-
extracted text
-
MODELING COLUMBIAN SHARP-TAILED GROUSE LEK
OCCUPANCY TO GUIDE SITE SELECTION FOR
TRANSLOCATIONS AND SPECIES POPULATION RECOVERY
by
Stacey A. Plumley
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2014
©2014 by Stacey A. Plumley. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Stacey A. Plumley
has been approved for
The Evergreen State College
by
________________________
Dina Roberts, Ph. D.
Member of the Faculty
________________________
Date
ABSTRACT
Modeling Columbian Sharp-tailed Grouse lek
occupancy to guide site selection for
translocations and species population recovery
by Stacey A. Plumley
A fundamental step in conserving biodiversity is identification of existing habitat
areas that could potentially sustain populations of target species. Animals can be
translocated to suitable habitat to re-establish populations and expand the target
species’ range. Columbian Sharp-tailed Grouse (Tympanuchus phasianellus
columbianus) (hereafter STG) were historically abundant in eastern Washington
before native vegetation was converted to agriculture and livestock grazing in the
early 20th century. Currently, there are only seven isolated STG populations in
Washington that total less than 1,000 birds. This study identified potential habitat
for STG translocations within their historical range in Washington. Two logistic
regression models were used to compare the influence of environmental variables
on STG occupancy at active (n = 40) and inactive lek complexes (n = 41). Each
environmental variable was assessed at three scales from the leks: typical STG
flight distance (1 km radius), spring/summer habitat area (3 km radius), and
winter habitat area (10 km radius). Habitat identified by the first model was
analyzed for habitat patch metrics at the three scales. The patch metrics, percent
area of habitat and maximum habitat patch size, were included with the original
environmental variables to develop the second model. Both models selected for
mean elevation at two scales (10 and 1 km) and for road density (1 km). The first
model also included percent grassland habitat (3 km) and the second model
included percent habitat area (3 km). Ten potential habitat areas greater than
50,000 ha were identified. These areas were located on the periphery of the STG
historical range, at higher elevations, similar to habitat locations of extant STG
populations. Some potential habitat contained areas of agriculture, forest, and
steep slopes–these areas would not be suitable for STG and further on-the-ground
assessment is needed to determine the overall potential of the habitat.
TABLE OF CONTENTS
LIST OF FIGURES ............................................................................................... vi
LIST OF TABLES ............................................................................................... viii
ACKNOWLEDGEMENTS ................................................................................... ix
CHAPTER 1 LITERATURE REVIEW ................................................................. 1
INTRODUCTION ............................................................................................... 1
LOSS OF WILDLIFE BIODIVERSITY ............................................................ 3
EFFORTS TO PROTECT AND RESTORE WILDLIFE BIODIVERSITY ..... 6
The U.S. Endangered Species Act ................................................................... 6
Habitat Restoration, Translocations, and Habitat Models ............................... 7
COLUMBIAN SHARP-TAILED GROUSE ...................................................... 9
Habitat Range and Population ....................................................................... 10
Emerging Threats .......................................................................................... 15
Efforts to Recover Columbian Sharp-tailed Grouse Populations in
Washington .................................................................................................... 17
CONCLUSION ................................................................................................. 23
CHAPTER 2 ARTICLE MANUSCRIPT ............................................................. 26
ABSTRACT ...................................................................................................... 26
INTRODUCTION ............................................................................................. 27
METHODS........................................................................................................ 32
Study Area ..................................................................................................... 32
Columbian Sharp-tailed Grouse Database ..................................................... 35
Scale of Analysis ........................................................................................... 36
Environmental Variables ............................................................................... 37
Model Development ...................................................................................... 39
Model Assessment and Validation ................................................................ 41
RESULTS.......................................................................................................... 42
DISCUSSION ................................................................................................... 52
iv
Columbian Sharp-tailed Grouse Population Goals and Habitat Requirements
....................................................................................................................... 52
Final Model Selection .................................................................................... 54
Analysis of Potential Columbian Sharp-tailed Grouse Habitat ..................... 56
Modeling Considerations ............................................................................... 60
MANAGEMENT RECOMMENDATIONS .................................................... 62
CHAPTER 3 ADDITIONAL RECOMMENDATIONS ...................................... 64
FURTHER ANALYSIS OF POTENTIAL COLUMBIAN SHARP-TAILED
GROUSE HABITAT IN OKANOGAN COUNTY ......................................... 64
Land Cover .................................................................................................... 64
Land Ownership ............................................................................................ 67
Challenges and Opportunities ........................................................................ 68
Recommendations and Conclusions .............................................................. 69
IMPROVING THE COLUMBIAN SHARP-TAILED GROUSE MODEL .... 72
REFERENCES ..................................................................................................... 75
APPENDIX A LEK COMPLEX DATA .............................................................. 84
APPENDIX B LAND COVER VARIABLE CATEGORIES ............................. 87
APPENDIX C SOURCES AND DESCRIPTIONS OF LAND COVER
CLASSES ............................................................................................................. 88
v
LIST OF FIGURES
Figure 1. Photo of Columbian Sharp-tailed Grouse.............................................. 10
Figure 2. Current and Historical Range of Columbian Sharp-tailed Grouse in
North America. ..................................................................................................... 11
Figure 3. Current and Historical Range of Columbian Sharp-tailed Grouse in
Washington. .......................................................................................................... 13
Figure 4. The Study Area ...................................................................................... 34
Figure 5. Phase I Model Receiver Operating Characteristic Curve. ..................... 44
Figure 6. Phase I Model Overlay Plot of Sensitivity and Specificity Versus
Probability Cutoffs. ............................................................................................... 46
Figure 7. 315 Phase I Model Columbian Sharp-tailed Grouse Potential Habitat
Patches with a Total Area of 2,169,954 ha. .......................................................... 48
Figure 8. Phase II Model Receiver Operating Characteristic Curve. ................... 49
Figure 9. Phase II Model Overlay Plot of Sensitivity and Specificity Versus
Probability Cutoffs. ............................................................................................... 51
Figure 10. 316 Phase II Model Columbian Sharp-tailed Grouse Potential Habitat
Patches with a Total Area of 1,739,212 ha. .......................................................... 52
Figure 11. Ten Habitat Patches Identified by the Final Model Greater Than 50,000
ha. .......................................................................................................................... 54
Figure 12. Mean Annual Precipitation and Columbian Sharp-tailed Grouse
Potential Habitat Identified by the Phase I and II Models. ................................... 56
vi
Figure 13. A Composite of Two Model Habitat Patches in Okanogan County,
Greater Than 50,000 ha, and Adjacent Habitat Patches within Two Kilometers. 66
vii
LIST OF TABLES
Table 1. The Distribution of Current Populations of Columbian Sharp-tailed
Grouse in Washington........................................................................................... 14
Table 2. Land Ownership within the Historical and Current Range for Columbian
Sharp-tailed Grouse in Eastern Washington ......................................................... 35
Table 3. Summary of GIS Predictor Variables Used for Columbian Sharp-tailed
Grouse Modeling. ................................................................................................. 38
Table 4. Phase I Final Candidate Models Selection Criteria. ............................... 43
Table 5. Phase I Columbian Sharp-tailed Grouse Occurrence Model Wald z
Statistic Results. .................................................................................................... 45
Table 6. Phase I and Phase II Models Verification Test Scores. .......................... 49
Table 7. Phase II Columbian Sharp-tailed Grouse Occurrence Model Wald z
Statistic Results. .................................................................................................... 50
Table 8. A Comparison of Percent Area for Different Land Covers Within
Columbian Sharp-tailed Grouse Current Range and the Final Model Habitat
Areas. .................................................................................................................... 60
Table 9. A Comparison of Percent Land Cover for Current Columbian Sharptailed Grouse Habitat Range and Two Composite Potential Habitat Areas in
Okanogan Countyl. ............................................................................................... 67
Table 10. A Comparison of Land Ownership for Two Composite Potential Habitat
Areas in Okanogan County. .................................................................................. 68
viii
ACKNOWLEDGEMENTS
I would like to thank Joanne Schuett-Hames for setting this project in motion and
for her enthusiasm and encouragement throughout. This project would not have
been possible without the help and support I received from Washington
Department of Fish and Wildlife (WDFW) biologists, Michael Schroeder and
Derek Stinson, and wildlife biologist, Leslie Robb. Thanks for providing me with
the Columbian sharp-tailed grouse lek complex database, and for your thoughtful
answers to my many questions. Many thanks to Brian Cosentino (WDFW) for
helping me track down Washington Wildlife Habitat Connectivity Group GIS
data and explaining the many intricacies that went into data development. A
special thanks to Andrew Shirk, University of Washington Climate Impacts
Group, for his willingness to share his modeling expertise. I would also like to
thank my thesis advisor, Dr. Dina Roberts for her enthusiasm and continued faith
in me that I would be able to complete this herculean task on time. Finally, I am
deeply grateful to my husband, Pete, for his unfailing support and encouragement
from my first MES class to the final page of my thesis.
ix
CHAPTER 1 LITERATURE REVIEW
INTRODUCTION
The current rate of species extinction due to human caused habitat loss,
fragmentation, and degradation is at least 100 times the rate of extinction
characterized by the fossil record (MEA 2005). According to the International
Union for Conservation of Nature and Natural Resources’ Red List of Threatened
Species, 41% of amphibian, 25% of mammal, and 13% of bird species are at high
risk for extinction (IUCN 2013). Many imperiled wildlife species and associated
habitats are actively managed in an attempt to increase threatened populations and
ensure their survival (Frankham, Ballou, and Briscoe 2002). In the U.S., the
Endangered Species Act and state laws protect imperiled species and require
actions to increase species’ populations such as species monitoring, management,
and habitat restorations (USFWS 2013).
In Washington State, Columbian Sharp-tailed Grouse (Tympanuchus
phasianellus columbianus) (hereafter STG) is a state listed threatened species.
STG were once an abundant game bird in eastern Washington until significant
portions of their habitat were converted to agriculture and livestock grazing in the
early 20th century (Yocom 1952). Currently, there are less than 1,000 STG in
Washington occurring in seven isolated populations (Stinson and Schroeder
2012). Efforts to increase the overall population numbers in Washington have
included habitat restoration and management, and translocation of birds from
larger populations to augment existing populations (Schroeder et al. 2012; Stinson
and Schroeder 2012). Additional translocations are planned to re-establish STG
1
populations within their historical range in Washington but the best locations for
the translocations still need to be identified (M. Schroeder, personal
communication, October 8, 2013). This thesis research was undertaken to identify
potential translocation sites by comparing the influence of environmental
variables on STG occupancy at current habitat areas and historical, unoccupied
habitat areas (Tack 2006; Aldridge et al. 2008).
The purpose of the literature review is to examine the relevant research
related to this thesis beginning with a brief overview of human land use change
and the subsequent negative impacts to wildlife populations. Efforts to protect and
restore imperiled species in the U.S. with enactment of the Endangered Species
Act and implementation of species recovery plans is discussed next. The last
section of the literature review focuses on STG and how this species and its
habitats are being managed in Washington to increase and restore healthy
populations. The last section concludes with a review of STG habitat studies to
summarize the current knowledge of the resource needs of the species and provide
a foundation for this research.
In addition to the literature review, this thesis contains two other chapters.
The second chapter is a complete account of the thesis research presented in a
manuscript format including a brief summary of the literature review and the
research methods and results, discussion, and management recommendations. The
third chapter provides further analysis of two potential habitat areas in Okanogan
County that were identified by the model and discusses ways to improve the STG
model.
2
LOSS OF WILDLIFE BIODIVERSITY
Globally, the most direct driver of biodiversity loss in terrestrial
ecosystems has been land use change resulting from the expansion of human
populations and activities (Sala et al. 2000; Gaston, Blackburn, and Goldewijk
2003; Foley et al. 2005; MEA 2005). As humans settle new areas, native
vegetation is converted to agriculture, pastures, or other uses to provide food, fuel,
shelter, and other natural resources. (MEA 2005; Primack 2010). Today,
agriculture and livestock pastures are the largest terrestrial land use, occupying
approximately 40% of the global land surface (Ramankutty, Foley, and
Olejniczak 2002; Asner, Elmore, Olander, Martin, and Harris 2004). Land use
change destroys and degrades habitat for many wildlife species which can
negatively impact their populations (Pimm and Raven 2000). In areas with higher
habitat loss, species show declining trends in global abundance compared to
species with increasing or stable trends for habitat (Donovan and Flather 2002).
When habitat is lost from land use modifications, the quality of the remaining
habitat is affected as it becomes fragmented into smaller patches that are
surrounded by a matrix of different land uses such as agriculture or development
(Wilcove, McLellan, and Dobson 1986; van den Berg, Bullock, Clarke, Langston,
and Rose 2001). The amount of habitat in these patches is further reduced by edge
effects, especially for specialist species that have specific habitat needs. Edge
effects extend up to 100 m into habitat patches and include increased predation,
increased numbers of invasive plant species, and additional changes to plant
3
community structure from differences in moisture, temperature, and sunlight
(Odell 2003; MEA 2005; Primack 2010).
Habitat loss and fragmentation may limit the ability of animals to move
among habitat patches to find food, mates, shelter, and other resources they need
(Crooks and Sanjayan 2006). A combination of the composition and configuration
of the landscape, a wildlife species’ ecological requirements, and its dispersal
ability, determines how well an animal is able to move among habitat patches in a
landscape (Crooks and Sanjayan 2006). This combination of factors establishes
the structure of wildlife populations in fragmented landscapes. Frankham (2006)
described several different wildlife population structures such as, source and sink,
stepping stones, metapopulations, and isolated populations. Source and sink,
stepping stones, and metapopulations are characterized by animals with medium
to high dispersal abilities that are able to move among habitat patches. Source and
sink wildlife populations structures are similar to a mainland and island structure
where animals on the island migrate from the mainland. Stepping stones wildlife
population structures consist of neighboring wildlife populations that are able to
migrate among habitat patches. Metapopulation structure consists of random
cycles of colonization, extinction, and recolonization of wildlife populations in
small fragmented habitat areas. Larger subpopulations of wildlife species provide
the source of migrants to recolonize the small areas of habitat. Dispersal among
subpopulations is an important component of managing wildlife species in
fragmented landscapes. Finally, some wildlife populations may become isolated
on “islands” of habitat. These wildlife populations are unable to move among
4
habitat patches because of the surrounding land use matrix, low dispersal abilities,
and/or special habitat needs that prevent them from dispersing in highly
fragmented landscapes (MEA 2005).
Small, isolated wildlife populations are at a greater risk for extinction due
to the combined effects of habitat loss, low genetic diversity, and environmental
variability or stochastic environmental events (Gilpin and Soule 1986; Pimm and
Raven 2000; Frankham 2005). Isolated wildlife populations may be at a higher
risk for developing low genetic diversity over time (Frankham 2006). Low genetic
diversity can result in inbreeding depression which is characterized by low mating
success, higher offspring mortality, and offspring that are weak or sterile
(Frankham et al. 2002). In addition, low genetic diversity can negatively affect the
ability of a population to adapt to short-term and long-term environmental change.
Short-term environmental stochastic events such as changes in the number of
predators, disease organisms, abundance of food, weather, etc, can lead to
fluctuations in population size (Frankham et al. 2002; Frankham 2005; Primack
2010). Species that have larger population size and genetic diversity have a
greater ability to recover from reductions in population size compared to small
populations with low genetic diversity (Frankham et al. 2002; Frankham 2005;
Primack 2010). Genetic diversity is also required for species to adapt to long-term
environmental change, such as climate change, through evolutionary processes.
Small populations with low genetic diversity have an increased risk of extinction
since they have little ability to evolve to cope with long-term environmental
change (Frankham et al. 2002; Primack 2010).
5
.
EFFORTS TO PROTECT AND RESTORE WILDLIFE BIODIVERSITY
The U.S. Endangered Species Act
“Under the Endangered Species Act, species may be listed as either
endangered or threatened. ‘Endangered’ means a species is in danger of
extinction throughout all or a significant portion of its range. ‘Threatened’
means a species is likely to become endangered within the foreseeable
future” (USFWS 2013).
