Inbreeding in red-cockaded woodpeckers: Effects of

B I O L O G I C A L C O N S E RVAT I O N
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biocon
Inbreeding in red-cockaded woodpeckers: Effects of
natal dispersal distance and territory location
Karin Schiegga,b,*, Susan J. Danielsa,z, Jeffrey R. Waltersa, Jeffery A. Priddyc,
Gilberto Pasinellia,b
a
Department of Biology, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061-0406, USA
Zoologisches Institut, Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland
c
Duke University, Nicholas School of the Environment, Marine Laboratory, 135 Marine Laboratory Road, Beaufort, NC 28516-9712, USA
b
A R T I C L E I N F O
A B S T R A C T
Article history:
Inbreeding depression constitutes a significant threat to the viability of small populations.
Received 15 August 2005
In addition to small size and isolation of populations, short distance dispersal may elevate
Received in revised form
risk of inbreeding, but empirical evidence is scarce. Inbreeding depression has been demon-
3 January 2006
strated in the highly endangered red-cockaded woodpecker Picoides borealis. It has been sug-
Accepted 3 March 2006
gested that conservation efforts to support extant populations should aim at spatially
Available online 18 April 2006
aggregating territories to enhance dispersal success. This however may aggravate inbreeding risk because distance between territories and hence dispersal distances become short.
Keywords:
We analysed empirical data from a long-term study of the demography of the red-cockaded
Inbreeding depression
woodpecker and found that inbreeding risk varied inversely with natal dispersal distance of
Individual-based
the mother. Using an individual-based, spatially explicit population model that incorporates
Modelling
simulations of environmental and demographic stochasticity and an empirically derived,
Picoides borealis
species-specific estimate of inbreeding costs, we demonstrated that inbreeding depression
Spatially explicit
significantly elevated extinction risk in this species. On the other hand, even though dispersal distances in populations with spatially aggregated territories were shorter and the
proportion of inbred individuals was higher than in other populations of the same size, such
populations were still more persistent. Despite the overall adverse effect of inbreeding
depression on viability of red-cockaded woodpecker populations, lowering interterritorial
distances can be viewed as a valuable conservation tool. Given the small size and isolated
location of most extant red-cockaded woodpecker populations however, our findings suggest that inbreeding depression represents a significant threat to the survival of this species.
2006 Elsevier Ltd. All rights reserved.
1.
Introduction
Evidence is accumulating that inbreeding depression, i.e., reduced fecundity owing to matings among relatives (Gibbs,
2001) can severely impair reproductive output in wild popula-
tions of naturally outbreeding species of plants and animals
(see Hedrick and Kalinowski, 2000; Keller and Waller, 2002
for recent reviews). It has been argued that non-genetic stochastic effects drive small animal populations to extinction
before inbreeding depression becomes important (Lande,
* Corresponding author: Tel.: +41 44 635 49 82; fax: +41 1 635 68 21.
E-mail addresses: [email protected] (K. Schiegg), [email protected] (J.R. Walters), [email protected] (J.A. Priddy), gpasi@zool.
unizh.ch (G. Pasinelli).
z
Deceased 11/04.
0006-3207/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biocon.2006.03.001
B I O L O G I C A L C O N S E RVAT I O N
1988; Caughley, 1994). Yet, recent studies document inbreeding depression and/or elevated risk of extinction even in
populations interconnected by dispersal (Saccheri et al.,
1998; Höglund et al., 2002) and modelling studies have demonstrated detrimental effects of inbreeding depression on
population viability in the presence of other stochastic events
(Tanaka, 2000; Brook et al., 2002; also see Reed (2004)). On a
regional level, small population size and insufficient exchange of individuals among populations are the prime
causes of inbreeding (Frankham et al., 2002). On a local level,
short dispersal distances within populations may also raise
the risk of inbreeding (Walters et al., 2004), because related
individuals cluster around their natal site (Daniels and Walters, 2000b; Cutrera et al., 2005) and variation in natal dispersal
distances is low. However, quantitative relationships between
dispersal distances and inbreeding levels have not been
established so far.
The cooperatively breeding red-cockaded woodpecker Picoides borealis is an endangered species endemic to the southeastern United States. Family groups consist of a breeding
pair and 0–4 mostly male helpers (Walters, 1990). Many helpers attain breeding status by inheriting their natal territory,
and those helpers that do leave the natal territory disperse
only one or two territories away (Walters et al., 1988; Daniels,
1997). Individuals that do not become helpers disperse in
search of a breeding vacancy. These juveniles move variable
distances that average longer than helpers, but still the
majority settle within three territories of the natal site (Daniels, 1997; Daniels and Walters, 2000b).
The species is confined to mature pine savannahs and its
abundance has dramatically declined within the past 30 years
due to destruction of its natural habitat. Fifty-nine of the
existing 73 populations (81%) contain fewer than 75 groups
and most populations are demographically isolated. Conservation efforts are extensive and include protection of remaining old-growth pine forests, translocations of individuals to
sustain dwindling populations and insertion of artificial cavities in live pine trees (Conner et al., 2001). Cavities have been
identified as the critical resource for this species (Walters,
1991), and recovery measures using artificial cavities to induce new group formation in new territories (termed recruitment clusters) have been tremendously successful (Watson
et al., 1995; USFWS, 2003; Costa, 2004).
Previous modelling work showed that persistence of
existing populations could be enhanced when territories
are spatially aggregated (Letcher et al., 1998; Schiegg et al.,
2002b; Walters et al., 2002a). Since helpers do not disperse
beyond their immediate neighbourhood, dispersal success
declines dramatically with increasing distances among territories in this species (Schiegg et al., 2002b). Therefore it is
generally recommended that recruitment clusters be placed
within close proximity to existing red-cockaded woodpecker
territories. Dispersal distances are correlated to distances
between populations or territories in this and other species
(Breininger, 1999; Matthysen, 1999; Pasinelli et al., 2004).
Hence, spatial aggregation of territories may in the long
run elevate inbreeding risk. Current management of redcockaded woodpecker habitat thus involves a trade-off between facilitating dispersal and risking inducing high levels
of inbreeding.
