Content uploaded by Jeff Keay
Author content
All content in this area was uploaded by Jeff Keay on Dec 29, 2014
Content may be subject to copyright.
U. S. Department of the Interior
U. S. Geological Survey
Alaska Biological Science Center
Anchorage, Alaska
GRIZZLY BEAR POPULATION ECOLOGY
AND MONITORING
DENALI NATIONAL PARK AND PRESERVE, ALASKA
Jeffrey A. Keay, Ph.D.
December 2001
GRIZZLY BEAR
POPULATION ECOLOGY AND MONITORING
DENALI NATIONAL PARK AND PRESERVE, ALASKA
Report of Project Development and Findings
2001
Jeffrey A. Keay, Ph.D.
Research Wildlife Biologist
Alaska Biological Science Center
U.S. Geological Survey
1
GRIZZLY BEAR POPULATION ECOLOGY AND MONITORING
DENALI NATIONAL PARK AND PRESERVE, ALASKA
Report of Project Development and Findings
December 2001
Abstract: The dynamics and ecology of a
naturally regulated grizzly bear population
were studied in Denali National Park and
Preserve, Alaska, from 1991 through 1998.
Bears were captured by helicopter darting and
fitted with radio-transmitters containing
mortality sensors. Bears were located at least
monthly during the non-denning period.
Density was determined by capture-mark-
resight procedures. Serum antibody
prevalence for infectious canine hepatitis,
canine distemper, and leptospirosis were low
indicating disease was not an important factor
affecting population dynamics. Grizzly bear
density, during fall 1995, was 27.1
independent bears/1000 km2 (95% C.I. =
25.1 - 30.2) with 61% of the population
marked. Adjusting survey area to reflect
forage-producing habitat for more direct
comparison with adjacent studies indicated
34.7 bears/1000 km2 (95% C.I. = 32.2 to
38.7). Sight-ability of both females and
males was high during the fall survey (60.7
% and 86.2%, respectively). The male age
structure was unimodal, was 10% subadults,
and included adults up to 16 years of age.
The female age structure was strongly
bimodal and 20% were subadults. The
female segment experienced a shift in age
distribution between 1991 and 1997; each
mode advanced 6 years but remained intact.
There were 14% subadults in the 1991 female
population and 3% in the 1997 population.
Cubs-of-the-year experienced the lowest
annual survival rates of all age classes at
0.341. Yearling survival was 0.455 and that
of dependent 2-year-olds was 0.785.
Subadult survival was 1.000 for females and
0.943 for males. Adult survival rates were
0.970 for females and 0.983 for males.
Average age of first reproduction was 7.6
years and varied from 6 to 10. Litter size of
cubs-of-the-year at the time of den emergence
averaged 2.1. Dependent bears separated
from their mothers at an average age of 2.9
years. Reproductive rates were 0.091 for 6-
year-olds, 0.307 for bears 7 to 9 years old,
0.384 for bears 10 to 26, and 0.071 for bears
27 to 30 years of age. Bears averaged 23% to
27% body fat prior to den entrance in
September. Females and subadult males had
less than 10% body fat following den
emergence in May. Adult males were 15%
fat during May. Females lost 30% to 35% of
total body mass and 18% to 20% of lean body
mass during winter hibernation. High
survival rates of independent bears and lack
of human interference during the past 80
years suggests the Denali population is likely
at carrying capacity. High dependent bear
mortality coupled with periodic recruitment
may allow the population to oscillate mildly
but, in the long term, is likely stable.
Implications for long-term monitoring are
discussed.
a Current address: U.S. Geological Survey, Eastern Re-
gional Office, 1700 Leetown Road, Kearneysville, WV
25430
Jeffrey A. Keay, U. S. Geological Survey,
Alaska Biological Science Center, P.O. Box 9,
Denali Park, AK 99755 a
Keay—Grizzly Bear Population Ecology and Monitoring
Grizzly bears (Ursus arctos) are an
important ecological resource and an integral
part of the naturally functioning predator-prey
system of Denali National Park and Preserve,
Alaska (Murie, 1981, Adams et al. 1995).
Visitors also consider them an important part
of their Denali experience, with 88% of the
visitors that ride one of the bus services into
the park observing grizzly bears in their
natural habitat (Miller and Wright 1998).
Human developments outside Denali National
Park, proposed alternative access routes into
the Park, and concerns regarding the impacts
of human activities within the park resulted in
the need for objective information regarding
the status and trend of grizzly bears in Denali
National Park and Preserve and for develop-
ment of cost-effective, non-invasive methods
to monitor the grizzly bear population.
The objectives of this paper are to 1)
document the development and direction of
grizzly bear research in Denali during the
period 1991 to 1998, 2) present findings on
grizzly bear population status and trend, and
3) describe efforts to develop noninvasive
monitoring techniques.
STUDY OBJECTIVES
Study objectives varied during the 8 years
of grizzly bear research in Denali. This work
originated in 1991 as a small study to exam-
ine the role of grizzly bears as predators of
caribou calves. Park staff redirected the study
(August 1992) to develop non-invasive
methods to monitor the grizzly bear popula-
tion. Recent research (Samuel et al. 1987)
had developed a technique for measuring
visibility bias associated with aerial surveys
that we felt could be applied in the open
habitats of the Park. Consequently, during
1993 through 1995 our objectives were to 1)
describe the characteristics of the existing
population and 2) develop field methods for
measuring visibility bias. Once developed,
the non-invasive survey techniques would be
tested on the known population.
As opportunity permitted, we also began
to collect data on factors that affected popula-
tion dynamics in preparation for a second
phase of broader, population ecology re-
search. The objective was to identify factors
that significantly affect grizzly bear popula-
tion dynamics for possible inclusion in the
Park’s long term monitoring program. Un-
usually high grizzly bear cub mortality was
noted early in the study. In 1994, we began
measuring percent body fat using bioelectri-
cal impedance analysis (Farley and Robbins
1994) to assess female physical condition.
The goal was to see if cub mortality was
related to female physical condition and,
hence, nutritionally affected. We simultane-
ously initiated cooperative studies with Dr.
Roseann Densmore and Dave Douglas, both
of the Alaska Biological Science Center, to
measure the abundance of important berry
foods for bears and to assess the potential of
Advanced Very High Resolution Radiometry
(AVHRR) satellite data to predict berry
abundance through correlation with plant
phenology, biomass abundance, and standard
weather parameters. Those studies will be
reported separately by Densmore and Doug-
las.
