Arizona State University LAB Seasons Tilt Orbit and Temperature Paper

PLEASE CLICK THE LINK BELOW TO ASSURE YOU ARE FAMILIAR WITH THE MATERIAL + THAT IT WORKS ON YOUR COMPUTER.

Ever wonder why we have seasons on earth? We will use

this interactive tool (Links to an external site.)Links to an external site.

to investigate how the tilt of the earth affects things like temperature, day length, and the path the sun takes through the sky.

Read the “Introduction” and “How To” tabs on the seasons interactive page.On the “Interactive” tab, watch a full orbit of the planet with the inclination set to the same as Earth. Record the maximum and minimum temperature (as given by the thermometer on the interactive – cold, cold 1/2, cool, mild, etc.) for Winter, Spring, Summer and Fall. Repeat this for inclination set to the same as Venus and then Uranus.Now pick an inclination other than 23, 2, or 86 degrees and repeat the process a fourth time.Turn in your observations of the relative temperatures for all four seasons for these 4 inclinations in a Word document (or similar) as part of your Lab 3 Report. You should include a data table summarizing the results of your 4 virtual experiments for the four seasons.You will discuss your analysis of these observations in Discussion Board 3 (below)

Rubric for Lab 3 Report (5 pts)

You should write a full lab report thatincludes: an “Introduction”, a “Materials and Methods”, “Results”,“Discussion” sections. This lab report should be a minimum of 300 words.Only the submission of a single Word document is required for this labreport. Please ensure that your document is either a .doc OR.docx file to ensure that we can easily access your file from both a PCand Mac computer.

Introduction (1 pt)

Describe what the term “tilt” means
Include characteristics of Earth, Venus, and Uranus

Include a hypothesis for the experiment

Materials and Methods (1 pt)(this should be written in past tense and in paragraph format. Writethis section as if someone else needs to complete the experiment solelyusing the information you’ve given here)

What materials do you need to complete this experiment?
What exact methods did you use to complete this experiment?

Results (2 pts)

Inclusion of data table that displays all collected data (1 pt)
Brief summary of results in paragraph format (1 pt)

Discussion (1 pt)

How did tilt impact minimum and maximum temperature
Relate these changes to day length and the path of sun through the sky
Was your hypothesis supported or rejected?

1 pt will be deducted if lab report doesnot meet word count requirements and up 1 pt may be deducted based onclarity of writing and grammar mistakes.

BELOW ARE RESOURCES INCLUDED IN THIS LESSON:

Sea Turtle Conservation and Climate Change Adaptation (Links to an external site.)Links to an external site.

Climate Change and Avian Population Ecology in Europe (Links to an external site.)Links to an external site. (French, Nature Education Knowledge 2011)

Predicting survival, reproduction and abundance of polar bears under climate change

