Range expansion of a declining forest species, the Western Gray Squirrel ( Sciurus griseus ), into semiarid woodland

Abstract

Globally, animals that are range-restricted are frequently becoming species of conservation concern, in part due to competitive exclusion by phylogenetically and ecologically similar species that are more tolerant of human disturbance. However, climate and land use changes to natural landscapes can create pockets of refugia for range-restricted species. Western gray squirrels (Sciurus griseus) are native to the west coast of North America, principally California and western Oregon. Over the past several decades, Western Gray Squirrel populations have declined in human-dominated areas, with increased competition from introduced congeneric species native to eastern North America cited as a primary driver. Despite declines in their established range west of the Pacific Crest in western North America, western gray squirrels are extending their range into the Great Basin, where they were not historically found. Using a network of remote camera traps deployed across the Sierra Nevada–Great Basin ecotone in northwestern Nevada, we detected western gray squirrels across 16 of 100 camera-trapping sites. The majority of detections were located in piñon–juniper woodland, a land cover type not previously occupied by this species. Occupancy modeling revealed that western gray squirrels were equally likely to occur in piñon–juniper woodland compared to mature pine forest that they occupy elsewhere in their range. A species distribution model parameterized with historical gray squirrel observations (pre-1950), indicated increased climatic suitability for the species on the eastern side of the Sierra Nevada in recent decades, which may have facilitated this range expansion. Our findings reveal the potential for species declining in their historical range to colonize novel habitats that become increasingly suitable as a result of human-driven changes to ecosystems.

Full text

Journal of Mammalogy, 2024, XX, 1–11
https://doi.org/10.1093/jmammal/gyae011
Advance access publication 12 February 2024
Research Article
© The Author(s) 2024. Published by Oxford University Press on behalf of the American Society of Mammalogists, www.mammalogy.org.
Submitted 21 November 2022; Accepted 25 January 2024
Research Article
Range expansion of a declining forest species, the
Western Gray Squirrel (Sciurus griseus), into semiarid
woodland
Sean M. Sultaire1,*,, Robert A. Montgomery2,3,, Patrick J. Jackson4, Joshua J. Millspaugh1
1Wildlife Biology Program, University of Montana, 32 Campus Drive, Missoula, MT 59812, United States
2Department of Biology, University of Oxford, 11a Manseld Road, Oxford OX1 3SZ, United Kingdom
3Wildlife Conservation Research Unit, Department of Biology, The Recanati-Kaplan Centre, University of Oxford, Tubney House, Abingdon Road, Tubney, Oxon OX13
5QL, United Kingdom.
4Nevada Department of Wildlife, 6980 Sierra Center Parkway, Suite 120, Reno, NV 89511, United States
*Corresponding author: Wildlife Biology Program, University of Montana, 32 Campus Drive, Missoula, MT 59812, United States. Email: sultaires@gmail.com
Associate Editor was Elizabeth Flaherty
Abstract
Globally, animals that are range-restricted are frequently becoming species of conservation concern, in part due to competitive exclu-
sion by phylogenetically and ecologically similar species that are more tolerant of human disturbance. However, climate and land
use changes to natural landscapes can create pockets of refugia for range-restricted species. Western gray squirrels (Sciurus griseus)
are native to the west coast of North America, principally California and western Oregon. Over the past several decades, Western
Gray Squirrel populations have declined in human-dominated areas, with increased competition from introduced congeneric spe-
cies native to eastern North America cited as a primary driver. Despite declines in their established range west of the Pacic Crest in
western North America, western gray squirrels are extending their range into the Great Basin, where they were not historically found.
Using a network of remote camera traps deployed across the Sierra Nevada–Great Basin ecotone in northwestern Nevada, we detected
western gray squirrels across 16 of 100 camera-trapping sites. The majority of detections were located in piñon–juniper woodland, a
land cover type not previously occupied by this species. Occupancy modeling revealed that western gray squirrels were equally likely
to occur in piñon–juniper woodland compared to mature pine forest that they occupy elsewhere in their range. A species distribution
model parameterized with historical gray squirrel observations (pre-1950), indicated increased climatic suitability for the species on
the eastern side of the Sierra Nevada in recent decades, which may have facilitated this range expansion. Our ndings reveal the
potential for species declining in their historical range to colonize novel habitats that become increasingly suitable as a result of
human-driven changes to ecosystems.
Key words: camera trap, dynamic occupancy model, nonanalog community, piñon–juniper woodland, range expansion, species dis-
tribution model.
Many range-restricted species are being displaced by phyloge-
netically related species that are more tolerant of human dis-
turbance and expanding their ranges (McKinney and Lockwood
1999; Dugger et al. 2011; Angerbjörn et al. 2013). This displace-
ment often results from direct competition for resources and
the superior competitive ability of range-expanding species in
human-modied landscapes (Dugger et al. 2011; Jessen et al.
2018). For example, introduced American Mink (Neogale vison)
have accelerated population declines of native European Mink
(Mustela lutreola; Põdra and Gomez 2018)—in the North American
southwest black-tailed jackrabbits (Lepus californicus) are displac-
ing endemic white-sided jackrabbits (Lepus callotis; Desmond
2004), and barred owls (Strix varia) have outcompeted native
spotted owls (Strix occidentalis) over much of the Pacic Northwest
(Franklin et al. 2021). These are a few examples of a much broader
phenomenon associated with adaptable, colonizing species limit-
ing the range of native species. However, changes to landscapes
occurring from climate and land use change may create unex-
pected pockets of refugia for declining, range-restricted species
(Morelli et al. 2012).
Anthropogenic land use modications and climate change are
dynamically altering species distributions. Many animal species,
for instance, have colonized habitat at higher elevations and lati-
tudes as warming makes climatic conditions more suitable (Chen
et al. 2011; Wilkening et al. 2011; Sultaire et al. 2016). Ranges of
many range-restricted species adapted to cold climates will inev-
itably contract in response to climate change (Tayleur et al. 2016;
Stewart et al. 2017; Zhang et al. 2021), but for lower-elevation
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

2 | Sultaire et al.
species climate change may create additional habitat in areas
not previously suitable. As species shift their ranges in response
to climate change at different temporal scales, novel communi-
ties and ecosystems are forming within which previously isolated
species co-occur (Williams and Jackson 2007; Hobbs et al. 2009).
Extensive research has focused on predicting range shifts (Lawler
et al. 2006; Early and Sax 2011; Sirén and Morelli 2020) and forma-
tion of novel communities in response to future climate change
(Alexander et al. 2016), but fewer studies have documented
species colonizing novel environments in response to ongoing
climate change, which could inform conservation decisions for
species with limited ranges.
Western gray squirrels (Sciurus griseus) are native to the west
coast of North America, principally west of the Pacic Crest in
California and Oregon, with a relatively restricted range com-
pared to species of the genus from eastern North America (Fig. 1;
Carraway and Verts 1994). Throughout this range they are patch-
ily distributed, occurring only in low to mid-elevation pine and
oak forests while absent from cold, high-elevation conifer forests
(Carraway and Verts 1994). Although quantitative data are lack-
ing, available evidence suggests that populations of western gray
squirrels have declined in human-dominated areas (Muchlinski
et al. 2009; Cooper and Muchlinski 2015). This decline is gener-
ally attributed to competition with introduced congeneric species
native to eastern North America, eastern gray squirrels (Sciurus
carolinensis) and eastern fox squirrels (Sciurus niger; Jessen et al.
