MOOSE HABITAT USE THROUGHOUT GROS MORNE NATIONAL PARK

Krystal Kerckhoff1, Brian E. McLaren1, Shane P. Mahoney2 and Tom W. Knight3

1Lakehead University, Faculty of Natural Resources Management, 955 Oliver Road, Thunder Bay, ON,
Canada P7B 5E1; 2Government of Newfoundland and Labrador, Sustainable Development and
Strategic Science Division, 2 Canada Drive, St. John’s, NL, Canada A1B 4J6; 3Gros Morne National
Park, P.O. Box 130, Rocky Harbour, NL, Canada A0K 4N0.

ABSTRACT: Previous research indicated high variability in availability and habitat use by female
moose in the lowlands of Gros Morne National Park (GMNP), Newfoundland and Labrador, an area
dominated by bogs and forest. Here, we extend the earlier analysis with an additional 7 female moose
(Alces alces americana) occupying the Park highlands, a region dominated by heath and shrub vegeta-
tion with forest limited to sheltered valleys, typical of interior and highland parts of the province.
Resource selection function (RSF) models with differences in habitat use between moose resident in
the 2 regions and 2 moose that migrated from the lowlands in winter to the highlands in summer
were rejected. In summer, more use of closed-canopy forest types occurred on the lowlands, while
more use of non-forest habitat types occurred on the highlands. As before, we found that selection
of disturbed forest is a winter phenomenon on the lowlands of GMNP; the same series of habitat types
associated with disturbance were avoided in summer. Summer migration by about 20% of GMNP
moose to the highlands suggests that foraging opportunities are better during that season than in winter,
a motivation for migration perhaps augmented by an overabundance of moose on the lowlands and
unfavourable temperatures in disturbed areas that might otherwise serve as lowland foraging areas.
An observation of more clustered relocations of moose on the highlands than on the lowlands of
GMNP is consistent with our conclusion that moose use habitats within the highlands and lowlands
of Newfoundland and Labrador very differently. We recommend 2 approaches to moose management
for these different landscapes, both within GMNP and elsewhere in Newfoundland and Labrador.

ALCES VOL. 49: 113–125 (2013)

Key words: Gros Morne National Park, habitat selection, moose, Newfoundland and Labrador,
resource selection functions.

According to habitat selection theory
(Fretwell and Lucas 1970), individuals dis-
tribute themselves in a manner proportional
to the quantity or quality of limiting
resources available in each of several fora-
ging patches, larger habitat units, and still
larger landscapes. For ungulates, habitat
selection should be driven by an individual’s
ability to sense and select higher-quality
food items or foraging areas (McNaughton
1985, Fryxell 1991). Habitat selection that
involves migration between 2 different land-
scapes can arise in a seasonal climate where
different fitness opportunities (or forage

availability) are offered by each landscape,
but where the difference is less during the
growing season (Holt and Fryxell 2011).
Moose (Alces alces americana) in New-
foundland and Labrador, Canada presumably
distribute themselves optimally according
to habitat selection theory in each of 2 typi-
cal landscapes in this province, the “high-
lands” and the “lowlands.” We explore this
idea with analysis of summer and winter
location data from GPS-collared moose,
using the assumption that more forested,
lowland landscapes are, on average (i.e.,
throughout the year), superior to the

113



highlands where some moose migrate during
summer.

This paper is motivated by previous
study of Gros Morne National Park
(GMNP), an area of 1,805 km2 in western
Newfoundland, where approximately 20%
of the female moose population migrates
within the Park from forested, coastal low-
lands (< 400 m above sea level) in winter to
relatively open highlands (between 400 and
800 m) during summer (McLaren et al.
2000). Our interpretation is that spending
summer in a highland landscape offers an
advantage to this fraction of the moose popu-
lation. We compared seasons of activity of
the resident moose in the highlands and
lowlands of GMNP, and described their
finer-scale activity in terms of frequency of
smaller movements and the densities in
which these smaller movement clusters occur
throughout, by comparing the 2 landscapes.

During winter, snow limits accessibility
to forage more on the highlands than in the
coastal lowlands (Martin 2004), a motivation
for migration that is consistent with empiri-
cal evidence from other studies of moose
(reviewed by Ball et al. 2001). As an addi-
tional explanation for the moose migration
within GMNP, as suggested in McLaren et al.
(2000), summer migration from the lowlands
may be a means to avoid black bear (Ursus
americana) predation on calves, because
the highlands may offer easier escape from
this predator given the longer sightlines in
open habitats. This second idea would be
similar to the explanation for why woodland
caribou (Rangifer tarandus caribou) often
migrate up mountain slopes (Bergerud et al.
1984), and for why elk (Cervus canadensis)
migrate between high and low elevations in
Alberta (Hebblewhite and Merrill 2007,
Hebblewhite et al. 2008). However, because
moose densities are about tenfold higher
in Newfoundland than in other parts of
their range in North America (McLaren
et al. 2004), creating obvious effects on

hampering regeneration in the forests of the
coastal plain of GMNP (Connor et al. 2000,
McLaren et al. 2004, Gosse et al. 2011,
Humber and Hermanutz 2011), and because
only a fraction of the population migrates,
we favour limited forage availability in the
lowlands as the primary factor for summer
moose migration. To explore this hypothesis,
which is consistent with the Holt and Fryxell
(2011) model for migration, we compare the
frequency of the fine-scale summer move-
ment on the lowlands and the highlands,
and compare resource selection functions
(RSFs) for highland and lowland moose
in GMNP.

