Log Agriculture Mean Patch Dimension (AGMPD),
Logilgriculture Edge Density(AGED), LogUrban Area(URAR)

 

 

 

 

 

 

 

Model R R Square Std. En'or F Change df1 df2 Sig. F Chang:
1 0.401 1 0.1609 0.0959 4.3737 1 76 0.0398
Model Coefficients Std. Error t 3L
1 (Constant) 0.0707 0.0587 1.2029 0.2328
AGMPD 2.2292 0.8047 2.7704 0.0070
AGED 0.0937 0.0422 2.2227 0.0292
URAR 0.0302 0.0144 2.0914 0.0398

 

 

The R value is the correlation observed between the independent and
dependent variable; it ranges between 0 and 1, with higher values indicating
stronger relationships. The significance of the multiple regression model is
0.0398, indicating that the model is a good predictor of red-headed woodpecker
change at a 95% confidence level. The bottom section in Table 7 analyzes the

coefficients. The t-statistics indicate the relative importance of each variable in

48

the equation and their associated significance level. Each variable is significant
at the 95 percent confidence level, therefore these variables were isolated for a

hierarchical cluster analysis to separate the most similar routes.

4.3.2 Hierarchical Clustering Analysis

The three land cover variables significantly related to the abundance
change of red-headed woodpecker were isolated into a table and evaluated
using a hierarchical cluster analysis. The clustering routine isolated four clusters,
which were most similar in land- cover characteristics significantly related to red-
headed woodpecker abundance. Interpretation of a cluster dendrogram is fairly
arbitrary. It is critical to look for gaps in the coefficients, as well as to visually
inspect the cluster dendrogram for breaks or groupings. Upon inspection of the
agglomeration schedule, Appendix B, and the dendrogram, Figure 10, I decided
on four clusters of similar routes occurring along distance of five in the
dendrogram. Each cluster is color coded in Figure 10. The routes of each
cluster were isolated into a table for analysis of variance between routes of high-
abundance and low-abundance. Figure 11 shows the location of these clusters

in the sample region.

49

Figure 10. Hierarchical Cluster Analysis Results.

Rescaled Distance Cluster Combine

 

 

 

r1 .1 ‘
4-034 a- —w
,. 1,, r
‘3 ‘v ‘ I {1. .4
a r) <
S 3‘ L I: L fl
#7 \,.,
-' l (.4 l —
q“ /:(.r) ’7
a ; f: x J
C ’ 1 5 --4
W 4‘ j? I —<
I '3. ‘ [I
A A) J. _
t
4 1 h _1
u, L 7' #14
3' f l :11 iii —4
,. F a
4 r _L H ——4
_ a. 1 d
- *1 - L- 2 1
3‘ 1 ' ”I.
1‘“? t —"i
.1 r 1 ‘ ‘
't~ l j i —_J
/, ~_
r“ ‘ .‘1 ; ‘, ——-t
1 ll r‘ ‘33 3 i —4
L ’. ‘ l'
< “2 "‘+‘ _' —'l
“i 4 e; .'
1
F i t (r _ ‘l a
4 if r T 4, _.
x A.
4 A 'I. '_l
L, A (_f/ (‘ i -——i
4 1; -.
4 L1 Cl A ‘1 .4

 

 

 

‘5‘. .
4-4. 4 .
4 « ..
“'1 ’"1J/ 2 ‘

 

 

f’i/“'/
’ik/L‘.) V‘
'1 ‘fi
CfLiét‘ Y

 

 

 

 

 

 

50

61043
61047
18014
66033
81010
91042
61063
34062
34064
91012
34010
36024
36027
91057
36008
34002
35002
91010
34015
36022
34045
91036
34024
91045
50003
66002
61072
34047
61052
36006
61065
61070
91011

("i Iq ”2 La
.i K} L. J

U1

2029
35010
91050
91070
50007
91015
36010

34009

Ui
\

Figure 10 (cont’d).

58
51
56
14
28
63
67
27
78
15
34
54
74
33
77
4O
68
19

1

2

9
46
43
71
53
24
47
38
50

L

 

 

 

 

 

 

 

 

 

 

51

 

 

Hierarchical Cluster Results

Q
’g'.
O .0:
. .‘.u.'.e
O
,O.. .0 . '0‘
.. a. O .0
... . .

‘ e
E

Figure 11. Hierarchical Cluster Distribution.

4.3.3 Analysis of Variance

The analysis of variance (ANOVA) looks for differences in means between

changes.

groups. If there is a significant difference in the means between the two groups,
then this indicates that the changes in red-headed woodpecker abundance
occurring along high-abundance routes are not the same as the changes
occurring along low-abundance routes in areas with similar land cover changes.

This would indicate that range location does play a role in determining population

Each route cluster (determined from the hierarchical cluster analysis) was

52

' divided into two groups; routes with a spatial ranking of one or two compared

with routes spatially ranked routes three or four. This divides the sample routes

into groups with similar land cover changes and similar abundance to determine
whether range location makes the species more susceptible to decline. In
addition, I separated routes with spatial ranking one or four for analysis of strictly
high-abundance routes versus low-abundance routes.

Cluster one’s analysis between ranking 1,2 versus 3,4 proved to be
insignificant. There was no significant difference between the two group means.
Table 8 is a detailed description of the ANOVA results for all clusters. The
variability within groups for cluster 1 was 6.03. This is the difference between
each variable and the group mean. The variability between groups was 0.22; this
is the difference between each group mean and the total mean (Sprinthall 1990).
The mean square (MS) is used to calculate the F ratio. The MS is the division of
each respective variability by their degree of freedom. The division of the MS of
the variability between groups by the MS of the variability within groups provides
the F ratio.

The F ratio for cluster one (1,2 versus 3,4) was 1.11, which is not
significant at the 95 percent confidence level. Therefore, the mean change
between high-abundance routes and low-abundance routes is not statistically
different.

ANOVA was completed for cluster one, spatial rank 1 versus 4. This
provided the same results, with an F ratio of 0, which is not significant at the 95
percent confidence level.

I was not able to use cluster two in ANOVA because it only contained

three routes, which all occurred within the spatial rank group 4.

53

Table 8. Analysis of Variance.

 

ANOVA. Cluster 1:

Spatial rankings 1,2 vs. 3,4.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sum of Squares df Mean Square F SkL
Between Groups 0.22 1 0.22 1.11 0.30
Within Groups 6.03 30 0.20
Total 6.26 31
ANOVA. Cluster 1: Spatial rankings 1 vs. 4.

Sum of Squares df Mean Square F SQ
Between Groups 0.00 1 0.00 0.00 0.96
Within Groups 1.52 13 0.12
Total 1.52 14
ANOVA. Cluster 3: Spatial rankings 1,2 vs. 3.4.

Sum of Squares df Mean Square F Sig.
Between Groups 0.00 1 0.00 0.00 0.99
Within Groups 7.79 31 0.25
Total 7.79 32
ANOVA. Cluster 3: Spatial rankings 1 vs. 4.

Sum of Squares df Mean Square F Sig.
Between Groups 0.01 1 0.01 0.04 0.85
Within Groups 5.23 15 0.35
Total 5.24 16
ANOVA. Cluster 4: Spatial rankings 1,2 vs. 3,4.

Sum of Squares df Mean Square F Sig.
Between Groups 0.01 1 0.01 0.08 0.78
Within Groups 1.88 10 0.19
Total 1 .89 1 1
ANOVA. Cluster 4: Spatial rankings 1 vs. 4.

Sum of Squares df Mean Square F Sig.
Between Groups 0.00 1 0.00 0.00 0.97
Within Groups 1.22 5 0.24
Total 1.22 6

 

Cluster three provided insignificant results as well. The mean change

between rankings 1 and 2 versus 3 and 4 yielded an F ratio of 0, which is not

 

significant. The mean between rankings 1 versus 4 had an F ratio of 0.04, which
is also not significant at the 95% confidence level.

Cluster four’s analysis of low-abundance versus high-abundance routes
resulted in an F ratio of 0.08, and the analysis of rankings 1 versus 4 resulted in
an F ratio of 0. Both failed to reject the null hypothesis at a 95 percent
confidence level.

The ANOVA failed to reject the null hypothesis in every cluster, which
states that the mean change between low-abundance and high-abundance
routes is equal. Therefore, the red-headed woodpecker is not sensitive to range
location in this study, and core and peripheral populations, in similar land cover

environments, are experiencing similar abundance changes.

55

CHAPTER 5. DISCUSSION AND CONCULSIONS

My study analyzed the impacts of land cover changes on the red-headed
woodpecker by aggregating routes with similar land cover changes associated
with woodpecker population change, and subsequently determining if variation in
population changes were occurring as a result of range location. Current
research in biogeography indicates that range location is not a key factor for
survival in range collapse studies of endangered species (Lomolino and Channell
1995). However, translocation success appears to be impacted by range release
location (Griffith et al. 1989, Wolf 1996). My study attempted to clarify this
contradiction by connecting environmental variables to a species’ population
change and evaluating locations with comparable environmental conditions in
order to isolate changes resulting from range location. It is important to look at
routes with similar land cover structures so as to isolate the impact of range
location on changes in species’ abundance. Thus, this study is unique; it
attempts to link abundance with land cover variables and analyze routes in the
core and periphery in an attempt to determine if location has an effect on
population change.

The red-headed woodpecker demonstrated no significant differentiation in
changes between populations at high and low-abundance range locations on
routes with similar land cover changes that were previously determined to be
statistically significant to red-headed woodpecker population change. The
ANOVA results range from significance at 1% through 70% confidence levels,

indicating that red-headed woodpecker reintroduction efforts would do equally

56

well in quality habitat locations in the core and periphery of it’s range. Quality
habitat locations should be evaluated based on environmental variables, with
emphasis upon life requirements and those variables that research has indicated
as a significant indicator of abundance. Three land cover variables were isolated
by this research as significant in causing red-headed woodpecker abundance
changes.