By the end of 1995 we had good estimates
of population density and sex and age struc-
ture. However, our final efforts to develop
field methods to measure visibility bias were
unsatisfactory. Dr. Pham X. Quang
(mathematician, University of Alaska, Fair-
banks) recommended using a binomial esti-
mator that as yet was untested and not previ-
ously reported in the biological literature. In
1996, we altered objective 2) to conduct
repeated aerial surveys, on the known density
population, to obtain point and variance
estimates to test the precision of the binomial
estimator. Inclement weather precluded
obtaining those data. Since we had com-
pleted a portion of objective 1) we also
Keay—Grizzly Bear Population Ecology and Monitoring 2
altered capture objectives to focus on females
to better measure cub production and survival.
Collars were maintained on all previously
radioed males and females to improve meas-
ures of survival for all age classes. However,
we did not attempt to replace males whose
transmitters were removed prematurely.
In late 1996 Park staff requested that we
postpone final efforts to develop techniques to
enumerate bears and begin more aggressively
studying nutritional ecology. A study plan was
developed in 1997 in cooperation with Dr.
Charles Robbins, Washington State University,
to address the new objectives. That study plan
was rejected by the Park. Consequently,
during 1997 we continued to 1) improve
measures of reproductive and survival rates,
and 2) assess female physical condition, using
bioelectrical impedance analysis, to see if it
could serve as a useful tool in elucidating the
role of nutrition in rates of cub production and
survival. Objectives during 1998 were limited
to monitoring cub production and survival.
STUDY AREA
The McKinley Slope Study Area encom-
passed 1,750 km2 along the north slope of the
central Alaska Range in Denali National Park
and Preserve, Alaska (formerly Mount
McKinley National Park, Fig 1). The climate
is generally cool and wet during the summer
with temperatures typically 10 C to 15 C.
Freezing temperatures and snow may occur
during any month. Snow accumulation usually
begins in October and dissipates from lowlands
and unshaded portions of foothills by mid-
May.
The area was delineated to include all
important foraging habitats, including large,
concentrated, berry patches on glacial mo-
raines. Elevations ranged from 600 m to 2,000
m; high elevations above the vegetation line
(>1500 m) were included as part of the study
area because bears often denned there.
The study area included the principal
calving grounds of the Denali Caribou Herd
(Rangifer terandus, Adams et al. 1995). The
Keay—Grizzly Bear Population Ecology and Monitoring 3
Figure 1. McKinley Slope grizzly bear study area, Denali National Park and Preserve, Alaska,
1991 through 1998.
Herd exceeded 20,000 animals in the early
1940's (Murie 1944), declined to around 5,000
by 1968, and varied from 1,000 to 3,000 during
the period 1970 to 1998 (Haber 1977, Adams
et al. 1989, Layne G. Adams, USGS, Biologi-
cal Resources Division, Alaska Biol. Sci.
Center, pers. commun.). Moose (Alces alces)
densities in the study area were low and aver-
aged 60 per 1000 km2, below 1050m, during
March 1992 (L. G. Adams, unpubl. data.).
There are no records or other indications of
bear poaching within the study area. Meat
hunting and trapping, which may have in-
cluded limited bear harvest, did occur prior to
establishment of the Park in 1917. Park-wide,
back-country visitor use increased from 2,200
user-nights in 1971, to 10,000 in 1978. It
averaged 11,400 from 1979 through 1990, and
15,500 during 1991-97. Within the study area,
back-country use averaged 2,204 user-nights
per year during this study. Most visitor use
(63%) was concentrated on the eastern 15% of
the study area (east of the Muldrow Glacier
and north of the McKinley River). Overnight
visitor densities averaged 8.2/1000 km2 during
the study but varied from 0.2 to 43.4 for
various back-country units. Grizzly bears in
the study area obtained human foods in no
more than 10 reported human encounters in the
past 18 years. There have been no manage-
ment kills or translocations of nuisance bears
to or from the study area. Consequently,
human activities have had virtually no impact
on grizzly bear population dynamics in the
study area for at least 80 years.
METHODS
Population Ecology
Grizzly bears were captured by helicopter
darting (Taylor et al. 1989). All captured bears
were fitted with radio transmitters containing
mortality sensors. Captures occurred annually
during May 1991 through 1998, and September
1993 through 1997. Dependent bears were
radio-collared at 2 years of age. We deter-
mined the ages of captured bears by counting
cementum annuli (Matson et al. 1993) from a
vestigial first premolar. The premolar was
extracted at each bear’s first capture to mini-
mize age-related bias (Keay 1995). Age
assignments were made by Matson's Laborato-
ries, Milltown, Montana but were adjusted in 1
case when the body size of a dependent bear
indicated otherwise (Keay 1995). Bears were
classified as adults if ≥ 6 and subadults
(reproductively immature) when independent
but ≤5 years old.
During 1991-93, bears were located weekly
from fixed-wing aircraft from the time of den
emergence until den entrance; during 1994-98,
weekly locations were limited to den emer-
gence through June and September until den
entrance. Cub production was determined by
the number of young observed with a female at
the time of den emergence. Cubs-of-the-year
(COY) and yearlings were assumed dead if
radio-marked females were later observed with
smaller litters.
Survival of radio-collared bears was calcu-
lated with a modified Kaplan-Meier method
(Pollock et al. 1989). Equation 2 (Pollock et
al. 1989) was used to calculate variance be-
cause it is less sensitive to animals added or
censored. Survival estimates for the 8-year
study were converted to the parameter of
interest, annual survival, by taking the 8th root
of the point and interval estimates (Amstrup
and Durner 1995). Bears were censored when
transmitters were removed or stopped function-
ing. The finite rate of increase, or growth
multiplier, was calculated using the Lotka
equation (Caughley 1974:110). Age of first
reproduction was calculated following proce-
dures outlined by Garshelis et al. (1998).
Reproductive rates (females produced per
reproductive female) were calculated as one
half of the number of cubs-of-the-year ob-
served at den emergence divided by the num-
ber of adult females monitored.