2018; Desharnais et al. 2022; Tran et al. 2022). At the northern
edge of their range in Washington state, for instance, Western
Fig. 1. Map of our study area near the border between the states of Nevada and California showing the location of 100 camera-trap sites (black
points). Sites that detected western gray squirrels from June 2018 to September 2020 are indicated in white. The dashed line is the border between
the Northwestern Forested Mountains Ecoregion (Sierra Nevada) to the west and the American Desert Ecoregion (Great Basin) to the east. This border
approximates the range of western gray squirrels after they colonized the Sierra Nevada foothills in the mid-1900s but before the species expanded
into piñon–juniper woodland in the Great Basin. The shaded area in map inset is the entire range of western gray squirrels in western North America.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

Journal of Mammalogy, 2024, Vol, XX, Issue XX | 3
Gray Squirrel populations have declined substantially enough
that they are listed as threatened in the state and only occur in
3 isolated populations (Ryan and Carey 1995; Stuart et al. 2018).
Despite declines in their established range west of the Pacic
Crest, western gray squirrels are extending their range into areas
where they were not previously documented (Matson et al. 2010;
Escobar-Flores et al. 2011; Rowe et al. 2015). In recent decades,
Western Gray Squirrel range has expanded into Ponderosa Pine
(Pinus ponderosa) forests in northeastern California (Matson et
al. 2010). The species has also established in a chaparral eco-
system > 250 km south of its historical range in Baja California,
Mexico, although a lack of historical surveys in the area may
explain this observation (Escobar-Flores et al. 2011). Western gray
squirrels were not considered part of the state of Nevada’s fauna
in the 1940s (Hall 1946; Orr 1949) but a museum record exists for
the species in the state from 1954 (www.vertnet.org), and later
publications included the small portion of the Sierra Nevada that
overlap Nevada as part of its range (Hall 1981; Carraway and Verts
1994; Mantooth and Riddle 2005; Fig. 1). As gray squirrels primar-
ily occupy pine and oak forest across their range, their presence
in the mixed-conifer forest of the Sierra Nevada is consistent with
their known habitat associations (Stuart et al. 2018; Mazzamuto
et al. 2020). However, east of the Sierra Nevada vegetation quickly
transitions to piñon–juniper woodland in semiarid mountain
ranges (Charlet 1996)—a land cover type that western gray squir-
rels are not reported to occur in, presumably limiting the eastern
edge of their range.
Here, we document the occurrence of western gray squirrels
within semiarid piñon–juniper woodlands in western Nevada via
a camera-trapping study conducted across a large spatial extent
(~15,000 km2). In addition to camera-trap detections, we used
museum records to demonstrate that these detections represent
a range extension into a novel habitat type. We further analyzed
gray squirrel camera detections in an occupancy model to quan-
tify differences in gray squirrel occurrence and detection between
piñon–juniper woodland and mixed-conifer forest. Finally, we
used a species distribution model (SDM) parameterized with his-
torical Western Gray Squirrel locations (pre-1950) to investigate
our prediction that the observed range expansion is related to a
recent trend toward warmer winters in the western Great Basin.
Our ndings reveal habitat exibility not previously documented
for this species declining across a large portion of its historical
range. We discuss the implications of continued environmental
change for the long-term persistence of western gray squirrels in
the Great Basin.
Materials and methods
Study area and eld methods
We conducted our camera-trapping study in west-central Nevada
and positioned our camera-trap network across the transi-
tion from the eastern Sierra Nevada to the Great Basin Desert
(Fig. 1). Vegetation in the Sierra Nevada consisted of primarily
mixed-conifer forest of Jeffrey Pine (Pinus jeffreyi) and White Fir
(Abies concolor) at lower elevations with Red Fir (Abies magnica)
and Western White Pine (Pinus monticola) more common at higher
elevations. Depending on re disturbance history, vegetation in
the Great Basin mountain ranges is a patchwork of piñon–juniper
woodland (Pinus monophyla and Juniperus osteosperma) and sage-
brush steppe (Aremesia tridentata). Lower elevations are domi-
nated by shrub-steppe and nonnative annual grasslands (Bromus
tectorum). Elevations in the region range from 1,100 to 3,400 m
and across study locations current average annual precipitation
(1991 to 2020) varied from 240 to 920 mm (PRISM Climate Group
2022). Reno, Nevada, is a large human population center (2020
population ~425,000) located at the northern end of the study
area (Fig. 1).
In Spring 2018, we deployed camera traps (Bushnell Trophy
Cam HD, Model 119776C) at 100 locations in the study area
(Fig. 1), as part of a regional mammal monitoring project, and
tracked animal detections until October 2020. Camera traps are
a demonstrated effective technique to survey tree squirrels, with
higher detection probability compared to the more traditionally
employed technique of live trapping (Diggins et al. 2016; Greene
et al 2016). As part of a more inclusive mammal community
assemblage study, scent lure was applied in the view of cameras
to increase detection of the target carnivore species, principally
black bears. Dates of luring ranged from June through August in
2018 and 2019 and July through September in 2020 and the appli-
cation of lure did not have a negative effect on the detection of
other smaller-bodied mammalian herbivores (e.g. lagomorphs;
Sultaire et al. 2023a). Camera-trap locations were set in a grid
with a 49 km2 resolution and camera traps were placed facing
north at the center of each grid cell (i.e. 7 × 7 km grid spacing). If
the grid cell center was inaccessible, we placed the camera trap at
the nearest accessible location. We did not sample locations that
were agricultural, had high human densities, or federally desig-
nated wilderness areas where we could not obtain permission
to deploy cameras. This sampling strategy resulted in deviations
from the 49-km2 grid, with the distance between a camera and the
closest neighboring camera averaging 6.75 km (range 3.99 to 17.6
km). Cameras were mounted ~1 m high on trees at sites located
in forest and woodland or on metal fence posts on sites located
in open areas. Cameras were aimed slightly toward ground level,
which better detected smaller-bodied species such as gray squir-
rels compared to aiming cameras parallel to the ground. During
summer each year, we revisited camera-trap locations on approx-
imately weekly intervals to reapply carnivore attractant. To max-
imize the period of time that cameras were operational between
checks, we set cameras to take a burst of 3 photos each time that
they were triggered and subsequently deactivate for 10 min.
Occupancy model
We identied species in all photos returned from our camera
traps using trained personnel. All photos classied as squirrels
were further reviewed by the lead author for accuracy because
photos of Tamiasciurus douglasii and Otospermophilus beecheyi were
occasionally classied as gray squirrels by personnel. We ana-
lyzed photo data with occupancy models, which are hierarchical
regression models that use repeat surveys of sites to estimate
species detection probability and provide estimates of site occu-
pancy probability corrected for false negatives (MacKenzie et al.
2002). We organized detections of western gray squirrels into
30-day detection histories, which resulted in ~8 secondary peri-
ods each year (excluding November through February, see below).
The choice of 8 secondary survey periods is consistent with the
number recommended for the scenario of low occupancy and
low detection probability (MacKenzie and Royle 2005), as might
be true for patchily distributed species at the edge of their range.
The average number of 30-day sampling periods that cameras
were operable across the 3 sampling years was 16 (SD = 4). All
sites were sampled in 2018 and 2019, with 85 of 100 sites sam-
pled in 2020 for a total of 457, 686, and 451 secondary periods in
each year, respectively. If a camera was inoperable during a given
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

4 | Sultaire et al.
month in the sampling season, we classied the secondary sam-
pling period for that site as missing data in detection histories,
which did not contribute to likelihood estimation.
To accommodate the inclusion of multiple years of data
from the same sites (2018 to 2020), we analyzed detection histo-
ries using a dynamic occupancy model (MacKenzie et al. 2003).