To maintain or restore wildlife biodiversity, imperiled species and their
habitats are actively managed and protected under the U.S. Endangered Species
Act (ESA) and/or similar state laws. The ESA prohibits the harm of listed species
including harassing, wounding, killing, or capturing and also actions that
significantly alter a species’ habitat resulting in impairment of the species’ ability
to survive (USFWS 2013). The ultimate goal of the ESA is to recover a species so
that it no longer needs protection (USFWS 2013). In pursuit of this goal, the ESA
mandates that science based recovery plans are written for each listed species to
determine and implement the steps needed to restore and ensure the long-term
survival of the species (USFWS 2013). Wildlife populations that have
experienced significant reductions in their statewide historical range may be state
listed as threatened or endangered even if they are not listed under the ESA. For
example, in Washington State, a species may be listed “when populations are in
danger of failing, declining, or are vulnerable, due to factors including but not
restricted to limited numbers, disease, predation, exploitation, or habitat loss or
change” (WAC 232-12-297). The intent of the law is to protect and ensure the
survival of the listed species as free-ranging populations in Washington. In
support of that goal, the law requires a recovery plan for all listed species that
6
provides information on recovery goals for target populations and implementation
strategies for reaching those goals. Implementation strategies can include:
regulation, mitigation, acquisition, incentives and compensation mechanisms,
public education, and species monitoring.
Habitat restoration, translocations, and habitat models
Implementation of the ESA and state recovery plans for listed species
includes measures such as monitoring populations through field surveys,
assessing existing habitat conditions, and restoring and managing the species’
habitat (WAC 232-12-297; Stinson and Schroeder 2012; USFWS 2013). Land
acquisitions may also be undertaken to increase habitat areas or improve
connectivity of habitat patches (WAC 232-12-297; Stinson and Schroeder 2012).
In addition, small populations that are isolated and have low genetic diversity may
be augmented by translocated animals from larger, more genetically robust
populations (Stinson and Schroeder 2012). Translocated animals can significantly
increase genetic diversity and restore reproductive fitness to populations that
exhibit inbreeding depression (Westemeier et al. 1998; Frankham et al. 2002).
Ongoing monitoring of wildlife populations provides information such as,
population size, seasonal movement patterns, and nesting, foraging, and breeding
habitat preferences (Giesen 1997; McDonald 1998; Boisvert, Hoffman, and Reese
2005; Goddard, Dawson, and Gillingham 2009; Stonehouse 2013). Information
from monitoring can be used to create models that identify species’ resource
7
needs (Giesen 1997; Stonehouse 2013) and suitability of habitat (Goddard et al.
2009).
Habitat suitability models provide important information about species’
resource requirements based on the characteristics, amount, and spatial
arrangement of habitats that are selected by existing populations (Brambilla et al.
2009). Habitat suitability models are based on species’ life stages and seasonal
habitat requirements for food, cover, water, and reproduction (USFWS 1981).
Recent habitat suitability models used a geographic information system (GIS) to
combine geospatial data of species’ locations with environmental parameters.
These models can identify the ecological minimums that limit occupancy by
comparing the differences between currently occupied habitat areas and
unoccupied areas (Aldridge et al. 2008; Wisdom, Meinke, Knick, and Schroeder
2011; Knick, Hanser, and Preston 2013). Habitat suitability models can inform
land management decisions about conservation or restoration efforts for habitat
areas that are important for maintaining or increasing wildlife populations (Edgley
2001; Rittenhouse et al. 2008; Aldridge, Saher, Childers, Stahlnecker, and Bowen
2012; Stonehouse 2013). In addition, habitat areas can be assessed based on
model outcomes to determine suitability for re-introducing wildlife populations in
those areas (Ramsey, Black, Edgley, and Yorgason 1999; Edgley 2001;
Fitzpatrick 2003).
8
COLUMBIAN SHARP-TAILED GROUSE
Columbian Sharp-tailed Grouse (hereafter STG) are one of six existing
subspecies of Sharp-tailed Grouse in North America (Johnsgard 1973). STG are
the smallest of the subspecies and have darker gray plumage, more pronounced
spotting on the throat, and narrower markings on the underside (Figure 1)
(Connelly, Gratson, and Reese 1998). They typically walk or fly short distances
(0.4–0.8 km) but are capable of flying longer distances (3.2–4.8 km) (Stinson and
Schroeder 2012). STG gather in the spring at leks where males engage in
elaborate courtship displays to attract mates (Giesen and Connelly 1993; Stinson
and Schroeder 2012). Male and female STG have high fidelity to leks, often
returning to the same lek every spring although lek locations can also shift over
time or be abandoned (Stinson and Schroeder 2012).
9
Figure 1. Photo of Columbian Sharp-tailed Grouse. By Michael A. Schroeder
(WDFW).
Habitat range and population
Western North America
STG historically ranged from British Columbia, Washington, Oregon,
California, Nevada, Idaho, Utah, Montana, Colorado, and Wyoming and were a
plentiful and important game bird (Figure 2) (Yocom 1952; Aldrich 1963).
Settlement and conversion of native vegetation to agriculture and livestock
grazing in the early 20th century coincided with a massive reduction in STG
populations and habitat range (Yocom 1952). Today, they are the rarest
subspecies, having lost 90% of their historical range in Idaho, Utah, Wyoming,
and Washington (Bart 2000). Three large populations that occur on the border
10
between Colorado and Wyoming, Idaho and Utah, and within British Columbia
comprise over 93% of all STG (Bart 2000). These three populations are reported
to be either stable or increasing (Bart 2000). Small populations also exist in
Washington and were re-introduced in Nevada and Oregon where they had
previously been extirpated (Stinson and Schroeder 2012). Populations have also
been extirpated from California and Montana (Stinson and Schroeder 2012).
Figure 2. Current and historical range of Columbian Sharp-tailed Grouse in North
America.
11
Washington
STG were historically abundant and widely distributed in the grassland
steppe, meadow steppe, and the shrub-steppe ecosystems in eastern Washington
(Figure 3) (Yocom 1952; Schroeder, Hays, Murphy, and Pierce 2000; Stinson and
Schroeder 2012). The grassland steppe and meadow steppe ecosystems of the
Palouse Prairie in southeastern Washington was characterized by deep, loess soils
that supported native grassland vegetation dominated by Idaho fescue (Festuca
idahoensis), bluebunch wheatgrass (Agropyron spicatum), and Sandberg
bluegrass (Poa secunda) (Bunting, Kingery, and Schroeder 2003). The Palouse
was rapidly settled and by 1895, most of the tillable land had been converted to
agriculture (Stinson and Schroeder 2012). By the middle of the 20th century, STG
had been extirpated from the Palouse region of eastern Washington (Stinson and
Schroeder 2012). Native shrub-steppe vegetation communities in eastern
Washington predominately contained big sagebrush (Artemisia tridentata) and
three-tipped sagebrush (Artemisia tripartita) in association with bluebunch
wheatgrass (Agropyron spicatum) (Daubenmire 1988). Shrub-steppe communities
once covered most dryland areas of eastern Washington extending from the forest
edge of the North Cascades to the Palouse Prairie (Dobler, Eby, Perry,
Richardson, and Vander Haegen 1996). The shrub-steppe ecosystems also had
substantial populations of STG historically before widespread conversion of the
land to agriculture and livestock grazing in the early 20th century (Yocom 1952;
Schroeder et al. 2000; Stinson and Schroeder 2012). Early settlers grew wheat
12
using dryland farming techniques and livestock grazing occurred in areas that
were not suitable for farming (Stinson and Schroeder 2012).
Figure 3. Current and historical range of Columbian Sharp-tailed Grouse in
Washington.
The combination of agriculture, livestock grazing, and other land use
change in eastern Washington greatly decreased and fragmented native shrubsteppe habitat. Sagebrush cover in the historical range of STG in Washington
decreased from approximately 44.1% in 1900 to 15.6% by 1990 (McDonald and
Reese 1998). Mean habitat patch size, comprised of sagebrush and grassland
habitats, also decreased by 36%, from 4,474 ha in 1900 to 2,857 ha in 1990
13
(McDonald and Reese 1998). Currently, STG occur in seven isolated shrub-steppe
habitats in Douglas, Lincoln, and Okanogan counties on a total area of
approximately 217,300 ha or approximately 2.7% of their estimated historical
range in Washington (Figure 3) (Schroeder et al. 2000; Stinson and Schroeder
2012). Estimated numbers of STG in Washington totaled 916 in 2013 (Table 1)
(Stinson and Schroeder 2012).
Table 1. The distribution of current populations of Columbian Sharp-tailed
Grouse in Washington.
Est. Population
Size (2013)
Area (ha)
Density
Chesaw
50
7,000
0.007
Crab Creek/Swanson Lakes
98
52,100
0.002
Dyer Hill
80
30,800
0.003
Greenaway Spring
60
34,000
0.002
Nespelem
438
51,300
0.009
Scotch Creek
54
7,900
0.007
Tunk Valley
136
34,200
0.004
Total
916
217,300
Population
The distance between existing STG populations ranges from 22.28 km to
46.19 km with a mean of 28.16 km. STG can move further than 20 km (Boisvert
et al. 2005). However, the intervening matrix of croplands, roads, and
transmission lines can make movement between highly fragmented habitat more
difficult for STG (Robb and Schroeder 2012). Existing STG populations in
14
Washington are unable to readily disperse beyond their habitat areas to interbreed
with other populations or find new sources of food or cover. Their small
population size and isolation put them at risk of extirpation from Washington
(Bart 2000; Stinson and Schroeder 2012).
Emerging threats
Wind energy development
In addition to habitat loss and fragmentation from agriculture and
livestock grazing, wind energy development is an expanding land use in the arid
lands of western U.S. that could negatively impact STG populations (Robb and
Schroeder 2012; Stinson and Schroeder 2012). In eastern Washington, there are
1,527 existing turbines and 300 more under construction within the historical
range of STG (Stinson and Schroeder 2012). Wind energy development includes
turbines and associated infrastructure such as roads and transmission lines which
can have direct and indirect impacts on STG populations. Wind turbines are a
potential threat to STG populations through direct mortality from collisions
(Manville 2004) and indirect effects over time (Harju, Dzialak, Taylor, HaydenWing, and Winstead 2010). Roads can negatively impact STG populations due to
habitat fragmentation, road avoidance behavior, noise, and direct mortality
(Manville 2004; Pruett, Patten, and Wolfe 2009; Robb and Schroeder 2012;
Stonehouse 2013). Roads are also conduits for invasive plant species which can
further degrade areas of native vegetation (Gelbard and Belnap 2003). Roads and
transmission lines from energy development can be a source of direct mortality
15
from collisions (Wolfe, Patten, Shochat, Pruett, and Sherrod 2007) and potentially
increase habitat fragmentation from avoidance behavior (Manville 2004; Pruett et
al. 2009; Stonehouse 2013). In addition, unimproved roads, transmission lines,
and right-of-ways may increase predation of STG by providing corridors for
mammalian predators and increase perching opportunities for raptors (Pitman,
Hagen, Robel, Loughin, Applegate 2005; Wolfe et al. 2007).
Climate change
Climate change is predicted to exacerbate species extinction especially for
species like STG that are already at high risk for extinction due to small
population size, low dispersal ability, and special habitat needs (MEA 2005).
Improving species ability to move through the landscape is the most often
recommended adaptation strategy for climate change (Heller and Zavaleta 2009
and references therein). However, even with improved connectivity between
habitat patches, species may be unable to migrate fast enough to adjust to climate
change and may need to be translocated to more suitable habitat areas (Davis and
Shaw 2001; Heller and Zavaleta 2009 and references therein). Climate change
models are predicting significant change to sagebrush habitats including higher
summer temperatures, more variable and severe weather events, and wetter winter
seasons (Neilson, Lenihan, Bachelet, and Drapek 2005). Warming temperatures
could reduce the distribution of sagebrush and change the vegetation composition
to favor expansion of invasive species such as cheatgrass (Bromus tectorum) and
woody vegetation which is more fire prone (Neilson et al. 2005). Big sagebrush in
16
shrub-steppe ecosystems does not re-sprout after burning and must be recolonized by seed (Stinson and Schroeder 2012). Reductions in the amount of
sagebrush and/or more frequent fires could have significant impacts on STG and
other shrub-steppe species (Connelly, Schroeder, Sands, and Braun 2000; Bunting
et al. 2003).
Efforts to recover Columbian Sharp-tailed Grouse populations in Washington
Endangered Species Act listing
STG were petitioned for federal listing under the Endangered Species Act
(ESA) in 1995. The U.S. Fish and Wildlife Service (USFWS) concluded that
listing was not warranted because three extant, large populations of the species,
located on the border between Colorado and Wyoming, Idaho and Utah, and
within British Columbia, were not currently at increased risk of extirpation
(USFWS 2000). They also cited the active management of populations by state
and federal agencies such as, improving and restoring habitat and re-introducing
birds to unoccupied areas within their historical range (USFWS 2000). In 2004,
the Columbian subspecies was once again petitioned for federal listing but listing
was not found to be warranted because of the existing metapopulations that “have
persisted for the last several decades with no discernible downward trend”
(USFWS 2006). However, Bart (2000), in a status review for the USFWS,
predicted that most small populations of STG, like those in Washington, would
likely be extirpated in a decade or two without federal protection.
17
Washington State listing
“Threatened species are any wildlife species native to the state of
Washington that are likely to become endangered within the foreseeable
future throughout a significant portion of their range within the state
without cooperative management or removal of threats” (WAC 232-12297).
STG were listed as a state threatened species by the Washington Fish and
Wildlife Commission in 1998 (Stinson and Schroeder 2012). A State of
Washington Columbian Sharp-tailed Grouse Recovery Plan was completed in
2012. The Recovery Plan sets goals and objectives for recovery of STG in
Washington and provides detailed, science-based information on their biology,
life-cycle needs, and current population status. The goal of the Recovery Plan is
to “restore and maintain healthy populations of Columbian Sharp-tailed Grouse in
a substantial portion of the species’ historical range in Washington” (Stinson and
Schroeder 2012). STG will be considered for down-listing to a status of state
sensitive when there is a 10-year period with a recognized metapopulation that
averages greater than 2,000 birds and the total number of birds in Washington
averages a minimum of 3,200 (Stinson and Schroeder 2012). Populations that are
separate but genetically connected by periodic dispersers would be combined in
assessing total numbers and the viability of the population for down-listing to
state sensitive (Stinson and Schroeder 2012).
Conservation reserve program
The Conservation Reserve Program (CRP), administered by the U.S.
Department of Agriculture, pays farmers to take their lands out of agricultural
18
production to achieve conservation objectives, including reduced soil erosion and
the provision of wildlife habitat (Rodgers and Hoffman 2005; Schroeder and
Vander Haegen 2006). The CRP is a voluntary program established by the
Federal Food Security Act of 1985 that generally targets marginal agricultural
lands (Rodgers and Hoffman 2005). In eastern Washington, CRP land increased
from 22,257 ha in 1986 to more than 607,028 ha in 2011 (Stinson and Schroeder
2012). Conservation goals for wildlife have been taken into consideration for
newer CRP fields in Washington which have been planted with a mix of native
grasses and forbs (Stinson and Schroeder 2012). Older fields are also being
planted with native vegetation to provide important habitat for STG and other
shrub-steppe species (Stinson and Schroeder 2012). However, the long-term
status of CRP is uncertain because it is a voluntary program and dependent on
congressional renewal, enrollments are affected by economic factors like the price
of wheat (Stinson and Schroeder 2012). In the highly fragmented, agriculture
dominated landscape of eastern Washington, CRP lands provide important habitat
for STG (Schroeder and Vander Haegen 2006). If these lands were put back into
crop production, STG population could be severely impacted and the risk of
extirpation from Washington would likely increase (Bart 2000; Schroeder and
Vander Haegen 2006).