This study discovered that the woodpecker was statistically sensitive to
changes in agricultural mean patch dimension, agricultural edge density, and
urban area. These variables explain 16% of the variation in red-headed
woodpecker population change. I expected the land cover variables to explain
more of the fluctuation in red-headed woodpecker abundance. The low R2
results may indicate that there are additional variables contributing to population
change that were not evaluated in this study, i.e., temperature, competition, and
disease. Errors in the bird or land cover data used in this study could also
contribute to low R square values. For example, errors in the land cover data
sets would result in miscalculations of real change occurring along the sample
BBS routes, and could lead to low R2 values when analyzed against red-headed
woodpecker abundance.

Agricultural mean patch dimension is the average complexity or
fragmentation measure of the agricultural parcels along a route. Woodpecker
abundance was positively related with agriculture mean patch dimension;
abundance increased with increases in agriculture mean patch dimension

measures. A positive relationship was also revealed with agricultural edge

57

density. These results were not surprising because the woodpecker has been
documented as utilizing forest patches adjacent to agricultural fields, which are
used as a food source (Page 1966, Brown et al. 1999). However, the
relationship unveiled between urban area and abundance changes was not
anticipated. A positive relationship, at the 95% confidence level, was calculated.
This contradicts current literature, which has shown that urban clean up efforts
and competition with the urban inhabitant European starling has resulted in
population declines (Page 1966, lngold 1989, Page 1996, Brown et al. 1999).
Urban area change was included in the hierarchical cluster analysis because my
data indicated it as a significant variable, however this result provoked some
questions regarding the accuracy of the land cover data that were used in the
analysis.

Aerial photography was the base for the 1978 data and satellite imagery
was the source for the 1993 data. Theoretically, both data sets should accurately
classify the level-1 land cover classes used in my analysis. The average
changes occurring along my BBS sample routes (see Table 6, page 47) caused
me to question this assumption. Most of the variables I measured produced
expected results; agriculture is declining, deciduous forest is increasing, but
becoming more fragmented, and residential patches are becoming larger in size.
The urban and residential classes exhibited declines between 1978 and 1993, an
average of —17.9 and —5.2 hectares per route, respectively. Built-up areas rarely
are converted back to forest land or agriculture. This leads me to believe that

commission errors are occurring in the 1993 data with some urban and

58

residential pixels classified as another land cover type. If true, this would result
in an inaccurate portrayal of the impact that these land cover variables are
having on red-headed woodpecker populations.

In addition, I think that the mixed forest class is either over classified in the
1978 data or omitted in the 1993 data because an average decline of
264 hectares! route was calculated for my study region. It seems unrealistic that
264 hectares of mixed forests have either become homogenous or were
converted to another land cover type in the span of 15 years. However, without
field-based error checking of this data set, I cannot accurately or concretely
determine whether these are realistic changes occurring along my sample routes
or not. Time and cost restraints prohibited an error checking effort, and I had to
trust that the land cover classifications were completed thoroughly and
accurately.

The unresponsiveness of red-headed woodpecker populations to range
location has some implications in support of biogeographic research results and
in advisement for species management. While my study only evaluates one
species, it adds another piece of evidence in support of the utilization of
peripheral range locations for endangered species management or
translocations. While optimal habitat areas may be more prevalent in core
locations, thus enabling the support of more individuals, optimal habitat regions
in the periphery should not be disregarded for management activities. In the
case study of the red-headed woodpecker, range location had no significant

affect on population change.

59

With all significant environmental variables similar, range location had no
impact on changes in red-headed woodpecker abundance. While the land cover
data utilized in this study yielded some questionable change results, it was the
best data to feasibly use and originated from the USGS, which I view as a
reliable source. If time and monetary constraints were not an issue in this study,
I would classify my own land cover data from strictly aerial photography or
satellite imagery and not a mix of the two. Additionally, I would sample the
maximum number of routes yielded in the power analysis, 200 routes, to ensure
that alpha and beta errors are at acceptable levels. The red-headed woodpecker
is one species on this planet, and more research is necessary to definitively
resolve the impact of range location on changes in species’ abundance.

Future research should continue to link species declines to their causes.
While we can learn from observing general patterns of species distribution and
decline, these patterns will not be useful to managers until the underlying
reasons for the patterns are determined.

The implications of the case study of the red-headed woodpecker
ovenivhelmingly suggest that wildlife managers should consider peripheral range
locations for management activities. I recommend that habitats be evaluated on
an individual basis. Management activities should be concentrated in areas that
contain necessary life requirements for the species being managed, whether this

location is in the core or periphery of the species’ historic range.

60

APPENDICES

61

APPENDIX A

LOG1o LAND COVER AND RED-HEADED WOODPECKER CHANGE

62

 

 

 

 

 

00... ?0.0 00. .. 0?? 50.0 00.0 0005.0 0000?
3.5 00.0 00... 00.? ?0.0 00.0 .0000 50000
00.0 00.0 ..5.0 00.? 50.0 00.0 503.0 50000
50. v ?0.0 05.0 00.? 00.0 00.0 .03 .0 ?0000
0...0 ?0.0 «0.0 00.0 00.0 00.0 5550.0 00000
?0.0 00.0 00.0 00.? 0...0 00.0 50.00 0.000
5...0 00.0 05.0 0w? 00.0 00.0 0?..00 00000
00.0 ?0.0 05.0 00.? 00.0 ?0.0 0050.0 00000
?0.0 00.0 ..0.0 00.? 00.0 00.0 030.0 00000
3. F 00.0 0?.0 00.0 00 .0 00.0 0000.0 00000
00. v 00.0 00.0 0...? 0.. .0 00.0 003.0 0 5000
?0.0 00.0 55.0 3.? 0p .0 00.0 0500.0 00000
0w.0 00.0 05.0 0?? 00.0 3.0 530.0 ?003
00. r ?0.0 05.0 0...? 00.0 5.0 ..000.0 0003
00... 00.0 00.5 0?? 0 F .0 05.0 0???.0 F003
.0... 00.0 00.0 00.? 00.0 00.0 0000.0 0003
00. _. 00.0 00.0 0 r .? 50.0 00.0 000?.0 0?03
.0... ?0.0 50.0 00.? ..0.0 00.0 030 .0 5?03
..0... 00.0 3... 00.? 00.0 3.0 030.0 0?03
?0.F 00.0 00.0 3.? 00.0 ?0.0 ?500.0 0?03
50. .. 00.0 ?0.0 00.? 00.0 00.0 000 .. .0 0?03
00.0 ?0.0 00.0 00.? 05.0 00.0 5?00.0 0003
00... 00.0 00.0 00.? 00.0 50.0 30?.0 5003
.0... 00.0 ?0.0 50.? 00.0 00.0 .00? .o 0003
00. _. 00.0 00 .o 00.0 50. F 00.0 5550.0 ?003
5... 00.0 ..0... 00.? 0...0 00.0 0500.0 0003
00.0 ?0.0 00.0 00.0 ..0.0 00.0 .00...o 0.03
00. F 00.0 ?5.0 00.? 0.. .0 00.0 ?000.0 0 503
00. .. 00.0 50.0 00.? F00 00.0 00?0.0 0003
.0... 00.0 05.0 5...? 00 .0 00.0 0000.0 0003
00... 00.0 0?.0 00.0 00.0 00.0 500 .. .0 300..
Exp—383:. eBacoEE sauna. :35 5.2.00 Guam swam 030 :3}. use! 02< a 2.0 3:01
2:230...

 

 

 

3320 .3388; 882-8m 2e .98 .83 Bus < 52%?

63

 

 

 

 

0... 3.0 00.0 ~00 .~0 00.~ -0¢0 0.0.0
00.. 00.0 00.0 0.... .~0 00.~ 830 0.0.0
.0.~ 00.0 00.0 00.0 30 3.~ 0000.0 -0~.
3.. 00.0 0.... 00... 0~0 0~.~ 0000.0 .0000
0... 00.0 ~0. .3 0.0 ~0.~ .000 00000
0... 00.0 00.0 0.... 000 00.~ 030.0 00000
00.. 3.0 .00 30 .~0 00.~ 00.00 8000
3.. 3.0 00.0 000 .~0 .0.~ 0000.0 ~83
00. 3.0 00.0 000 80 ~0.~ 0~0~.0 0.0.0
.0. 00.0 :0 00... ~00 .0.~ 0.2.0 000.0
3.~ 00.0 00.0 0.... .~0 00.~ ~00~0 000.0
~0.~ 3.0 .00 2.0 0~0 00.~ 0000.0 .020
00.. 3.0 00.0 00.0 -0 00.~ 0000.0 ~00.0
00.. ~00 0~.0 0.0 :0 00.~ .020 0.3.0
00.. 00.0 00.0 0.... .~0 00.~ 002.0 03.0
00.. ~00 3.0 0~... 00.~ .0.~ 30~0 0.03
.0. ~00 00.0 00... 00.~ 00.~ 3.~.0 ~33
0~.~ 3.0 0... 00... ~00 0~.~ ~...~.0 0~0~0
00.~ 3.0 3.. 00... 0.0 00.~ 0000.0 0~0~0
.0. 3.0 0... 00.0 0.0 00.~ 0~0~.0 0~000
00.~ 00.0 0~.. 0~.. 0~0 000 000~.0 088
00.. 00.0 00.. 03 t0 00.~ v.0~.0 0.000
-.. 3.0 00. £0 30 00.~ 00.00 0.000
00.. 00.0 B. 03 00.~ 0..~ 00-.0 00000
...~ 00.0 00.0 0~... t0 00.~ 000~0 .0000
0.0.. 0.0.0 3.0 30 000 00.~ 0~0~.0 00000
00.. 3.0 00.0 00.0 0~0 000 80.0 3000
0.. 3.0 0... 00.0 000 00.~ .33 00000
.0. 00.0 0... ~00 0.0 0~.~ 32.0 00000
.0. .00 0~.. 3... 00.~ 00.0 --0 ~000..
00.. 00.0 0.. 00.0 000 00.~ .080 -00..
0222.83... 02205.0 03.... :8: 00.2.8 00.0. am. 00.0 03.... :3: 32 2.0 .53.
22.32.30.