Dynamic occupancy models assume that the occupancy state of
a site is closed within years, but the occupancy state of a site
can change between years (MacKenzie et al. 2003). Thus, we t
these models to estimate occupancy probability in the initial year
of sampling and then transitions in site occupancy between sea-
sons (i.e. extinction and colonization). We excluded the 4 months
from November to February from detection histories to allow
time between primary sampling seasons for the occupancy state
of sites to change. Primary sampling seasons encompassed most
of the breeding season (dened as February to June by Linders et
al. 2004), excluding February when most cameras in the Sierra
Nevada were inoperable due to deep snow, and the months fol-
lowing when young remain close to natal areas (Vander Haegen
et al. 2005). We may have detected dispersing juveniles toward
the end of sampling seasons, but the coarseness of our sampling
grid (and occupancy covariates, see below) indicates these indi-
viduals would likely have dispersed from within the same study
grid cells.
Given this potential violation of the closure assumption, the
occupancy level of our model is best interpreted as the proba-
bility of site use at least once by western gray squirrels over a
sampling season (i.e. overlaps a home range; MacKenzie et al.
2006; Latif et al. 2015). The detection level of our model provided
inference on site-use frequency (i.e. within home range; Barry et
al. 2021)—hence, we quantied the effect of land cover at larger
spatial scales for occupancy covariates and comparatively ne
spatial scales for detection covariates. We used a 2.5-km buffer
surrounding cameras as a spatial scale to quantify occupancy
covariates. Within 2.5-km buffers, we calculated the proportion
of mixed-conifer forest (predominately in the Sierra Nevada),
piñon–juniper woodland, and the proportion of both land cover
types combined. We selected a 2.5-km radius because we pre-
dicted that western gray squirrels would be dependent on large
expanses of forest or woodland (i.e. piñon–juniper) in this sem-
iarid region and this distance approximates the mean recorded
dispersal distance for the species in the state of Washington (2.8
km; Vander Haegen et al. 2005). These occupancy covariates were
used to t 2 competing models: (1) a model that included a sin-
gle occupancy covariate that combined the proportion mixed-
conifer and piñon–juniper woodland; and (2) a model that
included separate occupancy covariates for proportion piñon–
juniper woodland and mixed-conifer forest. The 2 competing
models quantied whether gray squirrel occupancy varied in
relation to habitat type (i.e. piñon–juniper woodland vs. mixed-
conifer forest), or only in relation to habitat amount (i.e. propor-
tion conifer forest and piñon–juniper woodland combined).
Including habitat covariates into detection likelihood can
provide inference on site-use frequency if a site is occupied by
a species (i.e. within a home range; Barry et al. 2021). To capture
potential heterogeneity in detection probability driven by gray
squirrel responses to variation in forest structure, we included a
linear and quadratic effect of canopy height within 100 m of sur-
vey sites as detection covariates in both models. Including a quad-
ratic effect of canopy height explored differences in gray squirrel
site-use frequency between sparse piñon–juniper, mature piñon–
juniper (i.e. moderate canopy height), and mixed-conifer forest
(high canopy height) given that a site overlapped a Western Gray
Squirrel home range (i.e. occupied). Land cover and canopy height
data were sourced from the LANDFIRE 2016 database (https://
www.landre.gov/evt.php) at a 30-m resolution. We summarized
canopy height within a 100-m radius (~3 pixels) to characterize
the canopy height at camera locations while minimizing the
effects of potential individual pixel classication error. This spa-
tial scale (3.14 ha) is smaller than most reported summer home
range sizes of western gray squirrels (3.9 to 37.8 ha; Linders et
al. 2004) and approximates the size of core areas used by female
squirrels in Washington (4.29 ha; Vander Haegen et al. 2005).
Hence, this spatial scale was suitable to investigate within home
range patterns of habitat use by the species. We also included the
month of a secondary sampling period as a detection covariate, to
account for potential changes in gray squirrel behavior that could
inuence their detectability on cameras—for example, increased
movement during juvenile dispersal in the fall season. We did not
include covariates on colonization and extinction parameters (i.e.
intercept only) to limit model complexity.
We t the dynamic occupancy models in program R (R Core
Team 2023) using package “unmarked” version 1.1.1 (Fiske and
Chandler 2011). To rank the 2 competing occupancy models, we
used the version of Akaike Information Criterion adjusted for
small sample sizes, AICc. We assessed the t of the dynamic
occupancy model to detection data using the MacKenzie–Bailey
chi-squared test (MacKenzie and Bailey 2004), which compares
simulated detection histories under model parameters with true
detection histories and calculates whether the number of detec-
tions in each data set differs signicantly. The overdispersion
parameter ĉ is also calculated, with values of ĉ > 1.1 indicating
overdispersion of the data relative to model parameters. We per-
formed the goodness-of-t test using 999 simulations and the R
package “AICcmodavg” version 2.3.1 (Mazerolle 2020).
Species distribution model
To test our prediction that changes in regional climate over the
past century facilitated the colonization of western gray squirrels
into the western Great Basin, we t an SDM using historical obser-
vations of western gray squirrels and historical climate data. We
used the Global Biodiversity Information Facility (GBIF) to source
historical museum records of western gray squirrels prior to
1950 using the R package “rgbif” version 3.7.1 (Chamberlain et
al. 2022). This database yielded 669 total western gray squirrel
records, mostly from California and western Oregon, with a few
records from western Washington and northern Baja California
in Mexico. No pre-1950 records were found from Nevada, with the
earliest Nevada record occurring in 1954. We removed 9 records
that we considered unreliable because they came from desert
areas without habitat or offshore islands where the species does
not occur. As historical climate data did not extend outside of
the United States, we also removed 1 observation from northern
Baja California, Mexico. Of these 659 historical records, the earli-
est record was from 1800 and the median record year was 1913.
To develop a historical SDM for western gray squirrels we used
monthly historical weather data from PRISM (PRISM Climate
Group 2022), which provides monthly temperature (minimum
and maximum) and precipitation data for the contiguous United
States dating back to 1895 at a 4 × 4 km resolution. This resolu-
tion (16 km2) is comparable to the resolution at which we quan-
tied land cover covariates in the occupancy model (19.6 km2).
We used PRISM data to generate 30-year climate normals for
annual precipitation and minimum temperature of the coldest
quarter (December, January, and February) from 1910 to 1939 to
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

Journal of Mammalogy, 2024, Vol, XX, Issue XX | 5
characterize climate conditions across the range of western gray
squirrels prior to their range expansion into the eastern Sierra
Nevada. We used annual precipitation and minimum tempera-
ture of the colder quarter because we predicted that the distribu-
tion of western gray squirrels is limited by forest cover, which is
driven by precipitation in western North America, but also winter
severity because western gray squirrels are not a snow-adapted
species. Hence, including minimum temperature and precipita-
tion best represented the distribution of forest cover across their
range while also characterizing high-elevation forests with per-
sistent winter snow cover as unsuitable. We thinned the histori-
cal presence locations to the resolution of historical climate data,
1 per 4 × 4 km cell, which resulted in a total of 297 observations
used in the SDM. To t the SDM, we used a general additive model
algorithm (GAM), which can estimate nonlinear relationships
between presence and predictor variables for a species. To char-
acterize the background environment, we generated an equal
number of background points (297) that served as absence points
in the GAM. We constrained background point generation to ter-
restrial areas within a 100-km buffer around the distribution of
western gray squirrels as dened by the International Union for
the Conservation of Nature (NatureServe and IUCN 2008; Fig.