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
545
Two previous modelling studies of population viability in
red-cockaded woodpeckers identified inbreeding as a serious
threat to small populations of this species and emphasized
the need to incorporate inbreeding depression into population viability analyses (Haig et al., 1993b; Daniels et al.,
2000). Inbreeding depression has been documented in the
red-cockaded woodpecker (Daniels and Walters, 2000a; Schiegg et al., 2002a), but its impact at the population level has
not been investigated previously.
Our objective is to quantify the impact of inbreeding
depression on population viability of small red-cockaded
woodpecker populations. Specifically, we ask the following
questions: (1) Is the probability that offspring are inbred related to parental natal dispersal distance? (2) How large is
the potential impact of inbreeding depression on long-term
persistence of small red-cockaded woodpecker populations?
(3) Does the spatial arrangement of territories influence probability of inbreeding? We address question (1) by analysing
empirical data from a long-term study of a relatively large
red-cockaded woodpecker population, and questions (2) and
(3) using an individual-based, spatially explicit simulation
model that incorporates the effects of environmental and
demographic stochasticity as well as those of inbreeding
depression.
2.
Methods
2.1.
Empirical data
2.1.1.
Study system and inbreeding coefficients
We obtained empirical estimates of inbreeding in red-cockaded woodpeckers by constructing pedigrees using
demographic data collected from the North Carolina Sandhills population, located in south-central North Carolina,
USA. This population contains about 560 groups, 220 of
which have been monitored since 1980. All adult birds
and all fledglings in the monitored groups were marked
with unique combinations of colour bands. All these groups
were censused each year and breeding status was assigned
based on behavioural observations following Walters et al.
(1988). Measures of reproductive success such as clutch
size, number of nestlings and number of fledglings were
recorded in all groups each year. Territory sizes in this population average 84 ha (Walters et al., 2002b). For more
details of population monitoring and censusing see Walters
et al. (1988).
Inbreeding coefficients (F) were calculated based on the
pedigree including all known individuals (7350 birds) in
the study population between 1980 and 2000 using PROC
INBREED in SAS (SAS Institute Inc, 1999–2001). The
inbreeding coefficient is the probability that a randomly
chosen locus contains alleles that are identical by descent
(Falconer and Mackay, 1996). There is no evidence that helpers ever sire offspring in this species, and the frequency of
extra-pair fertilizations by individuals outside the group is
among the lowest yet recorded in birds (Haig et al., 1993a,
1994). All individuals at the beginning of data collection
and those immigrating into the study area were considered
unrelated; hence our estimates of inbreeding levels are
conservative.
546
2.1.2.
B I O L O G I C A L C O N S E RVAT I O N
Dispersal behaviour and distance
We categorized birds that bred on their natal territory as
philopatric (natal dispersal distance = 0), and birds that emigrated from the natal group before breeding as dispersers (natal dispersal distance > 0). The latter category included birds
that were helpers on their natal territory for 1–8 years before
dispersing to a nearby territory to breed, as well as birds that
dispersed in their first year. Natal dispersal distance was calculated as the straight-line distance between centres of the
natal and the first breeding territory (Pasinelli et al., 2004). Female red-cockaded woodpeckers cross on average only two
territories during natal dispersal (median, n = 603, Daniels
and Walters, 2000b) and the corresponding number for males
is even lower (Daniels, 1997). We hence decided to use Euclidean distances rather than number of territories crossed in our
analyses to ensure sufficient variation in the data set. Because
dispersal distances in this species are so short (median dispersal distance, excluding philopatric individuals: males:
2089 m, interquartile range iqr = 974–4380, n = 488; females:
3689 m, iqr = 1687–5921, n = 958) and our study area is large
(110,000 ha), we are confident that our data set is not substantially biased towards short dispersal distances (see also Pasinelli et al., 2004). This assumption is further corroborated by a
high recovery rate of fledglings (2518 (44.2%) out of 6260 nestlings banded between 1980 and 2000 were resighted as adults
within the study area, own unpubl. data).
2.2.
Model simulations
2.2.1.
Model description
The model we employed allows for stochastic, spatially explicit, individual-based simulations and was built on the extensive demographic database from the Sandhills population
described above. Model parameter values and algorithms
used are described in detail in Letcher et al. (1998) and Schiegg
et al. (2005).
Landscape size (32 · 24 km) was chosen to enable us to
match the range of territory densities observed among existing populations (Letcher et al., 1998). Territories were at fixed
locations, because in red-cockaded woodpeckers territories
are centered on clusters of cavities constructed in live mature
pines that are used for decades (Walters, 1991). When a territory was unoccupied for five consecutive years, it was considered unsuitable and could not be reoccupied thereafter
(Copeyon et al., 1991). New territories were formed by budding, i.e., the splitting of an existing territory into two, which
occurred with an annual probability of 1%. Budding is the primary means by which new territories are formed in real populations of red-cockaded woodpeckers, but occurs at very low
rates (Hooper, 1983; Doerr et al., 1989; Conner et al., 2001).
A bird’s behaviour depended on its sex and status class
(i.e., fledgling, disperser, helper, breeder, solitary male). In
their first year, males either remained on their natal territory
as helpers or dispersed, depending on an empirically derived
proportion set by model parameters. All females dispersed in
their first year, because female helpers are rare (Walters et al.,
1988). Dispersal of first-year birds occurred in random directions and in a straight line. Birds could emigrate from the
study area by crossing the landscape boundary, but we did
not allow for immigration, because most existing red-cock-
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
aded woodpecker populations are isolated from one another
(Conner et al., 2001). The percentage of unbanded birds found
each year is only 5% in our study area (Daniels and Walters,
2000b), even though it encompasses only about one-half of
the Sandhills population and is adjacent to additional territories along its eastern boundary. Dispersing males moved at a
speed of 2.3 km yr1 and competed for breeding vacancies
within their search radius (3 km) each time step (3 months)
as they moved across the landscape. Dispersing females
moved at a speed of 4.8 km yr1 and, like dispersing males,
competed for breeding vacancies within their search radius
(3 km) each time step as they traversed the landscape. Breeding females have been shown to actively avoid incest by leaving their territory in >90% of the cases in which a son attained
the breeding position (Daniels and Walters, 2000a). This
behaviour was included in the model by invoking female
breeding dispersal when a son replaced a female’s mate.
