Global Change Biology (2014) 20, 76–88, doi: 10.1111/gcb.12339

Variation in the response of an Arctic top predator
experiencing habitat loss: feeding and reproductive
ecology of two polar bear populations
KARYN D. RODE* ERIC V. REGEHR*, DAVID C. DOUGLAS†, GEORGE DURNER‡, ANDREW
E . D E R O C H E R § , G R E G O R Y W . T H I E M A N N ¶ and S U Z A N N E M . B U D G E k
*Marine Mammals Management, U. S. Fish and Wildlife Service, 1011 E Tudor Road, Anchorage, AK 99502, USA, †Alaska
Science Center, U. S. Geological Survey, 250 Egan Drive, Juneau, AK 99801, USA, ‡Alaska Science Center, U. S. Geological
Survey, 4210 University Drive, Anchorage, AK 99508, USA, §Department of Biological Sciences, University of Alberta,
Edmonton, AB T6G 2E9, Canada, ¶Faculty of Environmental Studies, York University, 4700 Keele Street, Toronto, ON M3J 1P3,
Canada, kProcess Engineering and Applied Science, Dalhousie University, Halifax, NS B3J 2X4 Canada

Abstract
Polar bears (Ursus maritimus) have experienced substantial changes in the seasonal availability of sea ice habitat in
parts of their range, including the Beaufort, Chukchi, and Bering Seas. In this study, we compared the body size, condition, and recruitment of polar bears captured in the Chukchi and Bering Seas (CS) between two periods (1986–1994
and 2008–2011) when declines in sea ice habitat occurred. In addition, we compared metrics for the CS population
2008–2011 with those of the adjacent southern Beaufort Sea (SB) population where loss in sea ice habitat has been
associated with declines in body condition, size, recruitment, and survival. We evaluated how variation in body condition and recruitment were related to feeding ecology. Comparing habitat conditions between populations, there
were twice as many reduced ice days over continental shelf waters per year during 2008–2011 in the SB than in the
CS. CS polar bears were larger and in better condition, and appeared to have higher reproduction than SB bears.
Although SB and CS bears had similar diets, twice as many bears were fasting in spring in the SB than in the CS.
Between 1986–1994 and 2008–2011, body size, condition, and recruitment indices in the CS were not reduced despite
a 44-day increase in the number of reduced ice days. Bears in the CS exhibited large body size, good body condition,
and high indices of recruitment compared to most other populations measured to date. Higher biological productivity and prey availability in the CS relative to the SB, and a shorter recent history of reduced sea ice habitat, may
explain the maintenance of condition and recruitment of CS bears. Geographic differences in the response of polar
bears to climate change are relevant to range-wide forecasts for this and other ice-dependent species.
Keywords: Body condition, body size, climate change, diet, feeding ecology, morphometrics, reproduction, Ursus maritimus
Received 4 September 2012 and accepted 20 July 2013

Introduction
In the Arctic, reductions in the extent and thickness of
sea ice have been observed (Maslanik et al., 2011) and
are projected to continue through at least 2100 based on
multiple global climate models, including all models
participating in the Intergovernmental Panel on Climate Change Fourth Assessment Report (Holland et al.,
2006; Stroeve et al., 2012). Because the direct effects of
climate change can be difficult to quantify, a species’
projected response is often estimated by the relationship of the organism with its habitat (Parmesan, 2006).
Polar bears (Ursus maritimus) are considered particuPresent address: Karyn D. Rode, Alaska Science Center, US Geological Survey, 4210 University Drive, Anchorage, AK 99508, USA
Correspondence: Karyn D. Rode, tel: +907 786 7106,
fax: +907 786 7150, e-mail: krode@usgs.gov

76

larly vulnerable to negative effects associated with
climate-induced sea ice loss because they rely on sea ice
to access their primary prey, seals (Amstrup et al., 2008;
Laidre et al., 2008).
Links between multiyear trends of declining sea ice
with declining polar bear condition (Stirling et al., 1999;
Obbard et al. 2006; Rode et al., 2010, 2012), reproduction
(Rode et al., 2010; Regehr et al., 2007), and survival
(Regehr et al., 2007; 2009; Peacock et al., 2012) have been
documented for five polar bear populations (Baffin
Bay, Davis Strait, Southern Hudson Bay, Southern
Beaufort Sea, and Western Hudson Bay). However,
declines in body condition and juvenile survival rates
in Davis Strait occurred simultaneous to an increase in
population density and could have been a result of density-dependent effects (Rode et al., 2012; Peacock et al.
2012) and other populations have experienced sea ice
loss with no apparent decline in population size or
© 2013 John Wiley & Sons Ltd

 P O L A R B E A R R E S P O N S E S T O C L I M A T E C H A N G E 77
survival rates (i.e., Southern Hudson Bay, Gagnon and
Gough 2005; Obbard et al., 2007; Northern Beaufort Sea,
Stirling et al., 2011). These studies suggest that, among
the world’s 19 subpopulations, the response of polar
bears to sea ice loss may vary temporally and geographically, due to variation in ecosystem function and
in the life history strategies (Amstrup et al., 2008). For
example, bears in some populations experience several
months of ice-free conditions and come onshore where
they fast until sea ice returns. In other areas, the majority of bears spend the entire year on the sea ice or a
small proportion of the population comes onshore during the annual sea ice minimum (Schliebe et al., 2008).
Reduced body size appears to be one of the proximate mechanisms by which climate change affects species (Gardner et al., 2011; Sheridan and Bickford, 2011)
and may be a useful indicator of organismal responses
to climate change. Size and body mass of ursids
respond to density-independent fluctuations in the
environment and density-dependent effects (Zedrosser
et al., 2006) and are linked to population density, reproduction, and cub survival (Noyce and Garshelis, 1994;
Hilderbrand et al., 1999). Studies of polar bears have
documented declines in body size with increases in the
duration of the ice-free period in seasonal ice habitats,
and with the annual sea ice minimum extent in circumpolar regions (mass relative to length; Stirling et al.,
1999; mass and skull size; Rode et al., 2010, 2012).
Although declines in body size could be an adaptive
response to reduce energy requirements, observed
declines in size were associated with declines in condition (i.e., measures of energy reserves relative to structural body size; Stirling et al., 1999; body mass and
skull width; Rode et al., 2010). For polar bears, trends in
body size and condition have been suggested as an
indicator of potential negative effects of climate warming (Stirling and Derocher, 2012), assuming that trends
in condition do not reflect population regulation (e.g.,
density dependence in the absence of declining habitat)
or natural short-term variation (Rode et al., 2012). One
of the most important food resources for polar bears,
ringed seal (Pusa hispida) pups, become available in
spring only (Pilfold et al., 2012) and continue to be an
important food source until breakup in early summer
(Stirling 2002). Because sea ice loss is occurring disproportionally in the summer and early autumn (Stroeve
et al., 2012), effects of sea ice loss on bear body condition, reproduction, and survival may depend on the
importance of summer foraging, vs. the contribution of
foraging during other times of the year, to annual variation in condition.
Little is known about the population dynamics and
ecology of the polar bear population that ranges in the
Chukchi, northern Bering, and eastern Siberian seas
© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

