ct E :.
V Lam

5...
.

. .. $.52.
. :5:

L . . \hb
51:3, a... ... JFPQHW

hfié.
a!
3..

. ;?§!
.51 , . ..
afifg 1

‘3‘
s..¢lx.tl:
i . . 7,

 

SJ. 5! {.3
. ,5!
3131321].

 

“fifiwémfla‘ 3%.. 5.; 33......“

0 11V

.. : . . _ A , , , m..w§.§£.,. fl.

 

/
’00

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

LANDSCAPE STRUCTURE AND THE BIOLOGICAL
DIVERSITY OF THE NORTH AMERICAN AVIFAUNA

presented by

JENNIFER JEANNE SKILLEN

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Department of Fisheries and

Wildlife and Program in
Ecology, Evolutionary Biology
and Behavior

 

 

mic/(IMAM

Major Professor’s Signature

August 4, 2005

 

MSU is an Affirmative Action/Equal Opportunity Institution

_‘_..—.-.--.-.-.-.—-‘--.-4—.-‘-.—.--.-.--.—~—n-—-- ~ ~—

 

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

 

JAN 1 8 2009

 

010509

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 czicficiomom.m15

LANDSCAPE STRUCTURE AND THE BIOLOGICAL DIVERSITY OF THE NORTH
AMERICAN AVIFAUNA

By

Jennifer Jeanne Skillen

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Fisheries and Wildlife and
Program in Ecology, Evolutionary Biology and Behavior

2005

ABSTRACT

LANDSCAPE STRUCTURE AND THE BIOLOGICAL DIVERSITY OF THE NORTH
AMERICAN AVIFAUNA

By

Jennifer Jeanne Skillen

Patterns of biological diversity are influenced by macroscale processes as well as
localized ecological conditions, including the spatial distribution of landscape elements.
Describing and quantifying the effect of landscape structure on ecosystem attributes such
as biodiversity patterns is a research priority in modern ecology. This dissertation
examines the connections between patterns of landscape structure and patterns of avian
biodiversity in North America. The first goal of my dissertation was to identify metrics of
landscape structure that are appropriate for describing the link between the spatial
heterogeneity of landscapes and the spatial patterns of biological diversity. Then, using
long-term species occurrence data from the North American Breeding Bird Survey, I
tested the hypothesis that the body size distributions of species assemblages display
discontinuities. I found that not only are there discontinuities, but the position of species
in the body mass distribution relative to the location of the discontinuities has ecological
consequences. I then used lacunarity analysis to test the hypothesis that discontinuities in
landscapes structure correspond to discontinuities in body mass distributions. To
accomplish this I again used species occurrence data from the Breeding Bird Survey and
forested habitat distribtion data from the National Land Cover Data set. The results of
this analysis suggest that the forest habitat of eastern North America is not hierarchically

structured and therefore is not the likely cause of the observed discontinuities in avian

body mass distributions. Finally, I examined the hypothesis that the composition of avian
species assemblages conforms to the distribution of ecoregions across North America.
Using seven beta-diversity metrics to describe the species assemblages, I found that the
distribution of the North American avifauna appears to be spatially autocorrclated and

independent of ecoregion distribution.

ACKNOWLEDGEMENTS

Throughout my research leading up to this dissertation, I have been fortunate to
work with many gifted ecologists. My advisor, Brian Maurer, has been an unending
source of ecological and statistical insight. His ability to synthesize concepts and develop
coherent explanations has provided me with both inspiration and a deeper understanding
of ecology. The members of my committee, Katherine Gross, Kelly Millenbah, and Scott
Winterstein have provided me with invaluable assistance, advice, and encouragement. I
am grateful for the research experience and funding provided by the Michigan
Department of Natural Resources and the Gap Analysis Program, as well as the teaching
experience and funding from the Michigan State University Biological Sciences Program.
I would also like to thank my colleagues at Michigan State University and the
Community College of Southern Nevada, especially Genny Nesslage, Katie Kahl, Anne
Axel, and Anita Morzillo. They have consistently been willing to put aside their own
tasks to help me solve problems and find resources and I deeply appreciate their support
and friendship. And finally, I thank my family, especially Bert and Jared, for their

unwavering support and patience that made this possible.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES ............................................................................................................ x
INTRODUCTION .............................................................................................................. 1
CHAPTER 1
QUANTIFYING ECOLOGICALLY MEANINGFUL LANDSCAPE STRUCTURE ..... 5
Pattern and Process ..................................................................................................... 5
Patterns of Biodiversity ............................................................................................... 6
Quantifying Landscape Structure ................................................................................ 8
Landscape composition ......................................................................................... 10
Diversity indices ............................................................................................... 10
Spatial configuration ............................................................................................. 12
Patch area, shape & perimeter .......................................................................... 12
Adjacency ......................................................................................................... 14
Proximity and isolation indices ......................................................................... 14
Contagion .......................................................................................................... 15
Connectivity ...................................................................................................... 16
Lacunarity ......................................................................................................... 18
Influence of Scale ..................................................................................................... 19
Conclusion ................................................................................................................ 23
CHAPTER 2
THE ECOLOGICAL SIGNIFICANCE OF DISCONTINUITIES IN BODY MASS
DISTRIBUTIONS ............................................................................................................ 26
Introduction ............................................................................................................... 26
The Species Assemblage ........................................................................................... 28
Identification and Measurement of Discontinuities .................................................. 31
Holling index ........................................................................................................ 31
Siemann & Brown index ....................................................................................... 32
Robust indices ....................................................................................................... 32
Lump analysis-gap rarity index ............................................................................ 32
Cluster analysis ..................................................................................................... 33
The Patterns .............................................................................................................. 3S
Discontinuity structure .......................................................................................... 35
Declining species .................................................................................................. 36
Non-indigenous species ........................................................................................ 37
Body size ............................................................................................................... 38
Region-species interactions .................................................................................. 39
Mechanisms Underlying Discontinuities .................................................................. 40
The landscape template ......................................................................................... 41
The population response ....................................................................................... 44
Conclusions ............................................................................................................... 49
CHAPTER 3
LINKING LANDSCAPE STRUCTURE WITH THE BODY SIZE DISTRIBUTIONS
OF SPECIES ASSEMBLAGES ....................................................................................... 67

Introduction ............................................................................................................... 67

Methods ..................................................................................................................... 70
Study area .............................................................................................................. 70
Land cover ............................................................................................................ 71
Structure ................................................................................................................ 71
Species assemblages ............................................................................................. 73
Body masses .......................................................................................................... 74
Finding discontinuities .......................................................................................... 74
Method 1 — Holling index ................................................................................. 74
Method 2 — Siemann & Brown index ............................................................... 75
Methods 3 and 4 — robust indices ..................................................................... 75
Method 5 — cluster analysis .............................................................................. 75
Linking lacunarity with body mass ....................................................................... 76
Results ....................................................................................................................... 77
Landscape structure .............................................................................................. 77
Discontinuities in body mass distributions ........................................................... 78
Correspondence of assemblage discontinuities with landscape discontinuities 79
Discussion ................................................................................................................. 81
CHAPTER 4
AVIAN SPECIES COMPOSITION AND NORTH AMERICAN ECOREGIONS ...... 102
Introduction ............................................................................................................. 102
Methods ................................................................................................................... 105
Data sources ........................................................................................................ 105
Data analysis ....................................................................................................... 106
Results ..................................................................................................................... 109
Discussion ............................................................................................................... 112
CONCLUSIONS ............................................................................................................. 133
Patterns of Biological Diversity .............................................................................. 133
Patterns of Landscape Structure .............................................................................. 133
Correspondence of Landscape Structure and Biodiversity ..................................... 134
Future Directions .................................................................................................... 135
APPENDIX A: SPECIES LISTS ................................................................................... 136
APPENDIX B: TAXONOMIC CHANGES ................................................................... 173
APPENDIX C: DISCONTINUITIES IN BODY SIZE DISTRIBUTIONS ................... 175
APPENDIX D: ECOREGIONS OF THE CONTINENTAL UNITED STATES .......... 190
LITERATURE CITED ................................................................................................... 201

vi

LIST OF TABLES

Table 1.1. Metrics quantifying aspects of landscape structure that likely influence the
formation and maintenance of species assemblages. ................................................ 25

Table 2.1. Summary of species lists for the six physiographic regions. .......................... 53

Table 2.2. Gaps identified by six methods for the Coastal F latwoods dataset. Consensus
gaps are based only on the Holling (HI) and Siemann & Brown (SB) indices and
their robust versions. ................................................................................................. 54

Table 3.1. Summary statistics for the National Ecological Unit Hierarchy sections used
in this analysis. .......................................................................................................... 87

Table 3.2. Land cover types identified by the National Land Cover Data Set and their
percent of occurrence in each section. A * indicates that cover type makes up <
0.01% of the total land cover within that section ...................................................... 88

Table 4.1. Beta-diversity metrics used in analyses; equations as presented in Koleff et al.
(2003). ..................................................................................................................... 118

Table 4.2. Hierarchical agglomerative clustering (UPGMA) results for species lists based
on sections (n = 162) and provinces (n = 35), and geographic distance between
sections (n = 163) and provinces (n = 35) ............................................................... 119

Table 4.3. Correlation of cophenetic distances for each of the beta-diversity matrices
based on section species lists (n = 162). ................................................................. 120

Table 4.4. Correlation of cophenetic distances for each of the beta-diversity matrices
based on province species lists (11 = 35) .................................................................. 121

Table 4.5. P-values for comparisons of beta-diversity matrices based on section species
lists and the section geographic distance matrix using Mantel tests with 1000
replicates. ................................................................................................................ 122

Table 4.6. P-values for comparisons of beta-diversity matrices based on province species
lists and the section geographic distance matrix using Mantel tests with 1000
replicates. ................................................................................................................ 123

Table 4.7. Province species lists (n = 35) correctly classified by PAM into 19 divisions
for each of the seven beta-diversity matrices (Note: Division 19 is located outside

the contiguous United States). ................................................................................ 124

Table A]. Species list for section 47, the Central Till Plains, Oak-Hickory Section. 137

vii

Table A2. Species list for section 59, the Central Ridge and Valley Section. ............... 143

Table A.3. Species list for section 133, the Coastal Plains and Flatwoods, Western Gulf

Section ..................................................................................................................... 149
Table A4. Species list for section 153, the Central Loess Plains Section ...................... 154
Table A5. Species list for section 170, the Mid Coastal Plains, Western Section. ........ 161

Table A6. Species list for section 172, the Southern Cumberland Plateau Section ....... 167
Table A6. Species list for section 172, the Southern Cumberland Plateau Section ....... 167
Table B. 1. Changes made to species names from those originally reported to the BBS. 174

Table C. 1. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, 81 = Siemann & Brown Index, RH = robust Holling Index, RS =
robust Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y =
discontinuity identified, N = discontinuity not identified) for the Section 47 dataset.
................................................................................................................................. 176

Table 02. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, SI = Siemann & Brown Index, RH = robust Holling Index, RS =
robust Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y =
discontinuity identified, N = discontinuity not identified) for the Section 59 dataset.
................................................................................................................................. 178

Table C.3. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, SI = Siemann & Brown Index, RH = robust Holling Index, RS =
robust Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y =
discontinuity identified, N = discontinuity not identified) for the Section 133 dataset.
................................................................................................................................. 180

Table C.4. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, SI = Siemann & Brown Index, RH = robust Holling Index, RS =
robust Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y =
discontinuity identified, N = discontinuity not identified) for the Section 153 dataset.
................................................................................................................................. 182

Table C.5. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, 81 = Siemann & Brown Index, RH = robust Holling Index, RS =
robust Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y =
discontinuity identified, N = discontinuity not identified) for the Section 170 dataset.
................................................................................................................................. 185

viii

Table C.6. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, SI = Siemann & Brown Index, RH = robust Holling Index, RS =
robust Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y =
discontinuity identified, N = discontinuity not identified) for the Section 172 dataset.
................................................................................................................................. 188

Table D. l. Provinces (in bold) and their component sections of the contiguous United
States delineated by the National Ecological Unit Hierarchy (Cleland et al. 1997)
with corresponding numeric codes and areas (kmz). .............................................. 191

LIST OF FIGURES

Figure 2.1. Physiographic regions delineated for the North American Breeding Bird
Survey (Bystrak 1981). 3 — Coastal Flatwoods, 4 — Upper Coastal Plain, 13 — Ridge
and Valley, 16 —- Great Lakes Plain, 24 — Allegheny Plateau, 31 - Till Plains ......... 55

Figure 2.2. Average linkage cluster analysis for the species lists of the six physiographic
regions based on Jaccard similarity values (cophenetic correlation = 0.87,
agglomerative coefficient = 0.43). Regions are numbered as in Figure 2.1. ............ 56

Figure 2.3. A. Size-ordered Holling Index values for the Ridge and Valley ecosystem.
Criterion line has been placed at the average native species index value, plus 1
standard error. B. Robust size-ordered Holling Index values for the Ridge and
Valley ecosystem. The 3 smallest and 3 largest species have been removed and the
criterion line re-adj usted. .......................................................................................... 57

Figure 2.4. Location of discontinuities in the body-mass spectra for the six physiographic
regions. Regions are numbered as in Figure 2.1. ...................................................... 58

Figure 2.5. Median distance to nearest discontinuity for declining species across all six
ecosystems. Error bars represent the range of values for those species whose
distance to nearest gap varied across ecosystems. 0 Species occurring in only 1
region (n=21), A species occurring in 2 regions (n=30), El species occurring in 3
regions (n=20), 0 species occurring in 4 regions (n=9), 0 species occurring in 5
regions (n=10), and V species occurring in 6 regions (n=2). Shaded vertical columns
represent approximate locations of discontinuities in the body size spectra. ........... 59

Figure 2.6. Empirical distribution function of distance to nearest discontinuity for all
species in the Ridge and Valley region (13). The horizontal line denotes the median
probability. Five of the six non-indigenous species have a distance to nearest
discontinuity that is greater than the median distance for this region ....................... 60

Figure 2.7. Average distance to nearest discontinuity for non-indi enous species (n=11)
across all six ecosystems with least squares regression line ( = 0.65). Names and
(number of regions of occurrence): A — house finch (6), B — Eurasian tree sparrow
(l), C — house sparrow (6), D -— monk parakeet (l), E — European starling (6), F —
Eurasian collared-dove (3), G — ringed turtle-dove (2), H — rock dove (6), I — gray
partridge (2), J — chukar (1), K — ring-necked pheasant (5). ..................................... 61

Figure 2.8. Average distance to nearest discontinuity for declining species (n = 92)
across all six ecosystems with least squares regression line (r = 0.55). Species with
unusually low distances to nearest discontinuity are (number of regions where the
species is declining in parentheses): A — cliff swallow (3), B - whip-poor-will (3), C
— common nighthawk (4), D - eastern meadowlark (2), E — sharp-shinned hawk (5),
F — long-cared owl (3), G — Mississippi kite (3), H — barred owl (2), I — peregrine

falcon (2), J — northern goshawk (3), K — Swainson’s hawk (1), L — greater prairie-
chicken (1). ............................................................................................................... 62

Figure 2.9. Empirical distribution functions of distance to nearest discontinuity for all six
ecosystems. These distributions are significantly different (Kolmogorov-Smimov
statistic = 0.11, n = 856, p < 0.005). ......................................................................... 63

Figure 2.10. Proportion of non-indigenous (A) and declining (O) species in each
ecosystem that occur at less than the median distance to nearest discontinuity.
Sample size indicated above each point. Regions are numbered as in Figure 2.1.... 64

Figure 2.1 1. Schematic representation of (A) the variation of the maximum per capita rate
of increase and (B) the negative effect of intraspecific competition across the
geographic ranges of two narrowly distributed species (A and B), and one widely
distributed species (C). The x-axis corresponds to the horizontal transect in Figure
2.12 ............................................................................................................................ 65

Figure 2.12. (A) A stylized representation of the discontinuous body mass spectra for two
ecosystems, with the position of species A, B and C indicated. (B) Schematic
representation of the geographic ranges of species A, B and C in relation to two
distinct ecosystems. The linear transect across the ecosystems corresponds to the x-
axis in Figure 2.11 ..................................................................................................... 66

Figure 3.]. Sections of the eastern United States delineated by an ecoregion classification
scheme (Bailey 1988, Cleland et al. 1997). Sections analyzed in this study are (47)
Central Till Plains, Oak-Hickory; (59) Central Ridge and Valley; (133) Coastal
Plains and F latwoods, Western Gulf; (153) Central Loess Plains; (170) Mid Coastal
Plains, Western; and (172) Southern Cumberland Plateau. ...................................... 90

Figure 3.2. Binary classification of section 172 for lacunarity analysis showing section
boundary and the boundary of the area of analysis (rectangle encompassing section).
Dark areas of the map indicate forest habitat within the section boundary. White
areas indicate non-forest habitat within the section boundary and all area between
section boundary and analysis boundary. Stippled square represents a moving
window of size 3 used in lacunarity analysis. ........................................................... 91

Figure 3.3. Size-ordered Holling Index values for the native terrestrial avian species of
Section 47. Black bars indicate index values removed for the calculation of the
Robust Holling Index. Criterion lines have been placed at the average index value
plus one standard error. Arrows indicate discontinuities identified by the Holling
Index (HI) and the Robust Holling index (RH). ....................................................... 92

Figure 3.4. Distribution of species clusters (dark bars) and discontinuities (white spaces)
in body masses of the six sections. ........................................................................... 93

xi

Figure 3.5. Percentage of land cover types across the six sections (with the disjunct areas
of sections 133 and 153 displayed separately) in each of five broad categories
(agriculture, upland forest, urban, water/wetland, and other), aggregated from the 20
land cover types present in these sections (Table 3.2). ............................................. 94

Figure 3.6. Dominance vs. diversity of land cover types in the six sections with least
squares regression line (r2 = 0.969). ......................................................................... 95

Figure 3.7. Lacunarity curves for the forest (0) and non-forest ( A) categories of section
47. Locations of discontinuities in home range size distribution as calculated by two
allometric equations are indicated by vertical lines. ................................................. 96

Figure 3.8. Lacunarity curves for the forest (0) and non-forest (A) categories of section
59. Locations of discontinuities in home range size distribution as calculated by two
allometric equations are indicated by vertical lines. ................................................. 97

Figure 3.9. Lacunarity curves for the forest (0) and non-forest (A) categories of section
133 east. Locations of discontinuities in home range size distribution as calculated
by two allometric equations are indicated by vertical lines. ..................................... 98

Figure 3.10. Lacunarity curves for the forest (0) and non-forest ( ‘) categories of section
133 west. Locations of discontinuities in home range size distribution as calculated
by two allometric equations are indicated by vertical lines. ..................................... 99

Figure 3.11. Lacunarity curves for the forest (0) and non—forest (A) categories of section
153 west. Locations of discontinuities in home range size distribution as calculated
by two allometric equations are indicated by vertical lines. ................................... 100

Figure 3.12. Lacunarity curves for the forest (0) and non-forest (A) categories of section
172. Locations of discontinuities in home range size distribution as calculated by
two allometric equations are indicated by vertical lines. ........................................ 101

Figure 4.1. Provinces (bold lines) and sections of the contiguous United States delineated
by the National Ecological Unit Hierarchy (Cleland et al. 1997). .......................... 125

Figure 4.2. The matching/mismatching components used in pairwise comparisons of
presence/absence data: (a) the total number of species common to both regions, (b)
the number of species that occur only in the neighboring region, and (c) the number
of species that occur only in the focal region .......................................................... 126

Figure 4.3. Section species richness vs. area (kmz) (n = 162) with least squares linear
regression (r’ = 0.001). ........................................................................................... 127

Figure 4.4. Province species richness vs. area (kmz) (n = 35) with least squares linear
regression (r2 = 0.461). ........................................................................................... 128

xii

Figure 4.5. Species richness and sampling intensity (BBS routes/kmz) for sections (n =
162) with least squares linear regression (r‘ = 0.313). ........................................... 129

Figure 4.6. Species richness and sampling intensity (BBS routes/kmz) for provinces (n =
35) with least squares linear regression (r2 = 0.001). ............................................. 130

Figure 4.7. Sampling intensity (BBS routes/kmz) vs. area for sections (n = 162) with least
squares linear regression (r2 = 0.273). .................................................................... 131

Figure 4.8. Sampling intensity (BBS routes/kmz) vs. area for provinces (n = 35) with least
squares linear regression (r2 = 0.248). .................................................................... 132

xiii

INTRODUCTION

The spatial heterogeneity of landscape elements can affect ecological processes
such as nutrient cycling, disturbance, and population dynamics (Turner et al. 2001). The
study of landscape structure and its spatial patterns is based on the link between pattern
and process (Gustafson 1998) with the goal of predicting ecological phenomena (Li and
Wu 2004). Assessing and quantifying this structure can complement landuse statistics
and enhance predictive abilities that are vital to the long term monitoring of ecosystems
and the establishment of functional environmental management strategies and achievable
goals (Herzog and Lausch 2001 ).

Patterns of biodiversity are influenced by both large-scale regional processes such
as the geologic and evolutionary history of a region as well as by contemporary
ecological conditions including interactions among individuals and the distribution of
resources in the environment (Turner et al. 2001, Ricklefs 2004). Humans are rapidly and
drastically altering the structure of landscapes on a global basis; these changes are having
dramatic effects on the planet’s biodiversity (Pimm and Raven 2000, Pimm 2001).
Understanding the relationship between the patterns of landscape heterogeneity and
patterns of biodiversity has become increasingly important to the success of regional and
global conservation and ecosystem management efforts. If we can improve our ability to
describe and quantify the effects of changes in landscape pattern on changes in
biodiversity, it will enable us to predict the effects of environmental modifications on

species assemblages, providing an opportunity to either enhance or prevent those

changes. Using data from North America, this dissertation examines the connection
between patterns of landscape structure and patterns of avian biodiversity.

My first chapter addresses the issue of describing landscape structure. There are
many available metrics that quantify various aspects of landscape pattern, though not all
metrics provide information useful to every question (Turner et a1. 2001). The overall
goal of my dissertation was to examine the link between patterns of spatial heterogeneity
in landscapes and patterns in the distribution of species. Therefore, I have focused this
first chapter on identifying those metrics that best describe aspects of landscape structure
that are likely to influence the distribution of species.

The second chapter of my dissertation describes the patterns of the distribution of
body sizes of birds and the ecological significance of those patterns in six physiographic
regions of eastern North America (Bystrak 1981). Holling (1992) has suggested that
ecosystems are hierarchically structured by only a few physical and biotic processes that
each operate at different spatial and temporal scales, resulting in a discontinuous
distribution of resources within a landscape. If such a resource structure exists, then the
distribution of the body masses of the species using those resources should also exhibit
discontinuities. Discontinuities in body mass distributions have been referred to as “zones
of crisis and opportunity” (Allen et al. 1999). Using species occurrence data from the
Breeding Bird Survey (BBS) (Robbins et a1. 1986), I tested the hypothesis that in each
assemblage body sizes are discontinuously distributed, using several different methods to
identify discontinuities. To address the suggestion that discontinuities may represent
either zones of crisis or opportunity, I then identified all declining and non-indigenous

species within each assemblage, and tested the hypothesis that these species have body

masses that place them closer to discontinuities in the body mass distribution than other
species in the assemblage.

My third chapter tests a proposed connection between landscape structure and the
distribution of species. As described in chapter 2, Holling (1992) has suggested that the
discontinuous structure of landscapes entrains attributes of the species living in those
landscapes, causing them to also be discontinuous. However, there have been no
quantitative tests of this connection in terrestrial environments. Using land cover data
from the US. Geological Survey (Vogelmann et al. 2001) and avian species occurrence
data from the BBS (Robbins et al. 1986), I tested the hypothesis that discontinuities in
landscape structure should correspond to discontinuities in body mass distribution. To
accomplish this, I applied the technique of lacunarity analysis (Plotnick et al. 1993,
Plotnick et al. 1996) to land cover and species data from six ecoregions in eastern North
America (Cleland et al. 1997).

The final chapter of my dissertation addresses the presumed congruence between
the distribution of ecoregions, which generally reflect geology, climate, and natural
vegetation, and the distribution of other taxa (Groves et al. 2000, Olson et al. 2001). The
concept of ecoregions has become a critical part of many conservation policies
(Government of Canada 1996, Olson and Dinerstein 1998, Groves et al. 2000), in part
because of their presumed correspondence to the patterns of distribution of other types of
taxa (Spector 2002). In this chapter I used land cover data from the US. Geological
Survey (Vogelmann et a1. 2001) and avian species occurrence data from the BBS
(Robbins et al. 1986) to test the hypothesis that the distribution of avian species

composition corresponds to the distribution of ecoregions in the continental United

States. I used seven beta-diversity metrics to assess this correspondence, and compared
the performance of these different beta-diversity metrics to determine if they differently

assessed the correspondence of species composition to ecoregion distribution.

CHAPTER 1

QUANTIFYING ECOLOGICALLY MEANINGF UL LANDSCAPE STRUCTURE

Pattern and Process

Ecologists have long argued that the spatial heterogeneity of landscape elements
has a significant influence on ecological processes such as nutrient transfer, gas flux,
evapotranspiration, and primary production across landscapes, as well as on species
abundances, population persistence, metapopulation dynamics, and local community
structure (Forman and Godron 1986, Turner 1989, Dunning et al. 1992). However,
despite this awareness of the interactions between landscape structure and ecological
processes, little progress has been made in developing a body of theory to explain the link
between pattern and process (Gustafson 1998, Wiens 2002, Li and Wu 2004).

The relationship between pattern and process can be reciprocal; the pattern of
landscape elements can affect the rate of a process and that same process can affect the
spatial distribution of landscape elements in both ecological and evolutionary time
(McCoy and Bell 1991, Li and Wu 2004). However, the connection between landscape
patterns and ecological processes is not always reciprocal, and there are undoubtedly
many spatial patterns in landscapes that are biologically meaningless (Li and Wu 2004),
or at least irrelevant to a particular ecological process (Gustafson 1998). A tendency to
focus on descriptive studies of landscape structure without linking this to quantitative
measurements of the functioning of any ecological process has slowed the maturation of

the discipline of landscape ecology (Hargrove and Pickering 1992, Li and Wu 2004).

The description of landscape pattern and the subsequent influence of that pattern
on ecosystem processes has been referred to as one of the fundamental themes of ecology
and as a research priority of landscape ecology (Levin 1992, Wu and Hobbs 2002).
Landscape metrics provide a means of quantifying landscape pattern, but linking pattern
with process has proven to be more difficult. One challenge facing landscape ecologists is
the inability to conduct randomized and replicated experiments at the relatively large
scale of most landscapes and the consequent need to rely on pseudoreplication and
inference instead of unambiguously establishing causal links between pattern and process
(Hargrove and Pickering 1992, Noon and Dale 2002). Though analysis of simulated
landscapes may partially alleviate this problem, these artificial landscapes may not
accurately reflect the configurations found in natural systems (Schumaker 1996).
Additionally, although the connection between a process and the relative value of a
landscape structure metric may be known, there is often no ecological interpretation for
any given absolute value of a metric (Gustafson 1998). Li and Wu (2004) suggest three
general topics that may be contributing to this difficulty: flaws in pattern analysis,
limitations of landscape structure indices, and improper use of those indices. To develop
the ability to predict likely trajectories of ecosystem processes, we must acquire an

understanding of the relationship between pattern and process.

Patterns of Biodiversity
As the Earth’s biodiversity continues to decline (Wilson 1992, Pimm and Raven
2000), many regional and global conservation efforts aimed at monitoring and preserving

biodiversity have been developed (Scott 1993, Olson and Dinerstein 1998, Groves et al.

2000, Myers et al. 2000). All of these efforts require some assessment of the biodiversity
present in an area before a conservation or restoration plan can be developed. However,
such assessments are often difficult to accomplish in a timely manner due to lack of
funding, lack of taxonomic expertise, and the logistical constraints of weather, political
upheaval, and access to remote locations.

Patterns of biodiversity, the spatial distribution of species, are influenced by the
structure of landscapes. From an evolutionary perspective these patterns are also linked to
macroevolutionary, geomorphological, and biogeographical events. However, in
ecological time the proximate factor involved is the dynamics of populations; the
distribution, movement, and persistence of organisms in a landscape. The formation of
species assemblages is affected by the availability of suitable resources (e.g. habitat
types) in patches of the appropriate size and shape, and distributed across the landscape
in such a way as to facilitate the ability of organisms to move, forage, and reproduce.

Describing the structure of landscape elements and understanding the influence of
that structure on ecosystem characteristics such as biodiversity patterns is an important
aspect of modem ecological research (Levin 1992, Ricklefs 2004). Structure refers to the
spatial pattern, including size, shape, and configuration of all biotic and abiotic
components of a landscape (Turner and Gardner 1991). All landscapes exhibit a
heterogeneous spatial pattern at some scale; even barren sand dunes are composed of a
heterogeneity of sand grain sizes and chemical composition. This heterogeneity results in
landscapes with complex structure, defined by the types and abundances of resources and
their spatial distribution across the landscape (With et al. 1997) and influenced by the

interactions of biotic, abiotic, and human social forces in the region (Turner 1989).

Quantifying Landscape Structure

Geographic information systems (GIS) are widely used to describe and analyze
land cover and land use across landscapes (Bolstad 2002). Using GIS, immense data sets
based on sources such as aerial photography, census information, and remote sensing can
be characterized by a variety of metrics that can be used to monitor ecological change
across broad geographic regions (O'Neill et al. 1988a, Zheng et al. 1997). There are two
commonly used formats for presenting data in a GIS, raster and vector (Clarke 1999). A
raster map uses a grid, where each grid cell is equivalent to one map unit and has only
one attribute value (e. g. land cover type or land use). A vector-based map represents
features as points on an x-y plane. Lines (sequences of points) create polygons that can be
characterized by a particular attribute value.

There are several publications containing broad overviews of metrics for
describing landscape structure and spatial pattern (Forman and Godron 1986, McGarigal
and Marks 1995, Haines-Young and Chopping 1996, Gustafson 1998, Turner et al.
2001). Some metrics are specific to categorical data, while others require point data.
Landscapes are typically described using categorical variables, although conditions
within any given habitat patch are rarely uniform, and the transition from one type of
patch to another might be more accurately represented as a gradient (ecotone) than as a
hard boundary (Gustafson 1998). However, most landscape metrics are designed for use
with categorical variables. Some metrics focus on characteristics of the landscape (6. g.
habitat diversity and contagion), while others describe individual landscape elements

(e. g. patch cohesion, mean patch size) (Tischendorf 2001).

Not all landscape metrics provide information useful to understanding
biodiversity patterns (Li and Wu 2004). A landscape is best described by a small number
of non-redundant and independent metrics that both quantify different aspects of pattern
and structure and have a relationship with the ecological pattern or process of interest
(O'Neill et al. 1988a, Riitters et al. 1995, Li and Wu 2004). Many metrics, such as
evenness and dominance (Turner et al. 2001), or edge density and contagion (Hargis et al.
1998) are derived from similar calculations; there is little advantage to including all of
them in an analysis as they provide redundant information. Riitters et al. (1995) used
factor analysis and principle components analysis to reduce an initial set of 55 metrics to
six, representing independent and orthogonal axes in metric state space. Similarly, I
McGarigal and Marks ( 1995) used principle components analysis to select a subset of
measurements from an initial set of 30 metrics. To simplify the comparison of landscapes
over space and time, O’Neill et al. (1996) combined their selected metrics of dominance,
contagion, and fractal dimension into one geometric distance measurement. Although the
subsets of metrics selected in these studies may not fully describe the important aspects
of pattern for all landscapes, the methods used to select non-redundant metrics are
broadly applicable.

Patterns of biodiversity are influenced by population dynamics; the distribution,
movement, and persistence of organisms in a landscape. There are many metrics
quantifying aspects of landscape structure that likely influence the formation and
maintenance of species assemblages in a landscape (Table 1.1). These metrics are
conventionally divided into two categories - those that characterize non-spatial landscape

composition and those that characterize spatial configuration of landscapes.

Landscape composition

Diversity indices
A variety of metrics can be used to quantify the diversity (richness, dominance,

and evenness) of land cover (habitat) types that occur in a landscape, such as Simpson’s
index and the Shannon-Wiener function (Krebs 1989, Riitters et a1. 1995). Metrics such
as these can be used to describe the general ability of a landscape to support organisms.
For example, the persistence of populations in a landscape will depend on the presence of
a sufficient quantity of one or several particular habitat types. The lack of large amounts
of unsuitable habitat may be equally important to population persistence.

Diversity comprises both richness (number of types present) and evenness
(relative area of each type), each of which can be quantified separately (Romme 1982,
Krebs 1989). Complementing evenness is dominance, which measures the deviation of an
area from the maximum possible diversity, or the extent to which one or a few land cover
types dominate the area (O'Neill et al. 1988a). The richness of land cover types in a
region is a measure of habitat availability. Habitat availablity influences resource
availability, and therefore is related to the biodiversity of a landscape. Evenness may be
an important metric to consider when assessing the availability of some critical resource
for a species (O'Neill et al. 1988a, Li and Reynolds 1994). However, these metrics —
diversity, richness, evenness, and dominance — are limited in their analytic value because
they are redundant; landscapes with very different compositions may have nearly
identical diversity and dominance values (Turner et al. 2001).

A relatively new diversity metric is mosaic diversity, m, which is used to measure
landscape complexity (Istock and Scheiner 1987). Mosaic diversity allows distinction

between simple landscapes (dominated by only a few species and controlled by a few

10

environmental gradients) and more complex landscapes because it is a function of the
variation in species richness among communities and the variation in evenness among
species (Istock and Scheiner 1987, Scheiner 1992). Mosaic diversity is potentially
important in diversity analyses because it allows a quantitative assessment of gamma
diversity, which is a combination of both alpha and beta diversity describing the species
richness of a large region (Sweeney and Cook 2001). Mosaic diversity also permits the
comparison of compositional pattern diversity, or the arrangement of communities across
landscapes (Scheiner 1992, Alard and Poudevigne 2000).