Seminaland Endocrine Characteristics of Male Pallas’ Cats (Otocolobus manul)Maintained Under Artificial Lighting With Simulated Natural Photoperiods Biological Conservation 143 (2010) 1612–1622
Contents lists available at ScienceDirect
Biological Conservation
journal homepage: www.elsevier.com/locate/biocon
Predicting survival, reproduction and abundance of polar bears
under climate change
Péter K. Molnár a,b,*, Andrew E. Derocher b, Gregory W. Thiemann c, Mark A. Lewis a,b
a
Centre for Mathematical Biology, Department of Mathematical and Statistical Sciences, University of Alberta, Edmonton, AB, Canada T6G 2G1
Department of Biological Sciences, University of Alberta, Edmonton, AB, Canada T6G 2E9
c
Faculty of Environmental Studies, York University, Toronto, ON, Canada M3J 1P3
b
a r t i c l e
i n f o
Article history:
Received 12 January 2010
Received in revised form 1 April 2010
Accepted 7 April 2010
Keywords:
Climate change
Dynamic energy budgets
Ursus maritimus
Population viability analysis
Starvation
Mechanistic models
a b s t r a c t
Polar bear (Ursus maritimus) populations are predicted to be negatively affected by climate warming, but
the timeframe and manner in which change to polar bear populations will occur remains unclear. Predictions incorporating climate change effects are necessary for proactive population management, the setting
of optimal harvest quotas, and conservation status decisions. Such predictions are difﬁcult to obtain from
historic data directly because past and predicted environmental conditions differ substantially. Here, we
explore how models can be used to predict polar bear population responses under climate change. We
suggest the development of mechanistic models aimed at predicting reproduction and survival as a function of the environment. Such models can often be developed, parameterized, and tested under current
environmental conditions. Model predictions for reproduction and survival under future conditions could
then be input into demographic projection models to improve abundance predictions under climate
change. We illustrate the approach using two examples. First, using an individual-based dynamic energy
budget model, we estimate that 3–6% of adult males in Western Hudson Bay would die of starvation before
the end of a 120 day summer fasting period but 28–48% would die if climate warming increases the fasting
period to 180 days. Expected changes in survival are non-linear (sigmoid) as a function of fasting period
length. Second, we use an encounter rate model to predict changes in female mating probability under
sea ice area declines and declines in mate-searching efﬁciency due to habitat fragmentation. The model
predicts that mating success will decline non-linearly if searching efﬁciency declines faster than habitat
area, and increase non-linearly otherwise. Speciﬁcally for the Lancaster Sound population, we predict that
female mating success would decline from 99% to 91% if searching efﬁciency declined twice as fast as sea
ice area, and to 72% if searching efﬁciency declined four times as fast as area. Sea ice is a complex and
dynamic habitat that is rapidly changing. Failure to incorporate climate change effects into population
projections can result in ﬂawed conservation assessments and management decisions.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
Climate change effects on species and ecosystems have been
identiﬁed as critical problems for conservation biology (McCarty,
2001; Mawdsley et al., 2009). Describing, understanding, and
anticipating these effects are precursors to identifying mitigation
strategies (Harley et al., 2006; Root and Schneider, 2006). Anticipation can be particularly challenging and requires a combination of
good quantitative data along with precise hypotheses on the mechanisms by which climate change will affect a species (Ådahl et al.,
* Corresponding author at: Centre for Mathematical Biology, Department of
Mathematical and Statistical Sciences, University of Alberta, Edmonton, AB, Canada
T6G 2G1. Tel.: +1 780 492 6347; fax: +1 780 492 8373.
E-mail addresses: pmolnar@ualberta.ca (P.K. Molnár), derocher@ualberta.ca (A.E.
Derocher), thiemann@yorku.ca (G.W. Thiemann), mlewis@math.ualberta.ca (M.A.
Lewis).
0006-3207/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biocon.2010.04.004
2006; Krebs and Berteaux, 2006). Mathematical models can be a
powerful tool in this process, and they can inform research, monitoring, and conservation planning by indicating where and how
change in a population is most likely to occur. The type of projection model that can be applied depends to a large degree on how
similar predicted environmental conditions are to the ones observed. Berteaux et al. (2006) discuss constraints to projecting
the ecological effects of climate change, and they suggest a distinction between forecast and prediction models. Forecast models are
based on correlational relationships between explanatory and
dependent variables (e.g., environmental conditions and vital
rates) and are useful if there is no extrapolation beyond the
observed range of explanatory variables. In contrast, predictive
models mechanistically describe the cause-effect relationships
determining change (e.g., the link between environmental conditions and vital rates via energetic constraints), and can be used beyond the observed ranges.
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
The Arctic is warming faster than many other areas (IPCC,
2007), and habitat alteration is well underway. One Arctic habitat
showing profound effects is the sea ice, with the perennial and annual ice cover shrinking, and sea ice thickness decreasing (Comiso,
2002; Maslanik et al., 2007; Comiso et al., 2008). The sea ice is
declining at rates faster than expected (Stroeve et al., 2007), and
declines are projected to accelerate (Holland et al., 2006; Serreze
et al., 2007). Variability in predictive sea ice models exist but it is
possible that the Arctic Ocean will be ice-free in summer by the
middle to the end of the 21st century (Holland et al., 2006; Zhang
and Walsh, 2006; Serreze et al., 2007; Boé et al., 2009). Among the
most vulnerable to these warming trends are ice-obligate species,
such as polar bear (Ursus maritimus), walrus (Odobenus rosmarus),
bearded seal (Erignathus barbatus), and ringed seal (Pusa hispida)
(Laidre et al., 2008; Moore and Huntington, 2008). Polar bears in
particular have become the subject of intense political debate,
and public interest in the future of the species is increasing (e.g.,
Charles, 2008). The vulnerability of polar bears to climate warming
is clear (e.g., Stirling and Derocher, 1993; Derocher et al., 2004;
Stirling and Parkinson, 2006; Laidre et al., 2008; Wiig et al.,
2008), but few predictions exist to address how polar bear abundance might change numerically in response to a warming climate
(Amstrup et al., 2007; Hunter et al., 2007).
Prediction of polar bear population dynamics under climate
change is challenging, because observed and predicted environmental conditions differ substantially (Wiig et al., 2008). Consequently, few data exist to inform us how reproduction and
survival (and thus population abundance) might change under future conditions. To date, only two studies have incorporated climate change trends into quantitative projections of polar bear
abundance (Amstrup et al., 2007; Hunter et al., 2007), and each
of these studies had to rely on some form of extrapolation or expert judgment to parameterize suggested population models due
to the lack of data relating present to future conditions. These
analyses are important steps, and they provide new hypotheses
on how populations may respond to further warming. However,
their projections may lack accuracy if unexpected non-linearities
exist in vital rate response curves to future environmental
conditions.
Here, we follow the framework of Berteaux et al. (2006) to suggest how predictions of population abundance under climate
change could be improved. For this purpose, we ﬁrst review expected and observed climate change effects on polar bears with
speciﬁc focus on the biological mechanisms affecting survival
and reproduction. We then summarize previous attempts to forecast polar bear abundance under climate change and discuss limitations of these studies. To improve predictions of population
abundance, we suggest the development of mechanistic models
aimed at predicting reproduction and survival as a function of
the environment. Such predictions could inform demographic projection models to improve population viability analyses (PVA) under climate change. We illustrate the approach with two examples:
a dynamic energy budget (DEB) model to predict changes in survival due to prolonged summer fasts, and an encounter rate model
to predict changes in female mating success due to climate change
induced habitat fragmentation and sea ice area declines. To aid further development of such mechanistic models, we discuss data collection needs to augment ongoing monitoring projects.
2. Climate change threats to polar bears
Polar bears are vulnerable to climate warming primarily because they depend on sea ice as a platform to access their main
prey, ringed seals and bearded seals (Stirling and Archibald,
1977; Smith, 1980). Other marine mammals may locally comple-
1613
ment the diet, but in general all marine prey is expected to become
less accessible to polar bears as the sea ice declines. Terrestrial food
sources may be opportunistically exploited but are unlikely to substitute for the high energy diet polar bears obtain from seals (Derocher et al., 2004; Wiig et al., 2008; Hobson et al., 2009; Molnár,
2009). The sea ice is also used in other aspects of polar bear life history, including traveling and mating (Ramsay and Stirling, 1986;
Stirling et al., 1993). With rising temperatures, areas of open-water
and ice ﬂoe drift rates are expected to increase, and traveling in
such a fragmented and dynamic sea ice habitat would become
energetically more expensive because polar bears would have to
walk or swim increasing distances to maintain contact with preferred habitats (Mauritzen et al., 2003).
The combined effects of decreasing food availability and
increasing energetic demands are predicted to result in decreasing
polar bear body condition and a consequent cascade of demographic effects (Stirling and Derocher, 1993; Derocher et al.,
2004; Wiig et al., 2008). Pregnant females, for instance, give birth
in maternity dens, when food is unavailable for 4–8 months (Atkinson and Ramsay, 1995). To meet the energetic demands of survival,
gestation, and early lactation, females need to accumulate sufﬁcient energy stores before denning. The lightest female observed
to produce viable offspring weighed 189 kg at den entry (Derocher
et al., 1992), and the proportion of females below such a reproduction threshold will increase with ongoing food stress (Molnár,
2009). Females above the threshold may reproduce, but their
reproductive success would still decline with reduced body condition, because body condition is positively correlated with litter size
and litter mass, where the latter is also positively correlated with
cub survival (Derocher and Stirling, 1996, 1998). After den exit,
cubs are nursed for about 2.5 years, but maternal food stress may
reduce milk production, with negative consequences for cub
growth and cub survival (Derocher et al., 1993; Arnould and Ramsay, 1994). Adult survival rates, in contrast, are probably only affected under more severe conditions because polar bears can
survive extended periods without feeding (Atkinson and Ramsay,
1995). Subadult mortality, however, may increase before adult survival is affected, because young bears are less proﬁcient in ﬁnding
food (Stirling and Latour, 1978) and thus more vulnerable to adverse conditions. Such negative changes in reproduction and survival could lead to decreased population growth rates or
population declines.
There is evidence that some of these predicted changes are
underway. For example, polar bears in the Western Hudson Bay
population (Fig. 1) have shown declines in body condition, reproductive success, survival, and population abundance, and these declines are thought to result from increasing food stress associated
with prolonged open-water fasting periods (Derocher and Stirling,
1995; Stirling et al., 1999; Regehr et al., 2007). Appropriate time
series to detect changes in body condition, reproduction, and survival do not exist for most other populations (but see Regehr
et al., 2010). However, food stress has been documented for polar
bears in the Beaufort Sea (Fig. 1) (Cherry et al., 2009), and recent
incidents of cannibalism and an increased presence of polar bears
near human settlements may provide further indicators for food
stress in various populations (Amstrup et al., 2006; Stirling and
Parkinson, 2006; Towns et al., 2009).
Changes in energy availability and consequent demographic effects constitute the biggest concern for polar bears under climate
warming. However, energy-independent or only partially energyrelated effects of climate warming are also possible, such as increased exposure and vulnerability to pollutants, the emergence
of new diseases, loss of denning habitat, and conﬂict with humans
associated with industrial development. For reviews of climate
warming effects on polar bears, see Stirling and Derocher (1993),
Derocher et al. (2004) and Wiig et al. (2008).
1614
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
Fig. 1. Circumpolar polar bear populations. BB: Bafﬁn Bay; DS: Davis Strait; FB: Foxe Basin; GB: Gulf of Boothia; KB: Kane Basin; LS: Lancaster Sound; MC: M’Clintock
Channel; NB: Northern Beaufort Sea; NW: Norwegian Bay; QE: Queen Elizabeth Islands; SB: Southern Beaufort Sea; SH: Southern Hudson Bay; VM: Viscount Melville Sound;
WH: Western Hudson Bay. The ﬁgure is from Aars et al. (2006).
3. Towards an understanding of the future of polar bears
Qualitative predictions regarding the future of polar bears under changing environmental conditions abound (e.g., Stirling and
Derocher, 1993; Derocher et al., 2004; Rosing-Asvid, 2006; Stirling
and Parkinson, 2006; Moore and Huntington, 2008; Wiig et al.,
2008), and some of these predictions were outlined above. Such
assessments are useful to identify threats and to provide insights
into complex interactions between ecological dynamics, environmental variables, and anthropogenic inﬂuences, but they cannot
provide quantitative information on the manner and timeframe
in which polar bear populations will be affected. However, sound
quantitative projections of population abundances are necessary
to correctly assess conservation status, to proactively direct conservation efforts, and to set sustainable harvest quotas (Coulson
et al., 2001; Mace et al., 2008).
Currently, most projections of polar bear population abundance
are accomplished using RISKMAN, a population simulation model
that accounts for the 3-year reproductive cycle of female polar
bears (Taylor et al., 2002). In its basic components, the program
is equivalent to a stage-structured matrix population model with
parental care, such as the one developed by Hunter et al. (2007;
illustrated in Fig. 2). RISKMAN has been used to determine harvest
quotas (e.g., Taylor et al., 2002) and to assess polar bear conservation status in Canada (COSEWIC, 2008). Model parameters in these
studies were based on recent mean estimates of reproduction and
survival, and potential future changes in these demographic
parameters due to climate change were not considered. However,
our understanding of polar bear life history and ecology implies
that such changes are likely.
Quantitative predictions of population dynamics under environmental change must account for potential changes in reproduction and survival to be meaningful (Beissinger and Westphal, 1998;
Coulson et al., 2001), and are therefore possible if (a) predictions
for future environmental conditions exist, (b) the relationship between future conditions and demographic parameters can be
quantiﬁed, and (c) a population model integrating these effects
can be developed (Jenouvrier et al., 2009). In some species, such
as Emperor Penguins (Aptenodytes forsteri), a population viability
approach incorporating these three steps was possible because
reproduction and survival data exist for environmental conditions
similar to those predicted to occur (Jenouvrier et al., 2009). For polar bears, the approach is difﬁcult because few data exist to inform
us how demographic parameters might change in the future. The
only studies to attempt quantitative predictions of polar bear
abundance under climate change were consequently limited by
the need to extrapolate from present conditions (Amstrup et al.,
2007; Hunter et al., 2007) and the reliance on expert judgment
(Amstrup et al., 2007) when parameterizing proposed population
models.
Hunter et al. (2007) coupled general circulation models with
matrix population models (Fig. 2) to obtain population size
1615
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
4 (1- 4)
Females
2
3
3
(4 yr)
4
(5+ yr,
adult)
6
6
L1 f)
4
4
/2
5 (1-
(
(1- 5)
2
(3 yr)
L0)
1
1
(2 yr)
6
5
L0
(adult w/ yrlg)
5
(adult w/ coy)
Males
(
6
L1 f)
7
(2 yr)
/2
5 (1-
8
(3 yr)
7
9
(4 yr)
8
L0)
5
10
(5+ yr,
adult)
9
10
Fig. 2. Schematic representation of the polar bear life cycle, as modelled by Hunter et al. (2007), using a stage-structured two-sex matrix population model with parental