Habitat availability was determined by
Keay—Grizzly Bear Population Ecology and Monitoring 4
examining 199 randomly selected points in the
study area on 31 August 1994. The study area
boundary was digitized at a scale of 1 inch per
mile and entered into an ArcInfo database. A
digitized 1:250,000 elevation model, interpo-
lated from 200 foot contours to meters, with 80
m horizontal and 1 meter vertical resolution
(interpolated), was used to identify the 915 m
(3,000 feet) contour. Horizontal spacing
among candidate sample points was set at 100
m. Random points were selected by ArcInfo
from each strata, above or below 915 m eleva-
tion, in proportion to the area of that strata
(51% above 915 m and 49% below or equal to
915 m elevation). Each random point was
located using a commercial global positioning
system (GPS) mounted in a Piper supercub
airplane. Vegetation type (Vierek et al. 1992,
level 3) was recorded for a projected 30 m
radius around the random point. (If candidate
sample points were too far apart, some ground
locations would have no probability of being
sampled. If candidate sample points were too
close together, then over-lapping radii of
sample units would give some ground loca-
tions a higher likelihood of being sampled than
others. Error associated with locating a ran-
dom point on the ground using commercial
global positioning systems effectively in-
creases the sample unit radius with regards to
its influence on the probability of sampling a
specific ground location. To insure that all
ground locations had an equal probability of
being sampled, the horizontal spacing between
random points was based on a combination of
sample unit size plus anticipated random GPS
error.)
Serum was extracted from blood samples
collected during normal capture operations and
stored at -10 C. Sera were analyzed at the
Washington Animal Disease Diagnostic Labo-
ratory (WADDL) for prevalence of infectious
canine hepatitis, leptospirosis, and canine
distemper. Dr. James Everman (DVM at
WADDL) provided professional guidance and
directed the processing of samples. The
standard threshold titer of 1:5 was considered
evidence of previous exposure and referred to
hereafter as “positive.” All lower titers were
referred to as “negative.”
Brown bear density was determined from
capture-mark-resight techniques (Miller et al.
1997) on a 1,707 ha portion of the study area
during September 1995. Half of the study area
was surveyed each day for six days providing 3
replicates. Density was calculated using
maximum likelihood estimation (White 1996).
An adjusted density based on forage producing
habitat was estimated by subtracting the area of
extensive, unvegetated habitats above 1,500 m
(267 ha) and extensive glaciers and barren
river bars (106 ha) that were surveyed during
the density estimate. Bears were captured for 4
years prior to conducting the density estimate
to assure a representative sample of all sex,
age, and reproductive classes.
Sex and age structure were determined by
the number of radio-collared bears detected
within the study area while conducting cap-
ture-mark-resight density estimates (Miller et
al. 1997). Sex and age data were weighted by
the number of times each bear was in the study
area during independent surveys conducted on
different days, in order to minimize the bias of
over-emphasizing animals with large home
ranges. Female age structure in Denali during
1991 and 1997 was determined by the number
of marked females known to be alive during
those years; comparable data for males was not
available.
Analysis of variance and t-tests were used
to identify differences in body fat and weight
between sex, age, and reproductive classes of
bears (α = 0.05).
Population Monitoring
Aerial Surveys. The use of radio collars to
enumerate bears has become an established
practice because 1) bear densities are low, 2)
certain estimation procedures become available
Keay—Grizzly Bear Population Ecology and Monitoring 5
when a known proportion of a population is
observed, and 3) to at least partially compen-
sate for differences in habitat use patterns of
various sex, age, and reproductive classes of
bears both within and between surveys, which
affects the likelihood of their detection. Aerial
surveys without the use of radio transmitters on
bears would have to provide some adjustment
for the visibility bias associated with differ-
ences in habitat use patterns of bears. Increas-
ing survey intensity was used to measure
visibility bias for moose when there was
complete snow cover (Gasaway et al. 1986)
but would not have adequate success for bears
on snow-free ground. Double sampling has
been used for sheep and is being tested for
bears but will likely only work in high density
populations (E. Becker, Alaska Dept. Fish and
Game, pers. comm.).
Visibility bias associated with group size,
behavior, and habitat, was successfully meas-
ured for elk using a second aircraft to monitor
radioed animals that were missed by the survey
aircraft and to determine the sighting condi-
tions associated with missed observations
(Samuel et al. 1987). Comparing sighting
characteristics for animals missed to those for
animals seen provided a means to estimate the
proportion of animals observed under differing
sighting conditions. Those proportions were
then used to adjust results of aerial surveys.
This procedure is widely used for elk in north-
ern Idaho. Measuring visibility bias appeared
to be the most plausible technique for bear
populations at the low density and in the open
habitats of Denali National Park. Three ap-
proaches were tested for measuring visibility
bias which varied the way in which monitoring
aircraft determined sighting conditions for
animals missed during survey flights. Monitor
aircraft 1) followed survey aircraft, 2) preceded
and followed survey aircraft, and 3) conducted
simultaneous observation through high altitude
surveillance. Method development occurred in
conjunction with radio-tracking or density
estimation efforts during May 1993 and Sep-
tember 1994 and 1995.
The ability of biologists and pilots to
properly classify bears by sex and age category
was tested during the 1995 density estimate.
Survey aircraft were not informed of the sex,
age, or reproductive status of marked bears.
They were asked to record the radio frequency
of marked bears and their best estimate of the
bears sex, age, age class, and the number and
age of dependent young, which were then
compared to known values.
Thermal Infrared Imaging. Thermal
infrared videography showed promise for
aerial surveys of white-tailed deer (Wiggers
and Beckerman 1993) and was investigated as
a potential population monitoring technique for
bears. Preliminary surveys of grizzly bears
outside the Park, with thermal infrared video-
graphy, provided inconsistent results and left
many questions regarding the potential useful-
ness of the technique (Miller and McDonald
1996). A study proposal was developed with
Dr. Ernie Wiggers, University of Missouri, to
elucidate problems encountered in those
previous efforts and provide a more scientifi-
cally rigorous test of the technique than was
conducted by Miller and McDonald (1996).
Thermal infrared temperatures were obtained
from sedated bears and their surrounding
environment using an Omega OS85-EM
infrared thermometer (Omega engineering,
Inc., Stanford, CT 06907). This thermometer
measures surface temperature in the 8-14
micron wavelength, similar to aerial infrared
scanners. Bear surface temperatures were
measured at six locations with the thermometer
held 50 cm from and perpendicular to the
surface of interest: forehead, spine between
shoulders, spine at approximately 10 cm above
base of tail, side at mid-ribcage, one foreleg
just above the elbow, and one hind leg just
above the knee. At each temperature location
the degree of moisture (dry, moist, wet), and
Keay—Grizzly Bear Population Ecology and Monitoring 6
whether the measurement was taken in high,
direct sunlight, low-angle sunlight, or shade
was recorded. Coat color and rectal tempera-
ture were recorded. Guard hair length was
measured along the spine between the shoul-
ders as an indicator of the amount of insulation
between the skin and measured surface of the
bear. Background temperature (vegetation,
soil, etc.) was measured at each of the 4 cardi-
nal points 4 m away from the bear’s position.
