-
https://cpw.cvlcollections.org/files/original/c483c29d854d0672b82237376f44f673.pdf
75394f3c9d4117361a9409f009e005d8
PDF Text
Text
INTERIM PROGRESS REPORT
Greater Sage-Grouse Research in the Parachute-Piceance-Roan Region of Western Colorado
Part II: Multi-scale Habitat Selection and Seasonal Habitat Mapping
AUTHOR: Dr. Brett L. Walker
PRINCIPAL INVESTIGATOR: Dr. Brett L. Walker
PLEASE DIRECT QUESTIONS TO: Dr. Brett Walker, Colorado Division of Wildlife, 711 Independent Ave.,
Grand Junction, CO 81505. Phone: 970-255-6125; E-mail: brett.walker@state.co.us.
H
H
PERIOD COVERED: March 2006 - 25 July 2010
PROJECT PERSONNEL: Brian Holmes, Wildlife Conservation Biologist (CDOW - Meeker); Brad Petch,
Senior Wildlife Conservation Biologist (CDOW - Grand Junction); Tom Knowles, DWM (CDOW - Meeker);
Albert Romero, DWM (CDOW - DeBeque); J.T. Romatzke, AWM (CDOW - Grand Junction); Bill deVergie,
AWM (CDOW - Meeker).
**Information in this report is preliminary and subject to further evaluation. Information herein may
not be published or quoted without permission of the author. Manipulation of data beyond that
contained in this report is discouraged.**
Abstract. Loss and degradation of sagebrush habitat throughout western North America has led to
growing concern for conservation of greater sage-grouse (Centrocercus urophasianus) and repeated
petitions to list the species under the Endangered Species Act. Greater sage-grouse in the ParachutePiceance-Roan (PPR) region of western Colorado face at least two known potential stressors: increasing
energy development and a long-term decline in habitat suitability associated with pinyon-juniper
encroachment. In 2006, the Colorado Division of Wildlife (CDOW) and industry partners initiated a 3year study to obtain baseline data on seasonal habitat use, movements, vital rates, and genetics of greater
sage-grouse in the PPR. CDOW has since expanded that original project to include two new objectives:
(1) generate high-resolution maps showing high-priority seasonal use areas for the entire population, and
(2) assess the value of pinyon-juniper removal for increasing sage-grouse habitat. Industry, landowners,
and state and federal agencies need high-resolution maps showing where sage-grouse occur during each
season to streamline development planning, quantify mitigation needs, and guide on-the-ground sagegrouse conservation efforts. I conducted multi-scale habitat selection analyses and validation for the
breeding and summer-fall seasons using a total of 1130 breeding-season locations from 102 radiomarked individuals collected from 2006-2010 and 1367 summer-fall locations from 84 radio-marked
individuals collected from 2006-2009. I used logistic regression to test landscape-level habitat features
at six scales (100, 350, 740, 1000, 1600, and 3200 m). Models validated well against independent
locations (R2 = 0.912-0.984). Sage-grouse selected similar habitat features at similar scales in all both
seasons. They selected for greater proportion sagebrush at multiple scales (100-m and 350-m), higher
elevations, and flatter terrain. They selected against proportion forest (350-m or 740-m) and proportion
mountain shrub-only (740-m or 1600-m). Landscape-level guidelines for sage-grouse are based on 25%75% quartiles of values for key predictor variables measured at used locations. Breeding areas should
have: (a) less rugged topography within 100 m (roughness index = 4.82-9.55), (b) 57.6-96.2% sagebrushdominated habitat within 100m, (c) 90.4-98.4% sagebrush + grassland + mixed sagebrush-mountain
shrub habitat within 350 m, (d) 0.5-6.5% forested habitat within 350 m, (e) 0.0-1.2% mountain shrubonly habitat within 740 m, and (f) areas 140-314 m from forest. Summer-fall habitat should have: (a)
less rugged topography within 100 m (roughness index 5.20-10.31), (b) 50-92% sagebrush-dominated
habitat within 100 m, (c) 88.1-98.6% sagebrush + grassland + sagebrush-mountain shrub habitat within
�Document Date: 08-10-10
350 m, (d) 4.5-11.5% forested habitat within 740 m, (e) 0.0-1.3% mountain shrub-only habitat within
740 m, (f) 0.0% riparian habitat within 1000 m, (g) northeast and northwest-facing terrain, and (h) areas
98-268 m from forest. Breeding and summer-fall habitat largely overlapped in this population. Sagegrouse in both seasons selected landscapes with a mixture of sagebrush, grassland (or sparse sagebrush),
and mixed sagebrush-mountain shrub habitat types over habitats with sagebrush alone. These results
support three main concepts in sage-grouse habitat selection. First, sage-grouse require sagebrush yearround. Although sage-grouse used landscapes with a mosaic of habitats during breeding and in summerfall, > 95% of used locations had a sagebrush component, and sage-grouse consistently preferred local
patches dominated by sagebrush. Second, sage-grouse selected areas based on habitat features at
multiple scales. Models with sagebrush at multiple scales always outcompeted those with sagebrush at
only one scale, even after controlling for topography and other habitat types. Third, sage-grouse are a
“landscape” species. Birds consistently selected areas with more sagebrush habitat and less nonsagebrush habitat at large scales, even after controlling for local topography and the amount of sagebrush
within 100 m. Future assessments of habitat suitability for greater sage-grouse should consider not only
local-level metrics like shrub composition, height, and cover, but also topography and the amount of
sagebrush and non-sagebrush habitat at multiple scales (100 - 3200 m or more). Model results also
support ongoing efforts to reduce pinyon-juniper encroachment into sagebrush habitats in the PPR.
Other types of treatments may be appropriate if the habitat resulting from the treatment meets both locallevel and landscape-level sage-grouse habitat guidelines. CDOW can consult with landowners regarding
treatments on a case-by-case basis as needed. These maps provide a starting point for answering
additional ecological and management questions and for informing development planning, mitigation,
and conservation strategies for greater sage-grouse in the PPR. However, model results applied outside
the analysis area may not hold and should be interpreted with caution.
Large-scale changes to sagebrush ecosystems and historical population declines (Schroeder et al. 2004)
have raised concern about the status and conservation of greater sage-grouse (Centrocercus urophasianus) and
repeated petitions to list both the species and distinct population segments under the Endangered Species Act
(DOI 2005). Greater sage-grouse in the Parachute-Piceance-Roan (PPR) region of western Colorado are of
conservation concern due long-term declines in habitat suitability caused by pinyon-juniper encroachment and
potential impacts from increasing energy development. Both issues are listed as threat factors in the USFWS
listing decision in spring 2010 (DOI 2010). In 2006, the Colorado Division on Wildlife (CDOW) and industry
and agency partners initiated a study to obtain baseline data on sage-grouse seasonal habitat use, movements,
vital rates, and genetics for the PPR population. In 2008, the Colorado Greater Sage-grouse Conservation Plan
identified seasonal habitat mapping as a state-wide research priority. High-resolution maps showing
concentrated seasonal use areas would be valuable for improving sage-grouse conservation and development
planning. Pinyon-juniper removal has been proposed as a way to restore sage-grouse habitat and offset or
mitigate impacts of energy development in the PPR and elsewhere. However, we lack quantitative data on the
magnitude and timing of how sage-grouse respond to pinyon-juniper removal. The objectives of this study are
to: (1) use locations of marked sage grouse to generate high-resolution habitat-use maps for each season for the
entire PPR population and, (2) experimentally quantify greater sage-grouse response to removal of encroaching
pinyon-juniper using changes in winter track and pellet occupancy in a before-after control-treatment design.