The mosaic diversity index has been used to characterize spatial patterns of
diversity in a variety of aquatic and terrestrial regions (Istock and Scheiner 1987,
Nicholls 1994, Scheiner and Rey Benayas 1994, Famsworth and Ellison 1996, Qian et al.
1998, Alard and Poudevigne 2000, Ma et al. 2000, Sweeney and Cook 2001, Rey
Benayas and Scheiner 2002). However, there has been little critical assessment of the
relationship between m and underlying landscape complexity. Nicholls (1994) found only
weak correlations between mosaic diversity and the two environmental factors likely
responsible for structuring the landscape in his study region, casting some doubt on
Scheiner’s ( 1992) interpretation of the correlation of mosaic diversity values and
landscape complexity. The interpretation of m must be clarified before the mosaic

diversity metric can be of real use in the management of landscapes.

ll

Spatial configuration

Patch area, shape & perimeter
Metrics concerning the size and shape of habitat patches in a landscape can

provide information on the likelihood that organisms will use these individual habitat
elements. For some species patch size may be a crucial landscape characteristic. In
forested landscapes with small patch sizes area-sensitive bird species have been shown to
have low diversity, increased variability in diversity, and higher extinction rates than non-
area-sensitive species (Boulinier et al. 1998, 2001). Patch size can also influence the
response of vegetation to a disturbance, and disturbances will influence both patch size
and shape (Franklin and F orman 1987, Li and Archer 1997). Although mean patch size
can be used to describe the grain of the landscape (Herzog and Lausch 2001), simply
calculating the arithmetic average could result in a misleading descriptor of patch size
because patch sizes are often not normally distributed across a landscape, but instead
have a skewed distribution with many small patches and only a few large ones. A better
metric is an area-weighted average patch size (Turner et al. 2001), or alternatively, a
metric that weights each patch size by the probability that patches of that size would
randomly occur in a given number of locations in a landscape (Li and Archer 1997). This
weighted mean patch size (WMPS) index can be used to characterize the response of a
landscape across multiple scales to disturbance. Over time, changes in ecosystem
function or landscape organization should result in changes to the index value.

Patch shape, as well as area, can influence the movement of organisms because
depending on the direction of movement, shape can affect an organism’s perception of
the apparent size of the patch (Game 1980, Stamps et al. 1987). There are a variety of

metrics assessing patch shape (Patton 1975, Game 1980, Carrere 1990, Baker and Cai

l2

 

 

1992, Gustafson and Parker 1992). Patch shape can also indicate general characteristics
of the landscape that could affect the types and numbers of species than can be found
there. Landscapes that are dominated by agricultural land cover types tend to have simple
geometric patch shapes while more complex shapes are found in regions with varied
topography or convoluted borders, such as coastlines (O'Neill et al. 1988a). Shape and
area combine to affect the amount of edge habitat that is present in a patch. Edge habitat,
or ecotones, influence species by impacting microclimates (Chen et al. 1992), the
movement of organisms (Turner 1989), risk of predation and brood parasitism (Robinson
et al. 1995), and the amount of forest interior habitat (Pickett and Rogers 1997). Though
there are metrics to measure the amount of edge in a landscape, such as edge density
(Hargis et al. 1998), the value of simply quantifying the amount of edge for a particular
landscape is debatable. Organisms will perceive and respond to edge habitats
individually, depending on their actual and realized movement abilities and the scale at
which they interact with their environment (Lidicker and Koenig 1996).

Patch perimeter, in various combinations with patch area, provides useful
descriptions of the shape complexity of patches in the landscape and their edge
characteristics (Forman and Godron 1986, Turner et al. 2001). These perimeterzarea
ratios are sensitive to patch area as two patches of the same shape but different areas will
have different ratio values. However, these ratios can be standardized to account for
shapes of varying size (Baker and Cai 1992). Other indices derived from perimeter
measures include core area metrics, measurements of distance from the edge to the
interior of patches (Gustafson and Parker 1992, Gustafson 1998), and comparison of

patch shapes to shapes of equal diameter (isodiametric) (Herzog and Lausch 2001).

13

Adjacency
The probability that two land cover types, 1' and j, occupy adjacent grid cells, Pij,

can be used to describe the finer details of the spatial pattern of a landscape and the
likelihood that an organism can move from one habitat patch to another. The probability
of adjacency of grid cells of the same or different land cover types can be calculated
directionally to look for directionality (anisotropy) in the spatial pattern, or the average
probability of having cover type j adjacent to i in any direction can be determined (Turner
et al. 2001).

Proximity and isolation indices
The proximity of habitat patches to other patches in the landscape (or conversely,

their isolation from other patches) is important to many population processes such as
migration, juvenile dispersal and foraging. The distance between habitat patches is an
important aspect of two major approaches to the study of population dynamics, island
biogeography theory and metapopulation theory (MacArthur and Wilson 1967, Hanski
and Gilpin 1997). The importance of the proximity of habitat patches will vary with the
perceptive and movement abilities of the organism of interest. For example, in general
birds are highly vagile organisms, easily moving through a variety of habitat types to
reach preferred habitat patches. However, movement ability is not constant across all
species. Some forest birds will not cross a patch of non-forest habitat to colonize a further
patch of suitable forest (Diamond 1972). Organisms that can move through non-preferred
habitats may only do so if the distance is relatively small or if the target patch can be
identified from the occupied patch. Many metrics have been developed to assess
proximity, including the proximity index (Gustafson and Parker 1992), and various

distance and area based measures (McGarigal and Marks 1995). Bender et al. (2003)

14

present a detailed review of patch isolation studies and a comparison of the usefulness of
patch proximity and isolation metrics in predicting dispersal. One of the most commonly
used metrics is nearest neighbor distance, which reflects the spatial configuration of a
cluster of patches (Hargis et al. 1998). However, nearest neighbor distance is somewhat
limited in its applicability as a proximity metric because it does not describe the entire
landscape and comparisons between landscapes can be made only if the data are of a
similar extent and grain (Hargis et al. 1998).

Contagion
Contagion characterizes overall landscape patterns by describing the degree of

aggregation of any given land cover type, thereby quantifying both landscape
composition and spatial configuration which influences the ability of organisms to move
through a landscape (O'Neill et al. 1988a, O'Neill et al. 1988b, Turner 1989, Li and
Reynolds 1993, O'Neill et al. 1996). As more land cover types are added to the landscape,
the landscape becomes more dissected and the contagion decreases. For species that are
restricted to one or a few habitat types, landscapes exhibiting low contagion can present
formidable barriers to population processes such as foraging and dispersal. Contagion can
also be used to assess the potential for disturbance to spread across the landscape;
disturbances that only spread within a single habitat type will have a higher likelihood of
propagation in a landscape exhibiting high contagion (Turner 1989).

To determine the contagion of a landscape, attribute adjacency values must be
calculated using one of two methods. The order of grid cells in each pair can either be
preserved (ordered) or not preserved (unordered). The method chosen can impact the
sensitivity of the resulting contagion index to landscape composition and configuration

(Li and Reynolds 1993, Riitters et al. 1996). The contagion metric will exhibit high

15

values for landscapes in which there is a high frequency of adjacency between grid cells
of the same land cover type or any two land cover types (Riitters et al. 1996). This can
make interpretation difficult since high contagion values may reflect true aggregation of a
land cover type or unequal land cover type frequencies (Riitters et al. 1996).

He et al. (2000) have developed another method of describing the aggregation of
land cover types, the aggregation index. This index is based on a ratio of the actual
number of shared edges to the possible number of shared edges (unlike contagion, which
is a function of the areazperimeter ratio of patches (Hargis et al. 1998)). It is independent
of map unit, allowing the comparison of aggregation values among land cover types from
the same or different landscapes. However, it is possible for two land cover types to have
the same aggregation index value and yet have different spatial patterns. The limitations
of these contagion and aggregation metrics emphasize the importance of using multiple,
non-redundant metrics to accurately describe spatial pattern.

Connectivity
There are two distinct concepts of connectivity in landscape ecology: structural

and functional (Tischendorf and F ahrig 2000b). Structural connectivity refers to the
physical contiguity of habitat across a landscape, regardless of its suitability for any given
species. Functional connectivity describes the relationship between the spatial pattern of
habitat and the ability of organisms to respond to this pattern and move through the
landscape (Taylor et al. 1993). If an organism’s movement is limited to only one habitat
type, then the structural connectivity of this habitat within the landscape is synonymous
with functional connectivity. However, because the congruence of structural and
functional connectivity is not a universal phenomenon, any effort to detect a relationship

between patterns of landscape connectivity and processes involving the movement of

16

organisms through that landscape must focus on assessing functional, not structural,
connectivity.

Functional landscape connectivity results from the spatial pattern of the landscape
and the response of organisms to that pattern and is therefore both a species- and
landscape-specific phenomenon that can influence the ability of populations to persist
(Fahrig and Merriam 1985, With et al. 1997, Tischendorf and Fahrig 2000b). Functional
connectivity is typically measured by quantifying dispersal success (number of successful
immigrants) or search time (number of movements during dispersal) for a number of
individuals in a study population (Schumaker 1996, Tischendorf and F ahrig 2000a).
However, these data, which would require tracking multiple individuals and generations
across landscapes, are logistically difficult to obtain (Tischendorf and F ahrig 2000b).
Schumaker (1996) examined the ability of nine commonly used metrics of landscape
pattern to predict landscape connectivity. He found only weak correlations between
simulated dispersal success and the measurement of pattern in real landscapes, suggesting
that these metrics would be poor substitutes for actual dispersal data. Instead, Schumaker
(1996) developed a new index, patch cohesion, based on patch perimeterzarea ratio and a
shape index (see also Gustafson 1998). He found that patch cohesion performed better at
predicting landscape connectivity.

Other methods for assessing landscape connectivity include graph theory and ring
statistics (Wiegand et al. 1999, Urban and Keitt 2001). Graph theory applications are
relatively new to landscape ecology although they are widely used in other disciplines
such as geography and computer science. Graph theory represents landscapes as binary

systems (habitat and non-habitat). Patches of habitat (nodes) are depicted as connected by

17

lines (edges), creating systems and sub-systems of nodes that are connected by pathways
of varying length. The connecting edges do not have to reflect Euclidean distances
between patches, but instead can represent species-specific processes, such as the
probability of dispersal between patches. By adding or removing nodes and edges to
evaluate subsequent effects on the connectivity of the graph, graph theory can be used to
assess the potential effects on population dynamics of gaining or losing habitat patches or
dispersal corridors. The advantages of graph theory are that it requires only a minimum
of demographic data and can be used to analyze landscapes with largevnumbers of
patches as powerful algorithms have already been developed that can easily be modified
for use in addressing ecological questions (Bunn et al. 2000, Urban and Keitt 2001).
Landscape connectivity can also be described with ring statistics, which do not rely on a
binary patch-based analysis and provide a means of assessing connectivity at multiple
spatial scales (Wiegand et a1. 1999). Using raster GIS data the connectivity of grid cells
of suitable habitat or the interspersion of suitable habitat with unsuitable or poor quality
habitat can be described, providing an assessment of the ability of a species of interest to
move through the landscape. A ring of radius r is chosen such that r corresponds to the
perceptual or movement abilities of the species of interest and the probability that a grid
cell of habitat type j is located at distance r from a cell of habitat type i is calculated.

Lacunarity
Lacunarity analysis, first introduced by Mandelbrot (1983), has been developed as

a method of describing the “texture” or “gappiness” of a landscape; it measures the
distribution of gaps between habitat patches across a landscape (Plotnick et al. 1993).
This measurement provides a scale-dependent quantification of the spatial heterogeneity

of a landscape. The heterogeneity of a landscape, and the variation of the pattern of this

18

heterogeneity with changes in scale, affects the ecological processes that occur in the
landscape (Turner et al. 2001).

Lacunarity analysis is primarily conducted with binary data, though it can also be
applied to quantitative data (Plotnick et al. 1996). Using a mean:variance ratio, lacunarity
assesses the deviance of the landscape from translational invariance (objects that are self-
similar at all scales are translationally invariant), or how similar different parts of a
landscape are to each other (Plotnick et al. 1996). Although there are several methods of
calculating lacunarity (see references in Plotnick et al. 1993), typically the gliding-box
algorithm is used (Allain and Cloitre 1991). This is a moving window analysis that
measures the density of the habitat of interest within the window as it moves across the
landscape. The lacunarity value is dependent on the size of the box used, the fraction of
the map occupied by the habitat of interest, and the geometry of the map (Plotnick et al.
1993). By using a variety of box sizes a lacunarity curve can be plotted as log(lacunarity)
vs. log(box size); the curve can be standardized to allow comparisons among landscapes
(Plotnick et al. 1996, Dale 2000). The shape of this curve depends only on the spatial
pattern of the habitat and not its density in the landscape; habitat density is reflected in
the position of the curve (Plotnick et al. 1993). Sudden changes in the slope of the
lacunarity curve indicate shifts in the spatial pattern of landscape heterogeneity (Plotnick

et al. 1996).

Influence of Scale
The concept of spatial scale, which includes both grain (the resolution of the data)

and extent (the spatial coverage of the data) is critical to understanding and predicting the

19

interaction between landscape patterns and processes (Wu and Qi 2000). Temporal scale
is also important (Gustafson 1998), but is not considered explicitly here. The scale at
which landscape structure data are collected (e. g. satellite imagery at 1 km vs. 30 m
resolution) can influence the interpretation of those data. The spatial extent of a landscape
that is considered at any given time will influence what spatial patterns are observed
(Turner et al. 1989, Haines-Young and Chopping 1996). For example, in its entirety a
meadow may appear homogeneous, but within the meadow there may be many
irregularly spaced clumps of grasses and other herbaceous plants interspersed with small
patches of bare ground. The processes determining landscape pattern vary with area
(O'Neill and King 1998). Biotic interactions may be more important at structuring the
landscape at a local level, while geomorphological processes are more important across
broad regions (Meisel and Turner 1998). The scale at which data are collected and
analyzed can also influence the amount of information that can be gained. For example,
small dispersed patches of habitat can not be detected as the grain of data is increased
(Turner 1989, Haines-Young and Chopping 1996). Analyses conducted using low
resolution (large grain) data may underestimate the diversity of land cover types and the
number of habitat patches present in a landscape. To address this, O’Neill et al. (1996)
suggest that the resolution of the collected data should be two to five times smaller than
the smallest feature of interest, and the sample size (or area over which a metric is
calculated) should be two to five times larger than the largest land cover patch in the
landscape.

Many population dynamic and source-sink models condense the true variability of

real landscapes into two land cover types, suitable and unsuitable habitat, or habitat and

20

neutral matrix (Wiegand et a1. 1999). Although a meta-analysis of 61 studies (Mazerolle
and Villard 1999) found that both landscape composition and configuration variables
were important predictors of species presence and abundance, solely considering
composition and configuration may not be enough to accurately assess the functionality
of a landscape for a particular species (Taylor et al. 1993). Instead, the ability of
organisms to perceive and respond to the heterogeneity present in a landscape at a given
scale must also be incorporated into population dynamic models. Organisms interact with
different components of their environment based on their movement capabilities,
resource requirements, and ability to use available resources (O'Neill et al. 1988b, Milne
et al. 1989, Holling 1992). For any given pattern of spatial heterogeneity, organisms of
distinct species will respond differently because of their differential ability to detect and
respond to this heterogeneity (With 1994, Riitters et al. 1997). Therefore, landscape
structure must be observed and its patterns analyzed at a scale (both extent and
resolution) that is appropriate to the organism of interest.

As the field of landscape ecology has developed, researchers have realized the
importance of changing their focus from human-defined landscapes (kilometers across) to
landscapes whose extent is defined by the organisms and processes being studied (Wiens
1989, Wiens and Milne 1989). Typically, this landscape size is larger than an individual
organism’s home range and smaller than the geographic range of the species (Dunning et
al. 1992); However, the resolution of data collection is often determined by available
technology instead of an attempt to match the resolution with which the organism
perceives the environment. Additionally, to reduce analysis time researchers may

aggregate their data to a coarser resolution than the scale of observation. Neither of these

21

practices will necessarily lead to incorrect conclusions, but changes in scale should be
noted and tested for potential impacts on conclusions.

Many contemporary issues in science and public policy require a broad-scale
approach (Peterson and Parker 1998, Wu and Qi 2000). The resources required to
establish regionally coordinated environmental monitoring and data collection efforts are
substantial. The obvious alternative is the “scaling—up” of information collected at small
scales or in studies of “microlandscapes” to explain large-scale phenomena (Wiens 1989,
Wiens and Milne 1989, King 1991, Johnson et al. 1992, Wiens et al. 1993). For example,
the dynamics of populations — their reproduction, foraging, dispersal, and mortality — are
dependent on the movement of individuals through the landscape (Wiens et al. 1995). A
characterization of individual movements through a heterogeneous landscape can be
translated across scales to describe and predict the distribution of a population (Turchin
1991, Johnson et al. 1992, Wiens et al. 1995, Turchin 1996).

However, translating information across scales is not simple, nor is it always
advisable (Wiens 1989, Schneider 1998). The relationship between pattern and process
may remain constant or change linearly only within certain domains of scale. The
extrapolation of information from one scale to another within such a domain is valid.
Separating these domains are transitions, or scale-breaks, across which relationships may
exhibit sudden changes (Wiens 1989). Information cannot be extrapolated across
domains, as this changes the observed level of organization, known as a transmutation
(aggregation) error (O'Neill and King 1998). For example, With and Crist (1996) found
that extrapolating average individual grasshopper movement patterns did not accurately

simulate observed population distributions.

22

Transitions between domains of scale can result in sudden changes in the
responding ecological process; these are called critical thresholds (Turner and Gardner
1991). Such thresholds have been found in the response of species to changes in the
contagion (habitat fragmentation) and connectivity of landscapes, affecting dispersal and
population persistence (Gardner et al. 1987, With and Crist 1995). Understanding these
thresholds is important for accurately assessing the viability of populations in fragmented
landscapes.

There are a variety of methods for detecting scale breaks in a landscape and
characterizing its multi-scale structure (Turner et al. 1991, Gardner 1998, Wu et al.
2000). Multi-scale methods include scale variance, semivariance analysis, and lacunarity
analysis. Scale breaks can also be detected indirectly by re-sampling data at different
scales and repeatedly computing landscape metrics (such as contagion or perimeterzarea
ratios). Analysis of real landscape data has shown that some landscape metrics display
predictable and simple changes with changes in scale, while others have more complex or
even erratic responses (Wu et al. 2000, Wu et al. 2002, Wu 2004). The variance of

calculated values will also fluctuate with changes in scale (Wiens 1989).

Conclusion

As monitoring and managing the environment at the scale of landscapes has
become increasingly common, the importance of adequately predicting the effect of
variation in environmental heterogeneity on ecological processes has become apparent
(Stewart et al. 2000). The complex phenomenon of landscape structure can be described

and quantified with many available landscape metrics. When used appropriately, these

23

metrics can provide a powerful means of linking landscape patterns with ecological
processes, such as the formation and maintenance of species assemblages, that influence
patterns of biodiversity. However, a better understanding of the limitations, sensitivities,
and the range of potential values for landscape metrics needs to be developed (Haines-
Young and Chopping 1996, Hargis et al. 1998). The ability to develop functional and
successful conservation and ecosystem management strategies require the integration of
landscape ecology with spatially explicit models of population dynamics (Wiens et al.
1993). As anthropogenically induced changes in landscape structure increase in
magnitude and extent, gaining the ability to understand and predict the effect of these
changes on patterns of species distribution has become crucial to the efficiency and long-

term success of ecosystem management and conservation efforts.

24

Table 1.1. Metrics quantifying aspects of landscape structure that likely influence the

formation and maintenance of species assemblages.

 

Metric

Primary Literature

 

Landscape Composition
Diversity indices
Richness
Evenness
Dominance

Mosaic diversity

Landscape Configuration
Patch area

Patch shape & perimeter

Adjacency
Proximity & isolation

Contagion

Connectivity

Lacunarity

Krebs (1989), Riitters et al. (1995)
Krebs (1989), Riitters et al. (1995)
O’Neill et al. (1988a)

Istock & Scheiner (1987), Scheiner (1992), Nicholls
(1994)

Li & Archer (1997), Herzog & Lausch (2001)

Game (1980), Carrere (1990), Baker & Cai (1992),
Gustafson & Parker (1992)

Turner et al. (2001)
Gustafson & Parker (1992), Bender et al. (2003)

O’Neill et al. (1988a), Li & Reynolds (1993), He et al.
(2000)

Tischendorf & Fahrig (2000a, 2000b), Wiegand et al.
(1999), Urban & Keitt (2001), Schumaker (1996)

Mandelbrot (1983), Plotnick et al. (1993), Plotnick et al.
(1996)

 

25

CHAPTER 2
THE ECOLOGICAL SIGNIFICANCE OF DISCONTINUITIES IN BODY MASS

DISTRIBUTIONS

Introduction

Typically, the mean body sizes of species within an assemblage are not evenly
distributed along a continuum of possible body sizes (see references within Holling 1992,
Havlicek and Carpenter 2001). Rather, these distributions typically contain significant
discontinuities that indicate species of certain size ranges do not exist within the
assemblage (Holling 1992). Discontinuous body size distributions appear to be relatively
common in natural assemblages, though the position and number of discontinuities varies
among ecosystems and taxa (Schwinghamer 1981, Holling 1992, Allen et al. 1999,
Havlicek and Carpenter 2001). This implies that ecosystems vary in their ability to
support organisms of particular sizes, and may reflect the heterogeneity of environmental
structure (Holling 1992).

The size of an organism determines the manner in which it interacts with its
environment, and therefore what constitutes available resources (Holling 1992, Milne et
al. 1992). The combination of the structure of the environment and the ability of an
organism of a particular size to use and respond to that structure will influence population
dynamics (Milne et a1. 1992, Cuddington and Yodzis 2002). This ultimately determines
which species can have self-sustaining populations in any given geographic region and

will therefore contribute to the body size frequency distribution of that assemblage.

26

Holling (1992) proposed that a small number of biotic and abiotic processes, each
of which operates at a particular spatial and temporal scale, govern the structure of
terrestrial landscapes. These scale-specific processes create discontinuously structured
landscapes. Organisms can only interact with their environment at a single scale at any
given time, although some species may have the ability to shift their interactions between
two scales (Holling et al. 1996, Allen and Saunders 2002). Using this approach, Holling
(1992) suggested that organisms living in a discontinuously distributed landscape will
exhibit attributes, such as range size or body size, that are also discontinuously
distributed due to the hierarchical structure of resources in their environment.

This study examines the relationship between several characteristics of landscapes
across eastern North America and the avian species assemblages within them in order to
address three hypotheses. First, a species assemblage within an ecosystem will have
identifiable discontinuities in its body size spectrum. Since discontinuities in the body
size spectrum represent body sizes for which suitable resources are not accessible in that
landscape (Holling et al. 1996), species that are declining or present in very low numbers
in that landscape should be found close to discontinuities in the size spectrum. Human
impacts on landscapes result in changes in landscape structure, and therefore resources.
Second, the pattern of discontinuities in the body size spectrum will respond to the
alterations in landscape structure caused by human activities. It has also been suggested
that these changes will cause species turnover in an ecosystem, beginning with species
occurring close to discontinuities in the size spectrum as these areas are “zones of crisis

and opportunity” (Allen et al. 1999). This leads to the third hypothesis, that in ecosystems

27

that have been impacted by human activities, not only will declining species occur close

to discontinuities, but non-indigenous species will also be found there.

The Species Assemblage

In North America, birds are widespread in every ecosystem and their
characteristics are well known. North American ecosystems (north of Mexico) have been
represented by “physiographic regions” delineated for the North American Breeding Bird
Survey (BBS) (Bystrak 1981). These physiographic regions were delineated based on
physical characteristics and land use. Therefore, it is assumed that each region represents
a relatively uniform biotic area, as required by Holling’s (1992) hypothesis. Of the 73
physiographic regions described by the BBS, only 10 are sampled by more than 100 BBS
routes. Of these 10 regions, six that are located almost entirely within the eastern United
States were selected in which to examine size structure (Figure 2.1). BBS routes consist
of 50 sampling points located 0.8 km apart along secondary roads. Surveys are conducted
once per year during June, and all birds seen or heard at each sampling point are recorded
(Robbins et al. 1986).

A species list was compiled for each physiographic region based on all
observations of terrestrial birds made on each route in that region over the course of the
BBS. These regions are relatively small compared to the size of the geographic range of
many bird species, and in many cases are adjacent to one another, resulting in overlap
among the species assemblages of different regions. The similarity of these species
assemblages was assessed with a hierarchical agglomerative cluster analysis. The average

linkage (UMPGA) cluster analysis performed better than a Ward’s cluster analysis

28

(cophenetic correlations 0.87 and 0.79, respectively), though they differed only in the
placement of region 31, the Till Plains. The dendrogram (Figure 2.2) shows that the
species assemblages of the two coastal regions (3 and 4) are quite similar. The
assemblages of the four non-coastal regions, especially the two mountainous regions (13
and 24) are also similar. Despite these similarities among regions, no two regions are
inhabited by exactly the same complement of species.

The body size of birds can be measured with a variety of univariate and
multivariate methods, but the best single descriptor is probably adult body mass (Rising
and Somers 1989). Although using body mass as an estimator of overall size can be
problematic because of seasonal and intra-specific variation (Gaston and Blackburn
2000), it also has distinct advantages. Body masses can be compared among taxa, unlike
other size metrics such as wing-span or tarsus length. They can be used to estimate many
other species characteristics, such as metabolic and reproductive rates (Peters 1983).
Body masses have been documented for many of the world’s bird species. These analyses
use the masses provided by Dunning (1993).

The six selected regions have experienced varying degrees of human
modification. The native vegetation of the Allegheny Plateau (24) and Ridge and Valley
(13) ecosystems both consist primarily of upland deciduous forest. The Coastal
F latwoods (3) contains sand dunes, maritime forest, marsh, bottomland hardwoods, and
pine. The slightly more interior Upper Coastal Plain region (4) is predominantly a
mixture of upland pine forest and bottomland hardwoods. The Great Lakes Plain (16) is

dominated by glacial topography, and contains broadleaf forest, oak savannah, and prairie

29

interspersed in the landscape. The Till Plains (31) are a mosaic of tall-grass prairie,
savannah, and forest.

The species list for each region was analyzed to determine which species were
exotic or declining (Table 1.1). Non-indigenous species were defined as those introduced
by humans from some other locality, such as the European starling (Sturnus vulgaris), or
in eastern North America, the house finch (Carpodacus mexicanus). The term
“declining” refers to those species that are at risk of extinction or local extirpation. In
some cases the abundance of a species may actually be increasing due to conservation
efforts, though still below historical levels. Declining species were identified by
consulting lists of protected species for each state that intersected a physiographic region.
Species in the three highest risk categories (state endangered, state threatened, and
species of special concern) were considered to be declining. If a state’s natural resource
agency did not maintain a list of state-protected species, then the information was
obtained from the state’s natural heritage program. Because each physiographic region is
intersected by multiple states it was possible, though not typical, for a species to be
defined as declining within the region even though it was considered to be at-risk by only
one state in the region. Every state has a different process for determining at-risk species;
each of these processes is also influenced by local politics. As a result, the definition of a
declining species used in this study may not accurately reflect the regional biological
status of a species in all cases. However, the use of state lists is a practical method of
describing species status. Additionally, for each ecosystem the state lists that were used

to generate the regional list of declining species were remarkably similar.

30

Identification and Measurement of Discontinuities

Discontinuities in the body mass spectrum can be identified with several methods
(Holling and Allen 2002). Each method has its strengths and weaknesses, and therefore
criticism can be leveled at any particular method (e. g. Manly 1996, Siemann and Brown
1999). Confidence in the accuracy of identified discontinuities is increased if multiple
methods produce the same result. Consequently, the following six methods were used to
identify the discontinuities in the avian body size spectrums. Only the native species of
an ecosystem were used in these analyses, in order to determine where discontinuities

exist in the recent historical fauna of a region (Allen et al. 1999).

Holling index

Holling (1992) described an index to measure how similar a species is to those
species closest to it in size,

Mn+l - Mn—l
(MM

 

H]:

where Mn is the body mass (g) of the nth species in order of increasing size and k is a
constant that detrends the data. Holling found this constant to be taxon-specific;
equivalent to 1.3 for birds and 1.1 for mammals. A large value of H1 indicates a species is
quite different from those ranked closest to it in size. To determine how large an index
value must be to represent a true discontinuity in the distribution, a criterion line is placed
at the average index value plus one standard error (Holling 1992). Discontinuities in the
distribution are identified by those index values that are greater than the criterion (Figure

2.3A).

31

Siemann & Brown index

Siemann and Brown (1999) also developed a body mass difference index,

M 1
SI =log10[ n+ ]
M”

They argued that this is a more precise metric than the Holling Index, and it requires no a

 

posteriori defined exponent to correct the data.

Robust indices

Because of the concern that the beginning and ending values in each body mass
spectrum were producing artifactual high index values that would result in the
identification of false discontinuities, a robust version of the two indices was created.
Exceptionally high index values at each end of the distribution were identified visually,
removed from the calculation of the average index value, and the criterion line for

determining if a discontinuity was significant was recalculated (Figure 2.38).

Lump analysis-gap rarity index

The lump analysis-gap rarity index (Restrepo et al. 1997) uses the gap rarity index
statistic (GRI), which is the probability that the observed discontinuities in the body size
spectrum occur by chance alone. The GRI was calculated by comparing the actual data
with a null distribution. The null distribution was generated by using log—transformed

body masses to estimate a continuous unimodal kernel distribution (Silverman 1981,

32

1986). This null distribution was then sampled repeatedly and an absolute discontinuity

M
d :10 n+1
n glO£ Mn 1

was calculated for each species in each simulation. The distribution of the differences for

value,

 

the nth largest species from the simulations was compared to the actual value obtained
from the original data. The GRI for each species in the actual assemblage is the
proportion of the simulated discontinuity values that were smaller than the observed
discontinuity value. The significance of each GRI value was then determined by testing
the null hypothesis that the value was drawn from a continuous distribution with an alpha

of < 0.05. Significant GRI values represent discontinuities in the body size distribution.

Cluster analysis

Hierarchical cluster analysis using PROC CLUSTER (SAS Institute 1999) was
also used to identify the “clumps” in the body mass spectrum. Ward’s minimum variance
clustering method (Ward 1963), based on a within-cluster least squares criterion, was
used because of its high performance compared to other clustering methods (Milligan
1981). Determining the appropriate number of clusters is difficult, but can be aided by
looking for consensus among the three criteria calculated by PROC CLUSTER: the
psuedo F-statistic, the psuedo t2 statistic, and the cubic clustering criterion (CCC). The
clusters formed in this way represent a sequence of clusters from one containing the
smallest species to one containing the largest species. Boundaries between clusters are

formed by species that are next to one another in ranked body size. A discontinuity can

33

then be defined as existing between species adjacent to one another in ranked body size,

but belonging to different clusters.

Discontinuities identified by the Holling and Siemann & Brown indices, as well
as their robust versions, were compared. Consensus discontinuities were defined as those
identified by a majority (at least three) of these four methods (Table 2.2), and were used
in further analyses. For three ecosystems, these consensus discontinuities were also
compared with the results from the lump analysis-gap rarity index and the cluster
analysis. In general, the six methods were quite comparable. Other studies (Holling et al.
1996, Holling and Allen 2002) have found that several additional methods will also
identify the same discontinuities in a dataset. These methods include the split moving
window index (Cornelius and Reynolds 1991), the Silverman difference index
(Silverman 1986), a kernel estimator (Chambers et al. 1983, Silverman 1986), and
classification and regression tree analyses (Clark and Pregibon 1992).

The structure of body mass distributions can be described in terms of the distance
of each species in an assemblage to the nearest discontinuity (e. g., Figure 2.4). To
determine these distances, species that defined the edges of a discontinuity were assigned
a value of zero. Any species that occurred in a discontinuity was also given a value of
zero. Values for all other species were calculated as the absolute value of the difference
between the mass of that species and the mass of the closest discontinuity edge. The
untransformed body size distributions were positively skewed, causing the distribution of

distances to nearest discontinuity to also be positively skewed. Before log-transforming

34

the distance to nearest discontinuity metric, a small constant of 1/6 was added to each

value to allow the use of zero distances (Tukey 1977).

The Patterns

The expectations for the patterns of the body mass spectra were partially fulfilled.
In each of the six physiographic regions, discontinuities in the body mass spectra were
identified with several methods, providing further support for Holling’s (1992)
hypothesis. Data from other taxa, locations, and scales have also been shown to support
this hypothesis — birds and mammals of North American boreal forests and the short-
grass prairies of Alberta (Holling 1992); birds, mammals, and herpetofauna of the
Everglades, cave bats of Mexico, and birds and mammals of Mediterranean-climate
Australia (Allen et al. 1999, Allen and Saunders 2002), and marine sediment assemblages
(Raffaelli et al. 2000); but see also Manly (1996), Siemann and Brown (1999) and Leaper

et al.(2001).

Discontinuity structure

The overall pattern of discontinuities in the six physiographic regions is shown in
Figure 2.4. The smallest species in each body mass spectrum was the ruby-throated
hummingbird (Archilochus colubris) at the left of the graph, while the largest species was
the wild turkey (Melleagris gallopavo). There were consistently more discontinuities at
smaller body sizes than at large; this reflects the larger number of small species in each

assemblage.

35

With a few exceptions, discontinuities were located at the same point in the body
mass spectrum, regardless of ecosystem. In the two regions with fewer discontinuities,
the Coastal Flatwoods (3) and the Till Plains (31), the discontinuities that were not
observed were those at the smaller body sizes. This observation illustrates two points.
First, at a large spatial scale (represented by large body sizes) the discontinuity structure
should be similar across ecosystems within a biome, as the underlying large-scale
geomorphological processes will also be similar. Discontinuities at smaller spatial scales
would be expected to be similar across forested ecosystems within a biome, because the
small-scale structure of leaves and twigs is similar regardless of tree species. However,
the small-scale structure would be expected to be different between forested and
nonforested ecosystems. The Coastal F latwoods and Till Plains regions may have fewer
discontinuities at small body sizes because the forest component of these ecosystems
differs from the other four. Related to this, a change in the discontinuity structure
suggests that there has been a change in the complexity of the landscape (Holling et al.
1996). Due to differences in vegetation composition and the impact of human activities,
the landscapes of the Coastal F latwoods and Till Plains may have less complexity in their

structure, and therefore fewer discontinuities in their body size spectra.

