ChapterPDF Available

Abstract and Figures

Wherever wildlife management concerns the movement of individuals across structured habitat, its scale of operations will encompass metapopulation dynamics. The goal of this chapter is to review the potential applications of metapopulation concepts and models in reserve design and conservation management. Our perspective is forward-looking. We show how some key problems of where to direct conservation effort and how to manage populations can be addressed in the context of regional habitat structure and the survival and renewal of habitat patches. We also mention several cases of successful metapopulation management and point out practical problems. We emphasise (1) that the viability of a population may depend on surrounding populations, in which case metapopulation processes influence or determine reserve design and management options; (2) that understanding the dynamic processes requires models, which make assumptions that need validating; (3) that the principle limitation of metapopulation models is their single-species focus. Conservation strategies clearly depend on the particular social, economic and ecological circumstances of each region, and concepts such as the metapopulation can seem irrelevant to practical concerns. We aim to show, nevertheless, that an understanding of metapopulation dynamics can be vital to asking pertinent questions and seeking potential solutions. The conceptual framework of metapopulation dynamics tells us what information is needed in order to build case-specific models relevant to any of a wide range of issues. These issues include the potential disadvantages of habitat corridors, or hidden benefits of sink habitat; the optimal schedule for translocations or reintroductions; the relative merits of reducing local extinctions against increasing colonisations; the optimum distribution of habitat improvement; and the advantages of increasing life spans of ephemeral habitats.
Content may be subject to copyright.
The role of metapopulations
in conservation
H. Resit Akc¸akaya, Gus Mills
and C. Patrick Doncaster
Nothing in the world is single;
All things, by a law divine,
In one another’s being mingle.
(Percy Bysshe Shelley (1792–1822), ‘Love’s Philosophy’)
Wherever wildlife management concerns the
movement of individuals across structured
habitat, its scale of operations will encompass
metapopulation dynamics. The goal of this
essay is to review the potential applications of
metapopulation concepts and models in reserve
design and conservation management. Our
perspective is forward-looking. We show how
some key problems of where to direct conser-
vation effort and how to manage populations
can be addressed in the context of regional
habitat structure and the survival and renewal
of habitat patches. We also mention several
cases of successful metapopulation manage-
ment and point out practical problems (for
example, see Box 5.1)
We emphasize:
1. that the viability of a population may de-
pend on surrounding populations, in which
case metapopulation processes influence or
determine reserve design and management
2. that understanding the dynamic processes
of populations requires models, which
make assumptions that need validating;
3. that the principle limitation of metapopula-
tion models is their single-species focus.
Conservation strategies clearly depend on the
particular social, economic and ecological cir-
cumstances of each region, and concepts such
as the metapopulation can seem irrelevant to
practical concerns. We aim to show, neverthe-
less, that an understanding of metapopulation
dynamics can be vital to asking pertinent
questions and seeking potential solutions. The
conceptual framework of metapopulation dy-
namics tells us what information is needed in
order to build case-specific models relevant to
any of a wide range of issues. These issues in-
clude: the potential disadvantages of habitat
corridors, or hidden benefits of sink habitat;
the optimal schedule for translocations or re-
introductions; the relative merits of reducing
local extinctions against increasing coloniza-
tions; the optimum distribution of habitat im-
provement; and the advantages of increasing
life spans of ephemeral habitats.
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 64 6.5.2006 2:48am
We define a metapopulation as a set of dis-
crete populations of the same species, in the
same general geographical area, that may ex-
change individuals through migration, disper-
sal, or human-mediated movement (based on a
very similar definition by Hanski & Simberloff
1997). Older, more restrictive definitions of
metapopulation (e.g. Hanski & Gilpin 1991)
reflect particular approaches to modelling, for
example, by requiring that populations have
independent (uncorrelated) fluctuations, are
all equally connected by dispersal (Levins’ ‘is-
land–island’ model), or that one population is
much larger and less vulnerable than the others
(MacArthur and Wilson’s ‘mainland–island’
model). Most criticisms of the metapopulation
concept (e.g. Dennis et al. 2003) arise from
shortcomings of these more restrictive defin-
itions (Baguette & Mennechez 2004). Over
the past decade, the trend in metapopulation
concepts has moved from abstract models to-
ward real-world applications. Our more general
definition has only two requirements: (i) popu-
lations are geographically discrete; (ii) mixing of
individuals between populations is less than
that within them otherwise the regional as-
semblage of local populations may be more aptly
described as a single panmictic population.
Within these limits, the definition encompasses
all levels of variation between populations in
colonization rates (including the extreme of
‘source–sink’ systems, detailed later in this
essay) and in extinction rates (including syn-
chronous extinctions, detailed later in this
essay). We emphasize that a metapopulation is
a dynamic system of linked populations, as op-
posed to simply a patchy habitat, and many of its
demographic processes are visible only through
the filter of models.
Although the focus of this essay is on species
conservation in habitat fragmented by human
activities, metapopulations occur in a variety of
forms without any human intervention. Many
species depend on habitat patches created by
natural disturbances such as fires. Other ex-
amples of natural metapopulations include
species inhabiting discrete water bodies such
as ponds and lakes; despite the physical isol-
ation of freshwater habitats, their populations
of aquatic plants and invertebrates may be
widely interconnected by birds inadvertently
transporting propagules between them (Figuer-
ola & Green 2002), and their populations of
amphibians are often interconnected by sea-
sonal dispersal through the landscape. Amongst
mammals the Ethiopian wolf (Canis simensis)
is naturally confined to rodent-rich alpine
meadows, but is threatened with extinction by
the intervening terrain between plateaux be-
coming too hostile to allow safe passage (Mac-
donald & Sillero 2004). Mountain sheep (Ovis
canadensis) populations in southern California
inhabit mountain ‘islands’ in a desert (Fig.
5.1); this species cannot live for long in the
desert, but it can migrate through it (Bleich
et al. 1990).
A sink is a population with deaths exceeding
births and extinction only averted by immi-
grants exceeding emigrants. Conversely, a
source is a population with a net outflux of
individuals. The identification of sources and
sinks is complicated by temporal and spatial
variability, and density dependence in demog-
raphy and dispersal (detailed later in this
Habitat corridors are more-or-less linear
strips of habitat with a designed or incidental
function of increasing dispersal among popula-
tions. We focus specifically on human-modified
habitat, additional to natural linear features
(such as riparian habitat) that may already
link populations. Corridors such as field mar-
gins supplement hedgerows which were
planted to meet needs not directly related to
conservation, but which are increasingly nur-
tured for their conservation value. Corridors
may provide a continuous stretch of habitat
between populations, or discontinuous patches
that improve connectivity in ‘stepping-stone’
fashion. A corridor for movement in one direc-
tion may simultaneously act as a barrier in the
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 65 6.5.2006 2:48am
perpendicular direction (such as road verge:
Rondinini & Doncaster 2002).
Issues and options
Does conservation need metapopulation
Animals and plants may occupy metapopula-
tions wherever landscapes are either naturally
heterogeneous, or fragmented as a result of
human activities such as habitat loss to urban-
ization, agriculture and transportation routes.
Metapopulations are thus relevant to the con-
servation of any patchy or fragmented habitat.
They are also relevant to the conservation of a
single population if its dynamics depend on
those of neighbouring populations.
One misunderstanding is that the use of the
metapopulation concept in conservation re-
quires or implies the conservation or manage-
ment of species as multiple populations. In
some cases, maintaining more than one popu-
lation does increase the persistence of the
species as a whole, but this is neither universal,
nor a necessary result of using a metapopula-
tion approach. Thus, what conservation needs
is not necessarily metapopulations per se, but
the metapopulation approach and concepts,
which permit assessment of the persistence of
a species that happens to exist in a metapopula-
tion, either naturally or due to habitat loss and
fragmentation. The metapopulation concept is
important because species that exist in a meta-
population face particular issues related to en-
vironmental impacts, and have conservation
options that can be evaluated more completely,
or only, in a metapopulation context. These are
discussed in the next two sections.
Environmental impacts
in a metapopulation context
Metapopulations can be affected by impacts on
their entirety or on the individual components.
Impacts studied at the regional level include
roads and other dispersal barriers that decrease
0 50 km
Fig. 5.1 Populations of mountain sheep (Ovis canadensis) in southern California. Shaded areas indicate
mountain ranges with resident populations, arrows indicate documented intermountain movements; the
dotted lines show fenced highways. (After Bleich et al. 1990; reprinted from Akc¸akaya et al. 1999 with
permission from Applied Biomathematics.)
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 66 6.5.2006 2:48am
connectivity of populations, and habitat frag-
mentation that divides a homogeneous popula-
tion into several smaller populations. The
effects of such factors on the overall viability
of the species involve interactions among popu-
lations (e.g. dispersal and recolonization), and
as such they can be assessed or studied only in a
metapopulation context.
Impacts such as hunting or fishing may re-
duce reproduction or survival of individuals in
particular populations. For example, hunting
pressure or fishing mortality may differ be-
tween neighbouring populations, and failure
to incorporate the variation into quotas may
result in overexploitation, even if the regional
harvest is set at a conservative (precautionary)
level (Smedbol & Stephenson 2004). An overall
harvest level set for a metapopulation may
even lead to a series of local extinctions (or a
serial collapse of stocks), if most hunters (fish-
ermen) focus on the same few populations with
easiest access. After these are locally extinct,
the focus shifts to remaining populations with
the easiest access. Thus, many local extinctions
can occur serially, although the overall (re-
gional) harvest quota is precautionary and is
never exceeded. Dynamics of these sorts may
have contributed to the collapse of the New-
foundland cod fishery in 1992 with the loss of
40,000 jobs and no recovery in sight.
Conservation and management
in a metapopulation context
Conservation options for species that exist in
metapopulations include those that aim to
increase the size or persistence of individual
populations, as well as those that aim to benefit
the metapopulation.