(hereafter the Chukchi Sea “CS” population) where
some of the most rapid sea ice loss in the Arctic has
occurred (Rigor and Wallace, 2004; Douglas 2010). Durner et al., (2009) estimated that optimal polar bear habitat in the CS declined by 8% per decade between 1979
and 2006, which is higher than declines experienced by
the adjacent southern Beaufort Sea population (SB)
(4.8% per decade) where reduced sea ice has been associated with declines in condition, recruitment, and survival of polar bears (Regehr et al., 2009; Rode et al.,
2010). Observed and projected losses in sea ice habitat
based on general circulation models, qualitative relationships between sea ice and population dynamics,
and potential effects of other population stressors incorporated in a Bayesian Network model led Amstrup
et al. (2008) to hypothesize that populations in the
divergent ice ecoregion, which includes the CS population, may face extirpation by the mid-21st century
unless climate warming is significantly mitigated.
Although the CS has experienced loss of sea ice, this
region differs ecologically from the adjacent SB (Sigler
et al., 2011) and from western Hudson Bay (WH) where
polar bear populations have exhibited signs of nutritional stress and associated declines in recruitment
(Stirling et al., 1999; Regehr et al., 2007, 2009; Moln 
ar
et al., 2010). Polar bears in WH and other eastern Canadian regions, such as Davis Strait and Baffin Bay, experience seasonally ice-free conditions and spend longer
time on land as the duration of the ice-free season
increases (Stirling and Parkinson, 2006). In contrast, sea
ice habitat in the CS has historically persisted over part
of the continental shelf even during the annual sea ice
minimum (Douglas, 2010). Unlike the SB, which has a
narrow continental shelf, nearly the entire CS is <300 m
(Fig. 1). Productivity and polar bear prey density are
much higher over the continental shelf than in deeper
waters (Frost et al., 2002) and the CS has some of the
highest marine primary productivity in the Arctic
(Sakshaug, 2004). Furthermore, the CS contains a
greater diversity of marine mammals as potential food
resources than the SB, a factor that has been linked to
increased body mass in predatory carnivores (Gittleman, 1985). In addition to the bearded seal (Erignathus
barbatus), ringed seal, beluga whale (Delphinapterus leucas), and bowhead whale (Balaena mysticetus; primarily
from subsistence harvest) that are available to polar
bears in the SB, CS polar bears also overlap with spotted seal (Phoca largha), ribbon seal (Histriophoca fasciata),
and Pacific walrus (Odobenus rosmarus divergens). However, diet composition, and its potential role in affecting
body condition and reproduction have yet to be investigated for polar bears in the Chukchi Sea.
In this study, we compare body size, condition, and
reproductive indices in the CS between two sampling

 78 K . D . R O D E et al.

Fig. 1 Locations of captured polar bears in the Chukchi Sea (1986–1994 and 2008–2011) and in the southern Beaufort Sea (2008–2011)
relative to the International Union for the Conservation of Nature Polar Bear Specialist Group’s identified population boundaries (in
black) for the Chukchi (also referred to as Alaska-Chukotka) and southern Beaufort Sea populations. Boundaries used for summarizing
sea ice conditions for each population are outlined in red.

periods: 1986–1994 and 2008–2011 when sea ice habitat
declined. We also compare body size, condition, and
reproduction in the CS and SB for 2008–2011. We evaluated spatial and temporal habitat changes, using sea ice
metrics based on habitat preferences of polar bears, and
relationships between these metrics and body size, condition, and reproduction. We also examined how feeding ecology, including diet and fasting behavior, may
affect variation in condition and reproduction.

Materials and methods

Sea ice availability
Sea ice habitat was quantified for the years before (1985–1993
and 2007–2010) collection of spring body condition and reproductive data (Rode et al., 2010). SB and CS bears primarily use
ice over the continental shelf (Durner et al., 2004, 2009), which
likely provides better access to prey and has been associated
with higher body condition and reproduction than deep-water
regions of the Arctic Ocean (Rode et al., 2010). The study area
for ice habitat analyses was bounded to the west and east by

the International Union for the Conservation of Nature Polar
Bear Specialist Group population boundaries for the CS and
SB (Obbard et al., 2010), and to the north and south by the
range limits of radio-collared polar bears monitored between
1984 and 2011 (Fig. 1; US Geological Survey and US Fish and
Wildlife Service, unpublished data).
We used sea ice concentration estimates derived from raster-format 25 km 9 25 km resolution passive microwave satellite imagery (Cavalieri et al., 1996) to calculate two habitat
metrics: (i) the number of “reduced ice days” per year (ice),
quantified as the number of days in which there was
<6250 km2 (roughly ten 25-km resolution pixels) of sea ice over
the continental shelf within each population’s ice analysis area
(Fig. 1); and (ii) the mean daily minimum distance (hereafter
referred to as “distance” or mndist) from each 25 km resolution
pixel along the edge of the continental shelf (the 300 m isobath)
to the edge of the pack ice, averaged over all days in September
(the month of the annual sea ice minimum; Rigor and Wallace,
2004). When measuring the minimum distance between each
shelf edge pixel and the ice edge, distances were set to zero in
cases where the ice edge was south of the shelf break.
We assumed that the similar size of the two populations
(CS: 2000 bears based on extrapolation of den surveys; Aars

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

 P O L A R B E A R R E S P O N S E S T O C L I M A T E C H A N G E 79
et al., 2006; SB: 1526 Æ 315 bears based on capture-recapture
estimation; Regehr et al., 2006) would allow use of a common
area threshold for the “reduced ice days” metric by creating
an effectively similar metric of population density. Because
the area threshold representing foraging constraints is not
known, we initially examined trends in ice using three thresholds: 6250, 10 000, and 20 000 km2 (Figs. S1 and S2). Because
polar bears occasionally occupy ice as low as 15% concentration (Durner et al. 2009), we also examined two concentration
thresholds for defining the presence of ice: 15% and 50%. We
found that trends, differences between populations (Figs. S1
and S2), and relationships with morphometric data were similar for the ice metric thresholds mentioned above, so we chose
the most limiting area threshold of 6250 km2 for reported
analyses. We also chose to define ice habitat where concentrations were ≥50% because polar bears have exhibited strongest
selection for this concentration throughout much of the year
in this region (Durner et al., 2004, 2006). Mann–Whitney U
tests were used to determine if ice metrics differed between
the CS during 1986–1994 and 2008–2011 and between the CS
and SB during 2008–2011.

Polar bear measurement and sample collection
We analyzed data from polar bears captured and released on
sea ice between mid-March and early May, during 2008–2011
in the SB and CS, and during 1986–1994 in the CS (Fig. 1;
Table S1). Between 2008 and 2011, captures in both populations targeted all sex and age classes. However, females with
cubs-of-the-year (COY, <12 months old) were rarely encountered in the CS during 2008–2011 because captures did not
occur near key denning areas in Russia (Wrangel and Herald
islands). To standardize the data sets, females with COY were
excluded from all analyses. Adult males were avoided in the
CS during 1986–1994 and could not be compared with the contemporary samples in the CS.
We classified adults as ≥5 years old (Regehr et al., 2006);
subadults as independent (i.e., without their mother) bears
2–4 years old; and dependent young as 2-year olds, yearlings, and COY. Polar bears were located from a helicopter
and immobilized with a dart containing zolazepam-tiletamine (Telazolâ or Zoletilâ) (Stirling et al., 1989). Immobilized
bears were weighed to the nearest kilogram (kg) in the Alaskan portion of the SB and CS, but were not weighed in the
Canadian portion of the SB. For all bears, straight-line body
length, tail length, and zygomatic skull width were measured and age was determined as described in Rode et al.
(2010). We measured straight-line body length of adults as
an indicator of structural size. Body mass, skull width, and
energy density (described below), which can vary annually
and seasonally (Rode et al., 2010), were used as measures of
condition. Measures of body mass and body length were
used to calculate energy density (MJ kgÀ1) for each bear following Moln ar et al. (2009). Energy density quantifies the
‘energetic content of storage relative to the mass of tissue
that requires energy for somatic maintenance’ and standardizes body mass measures relative to structural size (Moln 
ar
et al., 2009, 2010).

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

Standard methods for monitoring reproduction and offspring survival in ursids (e.g., continuous monitoring of individuals over time via resighting or radio telemetry) were not
possible for the CS and SB populations because of the large
and remote geographic areas. Instead, we examined indices of
recruitment including yearling litter size (number of yearlings
in a litter), the proportion of adult females with at least one
yearling (females with yearlings/all females except those with
COY), and the number of yearlings per female (number of
yearlings in the sample/number of females unaccompanied
by COY in the sample). We also examined relationships
between maternal condition and the production and condition
of COY (for the SB only) and yearlings (for the SB and CS), to
investigate relationships between maternal condition and cub
production.
Fat biopsies were collected from CS bears captured 2008–
2011 to estimate diet. A 6-mm biopsy punch placed approximately 15 cm lateral to the base of the tail provided a
full-layer core from skin to muscle (Thiemann et al., 2006).
Blood was collected in no-additive tubes to identify fasting
behavior from bears captured in the CS during 2008–2011. Fat
and blood were either not collected or unavailable for CS
bears captured during 1986–1994 and for SB bears captured
during 2008–2011. Information on diet composition and
fasting behavior of SB bears were obtained from published
studies (Thiemann et al., 2008; Cherry et al., 2011).