 

 

.0308 50.000802, 8002.000 000 .28 003 0.0% .< 023%

 

 

 

 

 

Eur N0.0 OOO NN.¢ VON NON MOVNO ONOFO
Noe VOO ONO Oeé mN.m 5N FNm—aO moo—O
Ow... VOO tho m; ON.m NON OOONO FOO—O
ON? VOO FOO ONé 92m CON OOFVO @0me
SN VOO OOO 09v ONO OON NOONO moo—O
Fmé voO Nod NNé 22m NON NOONO «mo—O
Omé OOO NOé and 5d MON VOW—NO FOO—4O
mwé MOO POO 92¢ ed OON NVOOO OOOvO
NON MOO cod 3.? ON.m No.0 NOEO NVOFO
mmé VOO OVO mod _.N.m OON OONNO 03 PO
OON mod and OOd NNN OON wOFvO 955
EN 3.0 OOO 34‘ 3d OON OOmNO N33
3.? OOO mod wad NN.m NON NONOO ONO—O
NO.—. VOO mad 0?? O P .m mON rmmNO 30;
NON VOO _.O._‘ «#6 ONO OON wand ONO—O
OON VOO mod 0N6 ON.” NON NNSO 053
m5? MOO ONO Of? San NON mend NFOFO
NON VOO OYO find PN.O NON NOVOO :OFO
going—23:. 530-585 :33..— cuo! Eocen— ogm owum 03% count coo! 50.2 9.5 3:01
flagaotdfl

 

 

 

.09ch 558325 38933. new .900 95.. 3&3 .4. 59.83.

65

 

 

 

 

.0. 00.0 00.. ~00 00.~ 00.~ 0800
.~.. 8.0 3.. ~00 00.~ 0~.~ -000
00.. 00.0 3.0 00.~ 30 00.. -000
8.0 8.0 8.0 00.0 8.0 8.0 3000
00.0 8.0 00.0 00.0 00.0 8.0 -000
00.. 00.0 3.0 00.0 .00 00.~ 0.000
.0.. 0.0 00.0 8... ~00 .00 0800
00.. 00.0 00.0 00.0 00.~ ~00 8000
00.. 0.0 00.0 ~00 00.~ 00.~ 0800
00.. 90 ~00 3.0 00.~ 3.0 00000
00.. 0.0 0~0 0.0 00.~ 00.~ 0.000
-.. 00.0 ~00 00.0 ~0.~ .00 800
3.. 00.0 000 00.0 -.~ 3.. 303
0~.. 0.0 0... 0.... 00.~ 00.~ ~83
00.. 9.0 ~... 00.0 00.~ 0~.~ .003
8.. 000 ~00 0.... ~0.~ 00.~ 0003
00.. 00.0 ~00 0.... 00.~ 00.~ 033
00.. 0.0 .00 00.0 00.~ 00.~ ~33
00.. 000 ~... 00... 00.~ 00.~ 033
~0.. 00.0 0~.0 0.... 00.~ 00.~ 033
00.. 00.0 ~00 -.v ~0.~ 00.~ 033
00.. 00.0 00.0 0~... 00.~ 30 0003
B. 000 ~00 00.0 00.~ 00.~ ~83
80 8.0 8.0 8.0 8.0 8.0 083
00.. 000 00.0 00.0 ~0.~ ~00 «~03
0~.. 00.0 3.. 03V ~0.~ 00.~ -03
8.0 8.0 8.0 8.0 8.0 8.0 0.03
0~.. 00.0 ~00 0~.0 ~0.~ 00.~ 0.03
00.. 00.0 00.0 00.0 00.~ ~00 083
00.. 00.0 0~.0 0.0 00.~ 00.~ ~83
~0.. 00.0 .... ~00 8.0 00.. 0.00.

00.22.23... 0200205 03... 08: 8.0000 00mm 0&0. on.» 030.. 0...: 02¢ 3000.

30“— 03032000

 

 

00005 5.0000003 8002.000 000 .98 003 0.03 < 0882

66

 

 

 

 

00.. 00.0 -0 ~0... 00.~ 00.~ 0.0.0
8.0 8.0 8.0 8.0 00.0 00.0 0.0.0
00.. 00.0 00.. 00.0 00.~ -.~ -0-
00.. 00.0 0~.. 00.0 00.~ 0~.~ .800
8.0 000 ~... 00.0 00.~ 0~.~ 0800
0~.. 000 00.0 0~.0 00.~ 00.~ 00000
00.. 000 00.0 00.0 00.~ ~0.~ 8000
.0. 00.0 00.. 00.0 00.~ ~..0 -0.0
00.. 000 ~0.. ~00 00.~ 00.0 0~0.0
~0.. 00.0 ~00 00.0 00.~ 0~.~ 000.0
~0.. 00.0 00.0 00.0 00.~ 00.0 000.0
~0.. 00.0 00.. ~00 00.~ ~..~ .8.0
00.. 00.0 0~.. 0~.0 00.~ 00.~ ~00.0
00.. 00.0 0~.. 00.0 .0.~ 0~.~ ~00.0
00.~ 00.0 .0. 00.0 00.~ 00.~ 000.0
00.. 00.0 0~.0 00.0 00.~ ~0.~ 0.000
8.0 8.0 8.0 8.0 00.0 8.0 ~.000
~0.. 00.0 000 00.0 00.~ 00.~ 0~0~0
0~.. 00.0 .00 0~.0 00.~ ~0.~ 0~0~0
~0.. 000 0~.. 0~.0 .0.~ ~0.~ 0800
00.. 00.0 00.. 00.0 0~.. 00.~ 880
00.. 000 0... 00.0 00.~ 0~.~ 0.000
~0.. 00.0 00.0 -.0 00.~ 00.~ ~.80
00.. 000 .00 0.0 00.~ 00.~ 0080
~0.. 00.0 00.0 0.0 00.~ .~.~ ~800
00.. 00.0 000 00.0 00.~ 00.~ 0800
-.. 00.0 00.0 8.0 ~0.~ 0~.. ~080
0~.. 000 00.. 0~.0 00.~ 00.~ 00000
~0.. 00.0 0~.. 00.0 00.~ 00.~ 00000
00.. 00.0 00.. 00.0 00.~ 00.~ ~0000
.0. 000 0~.. .00 00.~ 00.~ -000

02200280. 02000.05 080.. 08: 00.0000 00.00. $0.". 200 0800 000: 002 0.00m

«850"— 03039—000

 

 

00005 000000082, 000000-000 000 .008 003 0.03 .< 000002

67

 

 

 

 

OOé OYO O0.0 ONé OVN VON ONO—O
OOé OOO OOO ONé ZEN OON OOOFO
OON OYO ON... NNé0 OON OON NOOwO
ON. 0 Ovd Ooé Ové NON NON OOO _.O
3.0 OYO OOO NOé NON OON OOOFO
OO... OYO VOO ONé OON NON OO 3
NO. .0 OYO 3.0 :10 OO N NNN _.OO 3
OOé Ovd OOO OP .v OON OON OOO 3
NO; O¢.O OO... FOé OON OON NvOFO
NO... OYO ONO OOJ0 OON OvN O35
ON. 0 OYO OOO OO.v OVN OON O3 PO
NO... OYO Nwé OYV OON ONN N35
5.? OYO ON.O OOé OON ONN OOOFO
ONé OYO NOO OOé OO N OVN OO 5
OON OOO NV... No.00 OvN ONO ONO—O
5.. OYO No.0 Ové O...N OON OFOFO
ON... OYO 3.0 OOé OON wNN NFOFO
:N O0.0 OV... OOé VON FOO :0me
cgeaflsfiap: .50—0605.0 =0uflt COO: 3.0—.00 GNU“ CHOW 05m Scufll =00: §< 050K
a020n— “3032000

 

 

.mdcmco 09.002000; gouacuuwm 0cm ..o>oU new; ovum.— .< Anzac/0.

68

 

 

 

 

8.0 8.0 8.0 8.0 8.0 8.0 0800
-.. .00 8.. 00.0 0... 00.~ -000
8.0 .00 00.0 .00 0~.0 .0. -000
8.0 8.0 8.0 ~0. 0 -0 0~.. 0800
8.0 .00 -.0. 0~. 0 0... 0... -000
8.0 .00 000 00. 0 00.0 0~.. 0.000
00.. 8.0 .~.0. 0~. 0 00.0 00.0 0800
8.0 .00 00.0. .00 ~00 ~00 00000
8.. .00 00.0 0.0 00.0 8.~ 0800
8.0 8.0 8.0 8.0 8.0 8.0 083
8.0 8.0 8.0 8.0 8.0 8.0 0.000
0~.. 8.0 0~.0 00.0 .~.0 .0. ~800
8.0 8.0 8.0 8.0 8.0 8.0 0003
8.0 8.0 8.0 8.0 00.0 8.0 ~83
~0.. 8.0 ~..0. ~00 ~00 ~... .83
8.0 8.0 8.0 8.0 8.0 80 0003
8.0 8.0 8.0 8.0 8.0 8.0 0003
8.0 8.0 8.0 8.0 00.0 00.0 ~003
8.0 .00 00.0. 00.~ .~.0 .~.0 0003
8.0 8.0 8.0 8.0 8.0 8.0 0003
8.0 8.0 8.0 8.0 8.0 8.0 0003
8.0 8.0 8.0 8.0 8.0 00.0 0003
8.0 8.0 8.0 8.0 00.0 8.0 ~003
8.0 8.0 8.0 8.0 8.0 00.0 083
8.0 8.0 8.0 8.0 8.0 8.0 083
8.0 00.0 8.0 8.0 00.0 00.0 -03
8.0 8.0 8.0 8.0 8.0 8.0 0.03
8.0 8.0 8.0 8.0 8.0 8.0 0.03
8.0 8.0 8.0 8.0 8.0 8.0 083
8. 0 8 0 8. 0 8. 0 8. 0 00.0 ~83
8. 0 8. 0 8. 0 8. 0 8.0 0.00.

022000.00... 00.2.2.5 080.. 000: 3.008 00mm 0000. 00.0 080.. 000: 02¢ 303.