1). We excluded climate data pixels with Western Gray Squirrel
records from background point generation. We performed 5 runs
of the GAM, withholding 20% of the presence–background data
each time for model evaluation (Thuiller et al. 2009). We used
the receiver operator characteristic, also known as area under
the curve (AUC) to evaluate the ability of the model to predict
withheld data (Fielding and Bell 1997). We then used the model to
predict historical presence of western gray squirrels across their
range and used the most recent climate normals, 1991 to 2020,
to predict their current range with the model t with historical
observations. We t this GAM model in R package “biomod2” ver-
sion 3.5.1 (Thuiller et al. 2016).
Results
Across 67,556 camera-trap nights from June 2018 through October
2020, we detected western gray squirrels 656 times across 16
sites (Fig. 1). Summarizing these detections into 30-day intervals
resulted in a total of 90 monthly detections used to construct
detection histories (excluding November to February), including
41 monthly detections in 2018, 25 in 2019, and 24 in 2020. The
only gray squirrel detection at 1 site in the Great Basin occurred
in January and consequently was not included in detection his-
tories. The proportion of conifer forest and piñon–juniper wood-
land combined within a 2.5-km radius of sites ranged from 0.00 to
0.97. The composition of habitat surrounding sites almost com-
pletely transitioned between the Sierra Nevada and Great Basin—
the maximum proportion of mixed-conifer forest surrounding a
site in the Great Basin was 0.02 and the maximum proportion
piñon–juniper woodland in the Sierra Nevada was 0.06. A quarter
of the 100 sites completely lacked either conifer forest or piñon–
juniper woodland within 2.5 km.
The model that combined the proportion of mixed-conifer
forest and piñon–juniper woodland into 1 occupancy covariate
was ranked higher, based on AICc, than the model that esti-
mated separate effects of the 2 land cover types (Table 1). The
proportion of mixed-conifer forest and piñon–juniper woodland
combined within 2.5 km of sites had a positive effect on Western
Gray Squirrel occupancy (Table 1), suggesting that occupancy
was similar between landscapes with mixed-conifer forest and
landscapes with piñon–juniper woodland. For detection proba-
bility (i.e. frequency of site use), both the linear (β = 3.31; 95%
condence interval [CI] = 2.25 to 4.37) and quadratic (β = −3.36;
95% CI = −4.49 to −2.23) effects of canopy height were signicant
(Fig. 2). The month of secondary sampling period had a posi-
tive effect on gray squirrel detection probability (β = 0.16; 95%
CI = 0.01 to 0.32), providing support for increased gray squir-
rel detection later in the year. For transition parameters, the
better-supported model estimated a site-level extinction proba-
bility of 0.24 (95% CI = 0.09 to 0.50) and a colonization probabil-
ity of 0.08 (95% CI = 0.02 to 0.31), suggesting that the species may
have declined over the sampling period. The nal dynamic occu-
pancy model estimated squirrels to be present at 23 sites (95%
CI = 11 to 43) in the initial year of sampling after accounting
for imperfect detection. The goodness-of-t test based on 999
simulations did not indicate a lack of t of the better-supported
occupancy model to the gray squirrel detection data (χ2 = 866; P
= 0.69; ĉ = 0.53).
The SDM t to pre-1950 gray squirrel observations and cli-
mate data had good predictive ability with a mean AUC value
across 5 GAM runs of 0.79 (Table 2). The SDM was better at deter-
mining historical gray squirrel presence versus absence as the
mean sensitivity was higher than the mean specicity (Table
2). Precipitation was a more important factor limiting historical
Western Gray Squirrel distribution than minimum temperature,
with an average relative variable importance for precipitation
across the 5 GAM runs of 0.69 and 0.29 for minimum tempera-
ture. The GAM predicted that Western Gray Squirrel suitability
peaked at approximately ~900 mm of precipitation a year (Fig.
3A) and suitability was low when minimum temperature ranged
below −10 °C (Fig. 3B). Historically, suitability for the species was
low along the eastern Sierra Nevada including our study area in
Nevada (Fig. 4A and C), but suitability increased for predictions
under current climate conditions (Fig. 4B, D and E).
Discussion
We documented a range expansion of western gray squirrels, a
forest species that has declined in historical areas of their range,
into a semiarid woodland habitat that they did not previously
Table 1. Summaries of 2 competing dynamic occupancy models t to Western Gray Squirrel detections on camera traps deployed
in western Nevada, United States, 2018 to 2020. We t 2 competing occupancy models: (1) a model with an occupancy covariate that
grouped all habitat within 2.5 km of survey sites; and (2) a model that separated mixed-conifer forest and piñon–juniper woodland as
occupancy covariates. Both models included a quadratic effect of canopy height and an effect of secondary sampling period date as a
detection covariates. Coefcient estimates (β) are presented on the logit scale, with 95% CIs in parentheses. Dashes indicate covariates
not included in a model. The model that did not differentiate between habitat types was most supported based on AICc.
β—All habitat β—Mixed conifer β—Piñon–juniper AICc w
Model 1 1.98 (0.68, 3.28) — — 331.53 0.77
Model 2 — 1.52 (0.17, 2.04) 3.84 (0.61, 3.18) 333.95 0.23
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

6 | Sultaire et al.
occupy. Furthermore, this range expansion coincided with an
increase in climatic suitability, as measured by the climate pres-
ent in the species historical range, in the area of expansion along
the eastern Sierra Nevada and western Great Basin. Colonization
of a tree squirrel into a community that previously lacked arbo-
real species exemplies earlier predictions regarding the forma-
tion of nonanalog ecological communities in response to climate
change (Williams and Jackson 2007) and represents a possible
refugium for the species that faces pressures from introduced
species elsewhere in its range (Desharnais et al. 2022; Tran et al.
2022).
Our study was not explicitly designed to detect gray squirrels,
and the coarse resolution of the camera-trap grid did not allow
us to investigate ne-scale correlates of gray squirrel occurrence
(aside from canopy height). However, camera traps are a proven
effective technique for detecting tree squirrel species (Diggins et
al. 2016; Greene et al. 2016) and our results indicate reasonably
high detection of the species that peaked in locations with mod-
erate canopy height (Fig. 2). Monthly detection peaked at can-
opy heights characteristic of mature piñon–juniper woodland,
suggesting frequent but not transient use of this habitat. Detection
probability of western gray squirrels also varied across the sam-
pling season, with higher detection probability in the fall, which
may correspond to the dispersal period. This suggests a potential
violation of the closure assumption which can bias occupancy
estimates upwards (MacKenzie et al. 2002). However, the coarse
resolution of our sampling relative to Western Gray Squirrel
dispersal suggests that gray squirrels detected during dispersal
were likely dispersing from within the same grid cells and not
between sampling locations. Such movements likely explain the
strong effect of conifer cover within 2.5 km of camera locations
on gray squirrel occupancy, which is larger than gray squirrel
home ranges but within their typical dispersal distance. Hence,
we interpret occupancy as the probability gray squirrels used a
location over a sampling period (i.e. year; Latif et al. 2015), which
was explained by landscape-scale habitat amount (i.e. not habitat
type), with canopy height determining how frequently squirrels
used a given location (i.e. monthly probability of detection).
Documenting range expansions of species in response to
global change has been hindered by a lack of historical data
(Tingley and Beissinger 2009). Although we did not have access
to historical survey data, the literature we used to determine
the historical absence of gray squirrels in Nevada (Hall 1946)
was derived from extensive surveys of the Nevada mammal
fauna in the early 1900s, considered “the standard against
which other state surveys of mammals have been measured”
(Lawlor 1995:xiii; Riddle et al. 2014). Furthermore, the diurnal
habits of gray squirrels mean that they are easily detectable by
casual observers, and we conclude that this species would not
have gone undetected historically in these woodlands directly
adjacent to the second largest urban area in Nevada (i.e. Reno).