Comparisons of body size, condition, and reproduction
Comparison of body size and condition over time and among
populations requires consideration of explanatory covariates,
such as age. To facilitate comparisons between samples (e.g.,
between 1985–1994 and 2008–2011 in the CS and between the
SB and CS samples during 2008–2011), we fit modified von
Bertalanffy growth curves, relative to bear age, to measures of
body mass, length, and skull size (Derocher and Wiig, 2002)
and used the residuals as a dependent variable in generalized
linear models (GLM; Laidre et al., 2006; Rode et al., 2010).
Comparisons between populations and periods were made
separately by including only data for the two periods in the
CS and examining a time effect in the model or only the data
for the two populations captured during 2008–2011 and examining a population (pop) effect in the model. We also related
condition and reproduction to two potential ice metrics, separately, in place of either time or pop in the models. Ice replaced
pop or time in the models to determine if ice might explain
potential differences between periods or populations and
because ice was likely to be correlated with those two factors.
In addition to pop, time, or ice effects, other covariates included
in candidate models included age, capture date, and litter
size (for yearlings). Details about model selection are in
Appendix S2.
We used asymptotic body mass, body length, and skull
width from growth curves to compare the mean maximum
sizes obtained by bears in different populations. Asymptotic
body mass and body length were used to calculate energy
density for bears in the SB and CS, as well as in the other
polar bear populations that were considered for broader

 80 K . D . R O D E et al.
comparison. Body mass for other populations (Derocher, 1991;
Derocher and Wiig, 2002) was calculated using girth and
equations reported by Kolenosky et al. (1989), except for the
Barents Sea population where a population-specific equation
was used (Derocher and Wiig 2002).

(a)

Diet composition and fasting behavior
We examined the contribution of different prey species to the
diet of independent bears captured in the CS during 2008–
2011 based on quantitative fatty acid signature analysis (Iverson et al., 2004, Thiemann et al., 2008). Levels of blood urea
nitrogen (BUN) and creatinine were measured in serum to
identify bears that had fasted for >7 days (Derocher et al.,
1990). Details on diet analyses are in Appendix S1. Generalized linear models were used to evaluate the contribution of a
food item to dietary biomass as a function of sex, age, and
body mass.

(b)

Results

Sea ice availability
The number of reduced ice days over the CS continental shelf averaged 44 days during 2007–2010 compared
to 0 days during 1985–1993 (Mann–Whitney U-test,
df = 1, P = 0.003; Fig. 2). The mean daily minimum distance in September between the CS continental shelf
and pack ice increased 445 km between 1985–1993
(10 km) and 2008–2011 (455 km; Mann–Whitney U-test,
df = 1, P = 0.003).
During 2007–2010, the reduced ice period was about
twice as long in the SB (94.0 Æ 3.9 days; mean Æ SD)
compared to the CS (44.5 Æ 36.7 days; paired t-test:
P = 0.068), and the mean minimum distance to the September ice edge was similar in the SB (465 Æ 54 km)
compared to that in the CS (455 Æ 75 km; paired t-test:
P = 0.90).

Comparisons of body condition and reproduction during
1986–1994 and those during 2008–2011 in the CS
During 2008–2011, CS bears were either larger and in
better condition, or similar in size and condition compared to CS bears during 1986–1994. Parameters of
growth curves are provided in Table S2. Comparing
growth curves directly (Fig. 3a) females ≥1 year old
during 2008–2011 were 27.9 kg larger in body mass
(GLM including age: v2 = 45.1, P < 0.0001) and 0.7 cm
larger in skull width (GLM including cdate: v2 = 17.0,
P < 0.0001) than females during 1986–1994.
Mean body mass of CS male and female yearlings
was 31.4 and 19.2 kg greater and mean skull width
was 1.5 cm greater for both sexes during 2008–2011
than that during 1986–1994 (Tables 1 and S3).

Fig. 2 Annual variation during 1979–2010 in (a) the number of
days in which the extent of sea ice (≥50% concentration) was
≤6250 km2 (absence of bars in any year indicates 0 ice-free days)
and (b) September mean minimum distance between the continental shelf and sea ice of ≥50% concentration.

Differences are based on b coefficient values in Table
S3 and significance of P < 0.05 in general linear models. Body mass and skull width of subadult females
and body mass of subadult males did not differ, but
skull width of subadult males was 0.9 cm larger during 2008–2011 than that found during 1986–1994.
Mean body mass of adult females was 29.8 kg larger,
skull width was 1.0 cm larger, and body length was
10.0 cm larger during 2008–2011 than that during
1986–1994, but energy density did not differ (Table
S3). Because few adult males were captured in the
CS during 1986–1994, sample sizes were insufficient
for most comparisons. Mean skull width of six adult
males captured during 1986–1994 did not differ from
adult males captured during 2008–2011.
There was no difference in the number of yearlings
per female, yearling litter size, or the annual percentage
of females with yearlings between periods in the CS
(Tables 2 and S3; P > 0.50 for all tests).

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

 P O L A R B E A R R E S P O N S E S T O C L I M A T E C H A N G E 81
(a)

(b)

Fig. 3 Relationships between age and body mass for (a) female
and (b) male polar bears in the Chukchi Sea (1986–1994 for adult
females only; 2008–2010 for both sexes) and southern Beaufort
Sea (2008–2011) fit with modified von Bertalanffy growth
curves.

Replacing the time effect in general linear models
with one of the two ice metrics (ice or mndist) resulted
in a positive relationship between body condition and
reduced ice conditions for most sex and age classes
(Table 3, S3). Both ice metrics declined between 1986–
1994 and 2008–2011 simultaneous to maintenance or
increases in body condition. The number of yearlings
per female declined as sea ice was reduced, but there
was no effect of ice conditions on yearling litter size
when including data over the two periods in the CS.

skull width was 1.0 cm greater (v2 = 26.6, P < 0.0001)
than SB males.
CS female and male yearlings weighed 18.2 and
39.5 kg more and had skull widths 0.8 and 1.3 cm larger than their SB counterparts, respectively (Tables 1
and S4). CS subadult males weighed 44.3 kg more than
SB subadult males, but subadult females did not exhibit
a difference in mass. CS subadult females and males
had skull widths that were 1.0 cm and 0.9 cm wider
than SB subadult females and males, respectively.
Adult females in the CS weighed 30.5 kg more, had
0.9 cm larger skull widths, and had 25% higher energy
density than SB adult females, but body lengths were
similar. CS adult males weighed 52.6 kg more, had
1.2 cm larger skull widths, and had 14% higher energy
density than SB adult males, but body length did not
differ.
The number of yearlings per female in spring was
34.8% and 43.5% higher (GLM: v2 = 8.7, P = 0.003) in
the CS than in the SB 2008–2011 sample (Tables 1, 2,
and S4). There was no difference in yearling litter size.
The percent of females with yearlings across all years
was 10.9% higher in the CS than in the SB during 2008–
2011, but this was not statistically different (F1,6 = 0.08,
P = 0.79). Variation between years was large
(SD = Æ 20%) and annual sample sizes of females were
low (20 Æ 10 STD females). Yearling and COY body
masses for both CS and SB were related to maternal
body mass (Fig. 4).
When the pop effect in general linear models of the
two populations was replaced with one of the two ice
metrics, most sex and age classes had reduced body
condition (body mass, skull width, and energy density)
when the number of reduced ice days or the distance to
sea ice in September increased (Table S4). However,
subadults did not exhibit a relationship with reduced
ice days; and mass of subadult males and females and
skull width of subadult females was larger when there
was a greater distance to ice in September. Yearlings
per female and yearling litter size were lower when ice
conditions were reduced.

Comparisons of body condition with other populations
Comparisons between the SB and CS populations (2008–
2011)
Overall, CS bears during 2008–2011 were larger and in
better condition than SB bears during the same period.
Comparing growth curves for body mass and skull
width, females ≥ 1 year old weighed 30.2 kg more in
the CS than the SB (Fig. 3a: v2 = 24.4, P < 0.0001) and
had skull widths 0.8 cm larger (v2 = 15.0, P < 0.0001).
Similarly, body mass of CS males ≥1 year old was
48.5 kg greater (Fig. 3b: v2 = 58.1, P < 0.0001) and their
© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

Scale-measured and calculated body masses and measured skull widths of CS polar bears during 2008–2011
were among the highest reported for polar bear populations (Fig. 5). Similarly, energy density of adult males
(23.0 MJ kgÀ1) was higher than that in any population
for which data were available (≤22.8 MJ kgÀ1) and only
adult females in Foxe Basin (22.1 MJ kgÀ1) and the
Central Arctic (21.7 MJ kgÀ1) exhibited higher energy
densities
than
CS
bears
during
2008–2011
(19.6 MJ kgÀ1; Table S5).