Helpers attained breeding status by inheriting their natal
territory or by competing successfully for a breeding vacancy
in their vicinity (i.e., within their 3 km search range, Walters
et al., 1988). In case of competition for male vacancies, the
oldest resident helper wins, as IT invariably occurs in the real
population (Walters, 1990; Daniels and Walters, 2000a). If no
resident helpers are present, helpers from nearby territories
and dispersing birds compete for the vacancy and the closest
bird wins. Among equidistant competitors, the oldest bird
wins. Competition for vacancies among females was based
on age alone.
Mortality parameter values varied with sex and status
class based on Walters (1990). The number of fledglings produced on a territory each breeding season depended on the
probability of nesting successfully and the probable number
of young fledged if successful. These parameters were calculated as functions of breeder age and number of helpers
(Walters, 1990) (for equations and input parameters see Schiegg et al. (2005)). Variation in brood size was produced by generating the number of fledglings produced by a breeding pair
in a given year as the mean of a normal distribution with a
standard deviation estimated from empirical data (see Letcher et al. (1998) for additional justification of model
parameters).
Demographic stochasticity was simulated by applying annual survival probabilities, annual status transition probabilities of male fledglings, and probabilities of producing
different numbers of offspring to each individual. Non-emergent transition probabilities occurred when a random number
drawn from a uniform distribution [0;1] was less than the
appropriate transition probability value. Mortality probabilities and probabilities of producing different numbers of offspring
varied
annually
reflecting
environmental
stochasticity. The variance in these parameters was estimated from 14 years of data from the North Carolina Sandhills population, and was drawn randomly each year from
the resulting distribution to determine that year’s probability
value.
2.2.2.
Simulation of inbreeding depression
To incorporate inbreeding depression into the simulation
model, we estimated the number of lethal equivalents (Morton et al., 1956) using demographic data collected from the
B I O L O G I C A L C O N S E RVAT I O N
Sandhills and corresponding inbreeding coefficients as described above. A lethal equivalent is defined as ‘‘a group of
detrimental alleles that would cause on average one death if
homozygous’’ (Frankham et al., 2002). The probability of survival is dependent on inbreeding as follows:
S ¼ eðAþBFÞ
ð1Þ
A
where S is the probability of survival, e is fitness in an outbred population, F is the inbreeding coefficient, and B is a
measure of the increased deaths due to inbreeding, i.e., the
number of lethal equivalents per gamete (Hedrick, 1992). We
estimated values for A and B for survival from egg to age
one year, using data from 1985 to 2000 and one observation
per nest (n = 2278). We excluded observations prior to 1985 because that was the first year in which inbreeding was detected. We categorized each nest (with known clutch size) in
each year into one of seven levels of inbreeding (i.e., the kinship coefficient of the breeding pair) and calculated the proportion of eggs that survived to age one for each nest. We
then averaged these proportions within each level of inbreeding and performed a regression on the means weighted by
sample size (ln S = 1.49 + 4.93 * F; adj. R2 = 0.77, n = 2278), following Keller and Arcese (1998). The slope of this regression
model (B = 4.93) is the estimated number of lethal equivalents
per gamete. To incorporate this level of inbreeding depression
into the simulation model, we assigned all costs of inbreeding
(rounded to five lethal equivalents) to production of fledglings, for simplicity. Annual mean number of fledglings per
pair was reduced for related pairs by a value r (Hedrick,
1992), where
r ¼ 1 eðBFÞ
ð2Þ
and F is the coefficient of kinship of the breeding pair, equal
by definition to the inbreeding coefficient of the offspring (Falconer and Mackay, 1996). We treated breeding pairs with kinship coefficients <0.0625 as unrelated (F = 0), because most
pairs in populations simulated over long time periods will
be at least distantly related (Daniels et al., 2000), and reproduction of distantly related pairs was not found to be significantly different than that of unrelated pairs in the real
population (Daniels, 1997).
In calculating kinship coefficients for breeding pairs during a model simulation, all individuals were assumed to be
unrelated at the beginning of each run. Once a new pair
was formed during the course of a simulation, a kinship coefficient was calculated using a recursive path analysis algorithm and stored for later years to speed calculations.
2.2.3.
Model runs
We used six landscapes differing in territory number (25, 49 or
100 territories) and aggregation (scattered or clumped). Our
aim was to mimic features of current populations in terms
of size and territory distribution.
Spatial arrangement of territories was set using the k
parameter of the negative binomial distribution. The scattered arrangement was achieved by distributing territories
randomly in the landscape and this distribution was used to
define maximum k. Minimum k was set by maximal spatial
clumping of territories. Exact values of k were calculated by
overlaying a grid, 4 km to a side, over the landscape, counting
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
547
territory centres in each cell and using a maximum likelihood
estimate of k (Bliss and Fisher, 1953). This process was repeated for each particular number of territories used in the
simulations. To create the clumped aggregation level, we selected a k value 25% of the distance between log kmin and
log kmax (also see Letcher et al., 1998).
Runs started with all territories being occupied, either by
breeding pairs (90%) or by solitary males (10%). The number
of helpers was set equal to half of the number of territories.
Helper males were added to the landscape by assigning them
to territories containing breeding pairs selected at random.
Territories could be selected more than once, hence nearly
half the territories had one helper, and a few had more than
one, which mimicked the natural situation (Walters et al.,
1988). No dispersing birds or fledglings were present initially.
Each simulation was run over 100 years and was replicated
100 times. The stochastic algorithms included make each
run unique and ensure statistical independence among replicated runs of the same treatment.
Classification of dispersal behaviour and calculation of natal dispersal distances were performed as in the empirical
data set. To evaluate the impact of inbreeding depression on
extinction probability and time to extinction, we conducted
two sets of runs for each landscape. One set included inbreeding depression effects as described above, whereas the other
did not.
2.3.