”BC—008°

 

 

00005 00.000083 000000-000 000 .030 000.. 0.00.. 0 000002

_69

 

 

 

 

 

O0.0 O0.0 O0.0 O0.0 SO SO OFOFO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 OwOPO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 NNONN
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 FOOOO
SO SO SO O0.0 O0.0 O0.0 OOOOO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 OOOOO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 NOOOO
SO SO O0.0 O0.0 O0.0 O0.0 NNOPO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 ONO—.O
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 OOOFO
O0.0 O0.0 O0.0 SO SO O0.0 OOOFO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 .OOFO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 NOOFO
O0.0 O0.0 O0.0 O0.0 SO SO NVOFO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 O¢OFO
SO 86 N p .O. ON.O ON.O OO... O vOVO
OOé O0.0 vOO ONé O0.0 OFN NvOVO
OO.? POO N0.0 ONé OOé NNN ONONO
OV... O0.0 NvO OOO OOO OO... ONONO
SO SO SO SO O0.0 O0.0 ONOOO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 ONOOO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 OPOOO
O0.0 SO SO O0.0 O0.0 O0.0 NwOOO
O0.0 VOO OO. 0. OvN ON.O. ONO. OOOOO
O0.0 SO SO SO SO SO NOOOO
O0.0 O0.0 SO SO SO O0.0 OOOOO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 NOOOO
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 OOOOV
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 OOOOv
O0.0 O0.0 O0.0 O0.0 O0.0 O0.0 NOOOQ
10. F FOO N0.0 OOO NO. —. NON NNOOv
sci—323.... 0.0.2.250 :33. :3! 3.2.00 ouum cam on.» 0.3.0. :00! 02¢ 3:01
UGO—883

 

 

000000 00008082, 000000.000 000 0008 0000 013% .< 0000004.

70

 

 

 

 

 

~0.. .00 00.0. ~0.~ 00.0 00.0 0~0.0
8.0 8.0 0~..- 0~.. 00..- 00..- 000.0
8.0 8.0 00.0 00.0 00.0 8.0 ~00.0
~0.. 8.0 .00. ~0.~ 00.0 00.0 000.0
00.0 8.0 8.0 8.0 8.0 8.0 000.0
00.0 8.0 8.0 00.0 00.0 00.0 000.0
8.0 8.0 8.0 00.0 00.0 8.0 .00.0
8.0 8.0 00.0 00.0 00.0 8.0 000.0
8.0 8.0 00.0 8.0 00.0 00.0 ~00.0
00.. 8.0 .0? ~0.~ 00.0 00.0 000.0
00.0 8.0 00.0 00.0 00.0 8.0 000.0
00.0 8.0 8.0 00.0 8.0 8.0 ~00.0
8.0 8.0 8.0 00.0 8.0 8.0 000.0
8.0 8.0 8.0 00.0 00.0 8.0 30.0
8.0 8.0 00.0 8.0 00.0 8.0 0~0.0
00.0 8.0 8.0 8.0 00.0 8.0 0.0.0
8.0 00.0 8.0 00.0 00.0 00.0 ~.0.0
00.. .00 .00 00.0 00.0 00.. .83
00.20.0230. 02000.05 030.. 000: 00.008020 00.0 030.. 000: 002 33¢
3.50.0080

 

 

000000 000000082, 000000-00¢ 000 .030 000.. 0% .< 000002

71

 

O0.0
ON...
O0.0
86
86
86
SO
O0.0
O0.0
O0.0
O0.0
O0.0
cod
9.:
OO.?
OON
SO
SO
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
86
8.0
O0.0
O0.0
NON

0.0.8323:— :o_0:oE_n_ :39... can! 3.2.00 anew Guam on.» 0.3.0— :00!

86
O0.0
O0.0
O0.0
SO
86
SO
SO
O0.0
O0.0
O0.0
O0.0
O0.0
POO
FOO
OVO
O0.0
SO
SO
O0.0
SO
SO
SO
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0

O0.0
0N0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
VNé
Ow...
NN...
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
O0.0
OV...

O0.0 O0.0 O0.0 ONOOV
OO.? OON VNO NNOOO
8.0 86 006 NNOOO
O0.0 O0.0 O0.0 VNOOO
O0.0 O0.0 O0.0 NNOOO
O0.0 O0.0 O0.0 O FOOO
O0.0 O0.0 O0.0 OOOOO
O0.0 O0.0 O0.0 OOOOO
O0.0 SO SO OOOOO
O0.0 O0.0 O0.0 ONOOO
O0.0 O0.0 O0.0 O FOOO
O0.0 O0.0 O0.0 NOOOO
O0.0 O0.0 O0.0 VOOVO
FOé OON ONO NOovO
OO.v OON ON.O FOOVO
OOé OON VNO OOOVO
O0.0 SO SO OvOVO
O0.0 SO SO NVOVO
O0.0 SO 86 OVOVO
O0.0 O0.0 O0.0 OVOQO
O0.0 O0.0 O0.0 OVOVO
O0.0 O0.0 O0.0 OOOVO
SO SO O0.0 NOO¢O
O0.0 O0.0 O0.0 ONO¢O
O0.0 O0.0 O0.0 VNOVO
86 SO O0.0 NNOVO
SO SO O0.0 O FOVO
O0.0 O0.0 O0.0 OFOVO
O0.0 O0.0 O0.0 OOOVO
O0.0 O0.0 O0.0 NOOVO
OOé ONN FOO 3.8..

E 3:01

 

 

 

.020". 00.0:

 

 

000000 000000082. 000000-00¢ 000 000,00 0000 0.00.. .< 0000000.

72

 

 

 

 

 

.0. 00.0 0~.. ~00 00.~ 0~.0 0.0.0
8.0 8.0 00.0 8.0 8.0 8.0 0.0.0
00.. 00.0 -.. 00.0 00.~ 0~.0 -0-
8.0 8.0 8.0 00.0 8.0 8.0 .0000
8.0 8.0 00.0 00.0 8.0 00.0 00000
0~.. 00.0 0... 00.0 ~0.~ -.0 00000
8.0 8.0 8.0 8.0 8.0 8.0 ~800
~0.. .00 ~00 00.0 ~0.~ 00.~ -0.0
0~.. .00 00.0 00.0 ~0.~ 00.0 0~0.0
0~.. ~00 0... 00.0 00.~ .~.0 08.0
8.~ ~00 00.. ~00 ~0.~ ~00 08.0
00.. ~00 ~0.. 00.0 00.~ .00 .00.0
~0.. .00 0... 00.0 00.~ ~..0 ~00.0
00.. .00 8.0 8.0 00.~ 0.0 ~00.0
0~.. ~00 0... 00.0 00.~ 0.0 000.0
8.0 8.0 8.0 8.0 8.0 00.0 0.000
8.0 8.0 8.0 8.0 8.0 8.0 ~.000
00.0 8.0 8.0 8.0 8.0 00.0 0~0~0
8.0 8.0 8.0 8.0 8.0 8.0 0~0~0
~0.. 00.0 ~00 .00 00.~ 00.~ 0~80
00.. 00.0 0~.0 00.0 00.~ 00.~ 0~80
8.0 8.0 ~0.. 00.0 00.~ 0.0 0.80
0~.. 8.0 ~... 00.0 00.~ -.0 ~.80
8.0 00.0 00.0 8.0 8.0 00.0 0080
0~.. 00.0 0~.. ~00 00.~ 0~.0 ~800
8.0 8.0 00.0 8.0 8.0 8.0 0080
8.0 8.0 8.0 8.0 8.0 8.0 ~800
00.. 8.0 ~0.. .00 00.~ 0~.0 00000
8.0 8.0 00.0 00.0 8.0 00.0 00000
0~.. 00.0 0... 00.0 00.~ 0~.0 ~0000
8.0 8.0 8.0 8.0 8.0 8.0 -80

00.200230. 00.00305 030.. 000: 000000 00.0200 0~.0 03.... 000: 002 33¢

.020“. 00.0:

 

 

000000 000000082, 000000-00¢ 000 .008 0000 0.03 .< 0000090

73

 

 

 

 

 

cod cod 00.0 8.0 cod cod one;
cod cod cod cod cod cod moo—b
8.0 mad we... 56 wed «fin hmowm
mmé rod or... mmé ~0.~ Nu.” one—b
mné 5.0 on... 09v mod mmd one 3
cod cod 8.0 00.0 cod cod vmowm
cod oo.o cod cod cod cod So 5
cod cod cod cod cod cod omorm
mné omd 0~.0 59? mod vmd 53 5
no... mmd 0~.0 mod med and 03 3
an... and 3.... mvé mod ON...” m3 5
3... 5.0 8.0 0~.0. NON Sam N3 5
nae med mNé 09v and vmd one 3
hmé avd cw... Fwd mod vud «no 5
0~..? de 0N... and 3:0. cod 0N0;
mm... 5.0 vué Pod 00.~ gum mwowm
amé wmd owé 03.? and 0.2m N33
8.0 5.9 mm... end cod no... :03
20.23802: CO.QCOE.D :Oufll :00: 30:8 Out“ 03m 05m 580$ :00: g 8301
39.0..— 005!

 

 

000000 000000082, 000000-000 000 .050 0000 0.00.. .< 0000002

74

 

 

 

 

 

00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.. 00.0 00.0 .00 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. .00 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 .0000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. 00.0 00.0 00.0 00.0 .00 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
.00 00.0 00.0 00.0 00.0 .00 0.00.

00.200230. 00.00205 00.0.. 000: 5.008 0300. 00.0 020.. 000: 02¢ 0.000

0.20“. 002926

 

 

000000 000000082. 000000-000 000 .008 000.. 0.0% .< 0000000

75

 

 

 

 

 

00.0 00.0 00.0 00.0 00.0 00.0 0.0.0
00.0 00.0 00.0 00.0 00.0 00.0 0.0.0
00.0 00.0 00.0 00.0 00.0 00.0 0000.
00.. 00.0 00.0 .00 00.0 .00 .0000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. .00 00.0 00.0 00.0 00.0 000.0
00.. .00 00.0 000 00.0 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.0 00.0 00.0 000 00.0 00.0 .00.0
00.. 00.0 0.0 00.0 000 0.0.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.. 00.0 00.0 00.0 00.0 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
.0... .00 00.0 ...0 00.. 00.0 00000
00.. 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 .0000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
.0. 00.0 00.0 .00 00.0 .00 00000
00.. 00.0 00.0 00.0 00.0 00.0 00000
.0. 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 000 00.0 00.0 00000

00.20.00.000. 00.000000. 0800 000: 00008 am 00.0 00.00 000: 002 0.000

0.20". :02u3|>m

 

 

000000 000000082. 000000.000 000 .008 000.. 0.00. .< 0000......