The vegetation type (Vierek et al. 1992, level
3), whether it was in sunlight or shade, and its
phenology was recorded. Time of day, ambi-
ent air temperature, cloud cover, and precipita-
tion condition were recorded for each bear
measured.
RESULTS
Population Ecology
Habitat Availability. Low shrub habitats
were the predominant vegetation types and
occurred on 31.7% of the study area (Table 1).
Tall shrubs occupied 14%, herbaceous 16.1%,
and forest habitats covered 12.1% of the study
area. Sites devoid of vegetation occurred at
26.1% of the random points visited and were
primarily above 1,500 m. The most abundant
terrain features were flat terrain and gentle
slopes (< 30 degrees) at 33.7% and 24.1%,
respectively (Table 2). Steep slopes (30 – 60
degrees) and rolling terrain each occupied
16.6% of the study area. River bars occurred
on 9% of the random sites visited.
Disease. Serum antibody prevalence of
infectious canine hepatitis was 6% (n = 17) for
adult females (positive titer 1:5) and 0% (n =
5) for cubs-of-the-year captured in September
1994. Prevalence was 17% (n = 12) for sera
collected from adult females during May, 1991
and 1992 (positive titers 1:10 and 1:40). One
female tested positive in both 1992 and 1994.
Six females that tested negative in 1991-92
were also negative in 1994. Prevalence of
canine distemper was 8% (n = 12) for adult
females in 1991-92 (positive titer 1:10). There
was no evidence of exposure to leptospirosis
for adult females in 1991 and 1992 (n = 12).
Density. Grizzly bear density was 27.1
independent bears/1000 km2 (95% C.I. = 25.1 -
Keay—Grizzly Bear Population Ecology and Monitoring 7
Table 1. Percentage of each vegetation classification (Vierek et al. 1992, level 3) observed at
random points within the McKinley Slope grizzly bear study area, Denali National Park and Pre-
serve, Alaska, August 1994.
Vegetation Class > 3,000 ft. ≤ 3,000 ft Combined # Random Points
Forest, broadleaf, closed 0.0 2.1 1.0 2
Forest, conifer, closed 0.0 1.0 0.5 1
Forest, conifer, open 0.0 21.6 10.6 21
Scrub, low, closed 3.9 19.6 11.6 23
Scrub, low, open 13.7 26.8 20.1 40
Scrub, tall, closed 3.9 7.2 5.5 11
Scrub, tall, open 3.9 13.4 8.5 17
Herbaceous 28.5 3.1 16.1 32
Barren 46.1 5.2 26.1 52
TOTAL 100.0 100.0 100.0
# Random Points 102 97 199 199
30.2, Table 3) with 61% of the population
marked. Sight-ability of both females and
males was high during the fall survey (60.7 %
and 86.2%, respectively). The bounds on the
density estimate averaged 9% of the mean after
3 repetitions. The survey area included large
glaciers, extensive unvegetated river bars, and
substantially more unvegetated habitat above
1,500 m elevation than other studies in the
Alaska Range (Miller et al. 1997). Reducing
study area size to reflect forage-producing
habitat resulted in an adjusted density of 34.7
bears/1000 km2 (95% C.I. = 32.2 to 38.7).
Sex and Age Structure. In 1995, the Denali
subadult population was 21% males while 36%
of adults were male. The male age structure
was unimodal (Fig. 2), was 10% subadults, and
included adults up to 16 years of age. The
average age of all independent male bears was
8.9 years and the average age of adults was 9.5
years. The female age structure was strongly
bimodal (Fig. 2) and 20% were subadults. The
female segment experienced a shift in age
distribution between 1991 and 1997 (Fig. 3);
each mode advanced 6 years but remained
intact. There were 14% subadults in the 1991
female population and 3% in the 1997 popula-
tion.
Survival Rates. Cubs-of-the-year experi-
enced the lowest annual survival rates at 0.341
(95% C.I. = 0.242 to 0.440, n = 37 bear-years,
88 bears). Yearling survival was 0.455 (95%
C.I. = 0.284 to 0.627, n = 21.4 bear-years, 37
bears) and that of dependent 2-year-olds was
Table 3. Grizzly bear density on the McKinley Slope grizzly bear study area, Denali National Park
and Preserve, Alaska, September 1995.
Replication Dates Marked
Bears
Present
Marked
Bears Seen
Total
Bears
Seen
Marked Bears
in Area at
Least Once
Sightability
1 Sep 23, 24 29 18 33 30 62.1%
2 Sep 25, 27 28 20 34 30 71.4%
3 Sep 28, 30 28 21 30 30 75.0%
TOTAL 85 59 97 69.4%
Keay—Grizzly Bear Population Ecology and Monitoring 8
Table 2. Percentage of each terrain feature observed at random points within the McKinley
Slope grizzly bear study area, Denali National Park and Preserve, Alaska, August 1994.
Terrain Feature > 3,000 ft ≤ 3,000 ft Combined # Random Points
River bar 4.9 13.4 9.0 18
Flat 14.7 53.6 33.7 67
Rolling terrain 11.8 21.7 16.6 33
Gentle slope (< 30 degrees) 37.2 10.3 24.1 48
Steep slope (30-60 degrees) 31.4 1.0 16.6 33
TOTAL 100.0 100.0 100.0
# Random Points 102 97 199 199
0.785 (95% C.I. = 0.598 to 0.972, n = 15.2
bear-years, 22 bears). No mortality was
observed among 13 dependent 3-year-olds
observed for 5.3 bear-years. Subadult female
survival was 1.000 (n = 21.3 bear-years, 16
bears) and that of subadult males was 0.943
(95% C.I. = 0.821 to 1.005, n = 28.7 bear-
years, 20 bears). Adult survival rates were
0.970 (95% C.I. = 0.930 to 1.000, n = 125.2
bear-years, 29 bears) for females and 0.983
(95% C.I. = 0.947 to 1.013, n = 58.6 bear-
years, 21 bears) for males.
Two males died of natural causes; one 5-
year-old was killed by an adult male grizzly
bear and a 4-year-old died in mid-summer but
the specific cause was undetermined. One
adult male, in poor physical condition, was
killed in defense of life and property. Three
adult females died of unconfirmed natural
causes. A 19-year-old had 3 cubs-of-the-year
at the time of death, was consumed by another
bear, and may have died defending them. An
a)
b)
Figure 2. Age structure of female (a) and male (b) grizzly bears on the McKinley Slope study
area, Denali National Park and Preserve, Alaska, 1995.