Land acquisitions, habitat restorations, and translocations
Washington Department of Fish and Wildlife (WDFW) has been working
toward meeting the Recovery Plan goal for STG by purchasing land, restoring
19
habitat, and translocating birds to augment existing populations in eastern
Washington. WDFW has purchased more than 16,187 ha in eastern Washington
for the protection of STG (Stinson and Schroeder 2012; WDFW 2012). On
WDFW Wildlife Areas that are managed for STG, riparian areas are being
restored with plantings of native shrubs and trees and former agriculture fields
have been planted with a mix of native grasses and forbs to improve habitat
(Stinson and Schroeder 2012; WDFW 2012). Under the Washington State Acres
for Wildlife Enhancement program (SAFE), WDFW biologists also work with
area landowners to enhance older CRP fields with native plantings and to get
additional acres enrolled in the CRP program (Stinson and Schroeder 2012;
WDFW 2012). Despite habitat restoration, land acquisitions, and management of
land for STG and other shrub-steppe species, STG populations continued to
decline until 2005 (Schroeder et al. 2012). A genetic analysis of the STG
population at Swanson Lakes Wildlife Area indicated that they had approximately
25% lower genetic diversity than birds from large populations in British
Colombia (Warheit and Schroeder 2003). To increase population size and genetic
diversity, 329 STG from British Columbia, Idaho, and Utah were translocated to
four existing populations in eastern Washington from 2005 to 2012 (Schroeder et
al. 2012). The translocated birds have experienced high mortality rates (47%) in
the Swanson Lakes area but populations have increased overall at lek sites
(Schroeder et al. 2012).
20
Population monitoring and habitat studies
Monitoring of STG populations in Washington has been ongoing since the
1950’s (Stinson and Schroeder 2012). WDFW conducts annual surveys of all
active STG leks and searches for new leks to assess overall population status
(Schroeder et al. 2012). Currently, there are 39 active leks, while 87 lek sites
documented since 1954 are now inactive (D. Stinson, personal communication,
December 4, 2013). The locations of translocated STG are also monitored with
radio collars (Stinson and Schroeder 2012). The location data from collared birds
and field survey data have been used to identify the home ranges, resource use,
and suitable habitat for STG in Washington (McDonald 1998; Stonehouse 2013).
Home ranges for STG include active leks and the surrounding spring and
summer nesting/brood rearing habitat and winter habitat. Female STG typically
stay within 1–2 km of leks in the spring and summer during nesting and brood
rearing (Meints 1991; Meints, Connelly, Reese, Sands, and Hemker 1992; Giesen
1997; Boisvert et al. 2005; Stonehouse 2013) and males stay within 2 km of leks
in the summer (Marks and Marks 1987; Boisvert et al. 2005; Stonehouse 2013).
However, movements of up to 4.4 km from leks in the spring and summer have
been observed in Washington (McDonald 1998). There is more variation in the
distance STG move to winter habitat which may be based on the availability of
suitable winter habitat (Giesen and Connelly 1993) or because of intraspecific
competition between males and females (McDonald 1998). Maximum distance
STG moved from nest and lek sites to winter habitat varied from 2 km in western
Idaho (Marks and Marks 1988) to 11.5 km in Washington (McDonald 1998).
21
Meints (1991) observed movements of 20 km to winter habitats in southeastern
Idaho and STG moved up to 41 km to find winter habitat in CRP and mine
reclamation areas in Colorado (Boisvert et al. 2005).
STG require a mix of grass and shrub habitats near leks for nesting and
brood rearing, and riparian or upland deciduous shrubs and trees within close
proximity for escape cover in the spring, summer, and fall and for cover and
forage in the winter (Marks and Marks 1987; Meints 1991, Meints et al. 1992;
Giesen and Connelly 1993; Giesen 1997; McDonald 1998). Hofmann and Dobler
(1988) surveyed winter habitat use by STG in Lincoln, Douglas, and Okanogan
counties, Washington and estimated a density of 1 bird to 3 ha of riparian and
deciduous habitat. Riparian and upland deciduous trees and shrubs are a vital
component of STG habitat and are limited in shrub-steppe ecosystems due to
excessive grazing and the effects of conversion to cropland on hydrology
(Hofmann and Dobler 1988; Giesen and Connelly 1993; Stinson and Schroeder
2012). Marks and Marks (1987) reported that livestock grazing resulted in
degradation to riparian vegetation from trampling, browsing, and rubbing.
STG will nest under grasses or shrubs in steppe, meadow-steppe, and
shrub-steppe habitat types (Marks and Marks 1987; Meints 1991; McDonald
1998). In Colorado, females selected dense clumps of shrubs for nesting in
mountain shrub habitats (Giesen 1997) and big sagebrush and low sagebrush in
Idaho (Marks and Marks 1987; Meints 1991). In Washington and Montana,
grassland habitat with sparse shrub cover was selected for nesting and nests were
located under a variety of perennial grasses in Washington (Cope 1992;
22
McDonald 1998; Stonehouse 2013). CRP fields also provide suitable habitat for
nesting and brood rearing (Meints 1991; Edgley 2001; McDonald 1998;
Stonehouse 2013). In Washington, STG primarily selected grassy CRP fields for
nesting under bunchgrasses and CRP fields with sparse shrub cover for lek
locations (McDonald 1998; Stonehouse 2013).
STG diet includes grasses, seeds, forbs, and insects from spring through
fall and berries, buds, and catkins of shrubs and trees during the winter (Evans
and Dietz 1974; Marks and Marks 1987; Giesen 1997). Standing wheat or spilled
grain in agricultural fields is also an important fall and winter food source in some
locations (Meints 1991; Meints et al. 1992; McDonald 1998; Stinson and
Schroeder 2012).
Additional environmental features such as land ruggedness, slope,
elevation, and development also affect STG habitat selection. STG select less
rugged areas (Stonehouse 2013), slopes that are less than 30%, and elevations
between 300 m and 1350 m in Washington (Stinson and Schroeder 2012) and less
than 2200 m in Idaho for their home ranges, nest, and lek sites (Marks and Marks
1987; Ramsey et al. 1999). STG also avoid roads, distribution lines, and trees
(spring-summer) within their home ranges in Washington (Stonehouse 2013).
CONCLUSION
Habitat loss and fragmentation from human land use change have directly
reduced the biodiversity of Earth’s terrestrial ecosystems and these declines
continue to be dramatic (Sala et al. 2000; Gaston et al. 2003; Foley et al. 2005;
23
MEA 2005). Small, isolated populations in highly fragmented landscapes are at
high risk for localized extirpation and species’ extinction due to a combination of
genetic and stochastic events that has been called a “vortex of extinction” (Gilpin
and Soule 1986). Management of imperiled species and associated habitats may
reverse declines and restore more resilient populations. Wildlife management
strategies include habitat assessment, ecological restoration, land management,
land acquisition, and translocations of animals from more robust populations to
smaller populations in order to increase population size and genetic diversity
(Stinson and Schroeder 2012).
Populations of STG in Washington declined significantly after the
introduction and expansion of agriculture and livestock grazing in the early 20th
century greatly reduced and fragmented their habitat (Yocom 1952). The existing
small, isolated populations of STG in Washington are actively managed with the
goal of restoring healthy populations within a substantial portion of their
historical range (Stinson and Schroeder 2012). The Washington State Columbian
Sharp-tailed Grouse Recovery Plan specifies that a habitat suitability model be
developed to identify priority areas for habitat enhancement or restoration, and
areas of suitable habitat for re-establishing additional STG grouse populations
with translocations (Stinson and Schroeder 2012). Previous translocations of STG
from larger populations appear to have stabilized and slightly increased STG
populations in Washington (Schroeder et al. 2012). The next step is to identify
areas of suitable habitat within their historical range in Washington and
24
translocate birds to re-establish populations in those areas (Stinson and Schroeder
2012).
The goal of this study is to identify potential areas of STG habitat that may
be suitable for future translocations by modeling the influence of biotic, abiotic,
and anthropogenic variables on STG occupancy at active and inactive lek
complexes (Tack 2006; Aldridge et al. 2008).
25
CHAPTER 2 ARTICLE MANUSCRIPT
Formatted for submission to Ecological Applications
Modeling Columbian Sharp-tailed Grouse lek
occupancy to guide site selection for
translocations and species population recovery
ABSTRACT
A fundamental step in conserving biodiversity is identification of existing habitat
areas that could potentially sustain populations of target species. Animals can be
translocated to suitable habitat to re-establish populations and expand the target
species’ range. Columbian Sharp-tailed Grouse (Tympanuchus phasianellus
columbianus) (hereafter STG) were historically abundant in eastern Washington
before native vegetation was converted to agriculture and livestock grazing in the
early 20th century. Currently, there are only seven isolated STG populations in
Washington that total less than 1,000 birds. This study identified potential habitat
for STG translocations within their historical range in Washington. Two logistic
regression models were used to compare the influence of environmental variables
on STG occupancy at active (n = 40) and inactive lek complexes (n = 41). Each
environmental variable was assessed at three scales from the leks: typical STG
flight distance (1 km radius), spring/summer habitat area (3 km radius), and
winter habitat area (10 km radius). Habitat identified by the first model was
analyzed for habitat patch metrics at the three scales. The patch metrics, percent
area of habitat and maximum habitat patch size, were included with the original
environmental variables to develop the second model. Both models selected for
mean elevation at two scales (10 and 1 km) and for road density (1 km). The first
model also included percent grassland habitat (3 km) and the second model
included percent habitat area (3 km). Ten potential habitat areas greater than
50,000 ha were identified. These areas were located on the periphery of the STG
historical range, at higher elevations, similar to habitat locations of extant STG
populations. Some potential habitat contained areas of agriculture, forest, and
steep slopes–these areas would not be suitable for STG and further on-the-ground
assessment is needed to determine the overall potential of the habitat.
Key words: Columbian Sharp-tailed Grouse, Tympanuchus phasianellus
columbianus, translocations, logistic regression, geographic information systems,
GIS, habitat modeling
26
INTRODUCTION
Globally, the most direct driver of biodiversity loss in terrestrial
ecosystems has been land use change from the expansion of human populations
and human activities (Sala et al. 2000; Gaston et al. 2003; Foley et al. 2005; MEA
2005). Land use change destroys habitat for many wildlife species and fragments
the remaining habitat into smaller patches that are surrounded by a matrix of
different land uses such as agriculture or development (Wilcove et al. 1986; van
den Berg et al. 2001). Habitat loss and fragmentation can limit the ability of
animals to move among habitat patches to find food, mates, shelter, and other
resources they need (Crooks and Sanjayan 2006). Species with low dispersal
abilities or special habitat needs may be unable to move among habitat patches
and become isolated in highly fragmented landscapes (MEA 2005). Small,
isolated wildlife populations are at a greater risk for extinction due to the
combined effects of habitat loss, low genetic diversity, short-term environmental
variability, and long-term environmental change (Gilpin and Soule 1986; Pimm
and Raven 2000; Frankham 2005).
Species recovery efforts for small wildlife populations include
translocating animals from larger populations to increase the size and genetic
diversity of small populations (Frankham et al. 2002; Stinson and Schroeder
2012). Translocations can also be used to re-establish wildlife populations in areas
where they were previously extirpated. Populations of several grouse species have
been successfully re-established including, Greater Sage-grouse (Centrocercus
urophasianus) in Washington, Columbian Sharp-tailed Grouse (Tympanuchus
27
phasianellus columbianus) in Idaho, Plains Sharp-tailed Grouse (Tympanuchus
phasianellus jamesi) in Kansas, and Greater Prairie Chickens (Tympanuchus
cupido) in Illinois and Iowa (Snyder, Pelren, and Crawford 1999). The first step to
re-establishing wildlife populations is to identify potential habitat areas for
translocations. One method to identify potential habitat is to create a habitat
suitability model that combines geospatial data of environmental parameters with
occupied and unoccupied habitat areas to identify variables that best predict for
occupancy. The model parameters can be applied to the spatial data layers in a
geographic information system (GIS) to map potential habitat areas (Aldridge et
al. 2008; Wisdom et al. 2011; Knick et al. 2013). Landscapes can be assessed
based on model outcomes to determine suitability for re-introducing wildlife
populations in those areas (Ramsey et al. 1999; Edgley 2001; Fitzpatrick 2003). In
addition, habitat models can inform land management decisions about habitat
conservation or restoration of critical habitat areas that will maintain or increase
wildlife populations (Edgley 2001; Rittenhouse et al. 2005; Aldridge et al 2012;
Stonehouse 2013).
In Washington, Columbian Sharp-tailed Grouse (Tympanuchus
phasianellus columbianus) (hereafter STG) is a state listed threatened species.
STG were once an abundant game bird in eastern Washington until significant
portions of their habitat were converted to agriculture and degraded by livestock
grazing in the early 20th century (Yocom 1952). A conservative estimate by
Washington Department of Fish and Wildlife (WDFW), assigned pre-settlement
STG population size in Washington at greater than 100,000 birds (Stinson and
28
Schroeder 2012). Current estimated numbers of STG in Washington totaled 916
in 2013 (Stinson and Schroeder 2012).
Columbian Sharp-tailed Grouse (STG) are one of six existing subspecies
of Sharp-tailed Grouse in North America (Johnsgard 1973). STG are the smallest
of the subspecies and have darker gray plumage, more pronounced spotting on the
throat, and narrower markings on the underside (Connelly, Gratson, and Reese
1998). STG gather in the spring at leks where males engage in elaborate courtship
displays to attract mates (Giesen and Connelly 1993; Stinson and Schroeder
2012). Male and female STG have high fidelity to leks, often returning to the
same lek every spring (Stinson and Schroeder 2012). Home ranges for STG
include active leks and the surrounding nesting/brood rearing and winter habitat.
Female STG typically stay within 1–2 km of leks in the spring and summer during
nesting and brood rearing (Meints 1991; Meints et al. 1992; Giesen 1997;
Boisvert et al. 2005; Stonehouse 2013) and males stay within 2 km of leks in the
summer (Marks and Marks 1987; Boisvert et al. 2005; Stonehouse 2013).
However, movements of up to 4.4 km from leks in the spring and summer have
been observed in Washington (McDonald 1998). There is more variation in the
distance STG move to winter habitat which may be based on the availability of
suitable winter habitat (Giesen and Connelly 1993) or because of intraspecific
competition between males and females (McDonald 1998). Maximum distance
STG moved from nest and lek sites to winter habitat varied from 2 km in western
Idaho (Marks and Marks 1988) to 11.5 km in Washington (McDonald 1998).
Meints (1991) observed movements of 20 km to winter habitats in southeastern
29
Idaho and STG moved up to 41 km to find winter habitat in CRP and mine
reclamation areas in Colorado (Boisvert et al. 2005).
A variety of seasonal habitats within close proximity are required by STG
including dense grasses, forbs, or shrubs near leks for nesting and brood rearing in
the spring and summer, and riparian or upland deciduous shrubs and trees for
cover and forage in the winter (Marks and Marks 1987; Meints 1991, Meints et al.
1992; Giesen and Connelly 1993; Giesen 1997; McDonald 1998). Riparian and
upland deciduous trees and shrubs are a vital component of STG habitat and are
limited in shrub-steppe ecosystems due to excessive grazing and changes to
hydrology from land use conversion to cropland (Hofmann and Dobler 1988;
Giesen and Connelly 1993; Stinson and Schroeder 2012).
STG will nest under grasses or shrubs in steppe, meadow-steppe, and
shrub-steppe habitat types (Marks and Marks 1987; Meints 1991; McDonald
1998). In Washington, grassland habitat with sparse shrub cover was selected for
nesting and nests were located under a variety of perennial grasses in Washington
(McDonald 1998; Stonehouse 2013). CRP fields also provide suitable habitat for
nesting and brood rearing (Meints 1991; Edgley 2001; McDonald 1998;
Stonehouse 2013). In Washington, STG primarily selected grassy CRP fields for
nesting under bunchgrasses and CRP fields with sparse shrub cover for lek
locations (McDonald 1998; Stonehouse 2013).
