Ecology, 95(9), 2014, pp. 2526–2536
 2014 by the Ecological Society of America
Environmental correlates of temporary emigration for female
Weddell seals and consequences for recruitment
GLENN E. STAUFFER,1,3 JAY J. ROTELLA,1 ROBERT A. GARROTT,1 AND WILLIAM L. KENDALL2
1Department of Ecology, Montana State University, Bozeman, Montana 59717 USA
2U.S. Geological Survey, Colorado Cooperative Fish and Wildlife Research Unit, Fort Collins, Colorado 80523 USA
Abstract. In colonial-breeding species, prebreeders often emigrate temporarily from natal
reproductive colonies then subsequently return for one or more years before producing young.
Variation in attendance–nonattendance patterns can have implications for subsequent
recruitment. We used open robust-design multistate models and 28 years of encounter data for
prebreeding femaleWeddell seals (Leptonychotes weddellii [Lesson]) to evaluate hypotheses about
(1) the relationships of temporary emigration (TE) probabilities to environmental and
population size covariates and (2) motivations for attendance and consequences of
nonattendance for subsequent probability of recruitment to the breeding population. TE
probabilities were density dependent (bˆBPOP¼ 0.66, cSE¼ 0.17; estimated effects [b] and standard
errors of population size in the previous year) and increased when the fast-ice edge was distant
from the breeding colonies (bˆDIST ¼ 0.75, cSE ¼ 0.04; estimated effects and standard errors of
distance to the sea-ice edge in the current year on TE probability in the current year) and were
strongly age and state dependent. These results suggest that trade-offs between potential benefits
and costs of colony attendance vary annually and might influence motivation to attend colonies.
Recruitment probabilities were greatest for seals that consistently attended colonies in two or
more years (e.g., wˆage10 ¼ 0.56, SD ¼ 0.17) and lowest for seals that never or inconsistently
attended prior to recruitment (e.g., wˆage10 ¼ 0.32, SD ¼ 0.15), where wˆage10 denotes the mean
recruitment probability (over all years) for 10-year-old seals for the specified prebreeder state. In
colonial-breeding seabirds, repeated colony attendance increases subsequent probability of
recruitment to the adult breeding population; our results suggest similar implications for a marine
mammal and are consistent with the hypothesis that prebreeders were motivated to attend
reproductive colonies to gain reproductive skills or perhaps to optimally synchronize estrus
through close association with mature breeding females.
Key words: capture–mark–recapture; colonial breeding; density dependence; environmental variability;
marine mammal; McMurdo Sound, Antarctica; open robust-design multistate; pinniped; Ross Sea; trade-offs;
unobservable states.
INTRODUCTION
In long-lived colonial-breeding species, and some
noncolonial, territorial species, it is common for
juveniles to be absent from reproductive areas for one
or several years postbirth or posthatching and to then
subsequently attend these areas for a few years before
recruiting (here recruitment refers to first production of
offspring rather than first mating attempt) into the
breeding population (e.g., Cadiou et al. 1994, Beauplet
et al. 2005, Jenouvrier et al. 2008). When sampling
occurs only at breeding sites, individuals are in an
unobservable state when they are temporarily absent
from those sites, and entry into or remaining in such a
state is termed temporary emigration (TE; Kendall and
Bjorkland 2001). In some species (e.g., Halley et al.
1995, Dittmann and Becker 2003), most prebreeders
follow similar patterns of nonattendance/attendance,
but in other species the number of years absent from
breeding colonies or the number of years prebreeders
attend colonies before recruiting varies considerably
depending on characteristics of the individual or the
environment (e.g., Beauplet et al. 2005, Jenouvrier et al.
2008). Understanding the causes and consequences of
variation in temporary emigration is of considerable
ecological interest given possible implications of atten-
dance or nonattendance for age of recruitment or future
reproductive success (Cam et al. 2002, Dittmann et al.
2007).
In several colonial-breeding species where TE by
juveniles has been explicitly investigated, population-
level TE probabilities vary considerably with
individual- and population-level attributes (Beauplet et
al. 2005, Jenouvrier et al. 2008, Stauffer et al. 2013a). TE
probabilities typically decline with increasing age,
possibly because older individuals have increased
Manuscript received 18 October 2013; revised 14 February
2014; accepted 4 March 2014. Corresponding Editor: J. P. Y.
Arnould.
3 Present address: 419 Forest Resources Building, Penn-
sylvania State University, University Park, Pennsylvania
16802 USA. E-mail: ges162@psu.edu
2526
motivation or ability to navigate to colonies while
avoiding potential costs of attendance related to limited
foraging opportunities or intraspecific conflict at breed-
ing colonies (Siniff et al. 1977, Furness and Birkhead
1984, Dittmann and Becker 2003). Variation in food
supply or foraging opportunities driven by environmen-
tal conditions can influence growth rates or body
condition of individuals (Madsen and Shine 2000), and
thus individuals should be less likely to attend colonies
when unfavorable conditions compromise body condi-
tion (Frederiksen and Bregnballe 2000, Jenouvrier et al.
2008). There is circumstantial evidence that TE proba-
bilities might also be density dependent, although the
mechanism is unknown (Rotella et al. 2009). Despite the
prevalence of TE in the life history of most colonial-
breeding species, rigorous estimation of TE probabili-
ties, especially as a function of environmental and
population- or individual-level covariates, remains rare
in population studies.
One possible motivation for attendance at reproduc-
tive colonies by prebreeders is the opportunity to
prospect for information or acquire reproductive skills
by observing conspecifics. Possible prospecting behavior
and consequent benefits has been investigated mostly in
avian species, especially colonial seabirds (see review in
Reed et al. [1999]). Prebreeders that attend colonies for
multiple years are more likely to recruit and breed
successfully than individuals attending colonies for the
first time (Halley et al. 1995, Dittmann et al. 2007). The
age of recruitment and the extent to which prebreeders
practice breeding behavior have also been associated
with subsequent breeding success (Cam et al. 2002).
Prospecting Kittiwakes (Rissa tridactyla) sometimes
squat on active nests temporarily unoccupied by the
nest owners, and this behavior appears to facilitate
future mate pairing or acquisition of nearby nest sites
(Cadiou et al. 1994, Cam et al. 2002). Wandering
Albatross (Diomedea exulans) spend 2–8 yr interacting
with conspecifics at reproductive colonies before recruit-
ing and frequently pair up with a reproductive partner
.1 yr before making their first breeding attempt
(Pickering 1989). Prospecting might thus inform deci-
sions about where or even how to recruit, especially
when prebreeders participate in behavior typically
associated with reproduction. That is, attendance at
reproductive colonies by prebreeders and observation of
conspecifics might facilitate social learning (see review in
Galef and Laland [2005]) and acquisition of social skills
necessary for successful recruitment (Becker and Brad-
ley 2007).
Some pinniped species also have a life history of
prolonged absence of prebreeders followed by 1 yr of
colony attendance before recruitment (e.g., Beauplet et
al. 2005, Hadley et al. 2006), but the implications of
attendance for subsequent recruitment or breeding
success are not as well documented in this group as for
colonial seabirds. Prebreeders might benefit from
observing pup births, interactions between mothers
and pups during lactation, or when pups first begin to
swim, or by observation of mating interactions and
potential mate quality. Mate choice by females has been
shown for Antarctic fur seals (Arctocephalus gazella;
Hoffman et al. 2007) and is a common feature of lekking
species (see review in Balmford 1991).