This is the second of three reports. Here we summarize results of multi-scale seasonal habitat selection
analysis and seasonal habitat mapping efforts for breeding and summer-fall. Results of multi-scale seasonal
habitat selection analysis and seasonal habitat mapping efforts for winter will follow in a separate report.
SEASONAL HABITAT MAPPING
Identifying and delineating important seasonal habitats is critical for agencies, industry, and landowners
to make informed decisions about how to conserve key wildlife species in the face of landscape-level energy
development. However, in most areas, seasonal habitats for greater sage-grouse have not been adequately
mapped at a high enough resolution to be used in detailed planning, mitigation, and conservation efforts (Manly
2
�Document Date: 08-10-10
et al. 2002, Doherty et al. 2008). For this reason, mapping seasonal habitats is listed as a top priority in the
Colorado greater sage-grouse state-wide conservation plan (CGSSC 2008).
Managers also lack landscape-level guidelines for sage-grouse habitat. Sage-grouse are widely
considered a “landscape” species in that they require large areas of sagebrush habitat to persist (Schroeder et al.
1999; Connelly et al. 2000), but little quantitative data exists to evaluate that conclusion because most habitat
information for this species comes from studies that measure vegetation at small scales (e.g., 15 m radius plots;
Connelly et al. 2000, Hagen et al. 2007, CGSSC 2008). Research over the past four decades, including in the
PPR, has carefully documented local-level features of habitat used by sage-grouse in each season, including the
height, cover, and composition of shrubs, grasses, and forbs (for review, see Schroeder et al. 1999, Hagen et al.
2007, CGSSC 2008). However, only recently have studies begun to examine landscape-level habitat
requirements (Homer et al. 1993, Walker et al. 2007, Aldridge et al. 2008, Doherty et al. 2008, Doherty et al.
2010), and almost no published data are available to determine how much sagebrush habitat is required, and at
which scales, to maintain viable sage-grouse populations. It is also unclear how much non-sagebrush habitat at
which scales prevents use by sage-grouse. For that reason, it is imperative to test the influence of vegetation and
topography at multiple scales to determine which scale(s) sage-grouse use in selecting habitat and to generate
quantitative criteria for landscape-level conservation.
Recent advances in resource selection modeling and availability of high-resolution imagery allow
mapping the relative probability of habitat use at high resolution over large areas (Johnson et al. 2006, Doherty
et al. 2008). This approach also allows competing hypotheses to be addressed about the influence of local and
landscape factors at multiple scales on habitat selection and external validation of models against independent
datasets to ensure findings are robust (Boyce et al. 2002, Johnson et al. 2006).
The specific objectives of the multi-scale habitat selection analysis and seasonal habitat mapping
component of the PPR project are to: (1) generate high-resolution maps of important sage-grouse breeding,
summer-fall, and wintering habitat for the PPR population, (2) identify the appropriate scale at which habitat
features influence habitat use, and (3) quantify landscape-level habitat criteria. In this progress report, we
present results and maps from breeding and summer-fall habitat selection analyses.
METHODS
STUDY AREA
The study area encompassed the majority of occupied range of the PPR sage-grouse population as of
2006 as defined in the Colorado Greater Sage-grouse Conservation Plan (CGSSC 2008; Fig. 1) plus a 3 km
buffer to include some marked birds that moved outside this boundary as well as adjacent areas of unoccupied
habitat (Fig. 2). This area is a mix of public and private land with >20 major landowners (Fig. 3). Only part of
the study area could be used in analyses due to restrictions on access and logistical issues. For that reason, I
restricted use of data points to an “analysis area” within which radio-marked birds would have been regularly
available for tracking by field crews (Fig. 4). CDOW updated the occupied range boundary in July 2010 (Fig.
4). I applied results of the modeling to the entire 2010 occupied range boundary plus a 3 km buffer (including
the Magnolia section) because these areas were adjacent to and similar enough in vegetation and topography to
the analysis area that extrapolation seemed reasonable. However, model results applied outside the analysis area
may not hold and should be interpreted with caution.
FIELD METHODS
Capturing females in the PPR population is difficult due to poor access during the lekking season when
hens attend leks and because they are difficult to see in the tall, dense vegetation used by birds in spring,
summer, and fall. Field crews captured or recaptured and radio-marked 12 females from 3 Oct - 8 Dec 2008, 13
females from 1 Apr - 28 Apr 2009, and 22 females from 31 Aug - 10 Nov 2009 within the study area using
spotlights and hoop nets at night (Wakkinen et al. 1992), shoulder-mounted net-guns, and bumper-mounted net
launchers. All captured birds were sexed, aged, and fitted with aluminum, numbered leg bands and 17-g or 22-g
necklace-style radio collars (Advanced Telemetry Systems model A4060; Isanti, MN). Data from these birds
augmented an existing dataset of 85 radio-collared females and males monitored during 2006-2008 as part of the
original project. Field crews collected exact GPS locations (±10-15 m) on radio-collared birds approximately 12 times a week from September - November 2008, once approximately every 1-2 weeks from December 2008-
3
�Document Date: 08-10-10
March 2009, and approximately 1-2 times a week from April-November 2009. Field crews relocated missing
birds on telemetry flights from fixed-wing aircraft as needed; temporarily missing birds were monitored on
average less often than other birds. From September 2008 through July 2010, field crews recorded the major
habitat type and dominant and sub-dominant shrub species within 15 m of the exact location where marked birds
were found. Habitat types were classified as either aspen, barren, coniferous forest, grassland, mixed sagebrushmountain shrub, mountain shrub, sagebrush, road (dirt or gravel road or two-track), pipeline, or well pad.
MULTI-SCALE HABITAT SELECTION ANALYSES
Study design. I conducted habitat selection analyses using resource selection function approach (RSF). I
employed a used vs. pseudo-absence design rather than a used vs. available design to reduce contamination of
absence points (Keating and Cherry 2004). I pooled used locations of all marked individuals to make inferences
at the population level (Design II; Erickson et al. 2001, Manly et al. 2002). I then conducted logistic regression
on used (1) vs. pseudo-absence (0) points in R, version 2.11.0 (R Core Development Team, 2010).
Used vs. Pseudo-Absence Points. I included each location of a radio-marked bird once in the analysis as a used
location, with the exception that we considered each nest as only one location. I opted to retain locations in
which marked birds were found in a flock with another marked bird(s) because it is unclear statistically whether
such points are actually non-independent and doing so may bias analyses by giving less weight to flocks that had
more marked birds. I defined breeding-season locations as those during the pre-nesting, nesting, and early
brood-rearing periods (CGSSC 2008). I defined the start of the breeding season as 16 March in each year, as
that is when females begin moving from individual wintering areas to nesting areas. I identified the end date of
the breeding season in each year by adding 14 days to the date on which 95% of birds were estimated to have
completed nesting. I included locations from both successful and unsuccessful hens during this period. I
defined the end of the summer-fall season as November 30 in each year, as that when significant snowfall
occurs and birds shift from individual summer-fall ranges to wintering areas. I used locations collected from
2006-2009 to build models. Final data sets for building models contained 1072 breeding-season locations from
93 individuals and 1112 summer-fall locations from 67 individuals.