STUDY AREA
GMNP is located on the Gulf of the St.

Lawrence on the northern peninsula of New-
foundland. Its lowlands, which encompass
parts of the Western Newfoundland Forest
and the Coastal Plain sub-region of the
Northern Peninsula Forest (Damman 1983),
are characterized by weather influences
from the Gulf, producing moderate levels
of annual precipitation (900–1000 mm)
and cold and snowy winters (300–350 mm
is in the form of snow; Hare 1952). Its high-
lands, which are situated in the Long Range
Barrens (Damman 1983), are similarly influ-
enced by the Gulf, but with an orographic
effect that creates a harsher climate, having
annual precipitation and snowfall on average
double that of the lowlands (Watson 1974).
The mean annual temperature on the high-
lands is 4.5 °C colder than that of the low-
lands (Banfield 1983).

In 1878, one female and one male moose
were introduced to Newfoundland from
Nova Scotia, and in 1904, 2 male and 2
female moose were introduced from New
Brunswick (Pimlott 1953). Moose first
inhabited the northern peninsula of New-
foundland by the 1940s (Caines and Deich-
mann 1989). While moose are currently
found in all ecoregions of Newfoundland,

114

HABITAT USE THROUGHOUT GROS MORNE PARK – KERCKHOFF ET AL. ALCES VOL. 49, 2013



their density varies considerably. In the late
1970s when GMNP was being established,
moose increased first on the highlands and
by the 1980s moose had increased through-
out the Park (Connor et al. 2000). At the
time of the GPS collaring, surveys using
stratified random blocks estimated the
moose population at 7,377 ± 1,249 (4.1 ±
0.7 moose/km2; McLaren et al. 2000;
GMNP, unpublished data). In 2007, popula-
tion size was estimated separately for the
two landscapes, at 3,975 ± 1,287 in the low-
lands (4.2 ± 1.4 moose/km2) and 788 ± 223
SD in the highlands (0.9 ± 0.3 moose/km2);
densities in partial surveys of the park were
estimated in 2009 at 5.9 moose/km2 on the
lowlands and 1.1 moose/km2 on the high-
lands (GMNP, unpublished data).

METHODS
Habitat Classification

Taylor and Sharma (2010) classified
habitat types on the lowlands and highlands
of GMNP from a single-image subset of 2,
10-m multispectral SPOT-5 satellite images
(recorded 20 June 2006) with a K-means
unsupervised classification. Classes were
reorganized and described using information
from aerial photographs and forest inven-
tories, and local expert knowledge and field
visits. Ten habitat types resulted for the low-
lands (Table 1), and 6 for the highlands
(Table 2). Collectively, the lowlands com-
prise 938 km2 or 52% of the Park, of which
417 km2 or 44% is moose habitat in forest
or disturbed forest types; the highlands com-
prise 867 km2 or 48% of the Park, 641 km2 or
74% of which is moose habitat, but only a
fraction of which is forest (Table 3). The
classifications in Table 1 and 2 are the refer-
ence for our description of habitat use by
moose.

Moose Locations
In June 1997, 12 adult female moose (11

with at least one calf) were immobilized and

fitted with GPS collars (Lotek Engineering,
Inc.; McLaren et al. 2000; Table 4). The col-
lars were set to attempt a fix at 3-h intervals.
Remote downloading occurred in September
1997, November 1997, and March 1998.
The collars were removed in November
1998, and the remaining data records were
collected at that time. Location accuracy
was found to be dependent on collar position
in relation to topography and canopy, but
95% of all differentially corrected data
from test collars had ± 25 m accuracy
(Moen et al. 1997, McLaren et al. 2000).
All 2-dimensional fixes were removed from
the dataset, and only differentially corrected
locations were used in the current analysis.
Depending on collar functioning, locations
were recorded over a 4–15.5 month period
(Table 4). Five of the collared moose were
year-round residents in the lowlands, 5
were year-round residents in the highlands,
and the remaining 2 migrated seasonally
between the 2 landscapes.

Data Analysis
The dataset was divided into summer

and winter seasons following Vander Wal
and Rodgers (2009). Six moose were used
for calculation of seasonal transition dates;
3 in the lowlands and 3 in the highlands
with sufficient data records to span most of
a calendar year. For these moose, cumulative
distance travelled was calculated in ArcView
version 9 (ESRI, Redlands, California) and
plotted against time beginning with 1 Janu-
ary. Winter was defined as the period when
rate of travel was less than the mean rate,
estimated from the points of inflection of
the best-fit logistic curves to the plots, where
the estimated changes from winter to
summer and from summer to winter are sym-
metric around the inflection points. Curve-
fitting used the logistic regression program
in the Statistical Package for the Social
Sciences (SPSS), version 18 (also used for
all subsequent analysis). The median dates

ALCES VOL. 49, 2013 KERCKHOFF ET AL. – HABITAT USE THROUGHOUT GROS MORNE PARK

115



for the start and end of winter were estimated
from the 3 curves for each of the 2 land-
scapes and used to define the seasons for
all subsequent analysis.