Unlike birds, pinnipeds have long gestation periods
such that individuals producing young in a given year
must mate in the previous year. Reproductive colonies
are foci of mating activity, and thus another possible
motivation for colony attendance by female prebreeders
is initiation or synchronization of estrus cycles by close
association with reproductive adult females. Such
synchronization occurs in several mammalian taxa,
including humans (McClintock 1978, Handelmann et
al. 1980, Stern and McClintock 1998). Although some
seals likely mate while absent from reproductive colonies
(de Bruyn et al. 2011), attending reproductive colonies
might increase the probability that first-time estrus is
optimally timed.
We used 28 yr (1983–2010) of capture–mark–recap-
ture data for female prebreeders from a population of
Weddell seals (Leptonychotes weddellii ) breeding in
Erebus Bay, Antarctica to investigate (1) correlates of
variability in age-specific TE probabilities in prebreeders
and (2) possible motivations for attendance and
consequences of nonattendance for recruitment proba-
bilities. TE and recruitment probabilities in this popu-
lation are highly variable (Hadley et al. 2006, Stauffer et
al. 2013a), and annual environmental conditions in and
around Erebus Bay also are highly variable. Multiple
decades of encounter data are available for individually
marked seals, and because seals are easily approached,
they are highly detectable and reproductive status can be
reliably identified. We evaluated support for (1) the
hypothesis that age-specific probability of TE is related
to variability in the environment, local population size,
or both and (2) the nonexclusive hypotheses that
attendance by prebreeders is motivated by opportunities
to gather information or acquire skills, or to facilitate
timing or onset of estrus. We operationally defined a
prebreeder as any female individual that had not yet
produced her first pup, even though she may have
mated. Consequently, we defined recruitment as first
reproduction of offspring rather than first mating
attempt.
METHODS
Study area and population
Erebus Bay is located on the western side of Ross
Island in McMurdo Sound, Antarctica. Approximately
300–600 Weddell seal pups are born on the ice surface
each austral spring at several areas throughout Erebus
Bay where tidal action and glacial pressures on fast ice
result in perennial sea-ice cracks that provide reliable
access to the surface of the fast ice (Stirling 1969, Siniff
et al. 1977). Females give birth to a single pup after a
gestation period of about 11 months, including about
September 2014 2527EMIGRATION AND RECRUITMENT BY SEALS
two months of diapause, and then primarily use stored
reserves to sustain a period of lactation and close
association with the pup (see Plate 1) lasting about 30–
45 d (Smith 1966, Hill 1987). Males defend mating
territories under the ice, where mating takes place near
the end of the pupping season (Siniff et al. 1977, Hill
1987).
Food resources for seals are probably limited in
Erebus Bay (Testa et al. 1985), and adults and pups are
believed to move out into food-rich foraging areas in the
southwest Ross Sea following the pupping season
(Smith 1965, Burns et al. 1999, Stewart et al. 2000).
Females born in Erebus Bay typically emigrate tempo-
rarily for 1 yr following birth, then attend reproductive
colonies in Erebus Bay for 1 yr (but not always in
consecutive years) before recruiting into the pup-
producing portion of the population (Hadley et al.
2006, Cameron et al. 2007, Stauffer et al. 2013a). Despite
the fact that female Weddell seals physiologically can
reproduce by age 2–4 yr (Smith 1966), the mean age at
first reproduction is estimated as 7.6 yr (Hadley et al.
2006).
Data collection
In this study, we used data from 5550 female Weddell
seals tagged as pups in Erebus Bay during the pupping
seasons of 1983–2010. During these years, nearly all
pups born in Erebus Bay were individually marked with
a pair of tags in the interdigital webbing of each hind
flipper (four identical tags total); broken or missing tags
on adults were subsequently replaced as necessary.
Consequently, a large percentage of the females in
Erebus Bay each year were individually identifiable and
of known age. Resighting data primarily included
observations from five-to-eight intensive surveys con-
ducted during early November through mid-December
in Erebus Bay each year (3–6 d between surveys). To
maximize detection probability, we also included obser-
vations within Erebus Bay from the 2–3 d preceding and
following each intensive survey. Detections occurred
only within the Erebus Bay study area, the boundaries of
which remained consistent each year. We assumed that
mothers born within the study area produced their own
pups within the study area, and in Discussion we address
possible violation of this assumption.
Covariates of temporary emigration
We considered the following covariates of TE: (1)
winter sea ice extent (SIE) for the Ross Sea sector
measured in September of year t just before the pupping
season; (2) Southern Oscillation Index (SOI) averaged
for the winter (July–September) in year t; (3) sea-ice
concentration (SIC) averaged over eight 25-km grid cells
in and around McMurdo Sound for November of year t;
(4) shortest distance (DIST) from the study area to an
open water polynya at the edge of the fast ice during the
breeding season of year t; and (5) the size of the breeding
population of females (BPOP) in year t  1. Note that
SIE and DIST are uncorrelated because the distance to
the polynya at the edge of the fast ice is independent of
the extent of pack ice beyond the polynya. Winter sea ice
is important in the life cycle of krill (Loeb et al. 1997),
and increased krill abundance may benefit Antarctic
silverfish (Pleurogramma antarcticum), a primary food
source for Weddell seals (Burns et al. 1998). Addition-
ally, Weddell seals, given their excellent diving abilities
(Castellini et al. 1992), probably are better able to
exploit resources under extensive ice sheets than are
other Antarctic predators, and extensive winter sea ice
may reduce competition for prey. Thus, food resources
in the Ross Sea should be most plentiful when sea ice is
extensive. SOI is a broadscale climatic index commonly
used as a measure of El Nin˜o/Southern Oscillation
(ENSO) strength in the southern Ocean. SOI is
positively correlated with sea ice extent in the Ross
Sea (Kwok and Comiso 2002), and positive SOI phases
have been linked to increases in seal pup weaning mass,
population size, and reproductive rates (Testa et al.
1991, Proffitt et al. 2007, Rotella et al. 2009). We
predicted decreased TE probabilities associated with SIE
or positive SOI phases because seals should be in good
condition and more likely to forego foraging opportu-
nities while attending colonies when resources are
plentiful. We expected SIC and DIST to influence access
to reproductive colonies by determining the distance
seals have to swim under ice to reach haul-out areas in
Erebus Bay. Thus we predicted a positive relationship
between TE and DIST or SIC. Lastly, given the
suggestion by Rotella et al. (2009) that TE probabilities
of adult female Weddell seals might be density
dependent, we predicted a positive relationship between
TE and BPOP.
SIE and SIC data were obtained from the National
Snow and Ice Data Center and SOI data from the
Australian Government, Bureau of Meteorology SOI
Archive (data available online).4,5 DIST was measured
from georeferenced satellite images dated as close as
possible to 1 November. Images were not available for
all years so we used multiple imputation methods
(Schafer 1999) to evaluate the influence of DIST on
TE probabilities (details in Appendix A). Environmental
covariates were standardized (mean ¼ 0, SD ¼ 1), and
population sizes were log-transformed in all analyses.