I generated pseudo-absence points within a portion of the study area referred to as the “analysis area”
where field crews would have had regular, authorized access for monitoring radio-marked birds (Fig. 4). I
considered all habitats within the analysis area as available for two reasons. First, marked sage-grouse have
shown long-distance movements within the study area, so we felt that all birds were capable of accessing any
part of the analysis area at any time within a given season. Second, avoidance of certain types of habitats is as
important for mapping probability of use as is preference for other habitat types, so it is important to include
areas of known non-habitat in analyses. To generate pseudo-absence points for each season, we randomly
selected available points from within the analysis area with the restriction that they could not fall within a
dissolved average daily movement distance buffer around used points for that season. Average daily
movements were 165 m during the breeding season and 240 m during summer and fall. The rationale behind
using average daily movement distance is that a marked bird could have used any point within that distance on
the same day without the field crew detecting the bird at that point. This sampling design is essentially a sample
of available points with undetected used points removed. In other words, it represents a set of pseudo-absence
points that marked birds were highly unlikely to have used. I used equal numbers of used and pseudo-absence
points in the breeding and summer-fall analyses because we had sufficient sample sizes for pseudo-absence
points to fully represent habitat types within the analysis area. I selected pseudo-absence points only from
within the analysis area; otherwise locations that marked birds used but where field crews lacked access would
have been overrepresented in the pseudo-absence sample.
Hypotheses and variables tested. Sage-grouse typically occur in sagebrush habitat and largely avoid nonsagebrush habitats and areas with rugged terrain (Hupp and Braun 1989, Homer et al. 1993, Connelly et al.
2000, Doherty et al. 2008). For that reason, all models included effects of sagebrush and non-sagebrush habitat
at various scales and topography. To test the hypothesis that sage-grouse are a “landscape” species, I allowed
models with sagebrush measured only at the smallest scale (a “local-level model”) to compete against models
4
�Document Date: 08-10-10
with sagebrush measured at larger scales (“landscape” models), and against models with additive effects of
sagebrush at both scales or an interaction of sagebrush variables across scales. Because it is well-established
that sage-grouse prefer sagebrush habitat at small scales (Connelly et al. 2000, Hagen et al. 2007), the landscape
hypothesis predicts that there should be a positive effect of sagebrush at a larger scale over and above a positive
effect of sagebrush at the smallest scale (100-m), after controlling for avoidance of non-sagebrush habitats and
topography. There are also two competing hypotheses about the diversity of habitat types that sage-grouse
prefer. One hypothesis suggests that sagebrush habitat has enough within-habitat variation in diversity and
structure of shrubs, grasses, and forbs to accommodate all their seasonal habitat needs – this hypothesis predicts
that sagebrush-only or sagebrush-dominated habitats will be the best predictor of habitat use. An alternative is
that sagebrush habitat by itself does not vary enough meet all their seasonal habitat needs so sage-grouse prefer
a mosaic of sagebrush, grassland, and sagebrush mixed with other shrubs (Crawford et al. 2004). This
hypothesis predicts that sage-grouse will most strongly select some combination of sagebrush, grassland, and
sagebrush-mountain shrub habitats. Sage-grouse diets in spring, summer, and fall consist of a mix of sagebrush,
forbs, and insects (Wallestad et al. 1975, Drut et al. 1994, Gregg et al. 2008), so I predicted that a combination
of sagebrush, grassland, and mixed sagebrush-mountain shrub habitat would be selected more strongly than
sagebrush-only or sagebrush-dominated habitats. Finally, I hypothesized that sage-grouse would avoid forested
habitats because they are commonly used by raptors that prey on sage-grouse, such as northern goshawk
(Accipter gentilis) and great horned owl (Bubo virginianus). This hypothesis predicts that sage-grouse will
strongly avoid areas with greater forested habitat, specifically aspen, conifer, and pinyon-juniper, and they will
prefer areas further from forest. Based on general habitat preferences, I also predicted that sage-grouse would
strongly avoid areas with other non-sagebrush habitat at large scales as well, such as mountain shrub-only and
barren habitats. Mountain shrub-only habitat typically lacks any sagebrush component and consists of
serviceberry [Amelanchier utahensis], Gambel oak [Quercus gambelii], antelope bitterbrush [Purshia
tridentata], mountain mahogany [Cercocarpus spp.], snowberry [Symphoricarpus spp.], wild currant [Ribes
spp.], and wild rose [Rosa spp.]).
I considered 15 continuous cover-type variables in each analysis. The first nine variables are
“sagebrush” variables: (1) proportion sagebrush-only habitat, (2) proportion sagebrush-dominated habitat, (3)
proportion mixed sagebrush-mountain shrub habitat, (4) proportion sagebrush-only + grassland habitat, (5)
proportion sagebrush-dominated + grassland habitat, (6) proportion sagebrush-only + mixed sagebrushmountain shrub habitat, (7) proportion sagebrush-dominated + mixed sagebrush-mountain shrub habitat
combined, (8) proportion sagebrush-only, mixed sagebrush-mountain shrub, and grassland habitat combined,
and (9) proportion sagebrush-dominated, mixed sagebrush-mountain shrub, and grassland habitat combined. I
also tested (10) proportion forested habitat, (11) proportion grassland habitat (if not included in sagebrush
metrics), (12) proportion mountain shrub-only habitat, (13) proportion barren habitat, (14) proportion riparian
habitat, and (15) distance to forest (linear and quadratic). I calculated all habitat metrics from cover types in a
classified state-wide vegetation layer developed by the Colorado Vegetation Classification Project (CVCP; Fig.
5). I combined cover types in the CVCP layer to generate a smaller number of more general habitat classes
relevant to the hypotheses being tested (Table 1).
I also considered topographic variables derived from a 10-m digital elevation model, including
elevation, slope, an index of terrain roughness, and aspect to address the influence of topography. The index of
terrain roughness was calculated as the standard deviation of the elevation of pixels within the buffer (Doherty
et al. 2008). Aspect was converted from degrees to a scale representing extent of southern exposure from 0-1 (0
= north; 1 = south) using the transformation: [1-cos([2π * aspect]/360)]/2. I predicted that sage-grouse would
use higher elevation areas with gentle slopes and low values of terrain roughness. I also predicted that sagegrouse would use areas with greater southern exposure during breeding because snow melts off earlier in spring,
exposing forbs and nesting shrubs sooner, and sagebrush on north-facing slopes is more likely to remain buried
under snow for longer. I predicted greater use of north-facing slopes in summer and fall because they remain
mesic for longer and should have a higher abundance of forbs and insects for longer. I had no a priori reason to
anticipate an effect of terrain ruggedness at large scales, so it was only measured at the three smallest scales
(100, 350, and 740 m). I did not include oil and gas or other types of infrastructure (roads, power lines, vehicle
traffic, etc.) as predictors because data on the distribution of these features were not available for each year
across the entire study area.
5
�Document Date: 08-10-10
Scale. I measured habitat variables using a circular buffer at six scales around used and pseudo-absence points
for all analyses: 100, 350, 740, 1000, 1600, and 3200 m (values indicate radius of the buffer). I selected these
scales, in part, to match those of studies in other parts of the species’ range (Doherty et al. 2010, Walker 2010).
The 100-m scale represents patch-level habitat selection, the scale at which birds can see and assess the habitat
around them. We used 100 m to minimize the influence of GPS unit error (± 26 m). I used the 350-m and
larger scales to represent landscape-level habitat selection, scales at which other ecological factors such as the
availability of escape cover or the distribution and abundance of predators, might influence habitat selection.
Because there is no a priori reason to think that any specific scale would be most influential, I selected values of
1000, 1600, and 3200 m (0.6 mi. 1 mi., and 2 mi.) to make them relevant to management.