Summer and winter home ranges and
core-use areas were calculated using the
fixed-kernel method in Home Range Tools

(Rodgers et al. 2007) with Gaussian (bivari-
ate normal) distributions, reporting the 95%
and 50% isopleths for ranges and cores,
respectively. The bandwidth size was deter-
mined by finding the smallest proportion of
the reference bandwidth that allowed one
continuous outer line to encompass the

Table 1. Habitat descriptions from a lowlands classification of Gros Morne National Park, Newfoundland,
Canada.

Habitat type Description Category

Mature softwood
forest

Softwood dominated, especially balsam fir (Abies
balsamea); some mixed stands with white birch (Betula
papyrifera).

Closed-canopy forest

Closed spruce
forest

Softwood dominated (balsam fir and black spruce, Picea
mariana); other species include tamarack (Larix
laricina), trembling aspen (Populus tremuloides) and
alder (Alnus spp.); site condition can be wet. Some stands
of scrub forest.

Closed-canopy forest

Closed mixed forest Balsam fir dominated with some mixed stands (balsam
fir, white birch). Stem density can be very high. Younger
mixed stands (∼30 years since disturbance) are included.

Closed-canopy forest

Young softwood
forest

Softwood dominated with high content of hardwoods;
canopy >50% and 6–9 m in height.

Closed-canopy forest

Open softwood
forest

Balsam fir dominated with 25–50% open canopy; white
birch can be significant; some tree regeneration (heights
of 1–4 m).

Open-canopy forest

Open mixed forest Softwood dominated with 25–50% open canopy.
Sometimes wet. Trees shorter than in closed mixed
forest; some tree regeneration.

Open-canopy forest

Open hardwood
forest

Hardwood dominated with 25–50% open canopy. Often
originally a mixed forest where regeneration of balsam fir
does not occur.

Open-canopy forest

Sparse softwood
forest

Softwood dominated (balsam fir, black spruce)
with <25% canopy; limited regeneration; ferns and grass
very prominent (<50% of ground cover); forest canopy is
very broken consisting of mostly remnant forest from
past disturbance; low density young black spruce <6 m
height; pockets of conifer regeneration <4 m height can
be present.

Disturbed forest: sparse canopy
with herb/grass ground cover

Herb-hardwood
forest

Dominant plants include ferns, grass and raspberry
(Rubus spp.) >50% of ground cover; very sparse forest
canopy; some remnant white birch with alder or
elderberry (Sambucus racemosa). Very little balsam fir.
Scattered spruce <4 m height. Includes forested areas that
have not regenerated after severe disturbance.

Disturbed forest: sparse canopy
with herb/grass ground cover

Herb forest Dominant plants include ferns and grass (>50% of
ground cover); exposed soil is common; large amounts of
dead material (standing or fallen) and scattered remnant
trees. Little regeneration >30 cm height. Mostly forested
areas that have not regenerated after severe disturbance.

Disturbed forest: sparse canopy
with herb/grass ground cover

116

HABITAT USE THROUGHOUT GROS MORNE PARK – KERCKHOFF ET AL. ALCES VOL. 49, 2013



polygons (Worton 1989). Areas of open
water, wetlands, and rock barrens were
excluded from each of the resulting polygons
and the remaining area was divided into the
habitat types appropriate to the landscape.

Fine-scale habitat use examined areas where
a minimum of 3 consecutive GPS locations <
24 h apart occurred, with distances between
them of < 50 m. This definition of an impor-
tant habitat patch was arbitrary, but based on
an inference that foraging and other activities
such as bedding take place with shorter tra-
vel distances. Mean weekly travel distances,
as well as distances between the habitat
patches, were calculated for each moose,
for summer and winter separately, and then
compared across seasons using repeated-
measures Analysis of Variance (ANOVA).
Minimum travel distances were calculated
in all cases as straight lines between succes-
sive location points.

RSFs (Manly et al. 2002) were modelled
6 times each using logistic regression from
pooled locations of all individuals: 1) based
on number of locations in each habitat type
within the home range, compared to its area
on the surrounding landscape, for describing
summer habitat use by residents and
migrants using the highlands in a marginal
model; 2) in a similar marginal model for
describing winter habitat use by residents
and migrants using the lowlands; 3) in a con-
ditional model based on number of locations
for each moose in each habitat type within its

Table 2. Habitat descriptions from a highlands classification of Gros Morne National Park, Newfoundland,
Canada.

Habitat type Description Category

Open
softwood
forest

Balsam fir and some black spruce in a closed canopy ranging to <75%
open; dense pockets of krummholz (locally known as tuckamore). Open
heath and fen and bog interspersed.

Closed- to open-
canopy forest

Scrub forest Trees <4 m height. Open heaths, fens, and bogs throughout (>50%
of area).

Open-canopy forest

Shrub Predominantly low shrubs (<1 m height), interspersed with fens, bogs, and
small pockets of scrub forest. Associated with transition from fen and
tundra to scrub forest. Can be wet.

Non-forest

Tundra Low heath vegetation comprised of sedges (Carex spp.), caribou moss
(Cladonia spp.) and crowberries (Empetrum spp.); <20% rock, but few
shrubs or trees. Fairly dry.