The conservation options at the single popu-
lation level include habitat protection or im-
provement, regulation of harvest, reduction of
predation and removal of exotic species. Even
these measures that target individual popula-
tions may need to be evaluated in a metapopu-
lation context, because the presence of other
populations may change the relative effective-
ness of alternative options. An example of this
is the effectiveness of reducing seed predation
for Grevillea caleyi, an endangered understory
shrub of Australian eucalypt forest. The few
remaining populations of this species are
found within a small area at the interface be-
tween urban development and remnant native
vegetation, and are threatened by habitat de-
struction, adverse fire regimes and very high
seed predation (Auld & Scott 1997). Seed pred-
ators include weevils in the canopy and native
mammals at the soil surface. Seed germination
is triggered by fires, which also kill existing
plants. Thus, the frequency and intensity of
fires are important components of the species’
ecology. A study focusing on a single small
population (Regan et al. 2003) concluded that
predation reduction improved the chances of
long-term persistence of small populations sub-
stantially. However, a metapopulation study
(Regan & Auld 2004) concluded that manage-
ment of fires is crucial for the long-term
persistence of G. caleyi populations, and that
predation management was rather ineffective
by itself. The reason for this difference is that
the number of seeds entering the seed bank
after predation is extremely low for a single
small population, and there is a substantial
risk that all seeds will be depleted in the seed
bank due to viability loss and germination.
Reducing predation rates for a small population
would therefore substantially reduce its risk of
extinction. For the metapopulation, however,
its seed bank is large enough to always contain
available seeds, and a reduction in predation
rates does not have a substantial effect on its
risk of extinction. At the metapopulation level
it is more important to ensure adequate seed
production, regular germination and plant
survival in years when there are no fire events
(Regan & Auld 2004). Thus, for the regional
persistence of G. caleyi fire management appears
to be a much more important strategy, a
conclusion that was not as apparent when
only a single population was considered, even
though both actions fire management and
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 67 6.5.2006 2:48am
predation control can target a single popula-
tion or all populations in the metapopulation.
The conservation options at a metapopula-
tion level include reserve design, reintroduc-
tion and translocation, dispersal corridors and
management actions geared to local population
dynamics (such as sources and sinks). We
discuss these below.
Reserve design is a complex topic that almost
always involves multiple species, as well as
social, political and economic constraints. Here
we focus on only one aspect: directing conser-
vation effort at a subset of the populations of a
target species, in order to maximize the chances
of its survival. This issue is informed by predic-
tions and observations of generally higher
extinction rates in smaller populations, and
lower probabilities of rescue by immigration in
more isolated patches (Hanski 1994). It was
originally phrased as the ‘SLOSS’ debate, i.e.
whether a single large or several small
(SLOSS) populations are better to protect the
species. Although simplistic, this formulation
captures the nub of the issue, and underlines
the relevance to conservation of spatial struc-
ture and metapopulation dynamics.
On the one hand, several small populations
may have a lower extinction risk than one large
one if the rate of dispersal is high enough and
the degree of spatial correlation of environ-
ments is low enough. This is because a single
large population will not benefit from uncorrel-
ated environmental fluctuations; if it becomes
extinct, it cannot be recolonized. For example,
an important reason for establishing the wild
dog reserves discussed in Box 5.1 was to provide
a hedge against the possibility of a catastrophic
event hitting the single large Kruger population.
On the other hand, compared with a large
population, each of the small populations will
be more vulnerable to extinction due to
demographic stochasticity, higher mortality of
Box 5.1 Reintroduction of wild dogs in South Africa
Most metapopulations are the regional-scale expression of responses by individuals to patchiness in their habitat.
Persistence at the regional level is enhanced if individuals can retain some ability to move across the matrix to
prevent local extinctions or to recolonize empty patches. Here we describe a particularly extreme example of a
metapopulation, in which the habitat patchiness is caused by fences, and individuals have lost all intrinsic capacity to
mix freely between populations. The persistence of the metapopulation relies entirely on human-induced transloca-
tions, and corridors take the form of transportation vehicles.
A programme was initiated in 1997 to establish a second South African population of the endangered wild dog
Lycaon pictus apart from the only viable one in the Kruger National Park (Mills et al. 1998). As the Kruger population
fluctuates around 300 (Creel et al. 2004) it was thought prudent to bolster the small number of dogs in South Africa
and provide a hedge against the uncertainty of a catastrophic event hitting the Kruger population. At present South
Africa has no other protected area large enough to contain a self-sustaining population of wild dogs, so the strategy
has been to introduce them into a number of small widely scattered reserves separated by hundreds of kilometres
and to manage the various subpopulations as a single metapopulation.
Preliminary modelling of this wild dog metapopulation suggested that periodic, managed gene flow through
translocations should be implemented to reduce inbreeding and the resultant risks of meta- and subpopulation
extinction. The model indicated that by using a frequency of exchange based on the natural reproductive life span of
wild dogs (approximately 5 years) inbreeding could be reduced by two-thirds and population persistence could be
assured (Mills et al. 1998).
The guiding principle in reserve selection was to look for areas that reasonably can be expected to sustain at least
one pack of 10 to 20 animals. The average home range size for a pack is 537 km
in Kruger National Park (Mills &
Gorman 1997), which comprises a similar savannah woodland habitat to the habitat available in most of the
potential reserves for reintroduction. The range of sizes of the five reserves i nto which wild dogs have so far been
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 68 6.5.2006 2:48am
dispersers and edge effects (smaller patches
have a higher proportion of ‘edge’ to ‘core’
habitat). Thus, if they become extinct at the
same time, or if the extinct ones cannot be
recolonized from others, a metapopulation of
several small populations may have a higher
extinction risk than a single large population
(see Akc¸akaya et al. (1999) for an example).
In some cases, however, the choices are limited.
In the wild dog case, for example, available
habitat limited the size of the established popu-
lations to a maximum of three packs each,
resulting in a mixture of several small popula-
tions and one large (Kruger) population.
There is no general answer to the SLOSS
question. The answer depends not only on
the degree of correlation and chances for
recolonization, but also on other aspects of
metapopulation dynamics, such as the config-
uration, size and number of populations, their
introduced for the metapopulation is 370---960 km
. All reserves are enclosed with electrical f ences, to protect the
wild dogs and to minimize conflict with livestock farmers. Fences act as important barri ers to the m ovements of
the dogs, so that there is little emigration and even less immigration. The res erves are isolated from each ot her,
with no possibility at present to establish corrido rs, and almost all movement of wild dogs betwee n the reserves is
con duc ted through artificial i ntroductions and removals.
Apart from protecting the regional viability of the species, an important objective in the wild dog metapopulation
management programme is to promote biodiversity conservation. Biodiversity is a broad concept incorporating
compositional, structural and functional attributes at four levels of ecosystem organization: landscapes, communities,
species and genes (Noss 1990). A biodiversity objective for wild dogs that may be especially difficult to achieve in a
small reserve is to restore their ecological role as predator. Wild dog packs can produce large litters and more than
double in size within a year, posing a particularly challenging situation for managers because of the rapidly escalating
predation pressure, at least in the short term. This is exacerbated by the tendency for wild dogs to use fences as an aid
to hunting (van Dyk & Slotow 2003), which may artificially increase kill rate. An important aspect of the programme is
to research the viability of interactions between wild dogs and their prey in confined areas.
Following release of the first six to eight animals, the principle management strategy has been to continue to simulate
the natural dynamics of wild dog packs by moving single sex groups between reserves as and when necessary, so as to
maintain the genetic integrity of the metapopulation and, if necessary, to promote new pack formation as originally
recommended (Mills et al. 1998). In the reserves, regular maintenance and daily patrolling of the fences is essential. In
spite of this weaknesses do occur. Holes dug by other species such as warthogs (Phacochoerus africanus), flood damage
along drainage lines and occasions when predators chase prey through a fence are among the ways in which breaches
can occur. These are most likely to be exploited during dispersal events by the dispersing animals. Escapes are most
likely to happen if there are no suitable dogs of opposite sex available with which dispersers can form a new pack, or if
the reserve is too small to allow for the formation of another pack. The obvious solution to dispersers escaping from a
reserve is to remove dogs before they break away, but it is difficult to know which dogs to remove and when. The
preferred solution would be to remove dogs only after they have naturally split off from the pack. Managers decide on
the removal of dogs when they are concerned about the impact of increasing numbers on the prey, or in order to
decrease the risk of dogs escaping from a reserve. Behavioural observations may help to predict when a breakaway is
about to occur and which dogs are involved, in which case management intervention can thus be applied pre-
emptively based on this behavioural research.
Financial costs of the wild dog management programme have as much influence on strategy as do ecological
imperatives. Costs include upgrading reserve fences, constructing a holding facility, radio-telemetric apparatus for
monitoring the dogs, running vehicles, veterinary costs of capture, vaccination and transportation of the dogs, and
liability insurance against escaped dogs causing damage to neighbours’ domestic animals. Almost $380,000 was spent
on wild dog conservation in South Africa between 1997 and 2001, of which c.75% was spent on establishing the
metapopulation (Lindsey et al. 2005).
Despite the complexities outlined above, the extremely artificial nature of this metapopulation’s spatial structure,
and a general lack of knowledge about the dynamics of this species in small reserves, several aspects of this case are
closely related to the metapopulation issues we will discuss in this essay.
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 69 6.5.2006 2:48am
rates of growth, density dependence, carrying
capacities, etc.