Instead, the strong relationship we found between pre-1950
Western Gray Squirrel records and the climate of that time
period suggests that the region had low climatic suitability for
the species historically.
Fig. 2. Predicted relationship between average canopy height within 100 m of sampling locations and monthly Western Gray Squirrel detection.
Prediction was generated with a dynamic occupancy model parameterized with 3 years of camera-trap data, 2018 to 2020, from 100 sites at the Sierra
Nevada–Great Basin Desert transition in Northern Nevada. Solid line represents the predicted relationship, and the dashed lines show the upper and
lower limits of the 95% CI. Tick marks represent values of canopy height at sampling locations where gray squirrels were detected at least once (top)
and where they were never detected (bottom).
Table 2. Evaluation criteria for our GAM SDM across the 5-fold
cross-validation. Sensitivity is the true positive rate; specicity is
the true negative rate.
Iteration AUCaSensitivity Specicity
Run 1 0.75 93.22 54.24
Run 2 0.81 86.44 66.10
Run 3 0.77 89.83 59.32
Run 4 0.79 77.97 71.19
Run 5 0.82 79.66 72.88
Mean 0.79 85.42 64.75
aArea under the curve.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

Journal of Mammalogy, 2024, Vol, XX, Issue XX | 7
SDMs are correlative and do not provide mechanistic insight
into climatic effects on species distributions (Pearson and Dawson
2003). Western gray squirrels are a forest species but like other
species of the genus Sciurus, they do not occur in cold environ-
ments characterized by deep and persistent winter snow cover.
This effect was well captured by our SDM, with low estimated
climatic suitability for western gray squirrels in areas with win-
ter temperatures colder than −10 °C and precipitation >1,500 mm
(i.e. high-elevation zones; Fig. 3). Piñon–juniper woodland grows
in elevations with semi-arid climate in the Great Basin, but the
ecosystem nonetheless regularly experiences below freezing
temperatures and snow cover (Koehn et al. 2021), unlike lower-
elevation forests on the west side of Pacic Crest where winters
are wet and mild. Across all study locations, average minimum
winter temperatures increased by an average of 2.4 °C (range =
1.6 to 3.4 °C) from the historical period used to t the SDM (1910
to 1939) and the current time period (1991 to 2020). Trends in pre-
cipitation between time periods were less dramatic and variable,
increasing at some sites while decreasing at others, with an aver-
age 30-mm increase across all sites between time periods. These
trends are consistent with those reported for the Great Basin over
the past 3 decades, with winters warming faster than summers
(Snyder et al. 2019). Given these strong temperature trends but
weak precipitation trends, the recent increase in suitability in the
study area suggests that the western Great Basin had sufcient
precipitation but was too cold to support the species historically
(Fig. 4C). We speculate that increases in annual minimum tem-
peratures in the Great Basin are potentially more conducive to
Western Gray Squirrel survival, allowing the species to colonize
the area.
In addition to recent winter temperature increases in the
Great Basin, warming in the high-elevation Sierra Nevada could
have facilitated gray squirrel colonization from west of the Sierra
Nevada crest. However, our SDM indicated reduced climatic suit-
ability in recent decades for the species along the high-elevation
crest of the Sierra Nevada west of the study area (Fig. 4C–E), likely
driven by increases in precipitation in the Sierra Nevada over
the past century (Dolanc et al. 2013). Consistent with this nd-
ing, western gray squirrels have not moved upslope in the high-
elevation central and southern Sierra Nevada over the past
century (Rowe et al. 2015). Hence, changing climate conditions
in the high-elevation Sierra Nevada were not responsible for the
squirrels colonizing east of the Sierra Nevada Crest.
Our SDM indicates that regional warming is likely related
to gray squirrels colonizing the western Great Basin, but other
changes to the landscapes in the region could also have facilitated
this expansion. Reno, Nevada, is a fast-growing urban area, with
recorded population growth > 160% since 1980 (Moffat 1996; U.S.
Census Bureau 2020). These rates of human population growth
have led to expansions in exurban development, particularly in
the Sierra Nevada (Raumann and Cablk 2008). Potential increases
in anthropogenic subsidies associated with the increased human
population, such as birdfeeders (Reed and Bonter 2018) and a lack
of introduced congeners, may have facilitated gray squirrel colo-
nization of the region, especially considering the colonization of
the eastern Sierra and Lake Tahoe area in the mid-1900s. However,
human population density was low across the areas of the Great
Basin that we sampled, with an average of 1% developed land
cover within 2.5 km of survey sites, indicating that gray squirrel
occurrence in these areas was not driven by human presence.
Furthermore, re suppression and overgrazing by livestock,
which removes competing herbaceous vegetation, have increased
the extent and shifted the structure of piñon–juniper woodlands to
higher tree densities since European settlement (Miller and Wigand
1994). As western gray squirrels are sensitive to habitat fragmen-
tation (Jessen et al. 2018), this densication may have increased
habitat suitability to arboreal squirrels. However, these changes
were underway decades before gray squirrels colonized these wood-
lands, with piñon–juniper increasing 33% in areas of the Great Basin
between the 1966 and 1995 (Weisberg et al. 2007). Other gray squirrel
species expanded rapidly into available habitat (e.g. S. carolinensis in
Europe; Teangana et al. 2000) and we would not expect a lag of sev-
eral decades for western gray squirrels to colonize these increasingly
dense woodlands. Hence, gray squirrel expansion in the western
Great Basin may be related to changes in the amount and structure
of woodland habitat but our historical SDM suggests that observed
changes in underlying climate were also necessary for western gray
squirrels to colonize the region.
Rather than being a localized phenomenon, our SDM, previ-
ous natural history publications, and citizen science observa-
tions indicate that this expansion has occurred across much
of the eastern edge of Western Gray Squirrel range. Our SDM
predicts that the Warner Mountains in far northeast California,
where a range expansion was previously documented (Matson
et al. 2010), is currently highly suitable (Fig. 4B) and there are
Fig. 3. Estimated response curves for the effect of average annual
precipitation (A) and average annual minimum temperature (B) on
Western Gray Squirrel climatic suitability during the historical period
(pre-1950) across 5 SDM runs (shading). For each run, 20% of the
presence and background data were withheld for model evaluation. Tick
marks represent values of climate variables at gray squirrel presence
(top) and background locations (bottom).
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

8 | Sultaire et al.
a few observations on the iNaturalist citizen science data-
base from recent years from the western Great Basin south of
the study area (www.inaturalist.org/observations/16410034).
Furthermore, areas in Northeastern Oregon predicted as cur-
rently suitable (Fig. 4B) also have iNaturalist observations of
the species from recent years (www.inaturalist.org/observa-
tions/15349967). Hence, climatic changes occurring in mid-
elevation areas east of the Pacic Crest appear to be facilitating
a broadscale shift in the distribution of this species to the east.
Our study further reveals that this expansion is not limited to
conifer forest but can also occur into piñon–juniper woodland
where available. Because we detected gray squirrels at 1 loca-
tion in pinyon–juniper >70 km from mixed-conifer forests in the
Sierra Nevada, the species appears to be reproducing in piñon–
juniper and these woodlands are not a population sink for the
species.
Ongoing environmental changes, however, may threaten the
long-term persistence of the species in the western Great Basin.