Non-forest

Fen Sedge meadows with fens throughout. Non-forest

Rock barren Boulder fields and exposed rock. Very little vegetation. Non-forest

Table 3. Habitat availability in Gros Morne
National Park, Newfoundland, Canada.

Habitat type

Availability on landscape

Area (km2) Percent

Lowlands

Mature softwood forest 69.6 7

Closed spruce forest 27.6 3

Closed mixed forest 65.5 7

Young softwood forest 56.7 6

Open softwood forest 62.5 7

Open mixed forest 43.6 5

Open hardwood forest 39.0 4

Sparse softwood forest 19.3 2

Herb-hardwood forest 20.0 2

Herb forest 13.5 1

Highlands

Open softwood forest 184.7 21

Scrub forest 133.5 15

Shrub 130.0 15

Tundra 130.3 15

Fen 62.6 7

ALCES VOL. 49, 2013 KERCKHOFF ET AL. – HABITAT USE THROUGHOUT GROS MORNE PARK

117



Table 4. First and last dates of collaring, record length, and home range area in summer and winter from fixed-kernel estimates using a 95% isopleth for 12 GPS-
collared moose. This table also shows median seasonal transition dates, and lengths of summer and winter for moose using the two landscapes year-round in
Gros Morne National Park, Newfoundland, Canada. Migrating moose are identified by an asterisk.

ID Landscape First day collared Last day recording Record length (days)

Home range
size (km2) Median seasonal transition dates

Median season
length (days)

Summer Winter Winter to summer Summer to winter Summer Winter

15 Lowlands 25-Jun-97 13-Oct-98 468 11.9 12.1 18-Apr-98 11-Oct-98 173 181

16 Lowlands 25-Jun-97 13-Oct-98 468 13.2 13.1 18-Apr-98 11-Oct-98 173 181

19 Lowlands 25-Jun-97 05-Nov-97 130 2.9 1.8 18-Apr-98 11-Oct-98 173 181

21* Lowlands 25-Jun-97 16-Jan-98 201 — 5.5 18-Apr-98 11-Oct-98 173 181
22* Lowlands 26-Jun-97 18-Jun-98 352 — 12.3 18-Apr-98 11-Oct-98 173 181
25 Lowlands 26-Jun-97 21-Jun-98 355 8.3 12.1 18-Apr-98 11-Oct-98 173 181

26 Lowlands 26-Jun-97 15-Nov-97 139 4.2 2.4 18-Apr-98 11-Oct-98 173 181

17 Highlands 25-Jun-97 13-Oct-98 468 11.3 10.8 30-Apr-98 24-Oct-98 174 180

18 Highlands 25-Jun-97 27-Feb-98 242 6.8 8.7 30-Apr-98 24-Oct-98 174 180

20 Highlands 25-Jun-97 17-Mar-98 262 6.6 8.2 30-Apr-98 24-Oct-98 174 180

21* Highlands 25-Jun-97 16-Jan-98 201 5.7 — 30-Apr-98 24-Oct-98 174 180
22* Highlands 26-Jun-97 18-Jun-98 352 8.0 — 30-Apr-98 24-Oct-98 174 180
23 Highlands 26-Jun-97 13-Oct-98 467 9.2 7.2 30-Apr-98 24-Oct-98 174 180

24 Highlands 26-Jun-97 01-Jun-98 335 7.0 4.6 30-Apr-98 24-Oct-98 174 180

1
1
8

H
A
B
IT
A
T
U
S
E
T
H
R
O
U
G
H
O
U
T
G
R
O
S
M
O
R
N
E
PA

R
K

–
K
E
R
C
K
H
O
F
F
E
T
A
L
.

A
L
C
E
S
V
O
L
.
4
9
,
2
0
1
3



home range, compared to its area on the sur-
rounding landscape, for describing summer
and winter habitat use by residents of the
lowlands; 4) in a similar conditional model
for describing summer and winter habitat
use by residents of the highlands; 5) in a con-
ditional model based on number of locations
for each moose in each habitat type within its
core-use area, compared to its area in the
home range, for describing finer-scale sum-
mer and winter habitat use by residents of
the lowlands; and 6) in a similar conditional
model for describing summer and winter
habitat use within the core-use areas of resi-
dents of the highlands. In the first 2 (mar-
ginal) models, one moose resident on the
highlands was removed because of too few
locations (ID 18, Table 4).

Habitat use by residents in the 2 land-
scapes, habitat use by the 2 migrant moose,
and differences in habitat use between the
winter and summer seasons were statistically
compared in a mixed-effects model with ran-
dom intercepts and coefficients (Gillies et al.
2006). To determine the most parsimonious
regression models, corrected Akaike’s Infor-
mation Criteria (AICc) and model deviance
were compared to a model with random vari-
ables for each individual moose. A com-
pound symmetric structure was assumed,
meaning that covariance among all
responses of an animal was assumed con-
stant (Skrondal and Rabe-Hesketh 2004)
and habitat availability was also assumed
constant over time (Manly et al. 2002).
These assumptions limit the applicability of
the RSFs to the time period studied. Random
intercepts and coefficients for all habitat
types experiencing some use were estimated,
and coefficients significantly > 1 were
defined as selection of a habitat type. Calcu-
lations were all relative to open softwood
forest as a reference habitat type, which
was defined similarly for both landscapes.

