ArticlePDF Available

Denning Ecology of Wolves in East-Central Alaska, 1993–2017


Abstract and Figures

Dens are a focal point in the life history and ecology of gray wolves (Canis lupus), and their location can influence access to key resources, productivity, survivorship, and vulnerability to hunting, trapping, and control efforts. We analyzed the selection of den sites and the phenology of their use inside the Yukon-Charley River National Preserve from 1993 to 2017 to enhance our understanding of this resource. At the landscape scale, we found that wolves in east-central Alaska selected den sites that were lower in elevation, snow free earlier in the spring, exposed to greater solar radiation, and closer to water. Den sites were also associated with areas that had burned less recently and had lower terrain ruggedness at the 1 km scale. These results supported our hypothesis that wolves would den relatively close to essential resources (water and prey) and in areas that are drier (melt earlier) in the spring. At the home range scale, wolves also selected den sites at lower elevations and showed a strong selection for the center of their home range. Furthermore, the average distance between active den sites was 37.3 km, which is slightly greater than the average radius (32.5 km) of a home range of a pack. Our results support our hypothesis that dynamic social factors modulate the selection of environmental factors for den site location. Wolves den away from other packs to reduce competition and exposure to intraspecific conflict. High-quality denning habitat does not currently appear to be a limiting factor for this population. Females, on average, entered their dens on 10 May, stayed inside the den for eight days, and remained less than 1 km from the den for an additional six days after emerging. We found that wolves denning at higher elevations entered their dens later than those at lower elevations, which also supported one of our hypotheses. Lastly, we documented limited evidence of earlier denning over time. Long-term monitoring projects, such as ours, are critical in identifying these types of trends.
Content may be subject to copyright.
VOL. 71, NO. 4 (DECE MBER 2018) P. 444 – 455
Denning Ecology of Wolves in East-Central Alaska, 1993 2017
Kyle Joly,1,2 Mathew S. Sorum1 and Matthew D. Cameron1
(Received 4 May 2018; accepted in revised form 19 July 2018)
ABSTRACT. Dens are a focal point in the life history and ecology of gray wolves (Canis lupus), and their location can
inuence access to key resources, productivity, survivorship, and vulnerability to hunting, trapping, and control efforts. We
analyzed the selection of den sites and the phenology of their use inside the Yukon-Charley River National Preserve from 1993
to 2017 to enhance our understanding of this resource. At the landscape scale, we found that wolves in east-central Alaska
selected den sites that were lower in elevation, snow free earlier in the spring, exposed to greater solar radiation, and closer
to water. Den sites were also associated with areas that had burned less recently and had lower terrain ruggedness at the 1 km
scale. These results supported our hypothesis that wolves would den relatively close to essential resources (water and prey) and
in areas that are drier (melt earlier) in the spring. At the home range scale, wolves also selected den sites at lower elevations
and showed a strong selection for the center of their home range. Furthermore, the average distance between active den sites
was 37.3 km, which is slightly greater than the average radius (32.5 km) of a home range of a pack. Our results support our
hypothesis that dynamic social factors modulate the selection of environmental factors for den site location. Wolves den away
from other packs to reduce competition and exposure to intraspecic conict. High-quality denning habitat does not currently
appear to be a limiting factor for this population. Females, on average, entered their dens on 10 May, stayed inside the den for
eight days, and remained less than 1 km from the den for an additional six days after emerging. We found that wolves denning
at higher elevations entered their dens later than those at lower elevations, which also supported one of our hypotheses. Lastly,
we documented limited evidence of earlier denning over time. Long-term monitoring projects, such as ours, are critical in
identifying these types of trends.
Key words: Canis lupus; den; habitat selection; natality; protected areas; pup rearing
RÉSUMÉ. Les tanières sont un point central du cycle biologique et de l’écologie du loup gris (Canis lupus). Leur emplacement
peut inuencer l’accès aux ressources principales, la productivité, la survie et la vulnérabilité à la chasse, au piégeage et aux
mesures de contrôle. An de mieux comprendre cette ressource, nous avons analysé la sélection des emplacements de tanières
et la phénologie de leur utilisation dans la réserve nationale Yukon-Charley Rivers pour les années allant de 1993 à 2017. À
l’échelle du paysage, nous avons trouvé que les loups du centre-est de l’Alaska choisissaient des emplacements de tanières
en moins grande altitude, plus près de l’eau, où la neige fondait plus vite au printemps et où le rayonnement solaire était plus
grand. Par ailleurs, les emplacements des tanières étaient caractérisés par des secteurs brûlés moins récemment et un relief
accidenté plus bas à l’échelle de 1 km. Ces résultats ont permis d’appuyer notre hypothèse selon laquelle les loups établiraient
leur tanière relativement près des ressources essentielles (eau et proies), dans des endroits plus secs (fonte hâtive) au printemps.
À l’échelle du domaine vital, les loups choisissaient aussi des emplacements de tanières en plus faible altitude, avec une forte
propension pour le centre de leur domaine. De plus, la distance moyenne entre les tanières actives était de 37,3 km, ce qui
est un peu plus grand que le rayon moyen (32,5 km) du domaine vital d’une meute. Nos résultats viennent appuyer notre
hypothèse voulant que les facteurs sociodynamiques modulent la sélection de facteurs environnementaux pour l’emplacement
des tanières. Les loups établissent leurs tanières à l’écart d’autres meutes an de réduire la compétition et les possibilités
de conits intraspéciques. En ce moment, la haute qualité de l’habitat pour l’établissement des tanières ne semble pas être
un facteur limitant pour cette population. En moyenne, les femelles s’installaient dans leur tanière le 10 mai, y restaient
pendant huit jours et demeuraient à moins d’un kilomètre de leur tanière pendant six autres jours après leur sortie. Nous avons
remarqué que les loups optant pour des tanières en plus haute altitude s’y installaient plus tard que ceux en plus faible altitude,
ce qui étayait aussi une de nos hypothèses. En dernier lieu, nous avons documenté les preuves restreintes d’établissement plus
hâtif dans les tanières au l des ans. Les projets de surveillance à long terme comme le nôtre jouent un rôle primordial dans la
détermination de ces types de tendances.
Mots clés : Canis lupus; tanière; sélection de l’habitat; natalité; zones protégées; élevage des petits
Traduit pour la revue Arctic par Nicole Giguère.
1 National Park Service, Yukon-Charley Rivers National Preserve and Central Alaska Inventory and Monitoring Network,
4175 Geist Road, Fairbanks, Alaska 99709, USA
2 Corresponding author: kyle_
© United States Government. Administered by the Arctic Institute of North America
Large carnivores often are apex predators that serve
important ecological functions in the environment. They
can affect large herbivore populations directly, through
predation (Gasaway et al., 1992; Sinclair et al., 2003; Ripple
and Beschta, 2012; Joly et al., 2017), and also indirectly, by
altering their behavior, movements, and habitat selection
(Lima, 1998; Laundré et al., 2001; Fortin et al., 2005;
Berger, 2007). These impacts, in turn, can cause cascading
effects across different trophic levels (i.e., Paine, 1980;
Carpenter et al., 1985; Beschta and Ripple, 2009; Prugh et
al., 2009). Therefore, dramatic changes to large carnivore
populations should be expected to cause far-ranging and
consequential impacts to the natural environment.
Large carnivores have experienced massive population
declines and range contractions globally (Ripple et al.,
2014). Vast, remote, sparsely populated, and relatively
intact ecosystems in Alaska have generally insulated
these carnivores from pressures such as habitat loss and
fragmentation, persecution by humans, depletion of
their prey base, and the excessive hunting and trapping
that are the ultimate causes of these losses. However,
even in portions of Alaska, predator control efforts have
substantively affected predator populations (i.e., Boertje et
al., 1996; Keech et al., 2011). In east-central Alaska, predator
control efforts outside the Yukon-Charley Rivers National
Preserve affected the wolf (Canis lupus) population inside
the preserve (Schmidt et al., 2017). The preserve was
designated, in part, to maintain the environmental integrity
of the region and to protect populations of wolves and other
wildlife species and their habitat (Alaska National Interest
Lands Conservation Act, 1980: Section 201 (10)). To
accomplish this, wildlife managers need to understand the
ecological requirements of the wolves relative to the overall
take of wolves to aid in their conservation.
Dens can be critical for survival and are a limiting
resource for some populations (McLoughlin et al., 2004;
Ross et al., 2010; Klaczek et al., 2015). Dens provide shelter
from inclement weather and protection from other predators.
The relatively stable microclimate dens provide is critical
for the survival of young (Laurenson, 1994; Fernández and
Palomares, 2000; Benson et al., 2008). The location of the
den site is important for several reasons. First, food resources
for wolves during summer can be a limiting factor (Metz et
al., 2012) and affect pup survival (Fuller, 1989; Benson et al.,
2013; Klaczek et al., 2015). Most pup mortality occurs within
the rst six months after birth (van Ballenberghe and Mech,
1975; Benson et al., 2013). Since movements away from the
den site are limited by the pups motility for the rst six
weeks after birth (Fritts and Mech, 1981; Mills et al., 2008;
Lake et al., 2013), locating the den close to an abundant food
base is crucial (Ciucci and Mech, 1992; Klaczek et al., 2015).
Second, dens are typically situated near fresh water (Ballard
and Dau, 1983; Person and Russell, 2009; Benson et al.,
2015; Jacobs and Ausband, 2018) so that the breeding female
can drink while attending the pups (Mech, 1970). Third, the
location of dens inuences the vulnerability to predation
of pups and adults alike (Benson et al., 2015; Jacobs and
Ausband, 2018). For adult wolves, inter-pack strife accounts
for a substantial number of mortalities (Murie, 1944; Mech et
al., 1998; Smith et al., 2015; Schmidt et al., 2017). To mitigate
these risks, wolves are thought to place their dens near the
center of their territory (Fritts and Mech, 1981; Ciucci and
Mech, 1992). While dens are not used year-round, in some
respect they act as a center of activity for the pack’s annual
home range. Thus, den and territory location could affect the
entire pack through activity, mortality, and recruitment (e.g.,
Borg et al., 2016).