76

 

 

 

 

 

cod cod cod cod cod cod Chara
cod cod cod cod cod cod moo—b
mod mvd mmd mwé amfi 5N smog
8.0 and 5.0 Ned 8.~ DEN one;
00.~ de 56 «ad omN 3N mac 5
cod 00.0 8.0 cod cod 00.0 #003
cod cod cod cod cod cod 50 «m
cod cod cod cod cod cod one—.0
«~... and mvd 05m smd and 53 3
cod cod cod 00.0 00.0 00.0 033
we... end wed mod mod 3N m3;
nmé and mud N30 and 00.~ Nvowm
mm... 90.0 de ms.” and «md 80 3
00.0 00.0 00.0 00.0 cod 8.0 305
Nod mvd and mud and ~0.~ ONO—b
mhé mad and end me 00.~ m3 5
wmé end mud mvd and 3N N3 5
$0.. and and mod 00.0 8.0 :03
0.0.2383... .00—0:055 53-0— :005 3.230 outm gum 0N.” :30... :00! 02< 3301
J. 380“. eggm

 

 

69.000 09.0308; 3800.30. 000 .260 000.. ovum. .< .0000me

77

 

 

 

 

 

00.. .00 00.0 0.0 0... 00.0 00000
00.. .00 00.0 00.0 00.. 0.0 00000
00.0 .00 00.0 0.0 00.. 00.0 0.0000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. .0.0 00.0 0.0 00.. 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. 00.0 00.0 0.0 .0. 00.0 00000
00.0 .00 00.0 0.0 00.. 0.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.. .00 00.0 00.0 00.. 0.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. .00 00.0 0.0 00.. 00.0 00000
00.. .0.0 00.0 0.0 00.. .00 .0000
00.. .00 00.0 0.0 .0. .00 00000
00.. .00 00.0 00.0 00.. 00.0 00000
00.. .00 00.0 00.0 00.. 0.0 00000
00.. 00.0 00.0 00.0 00.. 0.0 00000
00.. .00 .00 00.0 00.. 0.0 00000
...0 00.0 00.0 0.0 00.. 00.0 00000
00.. 00.0 .00 00.0 00.. 0.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. 00.0 00.0 00.0 00.. 0.0 00000
.0. 00.0 00.0 00.0 00.. 00.0 00000
00.. .00 00.0 0.0 0... 00.0 00000
00.0 .00 00.0 0.0 0.... 00.0 0.000
0.0 .00 00.0 0.0 .0. 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. 00.0 00.0 ...0 00.. 00.0 00000
00.. 00.0 00.0 00.0 .0. 00.. 0.00.

00.200230. 00.000000 00.00 0.00.... 0.000030. 00.0 00.0.. 000: 02¢ 0.000

30000.80

 

 

000000 00.000082. 000000-000 000 .008 000.. 0.00.. .< 000000..

78

 

 

 

 

00.0 00.0 00.0 00.0 00.0 00.0 0.0.0
00.0 00.0 00.0 00.0 00.0 00.0 0.0.0
00.. .00 00.0 00.0 00.. 0.0 00000
.0. .00. 00.0 00.0 00.. 0.0 .0000
00.. 00.0 00.0 00.0 00.. 00.0 00000
00.. 00.0 00.0 00.0 00.. 0.0 00000
00.. .00 00.0 0.0 00.. 00.0 00000
00.. .00 00.0 0.0 00.. 00.0 000.0
.00 .00 00.0 0.0 .0. 00.0 000.0
00.. .00 00.0 0.0 00.. 00.0 000.0
00.. .00 00.0 ..0 00.. 00.0 000.0
0.0 00.0 00.0 0.0 .0. 00.0 .00.0
00.. 00.0 00.0 00.0 00.0 0.0 000.0
00.. 00.0 00.0 .00 00.. ...0 000.0
00.. 00.0 00.0 00.0 00.. 00.0 000.0
00.. 00.0 .00 0.0 00.0 00.0 0.000
00.. .00 00.0 0.0 00.. 00.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. 00.0 00.0 0.0 00.. 0.0 00000
00.0 00.0 00.0 0.0 ...0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 .00 00.0 00.0 .0. 00.0 0.000
00.. .00 00.0 00.0 00.0 00.0 0.000
.0. 00.0 00.0 .00 00.. 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
.0. 00.0 00.0 00.0 00.. 0.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
0.0 00.0 00.0 0.0 00.. 00.0 00000
00.. 00.0 00.0 00.0 00.. 00.0 00000
00.. 00.0 00.0 .00 00.. 0.0 00000
00.. .00 .00 00.0 00.. 00.0 00000

00.200330. 00.000000 00.0.. 000... 0.0000 000m 000m 00.0 00.0.. 000.... 002 0.000

3.000.000

 

$0000 0.0000003 000000-000 000 .050 000.. 0.0.3 .< 0000002

 

 

79

 

 

 

 

00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.0 .00 00.0 0.0 00.. 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
0.0 00.0 00.0 0.0 00.. 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.. .00 00.0 0.0 00.. 00.0 000.0
00.. .00 00.0 00.0 00.. 0.0 .00.0
00.. .00 00.0 00.0 00.. 0.0 000.0
00.. .00 00.0 0.0 00.. 00.0 000.0
00.. .00 00.0 00.0 00.. 00.0 000.0
00.. .00 00.0 .00 00.. 0.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.. .00 00.0 00.0 00.. .00 000.0
00.. .00 00.0 00.0 00.. 0.0 000.0
00.. .00 00.0 0.0 00.. .00 000.0
00.. 00.0 00.0 0.0 00.. 00.0 0.0.0
00.0 00.0 00.0 00.0 00.0 00.0 0.0.0
00.0 00.0 00.0 00.0 00.0 00.0 ..0.0

00.200220. 00.000000 00.0.. 000: 000000 00.00 0W0 00.0 00.00 000: 02¢ 0.000

3.000.000

 

 

09.05 09.80.0025 000000-000. 000 .030 000.. 033 .< 00000.5(

80

 

cod
cod
cod
cod
8.0
8.0
cod
00.0
of?
cod
cod
mmé
cod
cod
9.:
N0.0
00.0
00.0
cod
cod
mm...
00.0
cod
00,—.
cod
cod
cod
00.0
cod
cod
00.0

:223235 5.2055 :3..- :002 Eocen— Guam Guam cum sun-m :00!

00.0
cod
cod
cod
00.0
8.0
8.0
00.0
no.0
00.0
00.0
no.0
ood
cod
00.0
mod
8.0
8.0
00.0
00.0
00.0
00.0
00.0
N0.0
cod
cod
cod
no.0
cod
cod
00.0

00.0
00.0
8.0
cod
cod
cod
cod
cod
and
8.0
00.0
05.0
8.0
00.0
and
and
cod
cod
8.0
8.0
«~..0
cod
cod
95.0
00.0
cod
00.0
00.0
00.0
cod
cod

cod cod cod 0~80
cod cod cod nuowm
00.0 00.0 00.0 .3000
00.0 cod cod 0300
00.0 00.0 00.0 «Noon
ocd cod cod 058
00.0 cod cod @800
8.0 cod 00.0 0008
9.0 no.0 ~0.~ 0800
8.0 cod cod mmomm
8.0 00.0 cod crown
0w .0 cm... 00.~ «comm
cod cod cod 00000
8.0 cod cod N800;
3.0 mm... ~0.~ 0800
m v.0 mmé 00.~ 0000M
cod 00.0 00.0 @0000
cod cod 00.0 5000.0.
cod cod cod 00000
cod cod cod m0o0m
«0.0 ..0.0 ~0.~ 00°00
cod cod cod 00000
8.0 cod 8.0 5000
o _. .0 5.? 00.~ 080»...
8.0 00.0 cod 0800
00.0 00.0 8.0 NNo0m
00.0 00.0 8.0 m 500
no.0 mné 0~.~ 050.0.
8.0 00.0 8.0 800m
00.0 00.0 cod 3000
8.0 cod cod 059

62‘ 8..-OK

 

 

 

inning-0...

 

 

09.000 09.000000; 000000001 000 .050 000.. am... .< 500002

81

 

 

 

00.0 00.0 00.0 00.0 00.0 00.0 0.0.0
8.0 00.0 00.0 00.0 00.0 8.0 0.0.0
«0.0 00.0 00.0 0.... 00.. 00.~ «0000
00.0 00.0 00.0 00.0 00.0 00.0 .0000
00.0 8.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 8.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 «..0.0
8.. ..0.0 00.0 00.0 00.. 00.~ 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 .00.0
00.0 00.0 00.0 00.0 00.0 00.0 ~00.0
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00.0 00.0 8.0 00.0 00.0 00.0 000.0
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.. 00.0 00.0 2.0 00.. «0.0 0.000
00.0 00.0 00.0 00.0 00.0 00.0 0~0~0
00.0 00.0 00.0 00.0 00.0 00.0 0~0~0
00.0 00.0 00.0 03 00.. 00.0 0800
00.0 00.0 00.0 .~.0 ~0.~ 00.0 00000
00.0 00.0 00.0 00.0 00.0 8.0 0.000
00.0 8.0 00.0 00.0 00.0 8.0 :80
00.0 00.0 8.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 8000
8.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 8.0 00000
00.0 00.0 8.0 00.0 00.0 00.0 0000..
00.0 00.0 00.0 00.0 00.0 00.0 0000..
020000230. 00.000000 00.00 000.... 000000 0300 00.0 00.00 000... 02¢ 0.00..
3:03—0:85.

 

 

 

 

000000 000000082. 000000-000 000 0008 000.. 0.09 .< 0000000.