0
1
2
3
4
2 4 6 8 101214161820222426
Age
Number of Bears
0
1
2
3
4
2
4
6
8
10
12
14
16
18
20
22
24
26
Age
Number of Bears
Keay—Grizzly Bear Population Ecology and Monitoring 9
18-year-old was also consumed by another
bear; it was solitary and her bone marrow was
73.9% fat, suggesting poor physical condition.
The specific cause of one 20-year-old female
was undetermined. The cause of death of one
cub-of-the-year was determined to be a
crushed skull when the animal was caught in a
snow avalanche. It’s sibling disappeared
simultaneously and may have been buried in
the avalanche. Three yearlings died in the den
of starvation and were consumed by their
mother at the time of den emergence. Remains
of 3 dependent 2-year-olds, all females,
showed they had been consumed by bears.
The bone marrow from one of them was red
and gelatinous indicating the animal was in
very poor physical condition at the time of
death. Marrow from one of the others was
90.8% fat. Four uncollared males, not in-
cluded in the above calculations, dispersed
from the study area and were legally killed by
sport hunters outside the park.
a)
b)
Figure 3. Age structure of female grizzly bears during 1991 (a) and 1997 (b) on the McKinley
Slope study area, Denali National Park and Preserve, Alaska.
0
1
2
3
4
5
2
4
6
8
10
12
14
16
18
20
22
24
26
28
Age
Number of Bears
0
1
2
3
4
5
2
4
6
8
10
12
14
16
18
20
22
24
26
28
Age
Number of Bears
Keay—Grizzly Bear Population Ecology and Monitoring 10
Reproduction. Average age of first repro-
duction was 7.6 years (n = 69 bear-years and
13 bears) and varied from 6 to 10. Two of the
13 females monitored had not produced first
litters by 1998. Only 2 females produced first
litters that survived to family break-up. Two
other young females were accompanied by
surviving COY through 1998. Assuming all
remaining young adult females produce surviv-
ing litters in 1999, the average minimum age of
females producing surviving
litters would be 10.3 years, but
will most likely be higher.
Litter size of cubs-of-the-
year at the time of den emer-
gence averaged 2.1 (n = 42
litters). Annual litter size varied
from 1.9 to 3.0 cubs. Litter size
of yearlings at den emergence
averaged 1.7 (n = 20 litters);
annual yearling litter size varied
from 1.0 to 2.0. The litter size
for 2-year-olds averaged 1.5 (n
= 15 litters) and varied annually
from 1.0 to 2.0. Dependent
bears separated from their
mothers at an average age of 2.9
years (n = 19, SD = 0.57).
Actual age varied from 2 to 4
years. Reproductive rates were
0.091 for 6-year-olds, 0.306 for
bears 7 to 9 years old, 0.384 for
bears 10 to 26, and 0.071 for
bears 27 to 30 years of age
(Table 4).
Population Growth Rate.
Three breeding-age females,
aged 18 to 20 years, died during
the 8-year study. All 3 were
reproductively active. Two
other females lived to exceed
the years of reproductive contri-
bution. Nine bears were ob-
served as subadults and subse-
quently recruited into the breeding population.
Two females, first observed as adults, were
known to have produced their first litters
during the study. The loss of 5 breeding
females and gain of 11 suggests an increasing
population. In contrast, observed survival and
reproductive rates yield a finite rate of increase
of 0.977, that of a population declining at
about 2% per year.
Table 4. Age specific reproductive rates of radio-collared fe-
male grizzly bears, McKinley Slope study area, Denali Na-
tional Park and Preserve, Alaska, 1991 through 1998.
No. of
Female Females No. of Female Female Cubs
Age Monitored Cubs Cubs per Female
6 11 2 1.0 0.0909
7 13 9 4.5 0.3462
8 11 8 4.0 0.3636
9 7 2 1.0 0.1429
10 5 9 4.5 0.9000
11 2 0 0.0 0.0000
12 1 0 0.0 0.0000
13 1 0 0.0 0.0000
14 3 0 0.0 0.0000
15 6 10 5.0 0.8333
16 7 3 1.5 0.2143
17 5 6 3.0 0.6000
18 8 7 3.5 0.4375
19 10 8 4.0 0.4000
20 9 7 3.5 0.3889
21 8 0 0.0 0.0000
22 7 5 2.5 0.3571
23 5 2 1.0 0.2000
24 4 5 2.5 0.6250
25 2 0 0.0 0.0000
26 3 4 2.0 0.6667
27 3 0 0.0 0.0000
28 2 1 0.5 0.2500
29 1 0 0.0 0.0000
30 1 0 0.0 0.0000
Total 135 88 44.0
Keay—Grizzly Bear Population Ecology and Monitoring 11
Physical Condition. Both subadult and
adult males had more body fat during Septem-
ber than in May (P = 0.001 for subadults and
0.000 for adults, Table 5). They did not differ
from each other during fall (P = 0.881).
Subadult males did have less body fat than
adult males during spring (P = 0.001). Simi-
larly, females of both age classes had more
body fat during September than in May (P =
0.000). They did not differ from each other
during either season (P = 0.292 for fall, P =
0.534 for spring). Adult females that pro-
duced cubs may have had more body fat the
previous fall than females that did not, al-
though, a larger sample of females without
cubs is necessary (P = 0.134, x¯ = 26.0 , n = 7
for females that were not observed with cubs
and x¯ = 29.6, n = 25 for females with cubs).
Adult females observed with COY weighed
more during the preceding September than
eligible females that were not observed COY
(P = 0.036, n = 25 with cubs and 7 without,
Table 6). Females aged 5 to 27 that were
weighed during a fall and subsequent spring
lost an average of 30% total body weight (SD
= 8.40, n = 13) and 18% lean body mass (SD =
7.67, n = 13). Females aged 5 to 29 captured
during a spring and fall of the same year
gained an average of 36% of their fall body
weight (SD = 7.710, n = 31) and 20% of their
fall lean body mass (SD = 9.994, n = 26).
Population Monitoring
Aerial Surveys. The first efforts to develop
field methods to measure visibility bias began
with 2 days of flying in May 1993. Six of 11
bears were observed during survey flights
under very poor sighting conditions. Monitor-
ing aircraft followed survey aircraft, but
movements of bears between flights prohibited
determination of important sighting conditions
for bears not detected by survey aircraft.