The timing of parturition is physiologically linked to the
timing of mating. The timing of these events is likely an
adaptation to long-term climatic patterns and phenological
cycles that allow for the optimal conditions to support
young (Sandell, 1990; Bowyer et al., 1998; Walsh et al.,
2016). Over the last few decades, winter snow in Alaska has
been melting earlier in the spring, and vegetative green-up
is also occurring earlier (Monahan et al., 2016; Cox et al.,
2017). How such dramatic changes in the timing of seasons
and related phenological cycles influence the denning
ecology and demography of wolves is unknown.
The goal of our study was to elucidate the denning
ecology of wolves in east-central Alaska. Our primary
objectives were to identify landscape characteristics and
societal factors associated with den site selection and
to document phenological patterns of den use. Our rst
hypothesis was that wolves would select dens sites that
have physical and environmental characteristics suitable
for digging the den, thermoregulation, and rearing young.
These characteristics would ensure that the den could be
dug and would remain dry. Often these criteria mean that
dens are located on knolls, eskers, and hillsides that are well
drained, have no permafrost, and are composed of ne-
grained sediments (Ballard and Dau, 1983; Klaczek et al.,
2015). These sediment types are associated with riparian
zones at lower elevations. Our second hypothesis was that
wolves would locate den sites near key resources, such as
accessible fresh water (Mech, 1970; Ballard and Dau, 1983;
Person and Russell, 2009; Benson et al., 2015) and prey
(Ciucci and Mech, 1992; Klaczek et al., 2015). Our third
hypothesis was that wolves centralize their den sites within
their home range to avoid other packs. Wolves are territorial
animals, so centralizing their den sites within their home
ranges could reduce competition and inter-pack conict
(Fritts and Mech, 1981; Ciucci and Mech, 1992; Mladenoff
et al., 1999). Our fourth hypothesis was that wolves would
enter dens later at higher elevations, where snowmelt would
occur later. Our nal hypothesis was that over time, as the
climate warmed, wolves would enter dens earlier.
The 23 166 km2 study area encompassed the entire
10 209 k m2 of the Yukon-Charley Rivers National
446 • K. JOLY et al.
Preserve and extended outwards a distance of 20 km from
its perimeter (Fig. 1). We clipped the 20 km buffer at the
international border with Canada and south of the preserve
to match the extent of the habitat map (NPS, 1997) for the
region. The region is quintessential boreal forest. Black
spruce (Picea mariana) is the most common tree species,
inhabiting areas with permafrost and poorly drained soils.
Aspen (Populus tremuloides) and birch (Betula papyrifera)
trees are common on south-facing slopes, whereas white
spruce (Picea glauca) and poplars (Populus balsamifera)
can be found in riparian corridors. Willow (Salix spp.),
dwarf birch (Betula glandulosa), and alder (Alnus spp.)
shrubs are often found lining the riparian corridors but
also climbing the lower-elevation slopes. There are also
extensive areas of wetland, tussock (e.g., Eriophorum spp.)
tundra, and alpine tundra communities. Mountain peaks
are generally lower than 2000 m. The Yukon and Charley
Rivers are the two main waterways (Fig. 1).
The full complement of native fauna exists within the
study area, including low-density populations of moose
(Alces alces; Sorum and Joly, 2016) and Dall’s sheep
(Ovis dalli; Joly, 2015). Caribou (Rangifer tarandus) from
the Fortymile caribou herd spend much of their time,
including the calving period, in the study area (Boertje et
al., 2017). During our 25-year study period (1993 2017),
the herd ranged in size from 22 000 to 71 400 individuals
(Boertje et al., 2017; Friedman, 2017). In addition to wolves,
other predators include grizzly (Ursus arctos) and black
(U. americanus) bears, wolverines (Gulo gulo), and red fox
(Vulpes vulpes). King salmon (Oncorhynchus tshawytscha),
northern pike (Esox lucius), and Arctic grayling (Thymallus
arcticus) are common sh species.
The region has a typical continental climate with long
(7 8 months), cold winters and short (2 – 3 months) but
warm, relatively dry summers. Snow typically begins to
accumulate in October, reaching maximum depths of about
50 cm in March (Sousanes and Hill, 2014). Temperatures
can drop to −51˚C. Average annual temperature is about
−4˚C, with summer temperatures reaching a maximum of
33˚C (Sousanes and Hill, 2014). During our 25-year study
FIG. 1. Wolf den study area (outlined in white) in east-central Alaska, 1993–2017. The Yukon-Charley Rivers National Preserve is shown in green. Black squares
indicate the villages of Circle (upper left) and Eagle (lower right). Black lines show roads.
period, annual precipitation was approximately 31.5 cm
(Sousanes and Hill, 2014). Warm, dry summers led to more
than 40% of the preserve being burned by wildre since the
mid-1980s (Schmidt et al., 2017).
Identication and Characterization of Den Sites
Wolves were found via aerial tracking and caught using
darting techniques outlined by Schmidt et al. (2017). Most
dens were located on radio-tracking ights, and their
locations were recorded using GPS units on the aircraft.
From 1993 to 2000, collars were equipped with VHF
transmitters only. After 2000, both VHF and GPS collars
were deployed. We calculated the Euclidean distance
between active den sites for each year using ArcGIS. We
assigned values for the following attributes to each den site:
distance from waterway, elevation, aspect, slope, terrain
ruggedness (Sappington et al., 2007) at the 180 m and
1 km scales, average day of year in spring that it becomes
snow-free (Macander and Swingley, 2017), probability of
permafrost (Pastick et al., 2015), time since last re (Alaska
Interagency Coordination Center, https://,
solar radiation index (Keating et al., 2007), habitat type
(NPS, 1997), and a forested:unforested ratio.
We lumped habitat types (30 altogether) into six
categories (closed forest, open forest, tall shrub, low shrub,
graminoids, and miscellaneous). To generate an index of
cover around each den site, we calculated the ratio of forest
to unforested habitat types by dividing the area of forested
habitat within a 1 km radius of the den by the total area
within the same radius.
Selection of Den Sites at the Landscape Scale
We investigated physiographic factors associated with
den site selection at the landscape scale using resource
selection functions (RSFs; Manly et al., 2002) to compare
den locations used by wolves to other available sites across
the study area. Over our 25-year study period, the wolves
of east-central Alaska, as a population, have had the
opportunity to den anywhere within a 20 km buffer zone
around the preserve. Since our goal was to understand
the static environmental attributes of den sites across our
study area, regardless of history of use, den locations were
used only once in this analysis. To dene availability, we
attributed 1000 random locations (available sites) with
physiographic data that did not vary annually in the same
manner as the den sites. When we clipped the buffer to t
the habitat map, 98 random locations were removed. We
also removed two random locations that had implausible
snow-free dates, leaving 900 available points within the
study area.
We performed logistic regression using generalized
linear models in R Version 3.4.3 (R Core Team, 2017),
with den use as the response. We logit transformed the
two covariates which were proportions, probability of
permafrost and open/closed ratio (Warton and Hui, 2011),
and standardized (subtracted the mean and divided by the
standard deviation) the two measures of terrain ruggedness.
We used ‘miscellaneous’ as our reference category for
habitat comparisons. We tested for multi-collinearity of
predictor variables using variance ination factors with a
cutoff value of 3 (Zuur et al., 2010), as well as a cutoff of
higher than 0.5 for correlation values. Model selection was
performed using Akaike’s Information Criterion corrected
for small sample sizes (AICc, Burnham and Anderson,
2002), starting with a global model of all covariates and
testing biologically plausible subsets. In total, we tested
46 models (see online Appendix 1: Table S1). To assess
the relative selection for each covariate, we reran the
top model with standardized continuous covariates and
interpreted coefcient values. We used the top-performing
model to generate a predictive map for denning habitat. We
evaluated the performance of our top model using leave-
one-out cross-validation (Boyce et al., 2002), and measured
the predictive capacity of our model with the area under the
receiver operating curve (ROC) using the package “pROC”
(Robin et al., 2011). ROC values from 0.7 to 0.8 indicate
acceptable levels of discrimination for a model and above
0.8 indicates excellent discrimination (Hosmer et al., 2013).
Selection of Den Sites at the Home Range Scale
On an annual basis, the home range of one wolf pack
is generally unavailable to another pack (Mladenoff et al.,
1999). Therefore, we also investigated den site selection
at the third order (use of a habitat component within a
home range) of selection (Johnson, 1980) using RSFs to
compare actual den locations to available sites within
a breeding female’s home range. We developed annual
home ranges using only GPS data from breeding females
to provide consistency. We delineated the annual home
range, as determined by a 95% minimum convex polygon
(MCP) to reduce the inuence of extra-territorial forays,
using the GPS data from the breeding female during the
biological year prior to denning as available denning
habitat. For example, we used GPS locations from 30
April 2003 to 1 May 2004 to create a home range from
which random (available) points were compared to the
2004 den site of wolf No. 192. We used 1 GPS location per
day and required a minimum of 300 locations in that year
in order to develop an MCP. We generated and attributed
1000 random locations within the home range with the
same environmental covariates as described for the
landscape-scale analyses above. As an index of exposure
to neighboring packs, we determined the distance from
the den and random points to the edge of the annual home
range. We then compared the random locations within the
home range to the actual den site in a matched case-control
framework. We limited our models to only two parameters
because of the limited number of events (see Results) in
448 • K. JOLY et al.
this more restricted analysis (Hosmer et al., 2013), and thus
did not include the six-level habitat variable. We performed
conditional logistic regression using the ‘clogit’ function
in the ‘survival’ package (Therneau, 2015) in R with the
matched case-control sets as strata and pack identier as
the cluster to account for correlations between multiple
years of denning for some packs.