RESULTS
Home-range size varied considerably

among individual moose, and there was no
consistent size difference by landscape for
either winter (F1,11 = 0.57, P = 0.58) or sum-
mer (F1,11 = 1.53, P = 0.06; Table 4). There
was also no difference in winter and summer
home-range sizes on either the lowlands
(F1,11 = 0.26, P = 0.88) or the highlands
(F1,11 = 0.33, P = 0.67). The mean distances
travelled during a one-year period were 309
km for residents on the lowlands and 267
km for residents on the highlands. Moose
travelled less in winter than in summer.
Weekly travel distances varied according to
season (F1,11 = 106.35, P < 0.001; Fig. 1).
There was no difference in weekly travel dis-
tances by landscape (F1,11 = 0.75, P = 0.47).
The summer season differed in length
between the 2 landscapes, but only by a
day (Table 4). Summer, defined by moose
travel rates, started and ended close to 2
weeks earlier on the lowlands.

The best-fit marginal RSF models
describing habitat use by residents and
migrants showed consistent selection of

Fig. 1. Weekly distance travelled (km) for
moose in Gros Morne National Park, New-
foundland, Canada.

ALCES VOL. 49, 2013 KERCKHOFF ET AL. – HABITAT USE THROUGHOUT GROS MORNE PARK

119



habitats in summer when they occupied the
highlands together, and variable selection of
habitats in winter when they occupied the
lowlands together (Table 5); however, for
both seasons, models with differences in
habitat use between residents and migrants
were rejected. For the remaining 4 RSFs,
conditional models with residents and
migrants pooled, the best-fit were those
including seasonal differences. In summer,
habitat types used less than expected based

on availability at both the home-range and
core-use scales were herb-hardwood forest
and herb forest (Table 6). Closed spruce for-
est, closed mixed forest, and young softwood
forest were among the top habitat types
selected in summer relative to open soft-
wood forest, but were not selected more
than expected. The pattern was generally
reversed in winter on the lowlands; herb-
hardwood forest and herb forest, along with
open hardwood forest and sparse softwood
forest, were all selected by resident moose
at both the home-range and core-use scales.
At the home-range scale, young softwood
forest and both closed forest types were
also selected in winter. In the RSFs calcu-
lated for moose resident on the highlands,
the fen, tundra, and shrub habitat types
were selected in summer at the home-range
scale, while only the fen and tundra types
were selected at the core-use scale. The pat-
tern was similar in winter, but scrub forest
was also selected at the home-range scale
and shrub, not tundra, was selected at the
core-use scale. The overall trend in summer
was more use of closed-canopy forest types
on the lowlands and more use of non-forest
habitat types on the highlands. Selection of
disturbed forest is a winter phenomenon on
the lowlands of GMNP; the same category
of habitat types is avoided in summer.

Defined by repeated occupation of an
area with travel distances < 50 m apart,
most fine-scale habitat patches on the low-
lands were categorized as disturbed forest
(50/127) or as open-canopy forest (47/127).
There were an additional 22 fine-scale habi-
tat patches identified in young softwood for-
est, while only 8 of the 127 fine-scale habitat
selections on the lowlands were in closed-
canopy forest. There were 13.5 habitat
patches per 100 km2 on the lowlands, but
more on the highlands (18.3/100 km2) where
the majority were in open softwood forest
(70/159). Straight-line distances travelled
between fine-scale habitat patches were

Table 5. Ranking of habitat types, from most to
least selected, for highland residents (N = 5; n =
3,252) and migrants (N = 2; n = 2,919) during
summer on the highlands, and lowland residents
(N = 5; n = 2,013) and migrants (N = 2; n =
1,018) during winter on the lowlands, where
lower-case n refers to total number of locations in
home ranges used to calculate resource selection
functions (RSFs); Gros Morne National Park,
Newfoundland, Canada. Habitat types signifi-
cantly selected (P < 0.05) by at least 4 of 5
residents or both of the migrants are shown in
boldface. The open softwood forest is a reference
habitat (shown in italics).

Residents Migrants

Summer on the highlands

(1) Fen Fen

(2) Tundra Tundra

(3) Shrub Shrub

(4) Open softwood forest Open softwood forest

(5) Scrub forest Scrub forest

Winter on the lowlands

(1) Closed spruce forest Herb-hardwood forest

(2) Herb forest Closed spruce forest

(3) Closed mixed forest Herb forest

(4) Mature softwood
forest

Mature softwood
forest

(5) Young softwood forest Closed mixed forest

(6) Herb-hardwood forest Sparse softwood forest

(7) Sparse softwood forest Open hardwood forest

(8) Open hardwood forest Young softwood forest

(9) Open mixed forest Open mixed forest

(10) Open softwood forest Open softwood forest

120

HABITAT USE THROUGHOUT GROS MORNE PARK – KERCKHOFF ET AL. ALCES VOL. 49, 2013



greater in summer than in winter (F1,120 =
36.28, P = 0.01), a consistent pattern for
moose in both landscapes (F1,120 = 0.08,
P = 0.93); there was no difference in travel
distances between patches by landscape
(F1,120 = 0.01, P = 0.99).