Warming to date may have facilitated expansion of the spe-
cies into the region but this could represent a nonequilibrium
state in which piñon–juniper woodland does not persist under
warmer conditions long term. Piñon–juniper woodland is not a
re-adapted ecosystem and increased frequency and extent of
wildres under warming may reduce the extent of this cover
type in the region. Although western gray squirrels can ben-
et from re in mixed-conifer forest (Mazzamuto et al. 2020),
Fig. 4. Predicted suitability for western gray squirrels for 2 different time periods and spatial extents from an SDM t solely with pre-1950
observations of the species. Panel A maps suitability for historical time period across the species range (pre-1950) and panel B maps predicted
suitability for the current time period (1991 to 2020 climate normals). The bottom panels map suitability for each time period (pre-1950 panel C; 2020
panel D) in the study area region, with panel E showing the change in suitability between time periods. Purple points are historical locations used to
t the SDM in panels A and C. Post-1950 Western Gray Squirrel observations from the study region in panel D. We show only post-2010 points in panel
B for interpretability. Locations from those time periods were not used to t the SDM. Blue points in center panel are camera-trap locations where
western gray squirrels were detected at least once, 2018 to 2020, and black points are locations where the species was not detected.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

Journal of Mammalogy, 2024, Vol, XX, Issue XX | 9
postre succession back to piñon–juniper woodland is a slow
process that takes several decades to centuries (Koniak 1985;
Gruell 1999). Already, 1 location where we frequently detected
gray squirrels burned in an 8,300-ha wildre in 2020 and will
not support gray squirrel habitat for at least several decades.
If wildre increases in frequency in the region and reduces the
area of piñon–juniper woodland, the range expansion of gray
squirrels could be negated. Furthermore, Piñon Pine is a mast-
ing species and there is evidence of declining mast production
during hotter years (Redmond et al. 2012). Piñon Pine nuts are
likely the main food resource that gray squirrels are exploiting
in piñon–juniper woodlands, as detected in our camera-trap
photos (Fig. 5), and continued persistence of this species in the
Great Basin is likely dependent on this resource. Regarding
the potential for continued range expansion for the species in
the Great Basin, the arid Lahontan Trough directly to the east
of the study area will likely act a dispersal barrier, preventing
gray squirrels from colonizing extensive areas of piñon–juniper
in the central Great Basin, as has been the case for other forest
species (Sultaire et al. 2023b). Continued monitoring will help
determine the sustainability of this recently established popu-
lation. Although western gray squirrels may continue to decline
where they were previously abundant, our results suggest the
presence of a possible refugium for the species east of their
historical distribution.
Acknowledgments
We thank eld technicians W. Ortiz, D. Heit, J. Eaton, S. Miller,
L. Margadant, N. Everett, and C. Blommel for maintaining the
camera-trap grid and classifying photos. We thank the Pyramid
Lake Paiute Tribe for permission to deploy cameras on tribal land.
Author contributions
RAM, JJM, and PJJ conceived the study; RAM and JJM designed the
eld sampling with input from PJJ and supervised eld data col-
lection; SMS led the nal year of data collection, conducted the
analysis, and wrote the initial manuscript draft; and all authors
contributed to revising the nal manuscript.
Funding
This research was supported by the Nevada Department of
Wildlife $3 Predator Fee Program and the US Fish and Wildlife
Service with Wildlife Restoration funds. This manuscript is solely
the responsibility of the authors and does not necessarily repre-
sent the ofcial views of the Nevada Department of Wildlife or
the US Fish and Wildlife Service.
Conict of interest
None declared.
Data availability
The data and code that support the ndings in this manuscript
are available from the corresponding author upon reasonable
request.
References
Alexander JM, Diez JM, Hart SP, Levine JM. 2016. When climate
reshufes competitors: a call for experimental macroecology.
Trends in Ecology and Evolution 31(11):831–841. https://doi.
org/10.1016/j.tree.2016.08.003
Angerbjörn A, Eide NE, Dalén L, Elmhagen B, Hellström P, Ims RA,
Killengreen S, Landa A, Meijer T, Mela M, et al. 2013. Carnivore
conservation in practice: replicated management actions on
a large spatial scale. Journal of Applied Ecology 50(1):59–67.
https://doi.org/10.1111/1365-2664.12033
Barry BR, Moriarty K, Green D, Hutchinson RA, Levi T. 2021.
Integrating multi-method surveys and recovery trajectories
into occupancy models. Ecosphere 12(12):e03886. https://doi.
org/10.1002/ecs2.3886
Carraway LN, Verts BJ. 1994. Sciurus griseus. Mammalian Species
474(474):1–7. https://doi.org/10.2307/3504097
Chamberlain S, Barve V, Mcglinn D, Oldoni D, Desmet P, Geffert L, Ram
K. 2022. rgbif: interface to the Global Biodiversity Information
Facility API. R package version 3.7.1. https://CRAN.R–project.
org/package=rgbif.
Charlet DA. 1996. Atlas of Nevada Conifers: a phytogeographic refer-
ence. Reno (NV, USA): University of Nevada Press.
Chen IC, Hill JK, Ohlemüller R, Roy DB, Thomas CD. 2011. Rapid
range shifts of species associated with high levels of climate
warming. Science 333(6045):1024–1026. https://doi.org/10.1126/
science.1206432
Cooper DS, Muchlinski AE. 2015. Recent decline of lowland populations
of the Western Gray Squirrel in the Los Angeles area of southern
California. Bulletin Southern California Academy of Sciences
114(1):42–53. https://doi.org/10.3160/0038-3872-114.1.42
Desharnais RA, Muchlinski AE, Ortiz JL, Alvidrez RI, Gatza BP. 2022.
Timescale analyses of uctuations in coexisting populations
of a native and invasive tree squirrel. Ecology and Evolution
12(4):e8779. https://doi.org/10.1002/ece3.8779
Desmond MJ. 2004. Habitat associations and co-occurrence of
Chihuahuan Desert hares (Lepus californicus and L. callotis).
The American Midland Naturalist 151(2):414–419. https://doi.
org/10.1674/0003-0031(2004)151[0414:haacoc]2.0.co;2
Fig. 5. A camera-trap photo of a Western Gray Squirrel (Sciurus griseus)
carrying a piñon pinecone (Pinus monophylla) at a site located in piñon–
juniper woodland in western Nevada. In this manuscript we present
the rst documentation of this species in piñon–juniper woodland,
representing a range extension for this species into the Great Basin that
may be facilitated by regional warming.
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

10 | Sultaire et al.
Diggins CA, Gilley LM, Kelly CA, Ford WM. 2016. Comparison of
survey techniques on detection of northern ying squirrels.
Wildlife Society Bulletin 40(4):654–662. https://doi.org/10.1002/
wsb.715
Dolanc CR, Thorne JH, Safford HD. 2013. Widespread shifts in the
demographic structure of subalpine forests in the Sierra Nevada,
California, 1934 to 2007. Global Ecology and Biogeography
22(3):264–276. https://doi.org/10.1111/j.1466-8238.2011.00748.x
Dugger KM, Anthony RG, Andrews LS. 2011. Transient dynamics of
invasive competition: barred owls, spotted owls, habitat, and
the demons of competition present. Ecological Applications
21(7):2459–2468. https://doi.org/10.1890/10-2142.1
Early R, Sax DF. 2011. Analysis of climate paths reveals potential
limitations on species range shifts. Ecology Letters 14(11):1125–
1133. https://doi.org/10.1111/j.1461-0248.2011.01681.x
Escobar-Flores JG, Ruiz-Campos G, Guevara–Carrizales AA,
Martínez-Gallardo R. 2011. Extension of southern range and
new specimens of the Western Gray Squirrel, Sciurus griseus
anthonyi (Mammalia: Sciuridae), in Baja California, México.