Denning Phenology
We used the visual inspection based estimates of
denning onset outlined by Walsh et al. (2016) on our GPS
data (2001 17). The technique relies upon reduced daily
movement rates and increased GPS location x failures
to determine when wolves begin to den. Using the same
technique, we determined when wolves emerged from
their dens (i.e., rst GPS relocation after the period of
failure to get a GPS x) and how long they stayed there
before moving more than 1 km from the site. This analysis
led to the discovery of additional potential den sites (see
Results). We used multiple linear regression to assess a
set of candidate models to determine what variables were
correlated with timing (day of year) of den entrance. We
used size of the pack in spring, elevation, snow-free date,
and year of the event as variables in the model, and model
selection was based on AICc. Latitude was strongly and
negatively correlated with elevation, so it was not included
as a variable. For duration spent in and at the den, we also
included date of den entrance and estimated number of
pups produced as variables.
Identication and Characterization of Den Sites
A total of 52 individual den site locations, from 26
different packs, were initially identied through direct
aerial observations and cursory observations of the
GPS collar data from 1993 to 2017. The average distance
between active den sites was 37.3 km (SD = 15.1; range
13.8 – 95.9). We identied active den sites in all years
of the study, except 2014 – 16 when limited numbers of
marked individuals reduced our ability to detect den sites.
Consecutive use of individual den sites ranged from one
to eight years, which led to 116 den-year records. On 10
occasions, a single pack used two different den sites in the
same year: the 70 Mile Pack in 2011 and 2012, the Copper
Mountain Pack in 2010, the Cottonwood Pack in 2003 and
2005, the Edwards Pack in 1996, the Flat Pack in 1995, and
the Webber Creek Pack in 1999, 2000, and 2001.
Selection of Den Sites at the Landscape Scale
Means, standard deviations, and ranges of covariates
used in the modeling process are displayed in Table 1. We
excluded the covariates of slope, aspect, and forested/
un-forested ratios since they had a variance inflation
factor greater than 3 and were correlated with the solar
radiation index and elevation. Correlation among all other
variables was less than 50%. The top two models for den
site selection consisted of distance to water, elevation,
time since last re, average snow-free date, solar radiation
index, terrain ruggedness at the 1 km scale, and probability
of permafrost (online Appendix 1: Table S1). The second
model, which differed from the top model only by the
addition of probability of permafrost, was within two
AICc of the top model. Parameter estimates for these two
models were nearly identical for shared covariates, and the
condence interval for probability of permafrost overlapped
zero. Following the recommendations of Burnham and
Anderson (2002) and Arnold (2010), we considered the
addition of the permafrost covariate as uninformative. Our
top model consisted of distance to water (β = −0.0002,
SE = 0.0001), elevation (β =0.0004, SE = 0.0002),
scaled terrain ruggedness at 1 km resolution = −0.3481,
SE = 0.2011), snow-free date (β = −0.0675, SE = 0.0256, day),
time since last re = 0.0169, SE = 0.0051, year), and solar
radiation index (β = 1.3665, SE = 0.6057). Of these, time since
last re exhibited the greatest relative selection, with wolves
selecting older stands as den sites (Fig. 2). The negative
relationship between snow-free date and denning was the
second greatest relative selection, with wolves selecting
sites that became snow free earlier in the spring. Overall,
den sites were characterized by settings that were closer to
water, lower in elevation, older in stand age, melted earlier
in the spring, received more solar radiation, and exhibited
less rugged terrain (Fig. 2). The 95% CIs for elevation, solar
radiation, and terrain ruggedness overlapped zero. We found
no effect of habitat type on wolf denning at the resolution we
considered. Figure 3 depicts relative probability of denning
across the study area. Our top model had an ROC score of
0.77, indicating acceptable discrimination.
Selection of Den Sites at the Home Range Scale
We developed 25 home ranges for breeding females from
eight different packs, spanning 2004 15, for which we had
a year of location data prior to denning and a corresponding
actual den location. Annual home ranges averaged
2830 km2 (SE: ± 514 km2; range 743 11765 km2). Distance
to home range edge (β = 0.0003, robust SE = 0.00006) and
elevation (β = −0.0066, robust SE = 0.0015) comprised
the top performing model (online Appendix 1: Table S2).
Wolves exhibited strong selection for areas near the center
of their home range and lower elevations within their home
range (Fig. 4).
Denning Phenology
From our GPS data of breeding females, we identied
a total of 55 denning events, which occurred in all years
from 2001 to 2017 except 2002 and 2016. Most (> 75%) of
the events matched known den sites; however, 13 did not
and likely represent previously unknown den sites. In 2008
for the Step Mountain Pack and in 2012 and 2013 for the
Snowy Peak Pack, more than one female per pack appeared
to den (all other events were from 1 female/pack/year).
Wolves entered their dens to give birth from 29 April to 30
May (day of year 119 150), with a mean date of 10 May
(day of year 130; SD = 6.5; n = 55 events). Wolves stayed an
average of 8.3 days (SD = 3.7; range 2 20) inside the den.
After emerging from the den, wolves remained within 1 km
of the den site for an additional 5.8 days (SD = 7.7; range
0 40). All 55 denning events showed the wolves returning
to a single location: either the natal den site, a secondary
den site (i.e., a den site to which the pups were moved after
birth), or a rendezvous location.
The top model explaining the timing of entrance
included elevation (β = 0.0039, SE = 0.0012) and year
= −0.3049, SE = 0.2272). The only other model within
2 AICc AICc = 0.11) retained only elevation
(β = −0.0043, SE = 0.0011). For two denning events, we did
not have spring pack size and since that variable was not
in the top models, we re-ran our analyses without it. This
only slightly modied our results as the top model was
elevation (β = 0.0045, SE = 0.0012) alone and the only other
model within 2 AICcAICc = 0.46) was composed of
elevation (β = 0.0040, SE = 0.0012) and year (β = −0.3695,
SE = 0.2398). Elevation was signicantly associated with
denning onset (R2 = 0.21, F = 13.96, df = 54, p < 0.01), with
den entrance occurring later at higher elevation sites but
earlier over time. Den entrances on 15 May or later have
not occurred since 2011 (Fig. 5), when there were two (on
17 and 19 May). Of the den entrances occurring on 15
May or later, 85% (11 of 13) occurred prior to 2009 (i.e., in
2001 08). This nding was not affected by any sampling
bias as 49% of the den onsets were detected from 2001
to 2008 and 51% from 2009 to 2017. We did not detect a
signicant correlation between duration of presence in or at
the den with any of the variables we examined.
Den sites represent a critical component of wolf
ecology, and understanding the process of selection for
these features is important so that wildlife managers can
make informed decisions regarding wolf management and
conservation. Here, we investigated both physiographic
and social factors associated with den site selection by
wolves in interior Alaska. Our results suggest that wolves
select lower-elevation river corridors that melt out earlier
in the season, but also areas well away from the edges of
their annual home ranges to reduce the risk of conspecic
competition and conict. This information is novel and
informative for this region because den site selection can
inuence survival of adults and young (Laurenson, 1994;
Fernández and Palomares, 2000; Benson et al., 2008, 2015;
TABLE 1. Parameters, with means, SD, and range, used to model selection of wolf denning habitat in east-central Alaska, 1993–2017.
Habitat type was also included as a categorical variable.
Parameter Denition Mean SD Range
A) Den sites:
DistH2O Distance to water (m) 1431 1378 39–5240
Elevation Height above sea level (m) 575 282 207–1151
ElevFromMean Absolute value height difference from landscape mean (m) 281 161 4–530
Terrain Terrain Ruggedness Index at 1 km scale 0.07 0.08 0.00–0.30
Snow Julian date when area became snow free 118 7 104–136
Fire Number of years since the area last burned 83 31 9–100
Solar Solar Radiation Index 0.53 0.27 -0.36–0.97
Permafrost Probability of the area being permafrost 0.70 0.28 0.00–0.97
B) Random locations
DistH2O Distance to water (m) 2259 1672 2–7991
Elevation Height above sea level (m) 737 331 172–1708
ElevFromMean Absolute value height difference from landscape mean (m) 278 179 0–970
Terrain Terrain Ruggedness Index at 1 km scale 0.11 0.09 0.00–0.43
Snow Julian date when area became snow free 124 10 12–160
Fire Number of years since the area last burned 72 36 1–100
Solar Solar Radiation Index 0.34 0.36 -0.59– 0.99
Permafrost Probability of the area being permafrost 0.74 0.21 0.00–1.00
FIG. 2. Relative inuence of six continuous covariates on modeling of wolf
denning habitat in east-central Alaska, 1993–2017. Dots indicate the means
and bars, the 95% condence intervals. Covariates were standardized by
subtracting the mean and dividing by the standard deviation.
450 • K. JOLY et al.
Ross et al., 2010; Jacobs and Ausband, 2018). We expect our
results will broadly inform management and conservation
by documenting where and when wolves den, which could
allow for data-driven decisions on appropriate hunting and
trapping seasons and area closures.
Snow melts earlier at lower elevations, and we found
that sites that became snow-free earlier were selected for
denning at the landscape scale. These sites also had greater
levels of solar radiation. Ballard and Dau (1983) found that
dens were preferentially located on south- and east-facing
slopes (79% of the den sites we recorded were similarly
facing), which is also related to higher levels of solar
radiation. All of these conditions should promote warmth
and dryness during the denning season. We suspect that
presence of permafrost and type of soil (e.g., sand) are
important factors in den site selection, but we did not have
data at the necessary resolution to capture this relationship.