Often the monetary or political cost of
acquiring a patch for a reserve might not be
related to its size; in other cases the size or carry-
ing capacity of a patch might not be directly
related to its value in terms of the protection it
offers. A small patch that supports a stable popu-
lation might contribute more to the persistence
of the species than would a large patch that is
subject to greater environmental variation or
human disturbances. Each case requires indi-
vidual evaluation, using all of the available em-
pirical information to evaluate as many as
possible of the potential impacts on the extinc-
tion time of the metapopulation. Predictions for
individual cases, however, will always depend
on a thorough understanding of the underlying
dynamic processes of density dependence and
interactions with the physical environment that
drive the case-specific mechanisms (Doncaster
& Gustafsson 1999). Although few of these
processes can be observed directly in nature,
the wider framework in which they operate is
provided by generic models of the conceptual
Wherever possible, design options should
consider less extreme alternatives than SLOSS.
A mixture of smaller and larger populations can
hedge against uncertainty in the scale of future
impacts, and it has potential genetic benefits.
Unless the small populations act as sinks, they
are likely to send out a greater proportion of
emigrants as well as receiving more immi-
grants, than larger populations. For example,
collared flycatchers (Ficedula albicollis) exhibit
this higher turnover in smaller populations,
which both reduces genetic drift and slows the
evolution of adaptations to local conditions
(Doncaster et al. 1997). The combination of
small (habitat generalist) and large (habitat
specialist) populations pre-adapts the metapo-
pulation for future environmental changes.
A related question for reserve design con-
cerns the optimum distribution of resources
between patches. Is the species better protected
by a more heterogeneous or more homoge-
neous distribution of resources? Temporal vari-
ability tends to stress populations near to
extinction thresholds, so reducing their sizes
(Hastings 2003). In contrast, spatial heterogen-
eity is likely to improve the predicament of
such species across both population and meta-
population scales (Doncaster 2001). This effect
arises because the abundance of rare consumers
generally decreases disproportionately with
degrading habitat quality, regardless of their
particular functional response to limiting re-
sources. For example, oystercatchers (Haemato-
pus ostralegus) will abandon beds of mussels
(Mytilus edulis) below a certain threshold of
available shellfish set by their foraging efficiency
(Caldow et al. 1999). The counter-intuitive
implication for metapopulations is that the
regional abundance of a target species can be
raised by redistributing resources between
patches even without any overall improvement
to habitat quality, so that those of intrinsically
higher quality are augmented to the detriment
of others already below the giving-up density.
Establishment of new populations through
translocation and reintroduction actions re-
quires many decisions: how often; how many
individuals, of which age classes or sexes; from
which population, to which existing population
or formerly occupied habitat patch? Each deci-
sion is potentially a trade-off, because it may
benefit one population while decreasing the
size of another one. Metapopulation models
can address these questions by finding strat-
egies that maximize the overall viability of the
metapopulation. This was especially important
in the wild dog reintroduction case (Box 5.1),
because almost all movement of wild dogs be-
tween the reserves is conducted through artifi-
cial introductions and removals. In this case, a
metapopulation model with genetic structure
would have helped to plan translocations in
such a way as to reduce inbreeding and main-
tain population structure, but in the event a
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 70 6.5.2006 2:48am
more needs-driven approach had to be taken in
terms of supply and demand of suitable dogs,
although always keeping in mind the genetic
history of the individuals concerned.
Metapopulation models may be particularly
important tools in decisions related to trans-
location and reintroduction of endangered
species, because the status of these species dis-
courages experimentation and makes a trial-
and-error approach less desirable. Using a
metapopulation model, McCarthy et al. (2004)
assessed various options for establishing a new
population of helmeted honeyeaters (Lichenos-
tomus melanops cassidix) from a captive popula-
tion. This bird is endemic to remnant riparian
forests in southern Victoria, Australia. Exten-
sive habitat destruction in the nineteenth cen-
tury led to a dramatic decline, and by 1990 the
only remaining population included 15–16
breeding pairs. As part of a recovery pro-
gramme initiated in 1989, a captive colony
was established to support the wild population
and to establish populations in new areas
(Smales et al. 2000). Because of uncertainties
about the fates of individuals and the difficulty
of integrating the available information from
numerous different sources, the optimal release
strategy is not immediately apparent. McCarthy
et al. (2004) ran simulations to determine how
the rate of release from the captive population
affects the probability of success of the reintro-
duction over 20 years. The optimal strategy was
to release individuals only when the captive
population contained at least four adult males,
and then to release 30% of the stock. The simu-
lations suggested that the chance of success of
the proposed reintroduction was moderately
good, with little chance that the new popula-
tion will have fewer than 10 males after 20
years (McCarthy et al. 2004). Although there
were several factors that could not be modelled
explicitly (e.g. whether the released birds
would remain where they are released, would
establish the same population behaviour, and
would have the same vital rates as the current
wild population), the modelling exercise pro-
vided valuable information that could not
have been obtained in any other way for this
extremely rare species.
In addition to human-mediated dispersal
through reintroduction and translocation, dis-
persal can be increased by conservation or res-
toration of the habitat lying between existing
populations, sometimes called the ‘landscape
matrix’. Matrix restoration can reduce local ex-
tinctions by facilitating the ‘rescue effect’ of
colonization, and it can increase the rate of
recolonization following local extinction. One
implementation of these efforts to increase the
overall persistence of the species is the building
or maintenance of habitat corridors. To answer
the question ‘Are corridors useful conservation
tools?’, we need to answer several subquestions
that are intimately bound to metapopulation
1. Are the habitat corridors used by the target
species? Use of a corridor depends not only
on its habitat, but also its shape, particularly
the width and length. For example, of the
mammalian predators native to California,
more species use creeks with wide margins
of natural vegetation as corridors than use
creeks with narrow or denuded margins
(Hilty & Merenlender 2004). European
hedgehogs (Erinaceus europaeus) dispersing
across arable habitat use road verges as
corridors, particularly on long-distance dis-
persals of as much as 10 km (Doncaster et al.
2. If used, do the corridors increase dispersal
rate? Perhaps individuals using the corridor
would have dispersed anyway; corridors are
more likely to affect dispersal rate where
dispersal is otherwise limited. For example,
if it were possible to build corridors between
the widely scattered wild dog reserves dis-
cussed in Box 5.1, the lack of natural con-
nections and the pack-forming behaviour of
the species suggest that such corridors
would have increased dispersal between re-
serves. Corridors are likely to benefit fast-
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 71 6.5.2006 2:48am
reproducing species in the short term and
slower reproducers in the long term, so their
value depends on the time scale of conser-
vation goals (Hudgens & Haddad 2003).
Experimental fragmentation of moss banks
has demonstrated rescue effects of artificial
corridors for moss-living micro-arthropods.
Figure 5.2 shows how corridors between
moss fragments arrested declines in the
abundance of most species (Gonzalez et al.
1998). It is worth noting that this experiment
on an abundant fauna cost little to run, yet
has provided invaluable quantification of the
positive relation between abundance and
distribution in connected landscapes, and
of the breakdown of this relation in the ab-
sence of corridors. Conservation needs more
such field tests of metapopulation theory by
3. Does an increased dispersal rate increase the
overall viability of the metapopulation?
Usually it does, by rescuing local popula-
tions from potential extinction. Colonists
can also bring hybrid vigour to isolated
populations that suffer from inbreeding
depression (e.g. of Daphnia: Ebert et al.
2002). However, increased connectivity
may also have ‘anti-rescue effects’ (Harding
& McNamara 2002), with documented ex-
amples due to the spread of infectious dis-
eases, or parasites or predators (Hess 1996;
Grenfell & Harwood 1997), or gene flow re-
ducing local adaptation (Hastings & Harrison
1994; Harrison & Hastings 1996). High dis-
persal can increase impacts of catastrophes
(Akc¸akaya & Baur 1996), and losses to sink
habitats. In other cases, the effectiveness of
dispersal in reducing extinction risks de-
pends on the correlation of environmental
fluctuations experienced by different popu-
lations. If the correlation is high, all popula-
tions decline simultaneously, reducing
recolonization rates of empty patches. For
example, if a major climatic shift caused a
region-wide decline in the prey base of the
wild dogs discussed in Box 5.1, all popula-
tions would decline or become extinct and it
would be difficult to recolonize them even
with a well-planned translocation pro-
gramme. If, on the other hand, the fluctu-
ations are at least partially independent,
some patches can act as sources of emigrants
(Burgman et al. 1993). Extinction risks are
often sensitive to spatial correlation in envir-
onmental fluctuations and the pattern of dis-
turbance, as demonstrated by models for a
variety of species, including the mountain
gorilla (Gorilla beringei beringei; Akc¸akaya &
Ginzburg 1991), spotted owl (Strix occidenta-
lis; LaHaye et al. 1994) and Leadbeater’s pos-
sum (Gymnobelideus leadbeateri; McCarthy &
Lindenmayer 2000).
4. Do corridors have any other effects on meta-
population viability? Negative impacts may
include increased mortality due to predation
along the corridor. All Dutch highways
constructed since 1990 have underpasses for
Pseudo corridor
Pseudo corridor
Initial abundance of each
50 cm
50 cm
79 cm
7 cm long
Fig. 5.2 Experimental fragmentation of moss banks
into small patches reduces the abundance of
micro-arthropods, but most species are saved from
substantial decline by corridors connecting the frag-
ments. (From Gonzalez et al. 1998. Reprinted with
permission. Copyright 1998 AAAS.)
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 72 6.5.2006 2:48am
European badgers (Meles meles), and these are
also used by other wildlife, including hedge-
hogs. This benefits the hedgehogs because
they too are a frequent casualty of roads,
but the benefit is undone because they are
also a favoured food of badgers, into whose
jaws they are channelled by the underpasses
(Bekker & Kanters 1997). Costs such as these
need to be weighed against the benefits of
dispersal. The Dutch Government spends
US$5 million each year on tunnels and
fences for wildlife along its highways (Teo-
dorascu 1997), so it makes economic sense to
evaluate the combined effect of all changes
on metapopulation viability.