The diet of STG consists of grasses, seeds, forbs, and insects from spring
through fall, and berries, buds, and catkins of shrubs and trees during the winter
(Evans and Dietz 1974; Marks and Marks 1987; Giesen 1997). Standing wheat or
30
spilled grain in agricultural fields is also an important fall and winter food source
in some locations (Meints 1991; Meints et al. 1992; McDonald 1998; Stinson and
Schroeder 2012).
Additional environmental features such as land ruggedness, slope,
elevation, and human infrastructure also affect STG habitat selection. STG select
less rugged areas (Stonehouse 2013), slopes that are less than 30%, and elevations
between 300 m and 1350 m in Washington (Stinson and Schroeder 2012) and less
than 2200 m in Idaho for their home ranges, nest, and lek sites (Marks and Marks
1987; Ramsey et al. 1999). STG also avoid roads, distribution lines, and trees
(spring-summer) within their home ranges in Washington (Stonehouse 2013).
The Washington State Columbian Sharp-tailed Grouse Recovery Plan
specifies that a habitat suitability model be developed to identify priority areas for
habitat enhancement or restoration, and areas of suitable habitat for reestablishing additional STG populations with translocations (Stinson and
Schroeder 2012). Previous translocations of STG from populations in British
Columbia, Idaho, and Utah, have stabilized and slightly increased STG
populations in Washington (Schroeder et al. 2012). The next step is to identify
habitat areas that are suitable for establishing new STG populations within their
historical range in eastern Washington (M. Schroeder, personal communication,
October 8, 2013). This study compared the influence of environmental variables
on the probability of occurrence at active and inactive STG lek complexes in
eastern Washington to identify potential habitat areas for translocations.
31
METHODS
Study area
The study area encompasses the historical range of STG that is within the
Columbia Plateau Ecoregion in eastern Washington (Figure 4). Pre-settlement
habitats of this area included grassland steppe, meadow steppe, and shrub-steppe
ecosystems (Yocom 1952; Schroeder et al. 2000; Stinson and Schroeder 2012)
(Figure 4). The grassland steppe and meadow steppe ecosystems of the Palouse
Prairie in southeastern Washington, were characterized by deep, loess soils that
originally supported native grassland vegetation dominated by Idaho fescue
(Festuca idahoensis), bluebunch wheatgrass (Agropyron spicatum), and Sandberg
bluegrass (Poa secunda) (Bunting et al. 2003). By 1895, most of the tillable land
in the Palouse had been converted to agriculture and STG were extirpated from
the region by the middle of the 20th century (Stinson and Schroeder 2012). Native
shrub-steppe vegetation communities in the Columbia Plateau region of eastern
Washington predominately contained big sagebrush (Artemisia tridentata) and
three-tipped sagebrush (Artemisia tripartita) in association with bluebunch
wheatgrass (Agropyron spicatum) (Daubenmire 1988). Shrub-steppe communities
once covered most dryland areas of eastern Washington extending from the
forested slopes of the North Cascades to the Palouse Prairie (Dobler et al. 1996).
This area also had substantial populations of STG historically before the area was
settled in the early 20th century (Yocom 1952; Schroeder et al. 2000; Stinson and
Schroeder 2012).
32
The combination of agriculture, livestock grazing, and other land use
change in eastern Washington greatly decreased and fragmented STG habitat.
Sagebrush cover in the historical range of STG in eastern Washington decreased
from approximately 44.1% in 1900 to 15.6% by 1990 (McDonald and Reese
1998). Sagebrush and grassland habitat patch size also decreased by 36%, from a
mean of 4,474 ha in 1900 to 2,857 ha in 1990 (McDonald and Reese 1998). The
loss and fragmentation of sagebrush habitat has also increased the distance
between patches and isolated extant populations of STG. Currently, STG occur in
seven shrub-steppe habitat areas in Douglas, Lincoln, and Okanogan Counties that
represent approximately 2.8% of their estimated historical range in Washington
(Schroeder et al. 2000; Stinson and Schroeder 2012) (Figure 4).
33
Figure 4. The study area. The study area was the historical range for STG in
Washington that was included within the extent of the Washington Wildlife
Habitat Connectivity Working Group’s Columbia Plateau Ecoregion spatial data
(Stinson and Schroeder 2012; WHCWG 2012b).
Land ownership within the historical and current range for STG is
comprised of private, tribal, and public lands (Table 2). The majority of lands in
STG historical (77.79%) and current (56.05%) range are privately owned.
(Stinson and Schroeder 2012). Tribal lands are the next greatest area of land
ownership within STG historic (8.51%) and current (28.12%) range, and the
remaining land areas are owned by federal, state, and other public entities such as
counties and universities (Stinson and Schroeder 2012). Within the current STG
34
range, the Colville Confederated Tribes owns the largest area of land (28.10%),
Washington State owns the next largest area including the Department of Fish and
Wildlife (6.90%) and the Department of Natural Resources (4.80%), and the other
areas of public land are owned by the U.S. Bureau of Land Management (4.08%)
and the U.S. Forest Service (0.04%) (Stinson and Schroeder 2012).
Table 2. Land ownership within the historical and current range for Columbian
Sharp-tailed Grouse in eastern Washington (Stinson and Schroeder 2012).
Historic Range
Land Owner or Manager
Private
Federal
State
Tribal
Other Public
Total
Current Range
Percent
Hectares
Percent
Hectares
77.79
5.48
8.07
8.51
0.16
3,925,001
276,260
406,990
429,168
8,044
56.05
4.12
11.71
28.12
0.00
121,047
8,898
25,290
60,718
0
100
5,045,463
100
215,953
Columbian Sharp-tailed Grouse database
Active and inactive STG lek complexes were used to predict the
probability of occurrence by comparing the influence of the surrounding
environmental parameters on occupancy (Appendix A). Lek complexes were
comprised of clusters of leks that were no further apart than 1 km (Schroeder et al.
2000). Leks are small areas, usually on knolls or ridges, where males gather in the
spring for elaborate courtship displays to attract mates (Giesen and Connelly
1993; Stinson and Schroeder 2012). Active and inactive lek complexes are a good
35
indicator of occupancy because male and female STG have high fidelity to leks,
often returning to the same lek every spring, although lek locations can shift over
time or be abandoned (Stinson and Schroeder 2012). Additionally, annual surveys
of leks by WDFW biologists have been conducted since 1954 and searches for
new lek sites have been conducted since 1970 (Schroeder et al. 2000; Stinson and
Schroeder 2012). Active lek complexes (1) were defined as those that had at least
one male displaying in the spring of 2013 (n = 40). Inactive lek complexes (0)
were selected to achieve approximately 50% prevalence (i.e., the frequency of
occurrence) and included lek complexes that were abandoned between 1994 and
2012 (n = 41). Studies have found that datasets which have prevalence of 50%
improved the predictive accuracy of a model (Manel, Williams, and Ormerod
2001; Liu, Berry, Dawson, and Pearson 2005).
Scale of analysis
Spatial scales were selected to represent the lifecycle habitat resource
needs of STG. Environmental variables were analyzed at three radii distances
from the lek complexes: 1, 3, and 10 km. One kilometer is a typical STG flight
distance (Stinson and Schroeder 2012), three kilometers is the average distance
from the lek that females move in the spring and summer for nesting and brood
rearing (McDonald 1998; Stonehouse 2013), and ten kilometers is the average
distance that male and female STG move from the lek to find suitable winter
habitat (McDonald 1998).
36
Environmental variables
Environmental variables that were known to impact STG populations and
other variables that may also influence STG occupancy of an area were selected
for the analysis based on a priori knowledge from a thorough literature review.
Candidate variables were divided into 3 broad categories: land cover, human
infrastructure, and physical geography (Table 3).
All of the spatial data layers, except for precipitation and percent slope,
were developed by the Washington Wildlife Habitat Connectivity Working Group
(WHCWG) for modeling select species’ habitat connectivity in the Columbia
Plateau Ecoregion (WHCWG 2012b). WDFW provided the WHCWG layers for
this study. WHCWG base layers were 30 m raster datasets and included land
cover (Appendix B), freeways, major highways, secondary highways, local roads,
four transmission line layers (< 230 KV, 1 line and 2 or more lines, and ≥ 230
KV, 1 line and 2 lines), ruggedness, soil depth, and elevation. The road density
layer (roads/ha) was created by combining the four WHCWG’s road category
layers. Precipitation was derived from the PRISM 30-Year Normals dataset which
is the average annual precipitation from 1981-2010 (PRISM 2014). The PRISM
data was resampled from 800 m to 30 m. Percent slope was calculated from a 30
m Digital Elevation Model (DEM) using ArcGIS Spatial Analyst.
Individual raster layers for each environmental variable were created using
a circular moving widow analysis at 1, 3, and 10 km radii. Percent area was
calculated for land cover and road density and mean values were calculated for
slope, elevation, ruggedness, soil depth, and precipitation. Individual Euclidean
37
distance raster layers were created to measure distances to the nearest agriculture
fields, pasture/hay fields, riparian/winter habitat, roads, and transmission lines.
Geographic Information Systems (GIS) processing was performed using ArcGIS
version 10.1 (ESRI 2012), Python 2.7 (Python 2010), and PythonWin 2.7.2
(Hammond 2011).
Table 3. Summary of GIS predictor variables used for Columbian Sharp-tailed
Grouse modeling. Significant variables were determined by a univariate logistic
regression analysis and are annotated for the scale(s) at which they were
significant.
Variable
Category
Land cover
Name
Description
Units
Source
ag³°
Agriculture
percent
gr¹³°
Grassland
percent
sh
Shrubland
percent
fr°
Forest
percent
ph°
Pasture hay (CRP)
percent
rw°
Riparian/winter habitat
percent
igr
Introduced grassland
percent
d_ag*
Distance to ag
m
d_ph*
Distance to ph
m
d_rw
Distance to rw
m
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
38
Table 3 (continued)
Variable
Category
Infrastructure
Physical
Geography
Name
Description
Units
Source
rd¹³°
Road density
roads/ha
d_mh*
Distance to major highway
m
d_sch
Distance to secondary highway
m
d_lr*
Distance to local roads
m
d_lt_1
Distance to transmission line less than 230KV 1 line
m
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
d_lt_2*
Distance to transmission line less than 230KV 2 lines
m
WHCWG
(2012b)
d_ge_1
Distance to transmission line greater or equal 230KV 1 line
m
WHCWG
(2012b)
d_ge_2
Distance to transmission line greater or equal 230KV 2 line
m
WHCWG
(2012b)
prc¹³°
Precipitation
cm
PRISM (2014)
slp³°
elv¹³°
Slope
Elevation
percent
m
rug°
Ruggedness
sd°
Soil depth
DEM
WHCWG
(2012b)
WHCWG
(2012b)
WHCWG
(2012b)
cm
The variable was significant in the univariate analysis: ¹1 km scale, ³3 km scale,
°10 km scale, and *distance
Model development
Two sets of models were developed (phase I and phase II) by comparing
the influence of environmental variables on STG occurrence at active and inactive
lek complexes in eastern Washington. The phase I and II models were developed
using a purposeful selection approach to select the most parsimonious model from
a set of candidate models (Hosmer, Lemeshow, and Sturdivant 2013). Both
39
models were developed using logistic regression to analyze the influence of
environmental variables at three scales, 1, 3, and 10 km radii distances from STG
active and inactive lek complexes. The suitable habitat areas identified by the
phase I model were analyzed using FRAGSTATS version 4 (McGarigal,
Cushman, and Ene 2012) to calculate habitat fragmentation metrics. The phase II
model included the original environmental variables and the habitat fragmentation
variables from the phase I model in the statistical assessment.
First, a univariate analysis on each environmental variable, at each scale
was conducted using Wald z statistic, P < 0.25 as the limit for inclusion in
candidate models (Hosmer et al. 2013). Two active lek complexes, 64 and 65,
were not used for the univariate analysis of variables at the 10 km scale because
the radius included areas that were outside the extent of the available spatial data
(Appendix A). All of the active and inactive lek complexes were used for the
univariate analysis at the1 km and 3 km radii. In addition, lek complexes, 64 and
65 were excluded from distance analysis because they were within 6 km of the
extent of the spatial data. The next closest lek complex was 27.06 km from the
extent of the spatial data. Therefore, a 27 km maximum buffer was applied to the
distance measurements.
All significant variables were assessed for correlation using Pearson’s r ≥
|0.70| as the cutoff value (Aldridge et al. 2008). All uncorrelated variables that
were significant in the univariate analysis were included in multivariable models
(Table 3). The least significant variables were sequentially dropped from the
model until all remaining variables were significant at Wald z statistic, P < 0.10
40
(Aldridge et al. 2008; Hosmer et al. 2013). All significant variables were also
assessed for linearity with a visual assessment of a lowess smoothing scatterplot
(Hosmer et al. 2013). Next, each uncorrelated variable that was not identified as
being significant in the univariate analysis was added one at a time to identify
variables that were important in combination with other variables (Hosmer et al.
2013). As variables were removed or added, P-values and the magnitude of
change in the coefficient was monitored to identify interactions between variables
and confounding variables for each candidate model (Hosmer et al. 2013).
Multicollinearity of the candidate models was checked using variance inflation
factors (VIF). Multicollinearity was considered a problem if VIF scores for
individual covariates were greater than 10 (Chatterjee and Hadi 2012).
Model assessment and validation
The candidate models were assessed and a final model selected that
surpassed the other models in more than one area. (Hosmer et al. 2013). Akaike
Information Criterion (AIC) was used to compare the fit of the candidate models
since they were derived from the same sample (Peng and So 2002). The overall
goodness of fit was assessed using a Pearson’s chi-square test which has a higher
power for small sample sizes (Hosmer et al. 2013). A Receiving Operating
Characteristic (ROC) curve was generated for each of the candidate models by
plotting the sensitivity versus 1-specificity over all possible cut off points to
assess the predictive ability of the models (Hosmer et al. 2013). The predictive
ability of a model was considered outstanding at an ROC of ≥ 0.9, excellent at ≥
41
0.8, acceptable at ≥ 0.7, and poor at < 0.7 (Hosmer et al. 2013). The final step in
assessing the candidate models was to check for overfitting of the model using kfold cross validation (10-fold). All statistical analysis was conducted in STATA
version 13 (StataCorp 2013).
Identifying potential Columbian Sharp-tailed Grouse habitat
The logistic regression model parameters for the final phase I and II
models were applied to the spatial predictors for each model to derive a
probability of occurrence map. Potential habitat areas were identified by applying
the model’s optimal threshold cutoff for occurrence, where sensitivity and
specific intersected, to the occurrence map. Areas that had values greater than or
equal to, the model’s cutoff were identified as potential STG habitat.
RESULTS
Many of the variables were highly correlated (r ≥ |0.70|) which limited the
combinations of variables that could be assessed in the multivariate analysis. For
example, mean elevation (10 km) was highly correlated with percent forest (10
km), mean precipitation (10 km), mean elevation (3 km), mean precipitation (3
km), and distance to pasture/hay fields. The only variables that were not highly
correlated with at least one other variable at all scales were road density and mean
soil depth.
Four phase I candidate models for STG occurrence were identified from
the multivariate analysis (Table 4). The variables in the models were all
42
statistically significant (P < 0.25) in the univariate analysis and not highly
correlated (Hosmer et al. 2013). All of the candidate models had a two-way
interaction between two of the variables that were significant at P < 0.05. The
interaction terms and the main effects variables were retained in the final
candidate models (Hosmer et al. 2013). Candidate model 2 was selected as the
final phase I model because it had the lowest Akaike Information Criterion (AIC)
score and was able to adequately discriminate between active and inactive lek
complexes based on the ROC score (Figure 5). Model 2 also had the second
lowest cross validation RMSE score which measures the difference between
values predicted by a model and the values actually observed (Tack 2006;
Aldridge et al. 2008; Hosmer et al. 2013).
Table 4. Phase I final candidate models selection criteria. Candidate models were
compared for the best model fit with Akaike Information Criterion (AIC),
predictive ability with Receiving Operating Characteristic (ROC) curve,
goodness of fit with Pearson’s chi-square test, and for over fitting with k-fold
cross validation.