Den sites were also located away from areas that had
recently been burned by wildres. Depending on edaphic
conditions, old stands tend to be associated with forest
habitat. Mature forests tend to be on well-drained soils in
this region and could provide shade to aid thermoregulation
of pups as summer arrives. However, as in other studies
(e.g., Theuerkauf et al., 2003), differential selection by
habitat type was not supported. Dens were also located in
areas of lower terrain ruggedness at the 1 km scale. Lower
ruggedness near the den site might increase sight lines for
wolves (depending on vegetation), allowing for more time
to retreat from or engage with other predators that might
threaten their young. Thus, we feel our results support
our hypothesis that den sites are selected for physical and
environmental characteristics that are suitable for digging,
thermoregulation, and rearing young.
At the landscape scale, we found that wolves selected
den sites close to water and at lower elevations, which
supports our hypothesis about resources and agrees with
other studies. Having easy access to a reliable source of
fresh water is critical for a breeding female attending
her newborn pups (Mech, 1970; this study). Sites that are
relatively lower in elevation tend to have greater access to
FIG. 3. Resource suitability map for wolf denning habitat in east-central Alaska, 1993–2017. Dark blue shades represent the lowest suitability (relative probability
of selection), lighter shades represent greater suitability, and red shades, the greatest suitability.
prey species such as moose (Sorum and Joly, 2016), beaver
(Castor canadensis), snowshoe hares (Lepus americanus),
sh, and waterfowl. Caribou are also commonly found
at lower elevations during winter (Boertje et al., 2017).
However, higher-elevation sites would have greater access
to Fortymile Herd caribou calves that are born just after
wolves emerge from their dens (Boertje et al., 2017). Thus,
the connections between critical resources and selection
of den sites still warrant additional ne-scale study; we
recommend assessing wolf prey distribution, abundance,
and availability as the next step.
High-quality denning habitat (Fig. 3) was relatively
abundant across the landscape, and we do not believe it is a
limiting factor for this population. Our landscape RSF map
(Fig. 3) depicting relative probability of use for denning
habitat is, we believe, the rst of its kind in the region.
Lower-elevation areas with greater solar radiation that melt
out earlier in spring, and which were near waterways but
away from recently burned areas, had the greatest relative
probability of use. Two areas of high relative probability of
use that did not have documented den sites stand out: the
rst is along the Yukon River downstream (northwest) of
Eagle, and the second is upstream (southeast) of Circle. The
rst may lack documented den sites because it is relatively
far away from our base of operations in the northwest
portion of the preserve and may thus be affected by reduced
sampling effort. Alternatively, the lack of den sites may
be related to human use of the area. The Yukon River
freezes solid in winter and people use it as a travel corridor.
Increased hunting and trapping pressure and disturbance
associated with proximity to the villages have the potential
to influence den site selection. The lack of den sites
upstream from Circle supports this latter line of reasoning,
as the area is close to our base of operations. Additionally,
other factors, both ecological (e.g., prey abundance, amount
of terrain conducive to aerial capture operations) and
behavioral (e.g., social dynamics within or between packs),
that we were not able to address at this scale may also be
inuencing where wolves select den sites and our ability to
detect them. Our landscape analysis was limited to static
physiographic aspects of den site selection, but a suite of
biological and climatic factors that vary annually (e.g., wolf
density, prey availability, and snow conditions) most likely
also inuence den site selection each year.
Our home range – based analysis suggests that the
dynamic social structure of the wolf population in a given
year modulates the selection of the physical landscape
attributes for a denning location. As in the landscape scale
analysis, we found that wolves selected for lower elevations
relative to what was available within their home range. As
noted above, use of lower elevations for denning is likely
related to earlier snowmelt and improved access to key
resources. Interestingly, we also found that wolves selected
den sites near the center of their home ranges, as has been
found in other studies (e.g., Trapp et al., 2008). This nding
suggests that wolves attempt to reduce competition and
conict with other packs while optimizing access to prey.
Centralizing den sites within home ranges and away from
other packs reduces competition and inter-pack strife,
which is a large contributor to wolf mortality (Murie,
1944; Mech et al., 1998; Smith et al., 2015; Schmidt et
al., 2017). Having a den near the center of a pack’s home
range may thus benet tness (Fritts and Mech, 1981;
Ciucci and Mech, 1992). We posit that a centralized den
site may optimize access to prey in multiple directions,
and thereby may improve hunting efciency and reduce
the vulnerability of wolves traveling alone during the
summer, when pack cohesion is lower. Our ndings agree
with studies of other canids; for example, Moorcraft et al.
(2006) found that coyote (Canis latrans) territories were
inuenced by prey availability as well as by avoidance of
neighboring packs.
FIG. 4. Relative inuence of standardized covariates on conditional logistic
regression of den site selection based on annually varying home ranges of
wolf packs in east-central Alaska, 2004–15. Dots indicate the means and bars,
the 95% condence intervals.
FIG. 5. Timing (day of year) when wolves entered dens in east- central Alaska,
2001–17. The vertical line indicates Day 135 (May 15) as a point of reference.
452 • K. JOLY et al.
Active den sites were located approximately 37.3 km,
on average, from the nearest active den site. The average
home range size of packs in the region is 3322 km2 (Burch,
2013). A circle with this area has a radius of 32.5 km.
Therefore, we believe that these gures add further support
to our hypothesis that wolves situate their dens centrally
within their home range and away from other packs. For
those populations for which den sites are well monitored,
but radio collaring is limited, the use of distance to nearest
active den has the potential to be an index of home range
size, though more study of this relationship should be
conducted. In the future, den site selection should be
evaluated using annually varying factors, including
distance to and overlap with the territories of other packs,
prey abundance, level of human activity, and climatic
variables, and alternative means to delineate home ranges
(see Potts and Lewis, 2014; Kittle et al., 2015).
The onset of denning ranged from 29 April to 30
May, with a mean date of 10 May, which is remarkably
similar to the dates reported elsewhere in Alaska (see
Walsh et al., 2016). Denning in Alaska appears to occur a
couple of weeks later than in Minnesota (second week of
April; Fuller, 1989), but a couple of weeks earlier than in
the Canadian Arctic (late May to early June; Heard and
Williams, 1992), which suggests a strong nexus with
latitude. Females stayed an average of eight days in the
den and remained close (< 1 km) to it for an additional six
days. This is about 10 days less than Fuller (1989) reported
for wolves in Minnesota. We suspect that much of this
difference could be accounted for by differences in method,
including the increased level of precision afforded by GPS
technology that was not available in previous studies.
Additional studies investigating whether the duration of
females’ stay at the den is related to available prey biomass
are in order.
Onset of denning occurred later at higher-elevation
sites, which may be related to delayed snowmelt or less
available biomass of prey. ‘Year’ was also in the top models
for timing of denning, with denning occurring earlier over
time. The 95% CIs overlapped zero, so earlier onset of
denning over time was not a strong relationship. However,
since 2011, the onset of denning has always occurred prior
to 15 May. We monitored onset of denning for 17 years
(2001 17) and found that 85% of the onset events that
occurred on 15 May or later were during the rst seven
years of the study (i.e., in 2008 or before). Rapidly warming
temperatures in the region have led to earlier snowmelt
and vegetative green-up (Monahan et al., 2016; Cox et al.,
2017). Here, we document evidence that these earlier events
may in turn be affecting the timing of denning of wolves
in east-central Alaska. Given the xed gestation period of
wolves, these factors may be indices of conditions wolves
face during breeding (February and March) or conditions
from the previous summer that in turn inuence the timing
of breeding and conception. We posit that the relationship
between onset of denning and elevation suggests that
wolves have the requisite plasticity to adapt to conditions
at very ne temporal and spatial scales (i.e., within their
home range), which may increase their resiliency to
climate change. Further study of how changes in denning
phenology affect the demography of wolves is warranted.
Funding for this project came from the National Park Service,
Yukon-Charley Rivers National Preserve, and the Central Alaska
Inventory and Monitoring Program. J. Burch led the wolf-
monitoring program for the majority of its existence; without
his efforts, this work would not have been possible. We thank
all the pilots for decades of safe ying in difcult conditions
and the scores of biologists that helped with project eldwork
over the years. A project of this duration would not be feasible
without managers supportive of science and conservation; thus
we thank G. Dudgeon, J. Rasic, M. MacCluskie, T. Liebscher,
D. Mills, P. Rost, P. Knuckles, and others for keeping this project
going. We thank J. Burch, M. MacCluskie, J. Rasic, J. Schmidt,
and anonymous reviewers for providing comments on previous
versions of this manuscript that greatly improved it. We thank
L. Sanford and J. Eisaguirre for generously sharing their time and
advice on statistical methods.
The following tables are available in a supplementary
le to the online version of this article at:
TABLE S1. Model results from all 46 generalized linear
models for wolf den site locations, east-central Alaska,
1993 – 2017.
TABLE S2. Model results from 16 conditional logistic
regression models for wolf den site locations, east-central
Alaska, 200 4 – 15.
Alaska National Interest Lands Conservation Act. 1980. Public
Law 96 487, 94 Stat. 2371.
Arnold, T.W. 2010. Uninformative parameters and model selection
using Akaike’s information criterion. Journal of Wildlife
Management 74(6):1175 – 1178.
Ballard, W.B., and Dau, J.R. 1983. Characteristics of gray wolf,
Canis lupus, den and rendezvous sites in southcentral Alaska.
Canadia n Field-Naturalist 97:299 – 302.
Benson, J.F., Lotz, M.A., and Jansen, D. 2008. Natal den
selection by Florida panthers. Journal of Wildlife Management
72(2):405 – 410.