care. Stages 1–6 are females, stages 7–10 are males. ri is the probability of survival for an individual in stage i from one spring to the next, rL0 and rL1 are the probabilities of
at least one member of a cub-of-the-year (COY) or yearling (yrlg) litter surviving from one spring to the next, f is the expected size of yearling litters that survive to 2 years,
and bi is the conditional probability, given survival, of an individual in stage i breeding, thereby producing a COY litter with at least one member surviving to the following
spring. The ﬁgure is redrawn from Hunter et al. (2007).
projections for the Southern Beaufort Sea (Fig. 1) under projections
for future sea ice. For model parameterization, the authors estimated the functional relationship between polar bear survival,
reproduction, and sea ice from 6 years of capture–recapture data
(2001–2006). By classifying these demographic data into ‘‘good”
and ‘‘bad” years and assuming that future vital rates could be represented by these estimates, they analyzed the effects of an increase in the frequency of bad years on population growth and
suggested a substantial extirpation risk for the Southern Beaufort
Sea population within 45–100 years. Although their conclusions
of extirpation risk were robust against parameter uncertainty,
the authors noted wide prediction intervals in their projections,
partially due to the limited range of sea ice conditions considered
when estimating demographic parameters.
Amstrup et al. (2007) took an alternative approach, coupling
general circulation models with a polar bear carrying capacity
model and a Bayesian network model, respectively, to project population trends throughout the Arctic. They suggested likely extirpation of polar bears in two broad regions (Southern Hudson
Bay, Western Hudson Bay, Foxe Basin, Bafﬁn Bay, and Davis Strait
populations, as well as Southern Beaufort Sea, Chukchi Sea, Laptev
Sea, Kara Sea, and Barents Sea populations; Fig. 1), substantial declines in all other populations, and an overall loss of approximately
two-thirds of the global population by mid-century given current
sea ice projections. However, a lack of appropriate data linking predicted environmental conditions to polar bear population dynamics forced the authors to estimate future carrying capacities by
extrapolating from present densities, and to rely on expert judgment for other stressors.
3.1. Using mechanistic models to predict changes in survival and
reproduction
Non-linear dynamics and process uncertainty can lead to spurious predictions of population dynamics and abundance, when vital
rate estimates are extrapolated outside observed ranges or when
future vital rate estimates are based on expert judgment only
(Beissinger and Westphal, 1998; Berteaux et al., 2006; Sutherland,
2006). This kind of problem is illustrated, for example, by the failure of demographers to accurately predict human population
growth (Sutherland, 2006). An example illustrating the limitations
of extrapolation in estimating future vital rates, speciﬁcally for polar bears, is given by Derocher et al. (2004). Based on linear advances in spring sea ice break-up, they calculated that most
females in Western Hudson Bay would be unable to give birth by
2100. The authors contrasted this estimate with alternative calculations based on extrapolating observed linear declines in mean female body mass, which implied unsuccessful parturition for most
females by 2012.
Rather than estimating demographic parameters from limited
data and attempting extrapolation, we suggest using mechanistic
models that explicitly consider the cause-effect relationships by
which environmental conditions affect reproduction and survival.
Such models would allow independent prediction of these demographic parameters for yet unobserved environmental conditions
(Berteaux et al., 2006), which could then be input into demographic projection models. In Sections 3.2 and 3.3, we discuss this
approach, ﬁrst for changes in reproduction and survival as a consequence of changes in individual energy intake and energy expenditure towards movement, and then for changes that are mostly
independent from an individual’s energy budget. For both cases,
we provide a simple example for illustration.
3.2. Predicting changes in survival, reproduction, and growth due to
changes in energy intake and movement
Changes in energy availability through decreased feeding
opportunities and an increased necessity for movement would
negatively affect individual body condition, and thereby survival,
reproduction and growth. Qualitatively, this causal relationship is
clear, but quantitative predictions of how body condition, survival,
reproduction and growth would be affected under changing environmental conditions do not exist. Empirical energetic studies on
feeding, movement, somatic maintenance, thermoregulation,
1616
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
reproduction and growth in polar bears are available (e.g., Øritsland et al., 1976; Best, 1982; Watts et al., 1987; Arnould and Ramsay, 1994; Stirling and Øritsland, 1995), but these studies alone are
insufﬁcient for predictive purposes, because it is impractical to
measure survival, reproduction and growth under all possible scenarios of energy intake and movement. For prediction, a mathematical energy budget framework is needed that synthesizes
such data in a model that mechanistically describes how available
energy is prioritized and allocated within the organism.
DEB models (sensu Kooijman, 2010) explicitly track how an
individual utilizes available energy, using mechanistic rules for energy allocation and prioritization between somatic maintenance,
thermoregulation, reproductive output, and structural growth.
DEB models thus have the potential to predict survival, reproduction and growth, in response to expected changes in energy intake
and movement associated with changing environmental conditions (Gurney et al., 1990; Nisbet et al., 2000; Kooijman, 2010),
and DEB models are particularly useful to predict an individual’s
response to food limitation (Zonneveld and Kooijman, 1989; Noonburg et al., 1998). To date, DEB models have been applied to invertebrates, ﬁsh, amphibians, reptiles, and birds (Kooijman, 2010, and
references therein), and more recently also to whales (Klanjscek
et al., 2007) and ungulates (De Roos et al., 2009).
Assuming strong homeostasis (Molnár et al., 2009), a 2-compartment DEB model that tracks changes in storage energy (E;
units: MJ) and structural volume (V; units: m3) through time (t)
can be written as follows:
dE
¼ F IE F EA F EM F ET F EG F ER
dt
dV
¼ g 1 F EG
dt
ð1Þ
where FIE represents the inﬂux of energy from the environment
through food acquisition and assimilation, and FEA, FEM, FET, FEG,
and FER represent the respective rates of storage energy utilization
for activity, somatic maintenance, thermoregulation, structural
growth, and reproduction. The parameter g represents the energetic
cost of growing a unit volume of structure (Klanjscek et al., 2007).
For simplicity, we assume additivity of ﬂuxes (Wunder, 1975),
and that all energy is channeled through storage (Kooijman,
2010), although other formulations are possible (e.g., Lika and Nisbet, 2000; Klanjscek et al., 2007). Note also that the ﬂuxes in Eq. (1)
are not independent from each other: energy intake (FIE), for example, likely depends on how much energy is allocated to movement
(FEA), and energy allocation to growth (FEG) is usually assumed possible only after maintenance requirements (FEM and FET) are met
(Kooijman, 2010).
The challenge in formulating a DEB model for a given species is
threefold. First, a method is needed that allows estimation of energy stores (E) and structural volume (V), second, the functional
forms of the ﬂuxes FXY need to be determined, and third, these
functions need to be parameterized. A full DEB model for polar
bears is currently lacking, but the ﬁrst step was taken by Molnár
et al. (2009) who described a polar bear body composition model
that distinguishes between storage and structure. Their model allows estimation of E from total body mass and straight-line body
length, and estimation of V from straight-line body length. Molnár
et al. also suggest that somatic maintenance rate (FEM) in polar
bears should be proportional to lean body mass (i.e., the mass of
all tissue that is not body fat), and they parameterize this DEB
model component from body mass changes in fasting adult males.
Below, we extend their model to include costs of movement (FEA)
and illustrate the usefulness of the DEB approach for prediction
by estimating future changes in adult male survival due to expected extensions of the summer open-water fasting period in
Western Hudson Bay. A full DEB model would also allow prediction
of polar bear reproduction and growth under food limitation, but
insufﬁcient data exist to fully determine the necessary model components FER, and FEG. Directed studies, however, may ﬁll these data
gaps, and we outline key data requirements below to aid further
model development.
3.2.1. Example: predicting changes in survival due to prolonged fasting
– time to death by starvation
Polar bears in the Western Hudson Bay population (Fig. 1) are
forced ashore when the sea ice melts in summer (Derocher and
Stirling, 1990). On-land, energetically meaningful food is unavailable, and bears rely on their energy stores for survival (Ramsay
and Stirling, 1988; Hobson et al., 2009). In recent years, spring
sea ice break-up in Western Hudson Bay has been occurring progressively earlier, resulting in shortened on-ice feeding and prolonged on-shore fasting for polar bears in this population
(Stirling and Parkinson, 2006). Further extensions to the openwater period are expected under continued climatic warming,
and polar bear survival rates for this period may eventually drop
if bears cannot accumulate sufﬁcient storage energy for the fast.
To illustrate how future changes in survival due to prolonged fasting can be predicted, we use a DEB model to estimate how long a
bear can survive on its energy stores before death by starvation.
For simplicity, we consider adult males only.
We apply the DEB model for fasting, non-growing and nonreproducing polar bears in a thermoneutral state from Molnár
et al. (2009), with an additional component to account for energy
allocated to movement:
dE
¼
dt
mLBM
|ﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄ}
Somatic maintenance
ðaMb þ cMd v Þ
|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}
ð2Þ
Movement
The model assumes a somatic maintenance rate proportional to
lean body mass, LBM, with m representing the energy required per
unit time to maintain a unit mass of lean tissue (Molnár et al.,
2009). Energy costs of movement, by contrast, are dependent on
total body mass, M, because both lean tissue and body fat need
to be moved. Movement costs are represented by an allometric
equation, where the ﬁrst part of the sum, aMb, represents the metabolic cost of maintaining posture during locomotion (in addition
to somatic maintenance). The second part, cMdv, reﬂects the positive linear relationship between energy consumption and velocity,
v (Schmidt-Nielsen, 1972; Taylor et al., 1982).
Using the body composition model of Molnár et al. (2009), Eq.
(2) can be rewritten as:
dE
3
¼ mða1 ð1 uÞE þ qSTR kL Þ
|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}
dt
Somatic maintenance
3
3
ðaða1 E þ qSTR kL Þb þ cða1 E þ qSTR kL Þd v Þ
|ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ}
ð3Þ
Movement
where a represents the energy density of storage, u the proportion
of storage mass that is fat, and qSTRk is a composite proportionality
constant to estimate structural mass from straight-line body length,
L. Body composition and maintenance parameters were estimated
a = 19.50 MJ kg1,
u = 0.439,
as
m = 0.089 MJ kg1 d1,
qSTRk = 14.94 kg m3 (Molnár et al., 2009), movement parameters
as a = 0, c = 0.0214 MJ km1, d = 0.684 (Molnár, 2009). For model
development and parameterization details, see Molnár (2009) and
Molnár et al. (2009).
Time to death by starvation can be estimated for a bear of
straight-line body length L and initial energy stores E(0) = E0 by
numerically integrating Eq. (3) and solving for time T when
E(T) = 0. Here, we considered two scenarios, one for resting bears
(v = 0) and one for bears moving at average speed v = 2 km d1,
which corresponds to observed on-land movement rates (Derocher
1617
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
and Stirling, 1990). For resting bears, energy density (E/LBM) was
the sole determinant of time to death by starvation, whereas for
moving bears starvation time also depended on L. However, variation due to changes in L was small, so we used the mean observed
length of 2.34 m in all subsequent calculations. For both scenarios,
time to death by starvation increased non-linearly with energy
density (Fig. 3).
Predictions for changes in adult male survival in Western Hudson Bay as a function of fasting period length were then obtained
by linking the time to death by starvation response curves to observed energy densities. For this purpose, we used mass and length
data from 97 adult male polar bears (7 years of age) caught onland in 1989–1996 in Western Hudson Bay (see Molnár et al.,
2009, for handling procedures). All animal handling protocols were
consistent with the Canadian Council on Animal Care guidelines.
Body masses were scaled to August 1 (mean on-shore arrival date
during 1990s; Stirling et al., 1999) using the mass loss curve in
Molnár et al. (2009). Energy densities on August 1 were calculated
from these body masses using the body composition model of Molnár et al. (2009).
Adult male survival rate as a function of observed energy densities can be estimated for any fasting period length by considering
the proportion of bears that would starve to death before the end
of the fasting period. For illustration we discuss survival predictions for a fasting period length of 120 days, typical of the 1980s,
and 180 days which reﬂects potential future conditions (the fasting
period has been increasing by about 7 days per decade since the
early 1980s; Stirling and Parkinson, 2006). Observed energy densities were normally distributed, and with a fasting period of
120 days about 3% of these bears are expected to die of starvation
before the end of the fasting period when resting (line A in Fig. 3)
and about 6% when moving (line B in Fig. 3). If the fasting period is
extended to 180 days (i.e., due to earlier spring ice break-up and
delayed fall freeze-up), about 28% of these males would die with
no on-land movement (line C in Fig. 3) and about 48% if moving
(line D in Fig. 3). Expected changes in adult male survival are
non-linear due to the normal distribution of energy densities,
and to a smaller degree due to the non-linearity of the time to
death by starvation curves. Estimates for changes in survival are
conservative because death may happen sooner if the strong
300
250
1
200
C
0.8
D
150
0.6
A
B
100
0.4
3.3. Predicting non-energy-related changes in demographic
parameters
Some effects of climate change will not be directly energy-related. Mechanistic models, speciﬁc to the proposed cause-effect
relationships, may nevertheless be used for prediction in many
cases, but a comprehensive discussion of all possible effects and
models is impossible. However, to illustrate the potential of mechanistic models in predicting changes in vital rates, even when the
primary mechanism for change is not energy-related, we explore
how habitat fragmentation and declines in sea ice area would affect female mating success.
3.3.1. Example: potential climate change impacts on female mating
success
Derocher et al. (2004) put forth two contrasting hypotheses
regarding changes in female mating success under climate warming. First, increased areas of open-water and increased ice ﬂoe drift
rates may impede mate-ﬁnding and result in reduced pregnancy
rates because adult males rely on contiguous female tracks for
mate location. By contrast, declines in sea ice area may facilitate
mate-ﬁnding to increase pregnancy rates by increasing bear density during the mating season. Here, we assess the respective
importance of these contrasting effects. Speciﬁcally, we use the
mating model of Molnár et al. (2008) to show how quantitative
predictions for changes in female mating success due to changes
in habitat fragmentation (mate-searching efﬁciency) and sea ice
area can be obtained.
Polar bear pairing dynamics during the mating season are driven by mate location, pair formation, and pair separation, and
can be described by the following system of differential equations
(Molnár et al., 2008):
dm
dt
|{z}
0.2
0
5
10
15
20
0
Energy density (MJ / kg)
Fig. 3. Estimated time to death by starvation for fasting adult male polar bears,
when resting (solid line) and when moving at average speed v = 2 km d1 (dotted
line). The horizontal dotted line indicates a fasting period of 120 days, the
horizontal dashed line a fasting period of 180 days. Crosses show the cumulative
distribution of energy densities at the beginning of the fasting period (right axis) for
97 adult males caught in 1989–1996 in the Western Hudson Bay population. Lines
A–D illustrate the proportion of these males that would die from starvation
following a fast of 120 days and 180 days, with and without movement, respectively (see text for details).
df
dt
|{z}
Breeding pairs