Declining species

For the six regions, there was a slight trend for the declining species to be closer
to discontinuities than the other native species (average of 52% of the declining species
were closer to discontinuities than the median distance to nearest discontinuity). The

same pattern was also found for declining species in the Everglades (Allen et al. 1999).

36

Over half (42 out of 72) of the declining species that are found in multiple
ecosystems had the same distance to nearest discontinuity, regardless of ecosystem
(Figure 2.5). The different distance measurements for 30 of the declining species reflect

the slight changes in discontinuity structure from region to region (Figure 2.4).

Non-indigenous species

Unlike the study by Allen et al. (1999), which also looked at the arrangement of
invasive and declining species in the body size spectrum, here non-indigenous species
were not associated with the discontinuities in the body mass distribution. Cumulative
distribution functions of the distances to nearest discontinuities for each of the regions
showed that an average of 72% (range of 43% to 83%) of the non-indigenous species
were further away from discontinuities than the median distance (Figure 2.6).

It is also apparent from these cumulative distribution plots that, as with the
declining species, the occurrence of non-indigenous species in relation to the
discontinuity structure was consistent from ecosystem to ecosystem. The house finch
(Carpodacus mexicanus) is widely distributed in all six ecosystems and was consistently
located at the edge of a discontinuity in each distribution. The house sparrow (Passer
domesticus) and the European starling (Sturnus vulgaris) are also widely distributed in all
regions but these species were always found near the median distance to the nearest
discontinuity. The Eurasian collared-dovel (Streptopelia decaocta), ringed turtle-dove (S.

risoria), rock dove (Columba livia), gray partridge (Perdix perdix), and the ring—necked

 

' This non-indigenous species was actually introduced into the Bahamas and then invaded Florida.

37

pheasant, Phasianus colchicus, had the largest distance to the nearest discontinuity in

each region.

Body size

The positively skewed distribution of body sizes in each ecosystem means that
there are far more small species than large species. Similarly, the pattern of discontinuity
occurrence in the distributions (Figure 2.4) shows that most discontinuities occurred at
small body sizes. As a result of these patterns, large-bodied species were usually farther
from the nearest discontinuity than small-bodied species. This is evident in a plot of the
average distance to nearest discontinuity for non-indigenous species (Figure 2.7), and in a
similar plot for declining species (Figure 2.8). However, it is worthwhile to examine
those species that appear to be outliers and consider what may make them unique.

The plot of average distance to nearest discontinuity for non-indigenous species
(Figure 2.7) shows that based on its body size, the monk parakeet (Myiopsitta monachus)
was located closer to a discontinuity than expected in the one region in which it occurs.
In fact, this species was actually located within a discontinuity. Unlike the other non-
indigenous species, which are all ground gleaners (Ehrlich et al. 1988), the monk
parakeet can utilize a variety of foraging strategies (South and Pruett-J ones 2000). The
house sparrow (Passer domesticus) is widely distributed in all six ecosystems, and was
always found near the median distance to the nearest discontinuity. However, its similarly
sized congener, the Eurasian tree sparrow (Passer montanus) was found within a
discontinuity. Although the Eurasian tree sparrow has been quite successful in its region

of introduction, it has not expanded its range in North America since its arrival in 1870.

38

The trend of increasing average distance to nearest discontinuity with increasing
body size is even clearer for the more numerous declining species (Figure 2.8). However,
there are some species that were much closer to discontinuities than expected for their
body size. Interestingly, the opposite pattern does not appear. Many of these species
defined the edges of a discontinuity in all regions in which they occur. These include the
Swainson’s hawk (Buteo swainsoni) and the greater prairie-chicken (T ympanuchus
cupido), which occur in only one region, as well as the cliff swallow (Petrochelidon
pyrrhonota), whip-poor-will (Caprimulgus vociferous), eastern meadowlark (Sturnella
magna), long-cared owl (Asio otus), and the barred owl (Strix varia), which are declining
in multiple regions. Other species, such as the common nighthawk (Chordeiles minor),
red-shouldered hawk (Buteo lineatus), peregrine falcon (F alco peregrinus), and northern
goshawk (A ccipiter gentilis) defined discontinuity edges in some ecosystems but not in

others.

Region-species interactions

A combined plot of the cumulative distribution functions of the distance to nearest
discontinuity for all ecosystems (Figure 2.9) shows that most ecosystems have a very
similar probability structure. However, the Till Plains were strikingly different. In this
region there was relatively little probability of a non-indigenous species occurring only a
short distance from a discontinuity, but at distances greater than the median distance to
nearest discontinuity, the probability increased rapidly. The Upper Coastal Plain appeared
to be somewhat intermediate between the Till Plains and the other ecosystems. There was

a significant interaction between region and species type (non—indigenous or declining;

39

maximum likelihood contingency table analysis, x2 = l 1.42, df = 5, P = 0.04). Generally,
a small proportion of the non-indigenous species in each ecosystem were found at a
distance less than the median distance to nearest discontinuity, and a relatively high
proportion of declining species were found at a distance less than the median (Figure
2.10). However, this relationship was reversed in the Till Plains.

This analysis of the body mass distributions of terrestrial breeding birds across six
ecosystems of eastern North America resulted in several clear patterns. Discontinuities in
the body mass spectra were clearly present in all six ecosystems. The distribution of these
discontinuities in each body mass spectrum was similar across ecosystems. Declining
species were generally closer to discontinuities than expected, while non-indigenous
species were generally farther from discontinuities than expected. Body size generally
increased with increasing distance to nearest discontinuity. There is an ecological
significance to the structure of these discontinuities, evidenced by the fact that there are
consistent patterns across ecosystems, and there are clear differences between successful

invasive and declining native species in an ecosystem.

Mechanisms Underlying Discontinuities

With so much environmental heterogeneity present among these six ecosystems,
it is difficult to envision a simple underlying mechanism. Yet the discontinuities in the
assemblages are non-randomly located, and the species react to these discontinuities in a
non-random manner, which suggests that there must be some common underlying
mechanism. One possibility is that the mechanism consists of the identifiable effects that

the template of landscape structure has on population dynamics.

40

The landscape template

The composition of any landscape is a result of the past and present co-occurrence
of biotic and abiotic elements in a certain spatial configuration. Climate, geology,
topography, competition, predation, disturbance, and succession all interact to influence
the components of an ecosystem. The structure of a landscape is described by the
geometries of the biologic and geologic components (Turner et al. 2001, Noon and Dale
2002). The presence and configuration of certain elements in a landscape will influence
the characteristics of the ecosystem that exists in that landscape (Wiens 1995).

The structure of the landscape is scaled in an hierarchical manner by the temporal
and spatial pattern of a few key processes (Holling 1992). For example, the dynamics of
boreal forests and their insect defoliators are governed by four dominant cycles. Each of
these cycles operates over a specific time period and spatial extent, such as the relatively
fast, small-scale interaction between insects and needles, and the slower cycle of pest
outbreaks that cause tree mortality over large areas (Holling 1992). This hierarchical
structure is linked to the distribution of available resources. Any given resource exists at
only one scale in the landscape. Landscapes that differ in structure will have differently
scaled resources. This discontinuous structure forms a template upon which the
geographic ranges of species are overlaid. The template is both spatially and temporally
dynamic, as the types and abundances of resources change between locations and time
periods.

The template of landscape structure is influenced by natural disturbances in the
environment. Natural disturbances can be biotic (the grazing of large herbivores) or

abiotic (fire, wind, drought). The spatial scale of disturbances varies over orders of

41

magnitude; a raindrop will only impact one seedling (Begon et al. 1990), while a drought
may affect a large geographic region. A small disturbance may create only an opening in
the landscape (tree canopy gaps), while a large disturbance can reset the successional
stage of the community. Natural disturbances also vary temporally, with occurrences
ranging from frequent (tidal cycles) to very rare (volcanic eruptions). Disturbances that
happen often enough can invoke selection pressures on populations (Begon et al. 1990).
The impacts of disturbance can be assessed by measures of ecological resilience (Holling
1973)

Although landscapes have always been impacted by natural disturbances,
anthropogenic disturbances are becoming increasingly common. Currently, most changes
in landscape pattern are a result of human activities, including deforestation, conversion
to agriculture, and urbanization (Turner et al. 2001). Many of the anthropogenic changes
to ecosystems are not unlike natural changes that have occurred over evolutionary time.
Indeed, humans have influenced the structure of landscapes for millennia by changing the
relative abundances of species (especially plants), altering the geographic ranges of
species, aiding the invasion of non-native species, altering soil composition, and
changing the distribution of cover types (Delcourt 1987). However, anthropogenic
changes to ecosystems, including biological invasions and extinctions, are now occurring
at a much faster pace than ever before (Vitousek et al. 1997).

As with other continents, human impacts in North America have been substantial.
Forest clearing in the east began with the original Native American inhabitants, and then
increased dramatically after the arrival of European settlers. In the late 18005 eastern

forests were being lost at a rate of 200,000 ka/yr. With the abandonment of farms in the

42

early 19003 and subsequent reforestation, rate of forest loss slowed to about 96,000
kmz/yr (Williams 1990). In the last 300 years, North America has lost about seven
percent of its forests (Richards 1990), although almost every forest in the east has been
cut at some point (Pimm and Askins 1995).

The structures of the six physiographic regions analyzed in this study (Figure 2.1)
have been affected in various ways by human activities. In many areas, not only has
native vegetation been replaced by crops, but mixed crops have been replaced by single-
crop monocultures, further reducing the structural heterogeneity of the landscape (Warner
1994). The Till Plains ecosystem historically consisted of a mosaic of prairie, savannah
and forest, but currently almost 70% of this landscape has been converted to annual row
crops (corn and soybeans) (Fitzgerald et al. 2000). About 15% of the region remains as
some sort of pasture/ grassland habitat, but the composition has been altered here too. For
example, the amount of land planted with a homogeneous monoculture of alfalfa has
increased in recent years. Only a small amount of the original habitat is estimated to
remain (Fitzgerald et al. 2000). Land conversion for agriculture and urbanization,
especially along the coastline, has led to the loss of a large fraction of native habitat in the
Coastal Flatwoods region (Hunter et al. 2001). Most forested wetlands have been
harvested and fragmented, and intense development along the coast has led to the loss of
maritime forest, dune and marsh. The eastern portion of the Upper Coastal Plain has been
converted to annual row crops (corn or soybeans) across approximately one-third of its
area. Most of the historically widespread southern pine habitat has been cut and
converted to pine plantations or non-forest uses (Hunter et al. 2001). Almost the entire

original savannah habitat in the Great Lakes Plain is gone, as well as over 80% of the

43

forest. Croplands (corn, soybeans, hay, pasture, grains) are now the dominant landcover
(Knutson et al. 2001).

The influence of humans on ecosystems has been dramatic and pervasive.
Urbanization and conversion to agriculture homogenize both the structure of landscapes
as well as the genetic, population, and species diversity of biota within them (Matson et
al. 1997, Vitousek et al. 1997, August et al. 2002). Agricultural practices have substantial
effects on ecosystem processes, including the activities of soil biota, nitrogen and other
nutrient cycling, and hydrological cycles (Matson et al. 1997). It is likely that as human
activities continue to cause additions and deletions in regional species assemblages that
ecosystem functioning will be changed. These changes may reduce the resistance and
resilience of communities to future alterations to the landscape template (Chapin et al.

1997).

The population response

In a broad sense, the pattern of environmental structure across the landscape
affects the response of species living in the landscape through variation in resources. This
influences the ability of species to survive and successfully maintain populations in a
region. Species vary in their requirements and ability to utilize resources; each species
has a unique niche (Hutchinson 1957). Because of this, each species will exhibit a unique
response to the landscape. Any given landscape will be capable of supporting only a
subset of species at any given time. Additionally, species and even life-stages within
species differ in their capacity to resist and recover from disturbances. For example,

species with high reproductive capabilities and broad environmental tolerances will often

44

recover their former population levels relatively quickly. However, species with lower
reproductive capabilities and narrower tolerances will be slower to recover.

The species assemblage present in a region is the sum of the overlapping
geographic ranges of species that occur in that region. A species’ geographic range is
located where the conditions (biotic and abiotic) of the environment match the
requirements of the species, allowing it to maintain populations across the landscape.
The geographic range is characterized by spatially autocorrelated gradients in population
abundance and density that decrease from the center to the periphery of the range, though
sometimes this gradient is cut off by a “hard” range boundary set by some physical
barrier, competition, or predator-prey interactions (Brown 1984, Brown et al. 1996,
Cumutt et al. 1996). Underlying these gradients is spatial autocorrelation in the suitability
of the environment for a species, which also decreases from the center to the periphery of
the range (Brown 1984). The variability of population abundances is relatively low at the
center of the range, and relatively high at the periphery of the range (Pimm 1986, Cumutt
et al. 1996). Combined with lower abundances and densities, this increased variability
increases the likelihood of extinction for population at the edge of its range. The
boundary of a range has been generally defined as the point at which populations cease to
contribute to the next generation (Caughley et al. 1988); the exact position of the
boundary is determined by the interaction of population processes with the spatial and
temporal variation in the environment (Brown et al. 1996).

The geographic range of a species commonly extends over more than one distinct
ecosystem (99 out of 206 species were found in all six ecosystems in this study). The

ability of species to maintain successful, stable populations within any given ecosystem is

45

influenced by the relationship between the structure of the geographic range and the
template of landscape structure. Each population within the geographic range of a species
has distinct demographic characteristics because each is influenced by a unique set of
environmental factors. The dynamics of a local population can be described by the
stochastic logistic population model,

{7% = r — uN + 02
where N is population abundance, r is the maximum per capita rate of increase, a is the
negative acceleration on population growth due to intraspecific competition, 2 is a white
(Gaussian) noise random variable with mean zero and variance one, and o is a scale
factor that represents the size of random fluctuations in the per capita rate of change
(Dennis and Patil 1984, Dennis and Costantino 1988, Dennis 1989). Spatial variation in
population dynamics can be modeled by assuming that the maximum per capita rate of
increase (r) and the effects of intraspecific competition (a) are spatially explicit
parameters. The influence of the environment causes these parameters to covary across
the geographic range of the species (Maurer and Taper 2002). At the center of the
geographic range, where populations are typically abundant and stable, the maximum per
capita rate of increase is high and decreases with distance from the range center.
Conversely, the effect of intraspecific competition typically has the lowest effect on
populations at the range center and increases with distance.

The biological complexity caused by the ecological individuality of species in a
spatially and temporally varying environment can be described by combining the effects

generated by hierarchical environmental structure and the effects generated by the

position of populations within the species’ geographic range on local population

46

processes. The relative unsuitability of resources near the periphery of a species’
geographic range will influence the dynamics of populations occurring there; they will
exhibit relatively low per capita rates of increase (r), and relatively large negative effects
of intraspecific competition (u), (Figure 2.10). This scarcity of resources should be
reflected in the structure of the body size spectra for ecosystems near geographic range
edges; a species in an ecosystem at the edge of its range will likely occur near a
discontinuity in the body size spectrum for that ecosystem. Consider the hypothetical
landscape depicted in Figure 2.12B. The geographic range of species A is centered on
Ecosystem l. The environmental conditions for A are most suitable at the center of its
range (Brown 1984), with appropriate resources to meet species requirements, and
therefore populations of A in Ecosystem 1 will likely be stable. This stability is depicted
in Figure 2.1 1, where populations of A exhibit high maximum per capita rates of increase
(r) and low effects of intraspecific competition (u) at the range center (Maurer and Taper
2002). As shown in Figure 2.12A, species A is relatively small, and is located at the
center of a lump in the body size spectrum, far from discontinuities. Moving towards the
boundary of the geographic range of species A, into Ecosystem 2, the suitability of the
environment and the availability of resources decrease. This decline is reflected in the
position of species A in the body mass spectrum; in Ecosystem 2, species A is located at
the edge of a discontinuity. At the scale at which species A is able to interact with the
landscape of Ecosystem 2, resources are scarce or inaccessible (Holling et al. 1996).
Consequently, in this region populations of A are likely to be relatively small and
scattered, as compared with those in Ecosystem 1 (Brown 1984). They are also likely to

be unstable, exhibit low maximum per capita rates of increase, large effects of

47

intraspecific competition, and relatively high variability (Figure 1 1), (Maurer and Taper
2002).

Species B is larger than species A (note its relative position in the body mass
spectra), and interacts with the landscape differently. Having different requirements, its
geographic range is centered on Ecosystem 2. Here its populations are stable, with high
per capita rates of increase and little negative effect from intraspecific competition
(Figure 2.11). In Ecosystem 2, Species B is located in the center of a lump in the body
mass spectrum. Species B also occurs in Ecosystem 1, but its range boundary intersects
this region. Consequently, here its populations are relatively unstable with lower per
capita rates of increase and greater negative effects of intraspecific competition (Figure
2.11). In Ecosystem l, at the edge of its geographic range, Species B is located at the
edge of a discontinuity in the body mass spectrum.

Species C is intermediate in size to A and B. However, it is a widely ranging
species whose geographic range boundary intersects neither of the two ecosystems
depicted here. Consequently, all populations of C in these two ecosystems have sufficient
resources available to them. Species C is not located near a discontinuity in either body
mass spectrum, but instead has stable populations throughout both regions with high rates
of increase and few negative effects from intraspecific competition.

The actual population mechanism that determines the geographic range boundary
for any species varies both spatially and temporally across populations, resulting in a
varying pattern of extinctions and colonizations across the landscape (Maurer and Taper

2002). The overriding factor influencing the demography of any population at any point

48

in time could be the maximum per capita rate of increase (r), the negative effect of

intraspecific competition (u), or both.

Conclusions

The discontinuous structure of the landscape influences geographic variation in
the abundances of species and therefore the composition of species assemblages. The
discontinuities in landscape structure represent areas of scarce or highly variable
resources (Allen and Saunders 2002) where species can maintain only low abundance
populations and are typically at the edge of their geographic range. This suggests that
there is an element of risk for those species that occur at the edge of discontinuities in a
body mass spectrum. Even those that are not currently declining (as defined by state lists
of protected species) or at the edge of their breeding range may be at risk in the future if
there is sufficient change in the landscape.

However, the discontinuities in the body size spectrum can also provide a zone of
opportunity (Allen et al. 1999) for introduced species. Although these results show that
most non-indigenous species were farther from discontinuities than the median distance
(Figure 2.6), the house finch (Carpodacus mexicanus) was the exception in all six
ecosystems. The house finch is a recent and widespread invader of eastern North America
(Elliott and Arbib 1953, Veit and Lewis 1996). It is important to note that this is a species
whose range is still expanding. As such, its population dynamics may be interacting
differently with the landscape than a non-indigenous species that has stopped expanding
its range (Maurer et al. 2001). It is quite possible that in their native western landscapes

house finches are not located at the edge of discontinuities in the body mass spectra.

49

Another logical, though currently untestable, conjecture is that perhaps when the
geographic range stabilizes in the east, natural selection will move the house finches
away from discontinuities by an increase or decrease in body size.

The alteration of a landscape through human activities will likely disrupt the
spatial pattern of population processes that occur in that landscape. Disturbed landscapes
commonly see an increase in the number of non-indigenous species (Vitousek et al. 1997,
Western 2001). As with the house finch across eastern North America, when these non-
indigenous species are still in the invasive stage, they might be found close to
discontinuities in the body mass spectrum where they can take advantage of opportunities
present at that scale. Although declining species are typically located near discontinuities
in the body mass spectra, they are occasionally found far from discontinuities, as in the
Till Plains (31) region (Figure 2.10). There are two possible explanations for this
anomalous result. First, the region may be adjusting to recent (and likely anthropogenic)
environmental changes that resulted in a scarcity of once common resources. The species
assemblage is in transition, although the discontinuity structure of the region’s body mass
spectrum has not yet changed. As the altered landscape causes additions and deletions to
the species pool, the structure of the body mass spectrum will change and declining
species will then display the typical pattern of placement close to the discontinuities.
Alternatively, those declining species that are located far from discontinuities could
simply be exhibiting a lagged response to some past change in the environment.
However, this second scenario seems unlikely in light of the results from the six

ecosystems analyzed to date (see Figure 2.8 and Figure 2.10).

50

The scarcity of suitable resources for populations at the edge of their geographic
range affects their population dynamics (low maximum per capita rates of increase, high
effects of intraspecific competition) as well as their position in the body mass spectrum
for the ecosystem. In the body mass spectrum, those species that are at the edge of their
geographic range in that ecosystem will be located close to discontinuities. This
interaction between local population processes, the discontinuous structure of the
environment, and the structure of the geographic range means that for any ecosystem, the
relative values of population demographic parameters can be predicted from descriptions
of the body mass spectrum of the ecosystem and the geographic ranges of the species in
the assemblage.

Holling (1992) outlined a research agenda for the analysis of cross-scale
dynamics in ecosystems (see also Holling and Allen 2002), that should ultimately lead to
an understanding of how a specific hierarchical landscape structure leads to a specific
discontinuous pattern of body sizes (Holling et al. 1996). The results from these six
ecosystems suggest that the structural complexity of a landscape may be the key to
understanding the discontinuity pattern of its species assemblage. The connection
between the architecture of the landscape and the species residing in that landscape has
important implications for ecosystem management and conservation. Holling (1992,
Holling et al. 1996) suggested that this relationship could be used to form an “ecoassay”.
That is, the knowledge that a landscape is changing at a particular scale, either due to
intentional or unintentional activities, could be used to predict which species in an
assemblage are likely to be affected. The opposite is also true; the knowledge that a group

of species of a particular body size are being similarly affected (e. g. population declines)

51

can be used to predict the scale at which a landscape is changing (Lambert and Holling

1998).

52

Table 2.1. Summary of species lists for the six physiographic regions.

 

 

Region Non-indigenous Declining Native Species Total Species
Species
Coastal Flatwoods 6 l 8 130
Upper Coastal Plain 9 52 163
Ridge and Valley 6 64 158
Great Lakes Plain 6 43 158
Allegheny Plateau 5 44 157
Till Plains 7 18 129

 

53

Table 2.2. Gaps identified by six methods for the Coastal F latwoods dataset. Consensus
gaps are based only on the Holling (HI) and Siemann & Brown (SB) indices and their

robust versions.

 

 

Body Size Robust Robust Cluster Consensus
Range (g) H1 H1 SB SB GRI Analysis Gaps (g)
3 - 8 Y N Y N Y Y
8 — 8.5 N N N Y N N
8.8 — 9.4 Y Y N Y N N 8.8 - 9.4
11 - 11.9 Y Y N N N N
13.2 - 14.1 N Y N Y Y Y
15 — 15.9 N N N Y N N
17 - 18 Y Y N Y N N 17 — l8
21.6—23.6 Y Y Y Y Y Y 21.6—23.6
38.7—41.7 Y Y N Y N“ N 38.7—41.7
55.3—61.5 Y Y Y Y Y Y 55.3—61.5
86.8 - 102 Y Y Y Y Y Y 86.8 — 102
245 - 300 Y Y Y Y Y Y 245 - 300
632 - 1028 Y Y Y Y Y Y 632 — 1028
2172 - 4130 Y N Y N N N

 

*This GRI value is slightly less than the criterion for significance.

54

 

 

 

 

 

Figure 2.1. Physiographic regions delineated for the North American Breeding Bird
Survey (Bystrak 1981). 3 — Coastal Flatwoods, 4 — Upper Coastal Plain, l3 — Ridge and
Valley, 16 — Great Lakes Plain, 24 — Allegheny Plateau, 31 — Till Plains.

55

 

0.30
I

 

 

 

0.25
1

“I
.l

 

 

1-Jaccard similarity
0 20
l

0.15
l

16

 

 

—l
J

13
24

Figure 2.2. Average linkage cluster analysis for the species lists of the six physiographic
regions based on Jaccard similarity values (c0phenetic correlation = 0.87, agglomerative
coefficient = 0.43). Regions are numbered as in Figure 2.1.

56

 

 

 

 

 

 

 

 

 

 

 

0.35
:lndex
0.30 Avg+1SE
-Gaps
0.25
x 0.20
(D
'C
E 0.15
0.10
0.05 n] _. ill. .l l llll] ll trill l
,0, lllllm II. null .."lliln. lla‘t..lllu.lll.. illllllu...lllll|rl.m.ulllllull IIIIIlllllllllllllllllllllllhlll
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Size Order
B.
X
G)
'U
E

 

 

Size Order

Figure 2.3. A. Size-ordered Holling Index values for the Ridge and Valley ecosystem.
Criterion line has been placed at the average native species index value, plus 1 standard
error. B. Robust size-ordered Holling Index values for the Ridge and Valley ecosystem.
The 3 smallest and 3 largest species have been removed and the criterion line re-adjusted.

57

 

31~—----_ — _

24* —----- — — _
C
016- —-I---- — — —
0')
£13. _-II-—--— — I _
4* _-III---- _ —

3-——-—--— _ _

 

 

 

T

1 2 3 4

Log1o(Body Mass)

Figure 2.4. Location of discontinuities in the body-mass spectra for the six physiographic
regions. Regions are numbered as in Figure 2.1.

58

Log1o(o + 1/6)
a

 

 

 

 

1 v v v ‘ 1 v v v 1 v v v v y t v v v 1 v v v v I v r v v V v v v v I v v v v I v v v v ‘ v Vfi V71

Species ranked by mass

Figure 2.5. Median distance to nearest discontinuity for declining species across all six
ecosystems. Error bars represent the range of values for those species whose distance to
nearest gap varied across ecosystems. 0 Species occurring in only 1 region (n=21), A
species occurring in 2 regions (n=30), CI species occurring in 3 regions (n=20), 0 species
occurring in 4 regions (n=9), 0 species occurring in 5 regions (n=10), and V species
occurring in 6 regions (n=2). Shaded vertical columns represent approximate locations of
discontinuities in the body size spectra.

59

1.0 -

.0
oo
1

     

p
O)
l

"on” finch European starling Eurasian collared-dove

\ / “yam”, Rock dove Ring-necked pheasant
_-_ _-.._____ “_ --__ ____-__ __-.._/-_.|7 ''''''''''

I
I
I
I
I
I
I
|
|
I
|
I
I
I
I
I
I
1

Cumulative probability

.0
N
l

 

b——---_——-—-———————
r—————-—--—————_————
-——-————---—--

P-—‘——-

 

‘lr———-—_—__————-——-

 

d

L
O
.3
N
09
J}.

log10(D + 1/6)

Figure 2.6. Empirical distribution function of distance to nearest discontinuity for all
species in the Ridge and Valley region (13). The horizontal line denotes the median
probability. Five of the six non-indigenous species have a distance to nearest
discontinuity that is greater than the median distance for this region.

60

3.0 1
2.5 a
2.0 ~
1.5 i
1.0 a

0.5 ~

Log1o(D + 1/6)

0.0 -

 

-0.5 *

 

 

“1.0 I I I T I I
1.0 1.5 2.0 2.5 3.0 3.5 4.0

Log1o(Body Mass)

Figure 2.7. Average distance to nearest discontinuity for non-indigenous species (n=1 1)
across all six ecosystems with least squares regression line (r2 = 0.65). Names and
(number of regions of occurrence): A — house finch (6), B — Eurasian tree sparrow (1), C
- house sparrow (6), D —- monk parakeet (l), E — European starling (6), F — Eurasian .
collared-dove (3), G — ringed turtle-dove (2), H — rock dove (6), I - gray partridge (2), J —
chukar (1), K - ring-necked pheasant (5).

61

‘i

Average Log1o(D + 1/6)

 

 

 

Log1o(Body Mass)

Figure 2.8. Average distance to nearest discontinuity for declining species (n = 92)
across all six ecosystems with least squares regression line (r2 = 0.55). Species with
unusually low distances to nearest discontinuity are (number of regions where the species
is declining in parentheses): A — cliff swallow (3), B — whip-poor-will (3), C — common
nighthawk (4), D — eastern meadowlark (2), E — sharp-shinned hawk (5), F — long-eared
owl (3), G — Mississippi kite (3), H — barred owl (2), I — peregrine falcon (2), J — northern
goshawk (3), K - Swainson’s hawk (l), L — greater prairie-chicken (1).

62

.o
oo
1

P
O)
l

   
 

Coastal Flatwoods

Cumulative probability

_0
N
I

 

 

Iog1O(D + 1/6)

Figure 2.9. Empirical distribution functions of distance to nearest discontinuity for all six
ecosystems. These distributions are significantly different (Kolmogorov-Smimov statistic
= 0.11, n = 856, p < 0.005).

63

 

 

 

 

8 0.7 —

C

.53

a) 64 44 7

-- 0.6 —

U 52

C

.9

'C

G) 0.5 - 18 43

E

v

3 04 -

'5 6 9

8. A.

‘0 0.3 ~

to—

g 5 18

o

'E 0.2 ~ 6 6

o

D.

9

D. 0.1 I T r I I r
3 4 13 16 24 31

Physiographic region

Figure 2.10. Proportion of non-indigenous (A) and declining (0) species in each

ecosystem that occur at less than the median distance to nearest discontinuity. Sample

size indicated above each point. Regions are numbered as in Figure 2.1.

64

 

Maximum per capita rate of increase (r)

 

Space

..

 

 

Negative effect of intraspecific competition (u)

Space

Figure 2.11. Schematic representation of (A) the variation of the maximum per capita rate
of increase and (B) the negative effect of intraspecific competition across the geographic
ranges of two narrowly distributed species (A and B), and one widely distributed species
(C). The x-axis corresponds to the horizontal transect in Figure 2.12.

65

Ecosystem

 

A C B
_--

 

Log10(Body Mass)

' ,"'SpeciesA “cg/SpeciesB ‘.

I \

 

’ I ’ \ \ ‘

I I

I

I
I
a

\

I
I

\
|
l

V

 

A

-01}---

 

Ecosysterri l

 

Ecbsystem 2

 

 

-~—p-'

Figure 2.12. (A) A stylized representation of the discontinuous body mass spectra for two
ecosystems, with the position of species A, B and C indicated. (B) Schematic
representation of the geographic ranges of species A, B and C in relation to two distinct
ecosystems. The linear transect across the ecosystems corresponds to the x-axis in Figure

2.11.

66

CHAPTER 3
LINKING LANDSCAPE STRUCTURE WITH THE BODY SIZE

DISTRIBUTIONS OF SPECIES ASSEMBLAGES

Introduction

The composition of species assemblages is determined by the dynamics within
and among populations that lead to the coexistence of species. The dynamics of
populations are influenced by interactions among individuals as well as the spatial
heterogeneity of the environment in which these individuals live (Kareiva 1990).
Recognizing the importance of the heterogeneous structure of landscape elements,
models of population dynamics are increasing in complexity to incorporate spatially
explicit landscape data (e.g. Liu 1993, Wiegand et al. 1999). Holling (1992) suggested
that the presence of some species, and the corresponding absence of others, was a
reflection of the texture of the landscape.

Holling (1992) proposed that a small number of biotic and abiotic processes, each
of which operates at a particular spatial and temporal scale, govern the texture of
terrestrial landscapes. The scales at which these processes operate are discrete, causing
discontinuities in landscape structure, or changes in landscape pattern with changes in
scale, to arise. Organisms can only interact with and obtain resources from their
environment at a single scale at any given time (Holling et al. 1996, Allen and Saunders
2002); therefore, Holling (1992) suggested that organisms living in a discontinuously
structured landscape will exhibit attributes, such as home range size or body size, that are

also discontinuously distributed. Holling (1992) hypothesized that these discontinuities in

67

organismal attributes correspond to scale changes in the underlying structure of the
landscape. These scale changes, or breaks, represent scales at which suitable resources
are not available for organisms of a particular size, therefore preventing the occurrence of
species of that size in the landscape.

Holling’s initial observations that assemblages of boreal bird and mammal species
exhibit discontinuous body mass distributions (Holling 1992) are supported by studies of
a variety of taxa, including frugivorous birds of the Andes (Restrepo et al. 1997),
terrestrial birds of eastern North America (Chapter 2), birds of Mediterranean-climate
Australia (Allen and Saunders 2002), mammals, birds, and herpetofauna of the Florida
Everglades (Allen et a1. 1999), two North American Pleistocene mammalian faunas
(Lambert and Holling 1998), cave-dwelling bats of Mexico (Allen et al. 1999), marine
invertebrates (Schwinghamer 1981, 1988, Raffaelli et al. 2000, Leaper et a1. 2001), and
phytoplankton, zooplankton, and fish of North American temperate lakes (Havlicek and
Carpenter 2001). However, Holling’s (1992) suggestion that discontinuities in body mass
distributions correspond to discontinuities in landscape structure has been more difficult
to test, due both to the difficulty of experimenting with landscape structure and the
variety of spatial patterns encompassed in the term “structure”.

Two tests of Holling’s hypothesis using experimentally tractable marine sediment
fauna have given conflicting results. The first study manipulated the composition of the
species assemblage but did not change environmental structure (Raffaelli et al. 2000).
This resulted in no changes in the distribution of body size discontinuities, providing
indirect support for Holling’s proposal. The second study did change environmental

structure, with replicates of uniformly sized artificial substrate as well as controls (Leaper

68

et al. 2001). This study found no consistency in the distribution of body size
discontinuities across identical treatment groups or control groups suggesting that, at least
at these small scales, the distribution of body sizes may be influenced by some factor
other than habitat architecture. Restrepo et al. (1997) also investigated the
correspondence of body mass discontinuities with landscape structure, though in a non-
quantitative manner. They compared the body mass discontinuities of frugivorous
Andean birds across elevational zones (a gradient of vertical vegetative structure and
growth form diversity), and across sites within elevational zones with varying amounts of
human disturbance (a gradient of horizontal vegetative structure). They found that both
vertical and horizontal gradients in vegetation structure seemed to influence the
distribution of body mass discontinuities, though this preliminary study could not rule out
other possibilities (Restrepo et al. 1997).

Holling’s original study (Holling 1992) provided little guidance as to what aspect
of landscape structure is key to the body size patterns in species assemblages. However,
he suggested that mesoscale processes should produce landscape-specific patterns of
textural and body mass discontinuities. The spatial pattern of land cover types is the
result of such a process. Though influenced through evolutionary time by macroscale
geomorphological processes, the distribution of cover types within ecosystems is strongly
dependent on the actions of other organisms (including humans), and abiotic disturbances
such as fire, wind, and water. The concept of lacunarity provides a means of quantifying
the discontinuities in landscape structure (Chapter 1). Lacunarity is a measure of the
distribution of gap sizes in some spatial phenomenon of interest, such as the occurrence

of patches of a particular land cover type, and can be used to detect the presence of

69

hierarchical structure (Plotnick et al. 1993). Here I propose to quantitatively test the
correspondence between the distribution of body mass discontinuities and the
discontinuities in landscape structure predicted by Holling’s (1992) hypothesis using
avian species assemblages from eastern North America (similar to those studied in
Chapter 2) in conjunction with land cover data derived from satellite imagery.