Model sensitivity and suitability
Previous sensitivity analyses of a former model version that
did not include inbreeding depression showed that model predictions were stable when basic input parameters varied by
±10% (Letcher et al., 1998) and a formal validation procedure
indicated high predictive accuracy of the model even for complex output parameters such as distributions of dispersal distances or age of breeding individuals (Schiegg et al., 2005). We
performed additional sensitivity analyses with the model version used here on the effect of inbreeding depression on
extinction risk by conducting 100 runs with the two landscapes
representing the extreme conditions (100 territories, clumped;
25 territories, scattered). When increasing/decreasing the effect of inbreeding depression by 10%, predicted extinction risk
deviated from the nominal case by less than 0.45% in both
landscapes. We therefore consider the model to be insensitive
to small errors in estimates of inbreeding depression.
Models used to assess population viability under various
environmental scenarios (PVA models) have been criticised
to be only of limited use, because such models do not cover
the entire range of potential (behavioural) responses to
changes in resource availability and quality (Sutherland and
Norris, 2002). In fact, performance of some PVA models has
been poor when environmental conditions other than those
used for model parameterisation are simulated (Bradbury
and Payne, 2001). Here we use the model to explore extinction
risk in relation to variation of territory location and inbreeding levels, which are both expected to cause responses on
the individual and ultimately on the population level that
are accounted for by the model algorithms. We are therefore
confident that the predictions made for each scenario can
be meaningfully compared to each other.
548
B I O L O G I C A L C O N S E RVAT I O N
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
2.4.
Statistics
2.4.3. Effects of territory number and arrangement on
inbreeding probability and dispersal distance (model data)
2.4.1.
Dispersal behaviour and distance (empirical data)
We used the set of runs including inbreeding effects described
above to evaluate effects of population size and territory
aggregation on the number of inbred birds (F > 0). Because
several populations went extinct within 30 years of simulation (see below), we only analysed data from year 20. Results
did not change when we repeated the analyses using data
from year 10 instead of year 20. We used the proportion of
inbred individuals per simulation as the dependent variable
resulting in a potential sample size of 600 (3 landscape
sizes · 2 clumping levels · 100 replicates). Because 120 populations went extinct before year 20, final sample size was
480. We included a logit link to account for binomial data
(trial/events structure in SAS terminology: number of inbred
birds versus total number of birds). To model natal dispersal
distances as a function of population size and territory
arrangement, we used median natal dispersal distance (by
excluding philopatric individuals) per treatment and run as
the dependent variable in a general linear model. All analyses
were done with SAS 8.02 (SAS Institute Inc, 1999–2001).
Our approach was to test whether an individual’s inbreeding
coefficient was related to natal dispersal distance of its parents. The distribution of inbreeding coefficients was skewed
due to many zeros (6255 of 7350 birds (85.1%) censused during the study period with F = 0), which is to a large extent
caused by our assumption that all individuals were unrelated at the beginning of the study (see above). We therefore
used probability of being inbred (i.e., whether an individual’s
inbreeding coefficient was zero or not) as the dependent
variable in logistic regression models. Because natal dispersal distance was zero by definition for the philopatric
class (i.e., for helpers that inherited their natal territory)
we first tested whether offspring of philopatric parents were
more likely to be inbred than young whose parents had dispersed. This was done by logistic regression with parental
dispersal behaviour (philopatric or disperser) as a class
and year as a continuous predictor variable. To explore a potential sex specific effect of dispersal behaviour, we also included gender of the parent and its interaction with
dispersal behaviour as additional predictor variables. The
dependent variable was the probability of offspring being
inbred.
We then tested for an effect of parental natal dispersal
distance on the probability that offspring were inbred by
using parental natal dispersal distance (if >0) and year as
continuous independent variables in another logistic regression model. Because natal dispersal distance is larger in females than in males (Walters et al., 1988; Pasinelli et al.,
2004), we analysed the sexes separately. Model fit was
checked using residual analyses and likelihood ratios (Chatterjee and Price, 1991). If natal dispersal distances of both
parents were available, one dispersal distance was randomly
selected and used in the analyses (PROC SURVEYSELECT in
SAS, using simple random sampling (SAS Institute Inc,
1999–2001)). This resulted in a sample size of 487 male and
958 female dispersal distances recorded in the years 1980–
2001. Sample sizes for males are lower than for females because the philopatric class has been removed from this data
set. Results did not change, however, if individuals with natal dispersal distance = 0 were included.
2.4.2.
data)
Extinction probability and time to extinction (model
We defined extinction probability as the proportion of the 100
replicated model runs in which the simulated population
went extinct before the end of the simulation, i.e., before
100 years. Time to extinction was the number of years these
populations persisted, i.e., had at least one breeding pair.
The influence of territory number (continuous variable), spatial arrangement (two levels: scattered or clumped) and
inbreeding (two levels: whether simulations were performed
with or without inbreeding depression) on extinction probability and time to extinction were tested with generalized linear models. Model fit was assessed by residual analyses
(McCullagh and Nelder, 1989). Sample size was 1200 (3 landscape sizes · 2 clumping levels · 2 inbreeding levels inbreeding yes/no · 100 replicates).
3.
Results
3.1.
Probability of being inbred as a function of parental
natal dispersal behaviour and dispersal distance
Inbreeding coefficients ranged between F = 0–0.280. Dispersal
behaviour of parents, i.e., whether a parent bred on the natal
territory or dispersed before breeding, did not affect probability of offspring being inbred (logistic regression, likelihood ratio test, v2 = 0.13, df = 1, p = 0.715), nor were there sex-specific
effects (sex · dispersal behaviour: v2 = 0.23, df = 1, p = 0.631,
n = 1669). Likelihood of offspring being inbred increased with
year (v2 = 174.36, df = 1, p < 0.001, n = 1669). Because our sample was biased towards offspring with F = 0 (1354 out of 1669
individuals), we repeated these analyses with a data set that
contained all offspring with F > 0 (n = 315) and an equal number of offspring with F = 0 that were randomly selected from
all offspring with F = 0. There was still no relation between
parental dispersal behaviour and probability of offspring
being inbred (dispersal behaviour: v2 = 0.56, df = 1, p = 0.454,
sex · dispersal behaviour: v2 = 0.23, df = 1, p = 0.629, n = 630).