5. Are there cheaper alternatives to corridors?
These may involve seeding new habitable
patches, or improving existing populations
by augmenting net growth rates or carrying
capacities. In simulations of metapopula-
tions prone to local extinction events, the
viability of the system is found to benefit
more from reducing local extinction prob-
abilities, particularly on patches with the
lowest probabilities, than from increasing
colonization probabilities (Etienne & Hees-
terbeek 2001). Birds using rainforest frag-
ments show evidence of this response
(Lens et al. 2002), but more empirical test-
ing is needed of this model, as with most
metapopulation models.
Although the importance of corridors has
long been recognized, it is only with the use
of metapopulation models that their advan-
tages can be quantified, for example, in terms
of increased persistence or viability of the spe-
cies, and compared with advantages of alterna-
tive strategies.
When conservation is geared to local populations
with dynamics of sources and sinks, manage-
ment options must consider many interdepend-
ent factors. Two general issues arise:
1. How do source–sink dynamics affect meta-
populations? The overall effect on metapo-
pulation persistence of dispersal from
sources to sinks depends on the cost to
source population (increased risk of local
extinction), the benefit to sink population
(decreased risk of local extinction), and the
changes with local population density in
dispersal, survival or reproduction. In the
presence of density dependence, the excess
of deaths to births that is characteristic of a
sink can be caused directly by the influx of
immigrants rather than being an inherent
property of the patch. The viability of such
a ‘pseudosink’ consequently need not de-
pend on the arrival of emigrants from
sources. It may even benefit from a reduced
influx, in contrast to a true sink which is
rescued by immigration (Watkinson &
Sutherland 1995). Management options
will differ for true and pseudosinks because
of this, yet the two types can be hard to
distinguish in field surveys. For example,
sources and pseudosinks in the highly frag-
mented Taita Hills forests of Kenya could be
identified only from a combination of demo-
graphic, genetic and behavioural work
(Githiru & Lens 2004). To sidestep these
complications, sinks can be defined as popu-
lations whose removal would increase the
overall viability of the metapopulation. This
approach, however, requires modelling of
the underlying dynamics of the metapopu-
lation, and therefore more data.
2. Should sink populations be protected? This
depends on various factors, the most im-
portant being what is meant by ‘protected’
and its alternatives. If ‘protected’ means that
fecundity or survival may increase to the
extent that the sink population can become
self-maintaining (i.e. have a low risk of ex-
tinction even in the absence of dispersal
from other populations), and the alternative
is continuation as a sink population, then
protection is probably justified (as Breinin-
ger & Carter (2003) demonstrated for the
Florida scrub jay (Aphelocoma coerulescens)).
If ‘protected’ means it is maintained as a
sink population and the alternative is that
individuals that would have dispersed to the
sink end up in a habitat patch with higher
survival or fecundity, then protection of the
sink is probably not justified (as Gundersen
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 73 6.5.2006 2:48am
et al. (2001) demonstrated for root voles
(Microtus oeconomus) ). In the wild dog case
(Box 5.1), for example, if mortality exceeds
reproduction in one of the reserves as a
result of a local decline in the prey base,
then it would be justified to attempt to
maintain this population by increasing the
prey base in that reserve (e.g. by enlarging
the reserve or by habitat improvement), but
it might not be justified to attempt to main-
tain this population only by increasing
translocations from healthier populations.
Other considerations include whether the
sink population can increase connectivity
(as a ‘stepping stone’ between other popu-
lations), its contribution to total abundance
and its function as a buffer against cata-
strophic events. Where conservation is
aimed at culling an invasive alien, its re-
gional decline can be hastened by allocating
culls to sinks as well as sources (as for the
European hedgehog introduced into the
Scottish Western Isles; Travis & Park 2004).
The point is that there are a lot of details,
and generalizations are difficult if not im-
possible. The only way to address such ques-
tions is to develop case-specific models that
incorporate all that is known about the dy-
namics of the metapopulation, including
survival, fecundity and dispersal for all
populations whether source or sink, as well
as temporal variability and density-depend-
ence in these parameters.
Do metapopulations need models?
The metapopulation concept lends itself to
modelling because its core dynamic of popu-
lations colonizing patches (and their potential
local extinction) bridges models of persistence
at the levels of the individual and the commu-
nity: of individuals consuming resources (and
their eventual death), and of species colonizing
niches (and their potential extinction: Doncas-
ter 2000). Metapopulations encompass land-
scape-level processes of patches being formed,
split and merged in habitat successions and
disturbance events. At all of these scales, models
are used to pare away as much of the complexity
inherent to nature as is necessary to reveal the
underlying patterns and to explore the range of
forces that shape these patterns. Models are par-
ticularly important to the conservation of meta-
populations, because the regional focus and
undesirability of experimental manipulations
usually rules out any other methods of distin-
guishing causes of endangerment from second-
ary effects. Most of the issues and decisions
regarding metapopulations concern interde-
pendent factors, such as number of populations,
spatial correlation, dispersal and density de-
pendence. Because many of these factors in-
volve interactions between populations, there
is no simple way of combining models on dy-
namics of individual populations into regional-
scale decisions. The only way to incorporate all
these factors is to simultaneously include all
populations and their interactions in one
model, in other words, to use a metapopulation
model. Models are particularly valuable tools in
cases where the endangered status of species
makes other (e.g. experimental) approaches
difficult or impossible.
Models are also useful in evaluating manage-
ment actions at large spatial scales, at which
experiments may not be feasible. Frequently,
management of a metapopulation means man-
agement of the species’ habitat. Habitat man-
agement may take many forms, including
controlling the rate and pattern of habitat alter-
ation through the effects of grazers or harvest
by humans. For example, Schtickzelle &
Baguette (2004) used a structured metapopula-
tion model to study the effect of grazing on the
bog fritillary butterfly (Proclossiana eunomia)in
south-eastern Belgium. This species has a very
restricted habitat; both larval and adult stages
feed on a single plant species that occurs
mainly in wet hay meadows along rivers of
some uplands scattered in western Europe.
Grazing by large herbivores is sometimes used
by conservation agencies to maintain early suc-
cessional stages in wet hay meadows. The
metapopulation model demonstrated that graz-
ing substantially increases the extinction risk
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 74 6.5.2006 2:48am
for the bog fritillary butterfly in south-eastern
Belgium. Its predictions led to modifications in
the management protocol of a nature reserve:
several grazing regimes are being tested and
half the area is now kept ungrazed.
Controlled timber harvest is another form of
habitat management. Regan & Bonham (2004)
developed a metapopulation model of the car-
nivorous land snail Tasmaphena lamproides
inhabiting native forests in northwest Tas-
mania. This species is listed as threatened due
to its small range, much of which is within
timber production forest. The model was
designed as a decision support tool for man-
agers to explore the trade-offs between timber
production requirements and conservation of
the species under various management scen-
arios. Future use of the area involves convert-
ing native forest to eucalypt plantations, or
harvesting native forest followed by burning
to promote regeneration. Burning is thought
to eliminate populations of this snail, but they
reinvade native forest areas once the required
habitat has formed with adequate level of litter
and food sources. The metapopulation model
combines geographical information system
(GIS) data on the distribution of forests and
demographic data on the dynamics of the spe-
cies, and allows the investigation of alternative
harvesting strategies which meet wood produc-
tion needs in the long term but minimize popu-
lation declines in the short term.
In aquatic systems, habitat management
often involves water regimes and barriers such
as dams. Changes in water regime were impli-
cated in the severe decline of the European
mudminnow (Umbra krameri) along the River
Danube during the second half of the twentieth
century (Wanzenbo
ck 2004). Water regulation
in the Danube has increased flow velocity and
caused the river to cut a deeper channel, low-
ering the groundwater level in the surrounding
floodplain. As a result, the original side channel
used by the mudminnow has been transformed
into a chain of disconnected, groundwater-fed
ponds. A simple metapopulation model was
used to demonstrate that reversing this declin-
ing trend in habitat capacity is critical to the
mudminnow’s persistence, and to recommend
increasing the habitat availability for, and con-
nectivity of, populations. To implement these
recommendations, groundwater levels are
being raised by opening some of the longitu-
dinal dams bordering the main river and recon-
necting some backwaters to the river. These
conservation efforts began in the late 1990s,
are continuing today, and their impact on the
mudminnow is being monitored closely.
There are several different types of metapo-
pulation models, each with their own set of
assumptions and restrictions (detailed in Akc¸a-
kaya & Sjogren-Gulve 2000; Breininger et al.
2002). Patch-occupancy models have the sim-
plest demographic structure, describing each
population as present or absent (e.g. within
regional distributions of butterflies or other
winged insects; Hanski 1994). Intermediate
complexity is found in structured (or, fre-
quency-based) models that describe each popu-
lation in terms of the abundances of age classes
or life-history stages (Akc¸akaya 2000a). These
models incorporate spatial dynamics by model-
ling dispersal and temporal correlation among
populations (e.g. of the land snail Arianta arbus-
torum; Akc¸akaya & Baur 1996). At the other
extreme are individual-based models, which
describe spatial structure within the location
of territories, or of each individual in the popu-
lation (e.g. of northern spotted owls (Strix occi-
dentalis caurina); Lamberson et al. 1996; Lacy
2000). Some models use a regular grid where
each cell can be modelled as a potential terri-
tory. For example, Pulliam et al. (1992) used
this approach in a region managed for timber
production to show that population sizes of
Batchman’s sparrow (Aimophila aestivalis)
depended more strongly on mortality rates
than on dispersal ability. Another approach
uses a habitat suitability map to determine the
spatial structure of the metapopulation (e.g.
of the helmeted honeyeater (Akc¸akaya et al.
1995) and California gnatcatcher (Akc¸akaya &
Atwood 1997)). All of these approaches have
been applied to specific conservation manage-
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 75 6.5.2006 2:48am
ment questions (Chapter 9). The appropriate
choice depends on the complexity of the prob-
lem at hand, the assumptions of the model in
relation to the ecology of the species (see
below) and the data available.