Based on Holling’s (1992) hypothesis and previous studies the following
predictions can be developed. First, different landscapes that have different spatial
processes occurring within them will vary in structure. Second, this variation in structure
among landscapes means that they will support species assemblages with varying body
mass distributions. Third, these body mass distributions will not be continuous, but
instead will exhibit discontinuities. Finally, the discontinuities in the body mass

distributions will correspond to scale changes in landscape structure.

Methods

Study area

The terrestrial habitats of the United States can be divided into biological units
using the ecoregion classification scheme devised by Bailey (1988) and subsequently
modified for use in the National Ecological Unit Hierarchy (Cleland et a1. 1997). In this
system, the contiguous United States is divided into 35 provinces comprised of 163
sections. Six small sections in eastern North America consisting of predominantly forest

and/or agricultural land cover were selected for this study (Figure 3.1, Table 3.1).

70

Land cover

The vegetation occurring in these sections is described by the National Land
Cover Data set (NLCD) from the US. Geological Survey. The NLCD covers the
coterminous United States with a consistent classification scheme and time period of data
collection (Vogelmann et al. 2001). This raster dataset has a resolution of 30 m and is
based on Landsat 5 Thematic Mapper satellite imagery from the early 19905, primarily
1992. Of the 21 land cover classes identified in the NLCD, 20 occur in the selected six

study areas (Table 3.2).

Structure

The structural characteristics of the six sections can be broadly characterized by a
comparison of the richness, diversity, and dominance of land cover types in each section.
Land cover type diversity was assessed with Shannon’s diversity index (Legendre and
Legendre 1998), which is the negative sum across n land cover types of the proportion of

the landscape, pi, occupied by that type multiplied by the natural log of that proportion.
n
H = *2 Pi In P1
1' =1

Dominance measures the deviation of the proportions of land cover types in a landscape
from the expectation that all cover types are present in equal proportions (O'Neill et al.

1988a).

n
D=lnn+Zpilnpi

i=1

71

Lacunarity is a landscape metric that allows the detection of hierarchical structure
and the presence of changes in the scale of that structure. Lacunarity analysis also
enables the comparison of the structure of multiple landscapes. Lacunarity was calculated
using the software package APACK Version 2.22 (Mladenoff and DeZonia 2002).
Calculating lacunarity with APACK requires a rectangular map. For each section, the
smallest rectangle encompassing the section was set as the area of analysis. The NLCD
land cover data for the six sections was re-classified as binary data: forest and non-forest.
This division was chosen because since the time that these species assemblages came into
existence until relatively recently, much of eastern North America was forested (Williams
1990). The forest category included only upland forest (deciduous, evergreen, and mixed)
located within each section. The non-forest category included all other land cover types
within each section and all areas of the map between the section boundary and the
encompassing rectangle, regardless of land cover type (Figure 3.2). Two sections, 133
and 153, are each composed of two noncontiguous portions (Figure 3.1). Lacunarity
analysis was conducted on only the western portion of section 153 and separately on the
eastern and western portions of section 133. Lacunarity analysis was not conducted on
section 170 or the eastern portion of section 153 due to their large size.

APACK uses the gliding-box algorithm (Allain and Cloitre 1991) to calculate
lacunarity. A window with dimensions s x s is moved one pixel at a time across the area
of analysis. At each window position, the density of each habitat category (forest and
non-forest) is calculated. For each window size, 5, the lacunarity is:

A 5 1+ varof forest denszty

 

mean of forest density2

72

With multiple window sizes a lacunarity curve can be plotted (Figure 3.7). Changes in the
slope of the curve indicate shifts in the spatial pattern of heterogeneity (Plotnick et al.
1996). Window sizes were chosen to range from s = 60 m to approximately one quarter
the extent of the smallest landscape (s = 27390 m), evenly spaced on a log scale. The
same set of window sizes was used for each area analyzed (s = 60, 90, 120, 150, 210,
270, 390, 510, 690, 930, 1290, 1740, 2370, 3210, 4350, 5940, 8040, 10950, 14850,

20160, and 27390 m).

Species assemblages

A species list was compiled for each section (Tables A.1 — A. 6) based on all
observations of birds made on each route lying completely within that section over the
course of the North American Breeding Bird Survey (BBS). The BBS is an annual survey
conducted by experienced volunteers throughout Canada and the United States. Surveys
began in 1966, though not all routes have been surveyed yearly since the inception of the
BBS. BBS routes consist of 50 sampling points located 0.8 km apart along secondary
roads. Surveys are conducted once per year during June, and all birds seen or heard
within 0.4 km of each sampling point during a 3 min period are recorded (Robbins et al.
1986).

Water birds and shorebirds were excluded from the final species lists, as were any
species recorded as “unidentified” (e. g. unid. Empidonax). Additionally, several
taxonomic changes have occurred since the inception of the BBS; therefore some names
and American Ornithological Union (AOU) codes for these species have been changed

from those originally reported to the BBS (Table B. l ).

73

Body masses

Body masses (g) for most species on the final species lists were obtained from a
digital version of Dunning’s CRC Handbook of A vian Body Masses (1993) provided by
the National Center for Ecological Analysis and Synthesis Body Size Working Group. In
instances where a variety of masses for separate subspecies were provided, the mass for

the subspecies most likely to occur within that geographic area was used.

Finding discontinuities

Discontinuities in the body mass spectrum can be identified with several methods
(Holling and Allen 2002). Confidence in the accuracy of identified discontinuities is
increased if multiple methods produce the same result. Consequently, five methods were
used to identify the discontinuities in the avian body size spectrums. Only the native
species of an ecosystem were used in these analyses in order to determine where
discontinuities exist in the recent historical fauna of a region (Allen et al. 1999).

Method 1 — Holling index
Holling (1992) described an index that measures how similar a species is to those

species closest to it in size,

Mn+l _ Mn—l
Wu)"

 

H]:

where M,, is the body mass (g) of the nth species in order of increasing size and k is a
constant that detrends the data. Holling found this constant to be taxon-specific;
equivalent to 1.3 for birds and 1.1 for mammals. A large value of H1 indicates a species is
quite different from those ranked closest to it in size. In order to determine how large an

index value must be to represent a true discontinuity in the distribution, a criterion value

74

is calculated as the average index value plus one standard error (Holling 1992).
Discontinuities in the distribution are identified by those index values that are greater
than the criterion and greater than the index value of the preceding and following species
(Figure 3.3).

Method 2 — Siemann & Brown index
Siemann and Brown (1999) also developed a body mass difference index,

M
SI: 10 "+1
gut M. )

They argued that this is a more precise metric than the Holling Index, and it requires no a

 

posteriori defined exponent to correct the data. A criterion value and discontinuities are
identified in the same manner as with the Holling Index.

Methods 3 and 4 — robust indices
Because of the concern that the beginning and ending values in each body mass

spectrum were producing artifactual high index values that would result in the
identification of false discontinuities, a robust version of the both the Holling Index (RH)
and the Siemann & Brown Index (RS) was created. The clusters of index values at each
end of the distribution were identified visually, removed from the calculation of the
average index value, and the criterion line recalculated (Figure 3.3). This new criterion
value was used to identify discontinuities for the RH and RS indices.

Method 5 — cluster analysis
A hierarchical cluster analysis using agglomerative nesting (AGNES) with

Ward’s method was also used to identify discontinuities in the body mass distributions.
This analysis was conducted using the cluster package in R version 2.0.0 (Kaufman and
Rousseeuw 1990, R Development Core Team 2004). Dendrograms of the clustering

results were used to depict the species (in order of increasing body mass) in a series of

75

nested clusters. Discontinuities were defined as existing between species adjacent to each
other in the body mass spectrum but belonging to different terminal clusters.
Discontinuities identified by the five methods were then compared. Consensus
discontinuities were defined as those identified by at least three of the five methods
(Tables C.1 - C.6). In some cases, three or more methods identified non-identical but
overlapping discontinuities. In these cases, only the overlapping portion of the
discontinuity was considered to be the consensus discontinuity (for example, see the
discontinuities at body mass range 21.6 — 23.6 and 21.6 — 25.5 in Table C.1). In general

the five methods detected comparable discontinuities in the body mass distributions.

Linking lacunarity with body mass

The fourth prediction tested by this study is that discontinuities in body mass
distributions will correspond to scale changes in landscape structure. Discontinuities in
landscape structure are indicated by changes in the slope of the lacunarity curve.
However, the locations of discontinuities in the body mass distribution of species
assemblages were measured in grams, while discontinuities in landscape structure were
measured in meters. This discrepancy in measurement units was corrected by using
known allometric relationships between the body size of an organism and its home range
size (Peters 1983, Calder 1996). Discontinuities in the body mass distribution of each
section were converted to discontinuities in home range size using the regression
relationship for both passerine and nonpasserine birds, H oc M”6 (Schoener 1968).
Holling (1992) expanded and reanalyzed Schoener’s (1968) bird data, finding an

exponent of ~1 .36 for avian carnivores, omnivores, and herbivores, suggesting the

76

possibility that the slope of the regression relating home range size to body size in birds
might be as high as 1.3 instead of closer to the mammalian exponent of 1.1. Holling
hypothesized that this higher exponent could be a function of the three-dimensional use
of space by birds, as opposed to mammals which primarily move in only two dimensions.
Based on this, discontinuities in the body mass distribution of each section were also
converted using a higher exponent, H at M”. Home ranges were assumed to be
approximately circular in shape, and the diameter of each home range was used as a

representation of the spatial extent of the landscape sampled by a species.

Results

Landscape structure

The six landscapes analyzed here, despite all being located in the humid
temperate domain of the eastern United States (Cleland et al. 1997) are not uniform in
their structural characteristics. They vary in area (8790 —— 97840 kmz; Table 3.1), shape,
and topography (Figure 3.1). Sections 59 and 172 are relatively small and narrow,
oriented northeast-southwest along the Appalachian Mountains. Sections 133 and 170 are
located in the southern coastal flatwoods, with section 133 completely bisected by the
lower Arkansas and Mississippi River valleys and section 170 nearly so. The final two
sections, 47 and 153, are located in the Great Plains. However, section 153 is divided into
two noncontiguous portions, with the larger area adjacent to section 47 and the smaller
area more than 450 km to the west.

The number of land cover types occurring in the six sections is relatively

consistent, with land cover richness varying only from 15 —19 (Tables 3.1 and 3.2).

77

However, the proportions of these cover types present in any given section are different.
Aggregating the cover types into five general categories (agriculture, upland forest,
urban, water/wetland, and other) revealed that each section is dominated by either
agricultural or upland forest cover types (Figure 3.5, Table 3.2). Sections 47 and 153
consist primarily of agricultural land (76% and 84% of total land cover, respectively),
while sections 59, 133, 170, and 172 are composed mostly of upland forest (66%, 65%,
65%, and 71%, respectively). Urban, water/wetland, and “other” land cover types do not
make up a significant portion of total land cover in any section, with two exceptions. In
both sections 133 and 170, water/wetland cover types are present in approximately the
same proportions as agricultural cover (section 133: 16% and 13%, section 170: 12% and
19%). Using the original 20 land cover types, the diversity and dominance of each section
was calculated (Figure 3.6). The two sections consisting primarily of agricultural cover
types, sections 47 and 153, had the lowest land cover type diversity and highest
dominance, while the remaining sections had relatively similar diversity and dominance

values.

Discontinuities in body mass distributions

The avian species assemblages (Tables A.l — A6) of the six sections were
broadly similar, with terrestrial species richness varying from 94 — 107 (Table 3.1), and
73 of these species occurring in all six sections. When arrayed in order of increasing
mass, all sections were identical in their smallest (ruby-throated hummingbird,
Archilochus colubris, 3 g) and largest species (wild turkey, Meleagris gallopavo, 7400 g).

Analysis of the avian species assemblages of the six sections with five different methods

78

showed that, as predicted, there are detectable discontinuities in the body mass structure
of each assemblage (Tables C.1 — C.6, Figure 3.4).

The number of discontinuities identified in each section was similar, varying from
11 — 15. The position of these discontinuities was also quite similar across sections
(Figure 3.4). Three discontinuities (21.6 — 23.6 g, 86.8 — 102 g, and 148 — 174 g) were
identical in all six sections, while four other discontinuities had similar zones of overlap
in all sections (~ 200 g, 400 g, 800 g, and 3000 g), though starting and ending points
varied slightly. The patterns of discontinuities varied slightly among sections. Section
172 had more evenly spaced discontinuities at smaller body sizes (< 100 g) and fewer
discontinuities at larger body sizes (> 174 g) than section 59, though these two sections
adjoin each other and are both composed primarily of upland forest (Figures 3.1 and 3.5).
The species assemblage of section 153 exhibited more discontinuities (seven) at small
body sizes than the other sections, while the assemblage of section 133 had the most

discontinuities (seven) at larger body sizes.

Correspondence of assemblage discontinuities with landscape discontinuities

Lacunarity values were calculated and curves plotted for both forest and non-
forest within each section (Figures 3.7 — 3.12). The forest curves have the highest
lacunarity values at the smallest window sizes and then decrease relatively smoothly as
window size increases. The curves for sections 47 and 153 west (Figures 3.7 and 3.11)
are nearly straight (linear regression r2 = 0.999 and 0.9949, respectively). In contrast, the
non-forest curves display the highest lacunarity values at intermediate window sizes and

exhibit many changes in slope. The non-forest curves for sections 47 and 153 west are

79

quite similar in shape, with the two largest slope changes occurring at the sixth and eighth
largest window sizes (s = 270 m and s = 510 m). Sections 59 and 172 (Figures 3.8 and
3.12) also have similar non-forest lacunarity curves; section 59 has its largest slope
changes at the seventh and tenth window sizes (s = 390 m and s = 930 m) while section
172 has large slope changes at the sixth and ninth window sizes (s = 270 m and s = 690
m). The eastern and western portions of section 133 (Figures 3.9 and 3.10), despite being
considered part of the same ecological section, display major slope changes at different
window sizes, though the overall shape of the non-forest lacunarity curves is quite
similar. The largest slope changes in section 133 east occur at the ninth and twelfth
largest window sizes (s = 690 m and s = 1740 m) while in 133 west they occur at the
seventh and ninth window sizes (s = 390 m and s = 690 m).

Sections 47 and 153 west have the largest range of lacunarity values for forest
habitat (Figures 3.7 and 3.1 1; log-lacunarity varies from 0.137 — 0.897 for section 47 and
from 0.194 — 1.533 for section 153 west), showing that although these areas consist
primarily of agricultural land (Table 3.2, Figure 3.5), the forest structure that is present is
more heterogeneous than that of other sections. The gaps, or distances between patches of
forest habitat in the remaining sections have less variation than that present in sections 47
and 153 west. The lacunarity values for forest habitat in sections 59, 133 east, 133 west,
and 172 have a lower maximum and span a smaller range of value, indicating that the
forest habitat has a more homogeneous structure (Figures 3.8 — 3.10 and 3.12). Sections
59 and 172, which are primarily forest (Table 3.2, Figure 3.5), have the largest range of
lacunarity values for the non-forest category (Figures 3.8 and 3.12; log-lacunarity varies

from 0.099 — 0.614 in section 59 and from 0.128 — 0.622 in section 172). This indicates

80

that the structure of non-forest habitat is more heterogeneous in these two sections than in
the remaining sections. Sections 133 east and 133 west, which also consist primarily of
forest land cover types (Table 3.2, Figure 3.5), have similar ranges of lacunarity values
for both forest and non-forest habitat (Figures 3.9 and 3.10), indicating that both habitat
categories display a similar degree of heterogeneity in their structure.

Discontinuities in the body mass distributions were converted to discontinuities in
the distribution of home range size diameters. The location of these discontinuities is
overlaid on the lacunarity curves for each section (Figures 3.7 — 3.12). There are many
more discontinuities in the avian home range size distributions than there are obvious

changes in slope for either the forest or the non-forest lacunarity curves.

Discussion

The six sections analyzed here are broadly distributed across eastern North
America in areas of widely differing topography and geologic history. These diverse
geographic settings indicate that the structure of these landscapes is likely to differ,
though each section predominantly consists of either agricultural or upland forest land
cover. Calculations of the dominance and diversity of the land cover types in the six
sections show that the sections currently composed mostly of human originated land
cover types (agriculture) have very high dominance and low diversity values. In contrast,
those sections consisting primarily of non-anthropogenic land cover types (upland forest)
have very low dominance and high diversity values.

Though there is some upland forest present in each section (Table 3.2), the

distribution of this forest differs between the highly human-modified, agriculturally

81

dominated sections and the forest dominated areas. Interestingly, the large range of
lacunarity values for forest habitat in sections 47 and 153 west, both consisting
predominantly of agriculture, show that there is much more heterogeneity in the
distribution of forest land cover in these sections than in the remaining sections. The
forest land cover in sections 47 and 153 west appears to be located primarily along
waterways, roads, and in patches along field edges, a distribution in which the gaps
between forest patches are quite varied in size. However, the other sections contain large
tracts of continuous forest and so have little heterogeneity in gap size.

Although avian species richness did not vary significantly across the six sections,
and there were 73 species that occurred in every section, there was variation in the body
mass distributions of these sections. As expected based on studies of other taxa, the
species assemblages of these six sections also clearly exhibit discontinuities in their
distributions of body masses. However, there is no clear pattern to the structure of these
body mass distributions (Figure 3.4). In contrast to the results described in Chapter 2,
here the most heavily human modified areas, sections 47 and 153, displayed conflicting
results in terms of body mass discontinuities. The species assemblage of section 47 had a
very simple structure, with only 11 identified discontinuities, while the assemblage of
section 153 had the most complex structure with 15 discontinuities. This discrepancy in
results may reflect the differences in the total area of analysis. The study described in
Chapter 2 used data from much larger physiographic regions (Bystrak 1981), which had
. correspondingly higher species richness values, than the smaller National Ecological Unit

Hierarchy sections (Cleland et al. 1997) used here.

82

To quantitatively test Holling’s (1992) hypothesis that the heterogeneity in the
structure of a landscape entrains the attributes of species living in that landscape, this
study aimed to assess the correspondence between discontinuities in the distribution of
land cover types and discontinuities in body mass distributions (converted to home range
size distributions). The age of North American avian species assemblages is unknown.
The continent has been in an interglacial period for approximately 10,000 years, so these
species assemblages could date back thousands of years. However, substantial habitat
changes in the past 300 years (Pimm and Askins 1995) could also have resulted in recent
changes to the composition of these assemblages. If the species assemblages of these
sections have been in existence much longer than the presence of European settlements in
North America, then the distribution of many of the anthropogenic land cover types (e. g.
agricultural and urban cover types) listed in Table 3.2 may not be appropriate predictors
of species body mass distributions. Much of eastern North America was heavily forested
prior to European settlement (Williams 1990) so it is likely that if these species
assemblages have remained relatively unchanged, they would have co-evolved in
primarily forested landscapes. However, lacunarity analyses of upland forest land cover
types show that there are very few discontinuities in the structure of this habitat
(lacunarity curves have very few changes in slope) and therefore little correspondence to
discontinuities in home range size. There are more instances of correspondence between
discontinuities in non-forest habitat and home range size, though in each section there are
many more discontinuities in the distribution of home range sizes than in non-forest

habitat. Since home range size discontinuities exist without a corresponding discontinuity

83

in non-forest habitat structure, it is unlikely that the heterogeneity of non-forest structure
is the key factor influencing the structure of avian assemblages.

This lack of any correspondence between discontinuities in the structure of the
species assemblages and the current forest structure of the landscape does not provide
support for Holling’s hypothesis. However, out of necessity this analysis used relatively
recent land cover data. If these species assemblages arose prior to European settlement,
the composition and configuration of forest patches across eastern North America is
likely different than it was when the assemblages originated, even if the land cover types
have remained relatively consistent. For example, since most eastern forests have been
logged at some point in the past 300 years (Pimm and Askins 1995), it is likely that the
structure of this second- or third-growth forest is now different from the pristine forests
that once existed. Additionally, many forests are now actively managed for timber
production; these forests may consist of even-aged stands of only one or a few species
with little undergrowth and architectural diversity. The NLCD data set used here does not
discriminate between highly human modified forested areas and naturally regenerated
forest.

It is probably more likely that the habitat changes that have occurred in the past
few centuries have also caused changes in the composition of species assemblages. If this
is true, that species assemblages have changed in response to environmental change, then
the expectation is that there should be a correlation between the current composition of
species assemblages and landscape structure. No such correspondence was found
between discontinuities in home range size and discontinuities in upland forest habitat.

Although it is possible that a correspondence might be found using some other subset of

84

the land cover types listed in Table 3.2, the results of the present analysis are inconsistent
with Holling’s (1992) hypothesis.

The requirements of the software APACK (Mladenoff and DeZonia 2002) used to
conduct the lacunarity analyses likely affected the results depicted in Figures 3.7 —— 3.12.
All the sections analyzed here have irregular boundaries, but APACK requires a
rectangular map. The irregular shapes of these sections meant that analyzing the largest
possible rectangular area that could be placed completely within the section boundaries
would result in the loss of data from a large proportion of each section, especially from
the areas near the section boundaries. In order to use all the data from each section, the
area of analysis was defined as the smallest rectangle that encompassed the entire section
(Figure 3.2). However, this definition resulted in a large proportion of each area of
analysis being populated by continuous areas of non-forest habitat. Therefore, many of
the habitat density calculations, especially at the smaller window sizes, resulted in a zero
density of forest habitat (or conversely for the non-forest lacunarity analysis, a window
saturated with non-forest habitat). These consistent density values may be artificially
depressing the lacunarity curves at small window sizes; the true structure could be more
heterogeneous than it appears.

The occurrence of discontinuities in body mass distributions is pervasive across
taxa and geographic locations, suggesting that there must be some common causal
explanation for the pattern. The importance of environmental structure in the dynamics of
populations, and therefore the composition of species assemblages, is well known. In
light of this, Holling’s textural discontinuity hypothesis seems reasonable, though it was

not supported by this study. Holling’s hypothesis is difficult to test given the challenge of

85

finding an appropriate quantification of landscape structure. Lacunarity is a promising
metric because of its ability to detect hierarchical structure in a landscape. The insights
gained from this study suggest several possible avenues for future investigations,
including modifying the APACK software to remove the requirement for a rectangular
map, testing other subsets of land cover types, and choosing study areas in regions that
have been little impacted by human activities and therefore more closely represent

conditions under which the species assemblage arose.

86

Table 3.1. Summary statistics for the National Ecological Unit Hierarchy sections used
in this analysis.

 

 

Section Area No. Land No. Bird
No. Section Name (1000 kmz) Cover Types Species
47 Central Till Plains, Oak- 41.43 18 107
Hickory
59 Central Ridge and Valley 12.97 15 100
133 east Coastal Plains and 8.79 16
F latwoods, Western Gulf
133 west Coastal Plains and 24.91 18 94
Flatwoods, Western Gulf
153 east Central Loess Plains 93.50 19
101
153 west Central Loess Plains 24.39 18
170 Mid Coastal Plains, Western 97.84 17 106
172 Southern Cumberland 22.53 15 105
Plateau

 

87

Table 3.2. Land cover types identified by the National Land Cover Data Set and their
percent of occurrence in each section. A * indicates that cover type makes up < 0.01% of
the total land cover within that section.

 

LandCoverType 47 59 l33e 133w 153e 153w 170 172

 

Open water 1.54 2.70 2.84 3.32 0.97 0.95 2.66 2.28
Perennial ice/snow 0 0 0 0 0 0 0 0
Low intensity 1.03 2.01 0.48 0.66 0.70 0.37 0.98 1.26
residential

High intensity 0.32 0.38 0.02 0.18 0.42 * 0.25 0.28
residential

Commercial/ 0.48 1.15 0.17 0.65 0.53 0.46 0.74 0.79
industrial

/transportation

Bare rock/sand/clay 0.03 0.02 0.01 0.10 0.02 0.01 0.09 *
Quarries/strip 0.18 0.09 0.05 0.06 0.04 0.02 0.14 0.35

mines /gravel pits
Transitional 0.02 0.61 3.19 4.41 * 0 1.64 0.95

Deciduousforest 14.02 25.92 16.72 11.14 8.90 2.81 19.40 32.59

Evergreen forest 0.52 18.76 33.70 32.03 0.25 0.08 23.91 14.65
Mixed forest 0.41 21.24 15.62 20.97 0.07 0.02 21.71 23.61
Shrubland 0.01 0 0 * * 0.02 0 0
Orchards/ vineyards 0 0 0 0 0.01 0 0 0
/other

Grasslands/ 0.97 0 0 0.09 0.49 15.75 0.09 0
herbaceous

Pasture/hay 37.18 21.14 1.89 7.94 13.37 11.73 15.95 14.48

88

Table 3.2 (cont’d).

 

 

Land Cover Type 47 59 133e 133w 153e 153w 170 172
Row crops 37.28 4.65 9.78 4.58 71.98 61.27 2.90 7.41
Small grains 1.17 0 0.42 1.35 0.19 5.12 0.17 0
Fallow 0 0 0 0 0 0.01 0 0
Urban/recreational 0.79 1.07 0.06 0.1 1 0.59 0.16 0.17 0.43
Grasses
Woody wetlands 3.72 0.20 14.57 11.52 1.22 0.15 8.43 0.82
Emergent 0.31 0.07 0.47 0.88 0.24 1.05 0.76 0.10
herbaceous
wetlands

 

89

 

 

 

 

 

Figure 3.1. Sections of the eastern United States delineated by an ecoregion classification
scheme (Bailey 1988, Cleland et al. 1997). Sections analyzed in this study are (47)
Central Till Plains, Oak-Hickory; (59) Central Ridge and Valley; (133) Coastal Plains
and Flatwoods, Western Gulf; (153) Central Loess Plains; (170) Mid Coastal Plains,
Western; and (172) Southern Cumberland Plateau.

90

 

 

 

 

 

 

 

Figure 3.2. Binary classification of section 172 for lacunarity analysis showing section
boundary and the boundary of the area of analysis (rectangle encompassing section).
Dark areas of the map indicate forest habitat within the section boundary. White areas
indicate non-forest habitat within the section boundary and all area between section
boundary and analysis boundary. Stippled square represents a moving window of size 5
used in lacunarity analysis.

91

 

 

 

 

 

 

 

 

 

 

 

 

 

                                             

 

 

 

 

 

0.50
V
a Index Value
Remover! for RH
— HI Criterion Line
/ --— RH Criterion Line /
/ v HI Discontlnwtjes /
0.20 / v RH DISCOntanItleS /
a: Y
E 0.15 .
.E’ v
6
I 0.10 -
0'05 ' v v V F .ljl I
_____ ____ ______ ____ 70E," 11111111 11
100 110

Size Order

Figure 3.3. Size-ordered Holling Index values for the native terrestrial avian species of
Section 47. Black bars indicate index values removed for the calculation of the Robust
Holling Index. Criterion lines have been placed at the average index value plus one
standard error. Arrows indicate discontinuities identified by the Holling Index (HI) and

the Robust Holling index (RH).

92

 

 

 

 

172~ —--—--I-l -- _ —
170- __---l-I .l- I - I
g 153- —IlI--II III I— I I I
.9
‘8
m133~ —_-I--IIII- I- I
59- —-I-l-I II I _ I
47, —--II-l I- l- I
O 1 2 3 4
Log,o(Body mass)

Figure 3.4. Distribution of species clusters (dark bars) and discontinuities (white spaces)
in body masses of the six sections.

93

 

“

rem flr
.meamm
rba
gor 6L
AFUWO

4%
/\mg

 

 

 

  

 

 

 

 

Hufl
_________________..w

 

 

 

 

____m_m%

 

_ .............. r///////////////////////////////

      

 

47 59 133 133e 133w 153 1536 153w 170 172

 

 

 

__ ////////////////////////////////A
__%////n

..m>oo new. .92 Lo Eoeon.

Section

Figure 3.5. Percentage of land cover types across the six sections (with the disjunct areas

of sections 133 and 153 displayed separately) in each of five broad categories

(agriculture, upland forest, urban, water/wetland, and other), aggregated from the 20 land

cover types present in these sections (Table 3.2).

94

 

Dominance

 

 

 

 

0.8 I l I I
1.0 1.2 1.4 1.6 1.8 2.0

Shannon's Diversity Index

Figure 3.6. Dominance vs. diversity of land cover types in the six sections with least
squares regression line (r2 = 0.969).

95

1.0

 

 

  
 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A —0— Forest
+ Non-forest
O H=M"1.16
0.8 - A I-I=MI~1.3
.3
5 0.6 .
c
3
o
to
..I
‘é
a) 0.4 "
O
_I
0.2 -
0.0 I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Log10(Window Size)

Figure 3.7. Lacunarity curves for the forest (0) and non-forest (A) categories of section
47. Locations of discontinuities in home range size distribution as calculated by two
allometric equations are indicated by vertical lines.

96

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0
+ Forest
0 ”centimeter 4p A 0 A + Non-forest
O H=M“1.16
0.8 ~ A H=M"1.3
E
a 0.6 -
C M
3
O
(U
:i
85,-” 0.4 .
..I
0.2 ~
W
0.0 T l l l l l
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Log,0(Window Size)

Figure 3.8. Lacunarity curves for the forest (0) and non-forest (A) categories of section
59. Locations of discontinuities in home range size distribution as calculated by two

allometric equations are indicated by vertical lines.

97

 

1.0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0A 6 A0 or: 49 oo MAM» o A A o A . Egg-item
O H=M*‘I.16
0.8 ‘ A H=M"1.3
S
a 0.6 ~
C
3
U
(U
:1 {N
O
05' 0.4 I W
3
H
0.2 - W
0.0 I I I T I I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Log,0(Window Size)

Figure 3.9. Lacunarity curves for the forest (0) and non-forest (A) categories of section
133 east. Locations of discontinuities in home range size distribution as calculated by two
allometric equations are indicated by vertical lines.

98

 

1.0

 

A A 4 0 AA A A A ForeSt

O
0 9 if f ”a“? " O +Non-forest
O H=M"1.16

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.8 - A H=M‘1.3
.3
E 0.6 -
c
3
0
(U
:i
C
8, 0.4 “ q.“ I
_I
0.2 - .
0.0 I I I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Log,o(Window Size)

Figure 3.10. Lacunarity curves for the forest (0) and non-forest (A) categories of section
133 west. Locations of discontinuities in home range size distribution as calculated by
two allometric equations are indicated by vertical lines.

99

 

1.8
16 ’9 TF‘MAOAQA To A A Q A —0— Forest

 

 

+ Non-forest

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O H=M“1.16
A H=M"1.3
1.4 _
S 1.2 -
“C
S
3 1.0 ~
8
..I K
E 0.8 " \.\.‘
8’
_J 0.6 - , .
0.4 ~ KL
1k
0'2 d I Viki—“RR
0.0 -——&—Ir-H- I I I I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Log,0(Window Size)

Figure 3.11. Lacunarity curves for the forest (0) and non-forest (A) categories of section
153 west. Locations of discontinuities in home range size distribution as calculated by
two allometric equations are indicated by vertical lines.

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0
000130 A0 43 (P G4 AGD 49 A Q A FOI'GSI
1) P + Non-forest
O H=M"1.16
0.8 -1 A H=M"1.3
7% I
0.6 - ‘
:3
o
to
:L
8? 0.4 - “H
_J
0.2 -
0.0 I I I I I I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Log,0(Window Size)

Figure 3.12. Lacunarity curves for the forest (0) and non-forest (A) categories of section
172. Locations of discontinuities in home range size distribution as calculated by two
allometric equations are indicated by vertical lines.

101

CHAPTER 4

AVIAN SPECIES COMPOSITION AND NORTH AMERICAN ECOREGION S

Introduction

Environmental management and conservation policies are increasingly being
constructed around the framework of ecoregions; however, the factors that influence the
distribution of ecoregions may not always be key influences on the distribution of taxa
inhabiting those ecoregions (Goldstein 1999, Margules and Pressey 2000). Humans often
describe landscapes with arbitrary sociopolitical boundaries, although we also have a
long history of attempting to divide landscapes along more natural boundaries (Omernik
1995, Bailey 1996). These boundaries typically delineate areas of similar climate and
resources and contain similar ecosystems (Omernik 1995, Bailey 2002). Different
systems for delineating ecoregions have developed, as there is not universal agreement
(Golledge et a1. 1982, Hart 1982b, 3, 1983, Healey 1983). Ecoregions have traditionally
been delineated with a combination of quantitative and qualitative methods (Omernik
1995, Bailey 1996, Olson et a1. 2001). This lack of quantitative purity has led to some
controversy over the validity over the concept of ecoregions (Gering 2004, Magnusson
2004).

As ecosystem-based approaches to conservation and management have become
more predominant, ecoregions have become a critical part of the framework of resource
management and conservation policies worldwide (Government of Canada 1996, Olson
and Dinerstein 1998, Groves et al. 2000). Ecoregions have become an important focus of

conservation efforts as they are large enough to define areas that can preserve not only

102

populations and genetic diversity, but also critical ecological and evolutionary processes
(Groves et al. 2000, Olson et a1. 2001). For example, the Native Seed Network affiliated
with the US. Bureau of Land Management uses ecoregions as a substitute for empirical
data on the genetic makeup of metapopulations of native plants (Native Seed Network
2005).

One of the often cited benefits of framing management policies around ecoregions
is that ecoregion boundaries not only represent areas of distinct geological landforms and
potential natural vegetation, but also likely reflect the distribution of both vertebrate and
invertebrate taxa (Groves et a1. 2000, Olson et a1. 2001). Although ecoregions are
undoubtedly more accurate representations of ecological landscape units than political
divisions (Olson and Dinerstein 1998, Groves et al. 2000, Magnusson 2004), their
delineation is still often at least partially based on human interpretations of the landscape
and therefore subject to the question of whether these divisions reflect actual natural
entities (Goldstein 1999). Ecoregions are often delineated based on some subset of
variables including potential natural vegetation, soil types, and climatic variables
(Bystrak 1981, Omemik 1995, Cleland et al. 1997, Williams and Pashley 1999, Olson et
a1. 2001). One of the assumptions underlying the use of ecoregions in conservation
planning is that diverse taxa are thought to exhibit similar patterns in relation to
ecoregion boundaries because they are subject to common influences of geology, climate,
and structure (Spector 2002). However, several authors have pointed out that it is
unlikely that any single system of dividing the landscape into regions will adequately

represent the distribution of all taxa (Olson et a1. 2001, Magnusson 2004).