 82 K . D . R O D E et al.
Table 1 Comparisons between the morphometric measures of polar bears captured in the Chukchi Sea 1986–1994 (CS1986–94) and
2008–2011(CS2008–11) and in the CS during 2008–2011 and the southern Beaufort Sea 2008–2011 (SB2008–11) using general linear models. Comparisons of body mass, skull width, and body length (cm) for adults and subadults were made using residuals of the difference between predicted measures from a fit modified von Bertalanffy growth curve and observed measures. Energy density
(MJ kgÀ1) was determined from Moln ar et al., (2009). Covariates, including capture date, litter size for yearlings, and cub age for
females, were included in models when P < 0.10. A list of models, test statistics, and P-values are provided in Table S3 and S4.

Sex and
age class
Yearling
Females
Yearling Males
Subadult Females
Subadult Males
Adult females
without coy

Adult Males

Measure

CS1986–94

Mass
Skull width
Mass
Skull width
Mass
Skull width
Mass
Skull width
Mass
Skull width
Length
Energy D
Mass
Skull width
Length
Energy D

80.5
14.0
101.7
15.6
156.1
17.6
213.2
19.4
215.1
20.6
188.3
20.6
437.1
24.5
225.4
15.4

Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ

20.3 (12)
1.2 (19)
17.5 (17)
1.0 (19)
26.5 (9)
1.2 (11)
35.2 (6)
1.3 (9)
36.5 (52)
1.4 (66)
9.5 (57)
5.3 (47)
43.5 (3)*
2.3 (6)
7.6 (4)*
1.3 (3)*

CS2008–11
99.7
15.4
132.8
17. 0
151.4
18.2
221.4
20.0
228.2
21.4
197.3
17.9
423.1
25.9
227.5
14.9

Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ

SB2008–11

22.4 (12)
1.0 (14)
29.7 (23)
1.1 (24)
21.2 (15)
1.0 (15)
42.0 (34)
1.2 (34)
39.6 (51)
1.6 (53)
11.3 (49)
4.1 (47)
97.9 (60)
2.7 (65)
11.7 (59)
2.6 (57)

81.5
14.6
93.3
15.7
152.6
17.6
213.4
19.9
194.1
20.3
197.4
13.9
374.4
24.8
228.0
12.9

Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ
Æ

17.1 (12)
0.8 (13)
20.7 (18)
0.9 (20)
33.3 (15)
1.3 (15)
43.6 (42)
1.2 (41)
31.8 (70)
1.1 (82)
8.5 (77)
3.9 (60)
75.5 (99)
2.0 (101)
13.0 (101)
2.4 (97)

CS time effects
CS2008–11 >
CS1986–94?

Population
differences
CS2008–11 >
SB2008–11?

Time

Ice effect?

Pop

Ice effect?

+
+
+
+
0
0
0
+
+
+
+
0
NA
0
NA
NA

0
+
+
0
+
+
0
+
+
0
NA
0
NA
+
NA
NA

+
+
+
+
0
+
+
+
+
+
0
+
+
+
No
+

À
À
À
À
+
+
+
0
À
À
NA
À
À
À
NA
0

*Adult males were not targeted in CS captures 1986–1994; thus sample sizes for most time comparisons were insufficient

Table 2 Comparison of spring reproductive measures for bears in the Chukchi and Bering (CS; this study) and southern Beaufort
Seas (SB; values reported by Regehr et al. 2006 and from this study). Sample sizes in parentheses

Reproductive rate
(yearlings per female)
% females with yearlings *
% females with 2-year olds
Cub-of-the-year litter size
Yearling litter size

CS

CS

SB

SB

SB

1986–1994†

2008–2011

1967–1989

1990–2006

2008–2011

0.62 (58)
38.5
20
1.85 Æ 0.57 (85)
1.64 Æ 0.49 (22)

0.66 (53)

0.46 (135)

41.5
22.6
1.90, 2.17‡ (39, 24)
1.59 Æ 0.67 (22)

26

1.54

24
9
1.34

30.6
7.1
1.79 Æ 0.52 (57)
1.38 Æ 0.58 (45)

*Values for the southern Beaufort Sea were corrected to exclude females with COY as the Chukchi and Bering Seas sample is biased
against sampling of females with COY.
†Measures exclude females captured during the spring near Wrangel Island as the Wrangel Island sample was biased toward
denning females
‡Estimated by Ovsyanikov and Menyushina (2010) for 2007 and 2009 on Wrangel Island

Diet composition and fasting behavior
Ringed seal was consumed by all independent bears
(n = 182) and contributed more than any other prey to
CS bear diets during 2008–2011 (79 Æ 18% of dietary

biomass; Table 3). Most bears consumed some bearded
seal (74 Æ 12% of bears) and bowhead whale
(85 Æ 11% of bears), but these species contributed only
12% and 6% of dietary biomass, respectively. Walrus
(41 Æ 12% of bears) and beluga whale (7 Æ 5% of
© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

 P O L A R B E A R R E S P O N S E S T O C L I M A T E C H A N G E 83
Table 3 Mean (Æ SD) contribution of prey items to diets of CS polar bears captured in the Chukchi and Bering Seas during the
spring based on fatty acid composition of fat biopsies. Data represent the% contribution to polar bear fatty acid profiles and thus
reflect the relative contribution to polar bear diets on a biomass basis. Subadults include bears aged 2–4 years
Bearded Seal
Adult Females (55)
Adult Males (61)
Subadult Females (13)
Subadult Males (25)
MEAN

6.5
20.7
2.6
6.8
11.9

Æ
Æ
Æ
Æ
Æ

9.2
17.3
3.5
7.5
14.5

Beluga Whale

Bowhead Whale

0.5 Æ 3.5
1.4 Æ 4.5
0.3 Æ 1.0
1.8 Æ 6.9
1 Æ 4.5

5.1
7.7
3.9
4.8
6.1

(a)

Æ
Æ
Æ
Æ
Æ

6.0
7.4
4.0
5.7
6.8

Ringed Seal
87.4
65.7
93.2
86.4
78.5

Æ
Æ
Æ
Æ
Æ

11.2
17.3
3.9
10.8
18.5

Walrus
0.6
4.4
0.0
0.2
2.4

Æ
Æ
Æ
Æ
Æ

1.4
6.0
0.1
0.5
6.7

(a)

(b)

(b)

Fig. 4 Relationships between maternal body mass and litter
mass of (a) COY and (b) yearlings for polar bears captured in
the spring in the Chukchi Sea (1986–1994 and 2008–2011) southern Beaufort Sea (2008–2011).

bears) were consumed less frequently and contributed
only 2% and 1% of dietary biomass, respectively.
The contribution of bowhead whale to dietary biomass of CS bears was 4% greater for subadults than
adults (generalized linear model: v2 = 6.8, P = 0.009),
3% greater for females than males (v2 = 3.7, P = 0.055),
and increased with body mass within age and sex classes (b = 0.0003, v2 = 11.7, P = 0.001). The contribution
of bearded seal to dietary biomass exhibited the opposite relationship being 12% greater in adults than
subadults and 12% greater in males than females.
© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

Fig. 5 Comparison of (a) asymptotic skull width and (b) body
mass of male and female polar bears measured in this study
(Chukchi Sea: CS 1986–1994, 2008–2011; southern Beaufort Sea:
SB 2008–2011) with reported values for other populations (Derocher and Stirling 1998, Derocher and Wiig 2002; Rode et al. in
press). Asymptotes were determined from fitting a modified
von Bertalanffy growth curve. Calculated and scale body mass
are shown for bears captured in the Chukchi and southern
Beaufort Seas. Central Arctic includes portions of the Gulf of
Boothia and M’Clintock Channel polar bear populations and the
High Arctic includes the Lancaster Sound polar bear population.