Current limitations and dilemmas
Single-species focus
Most metapopulation applications focus on a
single species, yet much of conservation man-
agement concerns communities. Even where a
single species is targeted for conservation, its
survival and fecundity will often depend on
competition within the trophic level or preda-
tion from higher trophic levels. For example, in
the case of wild dogs (Box 5.1), an important
objective is to restore their ecological role as
predator, which requires research into the via-
bility of wild dog–prey interactions in confined
areas. Metapopulation models tend to focus on
single-species dynamics because these are better
understood than foodweb and ecosystem pro-
cesses. Adding an extra species to the system
requires at least two extra dimensions in the
analysis (to account for both exclusive and
shared occupancy of suitable habitat), greatly
increasing the number of parameters for estima-
tion and thus model error. The general lack of
understanding and data on multispecies inter-
actions means that few empirical metapopula-
tion studies have sufficient parameter estimates
to model community dynamics. An astute use of
simplifying assumptions, however, can bring
theory within the grasp of empirical data.
Simple models have achieved some robust
predictions for competitive coexistence by re-
ducing the representation of competition to a
binary distinction between competitively dom-
inant and inferior (fugitive) species. For ex-
ample, habitat destruction is predicted to
disadvantage dominant species with slow dis-
persal to the benefit of fugitive species, and
the early loss of dominants has most effect on
community structure because of their potential
role as keystone species (Tilman et al. 1997).
The dominant–fugitive dichotomy applies par-
ticularly to plant diversity in prairie grasslands.
The generalized version of this patch-
occupancy approach explores the full range of
competitive asymmetries in regional coexist-
ence, and without needing extra dimensions
in the analysis if it can be assumed that both
residents and colonists experience similar
effects of density on survival (Doncaster et al.
2003). This model reveals that subdominant
species with poor dispersal are the most sensi-
tive to habitat degradation. Their loss from the
community provides a useful early warning of
regional disturbance and degradation, because
it will have less impact on community structure
than the subsequent disappearance of domin-
ant and potentially keystone species. In gen-
eral, faster reproducing communities (e.g.
invertebrate assemblages) are both predicted
and observed to have higher tolerance for dif-
ferences in growth capacity, compared with
slower reproducing communities (e.g. forest
trees), which have higher tolerance for com-
petitive interactions. Coexistence is even pos-
sible amongst tree species competing for
identical resources in the same metapopula-
tion, if they differ in their threshold conditions
for switching from vegetative growth to seed
production (e.g. Mexican rain forest trees;
Kelly & Bowler 2002). These low productivity
communities tend to be the most at risk from
human induced disturbance, and therefore the
most in need of predictive models.
Where conservation efforts are directed to-
wards a community of species, a practical ap-
proach to dealing with the single-species
limitation is to select a target species that is
representative of the natural community, that
is sensitive to potential human impact and
whose conservation will protect other species
(Noon et al. 1997). One danger here is that the
target species and others may have different
networks of habitat patches in the same region.
For example, from a large-bodied predator’s
point of view, there may be a few large habitat
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 76 6.5.2006 2:48am
patches, but for its small prey, there may be
hundreds of distinct patches. Or, the degree
of fragmentation may be different for each spe-
cies depending on their habitat requirements.
For example, roads fragment forest habitat
for song-birds in direct proportion to their
dependence on canopy-level vegetation for
nesting and feeding (St Clair 2003). Endangered
silver-studded blue butterflies (Plebejus argus)
and sand lizards (Lacerta agilis) both disperse
between heathland fragments, but the greater
capacity for the butterflies to use areas between
habitat patches (also called ‘matrix’) suggests
they will benefit most from climate warming,
at least in terms of increased patch connectivity
and metapopulation stability (Thomas et al.
1999). The best strategy in such habitat and
community studies is often to combine results
from different target species (Root et al. 2003).
For addressing most conservation questions
involving species in fragmented landscapes,
metapopulation models often have less severe
limitations than the available alternatives, such
as rule-based methods, expert opinion, reserve-
selection algorithms and habitat mapping.
However, these alternatives have the potential
of contributing to the realism of metapopula-
tion models or of complementing them (Akc¸a-
kaya & Sjogren-Gulve 2000; Breininger et al.
2002; Brook et al. 2002).
Definition and delineation
of populations in a metapopulation
Most metapopulation approaches represent the
landscape by discrete habitable patches within
a surrounding matrix that may allow dispersal
but does not support populations. To the extent
that there exist areas where a species can re-
produce and those where it cannot, this as-
sumption is not unrealistic. However, it does
require the definition of a population, and a
method for identifying these areas (patches) in
a given landscape.
A general definition of a population presents
dilemmas, regardless of the metapopulation
context. Considering the difficulty of defining a
species, a much more fundamental concept, this
is perhaps not surprising. A biological popula-
tion can be defined as a group of interbreeding
(i.e. panmictic) individuals. Assuming that the
distribution of a species is more-or-less continu-
ous across parts of the landscape, the question of
delineating a population can be rephrased as:
how far apart must two individuals be in order
to be considered to be in different populations?
This depends on the movement distance, home
range, or some other measure related to the
possibility of interbreeding. This approach, com-
bined with modelling and prediction of suitable
habitat, is used in habitat-based metapopulation
models to delineate populations (Akc¸akaya
2000b, 2005). In the wild dog metapopulation
(Box 5.1), populations are easily defined by
fenced reserves.
Assumptions of metapopulation models
All models assume certain constants, in order to
interpret the dynamics of interest. The useful-
ness of any model therefore depends on the
validity of its assumptions. Below we discuss
recent approaches to improving the fit of meta-
population models to data.
Some metapopulation models assume that the
metapopulation is in equilibrium with respect
to the extinction and recolonization of patches
(e.g. incidence function type of patch-occu-
pancy models; Hanski 1999). There is little evi-
dence to suggest that metapopulations of any
species are in fact at equilibrium (Baguette
2004), and small, highly variable metapopula-
tions are particularly unlikely to be so. How-
ever, metapopulations that persist over long
time-scales must be under some form of density
regulation at the regional scale, often assumed
to be in colonization rate, which implies at least
a deterministic attraction towards an equilib-
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 77 6.5.2006 2:48am
rium density of occupied habitat. Equilibrium
models therefore can play a useful role as null
hypotheses for analysing the processes that
may threaten viability, such as habitat loss, ex-
ploitation and alien invasions. A poor fit of
equilibrium models to the data can signal the
need to account for other factors, such as com-
petitive interference in addition to exploitation
(Doncaster 1999), or multiple equilibria
(Hanski et al. 1995), or it may result from ran-
dom fluctuations. Null models test these alter-
natives parsimoniously by seeking to explain
deviations from equilibrium predictions. Note
that density regulation of local populations in a
metapopulation does not guarantee the exist-
ence of an equilibrium at the metapopulation
level. The metapopulation may still decline if
the rate of local extinctions due to environmen-
tal fluctuations and demographic stochasticity
exceeds colonization rates, because of factors
such as limited dispersal, or Allee effects on
small populations, or correlated environments.
The equilibrium assumption is sometimes
mistakenly believed to apply to the metapopu-
lation concept in general, yet several meta-
population models and approaches do not
make this assumption (e.g. structured and indi-
vidual-based models, described in Akc¸akaya &
Sjogren-Gulve 2000). These are particularly use-
ful for predicting the extinction probability of
small metapopulations which may have unbal-
anced sex ratios or age structures, or low genetic
variability, and which are most prone to envir-
onmental fluctuations (e.g. some coral reef
fishes: Bascompte et al. 2002), or which may be
declining (e.g. California gnatcatcher (Polioptila
californica); Akc¸akaya & Atwood 1997).
Some metapopulation models assume that the
dynamics of local populations are independent
of each other. However, this assumption is vio-
lated in many metapopulations where local
populations are affected by regional environ-
mental factors that impose a correlation. For
example, fecundities of the California least
tern (Sterna antillarum browni) are correlated
across different subpopulations, presumably
due to the effects of large-scale weather pat-
terns such as the El Nin
o–Southern Oscillation
that may simultaneously affect the food re-
sources of many populations. The correlation
coefficients average 0.32 (range 0–0.6), and de-
cline with increasing distance between the
populations (Akc¸akaya et al. 2003a). When
correlations are based on population sizes ra-
ther than vital rates such as fecundity, it may be
difficult to untangle the relative contributions
of correlated environmental factors, dispersal,
and trophic interactions to the observed spatial
correlation in population dynamics (Ranta et al.
1999). However, it is clear that for many spe-
cies, subpopulations experience spatially cor-
related dynamics (Leibhold et al. 2004). In
these cases, results of simple models that as-
sume independence may be misleadingly opti-
mistic in their estimation of risks of extinction
and decline. However, it is possible to make
realistic and unbiased assessments by using
models that incorporate dependencies or spatial
correlations among populations (e.g. Harrison
& Quinn 1989; Akc¸akaya & Ginzburg 1991;
LaHaye et al. 1994).
Many metapopulation models assume a con-
stant number and location of habitable patches,
yet natural landscapes are inherently dynamic.
Spatial structure changes according to seasons,
climatic fluctuations and succession, as well as
human impacts (urban sprawl, global climate
change, agricultural expansion, etc.). The via-
bility of a metapopulation will depend on its
rate of patch turnover, as well as the static
quantity and quality of suitable habitat (Key-
mer et al. 2000). Under habitat succession
or age-dependent disturbance, for example,
a metapopulation is predicted to persist for as
long as the mean age of its constituent patches
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 78 6.5.2006 2:48am
exceeds the average interval between coloniza-
tion events (Hastings 2003). A metapopulation
with a slow turnover of patches thus may per-
sist even with a high extinction rate of local
populations, and managers should be wary of
underestimating its viability. Equally, manage-
ment action aimed at increasing the lifespan of
patches is likely to do more good than action
focused directly on the survival of local popu-
lations on the patches.