103

Conservation planning usually involves choosing a surrogate taxon or a few taxa
to represent the biodiversity of an area. Vascular plants, birds, or butterflies are often
used as surrogate taxa because their taxonomic and distribution status are relatively well
known world wide (Ricketts et a1. 1999, Margules and Pressey 2000). Natural vegetation
is perhaps the most commonly used criterion in defining landscape units, although there
have been relatively few studies looking at whether divisions based on vegetation
structure have meaning for other taxa (Goldstein 1999). Those studies that have been
conducted show inconsistent results (Poynton and Boycott 1996, Mac Nally et a1. 2002,
Gering et al. 2003). This study tests the hypothesis that the distribution of avian species
composition corresponds to the distribution of ecoregions in the continental United
States. That is, bird community assemblages in ecoregions composed of comparable
soils, climate, and vegetation are predicted to be similar to each other, and dissimilar to
assemblages in ecoregions with different characteristics.

Assessing the correspondence of species assemblages to ecoregion distribution
requires a method of measuring similarity and dissimilarity. The dissimilarity of species
assemblages has commonly been described by three measures of diversity: alpha, beta,
and gamma (Whittaker 1975). Of these, beta diversity measures the turnover, or change
in the species composition between two communities. Multiple measures of beta diversity
have been proposed (Koleff et al. 2003), each emphasizing different aspects of
community structure. These differences among beta diversity metrics should result in
different assessments of the correspondence of bird community assemblages to ecoregion

distribution.

104

Methods

Data sources

This study used the National Ecological Unit Hierarchy (Cleland et a1. 1997)
which is a classification of ecological units encompassing spatial scales ranging from <
50 hectares to millions of square kilometers. In order of increasing size, the ecological
units are landtype phase, landtype, landtype association, subsection, section, province,
division, and domain. Two units within this hierarchy were used in this study, provinces
and sections. Provinces are delineated primarily by continental weather patterns, and are
characterized by similar soil orders and potential natural vegetation. Provinces are
composed of sections, which are distinguished by similar sub-regional climates, geology,
and potential natural vegetation. Only provinces (n = 35) and sections (11 = 163) within
the contiguous United States were included (Table D. 1 , Figure 4.1). The Channel Islands
off the coast of southern California were excluded from analysis. A GIS coverage of this
ecoregion classification scheme was obtained from the US. Forest Service (USDA Forest
Service Inventory and Monitoring Institute 2005).

The composition of bird community assemblages across the continental United
States was determined using species occurrence data from the North American Breeding
Bird Survey (Sauer et al. 2001). Species occurrence data from all routes within the
continental United States from 1966-2000 were included. BBS routes consist of 50
sampling points located 0.8 km apart along secondary roads. Surveys are conducted once
per year during June, and all birds seen or heard at each sampling point are recorded

(Robbins et al. 1986). A GIS coverage of all survey routes monitored by the BBS was

105

obtained from the National Atlas of the United States® (US. Department of the Interior
2005).

Species lists for each of the 35 provinces and 163 sections were generated using
ArcView 3.2 (Environmental Systems Research Institute Inc. 1992-1999) to identify the
BBS routes occurring within each province and section. Routes that crossed province and
section boundaries were excluded from province and section species lists, respectively.
However, routes that crossed section boundaries within a province were included in the
list for that province. Consequently a species list for a province could contain more
species than any of the component section lists, but not fewer. The Northern Rio Grande
Interrnontane section (43) contained no BBS routes entirely within its boundaries.

Consequently a species list could not be generated for this section.

Data analysis

To compare the adequacy of BBS sampling across sections and provinces of
varying area, species-area plots were created based on avian species lists at both the
section and province level (Figures 4.3 and 4.4). Species richness was also plotted against
sampling intensity, calculated as the number of BBS routes per km2 in both sections and
provinces (Figures 4.5 and 4.6).

Seven beta-diversity metrics (Table 4.1) were used to assess the differences
between the bird community assemblages at two spatial scales, provinces and sections.
These seven metrics represent the four categories of metrics described by Koleff et a1.
(2003): measures of continuity and loss, species richness gradients, continuity, and gain

and loss. The seven selected also include metrics commonly used in diversity studies

106

(Bw, Bj, and [350) as well as those metrics which have been shown to perform well in
terms of ability to reflect community turnover, additivity, and independence from alpha
diversity and sample size (BW and fit) (Wilson and Shmida 1984, Chao et al. 2005).
Jaccard’s similarity index, Bj, here expressed as a diversity index, is widely used and
easily calculated. Values of Bj are strongly dependent on matching component a (Figure
4.2, Koleff et a1. 2003) and weight all species equally, which contributes to its poor
performance when the species assemblage contains a large proportion of rare species
(Chao et a1. 2005). Values of Simpson’s diversity index, Bsi, depend on matching
component a as well as the relative magnitude of b and c (Koleff et al. 2003). Unlike Bj,
Bsi is weighted towards common species and so may be inappropriate for assemblages
with a large component of rare species (Gering et a1. 2003). Local species richness
gradients do not affect calculated values of Bsi (Lennon et al. 2001). Ruggiero et a1.
(1998) developed a similarity index, Brlb, here expressed as a diversity index, which is
the only metric to fall into the category of continuity and loss (Koleff et al. 2003). Values
for Brlb are dependent on matching components a and c. Bgl is a measure of local species
richness gradients, and as such has been used to test the performance of other beta-
diversity metrics (Lennon et a1. 2001). Values for Bgl depend solely on matching
components b and c (Koleff et al. 2003). Sorensen’s similarity index, [330, here expressed
as a diversity metric, is also widely used and easily calculated. As with Bj, Bso weights
all species equally and so is inappropriate for assemblages containing many rare species
(Chao et al. 2005). Unlike [351, 1330 is affected by local species richness gradients (Lennon

et a1. 2001). Whittaker’s diversity index, Bw, is probably the most widely used beta-

107

diversity metric (Koleff et al. 2003). Values of Bw decrease with increases in the value of
matching component a. As with 1350, BW is affected by local species richness gradients
(Lennon et a1. 2001). Wilson and Shmida (1984) developed a measure of beta turnover,
fit, that has also become widely used, perhaps because it is often more easily interpretable
than Bw. Like 13w, Bt also decreases with increases in matching component a (Koleff et
al. 2003).

Using the equations presented in Koleff et al. (2003), which compare species
occurrence data based on some or all of the components depicted in Figure 4.2, distance
matrices were calculated for these seven beta-diversity metrics. All calculations and
analyses were conducted with the R statistical package (R Development Core Team
2004) unless otherwise indicated.

The Nearest Features ArcView extension (Jenness 2004) was used to calculate the
actual geographic distance between section edges of all possible pairs of the 163 sections.
Adjoining sections were defined as having a distance of zero. All small coastal polygons
(e.g. barrier islands) were excluded from these calculations. For noncontiguous sections
such as the Central Loess Plains, which is composed of two distinct polygons, the smaller
of the two distances between the Central Loess Plains polygons and any other section was
retained in the geographic distance matrix. The resulting 163x163 section geographic
distance matrix was also condensed into a 35x35 province geographic distance matrix.
The inter-province distance for each possible pair of provinces was defined as the
smallest distance between all possible pairs of sections comprising each province.

At the scale of both sections and provinces, the beta-diversity and geographic

distance matrices were classified using a hierarchical agglomerative technique,

108

unweighted arithmetic average clustering (UPGMA) (Legendre and Legendre 1998),
implemented in R (R Development Core Team 2004) using the cluster package
(Rousseeuw et a1. 2004). The beta-diversity matrices were also classified using a divisive
method, partitioning around medoids (PAM), at each spatial scale (Kaufman and
Rousseeuw 1990). Each of the seven beta-diversity matrices based on section species lists
(n = 162) was partitioned into 35 clusters since there are 35 provinces in the contiguous
United States (Cleland et al. 1997). At the province level, each of the seven beta-diversity
matrices based on province species lists (n = 35) was partitioned into 19 clusters since
there are 19 divisions in the contiguous United States (Cleland et al. 1997).

Mantel tests with 1000 permutations were used to assess the congruence of the
seven beta-diversity matrices with each other and with the corresponding geographic
distance matrices at both the section and province spatial scales (Legendre and Legendre
1998). A sectional species lists could not be generated for the Northern Rio Grande
Interrnontane section (43); therefore, for the Mantel tests at the section level all
geographic distance data associated with this section were deleted, creating a section
geographic distance matrix with the same dimensions (162x162) as the beta-diversity

matrices.

Results

The 163 sections of the contiguous United States vary in area from 100 km2 to
1170 km2 (Table D.1). The species richness of the 162 sections for which species lists
could be generated varies from 54 to 259 (Figure 4.3). The 35 provinces vary in area

from 120 km2 to 6702 kmz, while species richness varies from 115 to 314 species (Figure

109

4.4). Although the number of species does increase with increases in province area (r2 =
0.461), at the section level there is no significant change in species richness with
increases in area (r2 = 0.001, F 1,160 = 0.0994, p = 0.753). Species richness increases with
increases in sampling intensity at the section level (Figure 4.5; r" = 0.313), but there is no
significant change in species richness with changes in sampling intensity across
provinces (Figure 4.6; r2 = 0.001, F 1,33 = 0.037, p = 0.849). Sampling intensity decreases
with area for both sections and provinces (Figures 4.7 and 4.8; r2 = 0.273 and 0.248,
respectively).

UPGMA cluster analysis of the beta-diversity matrices from both section and
province species lists produced dendrograms that were all similarly correlated with the
original data (Table 4.2). However, clustering based on section species lists consistently
had lower cophenetic correlations and higher root mean squared errors than those based
on provinces, while the clustering of the geographic distance matrix based on sections
had a higher c0phenetic correlation and lower root mean squared error than that based on
provinces.

Three of the seven dendrograms produced by UPGMA analysis of the beta-
diversity matrices based on section species lists, 1330, BW, and Bt, were identical, with
pairwise correlations of cophenetic distances equal to 1.00 (Table 4.3). The dendrogram
based on the Bj matrix was very similar, with differences in the grouping of only a few
sections. The remaining three dendrograms, Brlb, Bgl, and Bsi were each distinct, with the
13g] dendrograms being notable for its particularly short branch lengths.

Six of the seven dendrograms produced by UPGMA analysis of the beta-diversity

matrices based on province species lists were broadly similar in that the largest two

110

clusters (n = 16 and n = 19) were composed of the same provinces. Within these largest
clusters, however, the Brlb and Bsi dendrograms had a different pattern of aggregation
than in the other four dendrograms. The clustering patterns of the 1350, BW, and Bt
dendrograms were identical (Table 4.4). The l3j dendrogram was very similar to the
dendrograms of 1350, BW, and fit; only the branch lengths varied slightly. The pattern of
aggregation in the Bgl dendrogram was distinct from all other section dendrograms. The
largest two clusters were unbalanced in size (n = 30 and n = 5), and neither was identical
in composition to clusters in any other dendrogram.

Mantel tests comparing dissimilarity matrices based on section species lists
indicate that, with two exceptions (the comparison of 133i to 13g], p = 0.938, and Bgl to
geographic distance, p = 0.313), the beta-diversity distance matrices were similar both to
each other and the geographic distance matrix (p < 0.001) (Table 4.5). At the province
level, with three exceptions (the comparison of fig] to Bsi, p = 1, fig] to Brlb, p = 0.055,
and fig] to geographic distance, p = 0.206), the beta-diversity matrices were again highly
correlated with each other and with the geographic distance matrix (p < 0.001) (Table
4.6).

The results of the PAM analysis for the species lists based on sections (n = 162)
show a consistently high rate of misclassification across all seven beta-diversity metrics
(misclassification rate ranged from 0.89 to 1.0). Of the 35 provinces that the 162 species
lists were expected to cluster into, only one was classified correctly in multiple analyses.
The six sections of province 47 (Southwest Plateau and Plains Dry Steppe and Shrub)
were grouped together by the cluster analysis of the Bj, 1330, BW, and 0t beta-diversity

matrices. In all other instances at least one of the sections comprising a province was

111

placed in a group with sections from another province. The results of the PAM analysis
for the species lists based on provinces (n = 35) show a slightly lower level of
misclassification (rate ranged from 0.58 to 0.79), with more provinces that were correctly

grouped together into divisions by multiple distance matrices (Table 4.7).

Discussion

The use of BBS data introduces several potential sources of bias. The BBS survey
methods are not equally successful for all bird species; for example, populations of
nocturnal species are not well-described by BBS data. Although the species lists
generated from BBS data may not accurately represent the complete species assemblage
of a section or province, the biases are likely to be consistent across all regions. The
number of species observed in an area commonly increases with the size of the area
surveyed (MacArthur and Wilson 1967). Although this trend was present at the province
level (Figure 4.4), it did not emerge at the section level (Figure 4.3). Larger provinces
tend to be composed of more sections than smaller provinces, and therefore contain more
habitat types, which increases species richness. BBS sampling effort is not constant
across the contiguous United States. When sampling effort is expressed as route density
(routes/kmz), species richness increases with increasing sampling intensity at the section
level (Figure 4.5), but there is no change in species richness with sampling intensity at
the province level (Figure 4.6). This suggests that, at the province level, the BBS survey
data adequately captures richness; that is, adding more routes within a province is not
likely to increase species richness unless sampling methods are also changed. However,

within sections increasing sampling effort does increase the number of observed species,

112

perhaps by increasing the likelihood of finding individuals of species more commonly
found in neighboring sections within the province.

Although there are many metrics that describe beta-diversity, this study was
limited to those appropriate for presence/absence data. Four of these metrics, Bj, 1330, BW,
and (St, were categorized by Koleff et al. (2003) as measures of continuity, or “broad-
sense” turnover, which focus on differences in composition but not on the relative
magnitude of species gains or losses. Variation in the values of these measures depends
primarily on differences in the matching component a (Figure 4.2). Reflecting this
common influence the Bj, 1350, BW, and 0t distance matrices based on both section and
province species lists performed almost identically in UPGMA cluster analysis, Mantel
tests, and PAM cluster analysis (Tables 4.2 and 4.5 — 4.7). The Bj distance matrix
clustered only a few sections differently in the UPGMA analysis of section species lists.

Of the other three beta-diversity metrics, Brlb was described by Koleff et al.
(2003) as a measure of continuity and loss, depending on the values of both a and c
(Figure 4.2). Though originally formulated to create an equiprobabilistic map based on
species occurrences across multiple quadrats (Ruggiero et al. 1998), it was re-expressed
by Koleff et a1. (2003) to reflect only a single pair-wise comparison. In this re-expressed
form it is mathematically comparable to the other metrics, though in practice cophenetic
correlations tended to be low and root mean square error tended to be high (Table 4.2).
Bsi was categorized as a measure of gain and loss (Koleff et al. 2003), or “narrow-sense”
turnover that focuses on changes in species composition rather than changes in species
richness. Unlike the “broad-sense” turnover measures, variation in Bsi depends on both

the value of a and the relative magnitude of b and c.

113

The last metric, 0g], is distinct from the other six. Described by Koleff et al.
(2003) as a species richness gradient measure, it was not originally intended as a measure
of beta-diversity but rather as a measure of local richness gradients, or alpha diversity
(Lennon et al. 2001). As such, it is not surprising that at the spatial scale of this study 13g]
performed rather poorly. The Bgl distance matrices were relatively poorly correlated with
the original data (Table 4.2) and with the geographic distance matrices (Tables 4.3 and
4.4).

The scale of a study also can affect beta-diversity values. As the sampling grain
increases, species richness tends to increase (MacArthur and Wilson 1967), and as the
spatial extent of a study area increases, diversity between observations also tends to
increase (Nekola and White 1999). A study of birds and butterflies in the North American
Great Basin demonstrated empirically that as sampling grain increased, the similarity of
the species composition of.two assemblages also increased (Mac Nally et al. 2004). The
range of the sampling grain in this Great Basin study — sites, canyons, and mountain
ranges — covered about five orders of magnitude. The smallest sample size (sites)
assessed the distribution of individuals while the largest sample size (mountain ranges)
assessed the distribution of populations. Similarly to Mac Nally et al. (2004), the present
study of North American ecoregions does not have a consistent scale as sizes of both
sections and provinces vary. However, the size variation from the smallest section (100
kmz) to the largest province (6702 kmz) is less than two orders of magnitude. This
relatively small amount of variation in sampling grain is unlikely to affect the calculated
beta-diversity values. Additionally, the section and province species lists generated from

BBS survey data describe the distribution of populations and species, not individuals,

114

over a much longer time frame than that in the Great Basin study (Mac Nally et al. 2004).
The results of the Great Basin study are much more likely to have been influenced by
short-term local ecological conditions, whereas the results of this study of North
American ecoregions are more likely to represent long-term regional ecological
conditions.

Ecoregions are frequently used as the foundation of conservation planning, based
in part on the assumption that ecoregions likely represent the distribution of taxa (Groves
et al. 2000, Olson et al. 2001). Under this assumption, a comparison of the composition
of the section species lists should enable the recreation of provinces. With a divisive
classification method like PAM, the expectation was that partitioning the 162 section
species lists into 35 clusters would result in clusters that conformed to the 35 provinces
found in the contiguous United States. However, such correspondence was the exception
rather than the rule. The Brlb matrix was the most successful, correctly matching four
groups of species lists to their respective provinces. The Bgl matrix was the least
successful, with no provinces correctly identified. Only one province, the Southwest
Plateau and Plains Dry Steppe and Shrub, was correctly classified using the data from
multiple beta-diversity matrices. At a larger spatial scale, the results were slightly better.
Using PAM, the expectation was that partitioning the 35 province species lists into 19
clusters would result in clusters that conformed to the 19 divisions found in the
contiguous United States. Though there were more correct classifications by multiple
beta-diversity matrices (Table 4.7), the error rate was still > 50% for each beta-diversity

metric.

115

For the birds of North America, it appears that avian beta diversity does not
correspond to biogeographic regions based on soils, climate, and vegetation. Instead, the
results of this study suggest that the geographic distance between species assemblages is
a stronger influence on their composition. With the exception of the fig] matrix, all beta-
diversity matrices were strongly correlated with the geographic distance matrix (Mantel
tests, p < 0.001). This fits with other work that has shown the distribution of North
American bird species to be spatially autocorrelated (Brown et a1. 1995). Many sections
that are grouped together into a province by the National Ecological Unit Hierarchy
(Cleland et al. 1997) actually occur over a wide geographic area. For example, the ten
sections of the Great Plains-Palouse Dry Steppe Province are distributed over a 1300 km
swath from the Canadian border south to northern Texas while the seven sections of the
Outer Coastal Plain Mixed Forest Province range over 1500 km from the Atlantic coast
west to southeastern Texas.

, The validity of the concept of ecoregions can only be assessed by addressing
complex issues such as human and non-human perceptions of landscape characteristics,
the permeability of boundaries between landscapes for various species, and the dynamic
nature of landscape characteristics. However, the results of this study do provide insight
into the utility of ecoregions for a particular purpose, environmental management and
conservation planning. The provinces and sections of the National Ecological Unit
Hierarchy, which are based on continental weather patterns, soils, and potential natural
vegetation, do not seem to reflect the distribution of birds in North America, though they
may be adequate for other taxa (Gering et a1. 2003). The distribution of bird species

across North America may follow a Gleasonian continuum pattern, where the range of

116

occurrence of each species is independent of the co-occurrence of a particular plant

association or ecoregion (Barbour et al. 1987). Although some species within a particular
avian assemblage may indeed have distributions that correspond to ecoregion boundaries,
the results of this study suggest that overall, North American avian species are distributed

independently of the distribution of ecoregion landscape units.

117

Table 4.1. Beta-diversity metrics used in analyses; equations as presented in Koleff et al.
(2003).

 

 

B-Diversity Metric Equation
Jaccard Bj = l — [a / (a + b + c)]
Simpson Bsi = [min{b,c)] / [(min(b,c)) + a]
Ruggiero et a1. Brlb = 1 — [a / (a + c)]
Lennon et al. [3g] = 2( lb - c| )/(Za + b + c)
Sorenson Bso = 1 — [2a / (23 + b + c)]
Whittaker Bw = (a + b + c) / [(2a + b + c) / 2]
Wilson & Shmida [it = (b + c) / (2a + b + c)

 

118

Table 4.2. Hierarchical agglomerative clustering (UPGMA) results for species lists based
on sections (n = 162) and provinces (n = 35), and geographic distance between sections
(n = 163) and provinces (n = 35).

 

 

B-Diversity Cophenetic Correlation Root Mean Square Error

Matrix Section Province Section Province
Bj 0.857343 0.8726094 0.0853477 0.07726224
Bsi 0.8163491 0.8298172 0.1078507 0.09663368
Brlb 0.7500869 0.7657905 0.1261695 0.1225102
Bgl 0.748039 0.8808222 0.1318996 0.1067753
Bso 0.8267704 0.8498973 0.09590017 0.0818542
13w 0.8267704 0.8498973 0.09590017 0.0818542
Bt 0.8267704 0.8498973 0.09590017 0.0818542
Geographic distance 0.699514 0.6650709 6950686 706924

 

119

Table 4.3. Correlation of cophenetic distances for each of the beta-diversity matrices
based on section species lists (11 = 162).

 

Bj Bsi Brlb Bgl Bso [3w [3t

 

Bi

Bsi 0.885891

Brlb 0.839922 0.862965

Bgl 0.141615 0.030986 0.052162

Bso 0.988724 0.896334 0.855323 0.129944

Bw 0.988724 0.896334 0.855323 0.129944 1.000000

Bt 0.988724 0.896334 0.855323 0.129944 1.000000 1.000000

 

120

Table 4.4. Correlation of cophenetic distances for each of the beta-diversity matrices
based on province species lists (11 = 35).

 

Bj Bsi Brlb Bgl 1350 BW Bt

 

131

Bsi 0.879076

Brlb 0.864968 0.894623

Bgl 0.186577 0.021726 0.047332

BSO 0.996568 0.899755 0.877283 0.176228

Bw 0.996568 0.899755 0.877283 0.176228 1.000000

[3t 0.996568 0.899755 0.877283 0.176228 1.000000 1.000000

 

121

Table 4.5. P-values for comparisons of beta-diversity matrices based on section species
lists and the section geographic distance matrix using Mantel tests with 1000 replicates.

 

 

Bj Bsi Brlb Bgl 1350 BW [3t GeoDist

131'

Bsi <0.001

Brlb <0.001 <0.001

Bgl <0.001 0.938 <0.001

Bso <0.001 <0.001 <0.001 <0.001

Bw <0.001 <0.001 <0.001 <0.001 <0.001

Bt <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
GeoDist <0.001 <0.001 <0.001 0.313 <0.001 <0.001 <0.001

 

122

Table 4.6. P-values for comparisons of beta-diversity matrices based on province species
lists and the section geographic distance matrix using Mantel tests with 1000 replicates.

 

 

Bj Bsi Brlb Bgl 1350 BW 0t GeoDist

Bi

[331 <0.001

Brlb <0.001 <0.001

Bgl <0.001 1.000 0.055

Bso <0.001 <0.001 <0.001 <0.001

Bw <0.001 <0.001 <0.001 <0.001 <0.001

0t <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
GeoDist <0.001 <0.001 <0.001 0.206 <0.001 <0.001 <0.001

 

123

Table 4.7. Province species lists (11 = 35) correctly classified by PAM into 19 divisions
for each of the seven beta-diversity matrices (Note: Division 19 is located outside the
contiguous United States).

 

 

Division Bj Bsi Brlb [lg] 1350 BW fit
1 I/
2 I/
3
4 I/ r I/ r r r
5 .r I/ .r r r r
6 J I/
7 Ir J
8 J
9
10 I/ I/ I/ I/
1 1 I/
12 J \/ I/ I/ \/ I/
13 r I/ r I/
14 V r/
15 l r .r r
16 I/
17
18
20 J I/ J r

 

124

 

 

 

 

 

 

 

 

 

Figure 4.2. The matching/mismatching components used in pairwise comparisons of
presence/absence data: (a) the total number of species common to both regions, (b) the
number of species that occur only in the neighboring region, and (c) the number of
species that occur only in the focal region.

126

 

 

 

 

 

 

2.6

8 2.4 -

(D

C
.C

.9

m o
.3 2.2 -

0

OJ

C.

(I) o
C

_9 2.0 -

‘5

OJ

2’,

S

8’ 1.8 ~

..I

o
1.6 I I I I I I
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2

Log,O(Section Area (km2))

Figure 4.3. Section species richness vs. area (kmz) (n = 162) with least squares linear
regression (r2 = 0.001).

127

 

2.6

2.5 a O

2.4 -

2.3 -

 

Log 10( Province Species Richness)

 

 

 

2.2 -
O
9 o
2.0 I I I I
2.0 2.5 3.0 3.5

Log,0(Province Area (km2))

Figure 4.4. Province species richness vs. area (kmz) (n = 35) with least squares linear
regression (r2 = 0.461).

128

 

2.6

2.4 -

2.2 -

2.0 ‘

 

Log 1o(Section Species Richness)

 

 

 

1.6 I I I
-3 -2 -1 0

Log,0(BBS routes/kmz)

Figure 4.5. Species richness and sampling intensity (BBS routes/kmz) for sections (11 =
162) with least squares linear regression (r2 = 0.313).

129

 

2.6

2.5 - O

2.4 ‘

 

 

2.3 -

2.2 '1 .

Log10(Province Species Richness)
o

. O

 

 

 

2.0 I I I I I I I I
-2.2 -2.0 -1.8 ~1.6 -1.4 .-1,2 -1.0 ~0.8 -0.6 -O.4

Log,0(BBS routes/kmz)

Figure 4.6. Species richness and sampling intensity (BBS routes/kmz) for provinces (n =
35) with least squares linear regression (r‘ = 0.001).

130

0.0

 

-0.5 ~

-1.0 -

-1.5 -

-2.0 -

-2.5 -

Log10(BBS Routes/kmz)

 

-3.0 ~

 

 

 

'3.5 I I I I I I
1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2

Log,o(Section Area (km2))

Figure 4.7. Sampling intensity (BBS routes/kmz) vs. area for sections (n = 162) with least
squares linear regression (r' = 0.273).

131

 

-0.6 ‘ O

-0.8 -

-1.0 4 O O

-1.2 -

-1.4 ~

-1.6 5

Log 10(BBS Routes/kmz)

-1.8 -

 

-2.0 ~ '

 

 

 

'2.2 I I I r I I T I I I

1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0
Log10(Province Area (km2))

Figure 4.8. Sampling intensity (BBS routes/kmz) vs. area for provinces (n = 35) with least
squares linear regression (r2 = 0.248).

132

CONCLUSIONS

Patterns of Biological Diversity

The results of my dissertation illustrate several patterns of biological diversity. In
both Chapter 2 and 3, avian species assemblages from eastern North America were
shown to exhibit discontinuous distributions of body masses. These patterns are robust;
with similar discontinuities identified using multiple methods. In Chapter 2 I identified
the potential ecological correlates of discontinuities in body mass distributions for the
birds of these assemblages. Declining species typically had body masses that placed them
near the discontinuities in the body mass spectra. This suggests that in a species
assemblage, having such a body mass presents challenges to maintaining a viable
population. Non-indigenous species typically had body masses that placed them far from
discontinuities in the body mass spectra, suggesting that the body mass of a species may
be a factor in determining the ability of a species to successfully invade an ecosystem.
The species composition of birds across the contiguous United States was analyzed in
Chapter 4. Using seven different beta-diversity metrics, I compared the composition of
species assemblages in 162 sections. The results showed that the distribution of bird
species is spatially autocorrelated, with sections closer together having assemblages that

are more similar in composition than those of sections that are farther apart.

Patterns of Landscape Structure

The structure of landscapes can be described with many different metrics, not all

of which are useful to understanding patterns of biodiversity. Chapter 1 identifies those

133

metrics that are most likely to affect the formation and maintenance of species
assemblages. Using some of these metrics (diversity, dominance, and lacunarity), the
land cover of six sections of eastern North America was characterized in Chapter 3. As
described in Chapter 2, the activities of humans tend to homogenize landscapes. The
results of Chapter 3 showing that agriculturally dominated landscapes had the lowest
diversity and highest dominance values were not surprising. Lacunarity, which can
identify the presence of hierarchical structure in landscapes, was used to describe the
heterogeneity of the distribution of forest and non-forest habitat in eastern North
America. Forest habitat was found to be quite homogeneously distributed in most
sections, but more heterogeneously distributed in areas dominated by agriculture. Unlike
the non-forest habitat lacunarity curves, the forest habitat lacunarity curves were all quite
smooth, indicating that there were few discontinuities in the structure of forest habitat in

any of the areas analyzed.

Correspondence of Landscape Structure and Biodiversity

Though it has long been known that the spatial pattern of landscape elements can
affect ecological processes occurring in landscapes, it has been difficult to quantify these
links between pattern and process. Lacunarity analysis provided me with a means of
testing such a proposed link, the entrainment of species attributes by the structure of the
landscape. In Chapter 3 I used lacunarity to test the hypothesis that the distribution of
discontinuities in body mass would correspond to the discontinuities in landscape
structure. Though the results presented in Chapter 3 do not support such a

correspondence, they do suggest several directions for future research. In Chapter 4 I

134

tested the hypothesis that the spatial distribution of ecoregions corresponds to the
distribution of avian taxa. I found that ecoregions could not be recreated based on the
composition of avian species assemblages, showing that the distribution of North
American bird species is independent of the distribution of ecoregions defined on the

basis of climate, soils, and natural vegetation.

Future Directions

The results of this dissertation suggest several directions for future research. I am
interested in determining if the ecological consequences of discontinuities in the body
mass distributions identified in Chapter 2 will be found in other regions and with other
taxa. Analysis of different types of non-indigenous species (e.g. those that are established
vs. those that are still actively invading, “natural” invaders vs. human-introduced species)
and declining species (e. g. those that have already been extirpated from a region vs. those
that are currently declining) may reveal different patterns. The results of Chapter 2
suggested that ecological consequences may vary depending on the amount of human
impact and landscape homogenization, but this has not yet been quantitatively
investigated.

The use of lacunarity analysis to quantify the link between landscape structure
and biodiversity patterns is promising. However, based on the results of Chapter 3,
modifying the APACK software program used to calculate lacunarity could more
accurately reflect the spatial pattern of habitat in the landscape of interest. Additionally,

the spatial distribution of habitat types other than upland forest should also be examined.

135

APPENDIX A: SPECIES LISTS

136

 

Table A. 1. Species list for section 47, the Central Till Plains, Oak-Hickory Section.

 

 

137

AOU Code Common Name Scientific Name Mass (g)
02890 Northern Bobwhite Colinus virginianus 178.0
03050 Greater Prairie-Chicken Tympanuchus cupido 999.0
03091 Ring-necked Pheasant Phasianus colchicus 1317.0
03100 Wild Turkey Meleagris gallopavo 7400.0
03131 Rock Dove Columba livia 369.0
03160 Mourning Dove Zenaida macroura 123.0
03250 Turkey Vulture Cathartes aura 1467.0
03310 Northern Harrier Circus cyaneus 358.0
03330 Cooper's Hawk A ccipiter cooperii 349.0
03370 Red-tailed Hawk Buteojamaicensis 1028.0
03390 Red-shouldered Hawk Buteo lineatus 475.0
03600 American Kestrel F alco sparverius 1 11.0
03680 Barred Owl Strix varia 632.0
03730 Eastern Screech-Owl Otus asio 174.0
03750 Great Horned Owl Bubo virginianus 1318.0
03870 Yellow-billed Cuckoo Coccyzus americanus 64.0
03880 Black-billed Cuckoo Coccyzus erythropthalmus 51.1
03900 Belted Kingfisher Ceryle alcyon 148.0
03930 Hairy Woodpecker Picoides villosus 70.0
03940 Downy Woodpecker Picoides pubescens 27.0

Table A] (cont’d).

 

 

138

AOU Code Common Name Scientific Name Mass (g)
04050 Pileated Woodpecker Dryocopus pileatus 308.0
04060 Red-headed Woodpecker Melanerpes erjt'throcephalus 71.6
04090 Red-bellied Woodpecker Melanerpes carolinus 67.2
04120 Northern Flicker Colaptes auratus 135.0
04170 Whip-poor-will Caprimulgus vociferus 55.3
04200 Common Nighthawk Chordeiles minor 61.5
04230 Chimney Swift Chaetura pelagica 23.6
04280 Ruby-throated Hummingbird Archilochus colubris 3.0
04440 Eastern Kingbird T yrannus tyrannus 43.6
04470 Western Kingbird T yrannus verticalis 39.6
04520 Great Crested Flycatcher Myiarchus crinitus 33.5
04560 Eastern Phoebe Sayornis phoebe 19.8
04610 Eastern Wood-Pewee Contopus virens 14.1
04650 Acadian Flycatcher Empidonax virescens 12.9
04660 Willow Flycatcher Empidonax traillii 13.1
04670 Least Flycatcher Empidonax minimus 10.3
04740 Horned Lark Eremophila alpestris 31.9
04770 Blue Jay Cyanocitta cristata 86.8
04880 American Crow Corvus brachyrhynchos 458.0
04930 Eur0pean Starling Sturnus vulgaris 84.7

Table A.1 (cont’d).

 

 

139

AOU Code Common Name Scientific Name Mass (g)
04940 Bobolink Dolichonyx oryzz'vorus 47.0
04950 Brown-headed Cowbird Molothrus ater 49.0
04970 Yellow—headed Blackbird Xanthocephalus xanthocephalus 79.7
04980 Red-winged Blackbird Agelaius phoeniceus 63.6
05010 Eastern Meadowlark Sturnella magna 102.0
05011 Western Meadowlark Sturnella neglecta 112.0
05060 Orchard Oriole Icterus spurius 19.6
05070 Baltimore Oriole [cterus galbula 34.3
05110 Common Grackle Quiscalus quiscula 127.0
05190 House Finch Carpodacus mexicanus 21.4
05290 American Goldfinch Carduelis tristis 13.2
05400 Vesper Sparrow Pooecetes gramineus 26.5
05420 Savannah Sparrow Passerculus sandwichensis 20.6
05460 Grasshopper-Sparrow Ammodramus savannarum 17.0
05520 Lark Sparrow Chondestes grammacus 29.0
05600 Chipping Sparrow Spizella passerina 12.3
05630 Field Sparrow Spizella pusilla 12.5
05810 Song Sparrow Melospiza melodia 21.0
05870 Eastern Towhee Pipilo erythrophthalmus 41.7
05930 Northern Cardinal Cardinalis cardinalis 45.4

Table A.1 (cont’d).