Data analysis
We used open robust-design multistate (ORDMS;
Kendall and Bjorkland 2001, Kendall 2004) models to
estimate transition probabilities (denoted with the
symbol w) of Weddell seals between seven demographic
states (see Appendix A for full modeling details). We
defined the seven states (Appendix A: Fig. A1) as P0
(pup); U0 (unobservable prebreeder, never previously
observed in Erebus Bay subsequent to birth); P1 (first-
4 ftp://sidads.colorado.edu/pub/DATASETS/
5 www.bom.gov.au/climate/current/soihtm1.shtml
GLENN E. STAUFFER ET AL.2528 Ecology, Vol. 95, No. 9
time observable prebreeder); U1 (unobservable pre-
breeder, previously observed once); P2 (observable
prebreeder, previously observed 1 time); U2 (unob-
servable prebreeder, previously observed 2 times); and
F (observable first-time mother). TE was defined as a
transition into any of the unobservable states (e.g.,
wP0U0 denoted transition probability from pup to
unobservable prebreeder) and recruitment as a transi-
tion into state F (e.g., wP1F denotes recruitment
probability from state P1). Our goal was to investigate
possible sources of variation in transition probabilities
of prebreeders, so we right-censored information for
seals after they recruited. Censoring simplified the
analysis by limiting the number of possible state
transitions and was justified by the near-perfect detec-
tion of mothers within Erebus Bay (Hadley et al. 2006,
Stauffer et al. 2013a).
Based on previous analyses (Hadley et al. 2006,
Stauffer et al. 2013a), we estimated TE for 11 nominal
age classes (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11 yr old)
and recruitment for eight nominal age classes (4, 5, 6, 7,
8, 9, 10, and 11 yr old), where age denotes a female’s
age class at the end of an interval. For example, wrzt;age5
denotes the probability of transition from state r at age 4
yr in year t 1 to state z at age 5 yr in year t. We fixed to
zero all transitions that were impossible based on our
state structure and knowledge of the earliest possible age
of recruitment.
To evaluate which covariates best explained TE
probabilities, we constructed 17 models that estimated
age-specific TE as logit-linear functions of covariates.
Our model set included all possible combinations of
covariates with the constraint that we did not allow SIE
and SOI or SIC and DIST to appear in the same model
because of pairwise correlations .0.4 for these pairs and
the a priori hypothesis of similar effects of these
covariates. In each of these models, we allowed state
dependency in TE and recruitment probabilities based
on whether or not females attended a reproductive
colony in year t  1. For recruitment probabilities, we
also included an additive year effect (year effects shared
across all ages and states). We compared estimated TE
probabilities from our best covariate model to estimates
from two baseline models that either included an
additive year effect unstructured by covariates (year
model) or specified equal TE rates for all years (null
model).
To evaluate hypotheses about motivations for atten-
dance by prebreeders and consequences for age-specific
recruitment, we constructed a suite of four models with
different state-dependent structures for recruitment
probabilities. The first structure modeled recruitment
as equal for all states and represented the null hypothesis
that attendance patterns have little consequence for
subsequent recruitment. The second structure modeled
recruitment probabilities in year t based on whether a
seal attended a reproductive colony in year t  1 but
ignored attendance information for years before t  1;
this structure represented the estrus hypothesis that
motivation to attend colonies is related to facilitation of
optimal timing of estrus in prebreeders. Embryonic
diapause synchronizes reproduction in pinnipeds (Boyd
1991), with consequent synchrony in ovulation and
presumably mating (Smith 1966, Hill 1987). Mating
synchrony might be advantageous if it allows most
females to mate with dominant males (Wartzok 1991).
Such advantages would be greatest for highly polygy-
nous species, but might also be important for moder-
ately polygynous Weddell seals where males defend
underwater territories (Siniff et al. 1977, Hill 1987).
Also, spermatogenesis in Weddell seals declines shortly
after the breeding season (Smith 1966), so females that
ovulate late might fail to mate successfully. Given that
prebreeders do not have a previous parturition to trigger
estrus, we suggest that close association with parous
females could provide chemical or social cues to trigger
timely ovulation. Therefore, we predicted greater
recruitment probabilities for observable than unobserv-
able prebreeders. To represent the experience hypothesis
that attendance was motivated by opportunities to
acquire information or skills, the third structure
modeled recruitment probabilities based on whether
seals had attended a reproductive colony in 2 yr, but
ignored information on when attendance occurred.
Given that the mean age of physiological sexual
maturity is considerably younger than the mean age at
recruitment (Smith 1966, Hadley et al. 2006), it seems
clear that there are other factors, perhaps behavioral,
that influence recruitment probability. The experience
hypothesis posits that repeated attendance at reproduc-
tive colonies facilitates observation and learning of
appropriate behavior that increases the probability of
successful subsequent mating. Thus recruitment proba-
bility should be greatest for individuals that attended at
least twice (e.g., wP2F ’ wU2F . wP1F ’ wU1F). Our
fourth model structure estimated different recruitment
rates for each state and represented the hypothesis that
synchronization of estrus and acquisition of experience
both might motivate attendance. In each model, annual
variation for transition probabilities was unstructured
(not a function of environmental covariates) with
patterns shared across age classes. Model structures
for survival probabilities and parameters representing
within-year dynamics (see Appendix A for details) were
based on previous analyses (Hadley et al. 2006, Rotella
et al. 2009, Stauffer et al. 2013b).
We performed analyses in program MARK (White
and Burnham 1999) through the RMark package
(Laake 2013) in program R (R Development Core
Team 2012). Currently there is no general goodness-of-
fit test for the type of ORDMS models we used in our
analysis, and the median cˆ procedure implemented into
program MARK (cˆ is an estimated overdispersion
parameter) is not available for robust design data. Thus,
we used Akaike information criterion corrected for
small sample sizes (AICc) rather than quasi-likelihood
September 2014 2529EMIGRATION AND RECRUITMENT BY SEALS
AICc (QAICc) as our model selection criterion. Howev-
er, we did assess the influence of possible overdispersion
on model-selection results by examining how model
rankings based on QAICc changed as we increased cˆ
(Appendix B: Table B1).
RESULTS
Of the 5550 female seals in our data set, 1333 were
observed as prebreeders in 1 yr after tagging, and 942
recruited into the study area’s breeding population
(mean age at first reproduction was 7.7, range of 4–16
yr) by 2010. Most females were present by the first
survey period each year, and the probability of being
detected at least once during each season was 0.91 for
prebreeders and .0.99 for pups and mothers (details in
Appendix A). Between the year when they were born
and the year when they gave birth to their first pup, 246
females (26% of first-time breeders) were never observed
in the breeding colonies as prebreeders (mean age at first
reproduction was 7.0, range of 4–15 yr), 278 (30%) were
observed as prebreeders in 1 yr (mean age at first
reproduction was 7.5, range of 5–13 yr), and 418 (44%)
were observed as prebreeders in .1 yr (mean age at first
reproduction was 8.3, range of 5–16 yr). For seals that
were observed as prebreeders, mean age at first
observation was 4.6 yr (range of 1–12 yr) and mean
observed number of years of attendance was 2.1 (range
of 1–5 yr).