Analyses. I first assessed support for each variable across scales to identify the scale(s) that best represented
sage-grouse habitat selection based on log-likelihood values. Variables > 2 AICc units below the best model and
for which 95% CIs of odds ratios overlapped 1.0 in univariate analyses were excluded from further
consideration. All other variables were used in various combinations to build the final a priori model set. All
models in the final model set represented biologically plausible alternative hypotheses for sage-grouse habitat
selection. I did not allow correlated variables (r > 0.7) or the same variable at different scales in the same
model, with the exception of sagebrush variables across scales. I checked for stability of regression coefficients
and associated standard errors in models with correlated variables (r > 0.4) and excluded models in which
regression coefficients switched signs or had grossly inflated standard errors. I used AICc values to assess
relative support for different models (Burnham and Andersen 2002). I then converted regression coefficients
from the best model into a spatially-explicit layer showing relative probability of use by applying them to GIS
layers using a resource selection function across the entire landscape. I conducted parametric bootstrapping in R
(version 2.11.0) to generate 95% confidence intervals for the effects of habitat variables on relative probability
of use. This involved generating a bootstrap dataset of 1000 sets of regression coefficients with the same
covariance structure as the best approximating model, then finding 2.5% and 97.5% cut-offs for those values.
Model Validation. I tested the robustness and predictive power of the best model for each season following
validation techniques outlined in Johnson et al. (2006). This involved: (1) dividing the RSF values into 5-6
ordinal bins, (2) calculating the midpoint RSF value and area for each bin, (3) calculating the expected number
of validation observations in each bin based on the area within that bin and probability of use from the best
approximating model, and (4) regressing the observed number of validation locations in each bin against the
expected number of locations in each bin. Models that fit the data should have a high R2 value, a slope of 1.0,
and an intercept not different than zero (Johnson et al. 2006). I used 58 locations collected in spring 2010 from
9 birds captured in fall 2009 (and not monitored in previous breeding seasons) to validate the breeding model.
For the summer-fall model, we randomly selected 20% of the individuals in the database and used 255 locations
from those 17 individuals for validation. No individuals or locations used to build models were used to test
models. I also overlaid seasonal habitat maps with locations of marked greater sage-grouse collected in the PPR
in 1997-1998 to see how well our maps predicted locations collected a decade earlier.
RESULTS
Seasonal Locations. Field crews visited 4370 locations of 114 marked females and 14 marked males (including
captures, nest visits, flight data, and mortalities) from Apr 2006 - July 2010 (Fig. 6). After removing duplicates
(e.g., multiple visits to the same nest), mortalities, imprecise flight locations, and locations outside the analysis
area, 3104 seasonal locations were available for analyses and model validation. Of those, 2434 locations were
collected during the breeding and summer-fall seasons.
Cover types and shrub species at used locations. At 1133 locations of marked birds visited during the breeding
and summer-fall seasons from 2008-2010, 45.3% had sagebrush as the primary habitat type within 15 m, 39.2%
had sagebrush-mountain shrub mix, 6.3% had grassland, 1.8% had pipeline cut, 3.1% had dirt or gravel roads or
two-tracks, 2.8% had mountain shrub, 0.6% had aspen, 0.5% had well pad (either old or new), and 0.3% had
6
�Document Date: 08-10-10
barren. Of 1120 locations with shrub species recorded, 84.9% had mountain big sagebrush (Artemisia tridentata
vaseyana) listed as the dominant shrub species, 8.1% had serviceberry, 1.6% had Gambel oak, 1.2% had yellow
(Chrysothamnus viscidiflorus) or rubber rabbitbrush (Ericameria nauseosa), 0.7% had aspen (Populus
tremuloides), 0.6% had snowberry, and the remaining 2.9% had no shrubs or other shrub species. Of the 50
locations in grassland habitat, 86% had mountain big sagebrush listed as a dominant or subdominant shrub,
indicating that most grassland habitat types included some sparse sagebrush within 15 m. Mountain big
sagebrush was present as a dominant or sub-dominant shrub within 15 m at 95.9% of 1120 breeding and
summer-fall locations with shrub species recorded; serviceberry was present at 54.9%; green rabbitbrush, rubber
rabbitbrush or broom snakeweed (Gutierrezia sarothrae) was present at 51.9%; Gambel’s oak was present at
20.2%; and aspen was present at 3.0%.
Breeding analyses. Ten variables were retained after univariate analysis and were incorporated into models in
the final model set: (1) proportion sagebrush-dominated habitat within 100 m, (2) proportion sagebrush +
grassland + mixed sagebrush-mountain shrub within 350 m, (3) proportion forest within 350 m, (4) proportion
mountain shrub habitat within 740-m, (5) proportion barren habitat within 100 m, (6) roughness within 100 m,
(7) distance to forest, (8) elevation, (9) slope, and (10) aspect (transformed). Two pairs of variables (roughness
within 100 m and slope; proportion sagebrush + grassland + mixed sagebrush-mountain shrub within 350 m and
forest within 350 m) were highly correlated; only one variable from each pair was allowed in each model and
those variables were allowed to compete. An additive model with sagebrush variables at the 100-m and 350-m
scales outcompeted models with effects of sagebrush at only one scale, quadratic sagebrush models, and models
with interactions of sagebrush variables across scales. However, in the final model set, a negative effect of
proportion forest within 350 m was a better predictor of use than was a positive effect of proportion sagebrush +
grassland + mixed sagebrush-mountain shrub combined within 350 m.
The best-supported model with stable coefficients and standard errors included additive effects of
proportion sagebrush-dominated habitat within 100 m, proportion forest within 350 m, roughness within 100 m,
distance to forest, and elevation (Table 2). Regression coefficients indicated that breeding sage-grouse preferred
patches of sagebrush-dominated habitat with locally flat topography (both at 100-m scale) within landscapes
further from forest, with less forest habitat (350-m scale), at higher elevation, and with less mountain shrub-only
habitat (740-m scale) (Table 2, Fig. 7). The resulting RSF map was divided into 6 ordinal bins that represent
areas with varying relative probability of use by breeding greater sage-grouse (Fig. 8). Together, RSF bins 5 and
6 (orange and red areas combined) represent 94.4% of predicted breeding habitat (Fig. 8). The remaining bins
each had < 5% relative probability of use.
Summer-fall analyses. Ten variables were retained after univariate analysis and were incorporated into models
in the final model set: (1) proportion sagebrush-dominated habitat within 100 m, (2) proportion sagebrush +
grassland + mixed sagebrush-mountain shrub within 350 m, (3) proportion forest within 740 m, (4) riparian
habitat within 1000 m, (5) proportion barren habitat within 100 m, (6) roughness within 100 m, (7) distance to
forest, (8) elevation, (9) slope, and (10) aspect (transformed). Two pairs of variables (roughness within 100 m
and slope; proportion sagebrush + grassland + mixed sagebrush-mountain shrub within 350 m and forest within
740 m) were highly correlated; only one variable from each pair was allowed in each model and those variables
were allowed to compete. An additive model with sagebrush variables at the 100-m and 350-m scales
outcompeted models with effects of sagebrush at only one scale, quadratic sagebrush models, and models with
an interaction of sagebrush variables across scales. However, a negative effect of proportion forest within 740
m was a better predictor of sage-grouse use than a positive effect of proportion sagebrush + grassland + mixed
sagebrush-mountain shrub within 350 m.
The best-supported model with stable coefficients and standard errors included additive effects of
proportion sagebrush-dominated habitat within 100 m, proportion forest within 740 m, proportion riparian
habitat within 1000 m, roughness within 100 m, elevation, and aspect (Table 3). Regression coefficients
indicated that sage-grouse in summer and fall preferred patches of sagebrush-dominated habitat (100-m scale)
within larger areas with less forest (740-m scale), locally flat topography (100-m scale), higher elevations, less
mountain shrub-only habitat (740-m scale), and less riparian habitat (1000-m) (Table 3, Fig. 9).