In contrast, an individual’s inbreeding coefficient did depend on the natal dispersal distance of its mother, but not
of its father (Table 1). The farther the mother had dispersed
from the natal territory the less likely it was that her offspring
were inbred. Again, year was positively related to likelihood of
offspring being inbred.
3.2.
Long-term stability of small populations
Extinction risk of simulated populations was significantly higher when inbreeding depression was included than when it was
not (estimate ± SE = 2.87 ± 0.28; v2 = 107.72, df = 1, p < 0.001). Increased population size (4.01 ± 0.45; v2 = 77.94, df = 1,
p < 0.001) and territory clumping (2.84 ± 0.28; v2 = 106.31,
df = 1, p < 0.001, n = 1200 in all cases, Fig. 1) diminished probability of extinction. In simulations without inbreeding, territory
clumping decreased extinction risk of populations containing
B I O L O G I C A L C O N S E RVAT I O N
549
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
Table 1 – Logistic model of the probability of an offspring being inbred as a function of parental natal dispersal distance
(distance, km) and time (year)
Parameter
DF
Females (N = 958)
Estimate ± SE
Intercept
Distance
Year
1
1
1
460.10 ± 44.50
0.06 ± 0.03
0.23 ± 0.02
Males (N = 487)
2
v
p
Estimate ± SE
v2
p
106.87
5.11
106.46
<0.001
0.024
<0.001
393.91 ± 59.37
0.05 ± 0.04
0.20 ± 0.03
44.03
2.19
43.80
<0.001
0.139
<0.001
Sexes are analyzed separately. Modelled was the probability that inbreeding occurred.
scattered
scattered
100
Median time to extinction (years)
100
Extinction probability
50
0
25
49
100
clumped
100
50
0
50
0
49
50
3.3.
Probability of being inbred as a function of spatial
arrangement of territories
Increasing spatial clumping of territories led to a significant
reduction in median natal dispersal distance in the simulated
25
49
100
population size
population size
100 territories from 77% to 2%. However, when inbreeding
depression was included in the simulation process, clumping
reduced extinction risk only from 100% to 78% (Fig. 1).
Inbreeding also significantly reduced time to extinction
(estimate ± SE = 7.32 years ± 0.82; F = 79.12, p < 0.001), while
increasing population size (20.41 years ± 0.51; F = 1613.87,
p < 0.001) and territory clumping (24.95 years ± 0.82;
F = 919.58, p < 0.001, n = 993 in all cases) increased time to
extinction. Populations with a high extinction risk, i.e., very
small, inbred populations with scattered territory arrangement, mostly went extinct within the first 40 years of simulation (Fig. 2). That time to extinction in populations containing
100 clumped territories was shorter in the absence of inbreeding depression is an artefact, because the value for populations without inbreeding depression was derived from only
two populations (see Fig. 1, clumped territory arrangement,
population size = 100). We conclude that inbreeding depression severely hampered persistence of small red-cockaded
woodpecker populations.
100
clumped
100
Fig. 1 – Extinction probability (percent populations extinct
within 100 years of simulation in 100 model runs) in
populations with (black bars) and without (white bars)
inbreeding, in relation to population size and territory
arrangement (scattered or clumped).
49
100
0
25
25
Fig. 2 – Time to extinction (medians and interquartile
ranges) in populations with (black bars) and without (white
bars) inbreeding in relation to population size and territory
arrangement. The number of populations that went extinct
within 100 years (n) is shown for each treatment in Fig. 1.
populations irrespective of their size (estimate ± SE =
0.59 ± 0.07; F = 64.40, p < 0.001, n = 480). The proportion of
inbred individuals was higher when territories were clumped
rather than scattered (estimate ± SE = 0.14 ± 0.04; v2 = 12.66,
df = 1, p < 0.001, n = 480).
4.
Discussion
4.1.
Natal dispersal distance and probability of inbreeding
In our natural study population birds that remained on their
natal territory to breed were not more likely to produce inbred
offspring than were birds that dispersed from the natal territory. This finding is surprising and attests to the effectiveness
of incest avoidance behaviour (Walters et al., 1988; Daniels
and Walters, 2000a). Daniels and Walters (2000b) documented
that females leave their breeding site when a son inherits
breeding status on the territory. While such behaviour reduces inbreeding risk among familiar individuals on the natal
site, dispersing females do not avoid territories held by closely related males with whom they were previously unfamiliar (Daniels and Walters, 2000b). As a consequence, we found
a negative relationship between natal dispersal distance of
the mother and the probability of her offspring being inbred.
To our knowledge this is the first time that a quantitative
550
B I O L O G I C A L C O N S E RVAT I O N
relationship between natal dispersal distance and inbreeding
probability has been established empirically. Presumably this
pattern is a result of the spatially restricted dispersal of helpers, who do not move beyond their neighbourhood (Walters
et al., 1988; Daniels, 1997). Previously Daniels and Walters
(2000b) showed that the probability that a female dispersing
from her natal territory will encounter a related male declines
once a female crosses a sufficient number of territories, and
that dispersal distance increases when territory density
decreases.
The positive association between year and likelihood of
offspring being inbred suggests that inbreeding probability increases with time. However, this effect may instead be a consequence of our assumption that all birds were unrelated at
the beginning of data collection.
4.2.
Persistence of small red-cockaded woodpecker
populations
Comparing population simulations with and without costs of
inbreeding revealed that inbreeding depression significantly
enhanced extinction risk of small red-cockaded woodpecker
populations. Inbreeding also affected time to extinction,
although this effect was less pronounced (Fig. 2). Most populations that went extinct did so within 40 years of simulation and within this time frame, inbreeding did not greatly
accelerate this process. Our model of population dynamics
was unusual and perhaps unique in that it simultaneously
incorporated effects of territory aggregation, environmental
and demographic stochasticity, and an empirically derived,
species-specific estimate of inbreeding costs. Thus, this
study provides evidence that inbreeding elevates extinction
risk of small populations beyond levels induced by environmental or demographic stochasticity. Our results support
those of Brook et al. (2002), who evaluated the role of
inbreeding in extinction by conducting population viability
analyses on 20 threatened species of various taxa. Inbreeding markedly decreased median times to extinction by
25–31% for populations ranging in size from 50 to 1000 individuals. Further, Spielman et al. (2004) compared heterozygosities of 177 threatened species with those of taxonomically
related non-threatened species and found that in the majority of comparisons heterozygosities were lowered in the
threatened taxa. They concluded that species generally are
not driven to extinction before genetic factors become
important.