As expected, TE probabilities were strongly age- and
state-dependent. Consistent with the hypothesis that
highly variable environmental conditions should influ-
ence motivation or ability of prebreeders to attend
colonies, year-to-year variability in TE probabilities was
substantial, particularly after 2000 (Fig. 1a, c). Our best-
supported covariate model (Table 1, model 9) supported
our predictions that TE probabilities would be greatest
when the current-year fast-ice edge was distant (bˆDIST¼
0.76, cSE¼ 0.04; estimated effects [b] and standard errors
of distance to the sea-ice edge in the current year on TE
probability in the current year) and when the previous
year’s population of adult seals was large (bˆBPOP¼ 0.66,
cSE ¼ 0.17; estimated effects and standard errors of
population size in the previous year). Annual variation
in DIST and BPOP was considerable and was strongly
related to predicted TE probabilities (Fig. 1b, d). For
example, in a year when DIST was minimal and the
previous year’s BPOP was small, predicted TE proba-
bility was 0.40 (cSE¼0.04) for a 5-yr-old seal that did not
attend a reproductive colony in the previous year. In
contrast, predicted TE probability was 0.91 (cSE ¼ 0.01)
in a year when DIST was maximal and BPOP was large.
These covariates appear to explain much of the annual
variation in TE probabilities: estimated TE probabilities
from our best-supported covariate model were similar in
most years to those from a more general model with
unstructured annual variation in TE probabilities (Fig.
1c). Our second and third best-supported models
included BPOP and DIST, with estimated coefficients
and annual TE estimates similar to those from the top
model, but also included either SIE or SOI as predictors.
However, estimated effect sizes were essentially zero
(bˆSIE ¼ 0.04, cSE ¼ 0.03; bˆSOI ¼ 0.005, cSE ¼ 0.04) and
model deviance improved little (Table 1). Thus, SIE and
SOI apparently explained little additional variation in
TE probabilities beyond that explained by BPOP and
DIST.
Estimates from the year model (Table 1, model 21)
indicated that TE probabilities were high for 1-yr-olds
(wˆ
P0U0
t;age1 ¼ 0.98, cSE ¼ 0.01, SD [wˆP0U0t;age1] ¼ 0.019),
subsequently decreased until age eight (e.g., wˆ
P1U1
t;age8 ¼
0.15, cSE ¼ 0.02, SD [wˆP1U1t;age8] ¼ 0.16), then gradually
increased for the older age classes (e.g., wˆ
P1U1
t;age11þ ¼ 0.29,
cSE ¼ 0.05, SD [wˆP1U1t;age11þ]¼ 0.22; Fig. 1a). Symbols with a
circumflex denote estimated parameters, and overbars
denote parameter estimates averaged over all years. As
predicted, seals that did not attend a reproductive
colony in year t  1 had higher TE probabilities than
seals that did attend (bˆattend¼1.35, cSE¼ 0.10, 95% CI
¼ 1.54 to 1.15; e.g., wˆU2U2t;age8  wˆ
P2U2
t;age8 ¼ 0.21, cSE ¼
0.024).
Our best-supported model in the recruitment suite
(Table 1) represented the hypothesis that both synchro-
ny of estrus and acquisition of information might
motivate colony attendance. This model allowed differ-
ent age-specific recruitment probabilities for each
prebreeder state and indicated strong age and time
dependence (Fig. 2) but provided mixed support for
both hypotheses. As predicted, observable prebreeders
were more likely to recruit than were unobservable
prebreeders, given that both had attended colonies in
two or more previous years (e.g., wˆP2Ft;age8  wˆ
U2F
t;age8 ¼ 0.19,cSE ¼ 0.05). However, this difference was not evident for
prebreeders that had only attended colonies in one
previous year (e.g., wˆP1Ft;age8  wˆ
U1F
t;age8 , 0.01, cSE ¼ 0.038).
Among observable prebreeders, those with more years
of colony attendance were most likely to recruit (e.g.,
wˆP2Ft;age8  wˆP1Ft;age8 ¼ 0.11, cSE ¼ 0.033). Previous colony
experience seemed less important for seals that were
unobservable the year prior to recruitment. Unobserv-
able prebreeders that attended in one previous year were
slightly more likely to recruit than seals that had no
previous colony experience (e.g., wˆU1Ft;age8  wˆ
U0F
t;age8 ¼ 0.08,
cSE ¼0.03), but additional years of colony experience did
not increase age-specific recruitment probabilities fur-
ther (e.g., wˆU2Ft;age8  wˆ
U1F
t;age8 ¼0.08, cSE ¼ 0.05). Average
recruitment probability was ,0.01 for seals at age four,
remained low (,0.1) for seals at age five, gradually
increased to a maximum at age 10 (e.g., wˆ
P2F
t;age10 ¼ 0.56,cSE ¼0.06, SD [wˆP2Ft;age10]¼0.17), and then declined slightly
in the oldest age class (e.g., wˆ
P2F
t;age11þ ¼ 0.41, cSE ¼ 0.06,
SD [wˆP2Ft;age11þ]¼ 0.16; Fig. 2a). Recruitment probabilities
GLENN E. STAUFFER ET AL.2530 Ecology, Vol. 95, No. 9
FIG. 1. Annual covariate values and temporary emigration (TE) rates of female prebreeder Weddell seals in Erebus Bay,
Antarctica. (a) Summary of annual TE rates estimated from year model (unstructured annual variation; Table 1, model 21) with
medians (dark bars), interquartile range (boxes), and whiskers bounding the 2.5th and 97.5th percentiles of annual point estimates;
wˆU0U0 denotes TE by seals never observed as prebreeders, wˆP1U1 denotes first-time observable prebreeder (P1) or unobservable
prebreeder, previously observed once (U1), where wˆ denotes estimated transition probability. (b) Predicted TE rates (point
estimates and 95% confidence intervals [CIs] for wˆU1U1t;age5 , where wˆ
U1U1
t;age5 indicates TE by seals previously observed once as prebreeders,
but not in the previous year, where t is year and age5 is age of seal) when distance to the fast-ice edge is close (solid line),
intermediate (dashed line), or far (dotted line), over a range of population sizes. Estimates were generated from the best-supported
covariate model (Table 1, model 9). (c) Comparison of annual TE rates (point estimates and 95% CIs for wˆP1U1t;age5, where wˆ
P1U1
t;age5
indicates first-time observable prebreeders) estimated from null model, year model, and best covariate model. (d) Annual breeding
population size in Erebus Bay and distance to fast ice. Distance values in some years represent multiple imputations of missing
values for those years; the line passes through the mean of imputed values.
TABLE 1. Models representing hypotheses about variation in transition probabilities for female Weddell seals in Erebus Bay,
Antarctica.