df
dt
|{z}
Fertilized females
ð4aÞ
ð4bÞ
Pair formation
Pair formation
¼
lp
|{z}
Pair separation
sq
mf
A
|ﬄﬄ{zﬄﬄ}
sq
mf
A
|ﬄﬄ{zﬄﬄ}
¼
þ
Pair formation
¼
Unfertilized females
dp
dt
|{z}
sq
mf
A
|ﬄﬄ{zﬄﬄ}
¼
Solitary available males
50
0
Proportion
Time to death by starvation (days)
350
homeostasis assumption is violated near death. Furthermore, with
progressively earlier spring sea ice break-up, energy densities at
on-shore arrival are expected to be reduced relative to those observed during the 1990s due to shortened on-ice feeding (Stirling
and Derocher, 1993), thereby further reducing expected time to
death by starvation. Such declines in body condition have already
been documented in Western Hudson Bay (Derocher and Stirling,
1995; Stirling et al., 1999).
Predictions of starvation time and resultant changes in survival
are also possible for other groups, such as subadults or adult females with offspring, if the additional energy expended on lactation and growth, respectively (FER and FEG in Eq. (1)), can be
quantiﬁed. Generally, adult males may be the least affected group
because they do not spend energy on growth or lactation. However, due to their proportionally higher lean tissue content in storage, they cannot fast as long as non-reproducing adult females
(Molnár et al., 2009).
lp
|{z}