Our simulations may underestimate the impact of
inbreeding depression on population viability of red-cockaded
woodpeckers. First, we ignored the known effect of inbreeding on timing of egg-laying (Schiegg et al., 2002a). Second, effects of genetic drift, which may diminish genetic variability
and hence additionally increase inbreeding risk, were not
considered. Finally, fitness costs of moderate inbreeding
(F < 0.0625), if any exist, were not considered. On the other
hand, our model does not allow for potential effects of purging, whereby recessive deleterious alleles are removed
through natural selection (Frankham et al., 2001). The impact
of purging on extinction risk is controversial (Miller and Hedrick, 2001), however, and in their review of inbreeding depression in wild populations, Keller and Waller (2002) suggest that
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
purging has only small effects on inbreeding depression in
fragmented populations (see also Brook et al., 2002).
Five lethal equivalents as used in this study represent a
substantial cost of inbreeding, higher than those reported
for other species (reviewed by Keller et al., 2002). Costs of
inbreeding included in the calculations, however, vary greatly
among studies, as different life stages are used in survival
estimates. Keller and Arcese (1998) calculated 3.21 lethal
equivalents from egg to age one in an insular population of
song sparrows Melospiza melodia when estimated as one unit
egg to age one as we did here. Yet, we were conservative in
our assessment of the consequences of inbreeding depression, both in the calculation of lethal equivalents and in
assigning those costs in model simulations. We conclude that
inbreeding depression poses a high risk for small populations
of red-cockaded woodpeckers, and we further suggest that
inbreeding costs and, consequently, the threat of those costs
to population viability, may be higher for this and perhaps
other species than currently recognized.
Short distances between territories have repeatedly been
shown to be essential for the (demographic) stability of small
red-cockaded woodpecker populations (Letcher et al., 1998;
Schiegg et al., 2002b; Walters et al., 2002a), but these studies
did not include possible negative effects of territory aggregation arising from inbreeding depression. Our simulation model
tends to overestimate natal dispersal distances in both sexes
(Schiegg et al., 2005). Given that real distances are even shorter and are negatively related to inbreeding risk at least in females, as shown above, the association of territory clumping
and inbreeding risk may even be stronger than predicted
here. Our results hence suggest that territory aggregation increases inbreeding and thus the benefits of short interterritorial distances may be lessened. On the other hand, extinction
risk was still reduced by a clumped territory arrangement. We
therefore suggest that measures to lower the risk of inbreeding should be undertaken, when spatial aggregation of territories is used to sustain small red-cockaded woodpecker
populations.
4.3.
Inbreeding in cooperative breeders
Avoidance of inbreeding is generally thought to be one of the
major forces in the evolution of dispersal (Pusey and Wolf,
1996; Perrin and Goudet, 2001). Our empirical data reveal
how increasing natal dispersal distances enhances the
chance of mating with an unrelated partner, thereby reducing the odds of incurring fitness costs due to inbreeding.
Nevertheless, natal dispersal distances in red-cockaded
woodpeckers and other cooperatively breeding bird species
are very short in general (Zack, 1990; Emlen, 1991), suggesting the existence of benefits of short distance dispersal that
outweigh costs of inbreeding. Benefits of short distance dispersal may include competitive advantages in obtaining
breeding vacancies (Zack and Rabenold, 1989), which usually
are scarce in cooperative breeders. Further, territory owners
usually are less aggressive towards related neighbours (Watson et al., 1994), and may even share parts of the territory to
gain inclusive fitness benefits by enabling a relative to breed
(Perrin and Goudet, 2001). Also, short distance dispersal, by
maintaining low levels of inbreeding, may increase inclusive
B I O L O G I C A L C O N S E RVAT I O N
fitness benefits to helpers, which are essential for the maintenance of cooperative behaviour by kin selection (Stevens
and Wiley, 1995). Finally, previous work revealed that lifetime production of fledglings declines with increasing dispersal distance in red-cockaded woodpeckers (Pasinelli
et al., 2004).
4.4.
Conclusions
Consideration of reductions in fitness due to inbreeding is still
the exception rather than the rule in population viability
analyses. We recommend including an estimate of inbreeding
depression in stochastic models whenever possible and using
sensitivity analyses to assess potential impacts of inbreeding
on population persistence.
Our findings raise particular concern about the persistence
of the red-cockaded woodpecker. Because most of the remaining populations contain fewer than 80 breeding groups
(USFWS, 2003), immigration is rare or nonexistent (Conner
et al., 2001) and dispersal is restricted, this species is especially
prone to inbreeding accumulation. We hence recommend
increasing efforts to reduce the potential for inbreeding in
the management regime for the red-cockaded woodpecker,
especially when spatial clumping of territories is used to facilitate dispersal within populations. For instance, the size of extant populations can be enlarged by creating artificial cavities
(Walters, 1991) provided that suitable habitat is available. This
strategy is already applied successfully in several populations
(Watson et al., 1995; USFWS, 2003; Costa, 2004). Increasing connectedness between populations may be another option, but
would require in most cases the reforestation of residential
areas. Finally, we support the recommendation in the species’
Recovery Plan (USFWS, 2003) to translocate individuals between populations to reduce inbreeding levels.
Acknowledgements
We thank the Swiss National Science Foundation for financial support (Grant 81EZ-57341 to K. Schiegg and
81ZH57449 to G. Pasinelli). We greatly appreciate the contributions of Jay Carter, Phillip Doerr and Kerry Brust, whose
continuing efforts ensured the extension of our data set.
We thank Benjamin Letcher and Larry Crowder for their
work on the model and Lukas Keller for helpful comments
on the manuscript. Collection of data was supported by
the National Science Foundation (BSR-8307090, BSR8717683) and the Department of Defence, Department of
the Army, Fort Bragg. Finally, we thank the many students
and technicians from Virginia Tech and North Carolina State
University, and the staff of the Sandhills Ecological Institute,
who participated in data collection.