82

 

 

 

 

mm... 00.0 {to 3.0 00.~ ~0.~ Che—.0
cod ood cod cod cod 8.0 @005
00.0 cod cod cod cod 8.0 smog
00.0 00.0 00.0 cod cod cod 0005
8.0 8.0 cod ood cod cod mmowm
cod cod cod cod 00.0 8.0 0903
cod cod ood cod cod 00.0 505
cod cod 00.0 00.0 cod 8.0 one;
cod 00.0 00.0 cod cod cod N00 5
cod cod cod cod cod 8.0 0005
00.0 00.0 8.0 00.0 8.0 cod 00.003
cod cod 00.0 00.0 cod ood N003
cod cod cod ood cod 8.0 0005
00.0 00.0 00.0 00.0 cod cod 0.8 5
cod cod cod cod cod cod omowm
cod 00.0 00.0 00.0 00.0 8.0 m 35
cod cod 00.0 00.0 cod cod N33
flaw 096 Car 00.0 mod and 2.03
00.22.23... 00.00055 00.00 :00! 3.2.00 0u0m 0d0m 00.0 00.0.. :00: 00.3 3.31
300.205....

 

 

09.000 00.000083 000000001 000 .050 000.. 9&3 .< 0000009..

83

 

 

 

 

 

.0. 00.0 00.0 00.0 0... 00.0 00000
00.. 00.0 00.0 00.0 00.. 00.0 .0000
0.0 00.0 00.0 ..0.0 0... ..0 .0000
00.. 00.0 ..0.0 00.0 0.... 00.0 00000
00.. 00.0 00.0 0~.0 00.. ~..0 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
00.. 00.0 00.0 00.. 00.. 0.0 00000
.0. 00.0 00.0 0... 00.. ~0.~ 00000
0~.~ .00 00.0 .0... 00.. 0~.~ 00000
00.. 00.0 0.0 ~.... 0.... 00.~ 00000
00.0 00.0 00.0 00.0 00.0 00.0 0.000
0.0 .00 00.0 ~0... 00. 0.0 00000
00.. 00.0 00.0 0~.. 0... ..0 00000
00. 00.0 ..0.0 0~... 00.. 00.~ 00000
00.. 00.0 00.0 00.0 N... 00.~ .0000
~.... ~00 0.0 0.... 00.. .00 00000
00.. ..0.0 .00 .0... 00.. .00 00000
0.0 00.0 ~00 0~.0 00.. 00.. .0000
00.. 00.0 0.0 0.... 0... 00.. 00000
00.. 00.0 00.0 0~.0 0.... ...~ 00000
~..~ ~00 00.0 00... 00.. 0.0 00000
0.0 00.0 .00 00.. 00. 0.0 00000
00.. 00.0 .00 .0... 00.. ~..~ .0000
0.0 00.0 .00 ~0... 00.. 0.0 00000
00.. 00.0 00.0 .~.0 0... 00.0 00000
00.. ..0.0 00.0 00.0 .0. 0~.~ «00.....
00.0 ..0.0 00.0 0~.0 0.... 0.0 0.000
0.0 00.0 00.0 0~... .0. ~..~ 0.000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.. 00.0 00.0 00... 00.. 0.0 00000
00.. 00.0 .0. 00.0 .0. 0~.~ 0.00.
00.200230. 00.000000 00.00 000... 0.40000 muwmlmufl 00.0 00.00 000: 00.< 0.000
0005

 

000000 000000002. 000000.000 000 .0000 000.. 0.00. .< 000000..

 

 

 

 

 

 

Sam 56 3.0 uni 8.0 0~.~ owopm
maé mod wad «Ni :4 mod owowm
~0.~ and 8.0 and no.0 9am muons
mmé mod and 0~.? :6 mod 080m
0~..? 3.0 and omé .0... mod 0008
mm... mod cod mad «5.. mod 8000
«we mod N50 m _. .V 3hr Nae «comm
8.0 and end 0:0 0~... Na... whore
ma... de cod cod 3... 8.0 0005
cod cod cod cod cod 86 085
0»... mod 3.0 vmé ow... 00.~ moo—b
8.~ wad nod mod an; omé .085
3.0 and and mfiv 5... 3d «more
mm... and mad an? an... Sam 209.0
5N mad and mmi oné No.0 0300
30 mad and mad 90.0 8... gown
3N and mod NNé nu... «ma ~53
cod cod cod cod cod cod awomm
gum and 3.0 56 3.? Sum «Noun
mm... and mad vué «ad 3 .N 0~80
cod cod 8.0 8.0 86 cod mwoom
cod and and 92¢ no... mod Soon
«9.. wad Rd 36 «~..0 no... Coon
no... and and and 8.. MNN mooom
cod 8.0 cod cod cod cod booom
SN 36 mad mud mm... ~0.~ mooom
ma... N0.0 Fwd omé 0~..? bod «coon
5... 8.0 and 9.40 mm... mm... 80$?
5... and and 3d mné rod mmomv
m _. .N and and and N90 m _. .N «08¢
om; de Nod 5d wné VN... macaw
32:50:23... 0.0.2.250 £300. :3! £2.00 Guam yum 0N5 5300. :00: no.2 330m.—
cue:

 

 

.mmcmso 09.030025 3030.001 000 0030 000.. 030.. .4. 50009;?

85

 

 

 

 

 

ddd ddd ddd cod ddd cod chord
cod and nod mod 3... mm... ”cord
mod Ed and mmé no... 3. .N hmo 5
mm... mod Nod 0N6 cod 3... 805
mm.. 8.0 had omé 5.? NF .N one 5
nae mod vwd vwé dmé add woo Pd
00.~ and odd cud 0~.0 8.~ woo;
dod Odd dod cod ddd ood dmdwd
00.~ mmd 05d 3..? 8... N0.0 sword
hmé and mod 840 d5? use @003
mm... and 3d mud No.0 8.~ 9005
mmé odd mmd mué 2.6 d _‘ .N N35
and mod and vmé do... nod mad 3
3.0 and add mud hm; 9N 3”de
NON and omd 5.00 do... 3N dwdwm
ood cod cod ddd ddd Odd 9.03
mmé odd vwd mNé NE? mod N3 5
Odd ddd ddd ddd ood cod :0 3
0.0.9.3830.- .00—0:055 £300. :00! 3.2.8 Guam ouum 03% £300. :00! 00.2 3:01
.395

 

 

mmcmco 09.009000; $802.81 can .900 new; 0304 .< fiocmmg

86

 

 

 

 

 

do. med mod ..Né m5. 00.~ ouomv
cod Odd dod add 8.0 ood Room
ood ood ood ood ood ood Room
.00.. «ad mm d «dd om. and vmocm
ood Odd ood ood cod ood «moon
cod ood ddd dod ood ood goon
ood ood ood ood ood ood 0008
Odd Odd ood 8d 8.0 ood 80mm
ood ood ood cod ood ood mooom
v0. and on d dud .6. 8.~ $33
8.0 ood ood ood ood ddd 0.9mm
ood ddd cod ood Odd cod «comm
mod dvd Ed o..v N9. .oN «nova
00.~ and add 3.0 00.. mmN «down
8.. mad Ed «of #0.. and .83
«m. . and mud d . .v «a. mod amovm
3.. avd .md on.” N... 8.~ 2003”
2.. GVd and «.6 on. SN 50900
3.. dVd mm d cod 0~.. .mN ovocm
3.. mvd and and I... and mvocn
8.. mad and 3..0 mm. 8.~ ovovm
8.. and .hd odd ood 5N gown
m..N .md dvd mud 8.. 30d smog
3.. 90d and mud I... and 8000
Odd dod cod cod ood cod <83
#0.. dvd and and 9... .mN «nova
dod ood ddd cod ood ddd 0.90m
mod .md mod mad no. and 0.90»
ddd ood ood ood ood ood add;
8.0 cod ood ddd ood cod «down
8.~ 90d 3d «~..0 2.. .dN v.00.
:Eflomloac. £280.50 :30.“— :00! 3.8.00 Guam ouudm. an.» :3..— .305 .92 8:01
tea—.03

 

 

.0036 000000083 000000-000 000 .060 003 0.00.. .< 00080.0

87

 

 

 

 

 

3.. and v.d dd.« 8.. mv.« 035
no. .md .w. o d... .0 m.. . «m. « d.o.m
do. 0 cod 8. o oo. d 8. o oo. o ««o«h
oo. 0 Odd cod ddd cod ood .800
do. 0 Odd ood Odd ood ood 0008
Odd ood ood ood ddd ood mmooo
ood ood ddd dod ood ood «come
no... and 5d and o... «m.« «no.0
mm. «md mmd .d.» am. .e.« one .m
«a. .nd «a d and «0.. mv.« mod .0
dd.« and cod 2.0m a... .m. « and;
cod ood cod 8. d cod 8. d .85
0... .nd 30d 3. n no. . av. « «new
«a. .0d .md ~... m 8. . om. « «3.0
«o.« med vmd «w. m an. . .m. « 900.0
ddd cod cod 8. o do. 0 8. d 203
so.« d«d 8.0 mm. 0 ms. . .m. « «.03
3.. dvd and on. n 2.. . .m. « d«o«m
ddd ood Odd oo. o oo. o oo. o o«d«n
8.. .0d d... .00.? 8.. an. « 080....
on. .md .«.. 8.? 00.. mod m«oom
od.« .md m... omé «Q. «n.« 0.8m
.9. «ad o«.. .56 .5. . mod «.08
8.. dcd cod m«. c 8. . SN mooom
md.« 90d .sd 8. v 8.. on.« .800
3.. .Md and .m. m «0.. om.« madam
v... 3.0 and and mm. ov.« «doom
0.0.. on d and 5d 00.. hv.« 30$?
«9. .m d end end 20.. d¢.« mmomv
«~... .nd 3.0 0nd 3.. «o.« «meow
90.. .md and and on. «m.« ««om¢
P3.32.9.3... €23.53 :3... :00! 3.2.00 Guam ouum 08m sou-0. coo! .5 330m.
Den-.03

 

 

69.2.0 09.20025 0032.-qu new .030 new.— o.wo._ d. 59.32

88

 

 

 

 