DISCUSSION
Despite variation among individual

moose in habitat selection in the Park’s
more diverse and forested lowlands, as
reported earlier (McLaren et al. 2009), we
are able to show with RSFs that more use
of closed-canopy forest occurs in summer,
likely as a means of heat avoidance. Conver-
sely, selection of disturbed and open-canopy
forest is a winter phenomenon on the

lowlands; the same category of habitat types
is avoided during summer. In the cooler
highlands, selection of non-forest habitat
types may reflect less need to escape heat
in the summer relative to the lowlands, and
perhaps a means to escape insects. In winter,
where moose populations are locally at
higher densities according to both aerial sur-
veys (GMNP, unpublished data) and the fre-
quency of our identified winter habitat
patches, selecting disturbed and open-
canopy forest on the lowlands may be
matched to optimal foraging, while selecting
fen and shrub on the highlands may be
matched to travel through areas where snow
is packed along trails that reduces the energy
cost of locomotion (Telfer and Kelsall 1979).

Table 6. Ranking of habitat types, from most to least selected, within home ranges and core-use areas for
resident moose in Gros Morne National Park, Newfoundland, Canada: 5 moose in lowlands and 4 moose
in highlands; lower-case n refers to total number of locations in home ranges or in core-use areas used to
calculate RSFs. Habitat types selected or avoided (P < 0.05) in proportion to their available area are
shown in boldface. The open softwood forest is a reference habitat (shown in italics).

Summer habitat ranking Winter habitat ranking

Lowlands

Home range (n = 3,765) Core-use area (n = 1,485) Home range (n = 2,013) Core-use area (n = 1,679)

(1) Closed spruce forest Closed mixed forest (1) Herb forest Herb-hardwood forest

(2) Young softwood forest Closed spruce forest (2) Herb-hardwood forest Herb forest

(3) Closed mixed forest Young softwood forest (3) Young softwood forest Sparse softwood forest

(4) Open softwood forest Mature softwood forest (4) Open hardwood forest Open hardwood forest

(5) Mature softwood forest Open softwood forest (5) Closed mixed forest Open mixed forest

(6) Sparse softwood forest Open hardwood forest (6) Closed spruce forest Closed spruce forest

(7) Open hardwood forest Open mixed forest (7) Sparse softwood forest Young softwood forest

(8) Open mixed forest Sparse softwood forest (8) Open mixed forest Closed mixed forest

(9) Herb-hardwood forest Herb-hardwood forest (9) Mature softwood forest Open softwood forest

(10) Herb forest Herb forest (10) Open softwood forest Mature softwood forest

Highlands

Home range (n = 2,954) Core-use area (n = 1,609) Home range (n = 1,914) Core-use area (n = 1,619)

(1) Fen Fen (1) Fen Fen

(2) Tundra Tundra (2) Shrub Shrub

(3) Shrub Shrub (3) Tundra Tundra

(4) Open softwood forest Open softwood forest (4) Scrub forest Scrub forest

(5) Scrub forest Scrub forest (5) Open softwood forest Open softwood forest

ALCES VOL. 49, 2013 KERCKHOFF ET AL. – HABITAT USE THROUGHOUT GROS MORNE PARK

121



In areas where snow is deep, reducing the
energy cost of travel may be more important
than avoiding competition for food.

The most straightforward way of
describing habitat use is in terms of density
(Holt and Fryxell 2011). To approximate
local moose densities, the total area in habi-
tat types selected by moose could be substi-
tuted for an average density over the entire
landscape areas. If habitat types selected dur-
ing winter, based on the core-use areas of
GPS-collared moose occupying the lowlands
in this season, are used to represent the best
habitat types (herb-hardwood, herb, sparse
softwood, and open hardwood forests), win-
ter density would be 20.6 moose/km2, almost
5 x larger than the density estimate across the
lowlands in the March 2007 survey (4.2
moose/km2). If moose remaining on the
highlands in winter similarly used only those
habitat types selected in core-use areas by
the GPS-collared subset (fen and shrub),
their density would be 1.7 moose/km2, about
twice the landscape density estimate (0.9
moose/km2) for the highlands. The lowland
winter habitat types are essentially aban-
doned during summer in favour of habitat
types providing thermal cover (closed-
canopy forest); this change, combined with
20% of the population migrating to the
highlands (McLaren et al. 2000), reduces
effective summer density on the lowlands.
Meanwhile, summer migrants, according
to the 2007 winter lowland population
estimate, should double the corresponding
winter estimate for the highlands, where tun-
dra, roughly equal in area to shrub, is simply
substituted as a preferred habitat in summer.
Presumably, seasonal abundance of forage
is one benefit to spending the summer on
the highlands.

Thus, 2 landscapes in GMNP provide
insight into habitat selection by moose in
Newfoundland and Labrador. We find that
moose adapt seasonally to the Park’s low-
lands and highlands. Moose adopting either

of 2 strategies, year-round residence in one
landscape or migration between landscapes,
do not appear to select habitat differently
when they occupy the same landscape. This
point parallels the consensus for migration
in a review and study in Sweden (Ball et al.
2001) that concluded that snow depth is the
likely driver for moose migration. What dif-
fers in our study is insight into the advan-
tages in summer for the fraction of moose
opting to return to an otherwise less hospita-
ble landscape, that being the snowy high-
lands. If we accept a conclusion from a
Québec study that movement rates for moose
are better indicators of forage availability
than home range size (Dussault et al. 2005),
and that the habitat patches in open softwood
forest on the highlands offer more forage in
summer than that provided on average in
the disturbed or open-canopy forest on the
lowlands, we are describing a situation simi-
lar to what has been described for predator-
free Svaldbard reindeer (Rangifer tarandus
platyrhynchus) (Bremset Hansen et al.
2009). In this case, populations in over-
grazed range move to areas of higher forage
biomass, not higher forage quality. Further,
although plant phenology from spring
through early summer is generally associated
with increasing forage quality (Klein 1990),
the nitrogen content in forage declines initi-
ally after snowmelt (Van der Wal et al.
2000). Thus, migrant moose may travel
upland in GMNP to maximize biomass con-
sumption while tracking delayed plant phe-
nology in the cooler highlands climate.