Benson, J.F., Mills, K.J., Loveless, K.M., and Patterson, B.R. 2013.
Genetic and environmental inuences on pup mortality risk
for wolves and coyotes within a Canis hybrid zone. Biological
Conser vation 16 6:133 – 141.
Benson, J.F., Mills, K.J., and Patterson, B.R. 2015. Resource
selection by wolves at dens and rendezvous sites in Algonquin
Park, Canada. Biological Conservation 182:223 232.
Berger, J. 2007. Fear, human shields and the redistribution of prey
and predators in protected areas. Biology Letters 3(6):620 – 623.
Beschta, R.L., and Ripple, W.J. 2009. Large predators and trophic
cascades in terrestrial ecosystems of the western United States.
Biological Conservat ion 142(11):2401 – 2414.
Boertje, R.D., Valkenburg, P., and McNay, M.E. 1996. Increases in
moose, caribou, and wolves following wolf control in Alaska.
Journal of Wildlife Management 60(3):474 489.
Boertje, R.D., Gardner, C.L., Ellis, M.M., Bentzen, T.W., and
Gross, J.A. 2017. Demography of an increasing caribou herd
with restricted wolf control. Journal of Wildlife Management
81(3):429 – 448.
Borg, B.L., Arthur, S.M., Bromen, N.A., Cassidy, K.A., McIntyre,
R., Smith D.W., and Prugh, L.R. 2016. Implications of harvest
on the boundaries of protected areas for large carnivore
viewing opportunities. PLoS One 11(4): e0153808.
Bowyer, R.T., van Ballenberghe, V., and Kie, J.G. 1998. Timing
and synchrony of parturition in Alaskan moose: Long-term
versus proximal effects of climate. Journal of Mammalogy
79(4):1332 – 1344 .
Boyce, M.S., Ver nier, P.R., Nielsen, S.E., and Schmiegelow, F.K.A.
2002. Evaluating resource selection functions. Ecological
Modelling 157(2-3):281 – 300.
ht t ps://
Burch, J. 2013. Annual report on vital signs monitoring of wolf
(Canis lupus) distribution and abundance in Yukon-Charley
Rivers National Preserve, Central Alaska Network: 2012
report. Natural Resource Technical Report NPS/CAKN/
NRTR-2012/736. Fort Collins, Colorado: National Park
Burnham, K.P., and Anderson, D.R. 2002. Model selection and
multimodel Inference: A practical information-theoretic
approach, 2nd ed. New York: Springer-Verlag.
Carpenter, S.R., Kitchell, J.F., and Hodgson, J.R. 1985. Cascading
trophic interactions and lake productivity: Fish predation
and herbivory can regulate lake ecosystems. BioScience
35(10):634 – 639.
Ciucci, P., and Mech, L.D. 1992. Selection of wolf dens in relation
to winter territories in northeastern Minnesota. Journal of
Mammalog y 73(4):899 – 905.
Cox, C.J., Stone, R.S., Douglas, D.C., Stanitski, D.M., Divoky,
G.J., Dutton, G.S., Sweeney, C., George, J.C., and Longenecker,
D.U. 2017. Drivers and environmental responses to the
changing annual snow cycle of northern Alaska. Bulletin of
the American Meteorological Society 98(12):2559 2577.
Fernández, N., and Palomares, F. 2000. The selection of breeding
dens by the endangered Iberian lynx (Lynx pardinus):
Implications for its conservation. Biological Conservation
94(1):51 – 61.
ht t ps://
Fortin, D., Beyer, H.L., Boyce, M.S., Smith, D.W., Duchesne,
T., and Mao, J.S. 2005. Wolves inuence elk movements:
Behavior shapes a trophic cascade in Yellowstone National
Park. Ecology 86(5):1320 – 1330.
Friedman, S. 2017. Census puts Fortymile caribou herd population
at more than 71,400, up from 51,000. Fairbanks Daily News-
Miner, November 10.
Fritts, S.H., and Mech, L.D. 1981. Dynamics, movements, and
feeding ecology of a newly protected wolf population in
northwestern Minnesota. Wildlife Monographs 80. 79 p.
Fuller, T.K. 1989. Population dynamics of wolves in north-central
Minnesota. Wildlife Monographs 105. 41 p.
Gasaway, W.C., Boert je, R.D., G rangaard, D.V., Kel leyhouse, D.G.,
Stephenson, R.O., and Larsen, D.G. 1992. The role of predation
in limiting moose at low densities in Alaska and Yukon and
implications for conservation. Wildlife Monographs 120. 59 p.
Heard, D.C., and Williams, T.M. 1992. Distribution of wolf dens
on migratory caribou ranges in the Northwest Territories,
Canada. Canadian Journal of Zoology 70(8):1504 1510.
Hosmer, D.W., Jr., Lemeshow, S., and Sturdivant, R.X. 2013.
Applied logistic regression, 3rd ed. Hoboken, New Jersey:
John Wiley and Sons, Inc.
Jacobs, C.E., and Ausband, D.E. 2018. Pup-rearing habitat use
in a harvested carnivore. Journal of Wildlife Management
82(4):802 – 809.
ht t ps://
Johnson, D.H. 1980. The comparison of usage and availability
measurements for evaluating resource preference. Ecology
61(1):65 – 71.
Joly, K. 2015. Aerial survey of Dall’s sheep: Yukon-Charley Rivers
National Preserve, July 2015. Natural Resource Report NPS/
YUCH/NRR—2015/1020. Fort Collins, Colorado: National
Park Service. 13 p.
Joly, K., Craig, T., Cameron, M.D., Gall, A.E., and Sorum, M.S.
2017. Lying in wait: Limiting factors on a low-density ungulate
population and the latent traits that can facilitate escape from
them. Acta Oecologica 85:174 183.
ht t ps:// 4
Keating, K.A., Gogan, P.J.P., Vore, J.M., and Irby, L.R. 2007. A
simple solar radiation index for wildlife habitat studies. Journal
of Wildlife Management 71(4):1344 – 1348.
454 • K. JOLY et al.
Keech, M.A., Lindberg, M.S., Boertje, R.D., Valkenburg, P.,
Taras, B.D., Boudreau, T.A., and Beckmen, K.B. 2011. Effects
of predator treatments, individual traits, and environment on
moose survival in Alaska. Journal of Wildlife Management
75(6):1361 – 1380.
Kittle, A.M., Anderson, M., Avgar, T., Baker, J.A., Brown, G.S.,
Hagens, J., Iwachewski, E., et al. 2015. Wolves adapt territory
size, not pack size to local habitat quality. Journal of Animal
Ecology 84(5):1177 – 1186.
htt p s:// 65-2 656.12366
Klaczek, M.R., Johnson, C.J., and Cluff, H.D. 2015. Den site
selection of wolves (Canis lupus) in response to declining
caribou (Rangifer tarandus groenlandicus) density in the
central Canadian Arctic. Polar Biology 38(12):2007 – 2019.
ht t ps:// -z
Lake, B.C., Bertram, M.R., Guldager, N., Caikoski, J.R., and
Stephenson, R.O. 2013. Wolf kill rates across winter in a
low-density moose system in Alaska. Journal of Wildlife
Management 77(8):1512 – 1522.
Laundré, J.W., Hernández, L., and Altendorf, K.B. 2001. Wolves,
elk, and bison: Reestablishing the “landscape of fear” in
Yellowstone National Park, U.S.A. Canadian Journal of
Zoology 79(8):1401 – 1409.
ht t ps:// 094
Laurenson, M.K. 1994. High juvenile mortality in cheetahs
(Acinonyx jubatus) and its consequences for maternal care.
Journal of Zoology (London) 234(3):387 – 408.
ht t ps: //doi.o rg /10.1111/j.1469-79 9 8.1994.tb04855.x
Lima, S.L. 1998. Nonlethal effects in the ecology of predator – prey
interactions: What are the ecological effects of anti-predator
decision-mak ing? Bioscience 48(1):25 – 34.
Macander, M.J., and Swingley, C.S. 2017. Landsat snow
persistence and snow regime mapping for Alaska, and lichen
cover mapping for Yukon-Charley Rivers National Preserve.
Final Report. Fairbanks, Alaska: ABR, Inc.
Manly, B.F.J., McDonald, L.L., Thomas, D.L., McDonald, T.L.,
and Erickson, W.P. 2002. Resource selection by animals:
Statistical design and analysis for eld studies, 2nd ed.
Dordrecht: Kluwer Academic Publishers. 221 p.
McLoughlin, P.D., Walton, L.R., Cluff, H.D., Paquet, P.C., and
Ramsay, M.A. 2004. Hierarchical habitat selection by tundra
wolves. Journal of Mammalogy 85(3):576 580.
Mech, L.D. 1970. The wolf: The ecology and behavior of an
endangered species. Minneapolis: University of Minnesota
Press. 384 p.
Mech, L.D., Adams, L.G., Meier, T.J., Burch, J.W., and Dale,
B.W. 1998. The wolves of Denali. Minneapolis: University of
Minnesota Press. 227 p.
Metz, M.C., Smith, D.W., Vucetich, J.A., Stahler, D.R., and
Peterson, R.O. 2012. Seasonal patterns of predation for gray
wolves in the multi-prey system of Yellowstone National Park.
Journal of Animal Ecology 81(3):553 – 563.
htt p s:// j.1365-2656.2011.01945. x
Mills, K.J., Patterson, B.R., and Murray, D.L. 2008. Direct
estimation of early survival and movements in eastern wolf
pups. Journal of Wildlife Management 72(4):949 954.
Mladenoff, D.J., Sickley, T.A., and Wydeven, A.P. 1999. Predicting
gray wolf landscape recolonization: Logistic regression models
vs. new eld data. Ecological Applications 9(1):37 – 44.[0037:PGWLRL]2.