More intensive efforts were scheduled in
conjunction with a multi-aircraft bear density
estimate during September 1994 but inclement
weather precluded completion of field work.
During one day of flying, 10 of 10 marked
bears were detected by survey planes. Sighting
conditions that day were ideal with 4 to 6
inches of fresh snow throughout the study area.
During September, 1995, we tested 2
methods of obtaining sighting conditions of
bears missed during survey flights. On the first
day we radio tracked in survey units before and
after survey flights to provide 2 known loca-
tions of each bear to test for movement.
Movements between locations were high so we
next used high altitude surveillance simultane-
Table 5. Percent body fat of grizzly bears of different age classes during each season, McKinley
Slope study area, Denali National Park and Preserve, Alaska, 1991 through 1998.
Percent
Sex Age Class a
Season Fat n SD Minimum Maximum
Female Adult May 8.3 37 4.100 2.0 17.0
Female Adult September 27.4 52 6.237 16.0 47.0
Female Subadult May 7.3 8 4.136 2.0 14.6
Female Subadult September 25.1 10 5.466 18.0 33.0
Male Adult May 14.5 21 3.773 8.0 24.0
Male Adult September 23.2 7 4.282 18.0 31.0
Male Subadult May 8.3 13 5.801 -4.0 17.0
Male Subadult September 23.8 6 10.572 12.0 40.0
a Adults were ≥ 6 years, subadults were independent and ≤ 5 years.
Keay—Grizzly Bear Population Ecology and Monitoring 12
ous to survey flights. Survey planes detected
14 of 21 bears monitored. Difficulties for high
surveillance aircraft included inability to
continuously observe bears in some habitats
and uncertainty about when survey aircraft
should have detected a bear, especially when
multiple passes were made near the bear.
These issues were important since sight-ability
conditions varied with bear movement, cloud
cover, and aircraft position, all of which could
change in a short period of time.
Four experienced biologist/pilot teams
correctly classified 69% of a blind sample of
marked bears as adult female, adult male,
subadult female, or subadult male (n = 59).
When subadult sexes were combined, classifi-
ers were correct 73% of the time. The most
experienced biologist correctly classified 81%
of observations.
Thermal Infrared Imaging. Thermal
infrared temperatures were obtained from 55
bears during 1995 through 1998. Visual
inspection of data suggests that for overcast
conditions, bear surface temperatures were
about 6 to 8 C higher than background tem-
peratures. Under high incident solar radiation,
however, bear surface temperatures can exceed
that of the background by as much as 56 C and
can vary by 50 C on the bear itself.
DISCUSSION
Population Ecology
Serum antibody prevalence for infectious
canine hepatitis, canine distemper, and lepto-
spirosis was similar to or lower than elsewhere
in Alaska (Zarnke and Evans 1989, Zarnke
1992). Zarnke and Evans (1989) used a thresh-
old titer level of 1:20 for infectious canine
hepatitis. At that level only 1 Denali grizzly
bear would have been considered positive
instead of 3. This low prevalence makes it
unlikely that disease was a significant factor
affecting grizzly bear cub survival during the
Keay—Grizzly Bear Population Ecology and Monitoring 13
Table 6. Weights (kg) of grizzly bears of different age classes during each season, McKinley
Slope study area, Denali National Park and Preserve, Alaska, 1991 through 1998.
Mean Mean
Sex Age Class a Season Age Weight n SD Minimum Maximum
Female Adult May 16.4 99.2 59 13.403 69.0 132.0
Female Adult September 14.2 154.3 53 20.590 115.0 203.0
Female Subadult May 3.9 72.0 16 15.476 47.5 96.0
Female Subadult September 4.1 135.8 13 17.574 104.0 172.0
Female COY May 0.0 5.0 3 1.732 4.0 7.0
Female COY September 0.0 28.5 4 2.380 25.0 30.0
Male > 9 years May 16.7 205.1 20 38.022 146.0 260.0
Male > 9 years September 18.0 201.5 2 10.607 194.0 209.0
Male Adult May 13.2 189.1 32 44.068 104.0 260.0
Male Adult September 9.5 194.6 8 16.062 163.0 213.0
Male Subadult May 4.1 92.5 30 25.238 48.0 159.0
Male Subadult September 3.9 136.7 10 32.857 96.0 195.0
Male COY May 0.0 7.3 3 0.577 7.0 8.0
Male COY September 0.0 25.0 3 5.568 19.0 30.0
a Adults were ≥ 6 years, subadults were independent and ≤ 5 years, COY refers to cubs-of-the-year.
study.
Brown bear density in Denali National
Park, adjusted to represent forage-producing
habitat, was comparable to those reported by
Dean (1987) and Murie (1981) on adjacent
study areas in the Park. It was higher than the
nearby harvested Susitna population during
1985 (Miller et al. 1997). Sight-ability of both
females and males was higher during the fall
Denali survey than reported in other areas of
Alaska (Miller et al. 1997). High sight-ability
of both sexes during fall density estimates was
due to greater activity of bears during the
period of hyperphagia as compared to spring
surveys (Stelmock and Dean 1986), and open
nature of habits used for foraging at that time.
Sight-ability was sufficient to markedly reduce
survey costs to obtain an adequate sample for
density estimation in areas where fall hunting
season does not compromise human safety or
the vulnerability of study animals to sport
harvest due to aircraft attention or disturbance.
The bimodal distribution of the female
population likely resulted from periodic re-
cruitment related to high dependent bear
mortality. Pulses of recruitment probably
occur either when foraging conditions become
exceptional or predation eases due to unknown
factors.
High density, high independent bear sur-
vival rates, and lack of human interference
during the past 80 years suggests the Denali
population is likely at carrying capacity. High
dependent mortality coupled with periodic
recruitment may allow the population to
oscillate mildly around a mean density, but in
the long term, the population is likely stable.
The negative rate of increase associated with
observed survival and reproductive rates is
probably a result of the study not monitoring
cub production and survival long enough to
include a recruitment pulse.
During the brief summer months, adult
females gain about 50% of their spring weight
in preparation for cub production and the long
winter hibernation. The loss of 18% to 20%
lean body mass during winter comes largely
from skeletal muscle and emphasizes the
marginal nutritional condition of females and
subadult males in this population. Starvation
loss and poor physical condition associated
with some dependent bear mortality suggests
that nutrition plays a role in the high mortality
rates of dependent bears.