Some metapopulation models incorporate
community succession, which tends to be par-
ticularly patchy in time and space at its early
stages and determines critical habitat for certain
species (Johnson 2000; Hastings 2003). Other
models incorporate changes in carrying cap-
acity over time, either deterministically, for
example to simulate forest growth, or stochas-
tically to simulate the effects of random
disturbances such as fires, or both, e.g. as a
deterministic function of time since a stochastic
disturbance event (Pulliam et al. 1992; Linden-
mayer & Possingham 1996; Stelter et al. 1997;
Akc¸akaya & Raphael 1998; Johst et al. 2002;
Keith 2004). A recently developed approach
addresses these issues by linking a landscape
model and a metapopulation model (Akc¸akaya
et al. 2003b, 2004, 2005).
An example of incorporating habitat change
in metapopulation dynamics involves the
woodland brown butterfly (Lopinga achine),
which lays its eggs at the edges of glades of
the partly open oak woodland pastures where
its host plant Carex montana grows. The habitat
quality for this species is related to the amount
of bush and tree cover within the pastures and
the occurrence of its host plant (Bergman
1999). As discussed above, grazing often helps
maintain grassland habitats in successional
stages that favour certain species. As grazing
ceases, the essential habitat of this species
(open glades with host plants) becomes over-
grown and deteriorates. Using a metapopula-
tion model, Kindvall & Bergman (2004)
calculated long-term extinction risks under
various landscape scenarios. An important as-
pect of this analysis was that the landscape
scenarios were dynamic; thus, the study inte-
grated the changes in the habitat (as a result of
succession and grazing) with changes in the
metapopulation, and demonstrated the import-
ance of landscape dynamics in affecting the
viability of this species.
Metapopulation models have been essential to
the management of many species. The listing of
several species on the Endangered Species List
in the USA, as well as the management and
recovery plans for a number of species, were
based in part on the analysis of their metapo-
pulation dynamics. For example, the draft re-
covery plan for the Pacific coast population of
the western snowy plover (Charadrius alexandri-
nus nivosus) included a metapopulation model
(Nur et al. 1999), which highlighted the need
for increased management of the species and its
habitats. This population is listed as threatened
in the USA, because habitat degradation caused
by human disturbance, urban development,
introduced beachgrass (Ammophila spp.), and
expanding predator populations have resulted
in a decline in active nesting areas and in the
size of the breeding and wintering populations.
Using a metapopulation structure that allowed
estimates for demographic parameters to vary
among subpopulations was considered an
important aspect of this model. The metapopu-
lation model predicted a high probability of
decline under existing conditions, which in-
cluded intensive management in some areas
by area closures, predator exclosures and
predator control. The model suggested that re-
covery at a moderate rate would be possible
with a productivity of 1.2 or more fledglings
per breeding male, but would require short-
term intensive management and long-term
commitments to maintaining gains. Other
species for which metapopulation models have
been used in recovery planning or listing in-
clude northern spotted owl (Strix occidentalis
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 79 6.5.2006 2:48am
caurina), California spotted owl (Gambelia silus),
south-western willow flycatcher (Empidonax
traillii extimus), marbled murrelet (Brachyram-
phus marmoratus), Florida scrub jay (Aphelocoma
coerulescens) and Florida panther (Felis concolor
coryii). These cases, and several examples dis-
cussed throughout this essay, illustrate our an-
swers to the two questions posed earlier in the
essay: conservation needs metapopulation ap-
proaches, and metapopulations need models.
Many species live in naturally heterogeneous
or artificially fragmented landscapes, and de-
cisions on their conservation and management
should consider metapopulation concepts and
models. Models make assumptions, however,
many of which await evaluation and should
not be tested on our most treasured wildlife.
The metapopulation literature is full of caveats
to the effect that more empirical data are
needed to distinguish between alternative pro-
cesses and mechanisms. These data must come
from field experiments, yet too often field
ecologists are pulled towards the expediency
of mission-oriented conservation with the re-
sult that we still lack a well-tried framework for
managing endangerment at the regional scale.
The wild dog case (Box 5.1) illustrates how the
principal function of metapopulation models in
conservation to evaluate alternative options
and scenarios depends on there being alter-
natives to choose from. Metapopulation models
stimulated the original concept of linked
reserves, and contributed to addressing poten-
tial problems of inbreeding at the planning
stage (Mills et al. 1998). Options at the con-
struction stage were severely limited by the
small number of sites and animals available,
favouring a pragmatic approach of adaptive
management for this large social species with
complex behavioural ecology. Population mon-
itoring and autecological studies are now pro-
viding data for optimizing population sizes and
translocation rates. Models will thus become
increasingly important decision tools in the
long-term management of the metapopulation.
Despite these caveats and limitations, we
believe current conservation efforts for many
species would benefit from a more explicit and
quantitative consideration of metapopulation
There is nothing in this world constant, but inconstancy.
( Jonathan Swift (1667–1745), A Critical Essay upon the Faculties of the Mind.)
Akc¸akaya, H.R. (2000a) Population viability analyses
with demographically and spatially structured
models. Ecological Bulletins 48: 23–38.
Akc¸akaya, H.R. (2000b) Viability analyses with
habitat-based metapopulation models. Population
Ecology 42: 45–53.
Akc¸akaya, H.R. (2005) RAMAS GIS: Linking Spatial
Data with Population Viability Analysis, Version 45.0.
Applied Biomathematics, Setauket, New York.
Akc¸akaya, H.R. & Atwood, J.L. (1997) A habitat-
based metapopulation model of the California
Gnatcatcher. Conservation Biology 11: 422–34.
Akc¸akaya, H.R. & Baur, B. (1996) Effects of population
subdivision and catastrophes on the persistence of a
land snail metapopulation. Oecologia 105: 475–483.
Akc¸akaya, H.R. & Ginzburg, L.R. (1991) Ecological
risk analysis for single and multiple populations. In
Species Conservation: A Population-Biological Approach
(Eds A. Seitz and V. Loeschcke), pp. 73–87.
Birkhauser Verlag, Basel.
Akc¸akaya, H.R. & Raphael, M.G. (1998) Assessing
human impact despite uncertainty: viability of the
northern spotted owl metapopulation in the north-
western USA. Biodiversity and Conservation 7: 875–94.
Akc¸akaya H.R. & Sjo
gren-Gulve, P. (2000) Popula-
tion viability analysis in conservation planning: an
overview. Ecological Bulletins 48: 9–21.
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 80 6.5.2006 2:48am
Akc¸akaya, H.R., McCarthy, M.A. & Pearce, J. (1995)
Linking landscape data with population viability
analysis: management options for the helmeted
honeyeater. Biological Conservation 73: 169–76.
Akc¸akaya, H.R., Burgman, M.A. & Ginzburg, L.R.
(1999) Applied Population Ecology: Principles and
Computer Exercises using RAMAS EcoLab 2.0, 2nd
edn. Sinauer Associates, Sunderland, Massachu-
setts, 285 pp.
Akc¸akaya, H.R., Atwood, J.L., Breininger, D., Collins,
C.T. & Duncan, B. (2003a) Metapopulation dynam-
ics of the California least tern. Journal of Wildlife
Management 67(4): 829–42.
Akc¸akaya, H.R., Mladenoff, D.J. & He, H.S. (2003b)
RAMAS Landscape: integrating metapopulation
viability with LANDIS forest dynamics model.
User Manual for version 1.0. Applied Biomathe-
matics, Setauket, New York.
Akc¸akaya, H.R., Radeloff, V.C., Mladenoff, D.J. & He,
H.S. (2004) Integrating landscape and metapopu-
lation modeling approaches: viability of the sharp-
tailed grouse in a dynamic landscape. Conservation
Biology 18: 526–537.
Akc¸akaya, H.R., Franklin, J., Syphard, A.D. &
Stephenson, J.R. (2005)Viability of Bell’s sage
sparrow (Amphispiza belli ssp. belli): altered fire
regimes. Ecological Applications 15: 521–31.
Auld, T.D. & Scott, J. (1997) Conservation of endan-
gered plants in urban fire-prone habitats. Proceed-
ings: Fire Effects on Rare and Endangered Species
and Habitats Conference, Coeur D’Alene, Idaho,
pp. 163–71. International Association of Wildland
Fire, Hot Springs, South Dakota.
Baguette, M. (2004) The classical metapopulation
theory and the real, natural world: a critical
appraisal. Basic and Applied Ecology 5: 213–224.
Baguette, M. & Mennechez, G. (2004) Resource and
habitat patches, landscape ecology and metapopu-
lation biology: a consensual viewpoint. Oikos 106:
Bascompte, J., Possingham, H. & Roughgarden,
J. (2002) Patchy populations in stochastic envir-
onments: critical number of patches for persist-
ence. American Naturalist 159: 128–37.
Bekker, G.J. & Canters, K.J. (1997) The continuing
story of badgers and their tunnels: In Proceedings
Habitat Fragmentation and Infrastructure (Eds K.
Canters, A. Piepers & D. Hendriks-Heersma),
pp. 344–53. DWW Publications, Delft.
Bergman, K-O. (1999) Habitat utilization by Lopinga
achine (Nymphalidae: Satyrinae) larvae and
ovipositing females: implications for conservation.
Biological Conservation 88: 69–74.
Bleich, V.C., Wehausen, J.D. & Holl, S.A. (1990)
Desert-dwelling mountain sheep: conservation
implications of a naturally fragmented distribu-
tion. Conservation Biology 4: 383–90.
Breininger, D.R. & Carter, G.M. (2003) Territory
quality transitions and source–sink dynamics in a
Florida scrub-jay population. Ecological Applications
13: 516–29.
Breininger, D.R., Burgman, M.A., Akc¸akaya, H.R. &
O’ Connell, M.A. (2002) Use of metapopulation
models in conservation planning. In Applying Land-
scape Ecology in Biological Conservation (Ed. K. Gutz-
willer), pp. 405–27, Springer-Verlag, New York.