Monahan, W.B., Rosemartin, A., Gerst, K.L., Fisichelli, N.A.,
Ault, T., Schwartz, M.D., Gross, J.E., and Weltzin, J. F. 2016.
Climate change is advancing spring onset across the U.S.
national park system. Ecosphere 7(10): e01465.
ht t ps:// cs2.1465
Moorcroft, P.R., Lewis, M.A., and Crabtree, R.L. 2006.
Mechanistic home range models capture spatial patterns and
dynamics of coyote territories in Yellowstone. Proceedings of
the Royal Society B: Biological Sciences 273(1594):1651 – 1659.
Murie, A. 1944. The wolves of Mount McKinley. Fauna of the
National Parks Series Number 5. Washington, D.C.: U.S.
Government Printing Ofce.
NPS (National Park Service). 1997. Land cover map of Yukon-
Charley Rivers National Preserve. Anchorage, Alaska: NPS.
Paine, R.T. 1980. Food webs: Linkage, interaction strength
and community infrastructure. Journal of Animal Ecology
49(3):6 67 – 685.
Pastick, N.J., Jorgenson, M.T., Wylie, B.K., Nield, S.J., Johnson,
K.D., and Finley, A.O. 2015. Distribution of near-surface
permafrost in Alaska: Estimates of present and future
conditions. Remote Sensing of Environment 168:301 – 315.
ht t ps://
Person, D.K., and Russell, A.L. 2009. Reproduction and den
site selection by wolves in a disturbed landscape. Northwest
Science 83(3):211 – 224.
Potts, J.R., and Lewis, M.A. 2014. How do animal territories form
and change? Lessons from 20 years of mechanistic modelling.
Proceedings of the Royal Society B: Biological Sciences
281(1784): 20140231.
Prugh, L.R., Stoner, C.J., Epps, C.W., Bean, W.T., Ripple, W.J.,
Laliberte, A.S., and Brashares, J.S. 2009. The rise of the
mesopredat or. BioScience 59(9):779 – 791.
R Core Team. 2017. R: A language and environment for statistical
computing. Vienna, Austria: R Foundation for Statistical
Ripple, W.J., and Beschta, R.L. 2012. Large predators limit
herbivore densities in northern forest ecosystems. European
Journal of Wildlife Research 58(4):733 742.
Ripple, W.J., Estes, J.A., Beschta, R.L., Wilmers, C.C., Ritchie,
E.G., Hebblewhite, M., et al. 2014. Status and ecological effects
of the world’s largest carnivores. Science 343(6167): 1241484.
Robin, X., Turck, N., Hainard, A., Tiberti, N., Lisacek, F.,
Sanchez, J.-C., and Müller, M. 2011. pROC: An open-source
package for R and S+ to analyze and compare ROC curves.
BMC Bioinformatics 12: 77.
Ross, S., Kamnitzer, R., Munkhtsog, B., and Harris, S. 2010. Den-
site selection is critical for Pallas’s cats (Otocolobus manul).
Canadian Journal of Zoology 88(9):905 – 913.
ht t ps://
Sandell, M. 1990. The evolution of seasonal delayed implantation.
Quarterly Review of Biology 65(1):23 42.
Sappington, J.M., Longshore, K.M., and Thompson, D.B. 2007.
Quantifying landscape ruggedness for animal habitat analysis:
A case study using bighorn sheep in the Mojave Desert.
Journal of Wildlife Management 71(5):1419 1426.
Schmidt, J.H., Burch, J.W., and MacCluskie, M.C. 2017. Effects
of control on the dynamics of an adjacent protected wolf
population in interior Alaska. Wildlife Monographs 198(1).
30 p.
ht t ps:// mon.1026
Sinclair, A.R.E., Mduma, S., and Brashares, J.S. 2003. Patterns
of predation in a diverse predator-prey system. Nature
425:288 – 290.
Smith, D.W., Metz, M.C., Cassidy, K.A., Stahler, E.E., Mcintyre,
R.T., Almberg, E.S., and Stahler, D.R. 2015. Infanticide
in wolves: Seasonality of mortalities and attacks at dens
support evolution of territoriality. Journal of Mammalogy
96(6):1174 – 1183.
Sorum, M.S., and Joly, K. 2016. Moose (Alces alces) population
survey in Yukon-Charley Rivers National Preserve, November
2015. Natural Resource Report NPS/YUCH/NRR—2016/1150.
Fort Collins, Colorado: National Park Service. 15 p.
Sousanes, P.J., and Hill, K.R. 2014. Annual climate summary
2012: Central Alaska Network. Natural Resource Data Series
NPS/CAKN/NRDS—2014/606. Fort Collins, Colorado:
National Park Service.
Therneau, T. 2015. A package for survival analysis in S, version
Theuerkauf, J., Rouys, S., and Jedrzejewski, W. 2003. Selection of
den, rendezvous, and resting sites by wolves in the Bialowieza
Forest, Poland. Canadian Journal of Zoology 81(1):163 167.
Trapp, J.R., Beier, P., Mack, C., Parsons, D.R., and Paquet, P.C.
2008. Wolf, Canis lupus, den site selection in the Rocky
Mountains. Canadian Field-Naturalist 122(1):49 56.
Van Ballenberghe, V., and Mech, L.D. 1975. Weights, growth,
and survival of timber wolf pups in Minnesota. Journal of
Mammalog y 56(1):44 – 63.
Walsh, P.B., Sethi, S.A., Lake, B.C., Mangipane, B.A., Neilson,
R., and Lowe, S. 2016. Estimating denning date of wolves with
daily movement and GPS location x failure. Wildlife Society
Bulletin 4 0(4):663 – 668.
Warton, D.I., and Hui, F.K.C. 2011. The arcsine is asinine: The
analysis of proportions in ecology. Ecology 92(1):3 10.
Zuur, A.F., Ieno, E.N., and Elphick, C.S. 2010. A protocol for data
exploration to avoid common statistical problems. Methods in
Ecology and Evolution 1(1):3 14.
... However, in spring, the pregnant female chooses a den, a hole in the ground, rocky nook or cave, or a shallow pit in the ground (Mech 1993). Wolves tend to locate dens optimally in relation to food supplies (Joly et al. 2018). ...
... 4. Breeding siblings or other camp wolves, the breeding pair scent-marks a territory that includes the camp and howl frequently, thus keeping wild wolves away (Mech & Boitani 2003). 5. The females den nearby, possibly in the camp where food is available (Joly et al. 2018) and where they are protected by their human 'pack'. Female wild wolf 7268, mentioned above, denned within 1 km of humans who were feeding her; wolves raised by humans and totally dependent on them for food would den closer to them. ...
Full-text available
The dog was the first domesticated animal. Its derivation from grey wolves Canis lupus is important to the study of mammalian domestication, and wolf domestication is an active area of investigation. Recent popular books have promoted a hypothesis that wolves domesticated themselves as opposed to the earliest hypothesis that featured pup collection, adoption, and artificial selection. Continuing research has produced a greater understanding of wolf ecology and behaviour, including new insights into the wolf’s interaction with humans. Several characteristics make the wolf conducive to domestication: its sociality, catholic diet, excellent individual and cultural memory, inbreeding tolerance, varied personalities, and adaptable lifestyle. The wolf’s fear of humans is the main impediment and that alone is a factor strongly disfavouring the self‐selection hypothesis. However, collecting young pups from dens and raising them would foster their socialising with humans as pack members. Neither hypothesis explains how wolves undergoing domestication were separated reproductively from their wild relatives, an important condition for domestication. We combine information from the literature with information from our own research on wild wolves, archaeology, and canid morphology. We explain how pup collection and deliberate or incidental selection and encouragement to breed with similarly raised wolves could keep incipient dogs separated reproductively from wild relatives. The key is humans regularly feeding the wolves and keeping only those able to live harmoniously with humans. Well‐fed, human‐dependent wolves would remain near their food supply and in the company of humans, thus increasing their bonds to humans and vice versa. Outbreeding with wild wolves would thus be avoided. Generation after generation of these human‐fed, raised, and selected wolves would become increasingly dependent on humans and shaped by them. The pup‐adoption hypothesis presented here is more in keeping with basic wolf ecology and behaviour than the self‐domestication hypothesis. Dogs were domesticated from wolves 15000–25000 years ago, and two theories prevail about how the domestication process originated: 1) wolves domesticated themselves by frequenting human camps and feeding on discarded food, and 2) wolf pups were collected from dens and raised by humans selecting those most tractable and suitable for living with humans. This review is the first to assess these theories in relation to the characteristics of wolf ecology and behaviour that make the wolf suitable for domestication. The second hypothesis of pup adoption followed by selection seems better supported.
... packs) within populations (Asa andValdespino 1998, Mech andBoitani 2010). Although very little is known about the underlying mechanisms associated with the timing of wolf reproduction, the synchrony in spring parturition, as well as the observed delay in parturition at higher elevations (Joly et al 2018) and latitudes (Mech and Boitani 2010), indicate strong selective pressures acting on wolf denning phenology with possible cuing by climate signals. Therefore, we evaluated potential climate mechanisms underlying two components of grey wolf reproduction using a large regional dataset from North America (tables 1 and 2). ...
... We estimated parturition as the initial date from the earliest denning cluster observed during the reported period for wolf parturition (March-June; Mech and Boitani 2010). Occasionally, movement data from a reproductive female expressed gaps in a location time series that lasted approximately 4 d to 8 d, an indication of GPS satellite signal occlusion while in underground dens (Joly et al 2018). Visual inspection of the time series helped to identify these gaps so that den initiation dates (i.e. ...