It is recommended that moose manage-
ment in GMNP consider 2 landscapes (the
lowlands and the highlands) as separate
management units due to differences both
in habitat types they offer and densities of
moose they support. Park management plans
should ensure landscape connectivity for
moose migrating between the highlands and
lowlands. On this note, management across
Newfoundland and Labrador that is both

122

HABITAT USE THROUGHOUT GROS MORNE PARK – KERCKHOFF ET AL. ALCES VOL. 49, 2013



effective and adaptable need not be depen-
dent on defining discrete populations of
moose, but should be in the context of the
2 very different landscapes the Province
offers to moose.

ACKNOWLEDGEMENTS
Dr. A. Rodgers, Centre for Northern For-

est Ecosystem Research, Ontario Ministry of
Natural Resources, and Dr. D. Eastman, Uni-
versity of Victoria, provided advice and con-
structive criticism on portions of this paper
submitted by its lead author as a MSc thesis
in Forestry at Lakehead University. Two
anonymous reviewers and Alces editor Dr.
E. Addison provided helpful comments to
improve this manuscript. S. Taylor, GMNP,
provided invaluable assistance in interpret-
ing habitat types and in analyzing GPS
data. This project was funded and coordi-
nated by the Institute for Biodiversity, Eco-
system Science and Sustainability of the
Government of Newfoundland and Labra-
dor, with additional funding and in-kind sup-
port provided by Parks Canada and
Lakehead University. The original fieldwork
and preliminary data analysis occurred under
the guidance of the Inland Fish and Wildlife
Division of the Government of Newfound-
land and Labrador, with assistance by D.
Anions and C. McCarthy (formerly of
GMNP) and technical advice from Dr. A.
Rodgers.

REFERENCES

BALL, J. P., C. NORDENGREN, and K. WALLIN.
2001. Partial migration by large ungu-
lates: characteristics of seasonal moose
Alces alces ranges in northern Sweden.
Wildlife Biology 7: 39–47.

BANFIELD, C. E. 1983. Climate. Pages 37–
106 in G. R. South, editor. Biogeography
and Ecology of the Island of Newfound-
land. W. Junk Publishers, Boston, Massa-
chusetts, USA.

BERGERUD, A. T., H. E. BUTLER, and D. R.
MILLER. 1984. Anti-predator tactics of

calving caribou: dispersion in mountains.
Canadian Journal of Zoology 62: 1566–
1575.

BREMSET HANSEN, B., I. HERFINDAL, R. AANES,
B.-E. SÆTHER, and S. HENRIKSEN. 2009.
Functional response in habitat selection
and the trade-offs between foraging niche
components in a large herbivore. Oikos
118: 859–872.

CAINES, P., and H. DEICHMANN. 1989.
Resource Description and Analysis:
Gros Morne National Park. Gros Morne
National Park, Parks Canada, Rocky Har-
bour, Newfoundland, Canada.

CONNOR, K. J., W. B. BALLARD, T. DILWORTH,
S. MAHONEY, and D. ANIONS. 2000.
Changes in structure of a boreal forest
community following intense herbivory
by moose. Alces 36: 111–132.

DAMMAN, A. W. H. 1983. An ecological
subdivision of the island of Newfound-
land. Pages 163–205 in G. R. South,
editor. Biogeography and Ecology of
the Island of Newfoundland. W. Junk
Publishers, Boston, Massachusetts, USA.

DUSSAULT, C., R. COURTOIS, J.-P. OUELLET, and
I. GIRARD. 2005. Space use of moose in
relation to food availability. Canadian
Journal of Zoology 83: 1431–1437.

FRETWELL, S. D., and H. L. LUCAS. 1970. On
territorial behaviour and other factors
influencing habitat distribution in birds.
Acta Biotheoretica 19: 16–36.

FRYXELL, J. M. 1991. Forage quality and
aggregation by large herbivores. Ameri-
can Naturalist 138: 478–498.

GILLIES, C. S., M. HEBBLEWHITE, S. E. NIELSEN,
M. A. KRAWCHUK, C. L. ALDRIDGE, J. L.
FRAIR, D. J. SAHER, C. E. STEVENS, and
C. L. JERDE. 2006. Application of random
effects to the study of resource selection
by animals. Journal of Animal Ecology
75: 887–898.

GOSSE, J., L HERMANUTZ, B. MCLAREN, P.
DEERING, and T. KNIGHT. 2011. Degrada-
tion of boreal forests by non-native
herbivores in Newfoundland’s National
Parks: recommendations for ecosystem

ALCES VOL. 49, 2013 KERCKHOFF ET AL. – HABITAT USE THROUGHOUT GROS MORNE PARK

123



restoration. Natural Areas Journal 31:
331–339.