 84 K . D . R O D E et al.
Walrus contributed twice as much to the dietary biomass of males (3.2%) than females (1.6%; v2 = 21.2,
P < 0.0001). Beluga whale similarly contributed more
to the dietary biomass of males (1.7%) than females
(0.4%: v2 = 10.2, P = 0.006). In contrast, the contribution
of ringed seal to dietary biomass of CS bears did not
differ between sex or age classes, but decreased with
increasing
body
mass
(b = À0.001,
v2 = 65.7,
P < 0.0001). No bears consumed ribbon seals and only
two bears consumed spotted seals.
Thirteen of the 150 CS bears (annual mean
8.0 Æ 4.0%) during 2008–2011 had urea:creatinine ratios
(U:C) indicative of fasting (≤10.0). Urea levels were
lower (6.3 Æ 1.1 mg dlÀ1; F1,148 = 16.9, P < 0.0001) and
creatinine levels were higher (1.8 Æ 0.3 mg dlÀ1;
F1,148 = 34.3, P < 0.0001) for bears that had U:C ratios
≤10.0 vs. bears with U:C ratios > 10.0 (urea:
21.0 Æ 12.7 mg dlÀ1; creatinine: 1.0 Æ 0.4 mg dlÀ1), a
pattern consistent with fasting. All fasting bears were
adult males except one and no bears were assigned a
body condition score of 1 that would suggest protein
catabolism and an elevated U:C ratio while fasting.

Discussion
Our comparative analysis of ecological indicators for
CS and SB polar bears, in relation to sea ice and other
factors provides insights into spatial and temporal variation in the effects of diminishing ice cover. Body size,
condition, and reproductive indices of CS polar bears
did not decline over time between 1986–1994 and 2008–
2011 despite a 44-day increase in the number of
reduced ice days. Furthermore, CS bears were larger, in
better condition, and appeared to have higher recruitment compared to the adjacent SB population during
2008–2011. These differences were biologically significant and could have important implications for population dynamics and resiliency to continued habitat loss:
especially in the short term. For example, yearling
females weighed over 18 kg more in the CS than in the
SB, and adult females weighed over 30 kg more. Relationships between female body mass and yearling
mass, and apparent higher recruitment in the CS, suggest that differences in body condition between populations are associated with observed differences in
recruitment.
Differences in the rate and timing of sea ice loss in
the CS and SB may explain the apparent variation in
the response of polar bears. Although the rate of
decline in sea ice habitat was twice as high in the CS
(À8.0% per decade) than the SB (À4.8% per decade)
(Durner et al., 2009), there were twice as many reduced
ice days per year in the SB (mean 90 days) than in the
CS (mean 45 days) during 2007–2010. Cumulatively,

there were 160 more reduced ice days in the SB than
the CS during 2008–2012, although comparisons were
not statistically significant due to high variability. In
addition, the SB has experienced an increase in the
number of reduced ice days since the mid-1990s,
whereas this has only recently begun to occur in the CS
(Fig. 2).
Differences between populations fit expected relationships of decreased body condition and recruitment
when or where sea ice habitat is reduced. In the CS,
however, body condition was maintained or improved
when sea ice declined suggesting that factors other than
the presence of sea ice affect temporal and geographic
variation in condition and recruitment. Primary productivity in the CS (250–300 g C mÀ2 yrÀ1; Walsh et al.,
1989) was 10 times higher than the Beaufort Sea
(10–25 g C mÀ2 yrÀ1; Alexander, 1974; Horner, 1984)
and among the highest of any world ocean (Grebmeier
et al., 2006). Climate warming has not only led to sea
ice loss in the CS but also resulted in warmer waters
and increases in primary productivity (Zhang et al.,
2010). Evidence of high biological productivity is
apparent for polar bear prey species in the CS. Condition and reproduction of ringed seals and bearded
seals, which comprise almost 80% of polar bear diets in
the CS, have either increased or remained stable since
the 1970s (Quakenbush et al., 2011a, b). Pregnancy rates
among ringed seals in the CS were the second highest
recorded and females are maturing at the youngest age
reported. Ringed seal pups are one of the most important food sources for polar bears (Pilfold et al., 2012)
and high proportions of ringed seal pups in subsistence
harvest samples in the CS suggest they are abundant
(Quakenbush et al. 2011a). Growth rates, condition, and
pregnancy rates of CS bearded seals were also average
or above average (Quakenbush et al., 2011a). In contrast, ringed seal body condition in Amundsen Gulf in
the eastern Beaufort Sea has declined in recent years
(Harwood et al., 2012). While continued high biological
productivity in the Chukchi and northern Bering seas
region may be allowing polar bears and their prey to
prosper despite habitat loss, declines in the condition of
SB bears may be a combination of reduced access to
prey resulting from declining ice habitat (Rode et al.,
2010; this study), declining prey body condition (Harwood et al. 2012), and the much smaller area of productive, shallow continental shelf.
Our evaluation of nutritional ecology for polar bears
is consistent with lower prey availability in the SB compared to the CS. Only 8% of polar bears captured in the
CS appeared to be fasting in spring, compared to 21.4
and 29.3% of bears fasting in the SB in spring 2005 and
2006, respectively (Cherry et al., 2009). In the CS, the
majority of fasting bears were adult males (92% of
© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

 P O L A R B E A R R E S P O N S E S T O C L I M A T E C H A N G E 85
fasting bears) who may forgo foraging to pursue
females during the mating season (Ramsay et al., 1991;
Cherry et al., 2009). The SB, in contrast, had 20–25% of
adult females (including separate categories of females
with and without dependent young) and >15% of subadults fasting.
Consumption of larger prey species or carrion or a
greater diversity of prey does not appear to explain
the difference in body size and condition observed
between CS and SB bears. Diets consisted of 80–90%
ringed and bearded seals and were relatively similar
between the two populations (SB data from Thiemann
et al., 2008 and CS data from this study). Although CS
polar bears were larger and in better condition, their
diet was composed more of the smallest prey species,
ringed seals (78.5%; 50–70 kg), and less of the larger
bearded seal (12%; 200–430 kg), than bears in the adjacent SB (60% ringed seal, 20% bearded seal; Fig 7b. in
Thiemann et al., 2008). CS polar bears did consume
walrus (2.4% of dietary biomass), which rarely occur
in the SB and could provide large amounts of food
from a single kill or carcass (800–1680 kg). An
increased availability of walrus carcasses associated
with summertime sea ice loss (Fischbach et al., 2009)
could be an important food resource for CS bears. Furthermore, summer observations on Wrangel Island
suggest that walrus are frequently successfully hunted
and consumed as carrion (Kochnev 2002). However,
such feeding may not be reflected fully in this study
because diets were based on spring-collected fat biopsies and turnover time of fatty acids in polar bears is
not known.
Our spring 2008–2011 observations of CS polar bears
in good condition and with high recruitment are consistent with autumn observations of bears from this
population on Wrangel Island during the same period
(Ovsyanikov and Menyushina, 2010) and with comparisons to other populations (Table S5). Between 100 and
400 polar bears use Wrangel Island as a summer refugia (Ovsyanikov 2012) and the majority of the CS population dens on Wrangel Island or nearby (70 km east)
smaller (11.3 km2) Herald Island (Garner et al. 1994;
US Fish and Wildlife Service, unpublished. data).
Based on a condition index (Stirling et al., 2008) only
3% of the bears that came onshore on Wrangel Island
in autumn of 2005–2009 appeared undernourished
(condition index of 1–2), 30% were categorized as normal condition (condition index of 3), and 67% were
categorized as fat (condition index of 4–5; Ovsyanikov
and Menyushina, 2010). Skull width, body mass, and
energy density of polar bears caught in the CS during
2008–2011 were also among the largest of any polar
bear population that has been measured to date
(Fig. 5).
© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