Model K AICc DAICc wi Deviance
TE model suite
(Model 9) wrFðaÞ, w
rU
ðBPOP þ DISTÞ 87 74 019.53 0.000 0.472 73 845.15
(Model 17) wrFðaÞ, w
rU
ðBPOP þ DIST þ SIEÞ 88 74 020.11 0.587 0.352 73 843.72
(Model 15) wrFðaÞ, w
rU
ðBPOP þ DIST þ SOIÞ 88 74 021.51 1.985 0.175 73 845.12
Recruitment model suite
(Model 22) wrFðaeÞ, w
rU
ðyearÞ 113 73 685.08 0.000 0.976 73 458.43
(Model 20) wrFðeÞ, w
rU
ðyearÞ 110 73 692.66 7.580 0.022 73 472.05
(Model 21) wrFðaÞ, w
rU
ðyearÞ 110 73 699.01 13.935 0.001 73 478.40
(Model 19) wrFðnonMarkovianÞ, w
rU
ðyearÞ 109 73 699.57 14.492 0.001 73 480.97
Notes: In the temporary emigration (TE) model suite, rankings were robust for cˆ  3.5, and in the recruitment model suite, the
top model retained its rank for cˆ, 2.3, where cˆ is an estimated overdispersion parameter. See Appendix B: Table B1 for a table of
complete model selection results for TE. AICc is Akaike’s information criterion. Models included 11 nominal age classes (1, 2, 3,
4, 5, 6, 7, 8, 9, 10, and 11þ) for TE (wrU) and annual variation and nine nominal age classes (1-3, 4, 5, 6, 7, 8, 9, 10, and 11þ) for
recruitment (wrF), where wrU is temporary emigration and wrF is recruitment. The Markovian structure for TE specified different
rates depending on colony attendance the preceding year, whereas the Markovian structure for recruitment in the recruitment
model suite specified dependence on either colony attendance (a), previous attendance experience at time of recruitment (e), both
(ae), or neither (non-Markovian). Covariate values for TE were distance from Erebus Bay to the edge of the fast ice (DIST),
population size of breeding female seals in the previous year (BPOP), winter sea ice extent (SIE), and Southern Oscillation Index
(SOI).
September 2014 2531EMIGRATION AND RECRUITMENT BY SEALS
were variable among years and tended to be low when
TE probabilities were high (Fig. 2b).
DISCUSSION
We fit multistate capture–mark–recapture models to
28 yr of encounter data for Weddell seals to evaluate
possible correlates of annual variation in age-specific TE
probabilities and implications of colony attendance and
nonattendance for age-specific recruitment probabilities.
Temporary emigration probabilities varied considerably
from year to year, and our results are consistent with
previous suggestions that such variation could be related
to colony access, population size, or body condition
(Rotella et al. 2009, Stauffer et al. 2013a). Our results
also provide mixed support for the hypotheses that
attending colonies for 1 yr allows prebreeders to
optimally time estrus or to acquire information or skills
that increase subsequent age-specific recruitment prob-
abilities.
The distance from Erebus Bay to open water varied
annually from ,10 km to .80 km (Fig. 1d), and, as
expected, distance to the edge of the fast ice was strongly
and positively related to TE probabilities. In years when
fast ice is extensive, it likely is difficult for younger or
smaller seals to gain access to Erebus Bay. Older and
heavier Weddell seals have better diving abilities than
younger and lighter seals, and diving behavior of young
seals is subject to physiological or morphological
constraints (Kooyman 1966, Burns 1999). Thus, a long
swim under extensive fast ice might restrict access to
Erebus Bay for prebreeders in poor physiological
condition. If prebreeders need to learn to navigate to
their natal sites, as has been anecdotally suggested for
the colonial-breeding seabird Common Tern (Sterna
hirundo; Dittmann and Becker 2003), then extensive fast
ice might make navigation difficult for young seals.
There is a possible alternative explanation for the
relationship between TE and DIST. Weddell seals are
vulnerable to predation from orcas (Orcinus orca;
Pitman and Durban 2012), but are safe from predation
when hauled out on fast ice. Thus, in years when fast ice
is minimal (DIST is small), more seals might be
motivated to attend areas with reliable fast ice around
Erebus Bay colonies, whereas when DIST is maximal
there are large areas of predator-safe fast ice outside
Erebus Bay. Until large numbers of seals are tracked
outside Erebus Bay, it will be difficult to separate
various motivations for prebreeders to attend reproduc-
tive colonies in Erebus Bay. It is likely that predator
avoidance is one motivation, but there are reasons we
believe that acquisition of information or social skills
might provide additional motivation.
We do not know the underlying reasons for observed
density dependence in TE probabilities, but suggest
three nonexclusive hypotheses: (1) emigration is moti-
vated by resource depletion related to large populations
(see review in Matthysen 2005); (2) large local popula-
tions of adults in breeding colonies increase the level of
social conflict between adults and prebreeders (Siniff et
al. 1977), thus discouraging attendance (increasing
motivation for TE) by prebreeders the following year;
and (3) observed density dependence in TE probabilities
is spurious (i,e., an artifact based on unmodeled
correlations). Mature females generally do not forage
extensively during the breeding season, but some
resource depletion apparently does occur in Erebus
Bay during the pupping season (Testa et al. 1985). When
the number of seals in Erebus Bay colonies is large, the
broader regional population might also be large if
broadscale environmental conditions drive dynamics.
However, that relationship is unknown to us. If true,
consequent broadscale competition for resources in
FIG. 2. Recruitment probabilities of female Weddell seals in Erebus Bay, Antarctica, estimated from the best-supported
recruitment model (Table 1, model 22). Presented are (a) a summary of probabilities with medians (dark bars), the interquartile
range (boxes), and whiskers bounding the 2.5th and 97.5th percentiles of annual point estimates; and (b) annual rates and 95%
confidence intervals for age 7 showing the temporal patterns shared by all ages and states in the model. In panel b, the dashed line
indicating temporary emigration (TE) by first-time prebreeders (wˆP1U1) is included to illustrate the strong inverse correlation
between recruitment and TE. The symbols wˆU0F, wˆP1F, wˆU1F, wˆP2F, and wˆU2F, respectively, denote recruitment by seals never
observed as prebreeders, observed for the first time in the previous year, observed once but not in the previous year, observed for at
least the second time in the previous year, and observed at least twice, but not in the immediate previous year (where wˆ denotes
estimated transition probability).
GLENN E. STAUFFER ET AL.2532 Ecology, Vol. 95, No. 9
major foraging areas could be intense and negatively
influence a seal’s condition and motivation to attend
breeding colonies. Given the potential of high-resolution
satellite imagery for assessing regional population size
(LaRue et al. 2011), it might soon be possible to address
this hypothesis. Crowding of seals at the local scale
appears to increase levels of conflict between individuals,
and prebreeders are found mostly in the periphery of the
reproductive colonies early in the breeding season
(Stirling 1969, Siniff et al. 1977). Intrasexual aggression
has also been documented in several other pinniped
species (e.g., McCann 1982, Ferna´ndez-Juricic and
Cassini 2007) and in some ungulates appears to motivate
natal dispersal (e.g., Long et al. 2008). Proposing that
intrasexual conflict in one year motivates TE the next
year does assume either physiological costs or psycho-
logical influences that decrease ability or motivation to
attend Erebus Bay the next year. The possibility remains
that the density dependence we estimated in TE
probabilities might be spurious. For example, popula-
tion size may be correlated with environmental condi-
tions. If favorable environmental conditions tend to be
followed by less favorable conditions, then a large
population size in one year would tend to be followed by
a smaller population size and increased TE in the
following year. Consequently, the apparent relationship
between BPOP and TE might derive from correlated TE
probabilities of prebreeders and adult seals (those
already recruited into the breeding population), both
driven by the same environmental factors. Analyses
currently in progress to estimate TE probabilities of
adult seals should provide insights about possible
correlations between TE probabilities of adults and
prebreeders.