7
�Document Date: 08-10-10
The resulting RSF map was divided into 5 ordinal bins that represent areas with varying relative
probability of use by greater sage-grouse in summer and fall (Fig. 10). Together, RSF bins 4 and 5 (dark green
and light green areas combined) represent 98.4% of predicted summer-fall habitat (Fig. 10). The remaining bins
each had < 5% relative probability of use.
Validation. We were unable to divide RSF values into more than 5 bins for the summer-fall analysis. The
majority of validation locations in all three models fell in the top RSF bin (bin 5 or 6, depending on season). All
three models had reasonably high R2 values (0.912-0.984), slopes close to 1.0, and intercepts not statistically
different than 0.0. The breeding model slightly underestimated locations in RSF bin 4. The breeding validation
resulted in R2 = 0.9124, slope = 1.272 [95% CI: 0.725 to 1.819], and intercept = -2.627 [95% CI: -12.345 to
7.091] (Fig. 13a). The summer-fall validation resulted in R2 = 0.9838, slope = 1.116 [95% CI: 0.853 to 1.379],
and intercept = -5.931 [95% CI: -30.911 to 19.049] (Fig. 13b). Validation results predicted that the top two RSF
bins for would contain 94.5% of breeding locations and 98.4% of summer-fall locations.
Models performed well in predicting an independent dataset of marked greater sage-grouse locations
collected a decade earlier by CDOW in 1997-1998. The top two bins were predicted to contain 94.5% of
breeding locations and 93.5% of summer-fall locations from 1997-1998 (Figs. 13-14).
Landscape-level habitat guidelines. The majority of breeding sage-grouse used areas at high mean elevation
(2470 m, or ~8100 ft.) with relatively flat topography within 100 m (roughness index 4.8-9.5), 58-96% (mean
74%) sagebrush-dominated habitat within 100 m, 90-98% (mean 93%) sagebrush + grassland + mixed
sagebrush-mountain shrub habitat within 350 m, 0.5-6.5% (mean 4.6%) forested habitat within 350 m, 0.0-1.2%
(mean 0.8%) mountain shrub-only habitat within 350 m, and areas 140-314 m from forest (Table 4).
In summer and fall, the majority of sage-grouse used areas at a slightly higher mean elevation (2506 m,
or ~8200 ft.) with relatively flat topography within 100 m (roughness index 5.2-10.3), with 50-92% (mean 69%)
sagebrush-dominated habitat within 100 m, 88-98% (mean 92%) sagebrush + grassland + mixed sagebrushmountain shrub habitat within 350 m, 4.4-11.5% (mean 8.7%) forested habitat within 740 m, 0.0-1.3% (mean
0.8%) mountain shrub-only habitat within 740 m, 0.0% (mean 0.0%) riparian habitat within 1000 m, and areas
with more northeastern (262°-315°) or northwestern (45°-98°) exposure (mean was equivalent to 74° or 286°
from true north) (Table 5).
DISCUSSION
Greater sage-grouse in the PPR were more likely to spend their time in areas that had both more
sagebrush in the immediate vicinity (e.g., within 100 m), more sagebrush habitat at larger scales (350-m), and
less forested and mountain shrub-only habitat at larger scales (e.g., 350-740 m) in both seasons. In combination
with vegetation, terrain roughness within 100 m was a key predictor of sage-grouse use and consistently a better
predictor than slope or aspect. Models that measured selection for sagebrush habitats at two scales were much
more strongly supported in both analyses than single-scale models. Numerous previous studies have
documented the importance of sagebrush habitat at even smaller scales than we measured here (e.g., 15-25 m
scale; Connelly et al. 2000, Hagen et al. 2007). In combination, our finding of selection for sagebrush variables
at two scales (100-m and 350-m) and published evidence for selection at local scales (e.g., within 15 m)
suggests that sage-grouse are influenced by, and select habitat features at multiple scales. The scale at which
sage-grouse selected sagebrush habitat was smaller in the PPR (350 m) than in NW Colorado (740-1000 m;
Walker 2010). This may be because fewer sagebrush-dominated landscapes are available for use in the PPR and
birds are using the largest patches of sagebrush-dominated habitat that remain.
Field observations indicate that over 95% of all habitat types used by sage-grouse in all seasons had at
least some sagebrush component (i.e., some mountain big sage within 15 m). However, at 43 of those locations
where sagebrush was present, it was sparse enough that they were classified as grassland instead of sagebrush in
the CVCP layer. Interestingly, in both the breeding and summer-fall models, avoidance of forest was a better
overall predictor of where grouse occurred than was selection for habitats with a sagebrush component (after
controlling for mountain shrub-only habitat). This is likely because most of the habitat surrounding good sagegrouse habitat in the PPR (other than mountain shrub-only) is either aspen, conifer, or pinyon-juniper forest.
8
�Document Date: 08-10-10
Aspect did not play a role in the breeding model, but selection for more north-facing aspects in the summer-fall
model is consistent with the hypothesis that grouse use more mesic areas as grasses and forbs on ridge tops
desiccate in summer and fall.
Seasonal habitat maps support the classification of greater sage-grouse in the PPR as non-migratory,
despite the fact that some individuals made long-distance movements within the study area, both within and
across seasons. High-priority breeding and summer-fall habitats largely overlapped, with slightly more
predicted summer-fall habitat than breeding habitat.
From a practical standpoint, high-priority areas shown on seasonal habitat maps should be considered an
indicator that the appropriate cover types and topography are present at the appropriate scales and sage-grouse
are likely to occur there. On our maps, an absence of locations of radio-marked sage-grouse does not
necessarily indicate absence of sage-grouse, particularly in parts of the study area inaccessible to field crews or
where it was logistically difficult to capture birds. Seasonal habitat maps based on landscape-level features
measured in GIS provide a foundation for identifying high-priority habitat over large areas at fairly high
resolution, but where sage-grouse actually occur on the ground within those high-priority areas will also depend
on local-level habitat features, including percent sagebrush (and other shrub) cover, shrub height, live and
residual grass height, forb abundance, etc. (Connelly et al. 2000, Hagen et al. 2007, CGSSC 2008).
Additionally, error inherent in the CVCP cover type layer may contribute to inaccuracy in model predictions.
No post-classification accuracy assessment has been completed for the CVCP layer, so overall accuracy of the
habitat layer is unknown. Regardless, results indicate that managers should also consider both local-level and
landscape-level vegetation and topography when assessing habitat suitability for greater sage-grouse in the PPR.
Future on-the-ground assessments of habitat suitability should also include a GIS assessment of habitat types
within 350-3200 m derived from the CVCP layer and terrain roughness from a digital elevation model.
Landscape-level habitat and topography guidelines can also help identify appropriate areas for habitat
treatment. For example, treatments will only be effective if they are implemented in areas with appropriate
topography for sage-grouse (e.g., gently sloped, high elevation ridges). Treatments should also focus on areas
where habitat does not currently meet landscape-level guidelines, rather than areas that already do. Treatments
to increase breeding and summer-fall habitat in the PPR should set a goal of achieving areas with 50-96%
sagebrush-dominated habitat within 100 m, 88-96% sagebrush + grassland + mixed sagebrush-mountain shrub
within 350 m, 0.5-11.5% forest with 350 m, and < 1.2% of mountain shrub-only habitat, as measured in the
CVCP layer. If treatments are implemented with the goal of reducing the proportion of non-sagebrush habitat
types on the landscape, they will only be effective if what remains after treatment is habitat that sage-grouse will
use (e.g., removing aspen or conifer forest from an area without a sagebrush understory would not be effective).