Spring COY litter sizes reported on Wrangel Island
between 2007 and 2009 were 1.90 (n = 39) and 2.17
(n = 24; Ovsyanikov and Menyushina, 2010), respectively, which is similar to litter sizes 20 years earlier
(late 1980s through early 1990s, 1.81–1.85; Vaisfeld and
Chestin, 1993; this study). These spring COY litter sizes
are among the highest reported for 18 of 19 polar bear
populations, which range from 1.50 to 2.27 for bears at
the den site and 1.30–1.94 for bears that had moved
away from dens in spring (Derocher, 1999). Similarly,
autumn litter sizes between 2004 and 2010 were
1.59 Æ 0.58 for COY (n = 170; Ovsyanikov, 2012) which
are on the high end of those reported for other populations (1.49–1.58 for SB 1986–1994, Southern Hudson
Bay, Western Hudson Bay, and Davis Strait; Regehr
et al., 2006; Peacock et al. 2013; Obbard et al., 2010).
Spring litter sizes of CS yearlings from this study (1.59)
were also higher than those in other populations
(SB: 1.34–1.54; Amstrup, 1995; Regehr et al., 2006; Svalbard: 1.40, Wiig, 1998; 1.52, Derocher, 2005; Baffin
Island: 1.57; Peacock et al. 2013). The larger body mass
of adult females in the CS corresponded not only with
larger litter sizes but also with heavier yearlings
(Fig. 5), which have a greater chance of survival (Derocher and Stirling, 1996).
Our evaluation of ecological indicators for the CS
population is based on 4 years of recent data in a
highly variable and changing environment. Continued
monitoring is needed to determine whether condition
and reproduction will be maintained with projected
continued loss of sea ice. The CS shelf is projected to be
largely ice-free in September and October by mid-century (Douglas, 2010; Wang et al., 2012), and the duration of the ice-free period is projected to reach
4 months by the end of the century (Douglas 2010).
Relationships between ice availability, reproductive
indices, and cub size (Stirling et al., 1999; Rode et al.,
2010; this study) suggest that long-term continuous
decline in sea ice might first affect body condition and
reproduction (Stirling and Derocher, 2012) making
these important metrics to be included in long-term
monitoring. Continued sea ice loss could not only create challenges for bears to access prey but also to reach
key denning habitats on Wrangel and Herald islands
and the Chukotkan coast where the majority of CS
pregnant females den. For example, changes in sea ice
formation resulted in reduced use and periodic loss of
a denning area in Svalbard (Derocher et al., 2011).
Recent observations on Wrangel Island suggest a
decline in the number of denning females and low
autumn yearling litter sizes of 1.31 Æ 0.54 (n = 55)
between 2004 and 2010, and suggests that cub survival
rates have been lower in recent years (Ovsyanikov,
2012). Sea ice loss may result in other mechanisms that

 86 K . D . R O D E et al.
result in direct or indirect mortality such as drowning
or increased durations spent swimming (Monnett and
Gleason, 2006; Pagano et al., 2012). Furthermore, the
ecological indicators considered in this study do not
reflect potential effects of human-caused mortality,
which could affect population growth but would be
unlikely to affect body condition and reproductive
indices.
Our findings of geographic, temporal, and other ecological variation among polar bears experiencing sea
ice loss, emphasize the challenge of developing general
—yet accurate—population projections, over the short
term. In our relatively short study, declines in sea ice
extent alone did not completely explain variation in
polar bear population productivity. This suggests that
polar bears may exhibit complex and nonlinear
responses to climate change, especially in the short
term. The amount of available habitat, the status of prey
populations, and ecosystem productivity likely all play
important roles in determining the timing and magnitude of polar bear responses to sea ice loss. Parallel
declines in the condition of polar bears (Regehr et al.,
2006; Rode et al., 2010) and ringed seals (Harwood
et al., 2012) in the SB signal the potential for negative
trophic effects that could additively affect polar bears
in this region. This contrasts to maintenance of body
condition and reproduction of polar bears and their primary prey (Quakenbush et al., 2011a, 2011b) in the CS.
Such regional covariation in the apparent response of
prey and predator suggest that ecosystem-based projections (in addition to range-wide projections) may be a
fruitful path of future research and conservation efforts
for individual species.

Acknowledgements
US-based polar bear captures were funded by the US Fish and
Wildlife Service (FWS), the US Geological Survey (USGS) (Ecosystems and Climate and Land Use Change Mission Areas),
National Science Foundation grant (OPP 0732713) to the University of Wyoming, and the US Bureau of Land Management. In
Canada, captures were funded by ArcticNet, US Bureau of
Ocean Energy Management, Environment Canada, Natural Sciences and Engineering Research Council of Canada, the Polar
Continental Shelf Project, Quark Expeditions, and World Wildlife Fund (Canada). This research was permitted under the Marine Mammal Protection Act and Endangered Species Act under
FWS permit MA046081 and USGS permit MA 690038 and followed protocols approved by Animal Care and Use Committees
of the FWS, USGS (assurance no. 2010–3), and the University of
Alberta. G. Garner (deceased) led capture efforts in the CS in
the 1980s and 1990s and S. Amstrup led the long-term capture
program in the SB that provided data for this study. E. Regehr
and K. Rode let capture efforts in the CS in the 2000s. In addition,, C. Gardner, S. Arthur, G. York, C. Perham, S. Miller, M. St.
Martin, J. Wilder, T. DeBruyn, M. Lockhart, N. Ovsyanikov, K.
Simac, A. Pagano, L. Peacock, J. Bromaghin, S. Cherry, A.

McCall, M. Auger-M eth e, J. Pongracz, S. Hamilton, and N. Pilfold assisted with data collection. R. Donaldson, J. Carie,
C. Stansbury, T. Evans and K. Knott assisted with sample processing. The community of Point Hope, the Selawik National
Wildlife Refuge, Red Dog mine, Teck Alaska Inc, and Nana provided significant field support. The USGS thanks the Arctic
National Wildlife Refuge, and the North Slope Borough Department of Wildlife Management for logistical support. We thank
our excellent pilots for ensuring safe capture seasons. C. Jay, T.
Fischbach, S. Iverson, and L. Quakenbush provided data or
samples. C.T. Robbins, J. Bromaghin, T. Atwood, S. Amstrup,
M. St. Martin, and anonymous reviewers provided comments
on manuscript drafts. Use of trade names does not imply
endorsement by the US Government. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the FWS.

References
Aars J, Lunn NJ, Derocher AE (2006) Polar bears: Proceedings of the 14th working meeting
of the IUCN/SSC Polar Bear Specialist Group, 20-24 June 2005, Seattle, Washington,
USA. Occasional Paper of the IUCN Species Survival Commission No. 32. IUCN, Gland,
Switzerland and Cambridge, UK.
Alexander V (1974) Primary productivity regimes of the nearshore Beaufort Sea with
reference to potential roles of ice biota. In: The Coast and Shelf of the Beaufort Sea
(eds Reed JC, Sater JE), pp. 609–632. Arlington, Arctic Institute of North America.
Amstrup SC (1995) Movements, distribution, and population dynamics of polar bears in the
Beaufort Sea. Ph.D. dissertation. University of Alaska, Fairbanks, Alaska, USA.
Amstrup SC, Marcot BG, Douglas DC (2008) A Bayesian network modeling approach
to forecasting the 21st century worldwide status of polar bears. In: Arctic Sea Ice
Decline: Observations, Projections, Mechanisms, and Implications (eds.DeWeaver ET,
Bitz CM, Tremblay LB), pp. 213–268 Geophysical Monograph 180. American Geophysical Union, Washington, DC.
Cavalieri D, Parkinson C, Gloersen P, Zwally HJ (1996) Sea ice concentration from Nimbus-7 SMMR and DMSP SSM/I-SSMIS Passive Microwave Data (1979-2010). National
Snow and Ice Data Center, Digital media, Boulder, Colorado, USA.
Cherry SG, Derocher AE, Stirling I, Richardson ES (2009) Fasting physiology of polar
bears in relation to environmental change and breeding behavior in the Beaufort
Sea. Polar Biology, 32, 383–391.
Cherry SG, Derocher AE, Hobson KA, Stirling I, Thiemann GW (2011) Quantifying
dietary pathways of proteins and lipids to tissues of a marine predator. Journal of
Applied Ecology, 48, 373–381.
Derocher AE, Nelson RA, Stirling I, Ramsay MA (1990) Effects of fasting and feeding
on serum urea and serum creatinine levels in polar bears. Marine Mammal Science,
6, 196–203.
Derocher AE (1991) Population dynamics and ecology of polar bears in western Hudson
Bay. Edmonton, University of Alberta, PhD dissertation.
Derocher AE (1999) Latitudinal variation in litter size of polar bears: ecology or methodology? Polar Biology, 22, 350–356.
Derocher AE (2005) Population ecology of polar bears at Svalbard, Norway. Population Ecology, 47, 267–275.
Derocher AE, Andersen M, Wiig Ø, Aars J, Hansen E, Biuw M (2011) Sea ice and
polar bear den ecology at Hopen Island, Svalbard. Marine Ecology Progress Series,
441, 273–279.
Derocher AE, Stirling I (1996) Aspects of survival in juvenile polar bears. Canadian
Journal of Zoology, 74, 1246–1252.
Derocher AE, Stirling I (1998) Geographic variation in growth of polar bears (Ursus
maritimus). Journal of Zoology, London, 245, 65–72.
Derocher AE, Wiig Ø (2002) Postnatal growth in body length and mass of polar bears
(Ursus maritimus) at Svalbard. Journal of Zoology, 256, 343–349.
Douglas DC (2010) Arctic sea ice decline: projected changes in timing and extent of
sea ice in the Bering and Chukchi Seas. U.S. Geological Survey Open-File Report
2010-1176, 32 p.
Durner GM, Amstrup SC, Neilson R, McDonald T (2004) The use of sea ice habitat by
female polar bears in the Beaufort Sea. U.S. Geological Survey, Alaska Science Center,
Anchorage, Alaska. OCS study, MMS 2004-014.
Durner GM, Douglas DC, Nielson RM, Amstrup SC(2006) A model for autumn
pelagic distribution of adult female polar bears in the Chukchi Sea, 1987–1994. USGS