Full-text available
Arctic and boreal ecosystems are experiencing rapid changes in temperature and precipitation regimes. Subsequent shifts in seasonality can lead to a mismatch between the timing of resource availability and species' life-history events, known as phenological or trophic mismatch. Although mismatch has been shown to negatively affect some northern animal populations, longer-term impacts across large regions remain unknown. In addition, animals may rely on climate cues during preceding seasons to time key life history events such as reproduction, but the reliability of these cues as indicators of subsequent resource availability has not been examined. We used remote sensing and gridded spatial data to evaluate the effect of climate factors on the reproductive phenology and success of a wide-ranging carnivore, the gray wolf (Canis lupus). We used GPS location data from 388 wolves to estimate den initiation dates (n = 227 dens within 106 packs) and reproductive success in eight populations across northwestern North America from 2000-2017. Spring onset shifted 14.2 days earlier, on average, during the 18-year period, but the regional mean date of denning did not change. Preceding winter temperature was the strongest climatic predictor of denning phenology, with higher temperatures advancing the timing of denning. Winter temperature was also one the strongest and most reliable indicators of the timing of spring onset. Reproductive success was not affected by timing of denning or synchrony with spring onset, but improved during cooler summers and following relatively dry autumns. Our findings highlight a disconnect between climate factors that affect phenology and those that affect demography, suggesting that carnivores may be resilient to shifts in seasonality and yet sensitive to weather conditions affecting their prey at both local and regional scales. These insights regarding the relationship between climate and carnivore demography should improve predictions of climate warming effects on the highest trophic levels.
... When dependent offspring are confined to a breeding site, adults must commute from the same location each day to obtain food, a process referred to as central place foraging (Chapman et al. 1989;Elliott et al. 2009). Den or nest site location has been linked to proximity of habitats containing resources for numerous species, such as racoons (Procyon lotor), wolverines (Gulo gulo), wolves (Canis lupus), and arctic foxes (Vulpes lagopus) (Magoun and Copeland 1998;Norris et al. 2002;Henner et al. 2004;Szor et al. 2008;Joly et al. 2018). In the presence of a predator, however, animals necessarily alter what would otherwise be an optimal foraging strategy to avoid these predators (Cowlishaw 1997), potentially increasing the energetic costs of foraging (Sih 1980). ...
Full-text available
Many species rear offspring in fixed sites, returning frequently to provision them, and the selection of these sites is a critical decision in the life cycle, as they may in some instances increase susceptibility to predators. African wild dogs are a groupliving large carnivore that rear their offspring in fixed sites, provisioning dependent pups in dens for 3 months post-birth. Where possible, African wild dogs select den sites in rocky terrain, and it is hypothesised that this is because lions, their main predators, generally avoid this habitat. In the Okavango Delta, Botswana, there is a lack of rocky terrain, providing an opportunity to assess whether lions drive den site selection. GPS collar data from 7 impala and 4 lions revealed that both species prefer to reside in grassland and mixed woodland habitats, demonstrating that these are high risk/reward areas for African wild dogs. Using GPS collar data from 16 African wild dog packs over 8 years, our study characterised 116 African wild dog den sites identified in the field. Packs showed a preference for denning in mopane woodland, which lions avoid, and packs commuted further from the den each day as the den’s distance to grassland and mixed woodland increased, suggesting a preference for hunting in this habitat. Our results suggest that African wild dogs trade-off the costs of commuting and predation risk, such that longer commuting costs confer increased safety. Significance statement Species which utilise dens, nests, or other fixed sites to rear offspring must balance the need to protect their young from predators with the need to acquire resources for themselves and their young. The selection of den sites is expected to be of considerable importance to enable the animal to meet these two requirements and successfully raise young. Our study of African wild dogs indicates that they select dens in resource-scarce areas which are likely to minimise interactions with their main predator, lions. This increases the distance to prey-rich areas and therefore the cost of hunting. Availability of appropriate habitat for both hunting and denning is therefore important when considering landscapes appropriate for African wild dog conservation, energetic constraints of breeding, and home-range indices.
... Nearly half of these papers sampled for only one or two years post-fire (43.5%). Black bears (Ursus americanus, Schwartz and Franzmann, 1991), grizzly bears (McLellan and Hovey, 2001), and wolves (Canis lupus, Joly et al., 2018) were the only species to have studies lasting ≥10 years; all of these studies were on wildfires. ...
Shifting fire regimes are substantially changing North American forests. It is thus critical to understand how wildfires affect forest wildlife, especially for species managed for harvest and for species at risk of extinction. In particular, many populations of carnivores and ungulates are actively managed, so being able to anticipate their responses to future fires, or as burned forests regenerate, would be valuable for management and conservation. We examine how well the current literature addresses these important needs. We reviewed 131 papers published from 1970 to 2019 that reported carnivore and ungulate responses to fires in North American conifer forests. We evaluated the study designs, fire attributes, species studied, and response variables measured. This literature contains considerable taxonomic bias, inconsistent reporting of fire and landscape characteristics, and a mismatch between research objectives and management needs. Of the 18 carnivore species studied post-fire, just three (grizzly bears Ursus arctos, Canada lynx Lynx canadensis, and American marten Martes americana) comprised nearly half of the literature (47.8%). Few papers reported the spatial extents of fires being studied (32.6% for carnivores and 38.2% for ungulates). Most studies focused on recent burns (<15 years post-fire) and monitored post-fire responses for fewer than five years. Few papers measured true population responses post-fire (6.7% for carnivores and 12.4% for ungulates); the majority of papers focused on animal movements or habitat use (presence/absence or relative activity) rather than estimates of survival, recruitment, or population size. Our review also indicates that studying fires with a more diverse array of spatial extents, ages, and severities would be useful for most species.
Full-text available
Wolves are important to keep ecosystems healthy. For wolf populations to thrive, pups need to survive into adulthood. Wolf pups can be harmed, even killed, by wolves from outside their pack. To protect the pups, some wolves in the pack must stay at the den and guard the pups. But some of the adults must leave the den sometimes, to hunt for food and to keep other wolves out of the pack’s territory. By monitoring and studying wolves for many years across North America and in Alaska’s national parks, we are learning how wolves divide these tasks. We know that the mother wolves care for their pups for the first several weeks while nursing, but once the pups no longer need milk, all pack members share in taking care of the pups. We have come to understand that all wolves are important to the success of the pack.
Full-text available
Knowledge of how disturbances such as fire shape habitat structure and composition, and affect animal interactions, is fundamental to ecology and ecosystem management. Predators also exert strong effects on ecological communities, through top‐down regulation of prey and competitors, which can result in trophic cascades. Despite their ubiquity, ecological importance and potential to interact with fire, our general understanding of how predators respond to fire remains poor, hampering ecosystem management. To address this important knowledge gap, we conducted a systematic review and meta‐analysis of the effects of fire on terrestrial, vertebrate predators worldwide. We found 160 studies spanning 1978–2018. There were 36 studies with sufficient information for meta‐analysis, from which we extracted 96 effect sizes (Hedge's g) for 67 predator species relating to changes in abundance indices, occupancy or resource selection in burned and unburned areas, or before and after fire. Studies spanned geographic locations, taxonomic families, and study designs, but most were located in North America and Oceania (59% and 24%, respectively), and largely focussed on felids (24%) and canids (25%). Half (50%) of the studies reported responses to wildfire, and nearly one third concerned prescribed (management) fires. There were no clear, general responses of predators to fire, nor relationships with geographic area, biome or life history traits (e.g. body mass, hunting strategy and diet). Responses varied considerably between species. Analysis of species for which at least three effect sizes had been reported in the literature revealed that red foxes (Vulpes vulpes) mostly responded positively to fire (e.g. higher abundance in burned compared to unburned areas) and eastern racers (Coluber constrictor) negatively, with variances overlapping zero only slightly for both species. Our systematic review and meta‐analysis revealed strong variation in predator responses to fire, and major geographic and taxonomic knowledge gaps. Varied responses of predator species to fire likely depend on ecosystem context. Consistent reporting of ongoing monitoring and management experiments is required to improve understanding of the mechanisms driving predator responses to fire, and any broader effects (e.g. trophic interactions). The divergent responses of species in our study suggest that adaptive, context‐specific management of predator‐fire relationships is required.
Full-text available
Predation, habitat, hunting, and environmental conditions have all been implicated as regulatory mechanisms in ungulate populations. The low-density equilibrium hypothesis predicts that in low-density populations, predators regulate their prey and that the population will not escape unless predation pressure is eased. We evaluated survival of adult and juvenile moose (Alces alces) in north-central Alaska to determine whether or not the population supported the hypothesis. We instrumented adult male and female moose with radiocollars and used aerial observations to track parturition and subsequent survival of juvenile moose. Generalized linear mixed-effects models were used to assess survival. Adult annual survival rates were high (∼89%), but may be negatively influenced by winter conditions. Migratory status did not affect moose survivorship or productivity. Approximately 60% of the calf crop died before 5 months of age. Productivity was significantly lower in the northern section of the study area where there is less high-quality habitat, suggesting that, even in this low-density population, nutrition could be a limiting factor. It appears that predation on young calves, winter weather, and nutritional constraints may be interacting to limit this population. Latent traits, such as overproduction of calves and migratory behavior, which do not currently enhance fitness, may persist within this population so that individuals with these traits can reap benefits when environmental conditions change.
Full-text available
Long-term wolf (Canis lupus) research programs have provided many insights into wolf population dynamics. Understanding the mechanisms controlling responses of wolf populations to changes in density, environmental conditions, and human-caused mortality are important as wolf management becomes increasingly intensive. Competition with humans for ungulate prey has led to large-scale wolf control programs, particularly in Alaska, and although wolf populations may sustain relatively high (e.g., 22–29%) rates of conventional harvest, control programs are specifically designed to have lasting population-level effects.