Goodness of Fit
Model
Log
Likelihood
AIC
ROC
1
2
3
4
-46.833
-45.217
-46.705
-45.311
103.666
100.434
105.409
102.621
0.736
0.761
0.739
0.770
Pearson's
Prob>chi2
chi2
75.07
76.74
75.00
77.93
0.443
0.391
0.413
0.325
K-fold Cross
Validation (10
fold average)
RMSE
R2
0.487
0.470
0.495
0.469
0.247
0.234
0.219
0.247
43
1.00
Sensitivity
0.75
0.50
0.25
0.00
0.00
0.25
0.50
0.75
1.00
1 - Specificity
Area under ROC curve = 0.7606
Figure 5. Phase I model Receiver Operating Characteristic curve.
The phase I model environmental variables that best predicted STG
occurrence at active and inactive lek complexes included, mean elevation (10
km), percent area of native grass habitat (3 km), road density (1 km), and a twoway interaction between mean elevation (10 km) and percent area of grass habitat
(3 km) (Table 5). The threshold for STG occurrence was estimated by plotting the
final model sensitivity and specificity against probability cutoffs. The intersection
of sensitivity and specificity is the optimal cutoff that minimizes false negatives
and false positives in correctly identifying STG occurrence at lek complexes
(Peng and So 2002). For this model, the optimal cutoff (0.51682) yielded 71.05%
correct classifications for sensitivity and 70.73% for specificity (Figure 6).
44
Table 5. Phase I Columbian Sharp-tailed Grouse occurrence model Wald z
statistic results. Estimated coefficients (β), standard errors (SE), intercepts (z), Pvalues (P) and 95% confidence intervals (95% CI).
β
SE
z
P
Mean elevation
(10 km)
0.030
0.011
2.650
0.008
0.008
0.052
% grassland (3
km)
36.376
15.948
2.280
0.023
5.118
67.635
Road density (1
km)
-94.705
41.235
-2.300
0.022
-175.524
-13.886
Interaction of
elevation &
grassland
-0.051
0.022
-2.310
0.021
-0.094
-0.008
Variable
95 % CI
45
1.00
0.75
0.50
0.25
0.51682
0.00
0.00
0.25
0.50
0.75
1.00
Probability cutoff
Sensitivity
Specificity
Figure 6. Phase I model overlay plot of sensitivity and specificity versus
probability cutoffs. The proportion of lek complexes that were correctly classified
as active or inactive is equal at the optimal cutoff where sensitivity and specificity
intersect (Peng and So 2002). For this model, the optimal cutoff (0.521682)
yielded approximately 71% correct classifications for both groups.
The phase I model logistic regression parameters were applied to the
spatial data layers for each environmental variable in the model to create a
probability of occurrence map. Probability of occurrence map raster values that
were greater than or equal to, the phase I model probability cutoff (0.51682), were
identified as potential STG habitat. The phase I model identified 315 potential
habitat patches with a total area of 2,169,954 ha (Figure 7). The habitat patches
46
had a mean area of 6,889 ha, a median area of 810 ha, and an area range of 0.81
ha to 466,182 ha.
Habitat patch metrics were calculated from the phase I model habitat
areas. A circular moving window analysis was used to calculate percent habitat
area and maximum patch area at 1, 3, and 10 km radii. Several of the variables
were highly correlated (r ≥ |0.70|), percent area (1 and 10 km) were both
correlated with percent area (3 km), and maximum area (3 km) was correlated
with maximum area (10 km). A univariate analysis of the habitat patch metrics
resulted in significant values at P < 0.25 for percent habitat area (1, 3, and 10 km)
and maximum patch area (1 km).
47
Figure 7. 315 Phase I model Columbian Sharp-tailed Grouse potential habitat
patches with a total area of 2,169,954 ha.
A phase II model was developed from multiple logistic regression analysis
of the original environmental variables and the phase I model habitat patch
metrics. The phase II model variables were all significant (P < 0.25) in the
univariate analysis and there were no significant interactions between the
variables (Hosmer et al. 2013). The phase I and phase II models had very similar
model verification test scores (Table 6). The phase I model had a slightly better
(lower) AIC score but the phase II model had a slightly higher predictive ability
(ROC = 0.765) (Figure 8) and cross-validation RMSE score.
48
Table 6. Phase I and phase II models verification test scores. The phase I model
had a slightly higher AIC score but the phase II model had a slightly higher
predictive ability (ROC = 0.765) and cross-validation RMSE score.
Goodness of Fit
Model
Log
Likelihood
AIC
ROC
Pearson's
Prob>chi2
chi2
Phase I
-45.217
100.434
0.761
76.74
Phase II
-46.705
100.532
0.765
77.23
K-fold Cross
Validation (10 fold
average)
RMSE
R2
0.391
0.470
0.234
0.471
0.469
0.240
1.00
Sensitivity
0.75
0.50
0.25
0.00
0.00
0.25
0.50
0.75
1.00
1 - Specificity
Area under ROC curve = 0.7646
Figure 8. Phase II model Receiver Operating Characteristic curve.
49
The phase II model variables that best predicted STG occurrence at active
and inactive lek complexes included percent area of suitable habitat (3 km), as
defined by the phase I model, average elevation (1 km), and road density (1 km)
(Table 7). Only one of the variables in the phase II model, road density, was also
included in the phase I model. The phase II model optimal cutoff for sensitivity
and specificity (0.49843) yielded 72.50% correct classifications for sensitivity and
73.17% for specificity (Figure 9).
Table 7. Phase II Columbian Sharp-tailed Grouse occurrence model Wald z
statistic results. Estimated coefficients (β), standard errors (SE), intercepts (z), Pvalues (P) and 95% confidence intervals (95% CI).
Variable
% potential
habitat (3 km)
Road density (1
km)
Mean elevation
(1 km)
β
SE
z
P
1.879
0.809
2.320
0.020
0.294
3.465
-69.536
40.738
-1.710
0.088
-149.380
10.309
0.005
0.002
2.000
0.045
0.000
0.010
95 % CI
50
1.00
0.75
0.50
0.25
0.49843
0.00
0.00
0.25
0.50
0.75
1.00
Probability cutoff
Sensitivity
Specificity
Figure 9. Phase II model overlay plot of sensitivity and specificity versus
probability cutoffs. The optimal cutoff (0.49843), where sensitivity and specificity
intersect, yielded approximately 73% correct classifications for both active and
inactive leks.
A probability of occurrence map was created by applying the phase II
model logistic regression parameters to the spatial data layers. The probability of
occurrence map raster values that were greater than, or equal to, the phase II
model probability cutoff (0.49843) were identified as potential STG habitat. The
phase II model identified 316 habitat patches totaling 1,739,212 ha (Figure 10).
The potential habitat had a mean area of 5,504 ha, a median area of 1,038 ha, and
an area range from 0.09 ha to 190,019 ha.
51
Figure 10. 316 Phase II model Columbian Sharp-tailed Grouse potential habitat
patches with a total area of 1,739,212 ha.
DISCUSSION
Columbian Sharp-tailed Grouse population goals and habitat requirements
STG in Washington will be considered for down listing from state
threatened to sensitive when there is one metapopulation that averages 2,000
birds, and an overall population that averages a minimum of 3,200 birds, for a 10
year period (Stinson and Schroeder 2012). An area greater than 400,000 ha of
interconnected habitats of shrub-steppe, grassland, and CRP would be required to
support 2,000 or more STG at densities of 0.005 birds/ha (Stinson and Schroeder
2012). The final model identified 10 habitat patches greater than 50,000 ha
52
(Figure 11). The largest habitat patch (190,017 ha) and another large patch
(51,174 ha) were located in Okanogan County, in the northern portion of STG
historical range, where extant STG populations are concentrated. These patches
border the Okanogan-Wenatchee National Forest, WDFW Wildlife Areas, and are
within the Colville Indian Reservation. Potential habitat in these areas should be
the focus of STG range expansion efforts since they are located near the extant
STG populations (Schroeder 1996; McDonald and Reese 1998)
Another cluster of habitat patches greater than 50,000 ha, were located on
the western edge of STG historical range, predominately within Kittitas and
Yakima Counties. These patches are located on the Yakima Training Center,
WDFW’s Colockum and LT Murray Wildlife Areas, the Yakama Indian Nation
lands, and private lands. McDonald and Reese (1998) identified this area as
having the largest grassland patches in the Columbia Plateau. Dobler et al. (1996)
concluded that these large areas of remaining shrub-steppe on the Yakima
Training Center, Hanford Nuclear Site, and the Yakama Indian Nation, may be
the most suitable sites for species, like STG, that have evolved in expansive
shrub-steppe habitats. These areas may have good potential for supporting STG
populations since they are predominately located on public and sovereign tribal
lands and are large, fairly intact areas of native habitat.
53
Figure 11. Ten habitat patches identified by the final model that were greater than
50,000 ha. Two patches are located in Okanogan County in the northern portion
of STG historic range where the extant Columbian Sharp-tailed Grouse
populations are clustered.
Final model selection
The final model was selected by comparing the suitability of habitat areas
for STG that were identified by the phase I and II models. The habitat areas that
were identified by the two models were both located on the periphery of STG
historical range in higher elevation areas which is similar to the locations of the
extant STG habitat areas. Overall, the potential habitat patches identified by the
54
two models were very similar. The similarity was most likely due to the phase II
model being developed with phase I model habitat patch metric variables. In
addition, several of the environmental variables were the same (road density—1
km) or similar (mean elevation at different scales) for both models.
The main difference between the two model’s habitat patches were the
size and location of the patches and the phase I model had more habitat patches
located in low precipitation areas. The phase II model habitat patches were
smaller and more fragmented than the phase I model habitat. The road density (1
km) variable was included in both models and probably influenced the greater
degree of habitat fragmentation for the phase II model potential habitat. Roads
negatively impact STG populations due to habitat fragmentation, road avoidance
behavior, noise, and direct mortality (Manville 2004; Pruett et al. 2009, Robb and
Schroeder 2012; Stonehouse 2013). Road density is especially important within a
1 km radius of lek complexes because this area typically includes nesting and
brood rearing habitat (McDonald 1998; Stonehouse 2013). The phase I model also
had more habitat patches located in areas that received less than 23 cm of annual
precipitation compared to the phase II model (Figure 12). Areas that receive less
than 23 cm of annual precipitation are typically too dry to support the diversity of
grasses, forbs, and deciduous trees and shrubs that STG need for nesting, brood
rearing, and for winter forage and cover (Stinson and Schroeder 2012). Based on
the phase I model having more habitat patches located in low precipitation areas,
the phase II model was selected as the final model.
55
Figure 12. Mean annual precipitation and Columbian Sharp-tailed Grouse
potential habitat identified by the phase I and II models. There were more phase I
than phase II model patches in areas of low precipitation. Areas that receive less
than 23 cm of average annual precipitation cannot support the grasses and
diversity of vegetation that STG need for habitat.
Analysis of potential Columbian Sharp-tailed Grouse habitat
The extant STG populations in Washington are located in habitat areas on
the northern periphery of their historic range with elevations that range from 289
m to 1,518 m ( x = 748 m) and that receive 22.62 cm to 51.75 cm ( x = 34.72 cm)
of annual precipitation. The habitat areas identified by the final model follow this
pattern of being located on the periphery of the Columbia Plateau region. Overall,
56
compared to the current STG habitat areas, the model habitat areas were located
in higher elevations, with a range from 348 m to 2,078 m ( x = 858 m), and
received higher amounts of annual precipitation, 20.77 cm to 192.48 cm ( x =
45.49 cm). In Idaho, STG home ranges, nests, and lek sites were found at
elevations less than 2,200 m which indicates that the upper end of the model
habitat elevation may still be suitable for STG (Marks and Marks 1987; Ramsey
et al. 1999). Minimum precipitation is a limiting factor for suitable STG habitat in
the Columbia Plateau region. Most historical records indicate that STG occurred
in areas with a minimum of 28 cm of precipitation or near rivers in areas with
lower precipitation (Stinson and Schroder 2012). The Washington State
Columbian Sharp-tailed Grouse Recovery Plan considers areas below 23 cm of
precipitation as unsuitable habitat for STG (Stinson and Schroeder 2012).
A comparison of land cover within the extant STG habitat areas and the
potential habitat areas identified by the final model, showed that current STG
habitat areas had a higher percentage cover of grass and shrub habitat, introduced
grassland, and pasture/hay fields, and a lower percentage cover of agriculture,
forest, and riparian/winter habitat (Table 8). Percent area of pasture/hay fields (10
and 3 km) and percent area of shrub habitat (10 and 1 km) had negative linear
relationships with the probability of occurrence in the univariate analysis. This
means that as the percent area of those land cover types increased, the probability
of STG occurrence decreased. This result may seem contradictory since studies
have shown that STG select CRP fields, which are represented by the pasture/hay
category, for nests and brood rearing sites and that native shrubs are also
57
important habitat for STG in Washington (McDonald 1998; Stonehouse 2013).
CRP fields that have been planted with native shrubs and grasses are an important
component of STG habitat in Washington. However, many older CRP fields were
planted with a monoculture of crested wheatgrass that does not provide suitable
habitat for STG (Stinson and Schroeder 2012). The pasture/hay category does not
distinguish between CRP fields that are beneficial or not beneficial to STG which
is one possible explanation for why these areas were negatively associated with
STG occurrence. Additionally, current information on the location of cropland
enrolled in CRP was not readily available when the WHCWG land cover layers
were being developed. A comparison of the National Agricultural Statistics
Service Crop Data Layer (USDA-NASS), to an older 2007 CRP dataset, showed
that most CRP fields were captured in the CDL pasture/hay class (Appendix C)
(B. Cosentino, personal communication May 19th, 2014).
The negative linear relationship between shrub habitat and STG
occurrence may be due to the selection of land cover classes that were included in
the shrub category. The WHCWG land cover/land use layer classes that were
selected to represent native shrub habitat were, shrubland—basin and scabland
(Appendix C). The shrubland—basin class included taller shrubs such as big
sagebrush (Artemisia tridentata) and sparse herbaceous cover that is found in the
hotter, drier areas of the Columbia Plateau (WHCWG 2012a). The scabland class
included areas of poor, rocky soils with sparse cover characterized by low or
dwarf shrubs such as stiff sagebrush (Artemisia rigida) and buckwheat species
(Eriogonum sp.) (WHCWG 2012a). Stonehouse (2013) found that STG in eastern
58
Washington predominately selected grass habitats with sparse shrub cover for
their home ranges and lek sites, and built nests primarily under taller
bunchgrasses. This type of vegetative cover is defined by the WHCWG shrubsteppe class (WHCWG 2012a). The shrub-steppe class was not included in the
shrub category but in the grass habitat category for the model.
The final model also identified areas of potential STG habitat that were
predominately agriculture such as the fragmented patches along the border with
Idaho. The percent agriculture variable, at all scales, had a negative linear
relationship with the probability of occurrence in the univariate analysis. Percent
agriculture was not included as a variable in either the phase I or phase II models.
However, there was a strong correlation between percent grass and percent
agriculture. Percent grass (10 km) was highly correlated with percent agriculture
(10 km) and percent grass (3 km). Percent grass (3 km), one of the variables in the
phase I model, had a strong correlation (r = - 0.63) with percent agriculture (10
km). Likewise, percent grass (10 km) was strongly correlated (r = - 0.62) with
percent agriculture (3 km). Even though agriculture was strongly correlated with
percent grass, areas that were predominately agriculture were still identified as
potential habitat which may be a function of the combination of the model
variables. Overall, the final model selected for higher elevation areas with low
road density based on the variables in the two models: road density (1 km), which
occurred in both models, and average elevation (10 and 1 km), which also
occurred in both models but at different scales.
59
Table 8. A comparison of percent area for different land covers within Columbian
Sharp-tailed Grouse current range and the final model habitat areas.