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

 P O L A R B E A R R E S P O N S E S T O C L I M A T E C H A N G E 87
Alaska Science Center, Anchorage, Contract Completion Report 70181-5-N240,
67 pp.

Ovsyanikov NG, Menyushina IE (2010) Number, condition, and activity of polar
bears on Wrangel Island during ice free autumn seasons of 2005-2009. Proceedings

Durner GM, Douglas CD, Nielson RM, et al. (2009) Predicting 21st-century polar
bear habitat distribution from global climate models. Ecological Monographs, 79,
25–58.
Fischbach AS, Monson DH, Jay CV (2009) Enumeration of Pacific walrus carcasses on
beaches of the Chukchi Sea in Alaska following a mortality event, September 2009.
US Geological Survey Open-File Report 1291, 10 p.
Frost KJ, Lowry LF, Pendleton G, Nute HR (2002) Monitoring distribution and abun-

of the Marine Mammals of the Holarctic 2010. Kalinigrad, Russia.
Ovsyanikov NG (2012) Occurrence of family groups and litter size of polar bears on
Wrangel Island in the autumns of 2004-2010 as an indication of population status.
Proceedings of the Marine Mammals of the Holarctic 2013. Suzdal, Russia.
Pagano AM, Durner GM, Amstrup SC, Simac KS, York GS (2012) Long-distance
swimming by polar bears (Ursus maritimus) of the southern Beaufort Sea during
years of extensive open water. Canadian Journal of Zoology, 90, 663–676.

dance of ringed seals in northern Alaska. OCS Study MMS 2002-043. Final report from
the Alaska Department of Fish and Game, Juneau, AK, for US Minerals Management Service, Anchorage, AK. 66 p + Appendices.
Gagnon AS, Gough WA (2005) Trends in the dates of ice freeze-up and breakup over
Hudson Bay, Canada. Arctic, 58, 370–382.
Gardner JL, Peters A, Kearney MR, Joseph L, Heinsohn R (2011) Declining body size:

Parmesan C (2006) Ecological and evolutionary responses to recent climate change.
Annual Review in Ecology and Evolutionary Systematics, 37, 637–669.
Peacock E, Laake J, Laidre KL, Born EW, Atkinson SN (2012) The utility of harvest
recoveries of marked individuals to assess polar bear (Ursus maritimus) survival.
Arctic, 65, 391–400.
Peacock E, Taylor MK, Laake J, Stirling I (2013) Population ecology of polar bears in

a third universal response to warming? Trends in Ecology and Evolution, 26, 285–
291.
Garner GW, Belikov SE, Stishov MS, Barnes VG, Arthur SM (1994) Dispersal patterns
of maternal polar bears from the denning concentration on Wrangel Island. International Conference on Bear Research and Management, 9, 401–410.
Gittleman JL (1985) Carnivore body size: ecological and taxonomic correlates. Oecologia, 67, 540–554.

Davis Strait, Canada and Greenland. Journal of Wildlife Management, 77, 463–476.
Pilfold NW, Derocher AE, Stirling I, Richardson E, Andriashek D (2012) Age and sex
composition of seals killed by polar bears in the eastern Beaufort Sea. PLoS ONE,
7, e41429. doi:10.1371/journal.pone.0041429.
Quakenbush L, Citta J, Crawford J (2011a) Biology of the bearded seal (Erignathus barbatus) in Alaska, 1961–2009. Report to the National Marine Fisheries Service, March.
Quakenbush L, Citta J, Crawford J (2011b) Biology of the ringed seal (Phoca hispida) in

Grebmeier JM, Cooper LW, Feder HM, Sirenko BI (2006) Ecosystem dynamics of the
Pacific-influenced Northern Bering and Chukchi Seas in the Amerasian Arctic.
Progress in Oceanography, 71, 331–361.
Harwood LA, Smith TG, Melling H, Alikamik J, Kingsley MCS (2012) Ringed seals
and sea ice in Canada’s Western Arctic: harvest-based monitoring 1992–2011. Arctic, 65, 377–390.

Alaska, 1960–2010. Report to the National Marine Fisheries Service, March.
Ramsay MA, Nelson RA, Stirling I (1991) Seasonal changes in the ratio of serum urea
to creatinine in feeding and fasting polar bears. Canadian Journal of Zoology, 69,
298–302.
Regehr EV, Amstrup SC, Stirling I (2006) Polar bear population status in the southern
Beaufort Sea. U.S. Geological Survey Open-File Report 2006-1337, U.S. Geological

Hilderbrand GV, Schwartz CC, Robbins CT, Hanley TA (1999) The importance of
meat, particularly salmon, to body size, population productivity, and conservation
of North American brown bears. Canadian Journal of Zoology, 77, 132–138.
Holland MM, Bitz CM, Tremblay B (2006) Future abrupt reductions in the summer
Arctic sea ice. Geophysical Research Letters, 33, L23503.
Horner R (1984) Phytoplankton abundance, chlorophyll a, and primary productivity
in the western Beaufort Sea. In: The Alaskan Beaufort Sea: ecosystems and environments

Survey, Alaska Science Center, Anchorage, AK, USA.
Regehr EV, Lunn NJ, Amstrup SC, Stirling I (2007) Effects of earlier sea ice breakup
on survival and population size of polar bears in western Hudson Bay. Journal of
Wildlife Management, 71, 2673–2683.
Regehr EV, Hunter CM, Caswell H, Amstrup SC, Stirling I (2009) Survival and breeding of polar bears in the southern Beaufort Sea in relation to sea ice. Journal of Animal Ecology, 79, 117–127.

(eds Barnes PW, Schell D, Reimnitz E), pp. 295–310. Academic Press, Orlando.
Kochnev A (2002) Autumn aggregations of polar bears on Wrangel Island and their
importance to the population. Proceedings of the Marine Mammals of the Holarctic 2002. Moscow, Russia.
Kolenosky GB, Lunn NJ, Greenwood CJ, Abraham KF (1989) Estimating the weight of
polar bears from body measurements. Journal of Wildlife Management, 53, 188–190.