Population Monitoring
Aerial Surveys. Fall surveys provided the
highest sight-ability and minimized potential
observation bias due to habitat selection and
secrecy among bears of differing sex and
reproductive status, compared to spring sur-
veys in similar habitats (Miller et al. 1997).
Fall surveys are not bias-free, however, since
more detection bias was observed between
individuals within rather than among sex and
reproductive groups (Miller et al. 1997).
Without some means to measure or correct for
sight-ability bias there will always be uncer-
tainty when comparing surveys between
locations and years, as sighting conditions
vary.
None of the methods we tested for measur-
ing sighting bias proved satisfactory. Simulta-
neous, high altitude surveillance was most
promising but still did not permit accurate
determination of sighting conditions of missed
bears at the desired scale. Coarser scale
studies, using detection rates in broad vegeta-
tion categories, may have been possible but the
loss of information and precision in estimating
the number of missed bears was a concern.
Inability to develop clean and efficient sam-
pling procedures resulted in abandonment of
efforts to measure visibility bias following
field work in 1995. Continued research on
double sampling techniques by Dr. Earl Becker
(Alaska Department of Fish and Game) may
prove helpful, but limitations due to low bear
density at Denali will likely be a concern.
High observed mortality rates of dependent
Keay—Grizzly Bear Population Ecology and Monitoring 14
bears and high between year variability in
production suggests that monitoring cub
production and survival from radio-collared
bears would only be valuable in extremely
long-term studies with reasonable conclusions
drawn about each decade. The high expense of
such efforts for the amount of information
received makes it impractical.
Alternatively, existing data could be used
to estimate a range of possible variances for
the binomial estimator. Modeling efforts could
display the range of precision likely under
varying conditions. If this technique appears
favorable, it should be validated on a known
density population. Then, traditional aerial
surveys using the binomial estimator, bears per
square kilometer or bears per hour, would
provide useful estimates where bears are not
harvested and human impacts not expected.
When the need for precise information regard-
ing population status is great, intensive, short-
term capture-mark resight studies using radio
transmitters are recommended.
Thermal Infrared Imaging. Bear surface
temperatures appeared sufficiently different
from background temperatures to permit
detection using currently available thermal
infrared equipment, during periods of overcast
skies or during early morning or late evening
hours when incident solar radiation is at a low
angle. Confirmation of these findings through
more rigorous quantitative analysis, coupled
with a theoretical assessment of survey height,
survey path, image size, and observation
procedures, for the terrain bears inhabit in
Denali National Park, should be accomplished
prior to conducting additional field work. The
expense of this technique, given the current
availability of equipment, plus the research and
development necessary to make this technique
operational, will limit its value in the near term
as a useful bear population monitoring tool.
Even if in the long term it does not provide
improved detection rates the fact that biases
associated with it differ from those associated
with visual surveys suggests it may still have
promise as a supplemental tool or to validate
other techniques. Consequently, development
of the technique should be supported but only
if the cost of development can be maintained at
reasonably low levels. Development should
occur at locations proximate to available
thermal infrared equipment.
MANAGEMENT
RECOMMENDATIONS
1. Conduct intensive, short-term studies of
grizzly bears in different habitats within and
around the park to 1) determine population
status at additional locations, and 2) identify
human factors that could affect grizzly bear
productivity or survival. Monitor those human
factors and important natural resources that
could be affected by humans. Continuous
monitoring of long-lived, low density animals
with low reproductive output is too expensive
for most monitoring programs. Understanding
ecological requirements of those animals and
the potential impacts human activities might
have on them would provide a means to moni-
tor causative factors at much reduced cost.
These studies would provide valuable baseline
data needed for future comparisons to docu-
ment whether change has occurred.
2. Use global positioning system and
geographic information system technologies to
determine grizzly bear habitat use patterns and
build a comprehensive model of prioritized
bear habitats which would identify potentially
important movement corridors and foraging
habitats in areas likely for human development
or activity. Models could predict the area and
proportion of various qualities of habitat that
would be affected by human activities, without
the complexity of attempting to quantify
impacts to carrying capacity or population
dynamics.
Keay—Grizzly Bear Population Ecology and Monitoring 15
3. Use global positioning system and
geographic information system technologies to
assess impacts of various human activities on
bear movements and use those data to enhance
the habitat model. The habitat quality model
could be fine-tuned to assess the amount of
impact based on the type and amount of human
activity proposed. Much of this work can be
drawn from studies outside Denali, although
some effort to elucidate impacts of human
activities unique to Denali may be necessary.
4. Conduct periodic autumn aerial sur-
veys, without marked bears, using the binomial
estimator, bears observed per hour and bears
observed per square kilometer, to provide a
measure of bear distribution, abundance,
status, and to determine important concentra-
tion areas, for occasional evaluations where
human influence on the population is expected
to be negligible, or when first considering an
area for intensive study. This technique is
least expensive and least intrusive on the bear
population, can be accomplished during one
field season, and can be implemented immedi-
ately. Highly experienced pilots and observers
can provide a reasonable assessment of sex and
age composition to suggest a healthy popula-
tion. Unexpectedly low bear observation rates
or a low proportion of dependent bears would
be cause for more intensive investigation.
5. Conduct intensive surveys, using radio
collars and capture-mark-resight techniques,
when there appears to be a change in factors
that affect bear reproduction and survival, for
harvested populations, or when new develop-
ments pose threats to an existing population.
When more precise, unbiased data are needed
for comparisons between populations or
between years, capture-mark-resight tech-
niques provide the only reliable and proven
method. Such effort will require two to three
years of pre-marking to assure all reproductive
classes of adult females are represented in the
marked population for the density estimate.
Such efforts could efficiently be combined
with other recommendations for concurrent
study.
ACKNOWLEDGMENTS
I would like to express my sincere appre-
ciation to the many people who assisted with
this research. P. Owen deserves special recog-
nition; she very capably managed data, assisted
with field work and was critical to the success
of this project. L. Adams provided tremendous
ecological advice and logistical support. Pilots
D. Miller, S. Hamilton, D. Glaser, H.
McMahan, J. Lee, J. Rood, M. Meekin, H.
Twitchell, R. Purdum, and J. Unruh provided
fixed-wing aircraft support for radio tracking
and surveys. H. McMahan, J. Lee, S. Hamil-
ton, and V. Barnes provided valuable insight in
identifying factors that affect sight-ability and
meaningful ways to record those data. Heli-
copter pilots K. Butters, P. Walters, R. Warbe-
low, L. Lingren, and G. Gulick did a great job
facilitating bear captures. Biologists V. Bar-
nes, and S. Miller, assisted with density esti-
mates and provided very useful guidance on
study design and techniques. J. Bryant, S.