Our results provided mixed support for the estrus and
experience hypotheses, but the overall picture is one
where age-specific recruitment probabilities were great-
est for seals that consistently attended colonies, which is
consistent with both hypotheses. On the one hand, seals
that attended in the preceding year and 1 additional
previous year were more likely to recruit than seals that
attended in only the preceding year (e.g.,wˆP2Ft;age . wˆ
P1F
t;age)
or seals that attended in a previous year, but then
subsequently skipped attendance (e.g.,wˆP2Ft;age . wˆ
U2F
t;age).
Weddell seals have a breeding system similar to lekking,
where males defend underwater territories focused on
cracks in the ice (Siniff et al. 1977). In such a system,
prebreeders that attend Erebus Bay might benefit from
observing mating behavior of mature seals, either
gaining experience in evaluating potential mates or
simply copying the behavior of other females (e.g.,
Dugatkin and Godin 1993, Valone 2007). The experi-
ence hypothesis is further supported by the observed
difference between age at physiological maturity in
female Weddell seals (2–4 yr; Smith 1966) and the
average age at which females actually produce their first
pup (7.6 yr; Hadley et al. 2006). There is also evidence
for selective pressure for synchronized birth dates in
PLATE 1. A Weddell seal mother and pup rest on the ice surface in Erebus Bay, Antarctica. Photo credit: Shawn C. Farry.
September 2014 2533EMIGRATION AND RECRUITMENT BY SEALS
Weddell seals and other pinnipeds (Boness et al. 1995,
Proffitt et al. 2010), which might be facilitated by
synchronized estrus. On the other hand, seals that
attended reproductive colonies once before producing a
pup, regardless of when that attendance took place, had
nearly equal recruitment probability (e.g., wˆU1Ft;age ’ wˆ
P1F
t;age),
and for seals that did not attend in the year preceding
recruitment, recruitment probabilities of experienced
prebreeders were not greater than those for inexperi-
enced prebreeders (i.e., wˆU0Ft;age ’ wˆ
U2F
t;age , wˆ
U1F
t;age). Possible
explanations for these seemingly conflicting results are
that estrus synchrony might not be very important given
that embryonic diapause can synchronize births (Boyd
1991) or a single year of attendance might adequately
and permanently synchronize estrus. Also, a segment of
the population that doesn’t consistently attend colonies
as prebreeders might be unlikely to ever recruit (Hadley
et al. 2006). This latter explanation suggests fitness links
to repeated TE and is consistent with the estimated
state-dependent differences in recruitment probabilities,
the slight increases in TE probabilities and decreases in
recruitment probabilities for the oldest age class, and the
fact that the mean number of years seals attended prior
to recruitment was only 2.1 yr. A comparison of
recruitment probabilities for state U2 and an additional
state U3 for individuals seen three times or more could
help evaluate a fitness hypothesis. However, we did not
include state U3 because our data set included relatively
few individuals that were ever in state U3, which made it
unlikely that we could have gained important insights
from such a complex model.
In our analysis, we assumed that females born in
Erebus Bay typically return to Erebus Bay to mate and
subsequently produce their first pup in Erebus Bay.
However, some seals might mate in other areas or even
at sea (de Bruyn et al. 2011) and/or subsequently recruit
outside Erebus Bay. If that situation is common, then
recruitment probabilities are underestimated, probably
for all states, but perhaps especially so for unobservable
states because they are most likely to have mated outside
Erebus Bay. Recruitment outside Erebus Bay thus might
be an alternative explanation for reduced recruitment
probability at older ages and for state U2. However,
several lines of evidence suggest that our assumptions
are likely reasonable. First, as mentioned in Hadley et al.
(2007), extensive searches during the pupping seasons of
1997–2000 throughout McMurdo Sound and beyond
(although certainly not as intensive as surveys conducted
in Erebus Bay) failed to find very many Erebus Bay-
born females pupping outside Erebus Bay, even though
substantial numbers of females were located without
pups. Second, annual visits, albeit limited in duration
and extent, during 2004–2013 to several large aggrega-
tions of females with young pups outside Erebus Bay
that we thought might be used by animals born in
Erebus Bay as alternate pup-rearing sites have repeat-
edly found minimal evidence of pupping outside of
Erebus Bay by females born in Erebus Bay. Third,
evidence of geographic variations in genetics and
underwater vocalizations of Weddell seals suggests
limited long-distance dispersal (Thomas and Stirling
1983, Davis et al. 2008). Fourth, although we did not
include an analysis of reproductively active seals in this
paper, there is strong evidence that most seals are
faithful to colonies during years when they skip
reproduction. Although prebreeders might not be as
site faithful as breeders, site fidelity by breeders and
skip-breeders at least suggests that fine-scale site fidelity
is an important component of the life history of this
population of Weddell seals.
Although uncertainty remains about events outside
Erebus Bay, our findings on several sources of variabil-
ity in TE probabilities and the consequences of colony
attendance and nonattendance provide some support for
previous suggestions that prebreeders trade off costs and
benefits of attending colonies (Stauffer et al. 2013a). The
balance of such trade-offs likely depends on annual
environmental conditions, characteristics of individuals,
and maybe population size, and may have strong
implications for the age at first reproduction. Our
understanding of mechanisms underlying hypothesized
tradeoffs should improve as it becomes technologically
feasible to monitor the movements of many individuals
over relevant spatial and temporal scales.
Trade-offs related to TE and implications for future
reproductive success likely are common among long-
lived colonial-breeding species. Evidence from Kitti-
wakes (Rissa tridactyla) suggests that prebreeding
behavior and experience can influence short-term and
long-term reproductive success and that recruitment at
intermediate ages is optimal (Cam et al. 2002). An
intermediate recruitment age also is optimal for future
reproductive success of northern elephant seals (Mi-
rounga angustirostrus; Reiter and Boeuf 1991). We
focused our investigation on implications of annual
attendance and nonattendence for age-specific recruit-
ment probabilities. An interesting question we did not
address was whether experienced prebreeders arrived
earlier than inexperienced prebreeders (e.g., Pickering
1989, Dittmann and Becker 2003), and, given that
prebreeders apparently trickled into Erebus Bay
throughout the breeding season (Appendix A: Fig.
A3), whether arrival date was related to future
recruitment probability. Also, there appears to be
considerable heterogeneity in long-term reproductive
fitness among individual female Weddell seals following
recruitment (Chambert et al. 2013), and it has been
suggested that the highest quality individuals recruit at
the youngest ages (Hadley et al. 2007). In the future, it
would be useful to evaluate whether attendance patterns
prior to recruitment might help explain heterogeneity in
lifetime reproductive output.
ACKNOWLEDGMENTS
The National Science Foundation, Office of Polar Programs
(grant no. ANT-0635739 to R. A. Garrott, J. J. Rotella, and
D. B. Siniff, and previous grants to D. B. Siniff and J. W. Testa)
GLENN E. STAUFFER ET AL.2534 Ecology, Vol. 95, No. 9
funded our work. Raytheon Polar Services Corporation,
Petroleum Helicopters International, and the New York
National Guard provided logistical support, and many field
assistants assisted with data collection. M. LaRue and C. Porter
extracted data for sea-ice covariates. T. E. McMahon, D. B.