 

 

140

AOU Code Common Name Scientific Name Mass (g)
05950 Rose-breasted Grosbeak Pheucticus ludovicianus 45.6
05970 Blue Grosbeak Guiraca caerulea 29.3
05980 Indigo Bunting Passerina cyanea 14.9
06040 Dickcissel Spiza americana 29.3
06080 Scarlet Tanager Piranga olivacea 28.6
06100 Summer Tanager Piranga rubra 28.2
061 10 Purple Martin Progne subis 49.4
06120 Cliff Swallow Hirundo pyrrhonota 21.6
06130 Barn Swallow Hirundo rustica 16.2
06140 Tree Swallow T achycineta bicolor 20.1
06160 Bank Swallow Riparia riparia 14.6
06170 Northern Rough-winged Stelgidopteryx serripennis 15.9

Swallow
06190 Cedar Waxwing Bombycilla cedrorum 30.6
06220 Loggerhead Shrike Lanius ludovicianus 47.4
06240 Red-eyed Vireo Vireo olivaceus 16.7
06270 Warbling Vireo Vireo gilvus 14.8
06280 Yellow-throated Vireo Vireo flavifrons 18.0
06310 White-eyed Vireo Vireo griseus 1 1.4
06330 Bell’s Vireo Vireo bellii 8.5
06360 Black-and-white Warbler Mniotilta varia 11.0

Table A.1 (cont’d).

 

 

141

AOU Code Common Name Scientific Name Mass (g)
06370 Prothonotary Warbler Protonotaria citrea 14.3
06410 Blue-winged Warbler Vermivora pinus 8.4
06480 Northern Parula Parula americana 8.6
06520 Yellow Warbler Dendroica petechia 9.8
06580 Cerulean Warbler Dendroica cerulea 9.5
06630 Yellow-throated Warbler Dendroica dominica 9.4
06710 Pine Warbler Dendroica pinus 11.9
06730 Prairie Warbler Dendroica discolor 8.0
06740 Ovenbird Seiurus aurocapillus 19.4
06770 Kentucky Warbler Oporornisformosus 14.3
06810 Common Yellowthroat Geothlypis trichas 10.1
06830 Yellow-breasted Chat Icteria virens 25.5
06840 Hooded Warbler Wilsonia citrina 10.8
06870 American Redstart Setophaga ruticilla 8.5
06882 House Sparrow Passer domesticus 28.0
06883 Eurasian Tree Sparrow Passer montanus 22.0
07030 Northern Mockingbird Mimus polyglottos 48.5
07040 Gray Catbird Dumetella carolinensis 36.9
07050 Brown Thrasher T oxostoma rufum 68.8
07180 Carolina Wren T hryothorus ludovicianus 18.7

Table A.1 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
07210 House Wren T roglodytes aedon 10.9
07240 Sedge Wren Cistothorus platensis 9.0
07270 White-breasted Nuthatch Sitta carolinensis 21.1
07310 Tufted Titmouse Parus bicolor 21.6
07350 Black-capped Chickadee Parus atricapillus 10.8
07360 Carolina Chickadee Parus carolinensis 10.5
07510 Blue-gray Gnatcatcher Polioptila caerulea 6.0
07550 Wood Thrush Hylocichla mustelina 47.8
07610 American Robin T urdus migratorius 77.3
07660 Eastern Bluebird Sialia sialis 31.6
03260 Black Vulture Coragyps atratus 2172.0
05470 Henslow's Sparrow Ammodramus henslowii 13.1
06760 Louisiana Waterthrush Seiurus motacilla 19.8

 

142

Table A2. Species list for section 59, the Central Ridge and Valley Section.

 

 

143

AOU Code Common Name Scientific Name 'Mass (g)
02890 Northern Bobwhite Colinus virginianus 178.0
03000 Ruffed Grouse Bonasa umbellus 621.0
03100 Wild Turkey Meleagris gallopavo 7400.0
03131 Rock Dove Columba livia 369.0
03160 Mourning Dove Zenaida macroura 123.0
03250 Turkey Vulture Cathartes aura 1467.0
03260 Black Vulture Coragyps atratus 2172.0
03320 Sharp-shinned Hawk A ccipiter striatus 103.0
03330 Cooper's Hawk Accipiter cooperii 349.0
033 70 Red-tailed Hawk Buteojamaicensis 1028.0
03390 Red-shouldered Hawk Buteo lineatus 475.0
03430 Broad-winged Hawk Buteo platypterus 420.0
03600 American Kestrel F alco sparverius 1 11.0
03 680 Barred Owl Strix varia 632.0
03730 Eastern Screech-Owl Otus asio 174.0
03750 Great Horned Owl Bubo virginianus 1318.0
03870 Yellow-billed Cuckoo Coccyzus americanus 64.0
03900 Belted Kingfisher C eryle alcyon 148.0
03930 Hairy Woodpecker Picoides villosus 70.0
03940 Downy Woodpecker Picoides pubescens 27.0

Table A2 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
04050 Pileated Woodpecker Dryocopus pileatus 308.0
04060 Red-headed Woodpecker Melanerpes erythrocephalus 71.6
04090 Red-bellied Woodpecker Melanerpes carolinus 67.2
04120 Northern Flicker Colaptes auratus 135.0
04160 Chuck-will’s-widow Caprimulgus carolinensis 120.0
04170 Whip-poor-will Caprimulgus vociferus 55.3
04200 Common Nighthawk Chordeiles minor 61.5
04230 Chimney Swift Chaetura pelagica 23.6
04280 Ruby-throated Hummingbird Archilochus colubris 3.0
04440 Eastern Kingbird T yrannus tyrannus 43.6
04520 Great Crested Flycatcher 1 Myiarchus crinitus 33.5
04560 Eastern Phoebe Sayornis phoebe 19.8
04610 Eastern Wood-Pewee Contopus virens 14.1
04650 Acadian Flycatcher Empidonax virescens 12.9
04660 Willow Flycatcher Empidonax traillii 13.1
04740 Horned Lark Eremophila alpestris 31.9
04770 Blue Jay Cyanocitta cristata 86.8
04880 American Crow Corvus brachyrhynchos 458.0
04930 European Starling Sturnus vulgaris 84.7
04950 Brown-headed Cowbird Molothrus ater 49.0

144

Table A2 (cont’d).

 

 

145

AOU Code Common Name Scientific Name Mass (g)
04980 Red-winged Blackbird Agelaius phoeniceus 63.6
05010 Eastern Meadowlark Sturnella magna 102.0
05060 Orchard Oriole Icterus spurius 19.6
05070 Baltimore Oriole Icterus galbula 34.3
051 10 Common Grackle Quiscalus quiscula 127.0
05190 House Finch Carpodacus mexicanus 21.4
05290 American Goldfinch Carduelis tristis 13.2
05460 Grasshopper Sparrow Ammodramus savannarum 17.0
05600 Chipping Sparrow Spizella passerina 12.3
05630 Field Sparrow Spizella pusilla 12.5
05810 Song Sparrow Melospiza melodia 21.0
05870 Eastern Towhee Pipilo erythrophthalmus 41.7
05930 Northern Cardinal Cardinalis cardinalis 45.4
05950 Rose—breasted Grosbeak Pheucticus ludovicianus 45.6
05970 Blue Grosbeak Guiraca caerulea 29.3
05980 Indigo Bunting Passerina cyanea 14.9
06040 Dickcissel Spiza americana 29.3
06080 Scarlet Tanager Piranga olivacea 28.6
06100 Summer Tanager Piranga rubra 28.2
061 10 Purple Martin Progne subis 49.4

Table A2 (cont’d).

 

 

146

AOU Code Common Name Scientific Name Mass (g)
06130 Barn Swallow Hirundo rustica 16.2
06140 Tree Swallow T achycineta bicolor 20.1
06160 Bank Swallow Riparia riparia 14.6
06170 Northern Rough-winged Stelgidopteryx serripennis 15.9

Swallow
06190 Cedar Waxwing Bombycilla cedrorum 30.6
06220 Loggerhead Shrike Lanius ludovicianus 47.4
06240 Red-eyed Vireo Vireo olivaceus 16.7
06280 Yellow-throated Vireo Vireo flavifrons 18.0
06290 Blue—headed Vireo Vireo solitarius 16.6
06310 White-eyed Vireo Vireo griseus 11.4
06360 Black-and-white Warbler Mniotilta varia 11.0
063 70 Prothonotary Warbler Protonotaria citrea 15.0
06380 Swainson's Warbler Limnothlypis swainsonii 18.9
06390 Worm-eating Warbler Helmitheros vermivorus 13.0
06410 Blue-winged Warbler Vermivora pinus 8.4
06420 Golden-winged Warbler Vermivora chrysoptera 8.7
06480 Northern Parula Parula americana 8.6
06520 Yellow Warbler Dendroica petechia 9.8
06580 Cerulean Warbler Dendroica cerulea 9.5
06630 Yellow-throated Warbler Dendroica dominica 9.4

Table A2 (cont’d).

 

 

147

AOU Code Common Name Scientific Name Mass (g)
06670 Black-throated Green Warbler Dendroica virens 8.8
06710 Pine Warbler Dendroica pinus 11.9
06730 Prairie Warbler Dendroica discolor 8.0
06740 Ovenbird Seiurus aurocapillus 19.4
06760 Louisiana Waterthrush Seiurus motacilla 19.8
06770 Kentucky Warbler Oporornisformosus 14.3
06810 Common Yellowthroat Geothlypis trichas 10.1
06830 Yellow-breasted Chat Icteria virens 25.5
06840 Hooded Warbler Wilsonia citrina 10.8
06870 American Redstart Setophaga ruticilla 8.5
068 82 House Sparrow Passer domesticus 28.0
07030 Northern Mockingbird Mimus polyglottos 48.5
07040 Gray Catbird Dumetella carolinensis 36.9
07050 Brown Thrasher T oxostoma rufum 68.8
07180 Carolina Wren T hryothorus ludovicianus 18.7
07190 Bewick’s Wren T htyomanes bewickii 9.9
07210 House Wren T roglodytes aedon 10.9
07270 White-breasted Nuthatch Sitta carolinensis 21.1
07310 Tufted Titmouse Paras bicolor 21.6
07360 Carolina Chickadee Parus carolinensis 10.5

Table A2 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
07510 Blue-gray Gnatcatcher Polioptila caerulea 6.0
07550 Wood Thrush Hylocichla mustelina 47.8
07610 American Robin T urdus migratorius 77.3
07660 Eastern Bluebird Sialia sialis 31.6

 

148

Table A.3. Species list for section 133, the Coastal Plains and Flatwoods, Western Gulf
Section.

 

 

149

AOU Code Common Name Scientific Name Mass (g)
02890 Northern Bobwhite Colinus virginianus 178.0
03100 Wild Turkey Meleagris gallopavo 7400.0
03131 Rock Dove Columba livia 369.0
03160 Mourning Dove Zenaida macroura 123.0
03210 Inca Dove Columbina inca 47.5
03250 Turkey Vulture Cathartes aura 1467.0
03260 Black Vulture Coragyps atratus 2172.0
03270 Swallow-tailed Kite Elanaidesforficatus 442.0
03290 Mississippi Kite Ictinia mississippiensis 245.0
03330 Cooper's Hawk Accipiter cooperii 349.0
03370 Red-tailed Hawk Buteojamaicensis 1028.0
03390 Red-shouldered Hawk Buteo lineatus 475.0
03430 Broad-winged Hawk Buteo platypterus 420.0
03 600 American Kestrel F alco sparverius 11 1.0
03620 Crested Caracara Caracara plancus 1 1 17.0
03680 Barred Owl Strix varia 632.0
03730 Eastern Screech-Owl Otus asio 174.0
03 750 Great Horned Owl Bubo virginianus 1318.0
03850 Greater Roadrunner Geococcyx californianus 376.0
03870 Yellow-billed Cuckoo Coccyzus americanus 64.0

Table A3 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
03900 Belted Kingfisher Ceryle alcyon 148.0
03930 Hairy Woodpecker Picoides villosus 70.0
03940 Downy Woodpecker Picoides pubescens 27.0
03950 Red-cockaded Woodpecker Picoides borealis 43.6
04050 Pileated Woodpecker Dryocopus pileatus 308.0
04060 Red-headed Woodpecker Melanerpes erythrocephalus 71.6
04090 Red-bellied Woodpecker Melanerpes carolinus 67.2
04120 Northern Flicker Colaptes auratus 135.0
04160 Chuck-will’s-widow Caprimulgus carolinensis 120.0
04200 Common Nighthawk Chordeiles minor 61.5
04230 Chimney Swift Chaetura pelagica 23.6
04280 Ruby-throated Hummingbird Archilochus colubris 3.0
04430 ScissorItailed Flycatcher T yrannus forficatus 43.2
04440 Eastern Kingbird T yrannus tyrannus 43.6
04520 Great Crested Flycatcher Myiarchus crinitus 33.5
04560 Eastern Phoebe Sayornis phoebe 19.8
04610 Eastern Wood-Pewee Contopus virens 14.1
04650 Acadian Flycatcher Empidonax virescens 12.9
04770 Blue Jay C yanocitta cristata 86.8
04880 American Crow Corvus brachyrhynchos 458.0

150

Table A.3 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
04900 Fish Crow Corvus ossifragus 300.0
04930 European Starling Sturnus vulgaris 84.7
04950 Brown-headed Cowbird Molothrus ater 49.0
04980 Red-winged Blackbird Agelaius phoeniceus 63.6
05010 Eastern Meadowlark Sturnella magna 102.0
05060 Orchard Oriole Icterus spurius 19.6
05070 Baltimore Oriole Icterus galbula 34.3
051 10 Common Grackle Quiscalus quiscula 127.0
05130 Boat-tailed Grackle Quiscalus major 214.0
05190 House Finch Carpodacus mexicanus 21.4
05460 Grasshopper Sparrow Ammodramus savannarum 17.0
05520 Lark Sparrow Chondestes grammacus 29.0
05 600 Chipping Sparrow Spizella passerina 12.3
05630 Field Sparrow Spizella pusilla 12.5
05750 Bachman’s Sparrow Aimophila aestivalis 20.2
05870 Eastern Towhee Pipilo erythrophthalmus 41.7
05930 Northern Cardinal Cardinalis cardinalis 45.4
05970 Blue Grosbeak Guiraca caerulea 29.3
05980 Indigo Bunting Passerina cyanea 14.9
06010 Painted Bunting Passerina ciris 16.1

151

Table A.3 (cont’d).

 

 

152

AOU Code Common Name Scientific Name Mass (g)
06040 Dickcissel Spiza americana 29.3
06100 Summer Tanager Piranga rubra 28.2
061 10 Purple Martin Progne subis 49.4
06120 Cliff Swallow Hirundo pyrrhonota 21.6
06130 Barn Swallow Hirundo rustica 16.2
06170 Northern Rough-winged Stelgidopteryx serripennis 15.9

Swallow
06220 Loggerhead Shrike Lanius ludovicianus 47.4
06240 Red-eyed Vireo Vireo olivaceus 16.7
06270 Warbling Vireo Vireo gilvus 14.8
06280 Yellow-throated Vireo Vireo flavifrons 18.0
06310 White-eyed Vireo Vireo griseus 11.4
06360 Black-and-white Warbler Mniotilta varia 11.0
063 70 Prothonotary Warbler Protonotaria citrea 15.0
063 80 Swainson's Warbler Limnothlypis swainsonii 18.9
06390 Worm-eating Warbler Helmitheros vermivorus 13.0
06480 Northern Parula Parula americana 8.6
06630 Yellow-throated Warbler Dendroica dominica 9.4
06710 Pine Warbler Dendroica pinus 1 1.9
06730 Prairie Warbler Dendroica discolor 8.0
06760 Louisiana Waterthrush Seiurus motacilla 19.8

Table A.3 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
06770 Kentucky Warbler Oporornisformosus 14.3
06810 Common Yellowthroat Geothlypis trichas 10.1
06830 Yellow-breasted Chat Icteria virens 25.5
06840 Hooded Warbler Wilsonia citrina 10.8
06870 American Redstart Setophaga ruticilla 8.5
06882 House Sparrow Passer domesticus 28.0
07030 Northern Mockingbird Mimus polyglottos 48.5
07040 Gray Catbird Dumetella carolinensis 36.9
07050 Brown Thrasher T oxostoma rufum 68.8
07180 Carolina Wren T hryothorus ludovicianus 18.7
07270 White-breasted Nuthatch Sitta carolinensis 21.1
07290 Brown-headed Nuthatch Sitta pusilla 10.2
07310 Tufted Titmouse Parus bicolor 21.6
073 60 Carolina Chickadee Parus carolinensis 10.5
07510 Blue-gray Gnatcatcher Polioptila caerulea 6.0
07550 Wood Thrush Hylocichla mustelina 47.8
07610 American Robin T urdus migratorius 77.3
07660 Eastern Bluebird Sialia sialis 31.6

 

153

Table A4. Species list for section 153, the Central Loess Plains Section.

 

 

154

AOU Code Common Name Scientific Name Mass (g)
02881 Gray Partridge Perdix perdix 398.0
02890 Northern Bobwhite Colinus virginianus 178.0
03050 Greater Prairie-Chicken T ympanuchus cupido 999.0
03091 Ring-necked Pheasant Phasianus colchicus 1317.0
03100 Wild Turkey Meleagris gallopavo 7400.0
03131 Rock Dove Columba livia 369.0
03160 Mourning Dove Zenaida macroura 123.0
03250 Turkey Vulture Cathartes aura 1467.0
03310 Northern Harrier Circus cyaneus 358.0
03320 Sharp-shinned Hawk Accipiter striatus 103.0
03330 Cooper's Hawk Accipiter cooperii 349.0
03370 Red-tailed Hawk Buteojamaicensis 1028.0
03390 Red-shouldered Hawk Buteo lineatus 475.0
03420 Swainson's Hawk Buteo swainsoni 908.0
03430 Broad-winged Hawk Buteo platypterus 420.0
03600 American Kestrel F alco sparverius 111.0
03670 Short-cared Owl Asioflammeus 315.0
03680 Barred Owl Strix varia 632.0
03730 Eastern Screech-Owl Otus asio 174.0
03750 Great Horned Owl Bubo virginianus 1318.0

Table A4 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
03 870 Yellow-billed Cuckoo Coccyzus americanus 64.0
03 880 Black-billed Cuckoo Coccyzus erythropthalmus 51.1
03900 Belted Kingfisher Ceryle alcyon 148.0
03930 Hairy Woodpecker Picoides villosus 70.0
03940 Downy Woodpecker Picoides pubescens 27.0
04050 Pileated Woodpecker Dryocopus pileatus 308.0
04060 Red-headed Woodpecker Melanerpes erythrocephalus 71.6
04090 Red-bellied Woodpecker Melanerpes carolinus 67.2
04120 Northern Flicker Colaptes auratus 135.0
04170 Whip-poor-will Caprimulgus vociferus 55.3
04200 Common Nighthawk Chordeiles minor 61.5
04230 Chimney Swift Chaetura pelagica 23.6
04280 Ruby-throated Hummingbird Archilochus colubris 3.0
04430 Scissor-tailed Flycatcher T yrannus forficatus 43.2
04440 Eastern Kingbird T yrannus tyrannus 43.6
04470 Western Kingbird T yrannus verticalis 39.6
04520 Great Crested Flycatcher Myiarchus crinitus 33.5
045 60 Eastern Phoebe Sayornis phoebe 19.8
04570 Say's Phoebe Sayornis saya 21.2
04610 Eastern Wood-Pewee Contopus virens 14.1

155

Table A.4 (cont’d).

 

 

156

AOU Code Common Name Scientific Name Mass (g)
04650 Acadian Flycatcher Empidonax virescens 12.9
04660 Willow Flycatcher Empidonax traillii 13.1
04661 Alder Flycatcher Empidonax alnorum 12.7
04670 Least Flycatcher Empidonax minimus 10.3
04740 Horned Lark Eremophila alpestris 31.9
04750 Black-billed Magpie Pica pica 189.0
04770 Blue Jay C yanocitta cristata 86.8
04880 American Crow Corvus brachyrhynchos 458.0
04930 European Starling Sturnus vulgaris 84.7
04940 Bobolink Dolichonyx oryzivorus 47.0
04950 Brown-headed Cowbird Molothrus ater 49.0
04970 Yellow-headed Blackbird Xanthocephalus xanthocephalus 79.7
04980 Red-winged Blackbird A gelaius phoeniceus 63.6
05010 Eastern Meadowlark Sturnella magna 102.0
05011 Western Meadowlark Sturnella neglecta 112.0
05060 Orchard Oriole Icterus spurius 19.6
05070 Baltimore Oriole Icterus galbula 34.3
051 10 Common Grackle Quiscalus quiscula 127.0
05190 House Finch Carpodacus mexicanus 21.4
05210 Red Crossbill Loxia curvirostra 36.5

Table A.4 (cont’d).

 

 

157

AOU Code Common Name Scientific Name Mass (g)
05290 American Goldfinch Carduelis tristis 13.2
05400 Vesper Sparrow Pooecetes gramineus 26.5
05420 Savannah Sparrow Passerculus sandwichensis 20.6
05460 Grasshopper Sparrow Ammodramus savannarum 17.0
05470 Henslow's Sparrow Ammodramus henslowii 13.1
05520 Lark Sparrow Chondestes grammacus 29.0
05600 Chipping Sparrow Spizella passerina 12.3
05610 Clay-colored Sparrow Spizella pallida 12.0
05630 Field Sparrow Spizella pusilla 12.5
05810 Song Sparrow Melospiza melodia 21.0
05840 Swamp Sparrow Melospiza georgiana 17.0
05 870 Eastern Towhee Pipilo erythrophthalmus 41.7
05930 Northern Cardinal Cardinalis cardinalis 45.4
05950 Rose-breasted Grosbeak Pheucticus ludovicianus 45.6
05970 Blue Grosbeak Guiraca caerulea 29.3
05980 Indigo Bunting Passerina cyanea 14.9
06040 Dickcissel Spiza americana 29.3
06050 Lark Bunting Calamospiza melanocorys 37.6
06080 Scarlet Tanager Piranga olivacea 28.6
06100 Summer Tanager Piranga rubra 28.2

Table A.4 (cont’d).

 

 

158

AOU Code Common Name Scientific Name Mass (g)
061 10 Purple Martin Progne subis 49.4
06120 Cliff Swallow Hirundo pyrrhonota 21.6
06130 Barn Swallow Hirundo rustica 16.2
06140 Tree Swallow T achycineta bicolor 20.1
06160 Bank Swallow Riparia riparia 14.6
06170 Northern Rough-winged Stelgidopteryx serripennis 15.9

Swallow
06190 Cedar Waxwing Bombycilla cedrorum 30.6
06220 Loggerhead Shrike Lanius ludovicianus 47.4
06240 Red-eyed Vireo Vireo olivaceus 16.7
06270 Warbling Vireo Vireo gilvus 14.8
06280 Yellow-throated Vireo Vireo flavifrons 18.0
06310 White-eyed Vireo Vireo griseus 11.4
06330 Bell's Vireo Vireo bellii 8.5
06360 Black-and—white Warbler Mniotilta varia 11.0
063 70 Prothonotary Warbler Protonotaria citrea 14.3
06410 Blue-winged Warbler Vermivora pinus 8.4
06480 Northern Parula Parula americana 8.6
06520 Yellow Warbler Dendroica petechia 9.8
06580 Cerulean Warbler Dendroica cerulea 9.5
06590 Chestnut-sided Warbler Dendroica pensylvanica 9.8

Table A.4 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
06630 Yellow-throated Warbler Dendroica dominica 9.4
06730 Prairie Warbler Dendroica discolor 8.0
06740 Ovenbird Seiurus aurocapillus 19.4
06760 Louisiana Waterthrush Seiurus motacilla 19.8
06770 Kentucky Warbler Oporornisformosus 14.3
06810 Common Yellowthroat Geothlypis trichas 10.1
06830 Yellow-breasted Chat Icteria virens 25.5
06840 Hooded Warbler Wilsonia citrina 10.8
06870 American Redstart Setophaga ruticilla 8.5
06882 House Sparrow Passer domesticus 28.0
06883 Eurasian Tree Sparrow Passer montanus 22.0
07030 Northern Mockingbird Mimus polyglottos 48.5
07040 Gray Catbird Dumetella carolinensis 36.9
07050 Brown Thrasher T oxostoma rufum 68.8
07180 Carolina Wren T hryothorus ludovicianus 18.7
07190 Bewick's Wren T hryomanes bewickii 9.9
07210 House Wren T roglodytes aedon 10.9
07240 Sedge Wren Cistothorus platensis 9.0
07250 Marsh Wren Cistothorus palustris 11.9
07260 Brown Creeper Certhia americana 8.4

Table A. 4 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
07270 White-breasted Nuthatch Sitta carolinensis 21.1
07310 Tufted Titmouse Parus bicolor 21.6
07350 Black-capped Chickadee Parus atricapillus 10.8
07360 Carolina Chickadee Parus carolinensis 10.5
07510 Blue-gray Gnatcatcher Polioptila caerulea 6.0
075 50 Wood Thrush Hylocichla mustelina 47.8
07610 American Robin T urdus migratorius 77.3
07660 Eastern Bluebird Sialia sialis 31.6

 

160

Table A5 Species list for section 170, the Mid Coastal Plains, Western Section.

 

 

161

AOU Code Common Name Scientific Name Mass (g)
02890 Northern Bobwhite Colinus virginianus 178.0
03100 Wild Turkey Meleagris gallopavo 7400.0
03131 Rock Dove Columba livia 369.0
03160 Mourning Dove Zenaida macroura 123.0
03210 Inca Dove Columbina inca 47.5
03250 Turkey Vulture Cathartes aura 1467.0
03260 Black Vulture Coragyps atratus 2172.0
03290 Mississippi Kite Ictinia mississippiensis 245.0
03320 Sharp-shinned Hawk Accipiter striatus 103.0
03330 Cooper's Hawk Accipiter cooperii 349.0
03370 Red-tailed Hawk Buteojamaicensis 1028.0
03390 Red-shouldered Hawk Buteo lineatus 475.0
03420 Swainson's Hawk Buteo swainsoni 908.0
03430 Broad-winged Hawk Buteo plaiypterus 420.0
03600 American Kestrel F alco sparverius 111.0
03680 Barred Owl Strix varia 632.0
03730 Eastern Screech-Owl Otus asio 174.0
03 750 Great Horned Owl Bubo virginianus 1318.0
03 850 Greater Roadrunner Geococcyx californianus 376.0
03870 Yellow-billed Cuckoo Coccyzus americanus 64.0

Table A.5 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
03900 Belted Kingfisher Ceryle alcyon 148.0
03930 Hairy Woodpecker Picoides villosus 70.0
03940 Downy Woodpecker Picoides pubescens 27.0
03950 Red-cockaded Woodpecker Picoides borealis 43.6
04050 Pileated Woodpecker Dryocopus pileatus 308.0
04060 Red-headed Woodpecker Melanerpes erythrocephalus 71.6
04090 Red-bellied Woodpecker Melanerpes carolinus 67.2
04120 Northern Flicker Colaptes auratus 135.0
04160 Chuck-will's-widow Caprimulgus carolinensis 120.0
04170 Whip-poor-will Caprimulgus vociferus 55.3
04200 Common Nighthawk Chordeiles minor 61.5
04230 Chimney Swift Chaetura pelagica 23.6
04280 Ruby-throated Hummingbird Archilochus colubris 3.0
04430 Scissor-tailed Flycatcher Tyrannusforficatus 43.2
04440 Eastern Kingbird T yrannus tyrannus 43.6
04470 Western Kingbird T yrannus verticalis 39.6
04520 Great Crested Flycatcher Myiarchus crinitus 33.5
04560 Eastern Phoebe Sayornis phoebe 19.8
04610 Eastern Wood-Pewee Contopus virens 14.1
04650 Acadian Flycatcher Empidonax virescens 12.9

162

Table A.5 (cont’d).

 

 

163

AOU Code Common Name Scientific Name Mass (g)
04740 Horned Lark Eremophila alpestris 31.9
04770 Blue Jay Cyanocitta cristata 86.8
04880 American Crow Corvus brachyrhynchos 458.0
04900 Fish Crow Corvus ossifragus 300.0
04930 European Starling Sturnus vulgaris 84.7
04950 Brown-headed Cowbird Molothrus ater 49.0
04980 Red-winged Blackbird Agelaius phoeniceus 63.6
05010 Eastern Meadowlark Sturnella magna 102.0
05060 Orchard Oriole Icterus spurius 19.6
05070 Baltimore Oriole Icterus galbula 34.3
05110 Common Grackle Quiscalus quiscula 127.0
05120 Great-tailed Grackle Quiscalus mexicanus 191.0
05190 House Finch Carpodacus mexicanus 21.4
05290 American Goldfinch Carduelis tristis 13.2
05460 Grasshopper Sparrow Ammodramus savannarum 17.0
05520 Lark Sparrow Chondestes grammacus 29.0
05600 Chipping Sparrow Spizella passerina 12.3
05 630 Field Sparrow Spizella pusilla 12.5
05 750 Bachman's Sparrow Aimophila aestivalis 20.2
05 870 Eastern Towhee Pipilo erythrophthalmus 41.7

Table A.5 (cont’d).

 

 

164

AOU Code Common Name Scientific Name Mass (g)
05930 Northern Cardinal Cardinalis cardinalis 45.4
05970 Blue Grosbeak Guiraca caerulea 29.3
05980 Indigo Bunting Passerina cyanea 14.9
06010 Painted Bunting Passerina ciris 16.1
06040 Dickcissel Spiza americana 29.3
06080 Scarlet Tanager Piranga olivacea 28.6
06100 Summer Tanager Piranga rubra 28.2
06110 Purple Martin Progne subis 49.4
06120 Cliff Swallow Hirundo pyrrhonota 21.6
06130 Barn Swallow Hirundo rustica 16.2
06140 Tree Swallow T achycineta bicolor 20.1
06170 Northern Rough-winged Stelgidopteryx serripennis 15.9

Swallow
06190 Cedar Waxwing Bombycilla cedrorum 30.6
06220 Loggerhead Shrike Lanius ludovicianus 47.4
06240 Red-eyed Vireo Vireo olivaceus 16.7
06270 Warbling Vireo Vireo gilvus 14.8
06280 Yellow-throated Vireo Vireo flavifrons 18.0
06310 White-eyed Vireo Vireo griseus 11.4
06330 Bell's Vireo Vireo bellii 8.5
06360 Black-and-white Warbler Mniotilta varia 11.0

Table A.5 (cont’d).

 

 

165

AOU Code Common Name Scientific Name Mass (g)
063 70 Prothonotary Warbler Protonotaria citrea 15.0
063 80 Swainson's Warbler Limnothlypis swainsonii 18.9
063 90 Worm-eating Warbler Helmitheros vermivorus 13.0
06480 Northern Parula Parula americana 8.6
065 80 Cerulean Warbler Dendroica cerulea 9.5
06630 Yellow-throated Warbler Dendroica dominica 9.4
06710 Pine Warbler Dendroica pinus 11.9
06730 Prairie Warbler Dendroica discolor 8.0
06740 Ovenbird Seiurus aurocapillus 19.4
06760 Louisiana Waterthrush Seiurus motacilla 19.8
06770 Kentucky Warbler Oporornisformosus 14.3
06810 Common Yellowthroat Geothlypis trichas 10.1
06830 Yellow-breasted Chat Icteria virens 25.5
06840 Hooded Warbler Wilsonia citrina 10.8
06870 American Redstart ' Setophaga ruticilla 8.5
06882 House Sparrow Passer domesticus 28.0
07030 Northern Mockingbird Mimus polyglottos 48.5
07040 Gray Catbird Dumetella carolinensis 36.9
07050 Brown Thrasher T oxostoma rufum 68.8
07180 Carolina Wren T hryothorus ludovicianus 18.7

Table A.5 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
07190 Bewick's Wren T hryomanes bewickii 9.9
07210 House Wren T roglodytes aedon 10.9
07270 White—breasted Nuthatch Sitta carolinensis 21.1
07290 Brown—headed Nuthatch Sitta pusilla 10.2
07310 Tufted Titmouse Parus bicolor 21.6
07360 Carolina Chickadee Parus carolinensis 10.5
07510 Blue-gray Gnatcatcher Polioptila caerulea 6.0
07550 Wood Thrush Hylocichla mustelina 47.8
07610 American Robin T urdus migratorius 77.3
07660 Eastern Bluebird Sialia sialis 31.6

 

166

Table A6 Species list for section 172, the Southern Cumberland Plateau Section.

 

 

167

AOU Code Common Name Scientific Name Mass (g)
02890 Northern Bobwhite Colinus virginianus 178.0
03100 Wild Turkey Meleagris gallopavo 7400.0
03131 Rock Dove Columba livia 369.0
03160 Mourning Dove Zenaida macroura 123.0
03200 Common Ground-Dove Columbina passerina 30.1
03250 Turkey Vulture Cathartes aura 1467.0
03260 Black Vulture Coragyps atratus 2172.0
03290 Mississippi Kite Ictinia mississippiensis 245.0
03320 Sharp-shinned Hawk Accipiter striatus 103.0
03330 Cooper's Hawk Accipiter cooperii 349.0
03370 Red-tailed Hawk Buteojamaicensis 1028.0
03390 Red-shouldered Hawk Buteo lineatus 475.0
03430 Broad-winged Hawk Buteo platypterus 420.0
03520 Bald Eagle Haliaeetus leucocephalus 4130.0
03 600 American Kestrel F alco sparverius 1 1 1.0
03640 Osprey Pandion haliaetus 1403.0
03680 Barred Owl Strix varia 632.0
03730 Eastern Screech-Owl Otus asio 174.0
03 750 Great Horned Owl Bubo virginianus 1318.0
03870 Yellow-billed Cuckoo Coccyzus americanus 64.0

Table A6 (cont’d).