Farley, G. Hilderbrand, B. Dale, P. DelVec-
chio, A. Yost, K. Stahlnecker, K. Fox, J.
Belant, and others assisted with occasional
radio tracking and bear captures. S. Wesser
and J. Paynter provided computer mapping
support. S. Cline provided graphics support.
C. Robbins contributed valuable guidance on
the role of nutrition in cub production and
survival. Veterinarians J. Blake, G. Ford, and
W. Taylor provided advise on chemical immo-
bilizations. J. Blake conducted necropsies and
G. Ford participated in bear captures and
provided useful guidance on animal handling.
LITERATURE CITED
Adams, L. G., F. J. Singer, and B. W. Dale.
1995. Caribou calf mortality in Denali
National Park, Alaska. Journal of
Wildlife Management 59: 584-594.
Keay—Grizzly Bear Population Ecology and Monitoring 16
Adams, L. G., B. W. Dale, and B. Shults.
1989. Population status and calf
mortality of the Denali Caribou Herd,
Denali National Park and Preserve,
Alaska - 1984-1989. National Park
Service Natural Resources Progress
Report AR-89/13.
Amstrup, S. C. and G. M. Durner. 1995.
Survival rates of radio-collared female
polar bears and their dependent young.
Canadian Journal of Zoology.
73:1312-1322.
Bunnell, F. E. and D. E. N. Tait. 1980. Bears
in models and reality—implications to
management. International Conference
on Bear Research and Management
3:15-23.
Caughley, G. 1974. Analysis of vertebrate
populations. John Wiley & Sons, New
York, New York.
Dean, F. C. 1987. Brown bear density, Denali
National Park, Alaska, and sighting
efficiency adjustment. International
Conference on Bear Research and
Management 7:37-43.
Farley, S. D. and C. T. Robbins. 1994. Devel-
opment of two methods to estimate
body composition of bears. Canadian
Journal of Zoology 72:220-226
Garshelis, D. L., K. V. Noyce, and P. L. Coy.
1998. Calculating average age of first
reproduction free of the biases preva-
lent in bear studies. Ursus 10:437-447.
Gasaway, W. C., S. D. DuBois, D. J. Reed, and
S. J. Harbo. 1986. Estimating moose
population parameters from aerial
surveys. Biological Papers, University
of Alaska 22.
Haber, G. C. 1977. Socio-ecological dynamics
of wolves and prey in a subarctic
ecosystem. Ph.D. Dissertation. Univer-
sity of British Columbia.
Keay, J. A. 1995. Accuracy of cementum age
assignments for black bears. California
Fish and Game, 81:113-121.
Kemp, G. A. 1974. The dynamics and regula-
tion of black bear, Ursus americanus,
populations in Northern Alberta.
International Conference on Bear
Research and Management, 3:191-197.
Matson, G., L. Van Daele, E. Goodwin, L.
Aumiller, H. Reynolds, and H. Hris-
tienko. 1993. A laboratory manual for
cementum age determination of Alaska
brown bear first premolar teeth. Mat-
son’s Laboratory, Milltown, Montana,
USA.
Miller, C. A. and R. G. Wright. 1998. Visitor
satisfaction with transportation services
and wildlife viewing opportunities in
Denali National Park and Preserve.
Idaho Cooperative Fish and Wildlife
Research Unit.
Miller, S. D. 1990. Impact of increased bear
hunting on survivorship of young bears.
Wildlife Society Bulletin 18:462-467.
Miller, S. D. and M. McDonald. 1996. Devel-
opment and improvement of bear
management techniques and procedures
in Southcentral Alaska. Alaska Depart-
ment of Fish and Game. Federal Aid in
Wildlife Restoration Final Research
Report, 1 July 1992 - 30 June 1996.
Miller, S. D., G. C. White, R. A. Sellers, H. V.
Reynolds, J. W. Schoen, K. Titus, V.
G. Barnes, Jr., R. B. Smith, R. R.
Keay—Grizzly Bear Population Ecology and Monitoring 17
Nelson, W. B. Ballard, and C. C.
Schwartz. 1997. Brown and black bear
density estimation in Alaska using
radio telemetry and replicated mark-
resight techniques. Wildlife Mono-
graph 133.
Murie, A. 1944. The wolves of Mount
McKinley. National Park Serv. Fauna
Series No. 5.
Murie, A. 1981. The grizzlies of Mount
McKinley. National Park Service
Scientific Monograph Series No. 14.
Pollock, K. H., S. R. Winterstein, C. M.
Bunck, and P. D. Curtis. 1989. Sur-
vival analysis in telemetry studies: the
staggered entry design. Journal of
Wildlife Management, 53:7-15.
Samuel, M. D., E. O. Garton, M. W. Schlegel,
and R. G. Carson. 1987. Visibility
bias during aerial surveys of elk in
northcentral Idaho. Journal of Wildlife
Management 51:622-630
Stelmock, J. J. and F. C. Dean. 1986. Brown
bear activity and habitat use, Denali
National Park - 1980. International
Conference on Bear Research and
Management, 6:155-167.
Taylor, W. P., H. V. Reynolds, III, and W. B.
Ballard. 1989. Immobilization of
grizzly bears with tiletamine hydro-
chloride and zolazepam hydrochloride.
Journal of Wildlife Management,
53:978-981.
Vierek, L. A., C. T. Dyrness, A. R. Batten, and
K. J. Wenzlick. 1992. The Alaska
vegetation classification. General
Technical Report. PNW-GTR-286.
Portland, Oregon: U.S. Department of
Agriculture, Forest Service, Pacific
Northwest Station.
White, G. C. 1996. NOREMARK: Popula-
tion estimation from mark-resighting
surveys. Wildlife Society Bulletin
24:50-52.
Wiggers, E. P. and S F. Beckerman. 1993.
Use of thermal infrared sensing to
survey white-tailed deer populations.
Wildlife Society Bulletin 21:263-268.
Zarnke, R. L. 1992. Alaska wildlife serologic
survey 1975-1992. Alaska Department
of Fish and Game.
Zarnke, R. L. and M. B. Evans. 1989. Sero-
logic survey for infectious canine
hepatitis virus in grizzly bears (Ursus
arctos) from Alaska, 1973 to 1987.
Journal of Wildlife Diseases 25:568-
573.
Keay—Grizzly Bear Population Ecology and Monitoring 18