Siniff, J. A. Schmutz, and two anonymous reviewers provided
useful comments on previous manuscript drafts. Animal
handling protocol was approved by Montana State University’s
Institutional Animal Care and Use Committee (Protocol #41-
05). Any use of trade, firm, or product names is for descriptive
purposes only and does not imply endorsement by the U.S.
Government.
LITERATURE CITED
Balmford, A. 1991. Mate choice on leks. Trends in Ecology and
Evolution 6:87–92.
Beauplet, G., C. Barbraud, M. Chambellant, and C. Guinet.
2005. Interannual variation in the post-weaning and juvenile
survival of subantarctic fur seals: influence of pup sex,
growth rate and oceanographic conditions. Journal of
Animal Ecology 74:1160–1172.
Becker, P. H., and J. S. Bradley. 2007. The role of intrinsic
factors for the recruitment process in long-lived birds.
Journal of Ornithology 148:377–384.
Boness, D. J., W. D. Bowen, and S. J. Iverson. 1995. Does male
harassment of females contribute to reproductive synchrony
in the grey seal by affecting maternal performance?
Behavioral Ecology and Sociobiology 36:1–10.
Boyd, I. L. 1991. Environmental and physiological factors
controlling the reproductive cycles of pinnipeds. Canadian
Journal of Zoology 69:1135–1148.
Burns, J. M. 1999. The development of diving behavior in
juvenile Weddell seals: pushing physiological limits in order
to survive. Canadian Journal of Zoology 77:737–747.
Burns, J. M., M. A. Castellini, and J. W. Testa. 1999.
Movements and diving behavior of weaned Weddell seal
(Leptonychotes weddellii ) pups. Polar Biology 21:23–36.
Burns, J. M., S. J. Trumble, M. A. Castellini, and J. W. Testa.
1998. The diet of Weddell seals in McMurdo Sound,
Antarctica as determined from scat collections and stable
isotope analysis. Polar Biology 19:272–282.
Cadiou, B., J. Y. Monnat, and E. Danchin. 1994. Prospecting in
the kittiwake, Rissa tridactyla: different behavioural patterns
and the role of squatting in recruitment. Animal Behaviour
47:847–856.
Cam, E., B. Cadiou, J. E. Hines, and J. Y. Monnat. 2002.
Influence of behavioural tactics on recruitment and repro-
ductive trajectory in the kittiwake. Journal of Applied
Statistics 29:163.
Cameron, M. F., D. B. Siniff, K. M. Proffitt, and R. A.
Garrott. 2007. Site fidelity of Weddell seals: the effects of sex
and age. Antarctic Science 19:149–155.
Castellini, M. A., G. L. Kooyman, and P. J. Ponganis. 1992.
Metabolic rates of freely diving Weddell seals: correlations
with oxygen stores, swim velocity and diving duration.
Journal of Experimental Biology 165:181–194.
Chambert, T., J. J. Rotella, M. D. Higgs, and R. A. Garrott.
2013. Individual heterogeneity in reproductive rates and cost
of reproduction in a long-lived vertebrate. Ecology and
Evolution 3:2047–2060.
Davis, C. S., I. Stirling, C. Strobeck, and D. W. Coltman. 2008.
Population structure of ice-breeding seals. Molecular Ecol-
ogy 17:3078–3094.
de Bruyn, P. J. N., C. A. Tosh, M. N. Bester, E. Z. Cameron, T.
McIntyre, and I. S. Wilkinson. 2011. Sex at sea: alternative
mating system in an extremely polygynous mammal. Animal
Behaviour 82:445–451.
Dittmann, T., and P. H. Becker. 2003. Sex, age, experience and
condition as factors affecting arrival date in prospecting
common terns, Sterna hirundo. Animal Behaviour 65:981–
986.
Dittmann, T., T. H. G. Ezard, and P. H. Becker. 2007.
Prospectors’ colony attendance is sex-specific and increases
future recruitment chances in a seabird. Behavioural Pro-
cesses 76:198–205.
Dugatkin, L. A., and J.-G. J. Godin. 1993. Female mate
copying in the guppy (Poecilia reticulata): age-dependent
effects. Behavioral Ecology 4:289–292.
Ferna´ndez-Juricic, E., and M. H. Cassini. 2007. Intra-sexual
female agonistic behaviour of the South American sea lion
(Otaria flavescens) in two colonies with different breeding
substrates. Acta Ethologica 10:23–28.
Frederiksen, M., and T. Bregnballe. 2000. Diagnosing a decline
in return rate of 1-year-old cormorants: mortality, emigration
or delayed return? Journal of Animal Ecology 69:753–761.
Furness, R. W., and T. R. Birkhead. 1984. Seabird colony
distributions suggest competition for food supplies during the
breeding season. Nature 311:655–656.
Galef, B. G., Jr., and K. N. Laland. 2005. Social learning in
animals: empirical studies and theoretical models. Bioscience
55:489–499.
Hadley, G. L., J. J. Rotella, and R. A. Garrott. 2007. Influence
of maternal characteristics and oceanographic conditions on
survival and recruitment probabilities of Weddell seals. Oikos
116:601–613.
Hadley, G. L., J. J. Rotella, R. A. Garrott, and J. D. Nichols.
2006. Variation in probability of first reproduction of
Weddell seals. Journal of Animal Ecology 75:1058–1070.
Halley, D. J., M. P. Harris, and S. Wanless. 1995. Colony
attendance patterns and recruitment in immature common
murres (Uria aalge). Auk 112:947–957.
Handelmann, G., R. Ravizza, and W. J. Ray. 1980. Social
dominance determines estrous entrainment among female
hamsters. Hormones and Behavior 14:107–115.
Hill, S. E. B. 1987. Reproductive ecology of Weddell seals
(Leptonychotes weddelli ) in McMurdo Sound, Antartica.
Dissertation. University of Minnesota, St. Paul, Minnesota,
USA.
Hoffman, J. I., J. Forcada, P. N. Trathan, and W. Amos. 2007.
Female fur seals show active choice for males that are
heterozygous and unrelated. Nature 445:912–914.
Jenouvrier, S., G. Tavecchia, J.-C. Thibault, R. Choquet, and
V. Bretagnolle. 2008. Recruitment processes in long-lived
species with delayed maturity: estimating key demographic
parameters. Oikos 117:620–628.
Kendall, W. L. 2004. Coping with unobservable and misclas-
sified states in capture–recapture studies. Animal Biodiversity
and Conservation 27:97–107.
Kendall, W. L., and R. Bjorkland. 2001. Using open robust
design models to estimate temporary emigration from
capture–recapture data. Biometrics 57:1113–1122.
Kooyman, G. L. 1966. Maximum diving capacities of the
Weddell seal, Leptonychotes weddeili. Science 151:1553–1554.
Kwok, R. and J. C. Comiso. 2002. Southern Ocean climate and
sea ice anomalies associated with the Southern Oscillation.