R E F E R E N C E S
Bliss, C.I., Fisher, R.A., 1953. Fitting the negative binomial
distribution to biological data and a note on efficiently fitting
the negative binomial. Biometrics 9, 176–200.
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
551
Bradbury, R.B., Payne, R.J.H., 2001. Predicting population
responses to resource management. Trends in Ecology and
Evolution 16, 440–445.
Breininger, D.R., 1999. Florida scrub-jay demography and
dispersal in a fragmented landscape. Auk 116, 520–527.
Brook, B.W., Tonkyn, D.W., Q’Grady, J.J., Frankham, R., 2002.
Contribution of inbreeding to extinction risk in threatened
species. Conservation Ecology 6 (article no. 16).
Caughley, G., 1994. Directions in conservation biology. Journal of
Animal Ecology 63, 215–244.
Chatterjee, S., Price, B., 1991. Regression Analysis by Example,
second ed. Wiley, New York, USA.
Conner, R.N., Rudolph, D.C., Walters, J.R., 2001. The Red-cockaded
Woodpecker: Surviving in a Fire-maintained Ecosystem.
University of Texas Press, Austin.
Copeyon, C.K., Walters, J.R., Carter, J.H., 1991. Induction of
red-cockaded woodpecker group formation by artificial
cavity construction. Journal of Wildlife Management 55,
549–556.
Costa, R., 2004. State of the red-cockaded woodpecker
world – highlights of the previous decade 1992–2002. In: Costa,
R., Daniels, S.J. (Eds.), Red-cockaded Woodpecker: Road to
Recovery. Hancock House Publishing, Blaine, WA, pp. 39–46.
Cutrera, A.P., Lacy, E.A., Busch, C., 2005. Genetic structure in a
solitary rodent (Ctenomys talarum): implications for kinship
and dispersal. Molecular Ecology 14, 2511–2523.
Daniels, S.J., 1997. Female dispersal and inbreeding in the
red-cockaded woodpecker. Master Thesis, Virginia Polytechnic
Institute and State University, Blacksburg, VA.
Daniels, S.J., Priddy, J.A., Walters, J.R., 2000. Inbreeding in small
populations of red-cockaded woodpeckers: insights from a
spatially-explicit individual-based model. In: Young, A.G.,
Clarke, G.M. (Eds.), Genetics, Demography and Viability of
Fragmented Populations. Cambridge University Press, London,
pp. 129–147.
Daniels, S.J., Walters, J.R., 2000a. Between-year breeding dispersal
in red-cockaded woodpeckers: multiple causes and estimated
cost. Ecology 81, 2473–2484.
Daniels, S.J., Walters, J.R., 2000b. Inbreeding depression and its
effects on natal dispersal in red-cockaded woodpeckers.
Condor 102, 482–491.
Doerr, P.D., Walters, J.R., Carter III, J.H., 1989. Reoccupation of
abandoned clusters of cavity trees colonies by red-cockaded
woodpeckers. In: Proceedings of the Annual Conference of the
Southeastern Association of Fish and Wildlife Agencies, vol.
43, pp. 326–336.
Emlen, S.T., 1991. Evolution of cooperative breeding in birds and
mammals. In: Krebs, J.R., Davies, N.B. (Eds.), Behavioral
Ecology: An Evolutionary Approach. Blackwell Scientific
Publications, Oxford, pp. 301–337.
Falconer, D.S., Mackay, T.F.C., 1996. Introduction to Quantitative
Genetics, fourth ed. Longman, Essex.
Frankham, R., Ballou, J.D., Briscoe, D.A., 2002. Introduction to
Conservation Genetics. Cambridge University Press,
Cambridge.
Frankham, R., Gilligan, D.M., Morris, D.R., Briscoe, D.A., 2001.
Inbreeding and extinction: effects of purging. Conservation
Genetics 2, 279–284.
Gibbs, J.P., 2001. Demography versus habitat fragmentation as
determinants of genetic variation in wild populations.
Biological Conservation 100, 15–20.
Haig, S.M., Belthoff, J.R., Allen, D.H., 1993a. Examination of
population structure in red-cockaded woodpeckers using DNA
profiles. Evolution 47, 185–194.
Haig, S.M., Belthoff, J.R., Allen, D.H., 1993b. Population viability
analysis for a small population of red-cockaded woodpeckers
and an evaluation of enhancement strategies. Conservation
Biology 7, 289–301.
552
B I O L O G I C A L C O N S E RVAT I O N
Haig, S.M., Walters, J.R., Plissner, J.H., 1994. Genetic evidence for
monogamy in the cooperatively breeding red-cockaded
woodpecker. Behavioral Ecology and Sociobiology 34, 295–303.
Hedrick, P.W., 1992. Genetic conservation in captive populations
and endangered species. In: Jain, S.K., Botsford, L.W. (Eds.),
Applied Population Biology. Kluwer, Dordrecht, Netherlands,
pp. 45–68.
Hedrick, P.W., Kalinowski, S.T., 2000. Inbreeding depression in
conservation biology. Annual Review of Ecology and
Systematics 31, 139–162.
Höglund, J., Piertney, S.B., Alatalo, R.V., Lindell, J., Lundberg, A.,
Rintamäki, P.T., 2002. Inbreeding depression and male fitness
in black grouse. Proceedings of the Royal Society of London,
Series B 269, 711–715.
Hooper, R.G., 1983. Colony formation by red-cockaded
woodpeckers: hypotheses and management implications. In:
Wood, D.A. (Ed.), Red-cockaded Woodpecker Symposium II.
Florida Game and Freshwater Fishery Commission, US Fish
and Wildlife Service, Atlanta, GA, pp. 72–77.
Keller, K.L., Arcese, P., 1998. No evidence for inbreeding avoidance
in a natural population of Song Sparrows Melospizia melodia.
The American Naturalist 152, 380–392.
Keller, L.F., Grant, P.R., Grant, B.R., Petren, K., 2002. Environmental
conditions affect the magnitude of inbreeding depression in
survival of Darwin’s finches. Evolution 56, 1229–1239.
Keller, L.F., Waller, D.M., 2002. Inbreeding effects in wild
populations. Trends in Ecology and Evolution 17, 230–241.