lp
|{z}
ð4cÞ
Pair separation
ð4dÞ
Pair separation
where m(t), f(t), p(t), and f(t) represent the respective numbers (at
time t) of solitary males searching for mates, solitary unfertilized
1618
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
females, breeding pairs, and solitary fertilized females. The lefthand sides of Eqs. (4a–d) represent the respective rates of change
in these quantities, and these rates depend on pair formation and
pair separation. Pair formation is modelled using the law of mass
action, and pairs are formed at rate sq/A, where s represents searching efﬁciency (units: km2 d1), q is the probability of pair formation
upon encounter (i.e., mate choice), and A is habitat area (units:
km2). Pairs remain together for l1 time units (units: d), thus separating at rate l. The mating season begins at t = 0, when m(0) = m0,
f(0) = f0, p(0) = 0, f*(0) = 0, and lasts T time units. Female mating success is deﬁned as the proportion of females fertilized at the end of
the mating season and is estimated as 1 f(T)/f0. To explore how
changes in sea ice area and habitat fragmentation would affect female mating success, we rewrote the model of Molnár et al.
(2008) considering bear numbers rather than densities, thereby
explicitly representing sea ice area, mate-searching efﬁciency and
mate choice. We also assumed maximal male mating ability (i.e.,
all solitary males search for mates at all times), considering a simpliﬁed version of the model in Molnár et al. (2008). However, it is
noteworthy that male mating ability may also decline under climate warming induced food stress, and such declines could reduce
female mating success (Molnár et al., 2008).
The model explicitly considers the mechanisms determining female mating success, describes observed pairing dynamics well,
and can thus be used to predict female mating success from initial
male and female numbers, m0 and f0, and model parameters s, q, A,
and l (Molnár et al., 2008). We consider changes in sea ice area (A)
and mate-searching efﬁciency (s), and illustrate predictions using
the example of Lancaster Sound (Fig. 1), where m0 = 489, f0 = 451,
sq/A = 0.00021 d1, l1 = 17.5 d and T = 60 d were estimated for
1993–1997, implying a female mating success of 99% (Molnár
et al., 2008). Female mating success depends on the ratio sq/A
and is predicted to decline non-linearly if searching efﬁciency s declines faster than habitat area A, and to increase non-linearly
otherwise. For example, assuming that m0, f0, l1, T, and q remain
constant in Lancaster Sound, female mating success is predicted to
decline from 99% to 91% if s declined twice as fast as A, and to 72% if
s declined four times as fast as A. By contrast, if A declined faster
than s, mating success would remain essentially unchanged at
around 100% in this population (Fig. 4).
Female mating success
1.0
0.8
0.6
0.4
0.2
0
0
0.0001
0.0002
0.0003
−1
s / A (searching efficiency relative to habitat area; units d )
Fig. 4. Potential climate change impacts on female mating success (the proportion
of females fertilized at the end of the mating season), arising from declines in matesearching efﬁciency, s, and sea ice habitat area, A, assuming constant mate choice.
Predictions are shown for the population of Lancaster Sound, with male and female
numbers assumed unchanged relative to 1993–1997, and the estimate of s/A for
this period marked by a circle. Also indicated are scenarios where s declines twice
(square) and four times (diamond) as fast as A, respectively. A scenario where A
declines faster than s by a factor of 1.5 is indicated by a triangle (see text for details).
The parameters s and A may change independent of each other
because mate-searching efﬁciency depends on movement speeds,
movement patterns, detection distance, and male tracking ability,
parameters that are affected more by the degree of habitat fragmentation (areas of open-water between ice ﬂoes) than by total
habitat area. The degree to which s and A will be affected by climate change cannot be predicted from the mating model itself.
However, such predictions could be obtained independently for s
from mechanistic encounter rate models that account for changes
in movement patterns, tracking ability and detection distance due
to habitat fragmentation (Kiørboe and Bagøien, 2005). The degree
of future habitat fragmentation and changes to sea ice area (A)
could in turn be predicted from sea ice models. Resultant predictions for s and A could then be input into the mating model to obtain more speciﬁc predictions of female mating success under
climate change than presented here. Potential future changes in
mate choice (q) should hereby also be considered, because mate
choice may vary adaptively as a function of male densities, sex ratios, and expected mating success (Kokko and Mappes, 2005). Potential declines in s may be compensated by increases in q,
because pair formation rate is determined by the composite term
sq/A (but note that q cannot be increased to values larger than 1).
The predictions outlined here are insensitive to the parameters
l1 and T, but may be affected signiﬁcantly by harvest-induced
changes in m0 and f0 (Molnár et al., 2008).
4. Integrating predicted changes in demographic parameters
into population models
The stage-structured population dynamics of polar bears can be
formalized in matrix models (Fig. 2), which are useful for population projections and PVAs (Hunter et al., 2007). However, such
analyses are only accurate if future vital rates (reproduction and
survival) are accurately represented by existing estimates, or if future changes in vital rates can be accurately predicted from present
conditions. The lack of data on vital rates under not yet experienced conditions has thus been a major limitation to PVA accuracy
(Beissinger and Westphal, 1998; Ludwig, 1999; Coulson et al.,
2001; Sutherland, 2006). To avoid this problem, we have advocated
mechanistic models to predict changes in survival and reproduction because such models can often be developed and parameterized independent of environmental conditions. A second
advantage of such mechanistic models is their ability to identify
expected non-linearities and threshold events in vital rate response curves to environmental conditions (Figs. 3 and 4), which
will affect PVAs (Ludwig, 1999; Harley et al., 2006).
The mathematical integration of vital rate predictions into matrix population models is often straightforward, and we outline
this process for the two examples considered above. Adult male
survival rate from one spring to the next (parameter r10 in the matrix model of Hunter et al. (2007); Fig. 2) can be written as the
product of adult male survival during the fasting and feeding periods, respectively. Expected changes in survival during the fasting
period (Fig. 3) can thus be incorporated to predict changes in r10
due to this survival component. The probability of a female without offspring breeding (b4 in Fig. 2) can similarly be decomposed
into the probabilities of successful mating, successful implantation,
successful parturition, and early cub survival. Expected changes in
mating success caused by habitat fragmentation and sea ice area
declines (Fig. 4) could thus also be incorporated into a matrix population model.
The biggest limitation to this component-wise approach of predicting changes in reproduction and survival relates to uncertainty
in initial conditions. For example, the distribution of energy densities at the beginning of the fasting period in any given year, and
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
thus the period-speciﬁc survival rate, may depend on the date of
sea ice break-up in that year (and thus the length of the preceding
on-ice feeding period), but also on the lengths of the feeding and
fasting periods in previous years (i.e., time lags). This problem of
uncertainty could be avoided if a full DEB model was available that
tracks the energy intake and expenditure of polar bears through
the entire year. Population projections would in that case be a matter of tracking individuals over time. However, until a fully predictive model becomes available, a component-wise analysis of
expected changes in vital rates and resultant effects on population
growth is possible because the direction of the expected changes in
initial conditions is often clear. For example, polar bear energy
densities at on-shore arrival in Western Hudson Bay are already
declining and are expected to decline further. Models that assume
all else equal (in particular, on-shore arrival energy densities as observed during the 1990s) to predict future fasting period survival
rates as a function of predicted fasting period lengths would thus
be conservative and could set boundaries to expected changes in
survival. Until different effects of climate change on vital rates, addressed by different mechanistic models, can be connected into a
single predictive framework, component-wise prediction of
changes in vital rates (treating different aspects of climate change
on polar bears separately) could provide a series of conservation
indicators that should be considered in conservation assessments
and population management.
5. A call for data
The type of data required to further mechanistic models for
reproduction and survival is in many cases different from data collected for monitoring these demographic parameters (such as
mark-recapture data). The development of such models will require the integration of ﬁeld research to speciﬁcally address the
mechanisms determining change in reproduction and survival.
The areas of investigation will be speciﬁc to the mechanisms considered, and as it is impossible to provide a comprehensive summary of all potential modelling approaches, it is similarly
impossible to outline all data that might prove useful for model
development, parameterization, and validation. However, because
most expected climate change effects on polar bears are energy-related, we believe that DEB models may provide one of the most
useful venues for understanding and predicting climate change effects on polar bears. Changes in growth, reproduction, and survival,
in response to expected changes in feeding and movement can be
predicted from DEB models, provided that sufﬁcient physiological
data can be gathered to specify energy allocation rules and parameterize model terms (Gurney et al., 1990; Noonburg et al., 1998;
Kooijman et al., 2008; Kooijman, 2010). Long-term research on polar bears has already provided much of the required physiological
data for DEB development, and missing pieces could be addressed
with directed studies. To aid the development of a full polar bear
DEB model, we next outline key data requirements.
DEB models consider two distinct components of energy ﬂow:
net energy intake from the environment (the difference between
terms FIE and FEA in Eq. (1)) and the allocation of assimilated energy
within the organism towards somatic maintenance, thermoregulation, reproduction, and growth (FEM, FET, FER, and FEG). The physiological terms FEM, FET, FER, and FEG can be understood independently
from the environment, and they could be determined under current conditions. In fact, the term for somatic maintenance (FEM)
has already been speciﬁed (Molnár et al., 2009; cf. also Eqs. (2)
and (3)), and the thermoregulation term FET can probably be determined from published data (e.g., Best, 1982). By contrast, insufﬁcient data exist to fully determine the model terms FER, and FEG,
which specify the magnitude of energy allocation towards repro-
1619
duction and growth and the conditions under which energy allocation to these processes ceases.
Reproduction in female polar bears consists of a short gestation
period (ca. 60 days; Derocher et al., 1992), and a lactation period
that normally lasts up to 2.5 years (Derocher et al., 1993). The
energetic costs of gestation are small compared to those of lactation (Oftedal, 1993), so that data collection should prioritize quantifying milk energy transfer. Milk energy transfer rates may depend
on maternal body condition (e.g., storage energy or energy density), cub demand, and cub age. Cub demand, in turn, may be determined by cub body condition, cub growth, and the amount of solid
food consumed (Lee et al., 1991; Oftedal, 1993; Arnould and Ramsay, 1994). Although it may be straightforward to formulate lactation within a DEB model (e.g., Klanjscek et al., 2007), relatively
large amounts of data may be required for model parameterization
due to the number of factors involved. Milk energy transfer data
covering a range of feeding conditions (e.g., on-shore fasting and
on-ice feeding) as well as a range of maternal and cub body conditions are required for model development. Data on the presence or
absence of lactation in relation to maternal energy stores, particularly during the on-shore fasting period in southern populations,
may provide further insight into the mechanisms determining cessation of lactation. Cessation of lactation has been reported for
food-stressed females (Derocher et al., 1993), implying a storage
energy (or energy density) threshold below which lactation stops.
The existence of such a threshold is supported by DEB theory (Lika
and Nisbet, 2000), and would have implications for lactation (and
consequent cub survival) for females food-stressed by climate
warming.
The allocation of energy to structural growth is probably the
least understood component in the energy budget of polar bears.
It may also be the most difﬁcult term to specify in a DEB model, because energy allocation to growth may depend on energy intake
(Lika and Nisbet, 2000; Kooijman, 2010), and may also be sizedependent (Nisbet et al., 2004). Structural growth data, estimated
through changes in straight-line body length, is needed for bears
of different ages, sizes and body conditions with known energy intake. Captive bears may aid in determining this model component
because energy intake is known and changes in storage energy and
body length could be determined. Growth in bears under food limitation should also be considered to specify the conditions under
which energy allocation to growth ceases. While growth data from
food-stressed bears may not be available from captive studies, such
data could also be obtained from cubs and subadults caught during
the on-shore fasting period in southern populations. Energy intake
for nursing cubs could in this case be measured through isotope
dilution methods (Arnould and Ramsay, 1994), or approximated
through changes in maternal energy stores. For both growth and
reproduction (and, in fact, for all DEB components), longitudinal
data (i.e., repeated measurements of individuals over weeks or
months) is preferable over population cross-sections because individual-based processes are assessed.