Journal of Climate 15:487–501.
Laake, J. L. 2013. RMark: an R interface for analysis of
capture-recapture data with MARK. AFSC processed report
2013-01. Alaska Fisheries Science Center, NOAA, National
Marine Fisheries Service, Seattle, Washington, USA.
LaRue, M. A., J. J. Rotella, R. A. Garrott, D. B. Siniff, D. G.
Ainley, G. E. Stauffer, C. C. Porter, and P. J. Morin. 2011.
Satellite imagery can be used to detect variation in abundance
of Weddell seals (Leptonychotes weddellii ) in Erebus Bay,
Antarctica. Polar Biology 34:1727–1737.
Loeb, V., V. Siegel, O. Holm-Hansen, R. Hewitt, W. Fraser, W.
Trivelpiece, and S. Trivelpiece. 1997. Effects of sea-ice extent
and krill or salp dominance on the Antarctic food web.
Nature 387:897–900.
Long, E. S., D. R. Diefenbach, C. S. Rosenberry, and B. D.
Wallingford. 2008. Multiple proximate and ultimate causes
September 2014 2535EMIGRATION AND RECRUITMENT BY SEALS
of natal dispersal in white-tailed deer. Behavioral Ecology 19:
1235–1242.
Madsen, T., and R. Shine. 2000. Silver spoons and snake body
sizes: prey availability early in life influences long-term
growth rates of free-ranging pythons. Journal of Animal
Ecology 69:952–958.
Matthysen, E. 2005. Density-dependent dispersal in birds and
mammals. Ecography 28:403–416.
McCann, T. S. 1982. Aggressive and maternal activities of
female southern elephant seals (Mirounga leonina). Animal
Behaviour 30:268–276.
McClintock, M. K. 1978. Estrous synchrony and its mediation
by airborne chemical communication (Rattus norvegicus).
Hormones and Behavior 10:264–276.
Pickering, S. P. C. 1989. Attendance patterns and behaviour in
relation to experience and pair-bond formation in the
wandering albatross Diomedea exulans at South Georgia.
Ibis 131:183–195.
Pitman, R. L., and J. W. Durban. 2012. Cooperative hunting
behavior, prey selectivity and prey handling by pack ice killer
whales (Orcinus orca), type B, in Antarctic Peninsula waters.
Marine Mammal Science 28:16–36.
Proffitt, K. M., R. A. Garrott, J. J. Rotella, D. B. Siniff, and
J. W. Testa. 2007. Exploring linkages between abiotic
oceanographic processes and a top-trophic predator in an
Antarctic ecosystem. Ecosystems 10:120–127.
Proffitt, K. M., J. J. Rotella, and R. A. Garrott. 2010. Effects of
pup age, maternal age, and birth date on pre-weaning
survival rates of Weddell seals in Erebus Bay, Antarctica.
Oikos 119:1255–1264.
R Development Core Team. 2012. R: a language and
environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. http://www.
r-project.org
Reed, J. M., T. Boulinier, E. Danchin, and L. W. Oring. 1999.
Informed dispersal: prospecting by birds for breeding sites.
Current Ornithology 15:189–259.
Reiter, J., and B. J. Le Boeuf. 1991. Life history consequences
of variation in age at primiparity in northern elephant seals.
Behavioral Ecology and Sociobiology 28:153–160.
Rotella, J. J., W. A. Link, J. D. Nichols, G. L. Hadley, R. A.
Garrott, and K. M. Proffitt. 2009. An evaluation of density-
dependent and density-independent influences on population
growth rates in Weddell seals. Ecology 90:975–984.
Schafer, J. L. 1999. Multiple imputation: a primer. Statistical
Methods in Medical Research 8:3–15.
Siniff, D. B., D. P. DeMaster, R. J. Hofman, and L. L.
Eberhardt. 1977. An analysis of the dynamics of a Weddell
seal population. Ecological Monographs 47:319–335.
Smith, M. S. R. 1965. Seasonal movements of the Weddell seal
in McMurdo Sound, Antarctica. Journal of Wildlife Man-
agement 29:464–470.
Smith, M. S. R. 1966. Studies of the Weddell seal (Leptony-
chotes weddelli Lesson) in McMurdo Sound, Antarctica.
Dissertation. University of Caterbury, Christchurch, New
Zealand.
Stauffer, G. E., J. J. Rotella, and R. A. Garrott. 2013a.
Variability in temporary emigration rates of individually
marked female Weddell seals prior to first reproduction.
Oecologia 172:129–140.
Stauffer, G. E., J. J. Rotella, and R. A. Garrott. 2013b. Birth-
year and current-year influences on survival and recruitment
rates of female Weddell seals. Population Ecology 55:405–
415.
Stern, K., and M. K. McClintock. 1998. Regulation of
ovulation by human pheromones. Nature 392:177–179.
Stewart, B., P. Yochem, T. Gelatt, and D. Siniff. 2000. First-
year movements of Weddell seal pups in the Western Ross
Sea, Antarctica. Pages 71–76 in W. Davidson, C. Howard-
Williams, and P. Broady, editors. Antarctic ecosystems:
models for wider ecological understanding. Proceedings of
the 7th SCAR Symposium on Antarctic Biology, Christ-
church, New Zealand.
Stirling, I. 1969. Ecology of the Weddell seal in McMurdo
Sound, Antarctica. Ecology 50:573–586.
Testa, J. W., G. Oehlert, D. G. Ainley, J. L. Bengtson, D. B.
Siniff, R. M. Laws, and D. Rounsevell. 1991. Temporal
variability in Antarctic marine ecosystems—periodic fluctu-
ations in the Phocid seals. Canadian Journal of Fisheries and
Aquatic Sciences 48:631–639.
Testa, J. W., D. B. Siniff, M. J. Ross, and J. D. Winter. 1985.
Weddell seal–Antarctic cod interactions in McMurdo Sound.
Pages 561–565 in W. R. Siegfried, P. R. Condy, and R. M.
Laws, editors. Antarctic nutrient cycles and food webs.
Proceedings of the 4th SCAR symposium on Antarctic
biology, Springer, New York, New York, USA.
Thomas, J. A., and I. Stirling. 1983. Geographic variation in the
underwater vocalizations of Weddell seals (Leptonychotes
weddelli) from Palmer Peninsula and McMurdo Sound,
Antarctica. Canadian Journal of Zoology 61:2203–2212.
Valone, T. 2007. From eavesdropping on performance to
copying the behavior of others: a review of public informa-
tion use. Behavioral Ecology and Sociobiology 62:1–14.
Wartzok, D. 1991. Physiology of behavior in pinnipeds. Pages
236–299 in D. Renouf, editor. Behavior of pinnipeds.
Chapman and Hall, New York, USA.
White, G. C., and K. P. Burnham. 1999. Program MARK:
survival estimation from populations of marked animals.
Bird Study 46(Supplement):120–138.
SUPPLEMENTAL MATERIAL
Appendix A
Details of open robust-design multistate (ORDMS) modeling, and estimates for survival and within-season parameters
(Ecological Archives E095-222-A1).
Appendix B
Complete model selection results (Ecological Archives E095-222-A2).
GLENN E. STAUFFER ET AL.2536 Ecology, Vol. 95, No. 9