Western North American Naturalist 71(1):119–120. https://doi.
org/10.3398/064.071.0117
Fielding AH, Bell JF. 1997. A review of methods for the assessment
of prediction errors in conservation presence/absence mod-
els. Environmental Conservation 24(1):38–49. https://doi.
org/10.1017/s0376892997000088
Fiske I, Chandler R. 2011. Unmarked: an R package for tting hier-
archical models of wildlife occurrence and abundance. Journal
of Statistical Software 43(10):1–23. https://doi.org/10.18637/jss.
v043.i10
Franklin AB, Dugger KM, Lesmeister DB, Davis RJ, Wiens JD, White
GC, Nichols JD, Hines JE, Yackulic CB, Schwarz CJ, et al. 2021.
Range-wide declines of Northern Spotted Owl populations in
the Pacic Northwest: a meta-analysis. Biological Conservation
259(2021):109168. https://doi.org/10.1016/j.biocon.2021.109168
Greene DU, McCleery RA, Wagner LM, Garrison EP. 2016. A comparison
of four survey methods for detecting fox squirrels in the south-
eastern United States. Journal of Fish and Wildlife Management
7(1):99–106. https://doi.org/10.3996/082015-jfwm-080
Gruell GE. 1999. Historical and modern roles of re in pinyon–juniper.
In: Monsen SB, Stevens R, editors. Proceedings: ecology and man-
agement of pinyon–juniper communities within the interior west:
15–18 Sep 1997, Provo (UT, USA). US Department of Agriculture,
Forest Service, Rocky Mountain Research Station; p. 24–28.
Hall ER. 1946. Mammals of Nevada. Berkeley (CA, USA): University
of California Press.
Hall ER. 1981. The mammals of North America. Vol. 1. 2nd ed. New
York (NY, USA): Wiley.
Hobbs RJ, Higgs E, Harris JA. 2009. Novel ecosystems: implications for
conservation and restoration. Trends in Ecology and Evolution
24(11):599–605. https://doi.org/10.1016/j.tree.2009.05.012
Jessen T, Wang Y, Wilmers CC. 2018. Habitat fragmentation provides
a competitive advantage to an invasive tree squirrel, Sciurus
carolinensis. Biological Invasions 20(3):607–618. https://doi.
org/10.1007/s10530-017-1560-8
Koehn CR, Petrie MD, Bradford JB, Litvak ME, Strachan S. 2021. Seasonal
precipitation and soil moisture relationships across forests
and woodlands in the southwestern United States. Journal of
Geophysical Research, Biogeosciences 126(4):e2020JG005986.
https://doi.org/10.1029/2020JG005986
Koniak S. 1985. Succession in pinyon–juniper woodlands follow-
ing wildre in the Great Basin. The Great Basin Naturalist
45(3):556–566.
Latif Q, Ellis MM, Amundson CL. 2015. A broader denition of
occupancy: a reply to Hayes and Monls. Journal of Wildlife
Management 80(2):192–194. https://doi.org/10.1002/jwmg.1022
Lawler JJ, White D, Neilson RP, Blaustein AR. 2006. Predicting cli-
mate‐induced range shifts: model differences and model reli-
ability. Global Change Biology 12(8):1568–1584. https://doi.
org/10.1111/j.1365-2486.2006.01191.x
Lawlor TE. 1995. Foreword to the 1995 edition. In: Hall ER, editor.
Mammals of Nevada. Reno (NV, USA): University of Nevada
Press; p. xiii–xvi.
Linders MJ, West SD, Vander Haegen WM. 2004. Seasonal variability
in the use of space by western gray squirrels in southcentral
Washington. Journal of Mammalogy 85(3):511–516. https://doi.
org/10.1644/1545-1542(2004)085<0511:svituo>2.0.co;2
MacKenzie DI, Bailey LL. 2004. Assessing the t of site-occupancy
models. Journal of Agricultural Biological and Environmental
Statistics 9(3):300–318. https://doi.org/10.1198/108571104x3361
MacKenzie DI, Nichols JD, Hines JE, Knutson MG, Franklin AB. 2003.
Estimating site occupancy, colonization, and local extinction
when a species is detected imperfectly. Ecology 84(8):2200–2207.
https://doi.org/10.1890/02-3090
MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle JA, Langtimm
CA. 2002. Estimating site occupancy rates when detection prob-
abilities are less than one. Ecology 83(8):2248–2255. https://doi.
org/10.1890/0012-9658(2002)083[2248:ESORWD]2.0.CO;2
MacKenzie DI, Nichols JD, Royle JA, Pollock KH, Bailey L, Hines JE.
2006. Occupancy estimation and modeling: inferring patterns
and dynamics of species occurrence. Amsterdam, Netherlands:
Elsevier.
MacKenzie DI, Royle JA. 2005. Designing occupancy stud-
ies: general advice and allocating survey effort.
Journal of Applied Ecology 42(6):1105–1114. https://doi.
org/10.1111/j.1365-2664.2005.01098.x
Mantooth SJ, Riddle BR. 2005. Annotated checklist of the recent
mammals of Nevada. Occasional Papers, Museum of Texas Tech
University 245:1–19.
Matson JO, Laudenslayer WF Jr, Trewhitt SA, Flores M. 2010. Range
expansion of the Western Gray Squirrel (Sciurus griseus) in
northeastern California. Western North American Naturalist
70(2):266–269. https://doi.org/10.3398/064.070.0219
Mazerolle MJ. 2020. AICcmodavg: model selection and multimodel
inference based on (Q)AIC(c). R package version 2.3-1. https://
CRAN.R-project.org/package=AICcmodavg.
Mazzamuto MV, Mazzella MN, Merrick MJ, Koprowski JL. 2020. Fire
impacts on a forest obligate: Western Gray Squirrel response to
burn severity. Mammalian Biology 100(3):305–305. https://doi.
org/10.1007/s42991-020-00034-w
McKinney ML, Lockwood JL. 1999. Biotic homogenization: a few win-
ners replacing many losers in the next mass extinction. Trends
in Ecology & Evolution 14(11):450–453. https://doi.org/10.1016/
s0169-5347(99)01679-1
Miller RF, Wigand PE. 1994. Holocene changes in semiarid pinyon–
juniper woodlands. BioScience 44(7):465–474. https://doi.
org/10.2307/1312298
Moffat RM. 1996. Population history of western US cities and towns,
1850–1990. Lanham (MD, USA): Scarecrow Press.
Morelli TL, Smith AB, Kastely CR, Mastroserio I, Moritz C, Beissinger
SR. 2012. Anthropogenic refugia ameliorate the severe
climate-related decline of a montane mammal along its trailing
edge. Proceedings of the Royal Society of London, B: Biological
Sciences 279(1745):4279–4286. https://doi.org/10.1098/
rspb.2012.1301
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024

Journal of Mammalogy, 2024, Vol, XX, Issue XX | 11
Muchlinski AE, Stewart GR, King JL, Lewis SA. 2009. Documentation
of replacement of native western gray squirrels by intro-
duced eastern fox squirrels. Bulletin Southern California
Academy of Sciences 108(3):160–162. https://doi.
org/10.3160/0038-3872-108.3.160
NatureServe, IUCN (International Union for Conservation of Nature).
2008. Sciurus griseus. The IUCN red list of threatened species.
Version 2022-2. [accessed 3 Mar 2022]. https://www.iucnredlist.org.
Orr RT. 1949. Mammals of Lake Tahoe. San Francisco (CA, USA):
California Academy of Sciences.