 

oo.« on ood 3d 3.. .m.« oho.o
no. .md hvd «~..0 «0.. om.« moo .o
«N. .md ood o«.v no. on.« go .o
#9. «ad .md 2.6 m... o¢.« omo.o
ood ood ood ood ood ood mmo.o
3.. .od mcd on...” 8.. om.« 32o
ood ood ood ood ood ood .mo.o
ood ood ood ood ood ood omo.o
oo.. .md and ood of. ood .33
00.. .md ood ood 8.. on.« 0.53
do. .0 d ood «ad on. m¢.« m3 .o
~.... .md «od .od «0.. .m.« «.53
«E. .md ood 3.0 «E. «m.« ood;
«0.. o...o vmd «dd h... .m.« 3o;
8.. .md ood .«.v 20.. and o«o.o
oo.. .md .md and av. mm.« m.o.o
00.. .md o«d mvd .m.. «v.« «.o.o
ood ood .md 3...” oo.. m¢.« ..o.o
60-83202..— cO_O—__OE_O Saul :00! £09.00 CHOW Gum .N—M flail 2.0: 65 3:0“
vac—.0;

 

 

09.000 000.2083 000000-000. 000 .900 000.. 03% .< 0000002

89

 

 

 

 

 

.o. . .o. o ««. o ««.m .o.. 3.. o«oov
.o. . «n. o at. o ood ood oo. . .«oom
oo. o oo. o oo. o oo. o ood oo o shown
oo. . of o av. o on. n no. on. . v«oon
oo. o oo. o oo. o oo. o ood oo. o «Noon
ood ood ood ood ood ood o.oom
oo. . .n. o mm. o mm. n oo.. oo.. oooom
oo. o oo. o oo. o oo. o ood ood oooom
oo. o ood oo. o ood ood ood mooom
oo. o ood ood ood ood ood o«omm
oo.. .o. o sad no.» oo.. oo.. o.omm
vo.« of o .m. o mud «o. no. . «comm
ood oo. o oo. o ood ood oo. o voovm
ood ood oo. o oo. o ood oo. o «coco
no. . .o. o end on. o oo.. oo. . .83
no. . .m. o ood oo. o 2... oo. o omovm
ood oo. o ood ood ood ood otovm
«o. of o vvd ood oo.. oo.. .33
o... ovd on mud oo.. 2... ovovm
5.. «ad and and .o.. .o.. ovowm
oo.. .nd ovd inn oo.. .o.. ovotm
oo.. omd mm d mm.» C... 0.0.. omovm
ood ood ood ood ood ood booed
ood ood ood ood ood ood o«o¢m
ood ood ood ood ood ood ««ovm
ood ood ood ood ood ood ««ocm
ood ood ood ood ood ood 9oz
ood «od omd 3.» o... «0.. 9on
ood ood ood ood ood ood oootm
ood ood ood ood ood ood «oovm
oo.. .md and on.» o... 2... god.
:28‘235 c2235....— zficgi 3.0.30 ouum outmf on.» scan... :00! a2< 3:3...
.033

 

 

000000 0000.008; 000000-000 000 00.60 000.. 0.0.0.. .< 000003

90

 

 

 

 

 

K... 00.0 N0.0 N0.0 2.... me... 0.03
00.~ 5.0 .00.0 00.0 0~..? 0%.? 0.05
N0.0 ..0.0 00.0 0.0.0 N0.0 00. .. «Noun
00.0 00.0 00.0 00.0 00.0 00.0 .0000
N0 .0 ..0.0 0~.0 00.0 E... 3... 00000
00.0 00.0 00.0 00.0 00.0 00.0 00000
00... ..0.0 00.0 N0.0 00... 00.0 N800
00.0 00.0 00.0 00.0 00.0 00.0 «50.0
N0... 30 00.0 F06 00.0 00.0 05000
00... 0.0.0 30.0 05.0 50.0 00... 000 ..0
00... 00.0 .00.0 .00.0 ..0.0 00... 00000
00.0 00.0 00.0 00.0 00.0 00.0 .0000
«~..0 00.0 N0.0 00.0 00.0 00.0 «00.0
00.0 00.0 00.0 00.0 00.0 00.0 ~35
00.0 00.0 00.0 00.0 00.0 00.0 03.0
00.0 00.0 00.0 00.0 00.0 00.0 0 .03
00... 00.0 00.0 00.0 ~0... 0~.0 «00.00
00.0 00.0 00.0 00.0 00.0 00.0 0~0~0
00.? 0.0.0 00.0 00.0 00.0 00.0 0~0~0
00... em .0 0.0.0 00.0 00... 00... 0~80
3..? N0.0 0.0.0 0~.0 «0... E... 0~80
t? 00.0 00.0 9.0 N0... 00... 00000
00... ..0.0 0?.0 0~.0 V0... 0~.0 5000
on... 00.0 00.0 N0.0 00... 00... 00000
00.0 00.0 00.0 00.0 00.0 00.0 3000
00... 0‘0 3.0 :0 00.9 00... 00000
00.0 00.0 00.0 00.0 00.0 00.0 N800
00.~ ..0.0 P0 .0 05.0 0~.0 50... 0000.0
05.0 ..0.0 0~.0 3..? 00... Cd 0000c
00... ..0.0 00.0 v0.0 #0... 3.0 «000.0
00.0 00.0 00.0 00.0 00.0 00.0 «~00.0
00.20%...30. 00.8.5.0 £800. :00! 5.2.00 Guam Guam 0~.0 :3... :3! 02¢ 3:3.
.303

 

 

09.05 00.000082. 000000-000. 000 .060 000.. 2% .< 0000002

91

 

 

 

 

 

t... ..0.0 «10 00.0 «0... 00... 050 ..0
00... N0.0 00.0 ..0.0 v0... 00.0 000.0
nvé ..0.0 010 05.0 #0.? 00... N00 00
00.0 00.0 00.0 00.0 00.0 00.0 00000
00.0 00.0 00.0 00.0 00.0 00.0 000.0
50... 00.0 00.0 0.20 00.? 00... V0000
00.0 00.0 00.0 00.0 00.0 00.0 003
00.0 00.0 00.0 00.0 00.0 00.0 000.0
00... ..0.0 010 05.0 00... 00.? ~33
at... ..0.0 00.0 00.0 0~..? 00... 0300
3.0 ..0.0 00.0 00.0 00.0 00... 03.0
00... ..0.0 00.0 00.0 00... N0... «.003
00.0 00.0 00.0 00.0 00.0 00.0 0003
00.0 00.0 00.0 00.0 00.0 00.0 .0003
R... ..0.0 00.0 00.0 00... 05.0 0~0.0
0~.... ..0.0 0?.0 0~.0 00... ~0.~ 000.0
00... N0.0 0Y0 NRO N0... 00... ~..0 ..0
00.0 ..0.0 00.0 00.0 00.0 00.0 ..0.0
cO-BOqcss =O_OCOE_O ..le 0:! £008 out” CHEW Cum gain :00! Is Gan—OK
53-;

 

 

.omcmco 09.081082, 00009.69... 0:» .050 0:3 903 .< 50:00.?

92

APPENDIX B

HEIRARCHICAL CLUSTERING ANALYSIS AGGLOMERATION SCHEDULE

93

 

ndix E. Hierarchical Clusterin Ana

 

 

lomeration Schedule.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

__ Stat? 6193*"
Cluster Combined Coefflclents First Ap‘pears Next 801931
322: Cluster 1 Cluster 2 Cluster 1 Cluster 2
1 41 60 115-04 0 0 19
_ 2 63 67 0.0001 0 . 0 14
3 f 14 _26 l 0.0001 0% 0 _ 22
.._4. ___ ‘__16___F 31 0.0001 0 _ _0_ _ __ 35 _
5 l 13 76 0.0002 0 ___‘_fi_ 0 7
6 33 77 0.0002 0 0 21
7 __ 13 61 0.0004 5 0 25
6 24_ _l 47 0.0004 _0 0 51
_9 .. 36 50 0.0005 0 ___0 23
10 6 25 0.0005 ._0 _ 0 43
11 27 76 0.0005 0 0 4 14
12 -L. 20 75 0.0005 __ 0 _ _ 0 32
W3 62 72 0.0006 0 0 27
14 27 _‘ 63 0.0006 11 ____ 2 22
15 1 2 0.0006 0 0 36
_16 43 71 0.0009 _ 0 _ __ 0 36
17 6 46 0.0009 0____ 0 45
16 29 42 0.0009 0 ___0 26
19 11 41 0.0009 0 1 33
20 23 36 0.0011 0_»_#__ _ 0_ 39
___21 33 40 0.0013 __ -__ 6_ __ __ _0 30
22 14 27 0.0017 3 _ 14 44
23 36 73 0.0016 9 fl 0 51
24 7 52 0.0016 _0___~ _ 0 32
25 13 60 l 0.02 7 _ l 0 43
267 15 34 0.002 0 ‘_ 0 40
27 62 70 0.0026 13 ¢_ __ 0 33
26 29 32 0.0026 16 0 35
29 9 46 0.0029 0 0 36
30 33_ 66 0.0033 21 0 49
31 54 74 0.0033 0 0 40
32 7 __ . 20 0.0033 24 12 37
.33.... -11 _ 62 0.0041 19 " _ 2.7 __49_
34 5 21 0.0046 0 0 56
35 1.6 29 0.005 4 26 56
36 1 9 0.0051 15 29 46
37 __, 7___ > 37 0.0054 32 0 53
36 43 53 0.0056 16 1 0 46
39 4 23 4_ 0.0059 0 - ___20 _ 57
40 , 15 54 0.0059 26 31__ _ 44
41 10 44 0.0066 0 0 J 64
42 16 _ 22 0.0066 0_ __ __ 0 l _ 63
_ 43 6 13 0.0066 __19_ _ _25 54
L34.-.“ __14 _ 15 0.011 22 40_ 59 _
45 f_ -43.--- > _ _69 _ 0.0123 17 0 f _ 55
46 11 49 0.0126 33 _ 0 ’ 60
47 12 17 0.0145 0 0 66

 

 

 

 

 

 

Appendix 1'3. Hierarchical Clustering Anawis Agglomeration Schedule.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_ __._ -.§tateguster ._