Habitat
Current
Range
Final
Model
Percent Land Cover
Pasture/
Shrub Forest
Hay
Agriculture
Grass
15.45
40.96
25.38
6.79
21.84
36.76
8.45
23.41
Riparian/
Winter
Intro
Grass
4.70
2.06
3.78
4.13
3.00
2.10
Modeling considerations
The models were developed from geographic spatial data that contained
inherent inaccuracies that were determined by the scale and age of the data, the
number and quality of data sources, and other potential sources of error. The data
that was used for this model came from several sources, each compiled at
different times. All of the data that was used for the analysis had 30 m resolution
except for the precipitation data from PRISM which was 800 m. Even at this
relatively fine-scale resolution, the WHCWG (2012b) recommended that the 30 m
raster data from the Columbia Plateau analysis be used for landscape level
planning at scales of 1:100,000 or coarser.
Other important considerations in evaluating the results of the final model
are the habitat conditions and small population dynamics of the extant STG
populations that were the basis for model development. The models compared the
influence of environmental variables at active and inactive STG lek complexes.
However, the quality of current habitat at existing lek complexes was not
60
evaluated. Current habitat conditions may not be ideal for extant STG populations
compared to the quality of habitat in the Palouse Prairie and other areas in the
Columbia Plateau that historically supported the highest densities of STG (Stinson
and Schroeder 2012). Dobler et al. (1996), emphasized that the suitability of the
remaining shrub-steppe habitat for wildlife has changed in eastern Washington
since it is now predominately located in areas of poor, rocky soils which play an
important role in determining the quality of vegetation cover. Historically, the
best habitat for STG was the deep soil, high precipitation areas of the Palouse
Prairie in southeastern Washington (Yocom 1952; Stinson and Schroeder 2012).
This was the first area in Washington where STG were extirpated when the native
vegetation was converted to farmland in the early 20th century (Dobler et al 1996).
The remaining STG populations are located in the northern portion of their
historic range in higher elevation areas that were less impacted by agriculture,
orchard, and livestock grazing (Schroeder 1996). Average elevations (10 and 1
km) were two of the model variables that best predicted for STG occurrence
based on their current locations. While the current conditions of these higher
elevation areas provide more suitable habitat compared to the more modified
areas in lower elevations, they also may be less suitable for winter habitat. This is
reflected in the higher winter fatalities that are currently experienced by the
remaining populations (Schroeder 1996). Historically, lower elevation areas in the
Columbia Plateau had better winter weather conditions and more suitable riparian
habitat (Schroeder 1996). Another factor to consider in the model development is
small population dynamics. The extant STG populations in Washington are small
61
and therefore, more likely to be adversely affected by random changes to
environmental conditions, such as variations in food, extreme weather, predation,
and disease (Primack 2010; Stinson and Schroeder 2012). Therefore, STG lek
complexes may become inactive due to local extinctions of small populations
from random environmental or other stochastic events that may not be directly
related to the suitability of the existing habitat (Shaffer 1981).
MANAGEMENT RECOMMENDATIONS
On-the-ground assessment of the final model’s two large potential habitat
areas in Okanogan County should be conducted to evaluate potential areas for
land acquisitions, habitat restoration, and future translocations to expand the range
of STG. Existing areas of suitable habitat that are near extant STG habitat areas
and at high risk for development in Okanogan County, should be prioritized for
acquisition. Continued shrub-steppe habitat restoration on WDFW, Colville
Tribal, and Okanogan National Forest lands should also be a high priority in these
areas. Shrub-steppe habitat restoration for STG would also benefit many other
shrub-steppe species such as, Greater Sage-grouse (Centrocercus urophasianus),
sagebrush sparrow (Artemisiospiza belli), sage thrasher (Oreoscoptes montanus),
and burrowing owls (Athene cunicularia). In addition to the potential suitable
STG habitat in Okanogan County, further consideration should be given to the
large, existing areas of shrub-steppe habitat located on the Yakima Training
Center, WDFW’s Colockum and LT Murray Wildlife Areas, and the Yakama
Indian Nation lands. Habitat areas on the Yakima Training Center are already
62
managed for an extant population of Greater Sage-grouse and may also be
suitable for STG. Historical ranges for Greater Sage-grouse and STG overlapped
in eastern Washington and currently, one population of STG and Greater Sagegrouse home ranges overlap by 72% on the WDFW Swanson Lakes Wildlife Area
in Lincoln County (Stonehouse 2013). Overall, consideration should also be given
to the current conditions of STG habitat that were the basis for model
development and which may be suboptimal to pre-settlement habitats that
supported the greatest densities of STG.
63
CHAPTER 3 ADDITIONAL RECOMMENDATIONS
FURTHER ANALYSIS OF POTENTIAL COLUMBIAN SHARP-TAILED
GROUSE HABITAT IN OKANOGAN COUNTY
The final model identified ten potential habitat patches for STG that were
greater than 50,000 ha. Two of these patches were located in Okanogan County,
in the northern portion of STG historical range, where extant STG populations are
concentrated. These areas should be the focus of shrub-steppe conservation,
restoration, and population augmentation to expand the range of STG (Schroeder
1996; McDonald and Reese 1998). Both of the large habitat patches were adjacent
to multiple smaller patches that were less than two kilometers apart. These
smaller patches were combined with the two larger patches for a composite
analysis of land cover and land ownership within those potential habitat areas
(Figure 13). It is important to note that the smaller habitat patches were often
separated from the next patch by a local road or secondary highway. These roads
may create potential barriers to STG movement among the patches (Robb and
Schroeder 2012).
Land cover
The two composite potential habitat patches in Okanogan County were
located on the eastern and western periphery of STG historical range. The east
habitat patch had a total area of 226,246 ha and could potentially support a
population of 1,131 STG at a density of 0.005 birds/ha (Table 9). The west patch
had a total area of 105,173 ha and could potentially support a population of 525
64
STG at the same density. However, both of these patches had a high percentage of
forest cover that would not provide suitable habitat for STG. The east patch had
41% forest cover and the west patch had 51% forest cover. In addition, these
habitat patches may contain rugged areas and steep slopes greater than 30% that
would also not be suitable habitat for STG (Stinson and Schroeder 2012;
Stonehouse 2013). Overall, both patches had less percent area of agriculture,
native shrubland, pasture/hay fields, and introduced grass compared to the extant
STG habitat. The east patch had a higher percent of grassland but the west patch
had a much lower percent of grassland than extant STG habitat. Both patches had
a higher percentage of riparian/winter habitats. The west patch had more than
twice as much percent area of riparian/winter habitat compared to current STG
habitat.
65
Figure 13. A composite of two model habitat patches in Okanogan County,
greater than 50,000 ha, and adjacent habitat patches within two kilometers. These
potential habitat patches are located on the eastern and western edge of STG
historic range in Okanogan County in the area where extant STG populations are
clustered.
66
Table 9. A comparison of percent land cover for current Columbian Sharp-tailed
Grouse habitat range and two composite potential habitat areas in Okanogan
County.
Percent Land Cover
Habitat
Area
(ha)
Ag
Grass
Shrub
Forest
Pasture
/Hay
Riparian
/Winter
Intro
Grass
Current
Range
217,300
15.45
40.96
25.38
6.79
4.70
2.06
3.78
East
Patch
226,246
0.35
42.83
8.97
41.31
0.12
3.87
2.17
West
Patch
105,173
1.22
27.06
11.43
50.80
1.78
5.97
1.14
Land ownership
The two composite habitat patches are within ten recovery units identified
in the Washington STG Recovery Plan (Stinson and Schroeder 2012). The
Recovery Plan, in general, described the ten recovery units as potential STG
habitat areas that were important for habitat connectivity, but also contained
private lands that were at risk for development (Stinson and Schroeder 2012). The
majority of the land in the west patch is privately owned (72.60%) whereas the
east patch has slightly more land that is owned by federal and state agencies and
the Colville Confederated Tribes (52.22%) (Table 10).
67
Table 10. A comparison of land ownership for the two composite potential habitat
areas in Okanogan County.
Land Ownership
Area (ha)
% Area
East Patch
Federal
State
Tribal
Private
33,972.12
2,442.91
81,717.65
108,113.31
15.02
1.08
36.12
47.79
West Patch
Federal
14,971.39
14.24
State
13,841.04
13.16
Tribal
0.00
0.00
Private
76,360.57
72.60
Challenges and opportunities
Land ownership and land use issues present some of the greatest
challenges and opportunities for shrub-steppe habitat restoration in eastern
Washington (Dobler et al. 1996). More than half (51%) of STG active lek
complexes are located on private land (Schroeder et al. 2000). Connelly (2010), in
an assessment of STG conservation needs in Okanogan County, concluded that
private landowners were critical to the recovery effort for this species. However,
shrub-steppe habitats on private land are vulnerable to current and new land use
and land management practices including livestock grazing and development.
Livestock grazing in general, negatively impacts the quality of shrubsteppe habitats for STG. Intensive livestock grazing changes the structure and
composition of shrub-steppe vegetation by increasing the spread of invasive
68
grasses and woody vegetation, and decreasing native grasses and forbs that STG
need for nesting and brood rearing (Stinson and Schroeder 2012). In addition,
livestock grazing compacts soils and destroys the shrub-steppe soil crust which
can be instrumental for survival of native grasses and forbs (Belnap et al. 2001).
Livestock grazing that is very low intensity and that is timed to affect vegetation
the least, may be sustainable in shrub-steppe habitats. However, further research
and adaptive management strategies are needed to determine if there is an optimal
threshold for grazing based on different plant communities, soils, precipitation,
etc. (Beck and Mitchell 2000).
It is also important that remaining shrub-steppe habitat is not further
divided into rural residential development (Connelly 2010). The Washington
Recovery Plan for STG lists development as one of the factors affecting continued
existence of STG in Washington (Stinson and Schroeder 2012). Four of the 10
Recovery Plan recovery units where the Okanogan County habitat patches are
located are at high risk for development. Currently there are no federal or state
regulations that protect STG or their habitats on private land (Stinson and
Schroeder 2012).
Recommendations and conclusions
On-the-ground assessment of the potential habitat areas in Okanogan
County should be conducted to identify relatively large intact shrubsteppe/grassland habitats. The number one priority should be to conserve and
protect these habitats for STG and other shrub-steppe species. Dobler et al. (1996)
emphasized that species tend to evolve in concert with their surroundings, and for
69
shrub-steppe wildlife, like STG, this would mean species adapted to expansive
landscape of steppe and shrub-steppe communities. Many shrub-steppe obligate
and grassland species, like STG, are state or federal listed under the Endangered
Species Act including, Greater Sage-grouse (Centrocercus urophasianus), which
is a state threatened and federal candidate species, sagebrush sparrow
(Artemisiospiza belli), sage thrasher (Oreoscoptes montanus), loggerhead shrike
(Lanius ludovicianus), and burrowing owls (Athene cunicularia), which are state
candidate species, and pygmy rabbit (Brachylagus idahoensis) which is a state
and federal endangered species. Large, intact shrub-steppe habitats should be a
priority for land acquisitions in Okanogan County for conservation to benefit
these species. When land acquisitions are not possible, shrub-steppe habitats on
private lands should be conserved and protected with long-range planning and
policies adopted at the county level or in conjunction with governmental and nongovernmental entities (Azarrad et al. 2011). Conservation incentive programs for
private land owners such as, conservation easements that transfer development
rights, and tax incentives are some options that are available to encourage shrubsteppe conservation (Azerrad et al. 2011).
Another high priority is continued shrub-steppe restoration and protection
on WDFW, Colville Indian Reservation, Bureau of Land Management, and
Okanogan National Forest lands that are within or adjacent to Okanogan County.
When Greater Sage-grouse, which is currently a candidate species for protection
under the ESA, is upgraded to a threatened or endangered status, there will be
more funding opportunities for shrub-steppe habitat restoration on federal and
70
state lands (USFWS 2013). Shrub-steppe habitat restoration and other population
recovery efforts for Greater Sage-grouse could benefit STG since the two species
can live sympatrically within the same habitat area. Currently there is one
population of Greater Sage-grouse that is sympatric with a STG population in the
vicinity of Swanson Lakes Wildlife Area in eastern Washington. Stonehouse
(2013) found that the spring and summer habitat home ranges of the Swanson
Lakes’ Greater Sage-grouse and STG populations overlapped by 72%.
Finally, restoration of shrub-steppe habitats on private lands either through
the CRP or other farmland conservation programs area also important for STG
recovery. STG use restored CRP fields for nesting, brood rearing, and lek
locations in eastern Washington (Stonehouse 2013). CRP lands currently
comprise 4.7% of land cover in STG current range but only 0.12% of land cover
in the east habitat patch and 1.78% in the west habitat patch in Okanogan County.
There may be opportunities to increase the number of CRP acres that are enrolled
in Okanogan County. Another farmland conservation program that could
potentially be used to restore shrub-steppe habitat, is the Washington State
Farmland Preservation Grants program (Azerrad et al. 2011). Cities, counties,
nonprofit conservation organizations, and the State Conservation Commission can
purchase conservation easements on farmland to help preserve farmland and
protect wildlife through habitat restorations (Washington State Recreation and
Conservation Office 2010; Azerrad et al. 2011).
71
IMPROVING THE COLUMBIAN SHARP-TAILED GROUSE MODEL
Developing the STG model was made possible by the availability of GIS
data layers from the WHCWG (2012b) Columbia Plateau Ecoregional analysis.
However, the extent of the WHCWG GIS data did not include areas in Ferry,
Stevens, and Pend Oreille Counties in northeast Washington that were historically
occupied by STG (Figure 4). In addition, the 10 km radius for two of the active
lek complexes extended into Canada which was also not a part of the WHCWG
GIS data. The difference between the extent of the WHCWG data and the STG
historic range in Washington changed the way the model was created and applied.
When the model was created, two of the active lek complexes’ 10 km radii
extended beyond the extent of the GIS data. As a result, only the 3 km and 1 km
scales were used for the univariate analysis of the environmental variables for
those two active lek complexes (n = 40) whereas the 10 km scale was not used for
those two active lek complexes (n = 38). The model was applied to the WHCWG
spatial layers to create a probability of occurrence map which excluded areas of
STG historic range. These areas may contain suitable habitat for translocations.
Expanding the extent of the WHCWG GIS data layers to include the entire STG
historic range and areas of Canada should be considered especially if the
WHCWG data layers are updated in the future or if another STG model is
developed from the current layers.
In addition to expanding the extent of the spatial data layers and updating
the layers, alternative modeling strategies should be considered for future STG
models. One approach that may work well, since so many of the environmental
72
variables were highly correlated, is a principal component analysis (PCA). A PCA
would combine some of the variables together to create new variables that were
no longer correlated but still retained the information from the original variables.
The data in a PCA have to be normally distributed and independent.
Autocorrelation between the lek complexes would have to be assessed to
determine independence which may result in elimination of some of the leks
complexes and a reduced sample size. This could be a potential drawback for this
type of analysis since the sample size was already small.
Another approach to mapping suitable habitat for STG would be to create
a fine-scale map of existing riparian/winter habitats using hydrology and
vegetation maps in combination with near infrared National Agriculture Imagery
Program (NAIP) orthophotos. STG rely on deciduous riparian trees and shrubs for
cover and winter habitat (McDonald 1998) and these habitats can be a limiting
factor for STG occupancy in the highly fragmented shrub-steppe ecosystem of the
Columbia Plateau (Hofmann and Dobler 1988; Giesen and Connelly 1993;
Stinson and Schroeder 2012). Habitats in the vicinity of mapped riparian/winter
habitat could be assessed for suitability for STG using either a priori knowledge
from the literature or modeling.
Optimal STG habitat could also be mapped based on the existing
environmental conditions of areas that had the highest historic densities of STG,
such as the deep soil, higher precipitation areas of the Palouse Prairie (Stinson and
Schroeder 2012). These areas are now predominately cropland, however, a careful
selection of environmental variables, and well documented historic occurrence of
73
STG in those areas, could provide support for land acquisition and habitat
restoration on land that is better suited to STG.