 

 

168

AOU Code Common Name Scientific Name Mass (g)
03880 Black-billed Cuckoo Coccyzus erythropthalmus 51.1
03900 Belted Kingfisher Ceryle alcyon 148.0
03930 Hairy Woodpecker Picoides villosus 70.0
03940 Downy Woodpecker Picoides pubescens 27.0
03950 Red-cockaded Woodpecker Picoides borealis 43.6
04050 Pileated Woodpecker Dryocopus pileatus 308.0
04060 Red-headed Woodpecker Melanerpes erythrocephalus 71.6
04090 Red-bellied Woodpecker Melanerpes carolinus 67.2
04120 Northern Flicker Colaptes auratus 135.0
04160 Chuck-will's-widow Caprimulgus carolinensis 120.0
04170 Whip-poor-will Caprimulgus vociferus 55.3
04200 Common Nighthawk Chordeiles minor 61.5
04230 Chimney Swift Chaetura pelagica 23.6
04280 Ruby-throated Hummingbird Archilochus colubris 3.0
04440 Eastern Kingbird T yrannus tyrannus 43.6
04520 Great Crested Flycatcher Myiarchus crinitus 33.5
04560 Eastern Phoebe Sayornis phoebe 19.8
04610 Eastern Wood-Pewee Contopus virens 14.1
04650 Acadian Flycatcher Empidonax virescens 12.9
04740 Horned Lark Eremophila alpestris 31.9

Table A6 (cont’d).

 

 

169

AOU Code Common Name Scientific Name Mass (g)
04770 Blue Jay Cyanocitta cristata 86.8
04880 American Crow Corvus brachyrhynchos 458.0
04900 Fish Crow Corvus ossifragus 300.0
04930 European Starling Sturnus vulgaris 84.7
04950 Brown-headed Cowbird Molothrus ater 49.0
04980 Red-winged Blackbird Agelaius phoeniceus 63.6
05010 Eastern Meadowlark Sturnella magna 102.0
05060 Orchard Oriole Icterus spurius 19.6
05070 Baltimore Oriole Icterus galbula 34.3
051 10 Common Grackle Quiscalus quiscula 127.0
05190 House Finch Carpodacus mexicanus 21.4
05290 American Goldfinch Carduelis tristis 13.2
05460 Grasshopper Sparrow Ammodramus savannarum 17.0
05 600 Chipping Sparrow Spizella passerina 12.3
05 630 Field Sparrow Spizella pusilla 12.5
05750 Bachman's Sparrow Aimophila aestivalis 20.2
05810 Song Sparrow Melospiza melodia 21.0
05870 Eastern Towhee Pipilo erythrophthalmus 41.7
05930 Northern Cardinal Cardinalis cardinalis 45.4
05970 Blue Grosbeak Guiraca caerulea 29.3

Table A6 (cont’d).

 

 

170

AOU Code Common Name Scientific Name Mass (g)
05980 Indigo Bunting Passerina cyanea 14.9
06040 Dickcissel Spiza americana 29.3
06080 Scarlet Tanager Piranga olivacea 28.6
06100 Summer Tanager Piranga rubra 28.2
061 10 Purple Martin Progne subis 49.4
06120 Cliff Swallow Hirundo pyrrhonota 21.6
06130 Barn Swallow Hirundo rustica 16.2
06140 Tree Swallow T achycineta bicolor 20.1
06160 Bank Swallow Riparia riparia 14.6
06170 Northern Rough-winged Stelgidopteryx serripennis 15.9

Swallow
06190 Cedar Waxwing Bombycilla cedrorum 30.6
06220 Loggerhead Shrike Lanius ludovicianus 47.4
06240 Red-eyed Vireo Vireo olivaceus 16.7
06280 Yellow-throated Vireo Vireo flavifrons 18.0
06290 Blue-headed Vireo Vireo solitarius 16.6
06310 White-eyed Vireo Vireo griseus 11.4
06360 Black-and-white Warbler Mniotilta varia 1 1.0
063 70 Prothonotary Warbler Protonotaria citrea 15.0
063 80 Swainson's Warbler Linmothlypis swainsonii 18.9
06390 Worm-eating Warbler Helmitheros vermivorus 13.0

Table A6 (cont’d).

 

 

171

AOU Code Common Name Scientific Name Mass (g)
06410 Blue-winged Warbler Vermivora pinus 8.4
06480 Northern Parula Parula americana 8.6
06520 Yellow Warbler Dendroica petechia 9.8
065 80 Cerulean Warbler Dendroica cerulea 9.5
06630 Yellow-throated Warbler Dendroica dominica 9.4
06670 Black-throated Green Warbler Dendroica virens 8.8
06710 Pine Warbler Dendroica pinus 11.9
06730 Prairie Warbler Dendroica discolor 8.0
06740 Ovenbird Seiurus aurocapillus 19.4
06760 Louisiana Waterthrush Seiurus motacilla 19.8
06770 Kentucky Warbler Oporornisformosus 14.3
06810 Common Yellowthroat Geothlypis trichas 10.1
06830 Yellow-breasted Chat Icteria virens 25.5
06840 Hooded Warbler Wilsonia citrina 10.8
06870 American Redstart Setophaga ruticilla 8.5
06882 House Sparrow Passer domesticus 28.0
07030 Northern Mockingbird Mimus polyglottos 48.5
07040 Gray Catbird Dumetella carolinensis 36.9
07050 Brown Thrasher T oxostoma rufum 68.8
07180 Carolina Wren T hryothorus ludovicianus 18.7

Table A6 (cont’d).

 

 

AOU Code Common Name Scientific Name Mass (g)
07210 House Wren T roglodytes aedon 10.9
07270 White-breasted Nuthatch Sitta carolinensis 21.1
07290 Brown-headed Nuthatch Sitta pusilla 10.2
07310 Tufted Titmouse Parus bicolor 21.6
07360 Carolina Chickadee Parus carolinensis 10.5
07510 Blue-gray Gnatcatcher Polioptila caerulea 6.0
07550 Wood Thrush Hylocichla mustelina 47.8
07610 American Robin T urdus migratorius 77.3
07660 Eastern Bluebird Sialia sialis 31.6
22860 Eurasian Collared-Dove Streptopelia decaocto 152.0

 

172

APPENDIX B: TAXONOMIC CHAN CBS

173

Table B. 1. Changes made to species names from those originally reported to the BBS.

 

F orrner Name

Current Name

 

Yellow-shafted Flicker
Red-shafted Flicker
unid. Flicker subsp.

Black-crested
Titmouse

Slate-colored Junco
unid. Junco subsp.
Solitary Vireo
Crested Caracara

Rufous-sided towhee

Ringed turtle-dove

Caracara plancus

Steptopelia
risoria

Northern Flicker
Northern Flicker
Northern Flicker

Tufted Titmouse

Dark-eyed J unco
Dark-eyed Junco
Blue-headed Vireo
Crested Caracara
Eastern Towhee

Eurasian collared-
dove

174

Colaptes auratus
Colaptes auratus
Colaptes auratus

Parus bicolor

Junco hyemalis
Junco hyemalis
Vireo solitarius
Caracara cheriway
Pipilo
erythrophthalmus

Streptopelia
decaocta

APPENDIX C: DISCONTINUITIES IN BODY SIZE DISTRIBUTIONS

175

Table C.1. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, S1 = Siemann & Brown Index, RH = robust Holling Index, RS = robust
Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y = discontinuity
identified, N = discontinuity not identified) for the Section 47 dataset.

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
3 - 6 N Y N N N N
3 — 8 Y N N N N N
8 — 8.5 N N Y N N N
8.6 — 9.4 Y N N N N N
9.5 — 10.1 N N Y N N N
11 — 11.4 N N N N Y N
11 — 11.9 N N Y N N N
13.2 — 14.1 N N N Y N N
13.2 — 14.3 N N Y N N N
14.9 — 15.9 N N N Y N N
14.9 — 16.2 N N Y N N N
17 — 18 N N N N Y N
17 — 18.7 N N Y N N N
21.6 — 23.6 N Y N Y N Y
21.6 — 25.5 Y N Y N N N
23.6 — 25.5 N N N N Y N
34.3 — 36.9 N N N Y Y Y
34.3 - 39. Y N Y N N N
51.1 — 55.3 N N N N Y N

176

Table C.1 (cont’d).

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
51.1 - 61.5 Y N Y N N N
55.3 — 61.5 N Y N Y N Y
71.6 — 77.3 N Y N Y Y Y
79.7 — 102 Y N Y N N N
86.8 — 102 N Y N Y Y Y

112 — 123 N Y N Y N N

135 - 174 Y N Y N N N

148 - 174 N Y N Y Y Y

174 — 308 Y N Y N N N

178 — 308 N Y N Y Y Y
349 — 458 Y N Y N N N

358 — 458 N Y N Y Y Y
475 — 999 Y N Y N N N

632 — 999 N Y N Y Y Y
1028 — 1318 N Y N Y Y Y
1067 — 2172 N N N N Y N
1467 — 7400 Y N N N N N
2172 — 7400 N Y N N Y Y

 

177

Table C.2. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, S1 = Siemann & Brown Index, RH = robust Holling Index, R8 = robust
Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y = discontinuity
identified, N = discontinuity not identified) for the Section 59 dataset.

 

 

Discontinuity H1 SI RH RS AG Consensus
Location (g) Discontinuity
3 — 6 N Y N N N N
3 — 8 Y N N N N N
8.7 — 9.4 N N Y N N N
8.8 — 9.4 N N N Y N N
10.1 — 10.8 N N Y N N N
11 — 11.4 N N N N Y N
11 — 11.9 N N Y N N N
13.2 — 14.1 N N N Y N N
13.2 — 14.3 N N Y N N N
15 — 16.2 N N Y N N N
17 — 18 N N N N Y N
17 — 18.7 N N Y N N N
21.6 — 23.6 N Y N Y N Y
21.6 — 25.5 Y N Y N N N
23.6 — 25.5 N N N N Y N
34.3 — 41.7 Y N Y N N N
36.9 — 41.7 N Y N Y Y Y
49.4 - 55.3 N Y N Y N Y
49.4 — 61.5 Y N Y N N N

178

Table C.2 (cont’d).

 

Discontinuity

Location (g)

H1

AG

Consensus
Discontinuity

 

55.3 -61.5

71.6—- 77.3

77.3 — 102

86.8 — 102

103— 120

Ill—120

127— 135

135—174

148— 174

174—308

178 —308

349 — 420

458—621

475 -621

621 —1028

632 — 1028

1467—2172

1467 — 7400

2172 — 7400

Z < Z Z -< Z *< Z Z -< Z -< Z Z Z Z ..< Z Z

..< Z Z -< Z < 2 < < Z -< Z Z Z Z -< Z Z Z

Z Z Z Z < 2 <1 2 Z ..< Z < Z Z '-< Z < Z Z

Z Z Z -< Z -< Z < '-< Z '-< Z Z -< Z ..< Z Z Z

~< Z -< < Z < Z '-< ..< 2 <1 Z < < Z < Z -< <

N

..< Z Z < Z -< Z < *< Z .< Z Z ..< Z -< Z Z

 

179

Table C.3. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, S1 = Siemann & Brown Index, RH = robust Holling Index, RS = robust
Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y = discontinuity
identified, N = discontinuity not identified) for the Section 133 dataset.

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
3 — 8 Y Y N N N N
8.6 — 9.4 N Y N Y N Y
8.6 — 10.1 Y N Y N N N
11 — 11.9 N N Y N N N
12.9 - 14.1 N N Y N N N
13 — 14.1 N N N Y Y N
17 -— 18.7 N N Y N N N
21.6 — 23.6 N Y N Y Y Y
21.6 — 25.5 Y N Y N N N
29.3 — 31.6 N N N Y N N
29.3 — 33.5 N N Y N N N
34.3 — 41.7 Y N Y N N N
36.9 -— 41.7 N Y N Y Y Y
49 - 61.5 Y N Y N N N
49.4 — 61.5 N Y N Y Y Y
71.6 — 77.3 N N N N Y N
77.3 — 102 Y N Y N N N
86.8 - 102 N Y N Y Y Y
111 — 120 N N N N Y N

180

Table C.3 (cont’d).

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
127 —l35 N N N N Y N
135 — 174 Y N Y N N N
148 — 174 N Y N Y Y Y
178 - 214 N Y N Y Y Y
214 — 300 Y N Y N N N
245 — 300 N Y N Y Y Y
308 — 349 N Y N Y Y Y
376 -— 420 N Y N Y Y Y
475 — 632 N N N N Y N
475 — 1028 Y N Y N N N
632 — 1028 N Y N Y Y Y
11 17 — 1318 N Y N Y Y Y
1467 — 2172 N N N N Y N
1467 — 7400 Y N N N N N
2172 — 7400 N Y N N Y Y

 

181

Table C.4. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, S1 = Siemann & Brown Index, RH = robust Holling Index, RS = robust
Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y = discontinuity
identified, N = discontinuity not identified) for the Section 153 dataset.

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
3 — 6 N Y N N N N
3 — 8 Y N N N N N
8 — 8.4 N N N Y N N
8.6 — 9.4 Y N Y N N N
11 - 11.9 N N Y N N N
13.2 — 14.1 N Y N Y N Y
13.2 - 14.3 N N Y N N N
14.9 — 15.9 N Y N Y N Y
14.9 — 16.2 N N Y N N N
17 — 18 N N N Y Y Y
17 — 18.7 Y N Y N N N
21.6 — 23.6 N Y N Y N Y
21.6 — 25.5 Y N Y N N N
23.6 — 25.5 N N N N Y N
29.3 — 31.6 N N Y N N N
31.9 — 33.5 N N N Y N N
33.5- 36.5 N N Y N N N
34.3 — 36.5 N Y N Y Y Y
37.6 — 39.6 N N N Y N N

182

Table C.4 (cont’d).

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity

 

37.6—41.7

55.1 - 55.3

51.1—61.5

55.3—61.5

71.6—77.3

79.7 — 102

86.8 — 102

103—111

112—123

112—127

135 — 148

135—174

148— 174

178 —308

189—308

315—349

358 —420

358 —458

475 — 632

< Z -< Z Z Z < Z -< Z Z Z Z Z < Z 2 < Z Z
Z Z 2 <1 < -< Z < Z Z Z < '-< < Z -< < Z Z Z
-< Z < Z Z Z ..< Z < Z -< Z Z Z '< Z Z *< Z <
Z Z Z -< <1 <2 Z < Z Z Z < < "<2 2 "<1 '-< Z Z Z
Z < Z -< Z -< Z '-< 2 <1 Z < Z '< Z < Z Z '-< Z
Z Z Z -< Z < Z -< Z Z Z -< 2 .<1 2 '-< < Z Z Z

475 - 908

183

Table C.4 (cont’d).

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
632 - 908 N Y N Y Y Y

999 — 1318 Y N N N N N

1028 — 1318 N Y N N Y Y

1318 — 7400 Y N N N N N

1467 — 7400 N Y N N Y Y

 

184

Table C.5. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, SI = Siemann & Brown Index, RH = robust Holling Index, RS = robust
Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y = discontinuity
identified, N = discontinuity not identified) for the Section 170 dataset.

 

 

Discontinuity HI SI RH RS AG Consensus
Location (g) Discontinuity
3 —— 6 N Y N N N N
3 — 8 Y N N N N N
8 — 8.5 N N N Y N N
8 — 8.6 N N Y N N N
8.5 — 9.4 Y N Y N N N
8.6 — 9.4 N Y N Y N Y
9.5 — 10.1 N N Y N N N
11 — 11.9 N N Y N N N
11.9 — 12.3 N N N N Y N
13 — 14.1 N N Y N N N
13.2 — 14.1 N N N Y N N
15 — 15.9 N N N Y N N
15 — 16.1 N N Y N N N
17 — 18.7 N N Y N N N
21.6 — 23.6 N Y N Y N Y
21.6 — 25.5 Y N Y N N N
23.6 — 25.5 N N N N Y N
34.3 — 36.9 N Y N Y N Y
34.3 — 39.6 Y N Y N N N

185

Table C.5 (cont’d).

 

Discontinuity HI S1 RH RS AG Consensus
Location (g) Discontinuity

 

36.9 — 39.6

49.4 — 55.3

49.4 — 61.5

55.3 —61.5

71.6— 77.3

77.3 — 102

86.8 — 102

103—120

111—120

127— 135

135—174

148— 174

191—245

191—300

308 — 349

308 - 376

376 — 420

376 — 458

475 — 632

-< Z Z Z Z Z -< Z Z '-< Z Z Z Z '-< Z Z < Z Z
Z Z Z < Z < Z ..< '-< Z Z '-< Z -< Z Z Z Z -< 2
<1 Z -< Z -< Z -< Z Z -< Z Z ..< Z -< Z Z I-< Z Z
Z Z Z < Z -< Z < < Z Z -< Z -< Z Z Z 2 <1 2
Z Z Z < Z -< Z < -< Z Z < Z < Z Z Z 2 <1 2

Z r<1 Z '-< Z ..< Z '-< -< Z < ..< Z -< Z -< r<3 Z Z -<

475 — 908

186

Table C.5 (cont’d).

 

 

Discontinuity HI SI RH RS AG Consensus
Location (8) Discontinuity
632 — 908 N Y N Y Y Y

1028 — 1318 N Y N Y Y Y

1467 — 2172 N N N N Y N

1467 — 7400 Y N N N N N

2172 -— 7400 N Y N N Y Y

 

187

Table C.6. Discontinuities in body size distribution identified by five methods (HI =
Holling Index, S1 = Siemann & Brown Index, RH = robust Holling Index, RS = robust
Siemann & Brown Index, AG = agglomerative nesting cluster analysis, Y = discontinuity
identified, N = discontinuity not identified) for the Section 172 dataset.

 

 

Discontinuity H1 SI RH RS AG Consensus
Location (g) Discontinuity
3 — 6 N Y N N N N
3 — 8 Y N N N N N
8 — 8.5 N N Y N N N
8.6 — 9.4 Y N Y N N N
8.8 — 9.4 N N N Y N Y
9.5 — 10.1 N N Y N N N
11 — 11.9 N N Y N N N
13 — 14.1 N N Y N N N
13.2 — 14.1 N N N Y Y Y
15 — 16.2 N N Y N N N
17 — 18.7 N N Y N N N
21.6 —- 23.6 N Y N Y N Y
21.6 — 25.5 Y N Y N N N
23.6 — 25.5 N N N N Y N
34.3 — 41.7 Y N Y N N N
36.9 — 41.7 N Y N Y Y Y
51.1 — 61.5 Y N Y N N N
55.3 — 61.5 N Y N Y Y Y
71.6 —- 77.3 N N N N Y N

188

Table C.6 (cont’d).

 

 

Discontinuity H1 SI RH RS AG Consensus
Location (g) Discontinuity
77.3 — 102 Y N Y N N N
86.8 — 102 N Y N Y Y Y
103 — 120 N N Y N N N
111 — 120 N N N Y Y Y
127 — 135 N N N N Y N
135 — 174 Y N Y N N N
148 — 174 N Y N Y Y Y
178 — 245 N Y N Y Y Y
178 — 300 Y N Y N N N
245 — 300 N N N N Y N
308 — 420 Y N Y N N N
349 - 420 N Y N Y Y Y
475 - 632 N N N N Y N
475 - 1028 Y N Y N N N
632 — 1028 N Y N Y Y Y
1028 — 1318 N N N N Y N
1467 — 2172 N N N N Y N
1467 — 4130 Y N N N N N
2172 — 4130 N Y N N Y Y

 

189

APPENDIX D: ECOREGIONS OF THE CONTINENTAL UNITED STATES

190

Table D1 Provinces (in bold) and their component sections of the contiguous United
States delineated by the National Ecological Unit Hierarchy (Cleland et al. 1997) with

corresponding numeric codes and areas (kmz).

 

 

 

 

 

 

Name Code Area (kmz)
Adirondack-New England Mixed Forest-Coniferous 1 1720.00
F orest-Alpine Meadow Province
Adirondack Highlands Section 1 270.00
Catskill Mountains Section 2 640.00
Green, Taconic, Berkshire Mountains Section 3 190.00
New England Piedmont Section 4 160.00
White Mountains Section 5 460.00
American Semi-Desert and Desert Province 5 350.00
Mojave Desert Section 12 120.00
Sonoran Colorado Desert Section 13 130.00
Sonoran Mojave Desert Section 14 100.00
Arizona-New Mexico Mountains Semi-Desert-Open 7 1300.00
Woodland-Coniferous F orest-Alpine Meadow
Province
Sacramento-Monzano Mountain Section 16 390.00
White Mountain-San Francisco Peaks Section 17 910.00
Black Hills Coniferous Forest Province 9 950.00
Black Hills Section 21 950.00
California Coastal Chaparral Forest and Shrub 11 360.88
Province
Central California Coast Section 24 240.1 1
Southern California Coast Section 25 120.77

191

Table D.1 (cont’d).

 

 

 

 

 

 

 

Name Code Area (kmz)
California Coastal Range Open Woodland-Shrub- 12 650.00
Coniferous F orest-Meadow Province
Central California Coast Ranges Section 26 320.00
Southern California Mountains and Valleys Section 27 330.00
California Coastal Steppe-Mixed F orest-Redwood 13 120.07
Forest Province
Northern California Coast Section 28 120.07
California Dry Steppe Province 14 500.15
Great Valley Section 29 500.15
Cascade Mixed F crest-Coniferous Forest-Alpine 15 1391.22
Meadow Province
Eastern Cascades Section 30 580.00
Oregon and Washington Coast Ranges Section 31 411.21
Western Cascades Section 32 400.01
Central Appalachian Broadleaf F orest-Coniferous 16 1760.00
Forest-Meadow Province
Allegheny Mountains Section 33 250.00
Blue Ridge Mountains Section 34 540.00
Northern Cumberland Mountains Section 35 140.00
Northern Ridge & Valley Section 36 830.00
Chihuahuan Semi-Desert Province 17 510.00
Basin and Range Section 37 190.00
Stockton Plateau Section 38 320.00

 

192

Table D.1 (cont’d).

 

 

 

 

Name Code Area (kmz)
Colorado Plateau Semi-Desert Province 19 2770.00
Grand Canyon Lands Section 41 590.00
Navajo Canyonlands Section 42 570.00
Northern Rio Grande Interrnontane Section 43 910.00
Painted Desert Section 44 250.00
Tonto Transition Section 45 450.00
Eastern Broadleaf Forest (Continental) Province 20 6702.00
Central Till Plains, Beech-Maple Section 46 700.00
Central Till Plains, Oak-Hickory Section 47 410.00
Erie and Ontario Lake Plain Section 48 481.53
Interior Low Plateau, Bluegrass Section 49 400.00
Interior Low Plateau, Highland Rim Section 50 760.00
Interior Low Plateau, Shawnee Hills Section 51 370.00
Lake Agassiz, Aspen Parklands Section 52 880.00
Minnesota & NE Iowa Morainal, Oak Savannah Section 53 760.00
North Central US. Driftless and Escarpment Section 54 430.00
Ozark Highlands Section 55 120.00
South Central Great Lakes Section 56 650.47
Southwestern Great Lakes Morainal Section 57 460.00
Upper Gulf Coastal Plain Section 58 280.00
Eastern Broadleaf Forest (Oceanic) Province 21 3912.60

193

Table D.l (cont’d).

 

 

 

 

 

 

Name Code Area (kmz)
Central Ridge and Valley Section 59 130.00
Hudson Valley Section 60 150.00
Lower New England Section 61 801.92
Northern Appalacian Piedmont Section 62 210.08
Northern Cumberland Plateau Section 63 360.00
Southern Cumberland Mountains Section 64 990.00
Southern Unglaciated Allegheny Plateau Section 65 820.00
Upper Atlantic Coastal Plain Section 66 140.59
Western Glaciated Allegheny Plateau Section 67 310.00
Everglades Province 22 205.96
Everglades Section 68 205.96
Great Plains Steppe and Shrub Province 23 460.00
Redbed Plains Section 69 460.00
Great Plains Steppe Province 24 2080.00
Nebraska Sand Hills Section 70 480.00
North-Central Great Plains Section 71 430.00
Northeastern Glaciated Plains Section 72 720.00
South-Central Great Plains Section 73 150.00
Western Glaciated Plains Section 74 300.00
Great Plains-Palouse Dry Steppe Province 25 3780.00
Arkansas Tablelands Section 75 610.00

194

Table D1 (cont’d).

 

 

 

 

Name Code Area (kmz)
Central High Plains Section 76 600.00
Central High Tablelands Section 77 670.00
Northern Glaciated Plains Section 78 700.00
Northwestern Glaciated Plains Section 79 110.00
Northwestern Great Plains Section 80 200.00
Palouse Prairie Section 81 170.00
Powder River Basin Section 82 120.00
Southern High Plains Section 83 440.00
Upper Rio Grande Basin Section 84 160.00
Intermountain Semi-Desert and Desert Province 27 3660.00
Bonneville Basin Section 86 730.00
Lahontan Basin Section 87 660.00
Mono Section 88 260.00
Northeastern Great Basin Section 89 330.00
Northern Canyon Lands Section 90 390.00
Southeastern Great Basin Section 91 380.00
Uinta Basin Section 92 910.00
Intermountain Semi-Desert Province 28 3800.00
Bear Lake Section 93 790.00
Bighorn Basin Section 94 200.00
Central Basin and Hills Section 95 410.00

195

 

Table D.1 (cont’d).

 

 

 

Name Code Area (kmz)
Columbia Basin Section 96 550.00
Greater Green River Basin Section 97 510.00
High Lava Plains Section 98 210.00
Northwestern Basin and Range Section 99 120.00
Owyhee Uplands Section 100 740.00
Snake River Basalts Section 101 270.00
Laurentian Mixed Forest Province 29 5336.66
Aroostook Hills & Lowlands Section 102 970.00
Central Maine Coastal & Interior Section 103 124.71
Fundy Coastal & Interior Section 104 592.73
Maine & New Brunswick Foothills & Central 105 170.00
Lowlands Section
Northern Glaciated Allegheny Plateau Section 106 550.00
Northern Great Lakes Section 107 797.02
Northern Minnesota & Ontario Section 108 240.00
Northern Minnesota Drift & Lake Plains Section 109 310.00
Northern Superior Uplands Section 110 180.75
Northern Unglaciated Allegheny Plateau Section 1 11 150.00
Southern Superior Uplands Section 112 921.26
St. Lawrence Valley Section 113 120.19
Western Superior Section 114 210.00

 

196

Table D.1 (cont’d).

 

 

 

 

 

Name Code Area (ka)
Lower Mississippi Riverine Forest Province 30 120.00
Mississippi Alluvial Basin Section 115 120.00
Middle Rocky Mountain Steppe-Coniferous Forest- 31 2740.00
Alpine Meadow Province
Beaverhead Mountains Section 116 510.00
Belt Mountains Section 117 320.00
Bitterroot Valley Section 118 200.00
Blue Mountains Section 119 450.00
Challis Volcanics Section 120 140.00
Idaho Batholith Section 121 430.00
Rocky Mountain Front Section 122 690.00
N evada-Utah Mountains-Semi-Desert-Coniferous 32 1130.00
Forest-Alpine Meadow Province
Central Great Basin Mountains Section 123 630.00
Tavaputs Plateau Section 124 180.00
Utah High Plateaus and Mountains Section 125 320.00
Northern Rocky Mountain F orest-Steppe-Coniferous 33 990.00
Forest-Alpine Meadow Province
Bitterroot Mountains Section 126 330.00
Flathead Valley Section 127 210.00
Northern Rockies Section 128 110.00
Okanogan Highlands Section 129 340.00

 

197

Table D.1 (cont’d).

 

 

 

 

 

 

 

Name Code Area (kmz)
Ouachita Mixed Forest - Meadow Province 34 230.00
Ouachita Mountains Section 130 230.00
Outer Coastal Plain Mixed Forest Province 35 3741.64
Atlantic Coastal F latlands Section 131 607.53
Coastal Plains and Flatwoods, Lower Section 132 231.20
Coastal Plains and F latwoods, Western Gulf Section 133 1130.00
Florida Coastal Lowlands (Eastern) Section 134 601.02
Florida Coastal Lowlands (Western) Section 135 351.50
Louisiana Coast Prairies and Marshes Section 136 364.24
Middle Atlantic Coastal Plain Section 137 456.15
Ozark Broadleaf Forest - Meadow Province 36 170.00
Boston Mountains Section 138 170.00
Pacific Lowland Mixed Forest Province 39 372.75
Willamette Valley and Puget Trough Section 147 372.75
Prairie Parkland (Subtropical) Province 40 2081.05
Blackland Prairies Section 148 340.00
Central Gulf Prairies and Marshes Section 149 241.05
Cross Timbers and Prairie Section 150 930.00
Oak Woods and Prairies Section 151 570.00
Prairie Parkland (Temperate) Province 41 2860.00
Central Dissected Till Plains Section 152 180.00

198

Table D.1 (cont’d).

 

 

 

 

Name Code Area (kmz)
Central Loess Plains Section 153 1170.00
Flint Hills Section 154 440.00
North-Central Glaciated Plains Section 155 130.00
Osage Plains Section 156 470.00
Red River Valley Section 157 470.00
Sierran Steppe-Mixed F orest-Coniferous F orest- 44 2390.00
Alpine Meadow Province
Klamath Mountains Section 160 420.00
Modoc Plateau Section 161 250.00
Northern California Coast Ranges Section 162 160.00
Northern California Interior Coast Ranges Section 163 700.00
Sierra Nevada Foothills Section 164 230.00
Sierra Nevada Section 165 470.00
Southern Cascades Section 166 160.00
Southeastern Mixed Forest Province 45 2510.00
Arkansas Valley Section 167 200.00
Coastal Plains, Middle Section 168 150.00
Eastern Gulf Prairies and Marshes Section 169 590.00
Mid Coastal Plains, Western Section 170 980.00
Southern Appalachian Piedmont Section 171 190.00
Southern Cumberland Plateau Section 172 230.00

199

 

Table D.1 (cont’d).

 

 

 

 

Name Code Area (kmz)
Southern Ridge and Valley Section 173 170.00
Southern Rocky Mountain Steppe-Open Woodland- 46 3340.00
Coniferous Forest-Alpine Meadow Province
Bighorn Mountains Section 174 130.00
North-Central Highlands Section 175 310.00
Northern Parks and Ranges Section 176 470.00
Overthrust Mountains Section 177 460.00
South-Central Highlands Section 178 460.00
Southern Parks and Ranges Section 179 190.00
Uinta Mountains Section 180 190.00
Southwest Plateau and Plains Dry Steppe and Shrub 47 2792.02
Province
Edwards Plateau Section 183 570.00
Pecos Valley Section 184 500.00
Rio Grande Plain Section 185 770.00
Rolling Plains Section 186 100.00
Southern Gulf Prairies and Marshes Section 187 722.02
Texas High Plains Section 188 130.00

 

200

LITERATURE CITED

Alard, D., and I. Poudevigne. 2000. Diversity patterns in grasslands along a landscape
gradient in northwestern France. Journal of Vegetation Science 11:287-294.

Allain, C., and M. Cloitre. 1991. Characterizing the lacunarity of random and
deterministic fractal sets. Physical Review A 44:3552-3558.

Allen, C. R., E. A. Forys, and C. S. Holling. 1999. Body mass patterns predict invasions
and extinctions in transforming landscapes. Ecosystems 2:114-121.

Allen, C. R., and D. A. Saunders. 2002. Variability between scales: predictors of
nomadism in birds of an Australian Mediterranean-climate ecosystem.
Ecosystems 52348-359.

August, P., L. Iverson, and J. Nugranad. 2002. Human conversion of terrestrial habitats.
Pages 198-224 in K. J. Gutzwiller, editor. Applying landscape ecology in
biological conservation. Springer, New York City, New York.

Bailey, R. G. 1988. Ecogeographic analysis: a guide to the ecological division of land for
resource management. Misc. Publ. No. 1465. USDA Forest Service, Washington,
DC.

Bailey, R. G. 1996. Ecosystem geography. Springer, New York.

Bailey, R. G. 2002. Ecoregion-based design for sustainability. Springer, New York.

Baker, W. L., and Y. Cai. 1992. The r.le programs for multiscale analysis of landscape
structure using the GRASS geographical information system. Landscape Ecology

7:291-302.

Barbour, M. G., J. H. Burk, and W. D. Pitts. 1987. Terrestrial Plant Ecology, 2nd edition.
The Benjamin/Cummings Publishing Company, Inc., Menlo Park, California.

Begon, M., J. L. Harper, and C. R. Townsend. 1990. Ecology: individuals, populations
and communities, 2nd edition. Blackwell Scientific, London.

Bender, D. J ., L. Tischendorf, and L. Fahrig. 2003. Using patch isolation metrics to
predict animal movement in binary landscapes. Landscape Ecology 18: 17-39.

Bolstad, P. 2002. GIS Fundamentals: A first text on geographic information systems.
Eider Press, White Bear Lake, Minnesota.

Boulinier, T., J. D. Nichols, J. E. Hines, J. R. Sauer, C. H. Flather, and K. H. Pollock.
1998. Higher temporal variability of forest breeding bird communities in

201

fragmented landscapes. Proceedings of the National Academy of Sciences, United
States of America 95:7497-7501.

Boulinier, T., J. D. Nichols, J. E. Hines, J. R. Sauer, C. H. Flather, and K. H. Pollock.
2001. Forest fragmentation and bird community dynamics: inference at regional
scales. Ecology 82:1159-1 169.

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

Brown, J. H., D. W. Mehlman, and G. C. Stevens. 1995. Spatial variation in abundance.
Ecology 76:2028-2043.

Brown, J. H., G. C. Stevens, and D. M. Kaufman. 1996. The geographic range: size,
shape, boundaries, and internal structure. Annual Review of Ecology and
Systematics 27:597-623.

Bunn, A. G., D. L. Urban, and T. H. Keitt. 2000. Landscape connectivity: A conservation
application of graph theory. Journal of Environmental Management 59:265-278.

Bystrak, D. 1981. The North American Breeding Bird Survey. Studies in Avian Biology
6:34-41.

Calder, W. A. 1996. Size, function, and life history. Dover Publications, Inc., Mineola
NY.