Cluster Combined Coefficients First Appears _ Next Ste 6

Sta e Cluster 1 Cluster 2 Cluster 1 Cluster 2 I
48 1 43 0.0149 36 38 61
49 19 33 0.0158 70 30 59
- 50 59_ 79 0.017 0 . 0 _ 70
51 24 ‘ 38 0.0181 _ 8 #_ _ _23 . 61
52 7__ ____ _30 64 0.0201 0_ 0 74
..53 7 66 0.0227 37 _L 60
54 8 __ _-__§7 0.026 43 0 70
55 __ 6 39 0.0266 45 0 63
56 __ _5 16 0.0274 34 _35 68
57 4 __ 45 0.0278 39 0 66
58 51_ 56 0.0297 0 0 72
59 -. 14 4 __ _1_ 9 0.0297 44 49 67
60 -_ 7 1 1 0.0422 53 46 69
61 1 24 0.049 _ 48 51 67
62 3 55 0.0509 0 0 71
63 6___ 18 0.0579 _ 55 42 69
64 10 65 0.0629 41 _ __0 76
65 _ 35 58 0.0635 0 _0 71
66 4 _ 12 0.0797 57 47 68
- 67 -0 ____1_ _;___ _1_4 0.0986 61 59 72
68 4 5 0.1109 66 56 73
69 _ _ _ _6 _ 7 .- 0.1778 63 60 73
70 __1_ __ __ _8 59 0.2308 _ _1 _ 54 - 50 76
71 3 35 0.2441 62 65 75
72 1_ _ __1 51 _ 0.2829 67 58 75
73 fl __ __4 _ 6% 0.293 68 69 77
74__ 1 26 30 0.4253 0 52 77
1115 1 _ 3 0.6845 _ 72_ 71 78
76 8_ 7 __1_0“ 0.689 70 _ 64 79
9.7—7 +1 7 4 26 1 .0788 73 _ 74 78
78 1 4 2.1051 75 77 79
79 1 8 5.9557 78 76 0

 

 

 

95

LITERATURE CITED

96

Adams, R. 2000. Oral interview. September, 2000. Kalamazoo Nature Center.

Anderson, J.R., E.E. Hardy, J.T. Roach, and RE. Witmer. 1976. A land use and
land cover classification system for use with remote sensor data.
Geological Survey Professional Paper 964. US. Government Printing
Office. Washington, DC, USA.

Bohning-Gaese, K. 1997. Determinants of avian species richness at different
spatial scales. Journal of Biogeography. 24: 49-60.

Brown, B., M. Koenen, and D.W. Mehlman. 1999. Wings of the Americas: Red-
headed woodpecker. http:/Iwww.tnc.orglwingslwingresouroelrhw02.htrnl.

Brown, J.H. 1984. On the relationship between abundance and distribution of
species. American Naturalist. 124: 255-279.

Brown, J.H. and M.V. Lomolino. 1998. Biogeography. Sinauer Associates,
Sunderland, Massachusetts, USA.

Burrough, PA. and RA. McDonnell. 1998. Principles of geographic information
systems. Oxford University Press. New York, New York, USA.

Campbell, NA. 1996. Biology. The Benjamin! Cummings Publishing Company,
lNC. Menlo Park, Califomia, USA.

Cunningham, WP. and B.W. Saigo. 1995. Environmental science; aglobal
oonoem. Wm.C. Brown Publishers. Dubuque, Iowa, USA.

Diekoff, George. 1992. Statistics for the social and behavioral sciences:
univariate, bivariate, multivariate. Wm.C. Brown Publishers. Dubuque, IA,
USA.

Ehrlich, RR, 0.8. Dobkin, and D. Wheye. 1992. Birds in jeopardy: the
imperiled and extinct birds of the United States and Canada, including
Hawaii and Puerto Rico. Stanford University Press. Stanford, California,
USA.

Elkie, P.C., R.S. Rempel, and AP. Carr. 1999. Patch analyst user’s manual: a
tool for quantifying landscape structure. Queen’s Printer for Ontario.
Ontario, Canada.

Environmental Systems Research Institute, Inc (ESRI). 2000. Arc/INFO 7.2.

Environmental Systems Research Institute, Inc (ESRI). 2000. ArcView 3.2.

97

Environmental Systems Research Institute, Inc (ESRI). 2000.ArcV1ew Export
71.

Environmental Systems Research Institute, inc (ESRI). 2000. Spatial Analyst
Extension.

ERDAS. 1999. IMAGINE 8.5.

EROS Data Center and United States Geological Survey. 2001. 1-km AVHRR
North American Land Cover Characteristics.
http://www.nationalatlas.gov/atlasfip.htrnl

Fallon, J., G. Gouch, J. Hines, J. Sauer, and l. Thomas. 2000. The North
American breeding bird survey, results and analysis 1966-1999. USGS
Patuxtent Wildlife Research Center. Laurel, Maryland, USA.

Flather, CH. and JR Sauer. 1996. Using landscape ecology to test
hypotheses about large-scale abundance patterns in migratory birds.
Ecology. 77: 28—35.

Griffith, B., J.M. Scott, J.W. Carpenter, and C. Reed. 1989. Translocation as a
species conservation tool: status and strategy. Science. 245: 477-480.

Guissard, V. 1999. MILA Grid Utilities Extension. UCU AGRO.
http://arcscripts.esri.coml

Harpstead, M., T.J. Sauer, and W.F. Bennett. 1980. Soil science simplified.
Iowa State University Press. Ames, Iowa, USA.

Hengeveld, R. and J. Haeck. 1982. The distribution of abundance. l.
Measurements. Journal of Biogeography. 9: 303-316.

Hutchinson, GE. 1957. A treatise on limnology, vol. 1. John Wiley & Sons.
New York, New York, USA.

lngold, DJ. 1989. Nesting phenology and competition for nest sites among red-

headed and red-bellied woodpeckers and European starlings. Auk. 106:
209217.

Keller, CM. and J.T. Scallan. 1999. Potential roadside biases due to habitat
changes along breeding bird survey routes. Condor. 101: 50.57.

Kilham, L. 1958. Territorial behavior of wintering red-headed woodpeckers.
Wilson Bulletin. 70: 347-358.

98

Link, W.A., B.G. Peterjohn, and JR. Sauer. 1994. Observer differences in the
North American breeding bird survey. Auk. 111: 50—62.

Lomolino, M.V. and R. Channell. 1995. Splendid isolation: patterns of
geographic range collapse in endangered mammals. Journal of
Mammalogy. 76: 335—347.

Lomolino, M.V. and R. Channell. 1997. Range Collapse, Re-lntroductions, and
Biogeographic Guidelines for Conservation. Conservation Biology. 12:
481-484.

Luijten, J. 2001. Manual Grid Editor Extension. http://arcscriptsesricom/
Microsoft Access. 2000. Microsoft Corporation.
Microsoft Excel. 2000. Microsoft Corporation.

Natural Resource Conservation Service (NRCS). 2001. Major Land Resource
Area (MLRA).
ftp.nhq.nrcs.usda.gov/publoutgoingllandlarc_exportlus48mIra.e00.zip

Page, A. 1996. Status report on the red-headed woodpecker Melanerpes
erythrocephalus in Canada. Committee on the status of endangered
wildlife in Canada. Ottawa, ONT.

Rempel, R. 2000. Patch Analyst (grid) 2.1 build 33.
httpzlfflash.lakeheadu.cal~rrempellpatch

Robert, M. 1989. The threatened birds of Quebec. Environment Canada.
Canada Wildlife Service.

Rome, A. 2001. The bulldozer in the countryside. Cambridge University Press.
Cambridge, United Kingdom.

Sauer, J. R., J. E. Hines, and J. Fallon. 2001. The North American Breeding Bird
Survey, Results and Analysis 1966 - 2000. Version 2001.2, USGS
Patuxent Wildlife Research Center, Laurel, MD. http:/Mvmmbr—
pwrc.usgs.govlbbsl.

Short, LL. 1982. Woodpeckers of the world. Weidner Associates, INC. New
York, New York, USA.

Sibley, DA. 2000. The Sibley guide to birds. Alfred A. Knopf. New York, New
York, USA.

99

Smith, AT. 1974. The distribution and dispersal of pikas: influence of behavior
and climate. Ecology. 55: 1368-1376.

Snyder, J. P. 1993. Flattening of the Earth. University of Chicago Press.
Chicago, Illinois, USA.

Sprinthall, RC. 1990. Basic statistical analysis. 3'“ Additon. Prentice-Hall Inc.
Englewood Cliffs, New Jersey, USA.

SPSS, Inc. 2001. SPSS 11.0.

United States Environmental Protection Agency (USEPA) and United States
Geological Survey (USGS). 1998. 1:250,000 Scale Quadrangles of
Landuse l Landcover GIRAS Spatial Data of CONUS in BASINS.
http://www.epa.govlostlBASlNSlb3webdwn.htm.

United State Geological Survey (USGS). 2001. National Atlas. Breeding Bird
Survey Routes. http://www.nationalatlas.gov/atlas.ftp

United States Geological Survey (USGS). 2001. National Land Cover
Characterization. http:/llandcover.usgs.govlnatllandcover.html

United States Geological Survey (USGS) and Patuxent Wildlife Research Center.
2001. Raw BBS Data. http://wwwmpz—
pwrc.usgs.govlbbslretrievallmenu.cfm

\fillard, M. and BA. Maurer. 1996. Geostatistics as a tool for examining
hypothesized declines in migratory songbirds. Ecology. 77: 59-68.

Whitney, G. 1994. From coastal wilderness to fruited plain. Cambridge
University Press. Cambridge, United Kingdom.

Vifilcox, BA. and DD. Murphy. 1985. Conservation strategy: the effects of
fragmentation on extinction. American Naturalist. 125: 879-887.

Williams, JB 1975. Habitat utilization by four species of woodpeckers in a
central Illinois woodland. American Midland Naturalist. 93: 354-367.

Winkler, H., D.A. Christie, and D. Numey. 1995. Woodpeckers: a guide to
woodpeckers of the world. Houghton Mifflin Co. New York, New York,
USA. -

Wolf, C.M., B. Griffith, C. Reed, and SA. Temple. 1996. Avian and mammalian

translocations: update and reanalysis of 1987 survey data. Conservation
Biology. 10: 1142-1154.

100