Rigor IG, Wallace JM (2004) Variations in the age of Arctic sea-ice and summer seaice extent. Geophysical Research Letters, 31, L09401.
Rode KD, Amstrup SC, Regehr EV (2010) Reduced body size and cub recruitment in
polar bears associated with sea ice decline. Ecological Applications, 20, 768–782.
Rode KD, Peacock E, Taylor M, Stirling I, Born EW, Laidre KL, Wiig Ø (2012) A tale
of two polar bear populations: ice habitat, harvest, and body condition. Population

Laidre KL, Estes JA, Tinker MT, Bodkin J, Monson D, Schneider K (2006) Patterns of
growth and body condition in sea otters from the Aleutian archipelago before and
after the recent population decline. Journal of Animal Ecology, 75, 978–989.
Laidre KL, Stirling I, Lowry LF, Wiig Ø, Heide-Jorgensen MP, Ferguson SH (2008)
Quantifying the sensitivity of arctic marine mammals to climate-induced habitat
change. Ecological Applications, 18, S97–S125.
Maslanik J, Stroeve J, Fowler C, Emery W (2011) Distribution and trends in Arctic sea

Ecology, 54, 3–18.
Sakshaug E (2004) Primary and secondary production in the Arctic Seas. In: The
Organic Carbon Cycle in the Arctic Ocean (eds Stein R, MacDonald RW), pp. 57–81.
Springer, Berlin.
Schliebe S, Rode KD, Gleason JS, Wilder J, Proffitt K, Evans TJ, Miller S (2008) Effects
of sea ice extent and food availability on spatial and temporal distribution of polar
bears during the fall open-water period in the Southern Beaufort Sea. Polar Biology,

ice age through spring 2011. Geophysical Research Letters, 38, L13502.
Moln ar PK, Klanjscek T, Derocher AE, Obbard ME, Lewis MA (2009) A body composition model to estimate mammalian energy stores and metabolic rates from body
mass and body length, with application to polar bears. Journal of Experimental Biology, 212, 2313–2323.
Moln ar PK, Derocher AE, Klanjscek T, Lewis MA (2010) Predicting climate change

31, 999–1010.
Sheridan JA, Bickford D (2011) Shrinking body size as an ecological response to climate change. Nature Climate Change, 1, 401–406.
Sigler MF, Renner M, Danielson SL, Eisner LB, Lauth RR, Kuletz KJ, Logerwell EA,
Hunt GL Jr (2011) Fluxes, fins, and feathers. Relationships among Bering, Chukchi,
and Beaufort seas in a time of climate change. Oceanography, 24, 112–127.

impacts on polar bear litter size. Nature Communications, 2, 186.
Monnett C, Gleason JS (2006) Observations of mortality associated with extended
open-water swimming by polar bears in the Alaskan Beaufort Sea. Polar Biology,
29, 681–687.
Noyce KV, Garshelis DL (1994) Body size and blood characteristics as indicators of
condition and reproductive performance in black bears. International Conference on
Bear Research and Management, 9, 481–496.

Stirling I (2002) Polar bears and seals in the eastern Beaufort Sea and Amundsen Gulf:
a synthesis of population trends and ecological relationships over three decades.
Arctic, Supp 1 59–76.
Stirling I, Spencer C, Andriashek D (1989) Immobilization of polar bears (Ursus maritimus) with Telazolâ in the Canadian Arctic. Journal of Wildlife Diseases, 25, 159–168.
Stirling I, McDonald TL, Richardson ES, Regehr EV, Amstrup SC (2011) Polar bear
population status in the northern Beaufort Sea, Canada, 1971–2006. Ecological

Obbard ME, Cattet MRL, Moody T, Walton LR, Potter D, Inglis J, Chenier C (2006)
Temporal trends in the body condition of Southern Hudson Bay polar bears. Climate Change Research Information Note, 3, 1–8.
Obbard ME, McDonald TL, Howe EJ, Regehr EV, Richardson ES (2007) Polar bear population status in southern Hudson Bay, Canada. Administrative Report. US Geological
Survey, Anchorage, Alaska, USA.

Applications, 21, 859–876.
Stirling I, Derocher AE (2012) Effects of climate warming on polar bears: a review of
the evidence. Global Change Biology, 18, 2694–2706.
Stirling I, Parkinson CL (2006) Possible effects of climate warming on selected populations of polar bears (Ursus maritimus) in the Canadian Arctic. Arctic, 59, 261–275.
Stirling I, Lunn NJ, Iacozza J (1999) Long-term trends in the population ecology of polar

Obbard ME, Thiemann GW, Peacock E, DeBruyn TD (eds). (2010). Polar bears: Proceedings of the 15th working meeting of the IUCN/SSC Polar Bear Specialist Group, Copenhagen, Denmark, 29 June – 3 July 2009. IUNC, Gland, Switzerland and Cambridge,
UK. vii + 235 pp.

bears in Western Hudson Bay in relation to climatic change. Arctic, 52, 294–306.
Stirling I, Thiemann GW, Richardson E (2008) Quantitative support for a subjective
fatness index for immobilized polar bears. Journal of Wildlife Management, 72,
568–574.

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88

 88 K . D . R O D E et al.
Stroeve JC, Kattsov V, Barrett A, Serreze M, Pavlova T, Holland M, Meier WN (2012)
Trends in Arctic sea ice extent from CMIP5, CMIP3 and observations. Geophysical

Wang M, Overland JE, Stabeno P (2012) Future climate of the Bering and Chukchi
Seas projected by global climate models. Deep-Sea Research II, 65–70, 46–57.

Research Letters, 39, L16502. doi:10.1029/2012GL052676, 2012.
Thiemann GW, Iverson SJ, Stirling I (2006) Seasonal, sexual, and anatomical variability in the adipose tissue of polar bears (Ursus maritimus). Journal of Zoology, 269,
65–76.
Thiemann GW, Iverson SJ, Stirling I (2008) Polar bear diets and arctic marine food
webs: insights from fatty acid analysis. Ecological Monographs, 78, 591–613.
Vaisfeld MA, Chestin IE (1993) Bears: brown bears, polar bears, Asian black bear. Distribu-

Wiig Ø (1998) Survival and reproductive rates for polar bears at Svalbard. Ursus, 10,
25–32.
Zedrosser A, Dahle B, Swenson JE (2006) Population density and food conditions
determine adult female body size in brown bears. Journal of Mammalogy, 87,
510–518.
Zhang J, Spitz YH, Steele M, Ashjian C, Campbell R, Berline L, Matrai P (2010) Modeling the impact of declining sea ice on the Arctic marine planktonic ecosystem.

tion, ecology, use and protection. Game animals of Russia and adjacent countries and their
environment. Nauka, Moscow.
Walsh JJ, McRoy CP, Coachman LK, et al. (1989) Carbon and nitrogen cycling within
the Bering/Chukchi Seas: source regions for organic matter effecting AOU
demands of the Arctic Ocean. Progress in Oceanography, 22, 277–359.

Journal of Geophysical Research, 115, 1–24.

Supporting Information
Additional Supporting Information may be found in the online version of this article:
Appendix S1. Detailed description of methods used to estimate diet and fasting behavior.
Appendix S2. Detailed description of statistical methods used to compare body condition and reproduction between the two polar
bear populations and periods.
Table S1. Sex and age class of bears captured in the Chukchi Sea during 1985–1994 and 2008–2011 and in the southern Beaufort Sea
during 2008–2011.
Table S2. Parameters of modified von Bertalanffy growth equations fit to total body length, skull width, and body mass of polar
bears captured in the Chukchi and southern Beaufort Seas during 2008–2011 and in the Chukchi Sea during 1986–1994.
Table S3. Results from comparisons of body size and condition for polar bears captured in the Chukchi Sea (CS) during 1986–1994
and 2008–2011 using general linear models.
Table S4. Results from comparisons of body size and condition for polar bears captured in the Chukchi Sea (CS) and the Southern
Beaufort Sea between 2008 and 2011 using general linear models.
Table S5. Estimated energy density (based on models by Moln 
ar et al., 2009) of male and female polar bears from various populations across the Arctic based on asymptotic body length and body mass from a modified von Bertalanffy growth curve.
Table S6. Contribution of prey items to diets of polar bears captured in the Chukchi and Bering Seas during the spring based on
fatty acid composition of fat biopsies. Data represent the % contribution to polar bear fatty acid profiles and thus reflect the relative
contribution to polar bear diets on a biomass basis. Subadults include bears aged 2–4 years.
Figure S1. Annual variation in the mean minimum distance between sea ice of 15% (a) and 50% (b) concentration and the continental shelf break (300 m isobath) between 1979 and 2010 in the Chukchi Sea and southern Beaufort Sea.
Figure S2. Annual variation in the number of reduced ice days defined as days in which there was less than 6250, 10 000, or
25 000 km2 of either 15% (a) or 50% (b) ice concentration over the continental shelf (<300 m ocean depth) in the Southern Beaufort
Sea and Chukchi Sea during 1979–2010.

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 76–88