Carrere, V. 1990. Development of multiple source data processing for structural analysis
at a regional scale. Photogrammetric Engineering and Remote Sensing 56:587-
595.

Caughley, G., D. Grice, R. J. Barker, and B. Brown. 1988. The edge of the range. Journal
of Animal Ecology 57:771-785.

Chambers, J. M., W. 8. Cleveland, K. Beat, and P. A. Tukey. 1983. Graphical methods
for data. Duxbury Press, Boston, Massachusetts.

Chao, A., R. L. Chazdon, R. K. Colwell, and T.-J. Shen. 2005. A new statistical approach
for assessing similarity of species composition with incidence and abundance
data. Ecology Letters 8: 148-159.

Chapin, F. S., III, B. H. Walker, R. J. Hobbs, D. U. Hooper, J. H. Lawton, O. E. Sala, and

D. Tilman. 1997. Biotic control over the functioning of ecosystems. Science
277:500-504.

202

Chen, J ., J. F. Franklin, and T. A. Spies. 1992. Vegetation responses to edge
environments in old-growth Douglas-fir forests. Ecological Applications 2:387-
396.

Clark, J. S., and D. Pregibon. 1992. Three-based models. Pages 377-419 in J. M.
Chambers and T. J. Hastie, editors. Statistical models in S. Wadsworth &
Brooks/Cole, Pacific Grove, California.

Clarke, K. C. 1999. Getting Started with Geographic Information Systems. Prentice Hall,
Upper Saddle River, New Jersey.

Cleland, D. T., P. E. Avers, H. McNab, M. E. Jensen, R. G. Bailey, T. King, and W. E.
Russell. 1997. National hierarchical framework of ecological units. in M. S.
Boyce and A. Haney, editors. Ecosystem management applications for sustainable
forest and wildlife resources. Yale University Press, New Haven, CT.

Cornelius, J. M., and J. F. Reynolds. 1991. On determining the statistical significance of
discontinuities within ordered ecological data. Ecology 72:2057-2070.

Cuddington, K., and P. Yodzis. 2002. Predator-prey dynamics and movement in fractal
environments. The American Naturalist 160:1 19-134.

Cumutt, J. L., S. L. Pimm, and B. A. Maurer. 1996. Population variability of sparrows in
space and time. Oikos 76:131-144.

Dale, M. R. T. 2000. Lacunarity analysis of spatial pattern: A comparison. Landscape
Ecology 15:467-478.

Delcourt, H. R. 1987. The impact of prehistoric agriculture and land occupation on
natural vegetation. Trends in Ecology & Evolution 2:39-44.

Dennis, B. 1989. Stochastic differential equations as insect population models. Pages
219-238 in L. McDonald, B. Manly, J. Lockwood, and J. Logan, editors.
Estimation and analysis of insect populations. Springer-Verlag, Berlin.

Dennis, B., and R. F. Costantino. 1988. Analysis of steady-state populations with the
gamma-abundance model - application to Tribolium. Ecology 69: 1200-1213.

Dennis, B., and G. P. Patil. 1984. The gamma-distribution and weighted multimodal
gamma-distributions as models of population abundance. Mathematical
Biosciences 68:187-212.

Diamond, J. M. 1972. Biogeographic kinetics: estimation of relaxation times for
avifaunas of southwest Pacific islands. Proceedings of the National Academy of
Sciences, United States of America 69:3199-3203.

203

Dunning, J. B., Jr., editor. 1993. CRC Handbook of Avian Body Masses. CRC Press,
Boca Raton, Florida.

Dunning, J. B., Jr., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that
affect populations in complex landscapes. Oikos 65:169-175.

Ehrlich, P. R., D. S. Dobkin, and D. Wheye. 1988. The birder's handbook: a field guide to
the natural history of North American birds. Simon & Schuster, New York.

Elliott, J. J., and R. S. Arbib, Jr. 1953. Origin and status of the house finch in the eastern
United States. The Auk 70:31-37.

Environmental Systems Research Institute Inc. 1992-1999. ArcView GIS 3.2. Redmond,
California.

F ahrig, L., and G. Merriam. 1985. Habitat patch connectivity and population survival.
Ecology 66: 1762-1768.

Famsworth, E. J ., and A. M. Ellison. 1996. Scale-dependent spatial and temporal
variability in biogeography of mangrove root epibiont communities. Ecological
Monographs 66:45-66.

Fitzgerald, J. A., J. R. Herkert, and J. D. Brawn. 2000. Partners in Flight bird
conservation plan for the Prairie Peninsula (physiographic area 31). Partners in
Flight, Brentwood, Missouri.

Forman, R. T. T., and M. Godron. 1986. Landscape Ecology. John Wiley & Sons., New
York.

Franklin, J. F., and R. T. T. Forman. 1987. Creating landscape patterns by forest cutting:
Ecological consequences and principles. Landscape Ecology 1:5-18.

Game, M. 1980. Best shape for nature reserves. Nature 287:630-632.
Gardner, R. H. 1998. Pattern, process, and the analysis of spatial scales. Pages 17-34 in
D. L. Peterson and V. T. Parker, editors. Ecological scale: theory and

applications. Columbia University Press, New York.

Gardner, R. H., B. T. Milne, M. G. Turner, and R. V. O'Neill. 1987. Neutral models for
the analysis of broad-scale landscape pattern. Landscape Ecology 1:19-28.

Gaston, K. J ., and T. M. Blackburn. 2000. Pattern and process in macroecology.
Blackwell Science, Oxford.

Gering, J. C. 2004. Reply to "Ecoregion as a pragmatic tool". Conservation Biology 18:5-
6.

204

Gering, J. C., T. O. Crist, and J. A. Veech. 2003. Additive partitioning of species
diversity across multiple spatial scales: implications for regional conservation of
biodiversity. Conservation Biology 17:488-499.

Goldstein, P. Z. 1999. Functional ecosystems and biodiversity buzzwords. Conservation
Biology 13:247-255.

Golledge, R. G., R. Church, J. Dozier, J. E. Estes, J. Michaelsen, D. S. Simonett, R.
Smith, T. Smith, A. H. Strahler, and W. R. Tobler. 1982. Commentary on "The
highest form of the geographer's art". Annals of the Association of American
Geographers 72:557-558.

Government of Canada. 1996. The state of Canada's environment - 1996. Government of
Canada, Ottawa.

Groves, C., L. Valutis, D. Vosick, B. Neely, K. Wheaton, J. Touval, and B. Runnels.
2000. Designing a geography of hope: a practitioner's handbook for ecoregional
conservation planning. The Nature Conservancy, Arlington, Virginia.

Gustafson, E. J. 1998. Quantifying landscape spatial pattern: what is state of the art?
Ecosystems 1:143-156.

Gustafson, E. J ., and G. R. Parker. 1992. Relationships between landcover proportion and
indices of landscape spatial pattern. Landscape Ecology 7: 101-1 10.

Haines-Young, R., and M. Chopping. 1996. Quantifying landscape structure: a review of
landscape indices and their application to forested landscapes. Progress in
Physical Geography 20:418-445.

Hanski, 1., and M. E. Gilpin. 1997. Metapopulation biology: ecology, genetics, and
evolution. Academic Press, New York, New York.

Hargis, C. D., J. A. Bissonette, and J. L. David. 1998. The behavior of landscape metrics
commonly used in the study of habitat fragmentation. Landscape Ecology 13:167-
186.

Hargrove, W. H., and J. Pickering. 1992. Pseudoreplication: a sine qua non for regional
ecology. Landscape Ecology 62251-258.

Hart, J. F. 1982a. Comment in reply. Annals of the Association of American Geographers
72:559.

Hart, J. F. 1982b. The highest form of the geographer's art. Annals of the Association of
American Geographers 72: 1-29.

205

Hart, J. F. 1983. More gnashing of false teeth. Annals of the Association of American
Geographers 73:441-443.

Havlicek, T. D., and S. R. Carpenter. 2001. Pelagic species size distributions in lakes:
Are they discontinuous? Limnological Oceanography 46: 1021-1033.

He, H. S., B. E. DeZonia, and D. J. Mladenoff. 2000. An aggregation index (Al) to
quantify spatial patterns of landscapes. Landscape Ecology 15:591-601.

Healey, R. G. 1983. Regional geography in the computer age: a further commentary on
"The highest form of the geographer's art". Annals of the Association of
American Geographers 73:43 9-44 1.

Herzog, F., and A. Lausch. 2001. Supplementing land-use statistics with landscape
metrics: some methodological considerations. Environmental Monitoring and
Assessment 72:37-50.

Holling, C. S. 1973. Resilience and stability of ecological systems. Annual Review of
Ecology and Systematics 4: 1-23.

Holling, C. S. 1992. Cross-scale morphologies, geometry, and dynamics of ecosystems.
Ecological Monographs 62:447-502.

Holling, C. S., and C. R. Allen. 2002. Adaptive inference for distinguishing credible from
incredible patterns in nature. Ecosystems 5:319-328.

Holling, C. S., G. Peterson, P. Marples, J. Sendzimir, K. H. Redford, L. Gunderson, and
D. Lambert. 1996. Self-organization in ecosystems: lumpy geometries,
periodicities and morphologies. Pages 346-384 in B. Walker and W. Steffen,
editors. Global change and terrestrial ecosystems. Cambridge University Press,
Cambridge, Massachusetts.

Hunter, W. C., L. H. Peoples, and J. A. Collazo. 2001. South Atlantic coastal plain
Partners in Flight bird conservation plan (physiographic area #03). US. Fish and
Wildlife Service in cooperation with Partners in Flight, Atlanta, Georgia.

Hutchinson, G. E. 1957. Concluding remarks. Cold Spring Harbor Symp. Q. Biol.
22:415-427.

Istock, C. A., and S. M. Scheiner. 1987. Affinities and high-order diversity within
landscape mosaics. Evolutionary Ecology 1:11-29.

Jenness, J. 2004. Nearest features (nearfeatavx) extension for ArcView 3.x, v. 3.8.
Jenness Enterprises.

206

Johnson, A. R., J. A. Wiens, B. T. Milne, and T. O. Crist. 1992. Animal movements and
population dynamics in heterogeneous landscapes. Landscape Ecology 7:63-75.

Kareiva, P. 1990. Population dynamics in spatially complex environments: theory and
data. Philosophical Transactions of the Royal Society of London, Series B
330:175-190.

Kaufman, L., and P. J. Rousseeuw. 1990. Finding groups in data: an introduction to
cluster analysis. John Wiley & Sons, Inc., New York.

King, A. W. 1991. Translating models across scales in the landscape. Pages 479-517 in
M. G. Turner and R. H. Gardner, editors. Quantitative Methods in Landscape
Ecology: Analysis and Interpretation of Landscape Heterogeneity. Springer-
Verlag, New York.

Knutson, M. G., G. Butcher, J. Fitzgerald, and J. Shieldcastle. 2001. Partners in Flight
bird conservation plan for the Upper Great Lakes Plain (physiographic area 16).
USGS Upper Midwest Environmental Sciences Center in cooperation with
Partners in Flight, LaCrosse, Wisconsin.

Koleff, P., K. J. Gaston, and J. J. Lennon. 2003. Measuring beta diversity for presence-
absence data. Journal of Animal Ecology 72:367-382.

Krebs, C. J. 1989. Ecological Methodology. Harper Collins Publishers, New York, New
York.

Lambert, W. D., and C. S. Holling. 1998. Causes of ecosystem transformation at the end
of the Pleistocene: evidence from mammal body-mass distributions. Ecosystems
1:157-175.

Leaper, R., D. Raffaelli, C. Emes, and B. F. J. Manly. 2001. Constraints on body-size
distributions: an experimental test of the habitat architecture hypothesis. Journal
of Animal Ecology 70:248-259.

Legendre, P., and L. Legendre. 1998. Numerical Ecology. Elsevier, Amsterdam.

Lennon, J. J ., P. Koleff, J. J. D. Greenwood, and K. J. Gaston. 2001. The geographic
structure of British bird distributions: diversity, spatial turnover and scale. Journal
of Animal Ecology 70:966-979.

Levin, S. A. 1992. The problem of pattern and scale in ecology. Ecology 73: 1943-1967.

Li, B.-L., and S. Archer. 1997. Weighted mean patch size: a robust index for quantifying
landscape structure. Ecological Modelling 1022353-361.

207

Li, H., and J. F. Reynolds. 1993. A new contagion index to quantify spatial patterns of
landscapes. Landscape Ecology 8: 155-162.

Li, H., and J. F. Reynolds. 1994. A simulation experiment to quantify spatial
heterogeneity in categorical maps. Ecology 75:2446-2455.

Li, H., and J. Wu. 2004. Use and misuse of landscape indices. Landscape Ecology
19:389-399.

Lidicker, W. Z., Jr., and W. D. Koenig. 1996. Responses of terrestrial vertebrates to
habitat edges and corridors. Pages 85-109 in D. R. McCullough, editor.
Metapopulations and Wildlife Conservation. Island Press, Washington, DC.

Liu, J. 1993. ECOLECON: an ECOLogical-ECONomic model for species conservation
in complex forest landscapes. Ecological Modelling 70:63-87.

Ma, K.-M., B.-J. Fu, X.-D. Gue, and H.-F. Zhou. 2000. Finding spatial regularity in
mosaic landscapes: two methods integrated. Plant Ecology 149: 195-205.

Mac Nally, R., A. F. Bennett, G. W. Brown, L. F. Lumsden, A. Yen, S. Hinkley, P.
Lillywhite, and D. Ward. 2002. How well do ecosystem-based planning units

represent different components of biodiversity? Ecological Applications 12:900-
912.

Mac Nally, R., E. Fleishman, L. P. Bulluck, and C. J. Betrus. 2004. Comparative
influence of spatial scale on beta diversity within regional assemblages of birds

and butterflies. Journal of Biogeography 31:917-929.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography. Princeton
University Press, Princeton, New Jersey.

Magnusson, W. E. 2004. Ecoregion as a pragmatic tool. Conservation Biology 18:4-5.

Mandelbrot, B. B. 1983. The Fractal Geometry of Nature. W.H. Freeman, San Francisco,
California.

Manly, B. F. J. 1996. Are there clumps in body-size distributions? Ecology 77:81-86.

Margules, C. R., and R. L. Pressey. 2000. Systematic conservation planning. Nature
405:243-253.

Matson, P. A., W. J. Parton, A. G. Power, and M. J. Swift. 1997. Agricultural
intensification and ecosystem properties. Science 277:504-509.

Maurer, B. A., E. T. Linder, and D. E. Gammon. 2001. A geographical perspective on the
biotic homogenization process: implications from the macroecology of North

208

American birds. Pages 157-177 in J. L. Lockwood and M. L. McKinney, editors.
Biotic Homogenization. Kluwer Academic/Plenum Publishers, New York City,
New York.

Maurer, B. A., and M. L. Taper. 2002. Connecting geographical distributions with
population processes. Ecology Letters 52223-231.

Mazerolle, M. J ., and M.-A. Villard. 1999. Patch characteristics and landscape context as
predictors of species presence and abundance: A review. EcoScience 6: 1 17-124.

McCoy, E. D., and S. S. Bell. 1991. Habitat structure: The evolution and diversification
of a complex topic. Pages 3-27 in S. S. Bell, E. D. McCoy, and H. R. Mushinsky,
editors. Habitat Structure: The physical arrangement of objects in space. Chapman
and Hall.

McGarigal, K. M., and B. J. Marks. 1995. FRAGSTATS: spatial pattern analysis program
for quantifying landscape structure. USDA Forest Service General Technical
Report PNW-351.

Meisel, J. E., and M. G. Turner. 1998. Scale detection in real and artificial landscapes
using semivariance analysis. Landscape Ecology 13:347-362.

Milligan, G. W. 1981. A review of Monte Carlo tests of cluster analysis. Multivariate
Behavioral Research 16:379-407.

Milne, B. T., K. M. Johnston, and R. T. T. Forman. 1989. Scale-dependent proximity of
wildlife habitat in a spatially-neutral Bayesian model. Landscape Ecology 2: 101-
110.

Milne, B. T., M. G. Turner, J. A. Wiens, and A. R. Johnson. 1992. Interactions between
the fractal geometry of landscapes and allometric herbivory. Theoretical
Population Biology 41:337-353.

Mladenoff, D. J., and B. E. DeZonia. 2002. APACK 2.22 User's Guide.

Myers, N., R. A. Mittermeier, C. G. Mittermeier, G. A. B. da Fonseca, and J. Kent. 2000.
Biodiversity hotspots for conservation priorities. Nature 403:853-858.

Native Seed Network. 2005. Ecoregion Information: http://www.nativeseednetwork.org.
Corvallis, Oregon.

Nekola, J. C., and P. S. White. 1999. The distance decay of similarity in biogeography
and ecology. Journal of Biogeography 26:867-878.

Nicholls, A. O. 1994. Variation in mosaic diversity in the forests of coastal northern New
South Wales. Pacific Conservation Biology 1:177-182.

209

Noon, B. R., and V. H. Dale. 2002. Broad-scale ecological science and its application.
Pages 34-52 in K. J. Gutzwiller, editor. Applying landscape ecology in biological
conservation. Springer, New York, New York.

Olson, D. M., and E. Dinerstein. 1998. The global 200: a representation approach to
conserving the Earth's most biologically valuable ecoregions. Conservation
Biology 12:502-515.

Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. V. N. Powell, E.
C. Underwood, J. A. D'Amico, I. Itoua, H. E. Strand, J. C. Morrison, C. Loucks,
T. F. Allnutt, T. H. Ricketts, Y. Kura, J. F. Lamoreux, W. W. Wettengel, P.
Hedao, and K. R. Kassem. 2001. Terrestrial ecoregions of the world: a new map
of life on earth. BioScience 51:933-938.

Omemik, J. M. 1995. Ecoregions: A spatial framework for environmental management.
Pages 49-65 in W. S. Davis and T. P. Simon, editors. Biological assessment and
criteria: tools for water resource planning and decision making. CRC Press, Boca
Raton, FL.

O'Neill, R. V., C. T. Hunsaker, S. P. Timmins, B. L. Jackson, K. B. Jones, K. H. Riitters,
and J. D. Wickham. 1996. Scale problems in reporting landscape pattern at the
regional scale. Landscape Ecology 11:169-180.

O'Neill, R. V., and A. W. King. 1998. Homage to St. Michael; or, why are there so many
books on scale? Pages 3-15 in D. L. Peterson and V. T. Parker, editors. Ecological
Scale: Theory and Applications. Columbia University Press, New York.

O'Neill, R. V., J. R. Krummel, R. H. Gardner, G. Sugihara, B. Jackson, D. L. DeAngelis,
B. T. Milne, M. G. Turner, B. Zygmunt, S. W. Christensen, V. H. Dale, and R. L.
Graham. 1988a. Indices of landscape pattern. Landscape Ecology 12153-162.

O'Neill, R. V., B. T. Milne, M. G. Turner, and R. H. Gardner. 1988b. Resource utilization
scales and landscape pattern. Landscape Ecology 2:63-69.

Patton, D. R. 1975. A diversity index for quantifying habitat "edge". Wildlife Society
Bulletin 3:171-173.

Peters, R. H. 1983. The ecological implications of body size, lst edition. Press Syndicate
of the University of Cambridge, Cambridge.

Peterson, D. L., and V. T. Parker. 1998. Dimensions of scale in ecology, resource
management, and society. Pages 499-522 in D. L. Peterson and V. T. Parker,
editors. Ecological scale: theory and applications. Columbia University Press,
New York.

210

Pickett, S. T. A., and K. H. Rogers. 1997. Patch dynamics: the transformation of
landscape structure and function. Pages 101-127 in J. A. Bissonette, editor.
Wildlife and Landscape Ecology: Effects of Pattern and Scale. Springer, New
York.

Pimm, S. L. 1986. Community stability and structure. Pages 309-329 in M. E. Soulé,
editor. Conservation biology: the science of scarcity and diversity. Sinauer
Associates, Inc., Sunderland, Massachusetts.

Pimm, S. L. 2001. The world according to Pimm: a scientist audits the Earth. R.R.
Donnelley & Sons.

Pimm, S. L., and R. A. Askins. 1995. Forest losses predict bird extinctions in eastern
North America. Proceedings of the National Academy of Sciences, United States
of America 92:9343-9347.

Pimm, S. L., and P. Raven. 2000. Extinction by numbers. Nature 403:843-845.

Plotnick, R. E., R. H. Gardner, W. H. Hargrove, K. Prestegaard, and M. Perlmutter. 1996.
Lacunarity analysis: A general technique for the analysis of spatial patterns.
Physical Review E 53:5461-5468.

Plotnick, R. E., R. H. Gardner, and R. V. O'Neill. 1993. Lacunarity indices as measures
of landscape texture. Landscape Ecology 8:201-211.

Poynton, J. C., and R. C. Boycott. 1996. Species turnover between Afromontane and
eastern African lowland faunas: patterns shown by amphibians. Journal of
Biogeography 23:669-680.

Qian, H., K. Klinka, and G. J. Kayahara. 1998. Longitudinal patterns of plant diversity in
the North American boreal forest. Plant Ecology 138:161-178.

R Development Core Team. 2004. R: A language and environment for statistical
computing. R Foundation for Statistical Computing. Vienna, Austria. ISBN 3-
900051-07-0, URL http://www.R-project.org.

Raffaelli, D., S. Hall, C. Emes, and B. F. J. Manly. 2000. Constraints on body size
distributions: an experimental approach using a small-scale system. Oecologia
122:389-398.

Restrepo, C., L. M. Renjifo, and P. Marples. 1997. F rugivorous birds in fragmented
neotropical montane forests: landscape pattern and body mass distribution. Pages
616 in W. F. Laurance and R. O. Bierregaard, Jr., editors. Tropical forest
remnants: ecology, management, and conservation of fragmented communities.
University of Chicago Press, Chicago, Illinois.

211

Rey Benayas, J. M., and S. M. Scheiner. 2002. Plant diversity, biogeography and
environment in Iberia: Patterns and possible causal factors. Journal of Vegetation
Science 13:245-258.

Richards, J. F. 1990. Land transformation. Pages 163-178 in B. L. Turner, 11, W. C.
Clark, R. W. Kates, J. F. Richards, J. T. Mathews, and W. B. Meyer, editors. The
earth as transformed by human action: global and regional changes in the
biosphere over the past 300 years. Cambridge University Press, Cambridge,
Massachusetts.

Ricketts, T. H., E. Dinerstein, D. M. Olson, and C. Loucks. 1999. Who's where in North
America? Patterns of species richness and the utility of indicator taxa for
conservation. BioScience 492369-381.

Ricklefs, R. E. 2004. A comprehensive framework for global patterns in biodiversity.
Ecology Letters 7: 1-15.

Riitters, K. H., R. V. O'Neill, C. T. Hunsaker, J. D. Wickham, D. H. Yankee, S. P.
Timmins, K. B. Jones, and B. L. Jackson. 1995. A factor analysis of landscape
pattern and structure metrics. Landscape Ecology 10:23-39.

Riitters, K. H., R. V. O'Neill, and K. B. Jones. 1997. Assessing habitat suitability at
multiple scales: A landscape-level approach. Biological Conservation 81: 191-202.

Riitters, K. H., R. V. O'Neill, J. D. Wickham, and K. B. Jones. 1996. A note on contagion
indices for landscape analysis. Landscape Ecology 11:197-202.

Rising, J. D., and K. M. Somers. 1989. The measurement of overall body size in birds.
The Auk 106:666-674.

Robbins, C. S., D. Bystrak, and P. H. Geissler. 1986. The Breeding Bird Survey: its first
fifteen years, 1965-1979. US. Fish Wild. Serv. Resour. Publ. 157.

Robinson, S. K., F. R. Thompson, 111, T. M. Donovan, D. R. Whitehead, and J. Faaborg.
1995. Regional forest fragmentation and the nesting success of migratory birds.
Science 267: 1987-1990.

Romme, W. H. 1982. Fire and landscape diversity in subalpine forests of Yellowstone
National Park. Ecological Monographs 52: 199-221.

Rousseeuw, P. J ., A. Struyf, and M. Hubert. 2004. Cluster: functions for clustering. R
package version 1.9.6.

Ruggiero, A., J. H. Lawton, and T. M. Blackburn. 1998. The geographic ranges of

mammalian species in South America: spatial patterns in environmental resistnace
and anisotropy. Journal of Biogeography 25: 1093-1 103.

212

 

SAS Institute. 1999. SAS User's Guide: Statistics, Version 5 Edition. SAS Institute, Inc.,
Cary, North Carolina.

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.

Scheiner, S. M. 1992. Measuring pattern diversity. Ecology 73:1860-1867.

Scheiner, S. M., and J. M. Rey Benayas. 1994. Global patterns of plant diversity.
Evolutionary Ecology 8:331-347.

Schneider, D. C. 1998. Applied scaling theory. Pages 253-269 in D. L. Peterson and V. T.
Parker, editors. Ecological Scale: Theory and Applications. Columbia University
Press, New York.

Schoener, T. W. 1968. Sizes of feeding territories among birds. Ecology 49: 123-141.

Schumaker, N. H. 1996. Using landscape indices to predict habitat connectivity. Ecology
77:1210-1225.

Schwinghamer, P. 1981. Characteristic size distributions of integral benthic communities.
Can. J. Fish. Aquat. Sci. 38:1255-1263.

Schwinghamer, P. 1988. Influence of pollution along a natural gradient and in a
mesocosm experiment on biomass-size spectra of benthic communities. Marine

Ecology Progress Series 46:199-206.

Scott, J. M. 1993. A geographical approach to protection of biological diversity. Wildlife
Monographs: 1-41.

Siemann, E., and J. H. Brown. 1999. Gaps in mammalian body size distributions
reexamined. Ecology 80:2788-2792.

Silverman, B. W. 1981. Using kernel density estimates to investigate multimodality.
Journal of the Royal Statistical Society, B 43:97-99.

Silverman, B. W. 1986. Density estimation for statistics and data analysis. Chapman &
Hall, London.

South, J. M., and S. Pruett-Jones. 2000. Patterns of flock size, diet, and vigilance of
naturalized monk parakeets in Hyde Park, Chicago. The Condor 102:848-854.

Spector, S. 2002. Biogeographic crossroads as priority areas for biodiversity
conservation. Conservation Biology 16: 1480-1487.

213

Stamps, J. A., M. Buechner, and V. V. Krishnan. 1987. The effects of edge permeability
and habitat geometry on emigration from patches of habitat. The American
Naturalist 129:533-552.

Stewart, A. J. A., E. A. John, and M. J. Hutchings. 2000. The world is heterogeneous:
ecological consequences of living in a patchy environment. Pages 1-8 in M. J.
Hutchings, E. A. John, and A. J. A. Stewart, editors. The ecological consequences
of environmental heterogeneity. British Ecological Society, Oxford, England.

Sweeney, B. A., and J. E. Cook. 2001. A landscape-level assessment of understory
diversity in upland forests of North-Central Wisconsin, USA. Landscape Ecology
16:55-69.

Taylor, P. D., L. F ahrig, K. Henein, and G. Merriam. 1993. Connectivity is a vital
element of landscape structure. Oikos 68:571-573.

Tischendorf, L. 2001. Can landscape indices predict ecological processes consistently?
Landscape Ecology 16:235-254.

Tischendorf, L., and L. Fahrig. 2000a. How should we measure landscape connectivity?
Landscape Ecology 15:633-641.

Tischendorf, L., and L. Fahrig. 2000b. On the usage and measurement of landscape
connectivity. Oikos 9027-19.

Tukey, J. W. 1977. Exploratory data analysis. Addison-Wesley, Reading, Massachusetts.

Turchin, P. 1991. Translating foraging movements in heterogeneous environments into
the spatial distribution of foragers. Ecology 72: 1253-1266.

Turchin, P. 1996. Fractal analyses of animal movement: a critique. Ecology 7722086-
2090.

Tumer, M. G. 1989. Landscape ecology: the effect of pattern on process. Annual Review
of Ecology and Systematics 20:171-197.

Turner, M. G., and R. H. Gardner. 1991. Quantitative methods in landscape ecology: an
introduction. Pages 3-14 in M. G. Turner and R. H. Gardner, editors. Quantitative
Methods in Landscape Ecology: The Analysis and Interpretation of Landscape
Heterogeneity. Springer-Verlag, New York.

Turner, M. G., R. H. Gardner, and R. V. O'Neill. 2001. Landscape ecology in theory and
practice. Springer, New York, NY.

Turner, M. G., R. V. O'Neill, R. H. Gardner, and B. T. Milne. 1989. Effects of changing
spatial scale on the analysis of landscape pattern. Landscape Ecology 3: 153-162.

214

Turner, S. J ., R. V. O'Neill, W. Conley, M. R. Conley, and H. C. Humphries. 1991.
Pattern and scale: Statistics for landscape ecology. Pages 17-49 in M. G. Turner
and R. H. Gardner, editors. Quantitative Methods in Landscape Ecology: Analysis
and Interpretation of Landscape Heterogeneity. Springer-Verlag, New York.

US. Department of the Interior. 2005. National Atlas of the United States:
http://www.nationalatlas.gov. Washington, DC.

Urban, D., and T. H. Keitt. 2001. Landscape connectivity: a graph-theoretic perspective.
Ecology 82: 1205-1218.

USDA Forest Service Inventory and Monitoring Institute. 2005. Ecoregions:
http://www.fs.fed.us/institute/ecolink.html. Fort Collings, Colorado.

Veit, R. R., and M. A. Lewis. 1996. Dispersal, population growth, and the Allee effect:
dynamics of the House finch invasion of eastern North America. The American
Naturalist 148:255-274.

Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melilo. 1997. Human
domination of Earth's ecosystems. Science 277:494-499.

Vogelmann, J. E., S. M. Howard, L. Yang, C. R. Larson, B. K. Wylie, and N. Van Driel.
2001. Completion of the 19905 National Land Cover Data Set for the coterminous
United States from Landsat Thematic Mapper data and ancillary data sources.
Photogrammetric Engineering and Remote Sensing 67:650-652.

Ward, J. H. 1963. Hierarchical grouping to optimize an objective function. Journal of the
American Statistical Association 58:236-244.

Warner, R. E. 1994. Agricultural land use and grassland habitat in Illinois: future shock
for midwestem birds? Conservation Biology 8: 147-156.

Western, D. 2001. Human-modified ecosystems and future evolution. Proceedings of the
National Academy of Sciences of the United States of America 98:5458-5465.

Whittaker, R. H. 1975. Communities and ecosystems. Macmillan, New York.

Wiegand, T., K. A. Moloney, J. Naves, and F. Knauer. 1999. Finding the missing link
between landscape structure and population dynamics: a spatially explicit
perspective. The American Naturalist 154:605-627.

Wiens, J. A. 1989. Spatial scaling in ecology. Functional Ecology 32385-397.

Wiens, J. A. 1995. Landscape mosaics and ecological theory. Pages 1-26 in L. Hansson,

L. Fahrig, and G. Merriam, editors. Mosaic landscapes and ecological processes.
Chapman & Hall, London.

215

Wiens, J. A. 2002. Central concepts and issues of landscape ecology. Pages 3-21 in K. J.
Gutzwiller, editor. Applying Landscape Ecology in Biological Conservation.
Springer-Verlag, New York, NY.

Wiens, J. A., T. O. Crist, K. A. With, and B. T. Milne. 1995. Fractal patterns of insect
movement in microlandscape mosaics. Ecology 76:663-666.

Wiens, J. A., and B. T. Milne. 1989. Scaling of 'landscapes' in landscape ecology, or,
landscape ecology from a beetle's perspective. Landscape Ecology 3:87-96.

Wiens, J. A., N. C. Stenseth, B. Van Horne, and R. A. Ims. 1993. Ecological mechanisms
and landscape ecology. Oikos 66:369-380.

Williams, E. J ., and D. N. Pashley. 1999. Designation of conservation planning units. in
R. Bonney, D. N. Pashley, R. J. Cooper, and L. Niles, editors. Strategies for bird
conservation: the Partners in F light planning process. Cornell Lab of Ornithology,
<http://birds.comell.edu/pifcapemay>.

Williams, M. 1990. Forests. Pages 179-201 in B. L. Turner, 11, W. C. Clark, R. W. Kates,
J. F. Richards, J. T. Mathews, and W. B. Meyer, editors. The earth as transformed
by human action: global and regional changes in the biosphere over the past 300
years. Cambridge University Press, Cambridge, Massachusetts.

Wilson, E. O. 1992. The Diversity of Life. The Belknap Press of Harvard University
Press, Cambridge, Massachusetts.

Wilson, M. V., and A. Shmida. 1984. Measuring beta diversity with presence-absence
data. Journal of Ecology 72:1055-1064.

With, K. A. 1994. Using fractal analysis to assess how species perceive landscape
structure. Landscape Ecology 9:25-36.

With, K. A., and T. O. Crist. 1995. Critical thresholds in species' responses to landscape
structure. Ecology 76:2446-2459.

With, K. A., and T. O. Crist. 1996. Translating across scales: Simulating species
distributions as the aggregate response of individuals to heterogeneity. Ecological
Modelling 93: 125-137.

With, K. A., R. H. Gardner, and M. G. Turner. 1997. Landscape connectivity and
population distributions in heterogeneous environments. Oikos 78:151-169.

Wu, J. 2004. Effects of changing scale on landscape pattern analysis: Scaling relations.
Landscape Ecology 19:125-138.

216

 

Wu, J ., and R. J. Hobbs. 2002. Key issues and research priorities in landscape ecology:
An idiosyncratic synthesis. Landscape Ecology 17:355-365.

Wu, J ., D. E. Jelinski, M. Luck, and P. T. Tueller. 2000. Multiscale analysis of landscape
heterogeneity: Scale variance and pattern metrics. Geographic Information
Sciences 626-19.

Wu, J ., and Y. Qi. 2000. Dealing with scale in landscape analysis: an overview.
Geographic Information Sciences 621-5.

Wu, J., W. Shen, W. Sun, and P. T. Tueller. 2002. Empirical patterns of the effects of
changing scale on landscape metrics. Landscape Ecology 17:761-782.

Zheng, D., D. O. Wallin, and Z. Hao. 1997. Rates and patterns of landscape change

between 1972 and 1988 in the Changbai Mountain area of China and North
Korea. Landscape Ecology 12:241-254.

217

 

111111111111111111111111

293 02736 4383

     

IIIIIIIII