Lande, R., 1988. Genetics and demography in biological
conservation. Science 241, 1455–1460.
Letcher, B.H., Priddy, J.A., Walters, J.R., Crowder, L.B., 1998. An
individual-based, spatially-explicit simulation model of the
population dynamics of the endangered red-cockaded
woodpecker, Picoides borealis. Biological Conservation 86, 1–14.
Matthysen, E., 1999. Nuthatches Sitta Europaea: Aves in forest
fragments: demography of a patchy population. Oecologia 119,
501–509.
McCullagh, P., Nelder, J.A., 1989. Generalized Linear Models,
second ed. Chapman & Hall, London, New York.
Miller, P.M., Hedrick, P.W., 2001. Purging of inbreeding depression
and fitness declines in bottlenecked populations of Drosophila
melanogaster. Journal of Evolutionary Biology 14, 595–601.
Morton, N.E., Crow, J.F., Muller, H.J., 1956. An estimate of the
mutational damage in man from data on consanguineous
marriages. Proceedings of the National Academy of Sciences
42, 855–863.
Pasinelli, G., Schiegg, K., Walters, J.R., 2004. Genetic and
environmental influences on natal dispersal distance in a
resident bird species. The American Naturalist 164, 660–669.
Perrin, N., Goudet, J., 2001. Inbreeding, kinship, and the evolution
of natal dispersal. In: Clobert, J., Danchin, J.E., Dhondt, A.A.,
Nichols, J.D. (Eds.), Dispersal. Oxford University Press, Oxford,
pp. 123–142.
Pusey, A., Wolf, M., 1996. Inbreeding avoidance in animals. Trends
in Ecology and Evolution 11, 201–206.
Reed, D.H., 2004. Extinction risk in fragmented habitats. Animal
Conservation 7, 181–191.
Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius, W.,
Hanski, I., 1998. Inbreeding and extinction in a butterfly
metapopulation. Nature 392, 491–494.
SAS Institute Inc., 1999–2001. SAS Proprietary Software Release
8.02. Cary, NC.
Schiegg, K., Pasinelli, G., Walters, J.R., Daniels, S.J., 2002a.
Inbreeding and experience affect response to climate change
by endangered woodpeckers. Proceedings of the Royal Society
of London, Series B 269, 1153–1159.
1 3 1 ( 2 0 0 6 ) 5 4 4 –5 5 2
Schiegg, K., Priddy, J.R., Walters, J.R., 2005. Testing a spatially
explicit, individual-based model of the population dynamics
of red-cockaded woodpeckers. Ecological Applications 15,
1495–1503.
Schiegg, K., Walters, J.R., Priddy, J.A., 2002b. The consequences of
disrupted dispersal in fragmented red-cockaded woodpecker
populations. Journal of Animal Ecology 71, 710–721.
Spielman, D.B., Brook, W.B., Frankham, R., 2004. Most species are
not driven to extinction before genetic factors impact them.
Proceedings of the National Academy of Sciences of the United
States of America 101, 15261–15264.
Stevens, E.E., Wiley, R.H., 1995. Genetic consequences of restricted
dispersal and incest avoidance in a cooperatively breeding
wren. Journal of Theoretical Biology 175, 423–436.
Sutherland, W.J., Norris, K., 2002. Behavioural models of
population growth rates: implications for conservation and
prediction. Proceedings of the Royal Society of London, Series
B 357, 1273–1284.
Tanaka, Y., 2000. Extinction of populations by inbreeding
depression under stochastic environments. Population
Ecology 42, 55–62.
USFWS, 2003. Red-cockaded Woodpecker Picoides borealis
Recovery Plan: Second Revision. US Fish and Wildlife Service,
Atlanta.
Walters, J.R., 1990. Red-cockaded woodpeckers: a ‘primitive
cooperative breeder. In: Stacey, P.B., Koenig, W.D. (Eds.),
Cooperative Breeding in Birds: Long-term Studies of Ecology
and Behavior. Cambridge University Press, Cambridge, UK, pp.
67–101.
Walters, J.R., 1991. Application of ecological principles to the
management of endangered species: the case of the
red-cockaded woodpecker. Annual Review of Ecology and
Systematics 22, 505–523.
Walters, J.R., Cooper, C.B., Daniels, S.J., Pasinelli, G., Schiegg, K.,
2004. Conservation Biology. In: Koenig, W.D., Dickison, J. (Eds.),
Ecology and Behaviour of Cooperative Breeding in Birds.
Cambridge University Press, Cambridge, pp. 197–209.
Walters, J.R., Crowder, L.B., Priddy, J.A., 2002a. Population viability
analysis for red-cockaded woodpeckers using an
individual-based model. Ecological Applications 12, 249–260.
Walters, J.R., Daniels, S.J., Carter III, J.H., Doerr, P.D., 2002b.
Defining quality of red-cockaded woodpecker foraging habitat
based on habitat use and fitness. Journal of Wildlife
Management 66, 1064–1082.
Walters, J.R., Doerr, P.D., Carter III, J.H., 1988. The cooperative
breeding system of the red-cockaded woodpecker. Ethology 78,
275–305.
Watson, A., Moss, R., Parr, R., Mountford, M.D., Rothery, P., 1994.
Kin landownership, differential aggression between kin and
non-kin, and population fluctuations in red grouse. Journal of
Animal Ecology 63, 39–50.
Watson, J.C., Hooper, R.G., Carlson, D.L., Taylor, W.E., Milling,
T.E., 1995. Restoration of the red-cockaded woodpecker
population on the Francis Marion National Forest: three
years post Hugo. In: Kulhavy, D.L., Hooper, R.G., Costa, R.
(Eds.), Red-cockaded Woodpecker: Recovery, Ecology and
Management. Center for Applied Studies in Forestry, College
of Forestry, Stephen F. Austin State University,
Nacogdoches, TX, pp. 172–182.
Zack, S., 1990. Coupling delayed breeding with short-distance
dispersal in cooperatively breeding birds. Ethology 86,
265–286.
Zack, S., Rabenold, K.N., 1989. Assessment, age and proximity in
dispersal contests among cooperative wrens: field
experiments. Animal Behaviour 38, 235–247.