Changes to the second component of an individual’s energy
budget, net energy intake (FIE FEA), under changing environmental conditions cannot be predicted from single-species DEB models.
Multi-species DEB models, modelling the ﬂow of energy between
trophic levels (Nisbet et al., 2000), may be able to provide such predictions, but insufﬁcient data on the polar bear-seal predator–prey
system currently prevents the construction of such models. Little is
known about Arctic seal abundance, distribution, and population
dynamics, and even less is known about the mechanisms regulating the polar bear-seal predator–prey system. To date, only a handful of studies have documented kill frequency and meal size in
polar bears, and these studies are restricted in space and time (Stirling, 1974; Stirling and Latour, 1978; Stirling and Øritsland, 1995).
Kill frequencies are unknown for most populations and almost all
1620
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
seasons. A mechanistic link between habitat characteristics, prey
population dynamics, and polar bear energy intake is also missing.
Comprehensive feeding data are needed to illuminate these links
and should become a research priority if we are to move towards
a predictive framework for changes in polar bear energy intake
(and consequent changes in reproduction and survival) under climate warming. The collection of detailed dietary information can
be difﬁcult because polar bears forage in remote sea ice habitats,
but new statistical methods, such as state-space models (Franke
et al., 2006) or behavioural change point analyses (Gurarie et al.,
2009), could be used to extract feeding events from GPS movement
data. Moreover, given longitudinal mass and length data, energy
intake could also be inferred from DEB models, provided that the
energy expenditure terms FEA, FEM, FET, FER, and FEG can be speciﬁed
a priori.
In addition to the new set of research priorities outlined here,
we advocate continued mark-recapture studies to estimate survival and reproduction. Although such studies may be of limited
use for predicting polar bear population dynamics under climate
change (given the lack of long-term studies for most populations
and the discussed problems associated with extrapolating vital
rates into yet unobserved environmental conditions), they are useful for monitoring past and current change, crucial to population
management, conservation status assessment, and the setting of
harvest quotas. Additionally, in the context outlined here, mark-recapture studies may provide valuable reproduction and survival
data that could be used to validate proposed DEB and other mechanistic models aimed at predicting these demographic parameters.
6. Conclusions
There is no doubt that climate warming is occurring, and climatologists and other scientists have provided a number of predictive
models for temperature, precipitation, sea ice, permafrost, and
other issues (IPCC, 2007). Ecologists, by contrast, are still facing
considerable challenges to obtain quantitative predictions for the
resultant effects on species and ecosystems. It is clear that many
species are already affected (Walther et al., 2002; Parmesan,
2006), but quantitative predictions are lacking for most species,
and existing predictions are often associated with large uncertainty, largely due to limited data and insufﬁciently understood
causal chains (Berteaux et al., 2006; Krebs and Berteaux, 2006;
Sutherland, 2006). The mechanistic framework advocated here
may help to incorporate cause-effect relationships into ecological
predictions, could link expected effects of climate change over various levels of biological organization, and could alert us to the presence of yet unobserved non-linearities in reproduction and
survival in response to changing environmental conditions.
Whether or not climate change effects on survival and reproduction are incorporated into PVAs may have signiﬁcant effects
on conservation status assessments and other aspects of population management. Polar bears were listed globally as ‘‘Threatened”
in 2008 under the US Endangered Species Act due to the threats
posed by climate change (Federal Register, 2009). In contrast, the
assessment of polar bears in Canada by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) did not account
for possible climate change effects, and their ﬁnding of ‘‘Special
Concern” (COSEWIC, 2008) identiﬁed a lower level of threat than
the US assessment. The US and Canadian assessments used similar
population projection models in their PVAs, but they differed in
their approaches towards model parameterization. The COSEWIC
report used mean reproduction and survival rates from earlier
studies and projected these forward, speciﬁcally stating that they
‘‘. . .do not account for the possible effects of climate change.”
(COSEWIC, 2008: page iii). The US approach included environmen-
tal trends in their PVA, but they assumed that future vital rates
would correspond to estimates from three ‘‘good” and two ‘‘bad”
habitat years observed between 2001 and 2005 (Hunter et al.,
2007). Mechanistic models for reproduction and survival were
not used in either approach, but may affect status assessments in
both countries. If there are non-linear relationships between environmental conditions and polar bear vital rates, as suggested by
the two models considered above, then population projections
may be direr than suggested by existing assessments.
Moreover, polar bear vital rates may also be affected by other
stressors, not always directly caused but possibly ampliﬁed by climate change, such as harvest, pollution, or the emergence of new
diseases. Harvest-induced changes in population composition, for
instance, may lead to a mate-ﬁnding Allee effect (Molnár et al.,
2008). Increased exposure of polar bears to persistent organic pollutants (Derocher et al., 2004) may affect their endocrine system
(Skaare et al., 2002), their immune system (Bernhoft et al., 2000),
and by extension survival and reproduction (Derocher et al.,
2003). Climate change may lead to the emergence of new diseases
in Arctic wildlife (Bradley et al., 2005). These stressors should also
be considered in status assessments and population management
(Amstrup et al., 2007) and the suggested approach for predicting
changes in reproduction and survival remains applicable. However,
the degree to which these effects will be amenable to prediction
depends on the level at which causal chains are understood and
the availability of data to develop appropriate mechanistic models
(Jonzén et al., 2005; Berteaux et al., 2006; Krebs and Berteaux,
2006). Molnár et al. (2008), for instance, developed a mechanistic
model for the polar bear mating system (cf. Eq. (4)) to predict female mating success from male and female densities for yet unobserved population compositions, and they showed that a sudden
reproductive collapse could occur if males are severely depleted.
Their results could be incorporated into a two-sex population matrix model and would allow predicting the effects of a continued
sex-selective harvest on female mating success, and thus population growth. The effects of increasing pollution levels on reproduction and survival could also be predicted with mechanistic models,
speciﬁcally pharmacokinetic models coupled with DEB models
(Klanjscek et al., 2007), but no such efforts are underway for polar
bears. By contrast, potential future effects of emerging diseases on
vital rates remain currently unquantiﬁable in polar bears due to
unclear causal chains and a lack of empirical data.
The methods we have outlined in this paper for polar bears are
broadly applicable to other species. Linking energy availability to
demographic parameters will be a key means of understanding
species responses to climate change. The increase in fasting period
modelled here can be considered a form of shifting phenology and
can be applied to any species. For example, breeding schedules in
birds are closely tied to the phenology of their food supplies, and
the disruption of this pairing can affect reproductive success (Visser et al., 1998; Thomas et al., 2001). DEB modelling may be a
means to explore these relationships to aid conservation planning.
It seems clear that not all species will be currently amenable to
the mechanistic framework outlined above. For mechanistic models to be successful in prediction, initial conditions must be well
described, all important variables must be included in the model,
and model variables must be related to each other in an appropriate way (Berteaux et al., 2006). Whether or not these conditions
are fulﬁlled cannot be known a priori (Berteaux et al., 2006). However, modelling is an iterative approach, where proposed models
should be tested against independent data to decide whether the
models were successful in predicting. Models can then be improved and tested again, until they converge to satisfactory performance. Arctic species, in particular, may be among the most
amenable to prediction because low species diversity, relatively
simple food webs, and a limited range of species interactions result
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
in comparatively simple relationships between environmental
variables and their effects on individuals and populations.
Mechanistic models are not the only means of predicting the
climate change effects on species, but given their potential to predict into yet unobserved conditions, we believe they have been
underutilized and present a fruitful line of research to address conservation challenges in a changing world.
Acknowledgements
Support for this study was provided by ArcticNet, Canadian
International Polar Year, Canadian Wildlife Federation, Environment Canada, Manitoba Conservation Sustainable Development
Innovations Fund, Polar Bears International, Polar Continental Shelf
Project, World Wildlife Fund (Canada) and the University of Alberta. We gratefully acknowledge a Canada Research Chair (M.A.L.)
and NSERC Discovery Grants (A.E.D., M.A.L., G.W.T.). The project
was aided by data collected by the late M. Ramsay. We would like
to thank J. Arnould, S. Atkinson, F. Messier and the Government of
Nunavut for access to data on bears caught in Western Hudson Bay
and Lancaster Sound.
References
Aars J., Lunn, N.J., Derocher, A.E. (Eds.), 2006. Polar bears: Proceedings of the 14th
Working Meeting of the IUCN Polar Bear Specialist Group. IUCN, Cambridge.
Ådahl, E., Lundberg, P., Jonzén, N., 2006. From climate change to population change:
the need to consider annual life cycles. Global Change Biology 12, 1627–1633.
Amstrup, S.C., Stirling, I., Smith, T.S., Perham, C., Thiemann, G.W., 2006. Recent
observations of intraspeciﬁc predation and cannibalism among polar bears in
the Southern Beaufort Sea. Polar Biology 29, 997–1002.
Amstrup, S.C., Marcot, B.G., Douglas, D.C., 2007. Forecasting the range-wide status of
polar bears at selected times in the 21st century. USGS Science Strategy to
Support US Fish and Wildlife Service Polar Bear Listing Decision. US Geological
Survey, Reston, Virginia.
Arnould, J.P.Y., Ramsay, M.A., 1994. Milk production and milk consumption in polar
bears during the ice-free period in Western Hudson Bay. Canadian Journal of
Zoology 72, 1365–1370.
Atkinson, S.N., Ramsay, M.A., 1995. The effects of prolonged fasting of the body
composition and reproductive success of female polar bears (Ursus maritimus).
Functional Ecology 9, 559–567.
Beissinger, S.R., Westphal, M.I., 1998. On the use of demographic models of
population viability in endangered species management. Journal of Wildlife
Management 62, 821–841.
Bernhoft, A., Skaare, J.U., Wiig, Ø., Derocher, A.E., Larsen, H.J.S., 2000. Possible
immunotoxic effects of organochlorines in polar bears (Ursus maritimus) at
Svalbard. Journal of Toxicology and Environmental Health, Part A 59, 561–574.
Berteaux, D., Humphries, M.M., Krebs, C.J., Lima, M., McAdam, A.G., Pettorelli, N.,
Réale, D., Saitoh, T., Tkadlec, E., Weladji, R.B., Stenseth, N.C., 2006. Constraints to
projecting the effects of climate change on mammals. Climate Research 32,
151–158.
Best, R.C., 1982. Thermoregulation in resting and active polar bears. Journal of
Comparative Physiology 146, 63–73.
Boé, J.L., Hall, A., Qu, X., 2009. September sea-ice cover in the Arctic Ocean projected
to vanish by 2100. Nature Geoscience 2, 341–343.
Bradley, M.J., Kutz, S.J., Jenkins, E., O’Hara, T.M., 2005. The potential impact of
climate change on infectious diseases of Arctic fauna. International Journal of
Circumpolar Health 64, 468–477.
Charles, D., 2008. Climate change: polar bear listing opens door to new lawsuit.
Science 320, 1000–1001.
Cherry, S.G., Derocher, A.E., Stirling, I., Richardson, E.S., 2009. Fasting physiology of
polar bears in relation to environmental change and breeding behavior in the
Beaufort Sea. Polar Biology 32, 383–391.
Comiso, J.C., 2002. A rapidly declining perennial sea ice cover in the Arctic.
Geophysical Research Letters 29, 1956.
Comiso, J.C., Parkinson, C.L., Gersten, R., Stock, L., 2008. Accelerated decline in the
Arctic sea ice cover. Geophysical Research Letters 35, L01703.
COSEWIC, 2008. COSEWIC assessment and update status report on the polar bear
Ursus maritimus in Canada. Committee on the Status of Endangered Wildlife in
Canada, Ottawa.
Coulson, T., Mace, G.M., Hudson, E., Possingham, H., 2001. The use and abuse of
population viability analysis. Trends in Ecology and Evolution 16, 219–221.
De Roos, A.M., Galic, N., Heesterbeek, H., 2009. How resource competition shapes
individual life history for nonplastic growth: ungulates in seasonal food
environments. Ecology 90, 945–960.
Derocher, A.E., Stirling, I., 1990. Distribution of polar bears (Ursus maritimus) during
the ice-free period in Western Hudson Bay. Canadian Journal of Zoology 68,
1395–1403.
1621
Derocher, A.E., Stirling, I., 1995. Temporal variation in reproduction and body mass
of polar bears in Western Hudson Bay. Canadian Journal of Zoology 73, 1657–
1665.
Derocher, A.E., Stirling, I., 1996. Aspects of survival in juvenile polar bears. Canadian
Journal of Zoology 74, 1246–1252.
Derocher, A.E., Stirling, I., 1998. Maternal investment and factors affecting offspring
size in polar bears (Ursus maritimus). Journal of Zoology (London) 245, 253–260.
Derocher, A.E., Stirling, I., Andriashek, D., 1992. Pregnancy rates and serum
progesterone levels of polar bears in Western Hudson Bay. Canadian Journal
of Zoology 70, 561–566.
Derocher, A.E., Andriashek, D., Arnould, J.P.Y., 1993. Aspects of milk composition
and lactation in polar bears. Canadian Journal of Zoology 71, 561–567.
Derocher, A.E., Wolkers, H., Colborn, T., Schlabach, M., Larsen, T.S., Wiig, Ø., 2003.