Brook, B.W., Burgman, M.A., Akc¸akaya, H.R.,
O’Grady, J.J. & Frankham, R. (2002) Critiques of
PVA ask the wrong questions: throwing the heur-
istic baby out with the numerical bathwater. Con-
servation Biology 16: 262–3.
Burgman, M.A., Ferson, S.& Akc¸akaya, H.R. (1993)
Risk Assessment in Conservation Biology. Chapman
and Hall, London, 314 pp.
Caldow, R.W.G., Goss-Custard, J.D., Stillman, R.A.,
Durell, S.E.A.L.D., Swinfen, R. & Bregnballe,
T. (1999) Individual variation in the competitive abil-
ity of interference-prone foragers: the relative import-
ance of foraging efficiency and susceptibility to
interference. Journal of Animal Ecology 68: 869–78.
Creel, S., Mills, M.G.L. & McNutt, J.W. (2004) Dem-
ography and population dynamics of African wild
dogs in three critical populations. In Biology and
Conservation of Wild Canids (Eds D.W. Macdonald
& C. Sillero-Zubiri), pp. 337–350. Oxford Univer-
sity Press, Oxford.
Dennis, R.L.H., Shreeve, T.G. & Van Dyck, H. (2003)
Towards a functional resource-based concept for habi-
tat: a butterfly biology viewpoint. Oikos, 102: 417–26.
Doncaster, C.P. (1999) A useful phenomenological
difference between exploitation and interference
in the distribution of ideal free predators. Journal of
Animal Ecology 68: 836–8.
Doncaster, C.P. (2000) Extension of ideal free re-
source use to breeding populations and metapopu-
lations. Oikos 89: 24–36.
Doncaster, C.P. (2001) Healthy wrinkles for popula-
tion dynamics: unevenly spread resources can
support more users. Journal of Animal Ecology
70: 91–100.
Doncaster, C.P. & Gustafsson, L. (1999) Density
dependence in resource exploitation: empirical
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 81 6.5.2006 2:48am
test of Levins’ metapopulation model. Ecology Letters
2: 44–51.
Doncaster, C.P., Clobert, J., Doligez, B., Gustafsson,
L. & Danchin, E. (1997) Balanced dispersal
between spatially varying local populations: an
alternative to the source-sink model. American
Naturalist 150 : 425–45.
Doncaster, C.P., Rondinini, C. & Johnson, P.C.D.
(2001) Field test for environmental correlates of
dispersal in hedgehogs Erinaceus europaeus. Journal
of Animal Ecology 70: 33–46.
Doncaster, C.P., Pound, G.E. & Cox, S.J. (2003)
Dynamics of regional coexistence of more or less
equal competitors. Journal of Animal Ecology 72:
Ebert, D., Haag, C., Kirkpatrick, M., Riek, M., Hot-
tinger, J.W. & Pajunen, V.I. (2002) A selective
advantage to immigrant genes in a Daphnia meta-
population. Science 295: 485–8.
Etienne, R.S. & Heesterbeek, J.A.P. (2001) Rules of
thumb for conservation of metapopulations based
on a stochastic winking-patch model. American
Naturalist, 158: 389–407.
Figuerola, J. & Green, A.J. (2002) Dispersal of aqua-
tic organisms by waterbirds: a review of past re-
search and priorities for future studies. Freshwater
Biology 47: 483–94.
Githiru, M. & Lens, L. (2004) Using scientific evi-
dence to guide the conservation of a highly frag-
mented and threatened Afrotropical forest. Oryx
38: 404–9.
Gonzalez, A., Lawton, J.H., Gilbert, F.S., Blackburn,
T.M. & Evans-Freke, I. (1998) Metapopulation
dynamics, abundance, and distribution in a micro-
ecosystem. Science 281: 2045–7.
Grenfell, B. & Harwood, J. (1997) (Meta)population
dynamics of infections diseases. Trends in Ecology
and Evolution 12: 395–9.
Gundersen, G., Johannesen, E. Andreassen, H.P. &
Ims, R.A. (2001) Source–sink dynamics: how sinks
affect demography of sources. Ecology Letters 4:
Hanski, I. (1994) Patch occupancy dynamics in frag-
mented landscapes. Trends in Ecology and Evolution
9: 131–5.
Hanski, I. (1999) Metapopulation dynamics. Nature
396: 41–9.
Hanski, I. & Gilpin, M. (1991) Metapopulation dy-
namics : brief history and conceptual domain. Bio-
logical Journal of the Linnean Society 42: 3–16.
Hanski, I. & Simberloff, D. (1997) The metapopula-
tion approach. In Metapopulation Biology: Ecology,
Genetics and Evolution (Eds I. Hanski & M. Gilpin),
pp. 5–26. Academic Press, San Diego, CA.
Hanski, I., Po
yry, J., Pakkala, T. & Kuussaari, M.
(1995) Multiple equilibria in metapopulation
dynamics. Nature 377: 618–21.
Harding, K.C. & McNamara, J.M. (2002) A unifying
framework for metapopulation dynamics. Ameri-
can Naturalist
160: 173–185.
Harrison, S. & Hastings, A. (1996) Genetic and evo-
lutionary consequences of metapopulation struc-
ture. Trends in Ecology and Evolution 11: 180–3.
Harrison, S. & Quinn, J.F. (1989) Correlated envir-
onments and the persistence of metapopulations.
Oikos 56: 293–8.
Hastings, A. (2003) Metapopulation persistence with
age-dependent disturbance or succession. Science
301: 1525–6.
Hastings, A. & Harrison, S. (1994) Metapopulation
dynamics and genetics. Annual Review of Ecology
and Systematics 25: 167–88.
Hess, G. (1996) Disease in metapopulation models
implications for conservation. Ecology 77: 1617–32.
Hilty, J.A. & Merenlender, A.M. (2004) Use of riper-
ian corridors and vineyards by mammalian pred-
ators in Northern California. Conservation Biology
18: 126–35.
Hudgens, B.R. & Haddad, N.M. (2003) Predicting
which species will benefit from corridors in frag-
mented landscapes with population growth
models. American Naturalist 161: 808–20.
Johnson, M.P. (2000) The influence of patch demo-
graphices on metapopulations, with particular ref-
erence to successional landscapes. Oikos 88: 67–74.
Johst, K., R. Brandl & S. Eber. (2002) Metapopula-
tion persistence in dynamic landscapes: the role of
dispersal distance. Oikos 98: 263–70.
Keith, D. (2004) Population viability analysis of the
endangered Australian heath shrub, Epacris bar-
bata, subject to recurring fires and disease epidem-
ics. In Species Conservation and Management: Case
Studies (Eds H.R. Akc¸akaya, M.A. Burgman,
O. Kindvall, et al.), pp. 90–103.. Oxford University
Press, Oxford.
Kelly, C.K. & Bowler, M.G. (2002) Coexistence and
relative abundance in forest trees. Nature (London)
417: 437–40.
Keymer, J.E., Marquet, P.A., Velasco-Hernandez,
J.X. & Levin, S.A. (2000) Extinction thresholds
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 82 6.5.2006 2:48am
and metapopulation persistence in dynamic land-
scapes. American Naturalist 156: 478–94.
Kindvall, O. & Bergman, K-O. (2004) Woodland
brown butterfly Lopinga achine in Sweden. In Spe-
cies Conservation and Management: Case Studies (Eds
H.R. Akc¸akaya, M.A. Burgman, O. Kindvall, et al.),
pp. 171–178. Oxford University Press, Oxford.
Lacy, R.C. (2000) Considering threats to the viability
of small populations with individual-based
models. Ecological Bulletins 48: 39–51.
LaHaye, W.S., Gutierrez, R.J. & Akc¸akaya, H.R.
(1994) Spotted owl meta-population dynamics in
southern California. Journal of Animal Ecology
63: 775–785.
Lamberson, R.H., McKelvey, K., Noon, B.R. & Voss,
C. (1996) A dynamic analysis of Northern Spotted
Owl viability in a fragmented forest landscape.
Conservation Biology 6: 505–12.
Lens, L., Van Dongen, S., Norris, K., Githiru, M. &
Matthysen, E. (2002) Avian persistence in frag-
mented rainforest. Science 298: 1236–8.
Liebhold, A., Koenig, W.D. & Bjornstad, O.N. (2004)
Spatial synchrony in population dynamics. Annual
Review of Ecology Evolution and Systematics 35: 467–
Lindenmayer, D.B. & Possingham, H.P. (1996) Rank-
ing conservation and timber management options
for Leadbeater’s Possum in southeastern Australia
using population viability analysis. Conservation
Biology 10: 235–51.
Lindsey, P.A., Alexander, R., du Toit, J.T. & Mills,
M.G.L. (2005) The cost efficiency of wild dog con-
servation in South Africa. Conservation Biology
19: 1205–14.
Macdonald, D.W. & Sillero-Zubiri, C. (2004) Biology
and Conservation of Wild Canids. Oxford University
Press, Oxford.
McCarthy, M.A. & Lindenmayer, D.B. (2000) Spa-
tially-correlated extinction in a metapopulation
model of Leadbeater’s possum. Biodiversity and Con-
servation 9: 47–63.
McCarthy, M.A., Menkhorst, P.W., Quin, B.R.,
Smales, I.J. & Burgman, M.A. (2004) Helmeted
Honeyeater (Lichenostomus melanops cassidix)in
Southern Australia: assessing options for establish-
ing a new wild population. In Species Conservation
and Management: Case Studies (Eds H.R. Akc¸akaya,
M.A. Burgman, O. Kindvall, et al.), pp. 410–20.
Oxford University Press, Oxford.