Full-text available
Linkages between atmospheric, ecological and biogeochemical variables in the changing Arctic are analyzed using long-term measurements near Utqiaġvik (formerly Barrow), Alaska. Two key variables are the date when snow disappears in spring, as determined primarily by atmospheric dynamics, precipitation, air temperature, winter snow accumulation and cloud cover, as well as the date of onset of snowpack in autumn that is additionally influenced by ocean temperature and sea ice extent. In 2015 and 2016 the snow melted early at Utqiaġvik due mainly to anomalous warmth during May of both years attributed to atmospheric circulation patterns, with 2016 having the record earliest snowmelt. These years are discussed in the context of a 115-year snowmelt record at Utqiaġvik with a trend toward earlier melting since the mid- 1970s (-2.86 days/decade, 1975-2016). At nearby Cooper Island, where a colony of seabirds, Black Guillemots, have been monitored since 1975, timing of egg laying is correlated with Utqiaġvik snowmelt with 2015 and 2016 being the earliest years in the 42-year record. Ice-out at a nearby freshwater lagoon is also correlated with Utqiaġvik snowmelt. The date when snow begins to accumulate in autumn at Utqiaġvik shows a trend towards later dates (+4.6 days/decade, 1975-2016), with 2016 the latest on record. The relationships between the lengthening snow-free season and regional phenology, soil temperatures, fluxes of gases from the tundra, and to regional sea ice conditions are discussed. Better understanding of these interactions is needed to predict the annual snow cycles in the region at seasonal to decadal scales, and to anticipate coupled environmental responses.
Full-text available
We used Global Positioning System (GPS) radiotelemetry data from 7 breeding female wolves (Canis lupus; n = 14 dennings) in 3 regions across Alaska, USA, during 2008–2011 to develop and compare methods for estimating the onset of denning, and thus infer timing of parturition. We developed and tested 2 estimators based on a combination of GPS radiocollar location-fix failure and distance traveled between locations. We developed a quantitative method employing Generalized Additive Models to smooth time series of wolf data to estimate denning onset. In contrast, 3 study authors with first-hand experience with the study wolves implemented a subjective method of estimating denning onset by visual inspection of detection and distance traveled data. We then tested the visual method for repeatability by subjecting it to 10 wolf experts not associated with this study. Side-by-side comparison of estimators indicates that denning onset can be precisely measured using GPS detection success and distance traveled. Furthermore, the visual-inspection method was simple and rapid to implement and yielded more accurate (relative to assumed dates of denning onset) and precise results compared to the quantitative estimator. Although the Generalized Additive Model based approach had the advantage of estimating denning onset objectively following a set of prescribed rules in a statistical inferential framework, we found the method required significant technical capacity to implement and did not represent an improvement over simple visual-inspection-based estimates of denning onset. Published 2016. This article is a U.S. Government work and is in the public domain in the USA.
Full-text available
Many U.S. national parks are already at the extreme warm end of their historical temperature distributions. With rapidly warming conditions, park resource management will be enhanced by information on seasonality of climate that supports adjustments in the timing of activities such as treating invasive species, operating visitor facilities, and scheduling climate-related events (e.g., flower festivals and fall leaf-viewing). Seasonal changes in vegetation, such as pollen, seed, and fruit production, are important drivers of ecological processes in parks, and phenology has thus been identified as a key indicator for park monitoring. Phenology is also one of the most proximate biological responses to climate change. Here, we use estimates of start of spring based on climatically modeled dates of first leaf and first bloom derived from indicator plant species to evaluate the recent timing of spring onset (past 10–30 yr) in each U.S. natural resource park relative to its historical range of variability across the past 112 yr (1901–2012). Of the 276 high latitude to subtropical parks examined, spring is advancing in approximately three-quarters of parks (76%), and 53% of parks are experiencing “extreme” early springs that exceed 95% of historical conditions. Our results demonstrate how changes in climate seasonality are important for understanding ecological responses to climate change, and further how spatial variability in effects of climate change necessitates different approaches to management. We discuss how our results inform climate change adaptation challenges and opportunities facing parks, with implications for other protected areas, by exploring consequences for resource management and planning.
Changes in behavior and habitat use are often influenced by the risk of predation, including harvest, and carnivores alter their habitat use and movements to minimize predation risk. Large carnivores are subject to harvest around the world; however, few studies examine whether habitat use is different between harvested and unharvested carnivore populations. We examined the effects of harvest on gray wolf (Canis lupus) use of pup-rearing habitat. We predicted that in comparison to an unharvested population of wolves, wolves subject to harvest would use less suitable pup-rearing habitat (i.e., sites with no standing water source and dense vegetation with no open areas) and locate pup-rearing sites in areas with lower human activity. We also predicted that wolves would use less suitable pup-rearing habitat following breeder turnover. Finally, we predicted that field surveys using a method for monitoring an unharvested wolf population would detect fewer active pup-rearing sites and document fewer detections of scat, tracks, and howls of a harvested wolf population. We tested whether a habitat model for predicting pup-rearing sites used by an unharvested wolf population accurately predicted sites used by a harvested wolf population. To examine the effect of human activity, we calculated the road density within a 500-m buffer around sites used by an unharvested and harvested population of wolves. We also evaluated the habitat suitability of pup-rearing sites following the death of a breeder. Finally, we conducted field surveys of a harvested wolf population with a monitoring technique used for unharvested wolf populations, and compared detections of pup-rearing sites, scats, tracks, and howls between the unharvested and harvested wolf populations. Harvest did not affect wolf use of pup-rearing habitat. Wolves subject to harvest used highly suitable habitat (i.e., areas with standing, ephemeral water) to raise pups, road density near pup-rearing sites did not differ between harvested and unharvested wolf populations, and breeder turnover did not result in packs choosing less suitable pup-rearing habitat. Finally, field surveys successfully detected pup-rearing sites of wolves subject to harvest but documented fewer detections of scats and tracks, likely because of a decrease in wolf density. Wolves subject to harvest chose highly suitable habitat to raise pups, indicating that such habitat provides optimal resources in a landscape where harvest is a dominant source of mortality. In areas where such habitat is limited, it is important to consider how environmental changes affect the availability of suitable habitat. © 2018 The Wildlife Society.
Within the 2700km² Beltrami Island State Forest, near the W edge of the primary range of Canis lupus in Minnesota, wolf population density was low at the start of the study in 1972 but increased substantially up to 1977 (end of study). At least 8 of 13 social units present in mid-1976 had formed since 1972. Size of litters of established packs averaged 4.6 pups, and those of newly-formed pairs averaged 4.1. Mortality decreased over the study period, and recruitment of young wolves exceeded mortality following legal protection. A high rate of dispersal of young from packs was documented. Dispersal peaked in autumn. Most wolves paired within a few days of leaving their packs. Average territory size decreased as both population and pack numbers increased. Behaviour of alpha males, alpha females and subordinate members of the packs is discussed. Deer and moose comprised 94% of animal biomass eaten by wolves, with deer along accounting for 67%. Seasonal differences in food taken and energy requirements are noted.-P.J.Jarvis
Understanding the limiting factors of a prey population is important before and during predator control programs, and optimal intensive management of an increasing prey population requires formal recognition of a sustainable population size. The migratory Fortymile caribou (Rangifer tarandus) herd in Alaska reached a low of approximately 6,000 caribou during 1973–1975. To regain peak numbers of approximately 50,000 caribou estimated in the 1960s, stakeholder groups gained approval for conservative harvest rates (1973–2013) and periods of restricted nonlethal (1998–2004) and lethal wolf (Canis lupus) control (2005–2013). We studied demography of the herd using radio-telemetry during 1990–2014, when herd size increased from about 22,000 to 52,000 caribou. Parturition rates in the early 1990s were among the highest reported, but parturition rates of primiparous females subsequently declined to a level indicating resource-limitation as caribou numbers approached and then exceeded 50,000. This and companion studies documented several other cautionary signals to an eventual decline, including declining October calf weights, early summer movement off the alpine and subalpine tundra to lower elevation spruce–moss taiga, relatively high caribou densities, a nearly 40-year history of increasing caribou numbers, and a return to previous peak numbers. We studied mortality of calves and older females during the 4 years before wolf control and the first 5 years of nonlethal wolf control. During those 9 years, annual mortality rates averaged 54% for calves and 9% for adult females. We detected no convincing support for decreased wolf predation during nonlethal control. We also detected no support for increased caribou survival during nonlethal or lethal wolf control. Based on counts of caribou during summer aggregations using a total search photocensus technique, rate of herd increase (λr) was negligible (λr = 1.00) during 1990–1995, highest during the 3 years immediately before nonlethal wolf control (λr = 1.11, 1995–1998), moderate during nonlethal wolf control (λr = 1.07, 1998–2003), and low during the period that included the first 5 years of lethal wolf control (λr = 1.02, 2003–2010). We combined observed cause of death with the 9 annual modeled starting populations (all newborn calves and adults) and estimated that wolves killed 10–15% of the populations annually, grizzly bears (Ursus arctos) killed 4–7%, other predators killed 2–4%, nonpredation factors killed 1–2%, and hunters killed ≤2%. Wolves killed 5–9% of the annual populations as calves and 5–6% as adults. In retrospect, nonlethal wolf control efforts were too localized to decrease wolf numbers (e.g., adjacent untreated wolf packs reached max. mean numbers). Lethal wolf control efforts had only seasonal and localized effects on wolf numbers. It is important that stakeholders focus on describing a preferred, sustainable herd size, or nutritional status and proceed toward managing this increasing herd in a sustainable manner because, when ungulates overshoot carrying capacity, the effects of high density, adverse weather, and increased predation can have synergistic negative effects on prey numbers and long-lasting negative effects on sustainable yields, contrary to the intended purpose of the wolf control programs.