Contaminants in Svalbard polar bear samples archived since 1967 and possible
population level effects. The Science of the Total Environment 301, 163–174.
Derocher, A.E., Lunn, N.J., Stirling, I., 2004. Polar bears in a warming climate.
Integrative and Comparative Biology 44, 163–176.
Federal Register, 2009. Endangered and threatened wildlife and plants;
determination of threatened status for the polar bear (Ursus maritimus)
throughout its range. United States Department of the Interior, US Fish and
Wildlife Service. (retrieved 19.12.09).
Franke, A., Caelli, T., Kuzyk, G., Hudson, R.J., 2006. Prediction of wolf (Canis lupus)
kill-sites using hidden Markov models. Ecological Modelling 197, 237–
246.
Gurarie, E., Andrews, R.D., Laidre, K.L., 2009. A novel method for identifying
behavioural changes in animal movement data. Ecology Letters 12, 395–408.
Gurney, W.S.C., McCauley, E., Nisbet, R.M., Murdoch, W.W., 1990. The physiological
ecology of Daphnia: a dynamic model of growth and reproduction. Ecology 71,
716–732.
Harley, C.D.G., Hughes, A.R., Hultgren, K.M., Miner, B.G., Sorte, C.J.B., Thornber, C.S.,
Rodriguez, L.F., Tomanek, L., Williams, S.L., 2006. The impacts of climate change
in coastal marine systems. Ecology Letters 9, 228–241.
Hobson, K.A., Stirling, I., Andriashek, D.S., 2009. Isotopic homogeneity of breath CO2
from fasting and berry-eating polar bears: implications for tracing reliance
on terrestrial foods in a changing Arctic. Canadian Journal of Zoology 87,
50–55.
Holland, M.M., Bitz, C.M., Tremblay, B., 2006. Future abrupt reductions in the
summer Arctic sea ice. Geophysical Research Letters 33, L23503.
Hunter, C.M., Caswell, H., Runge, M.C., Regehr, E.V., Amstrup, S.C., Stirling, I., 2007.
Polar bears in the Southern Beaufort Sea II: demography and population growth
in relation to sea ice conditions. USGS Science Strategy to Support US Fish and
Wildlife Service Polar Bear Listing Decision: US Geological Survey, Reston,
Virginia.
IPCC, 2007. Climate change 2007: synthesis report. Intergovernmental Panel on
Climate Change, Geneva, Switzerland.
Jenouvrier, S., Caswell, H., Barbraud, C., Holland, M., Stroeve, J., Weimerskirch, H.,
2009. Demographic models and IPCC climate projections predict the decline of
an emperor penguin population. Proceedings of the National Academy of
Sciences 106, 1844–1847.
Jonzén, N., Pople, A.R., Grigg, G.C., Possingham, H.P., 2005. Of sheep and rain: largescale population dynamics of the red kangaroo. Journal of Animal Ecology 74,
22–30.
Kiørboe, T., Bagøien, E., 2005. Motility patterns and mate encounter rates in
planktonic copepods. Limnology and Oceanography 50, 1999–2007.
Klanjscek, T., Nisbet, R.M., Caswell, H., Neubert, M.G., 2007. A model for energetics
and bioaccumulation in marine mammals with applications to the right whale.
Ecological Applications 17, 2233–2250.
Kokko, H., Mappes, J., 2005. Sexual selection when fertilization is not guaranteed.
Evolution 59, 1876–1885.
Kooijman, S.A.L.M., 2010. Dynamic Energy Budget Theory for Metabolic
Organisation, third ed. Cambridge University Press, Cambridge.
Kooijman, S.A.L.M., Sousa, T., Pecquerie, L., van der Meer, J., Jager, T., 2008. From
food-dependent statistics to metabolic parameters, a practical guide to the use
of dynamic energy budget theory. Biological Reviews 83, 533–552.
Krebs, C.J., Berteaux, D., 2006. Problems and pitfalls in relating climate variability to
population dynamics. Climate Research 32, 143–149.
Laidre, K.L., Stirling, I., Lowry, L.F., Wiig, Ø., Heide-Jorgensen, M.P., Ferguson, S.H.,
2008. Quantifying the sensitivity of Arctic marine mammals to climate-induced
habitat change. Ecological Applications 18, S97–S125.
Lee, P.C., Majluf, P., Gordon, I.J., 1991. Growth, weaning and maternal investment
from a comparative perspective. Journal of Zoology (London) 225, 99–114.
Lika, K., Nisbet, R.M., 2000. A dynamic energy budget model based on partitioning of
net production. Journal of Mathematical Biology 41, 361–386.
Ludwig, D., 1999. Is it meaningful to estimate a probability of extinction? Ecology
80, 298–310.
Mace, G.M., Collar, N.J., Gaston, K.J., Hilton-Taylor, C., Akcakaya, H.R., LeaderWilliams, N., Milner-Gulland, E.J., Stuart, S.N., 2008. Quantiﬁcation of extinction
risk: IUCN’s system for classifying threatened species. Conservation Biology 22,
1424–1442.
Maslanik, J.A., Fowler, C., Stroeve, J., Drobot, S., Zwally, J., Yi, D., Emery, W., 2007. A
younger, thinner Arctic ice cover: increased potential for rapid, extensive seaice loss. Geophysical Research Letters 34, L24501.
Mauritzen, M., Derocher, A.E., Pavlova, O., Wiig, Ø., 2003. Female polar bears, Ursus
maritimus, on the Barents Sea drift ice: walking the treadmill. Animal Behaviour
66, 107–113.
1622
P.K. Molnár et al. / Biological Conservation 143 (2010) 1612–1622
Mawdsley, J.R., O’Malley, R., Ojima, D.S., 2009. A review of climate-change
adaptation strategies for wildlife management and biodiversity conservation.
Conservation Biology 23, 1080–1089.
McCarty, J.P., 2001. Ecological consequences of recent climate change. Conservation
Biology 15, 320–331.
Molnár, P.K., 2009. Modelling future impacts of climate change and harvest on the
reproductive success of female polar bears (Ursus maritimus). Ph.D. Thesis,
University of Alberta.
Molnár, P.K., Derocher, A.E., Lewis, M.A., Taylor, M.K., 2008. Modelling the mating
system of polar bears: a mechanistic approach to the Allee effect. Proceedings of
the Royal Society B: Biological Sciences 275, 217–226.
Molnár, P.K., Klanjscek, T., Derocher, A.E., Obbard, M.E., Lewis, M.A., 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.
Moore, S.E., Huntington, H.P., 2008. Arctic marine mammals and climate change:
impacts and resilience. Ecological Applications 18, S157–S165.
Nisbet, R.M., Muller, E.B., Lika, K., Kooijman, S.A.L.M., 2000. From molecules to
ecosystems through dynamic energy budget models. Journal of Animal Ecology
69, 913–926.
Nisbet, R.M., McCauley, E., Gurney, W.S.C., Murdoch, W.W., Wood, S.N., 2004.
Formulating and testing a partially speciﬁed dynamic energy budget model.
Ecology 85, 3132–3139.
Noonburg, E.G., Nisbet, R.M., McCauley, E., Gurney, W.S.C., Murdoch, W.W., de Roos,
A.M., 1998. Experimental testing of dynamic energy budget models. Functional
Ecology 12, 211–222.
Oftedal, O.T., 1993. The adaptation of milk secretion to the constraints of fasting in
bears, seals, and baleen whales. Journal of Dairy Science 76, 3234–3246.
Øritsland, N.A., Jonkel, C., Ronald, K., 1976. A respiration chamber for exercising
polar bears. Norwegian Journal of Zoology 24, 65–67.
Parmesan, C., 2006. Ecological and evolutionary responses to recent climate change.
Annual Review of Ecology Evolution and Systematics 37, 637–669.
Ramsay, M.A., Stirling, I., 1986. On the mating system of polar bears. Canadian
Journal of Zoology 64, 2142–2151.
Ramsay, M.A., Stirling, I., 1988. Reproductive biology and ecology of female polar
bears (Ursus maritimus). Journal of Zoology (London) 214, 601–634.
Regehr, E.V., Lunn, N.J., Amstrup, S.C., 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, E.V., Hunter, C.M., Caswell, H., Amstrup, S.C., Stirling, I., 2010. Survival and
breeding of polar bears in the Southern Beaufort Sea in relation to sea ice.
Journal of Animal Ecology 79, 117–127.
Root, T.L., Schneider, S.H., 2006. Conservation and climate change: the challenges
ahead. Conservation Biology 20, 706–708.
Rosing-Asvid, A., 2006. The inﬂuence of climate variability on polar bear (Ursus
maritimus) and ringed seal (Pusa hispida) population dynamics. Canadian
Journal of Zoology 84, 357–364.
Schmidt-Nielsen, K., 1972. Locomotion: energy cost of swimming, ﬂying, and
running. Science 177, 222–228.
Serreze, M.C., Holland, M.K., Stroeve, J., 2007. Perspectives on the Arctic’s shrinking
sea-ice cover. Science 315, 1533–1536.
Skaare, J.U., Larsen, H.J., Lie, E., Bernhoft, A., Derocher, A.E., Norstrom, R., Ropstad, E.,
Lunn, N.J., Wiig, Ø., 2002. Ecological risk assessment of persistent organic
pollutants in the arctic. Toxicology 181, 193–197.
Smith, T.G., 1980. Polar bear predation of ringed and bearded seals in the land-fast
sea ice habitat. Canadian Journal of Zoology 58, 2201–2209.
Stirling, I., 1974. Midsummer observations on the behavior of wild polar bears.
Canadian Journal of Zoology 52, 1191–1198.
Stirling, I., Archibald, W.R., 1977. Aspects of predation of seals by polar bears.
Journal of the Fisheries Research Board of Canada 34, 1126–1129.
Stirling, I., Derocher, A.E., 1993. Possible impacts of climatic warming on polar
bears. Arctic 46, 240–245.
Stirling, I., Latour, P.B., 1978. Comparative hunting abilities of polar bear cubs of
different ages. Canadian Journal of Zoology 56, 1768–1772.
Stirling, I., Øritsland, N.A., 1995. Relationships between estimates of ringed seal
(Phoca hispida) and polar bear (Ursus maritimus) populations in the Canadian
Arctic. Canadian Journal of Fisheries and Aquatic Sciences 52, 2594–2612.
Stirling, I., Parkinson, C.L., 2006. Possible effects of climate warming on selected
populations of polar bears (Ursus maritimus) in the Canadian Arctic. Arctic 59,
261–275.
Stirling, I., Andriashek, D., Calvert, W., 1993. Habitat preferences of polar bears in
the western Canadian Arctic in late winter and spring. Polar Record 29, 13–24.
Stirling, I., Lunn, N.J., Iacozza, J., 1999. Long-term trends in the population ecology of
polar bears in Western Hudson Bay in relation to climate change. Arctic 52,
294–306.
Stroeve, J., Holland, M.M., Meier, W., Scambos, T., Serreze, M., 2007. Arctic sea ice
decline: faster than forecast. Geophysical Research Letters 34, L09501.
Sutherland, W.J., 2006. Predicting the ecological consequences of environmental
change: a review of the methods. Journal of Applied Ecology 43, 599–616.
Taylor, C.R., Heglund, N.C., Maloiy, G.M.O., 1982. Energetics and mechanics of
terrestrial locomotion. I. Metabolic energy consumption as a function of speed
and body size in birds and mammals. Journal of Experimental Biology 97, 1–21.
Taylor, M.K., Laake, J., Cluff, H.D., Ramsay, M., Messier, F., 2002. Managing the risk
from hunting for the Viscount-Melville Sound polar bear population. Ursus 13,
185–202.
Thomas, D.W., Blondel, J., Perret, P., Lambrechts, M.M., Speakman, J.R., 2001.
Energetic and ﬁtness costs of mismatching resource supply and demand in
seasonally breeding birds. Science 291, 2598–2600.
Towns, L., Derocher, A.E., Stirling, I., Lunn, N.J., Hedman, D., 2009. Spatial and
temporal patterns of problem polar bears in Churchill, Manitoba. Polar Biology
32, 1529–1537.
Visser, M.E., Van Noordwijk, A.J., Tinbergen, J.M., Lessells, C.M., 1998. Warmer
springs lead to mistimed reproduction in great tits (Parus major). Proceedings of
the Royal Society B: Biological Sciences 265, 1867–1870.
Walther, G.R., Post, E., Convey, P., Menzel, A., Parmesan, C., Beebee, T.J.C., Fromentin,
J.M., Hoegh-Guldberg, O., Bairlein, F., 2002. Ecological responses to recent
climate change. Nature 416, 389–395.
Watts, P.D., Øritsland, N.A., Hurst, R.J., 1987. Standard metabolic rate of polar bears
under simulated denning conditions. Physiological Zoology 60, 687–691.
Wiig, Ø., Aars, J., Born, E.W., 2008. Effects of climate change on polar bears. Science
Progress 91, 151–173.
Wunder, B.A., 1975. A model for estimating metabolic rate of active or resting
mammals. Journal of Theoretical Biology 49, 345–354.
Zhang, X.D., Walsh, J.E., 2006. Toward a seasonally ice-covered Arctic Ocean:
scenarios from the IPCC AR4 model simulations. Journal of Climate 19, 1730–
1747.
Zonneveld, C., Kooijman, S.A.L.M., 1989. Application of a general energy budget
model to Lymnaea stagnalis. Functional Ecology 3, 269–278.
Zoo Biology 26:187–199 (2007)
Research Article
Seminal and Endocrine Characteristics
of Male Pallas’ Cats (Otocolobus manul)
Maintained Under Artificial Lighting
With Simulated Natural Photoperiods
Annie Newell-Fugate,1 Suzanne Kennedy-Stoskopf,2 Janine L. Brown,3
Jay F. Levine,2 and William F. Swanson4
1
Veterinary Wildlife Unit, Faculty of Veterinary Science, University of Pretoria, Pretoria,
South Africa
2
Population Health and Pathobiology, College of Veterinary Medicine, North Carolina
State University, Raleigh, North Carolina
3
Conservation and Research Center, National Zoological Park, Smithsonian Institution,
Front Royal, Virginia
4
Center for Conservation and Research of Endangered Wildlife, Cincinnati Zoo and
Botanical Garden, Cincinnati, Ohio
Pallas’ cats (Otocolobus manul) have a pronounced reproductive seasonality
controlled by photoperiod. Previous studies of reproduction in captive Pallas’ cats
exposed to natural light showed a breeding season of December–April. This study
evaluated the impact of artiﬁcial lighting timed to simulate natural photoperiods
on male reproductive seasonality of four Pallas’ cats housed indoors. Semen
evaluation, blood collection, and body weight measurements were conducted
every 1–2 months from November 2000–June 2001. Fecal samples were collected
from each male twice weekly to assess testosterone and corticoid concentrations.
Mean values for reproductive traits (sperm attributes, testicular volume) were
highest from February–April, the deﬁned breeding season. Fecal testosterone
concentrations were highest from mid-January to mid-March. Male Pallas’ cats
managed indoors under simulated photoperiods experienced a delayed onset of
the breeding season by 1–2 months and a decreased length of the breeding season.
Grant sponsors: IAMS Company; Disney’s Animal Kingdom; Cincinnati Zoo Center for Conservation
and Research of Endangered Wildlife; North Carolina Zoological Society.
Correspondence to: Dr. Suzanne Kennedy-Stoskopf, North Carolina State University, College of
Veterinary Medicine, 4700 Hillsborough Street, Raleigh, NC 27606. E-mail: suzanne_stoskopf@ncsu.edu
Received 4 April 2006; Revised 20 November 2006; Accepted 29 December 2006
DOI 10.1002/zoo.20127
Published online 27 April 2007 in Wiley InterScience (www.interscience.wiley.com).
r 2007 Wiley-Liss, Inc.
188 Newell-Fugate et al.
Over the course of the study, fecal corticoid concentrations did not seem to differ
among seasons. Although mating attempts during this study were unsuccessful,
subsequent pairings of male and female Pallas’ cats in the same research colony
during the 2002 and 2003 breeding seasons produced viable offspring. These
results suggest that male Pallas’ cats, housed indoors under simulated
photoperiods, exhibit distinct reproductive cyclic patterns, characterized by a
delayed and truncated breeding season. Adrenocortical activity varied among
individuals, but did not adversely affect reproductive parameters. Housing Pallas’
cats indoors under simulated photoperiods may represent a viable strategy
for maintaining breeding success while limiting disease exposure. Zoo Biol