74
REFERENCES
Aldrich, J. W. (1963). Geographic orientation of American tetraonidae. Journal of
Wildlife Management, 27(4), 529-545.
Aldridge, C. L., Nielsen, S. E., Beyer, H. L., Boyce, M. S., Connelly, J. W.,
Knick, S. T., & Schroeder, M. A. (2008). Range-wide patterns of Greater
sage-grouse persistence. Diversity & Distributions, 14(6), 983-994.
Aldridge, C. L., Saher, D., Childers, T. M., Stahlnecker, K. E., & Bowen, Z. H.
(2012). Crucial nesting habitat for Gunnison sage-grouse: A spatially
explicit hierarchical approach. Journal of Wildlife Management, 76(2),
391-406.
Asner, G. P., Elmore, A. J., Olander, L. P., Martin, R. E., & Harris, T. (2004).
Grazing systems, ecosystem responses, and global change. Annual Review
of Environment & Resources, 29(1), 261-C-4.
Azerrad, J. M., Divens, K.A., Livingston, M. F., Teske, M. S., Ferguson, H. L., &
Davis, J. L. (2011). Management recommendations for Washington’s
priority habitats: managing shrub-steppe in developing landscapes.
Washington Department of Fish and Wildlife, Olympia, Washington.
Bart, J. (2000). Status assessment and conservation plan for Columbian sharptailed grouse. Forest and Rangeland Ecosystem Science Center, U.S.
Geological Survey, Boise, Idaho.
Beck, J. L., & Mitchell, D. L. (2000). Influences of livestock grazing on sage
grouse habitat. Wildlife Society Bulletin, 28(4), 993.
Belnap, J., Kaltenenecker, J. H., Rosentreter, R.,Williams, J., Leonard, S., &
Eldredge, D. (2001). Biological soil crusts: Ecology and management.
Technical Ref. 1730–2. USDI, BLM and USGS. 110 pp.
Boisvert, J., Hoffman, R. W., & Reese, K. P. (2005). Home range and seasonal
movements of Columbian sharp-tailed grouse associated with
conservation reserve program and mine reclamation. Western North
American Naturalist, 65(1), 36-44.
Brambilla, M., Casale, F., Bergero, V., Matteo Crovetto, G. G., Falco, R., Negri,
I., & ... Bogliani, G. (2009). GIS-models work well, but are not enough:
Habitat preferences of Lanius collurio at multiple levels and conservation
implications. Biological Conservation, 142(10), 2033-2042.
75
Bunting, S. C., Kingery, J. L., & Schroeder, M. A. (2003). Assessing the
restoration potential of altered rangeland ecosystems in the interior
Columbia basin. Ecological Restoration, 21(2), 77.
Chatterjee, S. & Hadi, A. S. (2012). Regression analysis by example. Hoboken,
NJ: John Wiley and Sons, Inc.
Connelly, J. W. (2010). Habitat needs and protection for Columbian Sharp-tailed
Grouse in Washington with emphasis on Okanogan County. Retrieved
October 27, 2013 from http://www.columbiainstitute.org/pdf/Okanogan_STG_Report_3-7-10.pdf.
Connelly, J. W., Gratson, M. W., & Reese, K. P. (1998). Sharp-tailed grouse
(Tympanuchus phasianellus). In The birds of North America online (A.
Poole, Ed.). Ithaca: Cornell Lab of Ornithology; Retrieved from the Birds
of North America Online: http://0bna.birds.cornell.edu.cals.evergreen.edu/bna/species/354
Connelly, J. W., Schroeder, M. A., Sands, A. R., & Braun, C. E. (2000).
Guidelines to manage sage grouse populations and their habitats. Wildlife
Society Bulletin, 28(4), 967-985.
Cope, M. G. (1992). Distribution, habitat selection and survival of transplanted
Columbian sharp-tailed grouse (Tympanuchus phasianellus columbianus)
in the Tobacco Valley, Montana. Doctoral Dissertation. Montana State
University, Bozeman, Montana.
Crooks, K. R. & Sanjayan, M. editors. (2006). Connectivity conservation.
Cambridge, MA: Cambridge University Press.
Daubenmire, R. F. (1988). Steppe vegetation of Washington. Pullman,
Washington: Washington State University.
Davis, M.B. & Shaw, R.G. (2001). Range shifts and adaptive responses to
quaternary climate change. Science, 292, 673–679.
Dobler, F. C., Eby, J., Perry, C., Richardson, S., & Vander Haegen, M. (1996).
Status of Washington’s shrubsteppe ecosystem: Extent, ownership, and
wildlife/vegetation relationships. Phase one completion report.
Washington Department of Fish and Wildlife. Olympia, Washington.
Donovan, T. M., & Flather, C. H. (2002). Relationships among North American
songbird trends, habitat fragmentation, and landscape occupancy.
Ecological Applications, 12(2), 364-374.
76
Edgley, E. F. (2001). Use of geographic information systems to model potential
Columbian sharp-tailed grouse nesting/brood rearing habitat. Master’s
Thesis. Utah State University, Logan, Utah.
ESRI (2012). ArcGIS Desktop: Release 10.1. Redlands, CA: Environmental
Systems Research Institute.
Evans, K. E., & Dietz, D. R. (1974). Nutritional energetics of sharp-tailed grouse
during winter. Journal of Wildlife Management, 38(4), 622.
Fitzpatrick, M. C. (2003). Use of GIS and a modified habitat suitability index
model to quantify Columbian sharp-tailed grouse habitats in the Upper
Blackfoot Valley, Montana. Master’s Thesis. The University of Montana,
Missoula, Montana.
Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R.,
&…Snyder, P. K. (2005). Global consequences of land use. Science,
309(5734), 570-574.
Frankham, R. (2005). Genetics and extinction. Biological Conservation, 126(2),
131-140.
Frankham, R. (2006). Genetics and landscape connectivity. In K. R. Crooks and
M. Sanjayan (Eds.). Connectivity conservation (pp. 72-96). Cambridge,
MA: Cambridge University Press.
Frankham, R., Ballou, J.D., & Briscoe, D.A. (2002). Introduction to conservation
genetics. Cambridge, MA: Cambridge University Press.
Gaston, K. J., Blackburn, T. M., & Goldewijk, K. K. (2003). Habitat conversion
and global avian biodiversity loss. Proceedings of the Royal Society of
London. Series B: Biological Sciences, 270(1521), 1293-1300.
Gelbard, J. L., & Belnap, J. (2003). Roads as conduits for exotic plant invasions
in a semiarid landscape. Conservation Biology, 17(2), 420.
Giesen, K. M. (1997). Seasonal movements, home ranges, and habitat use by
Columbian sharp-tailed grouse in Colorado. Special report No. 72.
Colorado Division of Wildlife, Fort Collins, Colorado.
Giesen, K. M., & Connelly, J. W. (1993). Guidelines for management of
Columbian sharp-tailed grouse habitats. Wildlife Society Bulletin, 21(3),
325.
Gilpin, M. E., & Soule, M. E. (1986). Minimum viable populations: Processes of
species extinction. In M. E. Soule (Ed.). Conservation biology: The
77
science of scarcity and diversity (pp. 19-34). Sunderland, MA: Sinauer
Associates.
Goddard, A. D., Dawson, R. D., & Gillingham, M. P. (2009). Habitat selection by
nesting and brood-rearing sharp-tailed grouse. Canadian Journal of
Zoology, 87(4), 326-336.
Hammond, M. (2011). PythonWin 2.7.2. PythonWin IDE and GUI Framework
for Windows.
Harju, S. M., Dzialak, M. R., Taylor, R. C., Hayden-Wing, L. D., & Winstead, J.
B. (2010). Thresholds and time lags in effects of energy development on
Greater sage-grouse populations. Journal of Wildlife Management, 74(3),
437-448.
Heller, N. E. & Zavaleta, E. S. (2009). Biodiversity management in the face of
climate change: A review of 22 years of recommendations. Biological
Conservation, 142, 14–32.
Hofmann, L. & Dobler, F. C. (1988). Observations of wintering densities and
habitat use by Columbian sharp-tailed grouse in three counties of eastern
Washington. Unpublished report, Washington Department Wildlife,
Olympia, Washington.
Hosmer, D. W., Lemeshow, S., & Sturdivant, R. X. (2013). Applied logistic
regression (3rd edition). New York, NY: Wiley.
IUCN (International Union for Conservation of Nature and Natural Resources).
(2013). The IUCN red list of threatened species. Retrieved November 29,
2013 from http://www.iucnredlist.org/.
Johnsgard, P. A. (1973). Grouse and quails of North America. Lincoln, Nebraska:
University of Nebraska Press.
Knick, S. T., Hanser, S. E., & Preston, K. L. (2013). Modeling ecological
minimum requirements for distribution of Greater sage-grouse leks:
Implications for population connectivity across their western range,
U.S.A. Ecology and Evolution, 3(6), 1539-1551.
Liu, C., Berry, P. M., Dawson, T. P., & Pearson, R. G. (2005). Selecting
thresholds of occurrence in the prediction of species distributions.
Ecography, 28, 385-393.
Manel, S., Williams, H. C., & Ormerod, S. J. (2001). Evaluating presence-absence
models in ecology: The need to account for prevalence. Journal of Applied
Ecology, 38, 921-931.
78
Manville, A. M., II. (2004). Prairie grouse leks and wind turbines: U.S. Fish and
Wildlife Service justification for a 5-mile buffer from leks; Additional
grassland songbird recommendations. Division of Migratory Bird
Management, USFWS, Arlington, Virginia.
Marks, J. S., & Marks, V. A. (1987). Habitat selection by Columbian sharp-tailed
grouse in west central Idaho. U.S. Department of the Interior, Bureau of
Land Management, Boise, ID.
Marks, J. S. & Marks, V. A. (1988). Winter habitat use by Columbian sharp-tailed
grouse in western Idaho. Journal of Wildlife Management, 52(4), 743-746.
McDonald, M. W., & Reese, K. P. (1998). Landscape changes within the
historical distribution of Columbian sharp-tailed grouse in eastern
Washington: Is there hope?. Northwest Science, 72(1), 34-41.
McDonald, M. W. (1998). Ecology of Columbian sharp-tailed grouse in eastern
Washington. Master’s Thesis. University of Idaho, Moscow, Idaho.
McGarigal, K., Cushman, S. A., & Ene, E. (2012). FRAGSTATS v4: Spatial
pattern analysis program for categorical and continuous maps. Computer
software program produced by the authors at the University of
Massachusetts, Amherst. Retrieved from
http://www.umass.edu/landeco/research/fragstats/fragstats.html.
MEA (Millennium Ecosystem Assessment). (2005). Ecosystems and human wellbeing: Biodiversity synthesis. Washington DC: World Resources Institute.
Meints, D. R. (1991). Seasonal movements, habitat use, and productivity of
Columbian sharp-tailed grouse in southeastern Idaho. Master’s Thesis.
University of Idaho, Moscow, Idaho.
Meints, D. R., Connelly, J. W., Reese, K. P., Sands, A. R., & Hemker, T. P.
(1992). Habitat suitability index procedure for Columbian sharp-tailed
grouse. Bulletin 55. Idaho Forest, Wildlife and Range Experiment Station,
University of Idaho, Moscow.
Neilson R, P., Lenihan J, M., Bachelet, D., & Drapek R, J. (2005). Climate
change implications for sagebrush ecosystems. North American Wildlife
and Natural Resources Conference. Transactions, 70145-159.
Odell, E. L. (2003). Incorporating ecology into land use planning. Journal of the
American Planning Association, 69(1), 72.
Peng, C., & So, T. (2002). Logistic regression analysis and reporting: A primer.
79
Understanding Statistics, 1(1), 31-70.
Pimm, S. L. & Raven, P. (2000). Biodiversity: Extinction by numbers. Nature,
403(6772), 843-845.
Pitman, J. C., Hagen, C. A., Robel, R. J., Loughin, T. M., & Applegate, R. D.
(2005). Location and success of Lesser Prairie-chicken nests in relation to
vegetation and human disturbance. Journal of Wildlife Management,
69(3), 1259-1269.
Primack, R. (2010). Essentials of conservation biology (5th ed.). Sunderland,
MA: Sinauer Associates Inc.
PRISM (PRISM Climate Group) (2014). 30-Year Normals. Oregon State
University, Corvallis, OR: Northwest Alliance for Computational Science
and Engineering. http://www.prism.oregonstate.edu/.
Pruett, C. L., Patten, M. A., & Wolfe, D. H. (2009). Avoidance behavior by
prairie grouse: Implications for development of wind energy.
Conservation Biology, 23(5), 1253-1259.
Python (2010). Python 2.7. Beaverton, OR: Python Software Foundation.
Ramankutty, N., Foley, J., & Olejniczak, N. (2002). People on the land: Changes
in global population and croplands during the 20th century. AMBIO - A
Journal of the Human Environment, 31(3), 251.
Ramsey, R. D., Black, A. T., Edgley, E., & Yorgason, N. (1999). Use of GIS and
remote sensing to map potential Columbian sharp-tailed grouse habitat in
southeastern Idaho. Idaho City, ID: Bureau of Land Management, United
States Department of Interior.
Rittenhouse, C. D., Millspaugh, J. J., Cooper, A. B., Hubbard, M. W., Sheriff, S.
L., & Gitzen, R. A. (2008). Modeling resource selection using polytomous
logistic regression and kernel density estimates. Environmental &
Ecological Statistics, 15(1), 39-47.
Robb, L. & Schroeder, M. (2012). Appendix A.1 habitat connectivity for sharptailed grouse (Tympanuchus phasianellus) in the Columbia Plateau
ecoregion. Washington Wildlife Habitat Connectivity Working Group
(WHCWG). Washington Connected Landscapes Project: Analysis of the
Columbia Plateau Ecoregion. Washington’s Department of Fish and
Wildlife, and Department of Transportation, Olympia, WA.
Rodgers, R. D. & Hoffman, R. W. (2005). Prairie grouse population response to
conservation reserve program grasslands: An overview. In A. W. Allen &
80
M. W. Vandever (Eds.), The conservation reserve program - Planting the
future: Proceedings of a national conference, scientific investigation
report 2005-5145 (pp. 120-128). Fort Collins, Colorado: U. S. Geological
Survey, Biological Resources Division.
Sala, O. E., Chapin, F. S., Armesto, J. J., Berlow, E., Bloomfield, J., Dirzo,
&…Wall, D. H. (2000). Global biodiversity scenarios for the year 2100.
Science, 287, 1770–1774.
Schroeder, M. A. (1996). Potential for sharp-tailed grouse management on the
Okanogan National Forest, Washington. Report to the Okanogan National
Forest. Washington Department of Fish and Wildlife, Olympia,
Washington
Schroeder, M., Atamian, M., Ferguson, H., Finch, M., Stinson, D., & Stonehouse,
K. (2012). Re-establishment of viable populations of Columbian sharptailed grouse in Washington: Progress report. Job progress report.
Washington Department of Fish and Wildlife, Olympia, Washington.
Schroeder, M. A., Hays, D. W., Murphy, M. A., & Pierce, D. J. (2000). Changes
in the distribution and abundance of Columbian sharp-tailed grouse in
Washington. Northwestern Naturalist, 81(3), 95-103.
Schroeder, M. A. & Vander Haegen, W. M. (2006). Use of conservation reserve
program fields by greater sage-grouse and other shrubsteppe-associated
wildlife in Washington State. Technical report for USDA Farm Service
Agency. Washington Department of Fish and Wildlife, Olympia,
Washington
Shaffer, M. L. (1981). Minimum population sizes for species conservation.
BioScience. pp. 131-134.
Snyder, J. W., Pelren, E. C., & Crawford, J. A. (1999). Translocation histories of
prairie grouse in the United States. Wildlife Society Bulletin, 27(2), 428.
StataCorp (2013). Small Stata 13. College Station, TX: StataCorp LP.
Stinson, D. W. & Schroeder, M. A. (2012). Washington state recovery plan for
the Columbian sharp-tailed grouse. Washington Department of Fish and
Wildlife, Olympia.