Results from both models support ongoing efforts to reduce pinyon-juniper encroachment into sagebrush
habitats by BLM and CDOW in partnership with private landowners and energy companies. Other types of
treatments may be appropriate if the habitat resulting from the treatment results in habitat for sage-grouse that
meets both local-level and landscape-level guidelines. For example, where landscape-level guidelines have
already been met (e.g., areas within RSF bins 5-6 for breeding or bins 4-5 for summer-fall), treatments may still
be required to meet guidelines for sagebrush canopy cover and height, grass cover and height, etc. (CGSSC
2008). CDOW encourages a cooperative approach to treatments and can consult with landowners on a case-bycase basis.
RESEARCH TIMELINE
Field crews have completed data collection on radio-marked birds. Vegetation sampling will continue
through September 2010. Winter track surveys and summer pellet surveys will continue through 2012 to assess
sage-grouse response to pinyon-juniper removal (Table 6).
LITERATURE CITED
Aldridge, C. L., S. E. Nielsen, H. L. Beyer, M. S. Boyce, J. W. Connelly, S. T. Knick, and M. A. Schroeder.
2008. Range-wide patterns of greater sage-grouse persistence. Diversity and Distributions 14:983-994.
9
�Document Date: 08-10-10
Barnett, J. K., and J. A. Crawford. 1994. Pre-laying nutrition of Sage Grouse hens in Oregon. Journal of Range
Management 47:114-118.
Boyce, M. S., P. R. Vernier, S. E. Nielsen, and F. K. A. Schmiegelow. 2002. Evaluating resource selection
functions. Ecological Modeling 157:281-300.
Burnham, K. P. and D. R. Anderson. 2002. Model selection and inference: a practical information-theoretic
approach. Second edition. Springer-Verlag, New York, New York.
Colorado Greater Sage-grouse Steering Committee (CGSSC). 2008. Colorado greater sage-grouse conservation
plan. Colorado Division of Wildlife, Denver, CO, USA.
Connelly, J. W., M. A. Schroeder, A. R. Sands, and C. E. Braun. 2000. Guidelines to manage sage-grouse
populations and their habitats. Wildlife Society Bulletin 28:967-985.
Crawford, J. A., R. A. Olson, N. E. West, J. C. Moseley, M. A. Schroeder, T. D. Whitson, R. F. Miller, M. A.
Gregg, and C. S. Boyd. 2004. Ecology and management of sage-grouse and sage-grouse habitat. Journal
of Range Management 57:2-19.
Department of the Interior (DOI). 2005. 12-month finding for petitions to list the greater sage-grouse as
threatened or endangered. Federal Register 70(8): 2244-2282.
Department of the Interior (DOI). 2010. 12-month findings for petitions to list the greater sage-grouse
(Centrocercus urophasianus) as threatened or endangered. Federal Register 75(55): 13909-14014.
Doherty, K. E., D. E. Naugle, and B. L. Walker. 2010. Greater sage-grouse nesting habitat: the importance of
managing at multiple scales. Journal of Wildlife Management. In press.
Doherty, K. E., D. E. Naugle, B. L. Walker, and J.M. Graham. 2008. Sage-grouse winter habitat selection and
energy development. Journal of Wildlife Management. 72:187-195.
Drut, M. S., W. H. Pyle, and J. A. Crawford. 1994. Technical note: diets and food selection of Sage Grouse
chicks in Oregon. Journal of Range Management 47:90-93.
Erickson, W. P., T. L. McDonald, K. G. Gerow, S. Howlin, and J. W. Kerr. 2001. Statistical issues in resource
selection studies with radiotracked animals. Pages 209–242 in J. J. Millspaugh and J. M. Marzluff,
editors. Radio-tracking and animal populations. Academic Press, San Diego, California, USA.
Gregg, M. A., J. K. Barnett, J. A. Crawford. 2008. Temporal variation in diet and nutrition of preincubating
greater sage-grouse. Rangeland Ecology and Management 61:535-542.
Hagen, C. A., J. W. Connelly, and M. A. Schroeder. 2007. A meta-analysis of greater sage-grouse Centrocercus
urophasianus nesting and brood-rearing habitats. Wildlife Biology 13:42-50.
Homer, C. G., T. C. Edwards, Jr., R. D. Ramsey, and K. P. Price. 1993. Use of remote sensing methods in
modeling sage-grouse winter habitat. Journal of Wildlife Management 57:78-84.
Hupp, J. W., and C. E. Braun. 1989. Topographic distribution of sage-grouse foraging in winter. Journal of
Wildlife Management 53:823-829.
Johnson, C. J., S. E. Nielsen, E. H. Merrill, T. L. McDonald, and M. S. Boyce. 2006. Resource selection
functions based on use-availability data: theoretical motivation and evaluation methods. Journal of
Wildlife Management 70:347-357.
Manly, B. F. J., L. L. McDonald, D. L. Thomas, T. L. McDonald, and W. P. Erickson. 2002. Resource selection
by animals: statistical design and analysis for field studies. Second edition. Kluwer Academic,
Dordrecht, the Netherlands.
Schroeder, M. A., C. L. Aldridge, A. D. Apa, J. R. Bohne, C. E. Braun, S. D. Bunnell, J. W. Connelly, P. A.
Deibert, S. C. Gardner, M. A. Hilliard, G. D. Kobriger, C. W. McCarthy. 2004. Distribution of Sagegrouse in North America. Condor 106:363-376.
Schroeder, M. A., J. R. Young, and C. E. Braun. 1999. Sage-grouse (Centrocercus urophasianus). Account 425
in A. Poole and F. Gill, editors. The birds of North America. The Academy of Natural Sciences,
Philadelphia, PA, USA.
Wakkinen, W. L., K. P. Reese, J. W. Connelly, and R. A. Fischer. 1992. An improved spotlighting technique for
capturing sage-grouse. Wildlife Society Bulletin 20:425-426.
Walker, B. L. 2010. Hiawatha Regional Energy Development Project and greater sage-grouse conservation in
northwestern Colorado and southwestern Wyoming. Phase I: Winter and breeding habitat selection and
maps. Unpublished interim progress report. Colorado Division of Wildlife. Grand Junction, CO.
10
�Document Date: 08-10-10
Walker, B. L., D. E. Naugle, and K. E. Doherty. 2007. Greater sage-grouse population response to energy
development and habitat loss. Journal of Wildlife Management 72:2644-2654.
Wallestad, R., J. G. Peterson, and R. L. Eng. 1975. Foods of adult sage grouse in central Montana. Journal of
Wildlife Management 39:628-630.
ACKNOWLEDGEMENTS
Jeff Beck supervised data collection in 2006. Tony Apa supervised data collection in 2007-2008 and
was instrumental in the success of the field project. Karin Eichhoff conducted all GIS analyses. I thank our
hard-working field crew leaders Brandon Miller, Evan Phillips, Kaylan Kemink, Chris Binschus, and Carl
Bullock, as well as numerous field technicians and volunteers for collecting field data. Brad Petch, Brian
Holmes, Dan Neubaum, Kellen Keisling, and Tom Knowles assisted with project logistics and field work. I
thank Heather Sauls, Ed Hollowed, and Kent Walters at the Bureau of Land Management’s White River Field
Office for their ongoing support of the project and for logistical assistance and funding for pinyon-juniper
removal. Larry Gepfert piloted telemetry flights. Encana, Shell, Williams, and Conoco-Philips provided
funding and support for this project. The Colorado Division of Wildlife provided additional funding, housing,
and vehicles for field work. I thank staff at Encana, Conoco-Philips, Exxon-Mobil, Shell, Williams, ConocoPhillips, Chevron, and numerous private landowners and lessees for generously allowing crews access to private
lands within the study area for research. I thank J. C. Rivale, Bill deVergie, and Ron Velarde for allowing us to
use Little Hills State Wildlife Area facilities and J. C. Rivale for helping with snowmobile recovery.