c 2007 Wiley-Liss, Inc.
26:187–199, 2007.
Keywords: felids; testosterone; cortisol; leptin; spermatozoa; reproductive seasonality
INTRODUCTION
The Pallas’ cat (Otocolobus manul), a small-sized felid species endemic to
Central Asia, is threatened with extinction in the wild due to habitat loss, hunting
and vermin control programs [Nowell and Jackson, 1996]. To bolster captive
breeding and conservation efforts, 24 wild-born Pallas’ cats were imported from
Russia and Mongolia in the mid-1990s to establish a founder population in North
American zoos [Caron, 2004]. In 2001, the American Zoo and Aquarium
Association (AZA) created a Species Survival Plan (SSP) to manage Pallas’ cats in
captivity as a self-sustaining, genetically viable population.
Pallas’ cats are known to have a pronounced reproductive seasonality with
breeding occurring only during the Northern-hemisphere winter months [Swanson
et al., 1996; Brown et al., 2002]. This extreme seasonality is controlled primarily by
circannual variation in photoperiod, possibly mediated through changes in
circulating melatonin or leptin concentrations. In seasonally breeding species,
melatonin secretion from the pineal gland increases with decreasing daily light
exposure to either stimulate or inhibit gonadotropin-releasing hormone (GnRH)
release from the hypothalamus. Similarly, in species that exhibit seasonal ﬂuctuations in body weight, variations in leptin secretion from adipose cells may inﬂuence
GnRH release and the onset of reproductive activity [Goodman, 1999]. Under
natural photoperiods, male Pallas’ cats experience weight gains beginning in autumn,
and exhibit increases in testosterone concentrations, sperm production, and sperm
quality from December to April [Swanson et al., 1996; Brown et al., 2002]. Improvements in male reproductive parameters tend to bracket the months (January–March)
of observed female estrus cyclicity and breeding activity [Mellen, 1998; Brown et al.,
2002].
The reproductive knowledge gained from scientiﬁc research on Pallas’ cats has
beneﬁted captive propagation efforts in North American zoos. Between 1996–2001,
the captive breeding program for Pallas’ cats was highly productive, with 65 kittens
born in 17 litters. Unfortunately, approximately 60% of these kittens died within
4 months of birth, due mostly to toxoplasmosis [Swanson, 1999]. In the wild, Pallas’
cats rarely encounter Toxoplasma gondii in their natural prey species [Brown et al.,
2005], but in captivity, exposure to this parasite can occur through diets consisting of
raw horse meat, whole prey items (mice, chicks), and live rodents and birds captured
by cats in their exhibits. To eliminate these potential sources of T. gondii exposure,
Zoo Biology DOI 10.1002/zoo
Reproductive Traits of Male Pallas’ Cats 189
the AZA’s Felid Taxon Advisory Group has recommended that zoos provide Pallas’
cats with a T. gondii-free diet and house animals in indoor exhibits with natural
lighting or, if possible, under simulated natural photoperiods [Swanson, 2000]. Using
artiﬁcial light to simulate natural photoperiods, it may prove feasible to induce
appropriate seasonal reproductive patterns in Pallas’ cats in an indoor environment.
However, before implementing this disease management strategy across institutions,
the potential impact on reproductive parameters must be investigated thoroughly.
In the present study, four male Pallas’ cats were housed under ﬂuorescent
lighting timed to simulate natural photoperiods and assessed temporally for seminal
traits, serum testosterone and leptin concentrations, and fecal androgen metabolite
proﬁles. In addition, fecal corticoids were measured to evaluate basal adrenocortical
activity. Our speciﬁc objectives were to: 1) determine if male Pallas’ cats housed
indoors under simulated natural photoperiods exhibit the species’ characteristic
reproductive seasonality, and 2) assess temporal variation in adrenocortical activity
in cats housed under artiﬁcial lighting timed to simulate natural photoperiods.
MATERIALS AND METHODS
Animals and Fecal Sample Collection
In July 2000, a Pallas’ cat research colony was established at the Laboratory
Animal Research facility at North Carolina State University’s College of Veterinary
Medicine. Six Pallas’ cats (4 males, 2 females) exhibited recurring clinical signs from
feline herpes virus (FHV-1). Because these cats represented disease risks to other
felids housed in close proximity, the herpes-infected Pallas’ cats were donated
to North Carolina State University. The research colony was created to facilitate
detailed studies of disease and reproduction in Pallas’ cats and to apply this
knowledge to rescue valuable genes offered by these founders to the Pallas’ cat SSP
population. To prevent potential disease transmission between herpes-infected
Pallas’ cats and other research cat populations, the Pallas’ cats were maintained in an
isolated, controlled-access indoor laboratory animal facility.
Four adult male Pallas’ cats were monitored in this study. Two males (SB ]291
and SB ]297) were wild-born (approximately 6–8 years of age) and two (SB ]390 and
SB ]391) were captive-born (2 years of age) offspring of one of the wild-born males.
One adult, wild-born female Pallas’ cat (approximately 6 years of age) was
maintained in the same facility during the study period. The other female died before
the start of the study with meningoencephalitis caused by T. gondii. Animals were
housed indoors individually in adjacent wire mesh pens, except during the expected
breeding season when one wild-born male was paired with the female.
Cats did not have access to natural lighting through skylights or windows, but
were housed under standard ﬂuorescent lighting (i.e., not full-spectrum), with lightdark cycles changed weekly to mimic the natural photoperiod of the colony’s
geographic latitude (35.46.591N). Ambient temperature and humidity in the colony
were controlled and kept within recommended ranges for laboratory animal facilities
(68–721F; 20–70% humidity). Animals were provided with a daily diet consisting of
raw horse meat, supplemented with vitamins and minerals (Millikin Diet, Toronto,
ON), in addition to whole mice (from a gnotobiotic mouse colony) once a week. All
cats were weighed once a month. Fecal samples were collected from each male cat
Zoo Biology DOI 10.1002/zoo
190 Newell-Fugate et al.
two times per week for 8 months (November 2000–June 2001). Fecal samples also
were collected two times per week for 5 months (November 2000–March 2001) from
the female. When the cats were paired for breeding, 1/4–1/2 tsp red food coloring
paste (Christmas Red, Wilton Industries, Woodridge, IL) was added to the female’s
diet to allow differentiation of the male and female fecal samples.
Semen Collection and Analysis
Reproductive evaluations were carried out on anesthetized male Pallas’ cats
at 2-month intervals from December 2000–June 2001. For anesthesia, males were
injected (i.m.) with a mixture of ketamine hydrochloride (Ketaset, Fort Dodge
Animal Health, Fort Dodge, IA; 2 mg/kg body weight), medetomidine hydrochloride (Domitor, Pﬁzer Animal Health, Exton, PA; 0.04 mg/kg), and butorphanol
tartrate (Torbugesic, Fort Dodge Animal Health; 0.4 mg/kg). If necessary,
anesthesia was supplemented with isoﬂurane (0.5–1%), administered via face mask.
After sample collection, medetomidine was reversed with atipamezole hydrochloride
(Antisedan, Pﬁzer Animal Health; 0.2 mg/kg; i.m.). Semen was collected from
anesthetized males via electroejaculation, using an AC 60 Hz electrostimulator (PT
Electronics, Boring, OR) and a three-electrode rectal probe (0.6 cm diameter). A
standardized collection protocol was followed, consisting of three to four separate
sets (20–30 stimuli/set) of electrical stimuli (2–5 V), with 5-min intervals between sets
[Howard, 1993; Swanson et al., 1996]. Testicular parameters (W, width; L, length)
were measured using calipers to calculate testicular volume (V 5 W2 (cm) L
(cm) 0.524). Recovered semen was assessed for volume (ml) and evaluated
microscopically (100 magniﬁcation) for presence or absence of spermatozoa,
percent sperm motility (0–100%), and rate of sperm forward progression (scale of
0–5; 0 5 non-motile, 5 5 rapid forward progression). A sperm motility index (SMI)
was calculated for each sample [SMI 5 (% sperm motility1(20 rate of progressive
movement)/2]. Sperm concentration ( 106/ml) was determined manually (hemacytometer method), and an aliquot of raw semen (5 ml) was ﬁxed in 0.3%
glutaraldehyde (50 ml) for later morphologic examination (100 sperm/sample;
640 magniﬁcation). After semen collection, a blood sample was collected via
jugular venapuncture for assessment of serum testosterone and leptin concentrations
(ng/ml).
Fecal and Serum Hormone Analysis
Fecal testosterone and corticoid metabolites were extracted from fecal samples
as described by Brown et al. [1994, 1996]. Fecal samples were lyophilized, pulverized,
and fecal powder (0.2 g) was boiled in 5 ml aqueous ethanol (90%) for 20 min. To
monitor extraction efﬁciency, 100 ml of 3H-testosterone was added to samples before
extraction. After centrifuging at 1,500 rpm for 10 min, the ethanol supernatant was
decanted and saved. The fecal pellet was resuspended in 5 ml of 90% ethanol,
vortexed (1 min), and re-centrifuged at 1,500 rpm (10 min). Ethanol extracts were
combined, dried under compressed air, and dissolved in 1 ml of methanol. Extracts
were diluted with EIA buffer (0.09 M NaH2PO4, 0.12 M Na2HPO4, 0.075 M NaCl,
pH 7.0) for analysis. Steroid extraction efﬁciency was 485%.
Fecal testosterone and corticoid metabolites were assessed with enzyme
immunoassays (EIA) [Munro and Lasley, 1988], using antibody and horseradish
peroxidase (HRP) labeled tracers provided by C. Munro (University of California,
Zoo Biology DOI 10.1002/zoo
Reproductive Traits of Male Pallas’ Cats 191
Davis, CA). The testosterone EIA used a polyclonal anti-testosterone antibody
(R156/7), a testosterone-horseradish peroxidase ligand, and testosterone standards
(Sigma-Aldrich, St. Louis, MO). The EIA was carried out in 96-well microtiter plates
(Nunc-Immuno, Maxisorp Surface; Fisher Scientiﬁc, Pittsburgh, PA) coated
14–18 hr previously with testosterone antiserum (50 ml per well; diluted 1:7,500 in
coating buffer; 0.05 M NaHCO3, pH 9.6). The testosterone antiserum had the
following cross reactivities: testosterone (100%), 5a-dihydrotestosterone (57.37%),
androstenedione (0.27%), androsterone, cholesterol, 17b-estradiol, progesterone,
pregnenolone and hydrocortisone (o0.05%). Fecal extracts, evaporated to dryness
and diluted 1:50–1:1,600 in steroid buffer (0.1 M NaPO4, 0.149 M NaCl, pH 7.0),
were assayed in duplicate. Testosterone standards (50 ml, range 5 2.3–600 pg/well,
diluted in assay buffer, 0.1 M NaPO4, 0.149 M NaCl, 0.1% bovine serum albumin,
pH 7.0) and sample (50 ml) were combined with testosterone-HRP (50 ml, 1:15,000
dilution in assay buffer). After incubation at room temperature for 2 hr, plates were
washed ﬁve times before 100 ml ABTS substrate [0.4 mM 2,20 -azino-di-(3-ethylbenzthiazoline sulfonic acid) di ammonium salt, 1.6 mM H2O2, 0.05 M citrate, pH
4.0] was added to each well. After incubation at room temperature on a shaker for
up to 60 min, the absorbance was measured at 405 nm. Parallel displacement curves
were obtained by comparing serial dilutions of pooled fecal extracts with the
testosterone standard preparation.
The cortisol assay used a cortisol-HRP ligand and polyclonal antiserum (R486)
and cortisol standards (hydrocortisone; Sigma-Aldrich). The polyclonal antiserum
cross reacted with cortisol (100%), prednisolone (9.9%), prednisone (6.3%),
cortisone (5%), and corticosterone (o1%) [Young et al., 2004]. The EIA was
carried out in 96-well microtiter plates coated 14–18 hr previously with cortisol
antiserum (50 ml per well; diluted 1:20,000 in coating buffer; 0.05 M NaHCO3, pH
9.6). Fecal extracts were diluted 1:50 in steroid buffer and assayed in duplicate.
Cortisol standards (50 ml, range 3.9–1,000 pg/well) and sample (50 ml) were combined
with cortisol-HRP (50 ml, 1:8,500 dilution in assay buffer). After incubation at room
temperature for 1 hr, plates were washed before addition of ABTS substrate as
described above. After incubation on a shaker for 10–15 min, the absorbance was
measured at 405 nm. Parallel displacement curves were obtained by comparing serial
dilutions of pooled fecal extracts with the cortisol standard preparation.
Serum testosterone was analyzed using a solid phase radioimmunoassay
(Diagnostic Products, Los Angeles, CA). The leptin assay (Multi-Species Leptin RIA
Kit, Linco Research, St. Louis MO) used in this study has been validated previously
for use in the domestic cat [Backus, 2000]. Serum assays were validated for Pallas’
cats by showing parallel displacement of serial serum dilutions compared to the
respective standard preparations. Intra- and interassay coefﬁcients of variation for
all assays were o10% and 15%, respectively. Assay sensitivity at 90% maximum
binding was 2.3 pg/well for testosterone, 3.9 pg/well for cortisol, and 1 ng/ml HE
(human equivalent) for leptin assays (Linco).
Statistical Analysis
For analysis of seminal values and other reproductive traits, the male breeding
season was deﬁned as the time period from February–April, based on observations
of breeding activity and sharp increases and declines in sperm production and
Zoo Biology DOI 10.1002/zoo
192 Newell-Fugate et al.
quality. The small number of animals examined (n 5 4) precluded an assessment of
data normality and subsequent valid statistical analysis. Access to adequate numbers
of animals is a common problem when working with non-domestic species in
captivity, but the descriptive data supports the objectives of the study.
RESULTS
Seminal Traits, Serum Hormones, and Breeding Activity
Male Pallas’ cats exhibited substantial temporal variation in seminal traits
during the study period (Table 1), with peak values observed typically during the
breeding season (February–April). Mean values for serum testosterone, serum leptin,
and body weight peaked in February, before increases in seminal values (Table 1).
The male Pallas’ cat that was paired with a female in 2001 was observed mating
on March 14. The female ovulated, based on an increase in fecal progesterone
metabolite concentrations (data not shown), but no fetuses were observed on
ultrasonographic exam at 30 days post-breeding and no kittens were produced.
Fecal Testosterone Metabolite Profiles
Fecal testosterone metabolite concentrations exhibited a marked increase from
mid-January through mid-March (Fig. 1). This time period preceded, by 1–2
months, improvements in seminal traits observed in February and declines in
seminal quality seen by June. Means, standard deviations, and ranges for individual
animals during the breeding and non-breeding seasons are presented in Table 2.
Notably, the male (SB ]297) that was paired with the female during the 2001
breeding season had the highest mean fecal testosterone metabolite concentration
among the four males.
Fecal Corticoid Metabolite Profiles
Fecal corticoid concentrations for each cat varied during the 8-month study
period (Fig. 2). A small spike in the fecal corticoid proﬁle of male SB ]297 occurred
on March 2, 2001 when this male was paired with female SB ]292. His fecal corticoid
concentration that day was 1,164.7 ng/g of feces (Fig. 3). Twelve days later he was
observed mating. Additionally, a much larger increase in fecal corticoids occurred
for all cats combined during the middle of May 2001, when repair work was being
conducted in the room where the cats were housed. Means, standard deviations, and
ranges for individual animals during the breeding and non-breeding seasons are
presented in Table 3.
DISCUSSION
Endocrine and Physiologic Reproductive Parameters
This study shows that male Pallas’ cats maintained under simulated natural
photoperiods exhibit pronounced temporal variation in morphologic, seminal, and
endocrine parameters. Our ﬁndings conﬁrm that the breeding season in males
is controlled primarily by seasonal changes in day length. Observed physiologic
changes mirror seasonal ﬁndings from earlier studies of Pallas’ cats housed
in outdoor exhibits [Swanson et al., 1996; Brown et al., 2002]. In both environments,
Zoo Biology DOI 10.1002/zoo
0–0.2
NAc
31–54.0
1.5–2.3
2.8–4.6
4.8–5.9
4.7–4.9
NAc
31.3711.7
2.070.4
3.370.8
5.270.5
4.870.1
38–88.0
0–7.0
Range
0.170.1a
51.7731.5
2.374.0a
a
Mean7SD
64.2725.3d
47.7717.1d
2.870.6
5.370.9
6.472.0
5.270.2
2.572.9
103.5745.5
29.0736.5
Mean7SD
35–80.0
28–59.0
2.1–3.5
4.2–6.2
4.3–9.1
4.9–5.4
0–6.7
72–171.0
0–82.0
Range
February 2001
70.875.2e
52.0716.5e
2.570.6
3.771.8
3.671.3
4.470.6
5.775.4e
164761.5
51.3761.5e
b
e
65–75
41–71
2.2–3.3
2.1–5.7
2.1–4.7
4.1–5.3
2–11.9
93–200
10–128
Range
April 2001
Mean7SD
One male had urine in the ejaculate (n 5 3).
SMI 5 [% sperm motility1(20 rate of forward progression)]/2.
c
NA, not assessed due to low sperm concentrations and urine contamination in one ejaculate.
d
One male had no sperm in the ejaculate (n 5 3).
e
One male did not produce an ejaculate (n 5 3).
f
Onl…

Turn in your highest-quality paper
Get a qualified writer to help you with

“ Arizona State University LAB Seasons Tilt Orbit and Temperature Paper ”

Get high-quality paper

Guarantee! All work is written by expert writers!

Still stressed from student homework?

Get quality assistance from academic writers!

Order now