Stonehouse, K. F. (2013). Habitat selection by sympatric, translocated Greater
sage-grouse and Columbian sharp-tailed grouse in eastern Washington.
Master’s Thesis. Washington State University, Pullman, Washington.
81
Tack, J. D. (2006). Sage-grouse and the human footprint: Implications for
conservation of small and declining populations. Master’s Thesis.
University of Montana, Missoula, Montana.
USFWS (U.S. Fish and Wildlife Service). (2000). Endangered and threatened
wildlife and plants: 12-Month findings for a petition to list Columbian
sharp-tailed grouse as threatened. Federal Register, 65(197), 60391–
60396.
USFWS (U.S. Fish and Wildlife Service). (2006). Endangered and threatened
wildlife and plants: 90-Day finding on a petition to list the Columbian
sharp-tailed grouse as threatened or endangered. Federal Register,
71(224), 67318–67325.
USFWS (U.S. Fish and Wildlife Service). (2013). ESA basics: 40 years of
conserving endangered species. Retrieved November 14, 2013 from
http://www.fws.gov/endangered/esa-library/pdf/ESA_basics.pdf
USFWS (U.S. Fish and Wildlife Service). (1981). Standards for the development
of habitat suitability index models. 103 ESM. USDI Fish and Wildlife
Service, Division of Ecological Services, Washington, D.C.
van den Berg, L. J. L., Bullock, J. M., Clarke, R. T., Langston, R. H. W., & Rose,
R. J. (2001). Territory selection by the Dartford warbler (Sylvia undata) in
Dorset, England: The role of vegetation type, habitat fragmentation and
population size. Biological Conservation, 101(2), 217-228.
WAC 232-12-297. Endangered, threatened, and sensitive wildlife species
classifications. Retrieved November 14, 2013 from
http://apps.leg.wa.gov/wac/default.aspx?cite=232-12-297.
Warheit, K. I. & Schroeder, M. A. (2003). Genetic survey of Columbian sharptailed grouse populations in western North America: Summary of
microsatellite results. Draft Report, Washington Department of Fish and
Wildlife, Olympia, Washington.
Washington State Recreation and Conservation Office. (2010). Farmland
preservation grants. Retrieved May 24, 2014 from
http://www.rco.wa.gov/grants/farmland.shtml
WDFW (Washington Department of Fish and Wildlife). (2012). Threatened and
endangered wildlife in Washington: 2012 annual report. Columbian
sharp-tailed grouse. Listing and Recovery Section, Wildlife Program,
Washington Department of Fish and Wildlife, Olympia, Washington. 124128.
82
Westemeier, R. L., Brawn, J. D., Simpson, S. A., Esker, T. L., Jansen, R. W.,
Walk, J. W… & Paige, K. N. (1998). Tracking the Long-Term Decline
and Recovery of an Isolated Population. Science, 282(5394).
WHCWG (Washington Wildlife Habitat Connectivity Working Group). (2012a).
Washington connected landscapes project: Analysis of the Columbia
Plateau ecoregion. Appendix D GIS data layer development and data
sources. Unpublished report. Washington’s Department of Fish and
Wildlife and Department of Transportation, Olympia, Washington.
WHCWG (Washington Wildlife Habitat Connectivity Working Group). (2012b).
Washington connected landscapes project: Analysis of the Columbia
Plateau ecoregion. Washington’s Department of Fish and Wildlife and
Department of Transportation, Olympia, Washington.
Wilcove D. S., McLellan C. H., & Dobson A. P. (1986). Habitat fragmentation in
the temperate zone. In M. E. Soule (Ed.), Conservation biology (pp. 237256). Sunderland,MA: Sinauer.
Wisdom, M. J., Meinke, C. W., Knick, S. T., & Schroeder, M. A. (2011). Factors
associated with extirpation of sage-grouse. In S. T. Knick and J. W.
Connelly (Eds.), Greater sage-grouse: Ecology and conservation of a
landscape species and its habitats. Studies in avian biology (vol. 38) (pp.
451–472). Berkeley, CA: University of California Press.
Wolfe, D. H., Patten, M. A., Shochat, E., Pruett, C. L., & Sherrod, S. K. (2007).
Causes and patterns of mortality in Lesser prairie-chickens Tympanuchus
pallidicinctus and implications for management. Wildlife Biology, 13(1),
95-104.
Yocom, C. F. (1952). Columbian sharp-tailed grouse (Pedioecetes phasianellus
columbianus) In the State of Washington. American Midland Naturalist,
48(1), 185-192.
83
APPENDIX A LEK COMPLEX DATA
Lek ID
Lek Status
Last Year
Recent Count
Max Count
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
1
0
0
0
0
0
0
0
1
0
0
0
1
0
1
1
1
0
0
1
1
0
1
1
1
1
1
0
1
1
0
0
1
0
0
0
0
2013
2002
2012
1996
2011
1996
1994
2002
2013
1997
2005
1997
2013
2012
2013
2013
2013
2002
2009
2013
2013
2000
2013
2013
2013
2013
2013
2002
2013
2013
2002
1998
2013
1997
2002
1998
2005
11
2
1
4
2
1
1
2
6
2
5
1
5
3
18
5
4
1
1
3
20
1
21
13
3
21
19
2
24
8
3
2
7
2
15
3
1
30
10
7
45
2
14
5
4
22
13
14
7
14
9
18
31
25
22
28
21
20
3
21
18
7
21
40
8
26
32
9
8
7
2
15
23
10
84
Appendix A (continued)
Lek ID
Lek Status
Last Year
Recent Count
Max Count
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
1
0
0
0
1
1
0
1
1
1
1
0
1
0
0
1
1
0
0
1
0
0
0
1
0
0
1
1
1
1
1
1
0
0
1
1
1
0
2013
2012
2012
2011
2013
2013
1994
2013
2013
2013
2013
2012
2013
2005
2001
2013
2013
1997
2006
2013
2009
1998
2012
2013
1994
2007
2013
2013
2013
2013
2013
2013
1994
2011
2013
2013
2013
2012
5
10
4
1
4
26
1
7
18
11
9
11
15
2
1
13
2
1
4
28
1
2
3
12
1
2
21
4
5
3
14
12
2
3
13
3
18
2
18
16
14
3
10
26
14
15
21
11
9
12
23
8
45
13
10
1
20
32
15
4
31
13
4
24
58
11
19
10
19
20
12
9
28
13
50
28
85
Appendix A (continued)
Lek ID
Lek Status
Last Year
Recent Count
Max Count
76
77
78
79
80
81
0
0
1
1
1
0
1995
2002
2013
2013
2013
2011
1
1
4
1
22
2
3
4
30
22
39
4
86
APPENDIX B LAND COVER VARIABLE CATEGORIES
Variable Land Cover
Name
Category
WHCWG (2012b) Land Cover/Land Use Class
Name
gr
Grassland
sh
Shrubland
rw
Riparian/winter
habitat
Grassland - basin
Grassland - mountain
Shrubsteppe
Meadow
Shrubland - basin
Scabland
Shrubland - mountain
igr
fr
ag
ph
Herbaceous wetland
Riparian
Aspen
Introduced grassland Introduced upland vegetation -annual grassland
Forest
Woodland
Forest
Agriculture
Nonirrigated cropland
Irrigated/not irrigated cultivated agriculture
buffer 0 - 250 m from native habitat
Irrigated cropland
Highly structured agriculture
Cultivated cropland
Irrigated/not irrigated cultivated agriculture
buffer 250-500 m from native habitat
Pasture/hay (CRP)
Pasture hay agriculture buffer 0-250 m from
native habitat
Pasture hay agriculture buffer 250-500 m from
native habitat
Pasture hay
87
APPENDIX C SOURCES AND DESCRIPTIONS OF LAND COVER
CLASSES
NW GAP Class
Columbia Basin Foothill and Canyon Dry Grassland
Columbia Basin Palouse Prairie
Columbia Plateau Steppe and Grassland
Inter-Mountain Basins Semi-Desert Grassland
Northern Rocky Mountain Lower Montane, Foothill
and Valley Grassland
North Pacific Alpine and Subalpine Dry Grassland
North Pacific Dry and Mesic Alpine Dwarf-Shrubland,
Fell-field and Meadow
North Pacific Herbaceous Bald and Bluff
North Pacific Montane Grassland
Northern Rocky Mountain Subalpine-Upper Montane
Grassland
Rocky Mountain Alpine Fell-Field
Rocky Mountain Alpine Tundra/Fell-field/Dwarfshrub Map Unit
Columbia Plateau Low Sagebrush Steppe
Columbia Plateau Silver Sagebrush Seasonally
Flooded Shrub-Steppe
Inter-Mountain Basins Big Sagebrush Steppe
Inter-Mountain Basins Montane Sagebrush Steppe
Inter-Mountain Basins Semi-Desert Shrub Steppe
Great Basin Xeric Mixed Sagebrush Shrubland
Inter-Mountain Basins Big Sagebrush Shrubland
Inter-Mountain Basins Mixed Salt Desert Scrub
Inter-Mountain Basins Curl-leaf Mountain Mahogany
Woodland and Shrubland
North Pacific Avalanche Chute Shrubland
North Pacific Montane Shrubland
Northern and Central California Dry-Mesic Chaparral
Northern Rocky Mountain Montane-Foothill
Deciduous Shrubland
Northern Rocky Mountain Subalpine Deciduous
Shrubland
Columbia Plateau Ash and Tuff Badland
Columbia Plateau Scabland Shrubland
WHCWG (2012a)
Class Name
Grassland - Basin
Grassland - Basin
Grassland - Basin
Grassland - Basin
Grassland - Basin
Grassland - Mountain
Grassland - Mountain
Grassland - Mountain
Grassland - Mountain
Grassland - Mountain
Grassland - Mountain
Grassland - Mountain
Shrubsteppe
Shrubsteppe
Shrubsteppe
Shrubsteppe
Shrubsteppe
Shrubland - Basin
Shrubland - Basin
Shrubland - Basin
Shrubland - Mountain
Shrubland - Mountain
Shrubland - Mountain
Shrubland - Mountain
Shrubland - Mountain
Shrubland - Mountain
Scabland
Scabland
88
Appendix C (continued)
NW GAP Class
Introduced Upland Vegetation - Annual Grassland
North Pacific Bog and Fen
Rocky Mountain Alpine-Montane Wet Meadow
Rocky Mountain Subalpine-Montane Fen
Rocky Mountain Subalpine-Montane Mesic Meadow
Temperate Pacific Montane Wet Meadow
Willamette Valley Wet Prairie
Columbia Plateau Vernal Pool
Inter-Mountain Basins Alkaline Closed Depression
Inter-Mountain Basins Playa
North American Arid West Emergent Marsh
Temperate Pacific Freshwater Aquatic Bed
Temperate Pacific Freshwater Emergent Marsh
Columbia Basin Foothill Riparian Woodland and
Shrubland
Great Basin Foothill and Lower Montane Riparian
Woodland and Shrubland
Inter-Mountain Basins Greasewood Flat
Inter-Mountain Basins Montane Riparian Systems
Introduced Upland Vegetation - Shrub
Introduced Upland Vegetation - Treed
Mediterranean California Foothill and Lower Montane
Riparian Woodland
North Pacific Lowland Riparian Forest and Shrubland
North Pacific Montane Riparian Woodland and
Shrubland
North Pacific Shrub Swamp
Northern Rocky Mountain Conifer Swamp
Northern Rocky Mountain Lower Montane Riparian
Woodland and Shrubland
Rocky Mountain Lower Montane Riparian Woodland
and Shrubland
Rocky Mountain Subalpine-Montane Riparian
Shrubland
Rocky Mountain Subalpine-Montane Riparian
Woodland
WHCWG (2012a)
Class Name
Introduced Upland
Vegetation - Annual
Grassland
Meadow
Meadow
Meadow
Meadow
Meadow
Meadow
Herbaceous Wetland
Herbaceous Wetland
Herbaceous Wetland
Herbaceous Wetland
Herbaceous Wetland
Herbaceous Wetland
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
89
Appendix C (continued)
NW GAP Class
Inter-Mountain Basins Aspen-Mixed Conifer Forest
and Woodland
Rocky Mountain Aspen Forest and Woodland
Columbia Plateau Western Juniper Woodland and
Savanna
Introduced Upland Vegetation - Perennial Grassland
and Forbland
North Pacific Broadleaf Landslide Forest and
Shrubland
North Pacific Maritime Mesic Subalpine Parkland
North Pacific Oak Woodland
North Pacific Wooded Volcanic Flowage
Northern California Mesic Subalpine Woodland
Northern Rocky Mountain Ponderosa Pine Woodland
and Savanna
Northern Rocky Mountain Subalpine Woodland and
Parkland
Northern Rocky Mountain Western Larch Savanna
Willamette Valley Upland Prairie and Savanna
East Cascades Mesic Montane Mixed-Conifer Forest
and Woodland
East Cascades Oak-Ponderosa Pine Forest and
Woodland
Mediterranean California Dry-Mesic Mixed Conifer
Forest and Woodland
Mediterranean California Red Fir Forest
Middle Rocky Mountain Montane Douglas-fir Forest
and Woodland
North Pacific Dry Douglas-fir-(Madrone) Forest and
Woodland
North Pacific Dry-Mesic Silver Fir-Western HemlockDouglas-fir Forest
North Pacific Lowland Mixed Hardwood-Conifer Forest
and Woodland
North Pacific Maritime Dry-Mesic Douglas-fir-Western
Hemlock Forest
North Pacific Maritime Mesic-Wet Douglas-fir-Western
Hemlock Forest
North Pacific Mesic Western Hemlock-Silver Fir Forest
WHCWG (2012a)
Class Name
Aspen
Aspen
Woodland
Woodland
Woodland
Woodland
Woodland
Woodland
Woodland
Woodland
Woodland
Woodland
Woodland
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
90
Appendix C (continued)
NW GAP Class
North Pacific Mountain Hemlock Forest
Northern Rocky Mountain Dry-Mesic Montane Mixed
Conifer Forest
Northern Rocky Mountain Mesic Montane Mixed
Conifer Forest
Rocky Mountain Lodgepole Pine Forest
Rocky Mountain Poor-Site Lodgepole Pine Forest
Rocky Mountain Subalpine Dry-Mesic Spruce-Fir
Forest and Woodland
Rocky Mountain Subalpine Mesic Spruce-Fir Forest
and Woodland
Cultivated Cropland
WHCWG (2012a)
Class Name
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Cultivated Cropland
USDA-NASS Crop Data Class
WHCWG (2012a) Class
Name
Pasture/Hay
Pasture/Hay
Alfalfa
Other Hay
Clover/Wildflowers
Pasture/Grass
Caneberries
Hops
Cherries
Peaches
Apples
Grapes
Christmas Trees
Other Tree Nuts
Other Tree Fruits
Walnuts
Pears
Nectarines
Plums
Apricots
Blueberries
Corn
Pasture_Hay
Pasture_Hay
Pasture_Hay
Pasture_Hay
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Highly Structured Agriculture
Irrigated Cropland
91
Appendix C (continued)
USDA-NASS Crop Data Class
Soybeans
Sunflower
Sweet Corn
Mint
Flaxseed
Mustard
Sugarbeets
Potatoes
Other Crops
Misc. Vegs. & Fruits
Watermelons
Onions
Peas
Tomatoes
Herbs
Carrots
Asparagus
Greens
Strawberries
Squash
Dbl. Crop WinWht/Corn
Dbl. Crop Oats/Corn
Lettuce
Cucumbers
Pumpkins
Cabbage
Radishes
Sorghum
Barley
Spring Wheat
Winter Wheat
Rye
Oats
Speltz
Canola
Safflower
Rape Seed
Camelina
WHCWG (2012a) Class
Name
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Irrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
92
Appendix C (continued)
USDA-NASS Crop Data Class
Dry Beans
Lentils
Sod/Grass Seed
Fallow/Idle Cropland
Triticale
WHCWG (2012a) Class
Name
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
Nonirrigated Cropland
93