11
�Document Date: 08-10-10
TABLE 2. Descriptions and habitat groupings for cover types in the Parachute-Piceance-Roan population,
Colorado from the Colorado Vegetation Classification Project (CVCP).
Class no.
6
24
31
34
104
109
110
111
112
114
11
13
18
19
32
103
22
33
23
35
40
28
48
49
27
46
47
43
44
53
54
55
56
57
58
62
63
66
67
69
78
79
81
84
105
108
91
92
93
Cover type description
Irrigated Ag
Greasewood
Sagebrush/Greasewood
Rabbitbrush/Grass Mix
Riparian
Shrub Riparian
Willow
Exotic Riparian Shrubs
Herbaceous Riparian
Water
Grass Dominated
Grass/Forb Mix
Foothill and Mountain Grasses
Disturbed Rangeland
Shrub/Grass/Forb Mix
Subalpine Grass/Forb Mix
Sagebrush Community
Sagebrush/Grass Mix
Saltbush Community
Sagebrush/Mesic Mountain Shrub Mix
Sagebrush/Rabbitbrush Mix
Snowberry/Shrub Mix
Mesic Mountain Shrub Mix
Serviceberry/Shrub Mix
Snowberry
Gambel Oak
Xeric Mountain Shrub Mix
Pinyon-Juniper
Juniper
Pinyon-Juniper-Oak Mix
Pinyon-Juniper-Sagebrush Mix
Pinyon-Juniper-Mountain Shrub Mix
Sparse Pinyon-Juniper/Shrub/Rock Mix
Sparse Juniper/Shrub/Rock Mix
Juniper/Sagebrush Mix
Aspen
Aspen/Mesic Mountain Shrub Mix
Engelmann Spruce/Fir Mix
Douglas Fir
Sub-Alpine Fir
Fir/Lodgepole Pine Mix
Douglas Fir/Engelmann Spruce Mix
Spruce/Fir/Aspen Mix
Douglas Fir/Aspen Mix
Forested Riparian
Conifer Riparian
Rock
Talus Slopes & Rock Outcrops
Soil
Habitat type
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Riparian
Grassland
Grassland
Grassland
Grassland
Grassland
Grassland
Sagebrush-only, sagebrush-dominated
Sagebrush-only, sagebrush-dominated
Sagebrush-dominated
Sagebrush-dominated, mixed sagebrush-mountain shrub
Sagebrush-dominated
Mixed sagebrush-mountain shrub
Mixed sagebrush-mountain shrub
Mixed sagebrush-mountain shrub
Mountain shrub
Mountain shrub
Mountain shrub
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Forest
Barren
Barren
Barren
12
�Document Date: 08-10-10
TABLE 2. Regression coefficients for variables in the best approximating model of greater sage-grouse
breeding habitat selection in the Parachute-Piceance-Roan population, Colorado, 2006-2009.
Variable
Estimate (β) ± SE
Intercept
-9.769 ± 1.831
Proportion sagebrush-dominated habitat within 100 m
1.807 ± 0.2561
Proportion forest within 350 m
-9.518 ± 0.9048
Proportion mountain shrub-only habitat within 740 m
-38.47 ± 4.481
Terrain roughness within 100 m
-0.1533 ± 0.01512
Elevation (m)
0.004767 ± 0.000716
Distance to forest (m)
0.001406 ± 0.0005747
TABLE 3. Regression coefficients for variables in the best approximating model of greater sage-grouse
summer-fall habitat selection in the Parachute-Piceance-Roan population, Colorado, 2006-2009.
Variable
Estimate (β) ± SE
Intercept
-36.186975 ± 3.081181
Proportion forest within 740 m
-16.169219 ± 1.168489
Proportion sagebrush-dominated habitat within 100 m
1.888706 ± 0.301398
Proportion mountain shrub-only within 740 m
-53.431828 ± 5.675204
Proportion riparian habitat within 1000 m
-99.478843 ± 50.591082
Terrain roughness within 100 m
-0.160485 ± 0.019555
Elevation (m)
0.016521 ± 0.001238
Aspect (transformed)
-2.548745 ± 0.281008
13
�Document Date: 08-10-10
TABLE 4. Summary of vegetation and topography variables at selected scales at 1072 breeding-season locations of marked greater sage-grouse in the
Parachute-Piceance-Roan population, Colorado, 2006-2009.
SageDom1
SageMixGrass1
Forest
Mtn. shrub
Roughness1
Elevation
Distance
100-m
350-m
350-m
740-m
100-m
(m)
to forest (m)
Mean ± SE
0.742 ± 0.008
0.930 ± 0.002
0.046 ± 0.002
0.008 ± 0.000
7.68 ± 0.12
2470 ± 2.06
244 ± 4.81
Median
0.816
0.956
0.024
0.003
6.90
2473
214
(25-75% quartiles)
(0.576-0.962)
(0.904-0.983)
(0.005-0.065)
(0.000-0.012)
(4.82-9.55)
(2427-2524)
(140-315)
(5-95% quantiles)
(0.250-1.000)
(0.784-1.000)
(0.000-0.0150)
(0.000-0.030)
(2.92-15.63)
(2351-2565)
(50-525)
1
SageDom 100-m = proportion sagebrush-dominated habitat within 100 m. SageMixGrass 350-m = proportion sagebrush + grassland + mixed sagebrush-mountain shrub habitat
within 350 m. Roughness 100-m = standard deviation of the elevation of pixels within 100 m.
TABLE 5. Summary of vegetation and topography variables at selected scales at 1112 summer-fall locations of marked greater sage-grouse in the
Parachute-Piceance-Roan population, Colorado, 2006-2009.
SageDom1
SageMixGrass1
Forest
Mtn. shrub
Riparian
Roughness1
Elevation
Distance
100-m
350-m
740-m
740-m
1000-m
100-m
(m)
to forest (m)
Mean ± SE 0.685 ± 0.008
0.923 ± 0.002
0.087 ± 0.002 0.008 ± 0.000 0.000 ± 0.000
8.12 ± 0.11
2445 ± 2.38
196 ± 3.96
Median
0.714
0.948
0.072
0.003
0.000
7.36
2444
173
(25-75%)
(0.500-0.922)
(0.881-0.986)
(0.044-0.115) (0.000-0.013) (0.000-0.000)
(5.20-10.31)
(2409-2487)
(98-268)
(5-95%)
(0.177-1.000)
(0.773-1.000)
(0.012-0.215) (0.000-0.030) (0.000-0.002)
(3.16-15.32)
(2397-2597)
(29-440)
1
SageDom 100-m = proportion sagebrush-dominated habitat within 100 m. SageMixGrass 350-m = proportion sagebrush + grassland + mixed sagebrush-mountain shrub habitat
within 350 m. Roughness 100-m = standard deviation of the elevation of pixels within 100 m. Aspect (transformed) = extent of southern exposure (1= due south, 0 = due north).
14
�Document Date: 08-10-10
TABLE 6. Updated timeline for greater sage-grouse research (seasonal habitat maps and assessment of pinyonjuniper removal) in the Parachute-Piceance-Roan population, Colorado.
Task
Initiation
Completion
Seasonal habitat use maps
GIS analyses and seasonal model development
31 Aug 2009
COMPLETE
Collect validation location dataset
1 Sep 2009
COMPLETE
Complete final model assessment and GIS map processing.