Pearson RG, Dawson TP. 2003. Predicting the impacts of climate
change on the distribution of species: are bioclimate envelope
models useful? Global Ecology and Biogeography 12(5):361–371.
https://doi.org/10.1046/j.1466-822x.2003.00042.x
Põdra M, Gómez A. 2018. Rapid expansion of the American Mink
poses a serious threat to the European Mink in Spain. Mammalia
82(6):580–588. https://doi.org/10.1515/mammalia-2017-0013
PRISM Climate Group. 2022. Oregon State University. [data created
2 Aug 2013; accessed 18 Mar 2022]. https://prism.oregonstate.
edu.
R Core Team. 2023. R: a language and environment for statistical
computing. Version 4.2.3. Vienna (Austria): R Foundation for
Statistical Computing. www.R-project.org/.
Raumann CG, Cablk ME. 2008. Change in the forested and developed
landscape of the Lake Tahoe basin, California and Nevada, USA,
1940–2002. Forest Ecology and Management 255(8–9):3424–
3439. https://doi.org/10.1016/j.foreco.2008.02.028
Redmond MD, Forcella F, Barger NN. 2012. Declines in pinyon pine
cone production associated with regional warming. Ecosphere
3(12):1–14. https://doi.org/10.1890/es12-00306.1
Reed JH, Bonter DN. 2018. Supplementing non-target taxa: bird
feeding alters the local distribution of mammals. Ecological
Applications 28(3):761–770. https://doi.org/10.1002/eap.1683
Riddle BR, Jezkova T, Hornsby AD, Matocq MD. 2014. Assembling
the modern Great Basin mammal biota: insights from molec-
ular biogeography and the fossil record. Journal of Mammalogy
95(6):1107–1127. https://doi.org/10.1644/14-mamm-s-064
Rowe KC, Rowe KM, Tingley MW, Koo MS, Patton JL, Conroy CJ,
Perrine JD, Beissinger SR, Moritz C. 2015. Spatially heteroge-
neous impact of climate change on small mammals of mon-
tane California. Proceedings of the Royal Society of London, B:
Biological Sciences 282(1799):20141857. https://doi.org/10.1098/
rspb.2014.1857
Ryan LA, Carey AB. 1995. Distribution and habitat of the Western
Gray Squirrel (Sciurus griseus) on Ft. Lewis, Washington.
Northwest Science 69(3):204–216.
Sirén AP, Morelli TL. 2020. Interactive range‐limit theory (iRLT): an
extension for predicting range shifts. Journal of Animal Ecology
89(4):940–954. https://doi.org/10.1111/1365-2656.13150
Snyder KA, Evers L, Chambers JC, Dunham J, Bradford JB, Loik ME.
2019. Effects of changing climate on the hydrological cycle
in cold desert ecosystems of the Great Basin and Columbia
Plateau. Rangeland Ecology and Management 72(1):1–12.
https://doi.org/10.1016/j.rama.2018.07.007
Stewart JA, Wright DH, Heckman KA. 2017. Apparent climate-
mediated loss and fragmentation of core habitat of the American
Pika in the Northern Sierra Nevada, California, USA. PLoS One
12(8):e0181834. https://doi.org/10.1371/journal.pone.0181834
Stuart KD, Vander Haegen WM, Jenkins KJ, Keren IN, West SD. 2018.
Western Gray Squirrel resource selection related to re fuel
management. The Journal of Wildlife Management 82(7):1478–
1490. https://doi.org/10.1002/jwmg.21499
Sultaire SM, Kawai-Harada Y, Kimmel A, Greeson EM, Jackson PJ,
Contag CH, Lackey CW, Beckmann JP, Millspaugh JJ, Montgomery
RA. 2023b. Black Bear density and habitat use variation at the
Sierra Nevada–Great Basin Desert transition. The Journal of
Wildlife Management 87(3):e22358. https://doi.org/10.1002/
jwmg.22358
Sultaire SM, Millspaugh JJ, Jackson PJ, Montgomery RA. 2023a. The
inuence of ne‐scale topography on detection of a mam-
mal assemblage at camera traps in a mountainous land-
scape. Wildlife Biology 2023(2):e01026. https://doi.org/10.1098/
rspb.2015.3104
Sultaire SM, Pauli JN, Martin KJ, Meyer MW, Notaro M, Zuckerberg
B. 2016. Climate change surpasses land-use change in the
contracting range boundary of a winter-adapted mammal.
Proceedings of the Royal Society of London, B: Biological Sciences
283(1831):20153104. https://doi.org/10.1098/rspb.2015.3104
Tayleur CM, Devictor V, Gaüzère P, Jonzén N, Smith HG, Lindström A.
2016. Regional variation in climate change winners and losers
highlights the rapid loss of cold-dwelling species. Diversity and
Distributions 22(4):468–480. https://doi.org/10.1111/ddi.12412
Teangana DO, Reilly S, Montgomery WI, Rochford J. 2000. Distribution
and status of the Red Squirrel (Sciurus vulgaris) and Grey Squirrel
(Sciurus carolinensis) in Ireland. Mammal Review 30(1):45–56.
https://doi.org/10.1046/j.1365-2907.2000.00054.x
Thuiller W, Georges D, Engler R, Breiner F. 2016. biomod2: ensemble
platform for species distribution modeling. R package version
3.5.1. https://CRAN.R-project.org/package=biomod2.
Thuiller W, Lafourcade B, Engler R, Araújo MB. 2009. BIOMOD—a plat-
form for ensemble forecasting of species distributions. Ecography
32(3):369–373. https://doi.org/10.1111/j.1600-0587.2008.05742.x
Tingley MW, Beissinger SR. 2009. Detecting range shifts from histori-
cal species occurrences: new perspectives on old data. Trends in
Ecology and Evolution 24(11):625–633. https://doi.org/10.1016/j.
tree.2009.05.009
Tran TT, Carter BE, Castillo Vardaro JA. 2022. Predicted threats to
a native squirrel from two invading species based on citizen
science data. Biological Invasions 24(11):3539–3553. https://doi.
org/10.1007/s10530-022-02859-7
U.S. Census Bureau. 2020. RACE. Decennial census, DEC redistricting
data (PL 94-171), Table P1. [accessed 4 Jan 2024]. https://data.
census.gov/table/DECENNIALPL2010.P1?g=160XX00US3260600.
Vander Haegen WM, Orth GR, Aker LM. 2005. Ecology of the west-
ern gray squirrel in south-central Washington. Progress report.
Olympia (WA, USA): Washington Department of Fish and
Wildlife.
Weisberg PJ, Lingua E, Pillai RB. 2007. Spatial patterns of pinyon–
juniper woodland expansion in central Nevada. Rangeland
Ecology and Management 60(2):115–124. https://doi.
org/10.2111/05-224r2.1
Wilkening JL, Ray C, Beever EA, Brussard PF. 2011. Modeling contem-
porary range retraction in Great Basin pikas (Ochotona princeps)
using data on microclimate and microhabitat. Quaternary
International 235(1–2):77–88. https://doi.org/10.1016/j.
quaint.2010.05.004
Williams JW, Jackson ST. 2007. Novel climates, no‐analog commu-
nities, and ecological surprises. Frontiers in Ecology and the
Environment 5(9):475–482. https://doi.org/10.1890/070037
Zhang J, Jiang F, Li G, Qin W, Wu T, Xu F, Hou Y, Song P, Cai Z, Zhang
T. 2021. The four antelope species on the Qinghai–Tibet plateau
face habitat loss and redistribution to higher latitudes under
climate change. Ecological Indicators 123(2021):107337. https://
doi.org/10.1016/j.ecolind.2021.107337
Downloaded from https://academic.oup.com/jmammal/advance-article/doi/10.1093/jmammal/gyae011/7606630 by guest on 12 February 2024