HARE, F. K. 1952. The climate of the island of
Newfoundland: a geographical analysis.
Geographic Bulletin 2: 36–88.

HEBBLEWHITE, M., and E. H. MERRILL. 2007.
Multiscale wolf predation risk for elk:
does migration reduce risk? Oecologia
152: 377–387.

———, ———, and G. MCDERMID. 2008.
A multi-scale test of the forage matura-
tion hypothesis in a partially migratory
ungulate population. Ecological Mono-
graphs 78: 141–166.

HOLT, R., and J. FRYXELL. 2011. Theoretical
reflections on the evolution of migration.
Pages 17–31 in E. Milner-Gulland, J.
Fryxell, and A. Sinclair, editors. Animal
Migration: A Synthesis. Oxford Univer-
sity Press, Oxford, United Kingdom.

HUMBER, J. M., and L. HERMANUTZ. 2011.
Impacts of non-native plant and animal
invaders on gap regeneration in a pro-
tected boreal forest. Biological Invasions
13: 2361–2377.

KLEIN, D. R. 1990. Variation in quality of car-
ibou and reindeer forage plants asso-
ciated with season, plant part, and
phenology. Rangifer Special Issue 3:
123–130.

MANLY, B. F. J., L. L. McDONALD, D. L.
THOMAS, T. L. McDONALD, and W. P.
ERICKSON. 2002. Resource Selection
by Animals, Second Edition. Kluwer
Academic Publishers, Dordrecht,
Netherlands.

MARTIN, C. 2004. Climatology and historical
snowcover of the Big Level Plateau, Gros
Morne National Park, Newfoundland.
MSc Thesis, Memorial University of
Newfoundland, St. John’s, Newfound-
land, Canada.

MCLAREN, B., C. MCCARTHY, and S. MAHO-
NEY. 2000. Extreme moose migrations in
Gros Morne National Park, Newfound-
land. Alces 36: 217–232.

———, B. A. ROBERTS, N. DJAN-CHÉKAR, and
K. P. LEWIS. 2004. Effects of

overabundant moose on the Newfound-
land landscape. Alces 40: 45–59.

———, S. TAYLOR, and S. H. LUKE. 2009.
How moose select forested habitat in
Gros Morne National Park, Newfound-
land. Alces 45: 125–135.

MCNAUGHTON, S. J. 1985. Ecology of a graz-
ing ecosystem: the Serengeti. Ecological
Monographs 55: 259–294.

MOEN, R., J. PASTOR, and Y. COHEN. 1997.
Accuracy of GPS telemetry collar loca-
tions with differential correction. Journal
of Wildlife Management 61: 530–539.

PIMLOTT, D. H. 1953. Newfoundland moose.
Transactions of the North American
Wildlife Conference 18: 563–581.

RODGERS, A. R., A. P. CARR, L. B.
HAWTHORNE, L. SMITH, and J. G. KIEM.
2007. Home Range Tools for ArcGIS.
Version 1.1. Ontario Ministry of Natural
Resources, Centre for Northern Forest
Ecosystem Research, Thunder Bay,
Ontario, Canada.

SKRONDAL, A., and S. RABE-HESKETH.
2004. Generalized Latent Variable Mod-
eling: Multilevel, Longitudinal, and
Structural Equation Models. Chapman
and Hall, New York, New York, USA.

TAYLOR, S., and R. SHARMA. 2010. Asses-
sing the Status of Forest Recovery in Dis-
turbance Affected Areas Using SPOT 5
Imagery in Gros Morne National Park.
Gros Morne National Park, Parks
Canada, Rocky Harbour, Newfoundland,
Canada.

TELFER, E. S., and J. P. KELSALL. 1979. Studies
of morphological parameters affecting
ungulate locomotion in snow. Canadian
Journal of Zoology 57: 2153–2159.

VANDER WAL, E., and A. R. RODGERS. 2009.
Designating seasonality using rate of
movement. Journal of Wildlife Manage-
ment 73: 1189–1196.

VAN DER WAL, R., N. MADAN, S. V. LIESHOUT,
C. DORMANN, R. LANGVATN, and S. D.
ALBON. 2000. Trading forage quality for
quantity? Plant phenology and patch

124

HABITAT USE THROUGHOUT GROS MORNE PARK – KERCKHOFF ET AL. ALCES VOL. 49, 2013



choice by Svalbard reindeer. Oecologia
123: 108–115.

WATSON, W. B. 1974. The Climate of Gros
Morne National Park, Newfoundland.
Atmospheric Environmental Service,

Department of Environment, St. John’s,
Newfoundland, Canada.

WORTON, B. J. 1989. Kernel methods for esti-
mating the utilization distribution in home
range studies. Ecology 70: 164–168.

ALCES VOL. 49, 2013 KERCKHOFF ET AL. – HABITAT USE THROUGHOUT GROS MORNE PARK

125


	MOOSE HABITAT USE THROUGHOUT GROS MORNE NATIONAL PARK
	STUDY AREA
	METHODS
	Habitat Classification
	Moose Locations
	Data Analysis

	RESULTS
	DISCUSSION
	ACKNOWLEDGEMENTS
	REFERENCES