1 Mar 2010
COMPLETE
Prepare final report on breeding habitat-use maps
15 Jul 2010
COMPLETE
Prepare final report on summer-fall habitat-use maps
15 Jul 2010
COMPLETE
Prepare final report on winter habitat-use maps
31 Mar 2010
IN PROGRESS
Assessing response to pinyon-juniper removal
Identification of plots for pinyon-juniper removal
Winter track surveys, pellet collection (annually)
Remove encroaching pinyon-juniper (2010)
Analysis of winter track data (annually)
Analysis of genetic samples (annually, depends on no. samples)
Analysis of genetic data (annually)
Prepare cumulative report (annually)
Prepare cumulative final report
COMPLETE
1 Jan
20 Oct
1 Mar
1 Apr
1 Jun
1 Aug
1 Aug 2012
COMPLETE
1 Mar
1 Dec
1 Jun
1 Jun
1 Aug
1 Oct
1 Oct 2012
15
�Document Date: 08-10-10
FIGURES
W
Meeker
hi
te
ve
Ri
Northwest Colorado
Population
r
64
rt
No
or
hF
kW
h it
eR
iv e
r
Magnolia
ur
lp h
Su
Cr
Meeker/White River
Population
ee k
Hu n
ter
C re
ek
Cr
a
ek
ck
Bl
st D
ou g
la
re
We
RIO BLANCO
sC
Fa
wn
sC
re
e
ou
g la
ee
k
s
tD
k
Ea
Pi
c
ea
nc
re
ek
r eek
eC
W
es
tC
13
Parachute/Piceance/Roan
Population
325
New Castle
Ca
rr C
R o an
ek
C ree
k
Glenwood
Springs
Silt
6
Para
tC
ch u
r
6
k
e
Rifle
re e
GARFIELD
East
S al
te
Cr
ee
k
ash
Parachute
gS
al
tW
Bi
Battlement Mesa
6
De Beque
139
70
6
Collbran
MESA
70
65
330
Fruita
COLORADO
Towns
Streams
Counties
Land Ownership/Management
Habitat Status
Occupied
Potential
Vacant/Unknown
Easements
Private
City
Landownership based
on CoMAP
CDOW
SLB & State
Federal
BLM
NPS
USFS
4
2
0
4
8
12
Miles
Figure 1. Distribution of the Parachute-Piceance-Roan greater sage-grouse population as of 2006, including the Magnolia portion (CGSSC 2008).
16
�Document Date: 08-10-10
Figure 2. Occupied range as of 2006 and the study area boundary for the Parachute-Piceance-Roan greater sage-grouse population, Colorado.
The study area excluded the Magnolia portion of occupied range because we did not attempt to capture or track birds there.
17
�Document Date: 08-10-10
Figure 3. Surface ownership (major landowners only) in the Parachute-Piceance-Roan population of greater sage-grouse as of 2009.
18
�Document Date: 08-10-10
Figure 4. Occupied range as of 2010, the analysis area, and the extent of seasonal habitat mapping overlaid with active, inactive, and
unknown status leks in the Parachute-Piceance-Roan greater sage-grouse population, Colorado.
19
�Document Date: 08-10-10
Figure 5. Major habitat types derived from cover types in the Colorado Vegetation Classification Project and occupied range as of 2010 for the
Parachute-Piceance-Roan greater sage-grouse population, Colorado. See Table 1 for how CVCP cover types were grouped into habitat types.
20
�Document Date: 08-10-10
Figure 6. Locations of marked greater sage-grouse collected from 2006-2010 overlaid with occupied range as of 2010 for the ParachutePiceance-Roan population, Colorado. Not all areas were accessible in all seasons due to land ownership or logistical constraints.
21
�Document Date: 08-10-10
Figure 7. Relationships between landscape-level habitat variables and relative probability of use during the breeding season for greater sage-grouse in
the Parachute-Piceance-Roan population, Colorado, 2006-2009. In all models, values for other variables were set to the mean value at used locations.
22
�Document Date: 08-10-10
Figure 8. Breeding habitat map for greater sage-grouse in the Parachute-Piceance-Roan population based on vegetation, topography, and
breeding-season locations of marked birds, 2006-2009. Model predictions may not hold outside the analysis area boundary.
23
�Document Date: 08-10-10
Figure 9. Relationships between landscape-level habitat variables and relative probability of use during summer and fall for greater sage-grouse in
the Parachute-Piceance-Roan population, Colorado, 2006-2009. In all models, values for other variables were set to the mean value at used locations.
24
�Document Date: 08-10-10
Figure 10. Summer-fall habitat map for greater sage-grouse in the Parachute-Piceance-Roan population based on vegetation, topography, and
summer-fall locations of marked birds, 2006-2009. Model predictions may not hold outside the analysis area boundary.
25
�Document Date: 08-10-10
a
b
Figure 13. Regression of observed vs. expected no. of independent greater sage-grouse locations in each of 5-6 RSF bins for (a) breeding, and (b)
summer-fall habitat in the Parachute-Piceance-Roan population, 2006-2009. Dashed lines shows expected pattern under perfect model fit. Breeding
validation data are from spring 2010.
26
�Document Date: 08-10-10
Figure 14. Breeding-season map for greater sage-grouse in the Parachute-Piceance-Roan population from 2006-2010 overlaid with 109
breeding-season locations of marked greater sage-grouse from 1997-1998. 94.5% of previous breeding locations occurred in bins 5-6.
27
�Document Date: 08-10-10
Figure 15. Summer-fall map for greater sage-grouse in the Parachute-Piceance-Roan population from 2006-2010 overlaid with 246
summer-fall locations of marked greater sage-grouse from 1997-1998. 93.5% of these locations occurred in bins 4-5.
28
�Document Date: 08-10-10
29
�
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
Reports
Text
A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text.
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
Greater sage-grouse research in the Parachute-Piceance-Roan Region of western Colorado. Part II: multi-scale habitat selection and seasonal habitat mapping
Description
An account of the resource
Loss and degradation of sagebrush habitat throughout western North America has led to growing concern for conservation of greater sage-grouse (Centrocercus urophasianus) and repeated petitions to list the species under the Endangered Species Act. Greater sage-grouse in the Parachute- Piceance-Roan (PPR) region of western Colorado face at least two known potential stressors: increasing energy development and a long-term decline in habitat suitability associated with pinyon-juniper encroachment. In 2006, the Colorado Division of Wildlife (CDOW) and industry partners initiated a 3- year study to obtain baseline data on seasonal habitat use, movements, vital rates, and genetics of greater sage-grouse in the PPR. CDOW has since expanded that original project to include two new objectives: (1) generate high-resolution maps showing high-priority seasonal use areas for the entire population, and (2) assess the value of pinyon-juniper removal for increasing sage-grouse habitat. Industry, landowners, and state and federal agencies need high-resolution maps showing where sage-grouse occur during each season to streamline development planning, quantify mitigation needs, and guide on-the-ground sage-grouse conservation efforts.
Creator
An entity primarily responsible for making the resource
Walker, Brett L.
Subject
The topic of the resource
Greater sage-grouse
<em>Centrocercus urophasianus</em>
Parachute-Piceance-Roan (PPR) region
Wildlife habitat improvement
Northwestern Colorado
Extent
The size or duration of the resource.
29 pages
Date Created
Date of creation of the resource.
08-10-2010
Rights
Information about rights held in and over the resource
<a href="http://rightsstatements.org/vocab/NoC-NC/1.0/">No Copyright - Non-Commercial Use Only</a>
Type
The nature or genre of the resource
Text
Format
The file format, physical medium, or dimensions of the resource
application/pdf
Language
A language of the resource
English