Mills, M.G.L. & Gorman, M.L. (1997) Factors affect-
ing the density and distribution of wild dogs in the
Kruger National Park. Conservation Biology 11:
Mills, M.G.L., Ellis, S., Woodroffe, R., et al. (Eds)
(1998) African Wild Dog (Lycaon pictus) Population
and Habitat Viability Assessment.ConservationBreed-
ing Specialist Group, Species Survival Commission,
International Union for the Conservation of Nature
and Natural Resources, Apple Valley, MN.
Noon, B.R., McKelvey, K.B. & Murphy, D.D. (1997)
Developing an analytical context for multispecies
conservation planning. In The Ecological Basis for
Conservation: Heterogeneity, Ecosystems, and Biodiver-
sity (Eds S.T.A. Pickett, R.S. Ostfeld, M. Shachak &
G.E. Likens), pp. 43–59. Chapman and Hall,
New York.
Noss, R.F. (1990) Indicators for monitoring biodiver-
sity: a hierarchal approach. Conservation Biology
4: 355–64.
Nur, N., Page, G.W. & Stenzel, L.E. (1999) Popula-
tion viability analysis for Pacific coast Snowy
plovers. Appendix D of ‘Western Snowy Plover
(Charadrius alexandrinus nivosus) Pacific Coast
Population Draft Recovery Plan’. U.S. Fish and
Wildlife Service, Region 1, 2001, Portland, Ore-
gon, xix þ 630 pp.
Pulliam, H. P., Dunning, J.B. & Liu, J. (1992) Popu-
lation dynamics in complex landscapes: a case
study. Ecological Applications 2: 165–77.
Ranta, E., Kaitala, V. & Lindstro
m, J. (1999) Spatially
autocorrelated disturbances and patterns in popu-
lation synchrony. Proceedings of the Royal Society of
London, Series B Biological Sciences 266: 1851–6.
Regan, H.M. & Auld, T.D. (2004) Australian shrub
Grevillea caleyi: recovery through management of
fire and predation. In Species Conservation and Man-
agement: Case Studies (Eds H.R. Akc¸akaya, M.A.
Burgman, O. Kindvall, et al.), pp. 23–35. Oxford
University Press, Oxford.
Regan, T.J. & Bonham, K. (2004) Carnivorous land
snail Tasmaphena lamproides in Tasmania: effects of
forest harvesting. In Species Conservation and Man-
agement: Case Studies (Eds H.R. Akc¸akaya, M.A.
Burgman, O. Kindvall, et al.), pp. 112–4. Oxford
University Press, Oxford.
Regan, H.M., Auld, T.D., Keith D.A. & Burgman,
M.A. (2003) The effects of fire and predators on
the long-term persistence of an endangered
shrub, Grevillea caleyi. Biological Conservation 109:
Rondinini, C. & Doncaster, C.P. (2002) Roads as bar-
riers for hedgehogs. Functional Ecology 16: 504–9.
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 83 6.5.2006 2:48am
Root, K.V., Akc¸akaya H.R. & Ginzburg, L.R. (2003) A
multi-species approach to ecological valuation and
conservation. Conservation Biology 17: 196–206.
Schtickzelle, N. & Baguette, M. (2004) Metapopula-
tion viability analysis of the bog fritillary butterfly
using RAMAS/GIS. Oikos 104: 277–90.
Smales, I., Quin, B., Krake, D., Dobrozczyk, D. &
Menkhorst, P. (2000) Re-introduction of helmeted
honeyeaters, Australia. Re-introduction News
(Newsletter of the Re-introduction Specialist
Group of the Species Survival Commission, Inter-
national Union for the Conservation of Nature and
Natural Resources) 19: 34–6.
Smedbol, R.K. & Stephenson, R.L. (2004) Atlantic
herring (Clupea harengus) in the northwest Atlantic
Ocean: dynamics of nested population compon-
ents under several harvest regimes. In Species Con-
servation and Management: Case Studies (Eds H.R.
Akc¸akaya, M.A. Burgman, O. Kindvall, et al.),
pp. 245–55. Oxford University Press, Oxford.
St Clair, C.C. (2003) Comparative permeability of
roads, rivers, and meadows to songbirds in Banff
National Park. Conservation Biology 17: 1151–60.
Stelter, C., Reich., M., Grimm, V. & Wissel, C. (1997)
Modelling persistence in dynamic landscapes: lessons
from a metapopulation of the grasshopper Bryodema
tuberculata. Journal of Animal Ecology 66: 508–18.
Teodorascu, D. (1997) Infra Eco Network Europe
an open European platform for cooperation and
exchange of information in the field of habitat
fragmentation caused by transportation infra-
structure: In Proceedings Habitat Fragmentation and
Infrastructure (Eds K. Canters, A. Piepers and
D. Hendriks-Heersma), pp. 442–50. DWW Publi-
cations, Delft.
Thomas, J.A., Rose, R.J., Clarke, R.T., Thomas, C.D.
& Webb, N.R. (1999) Intraspecific variation in
habitat availability among ectothermic animals
near their climatic limits and their centres of
range. Functional Ecology 13(Supplement 1): 55–64.
Tilman, D., Lehman, C.L. & Yin, C. (1997) Habitat
destruction, dispersal, and deterministic extinction
in competitive communities. American Naturalist
149: 407–35.
Travis, J.M.J. & Park, K.J. (2004) Spatial structure
and the control of invasive alien species. Animal
Conservation 7: 321–330.
Van Dyk, G. & Slotow, R. (2003) The effects of fences
and lions on the ecology of African wild dogs re-
introduced to Pilansberg National Park, South Af-
rica. African Zoology 38: 79–94.
ck, J. (2004) European Mudminnow
(Umbra krameri) in the Austrian floodplain of the
river Danube: conservation of an indicator
species for endangered wetland ecosystems in Eur-
ope. In Species Conservation and Management: Case
Studies (Eds H.R. Akc¸akaya, M.A. Burgman,
O. Kindvall, et al.), pp. 200–7. Oxford University
Press, Oxford.
Watkinson, A.R. & Sutherland, W.J. (1995) Sources,
sinks and pseudo-sinks. Journal of Animal Ecology
64: 126–30.
Macdonald/Key Topics in Conservation Biology 1405122498_4_005 Final Proof page 84 6.5.2006 2:48am
... Setyawan et al. (2018) first suggested that the large and widely distributed population of M. alfredi in the Raja Ampat archipelago might consist of several subpopulations, with limited exchange of individuals between these subpopulations. Such a situation is perhaps best described as a metapopulation (Hanski & Gilpin 1991, Wells & Richmond 1995, operationally defined by Akçakaya et al. (2007) as a set of discrete (sub)populations of the same species inhabiting the same general geographical region, between which individuals move through migration and dispersal. Key requirements of the Akçakaya et al. (2007) metapopulation definition include 1) (sub)populations are geographically discrete, and 2) mixing of individuals between (sub)populations is less than that within them (otherwise they should be considered a single panmictic population). ...
... Such a situation is perhaps best described as a metapopulation (Hanski & Gilpin 1991, Wells & Richmond 1995, operationally defined by Akçakaya et al. (2007) as a set of discrete (sub)populations of the same species inhabiting the same general geographical region, between which individuals move through migration and dispersal. Key requirements of the Akçakaya et al. (2007) metapopulation definition include 1) (sub)populations are geographically discrete, and 2) mixing of individuals between (sub)populations is less than that within them (otherwise they should be considered a single panmictic population). ...
... Though not used in the M. alfredi literature, the metapopulation concept (Hanski & Gilpin 1991) seems to aptly describe the M. alfredi population dynamics detailed in all of these studies and the present one. Akçakaya et al. (2007) define a metapopulation as a set of discrete (sub)populations of the same species inhabiting the same general geographical region, between which individuals move through migration and dispersal, with key requirements that (sub)populations are geographically discrete, and that mixing of individuals between (sub)populations is less than that within them. ...
Full-text available
The Bird's Head Seascape (BHS) in West Papua, Indonesia, is widely recognized as the global epicenter of coral reef biodiversity and is protected by an extensive network of 20 marine protected areas (MPAs) totaling over 4.7 million ha. It is home to large populations of both the reef manta ray Mobula alfredi (Krefft, 1868) and the oceanic manta Mobula birostris (Walbaum, 1792). We document the natural history of manta rays in the BHS and describe the demographics and spatial ecology of Mobula alfredi using underwater and aerial observations, a comprehensive photo-ID database, and passive acoustic telemetry. Manta rays were recorded from 127 sites across the BHS, including 70 aggregation sites (cleaning stations and routine feeding aggregations), with the largest feeding aggregation recorded consisting of 112 M. alfredi in the Dampier Strait in the Raja Ampat archipelago. We recorded 4,052 photographically identified M. alfredi sightings of 1,375 individuals between November 2004 and December 2019, with a biased female-to-male sex ratio of 1.58 to 1.0 and 67.4% exhibiting the chevron color morph vs. 32.6% melanistic. Over 85% of sightings came from the two large MPAs (>330,000 ha) of South East Misool and Dampier Strait. Importantly, 16 photo-IDs of somersault-feeding individuals were obtained using a drone, apparently the first report of UAVs used for manta photo-IDs. We resighted 642 individuals (46.7%) at least once during the period, with the two most-resighted individuals registering 67 and 66 resightings over periods of about 12 years. We observed 217 females pregnant at least once, with one having 4 consecutive pregnancies from 2013-16 (and a total of 5 pregnancies in 7 years) and 15 with at least two consecutive-year pregnancies. Four nursery sites were identified with a consistent presence of numerous young-of-the-year (YoY; i.e. ≤2 m disc width) over 3-14 years of observations: we recorded 65 YoYfrom Raja Ampat. The Raja Ampat population is best described as a metapopulation composed of 4-7 subpopulations inhabiting island groups separated by over-water distances of only 20-40 km, but which nonetheless exhibit limited exchange of individuals. We recorded 309 movement events among 7 hypothesized manta subpopulations in Raja Ampat based on photo-IDs between 2004 and 2019 and passive acoustic telemetry between 2013 and 2019, with the longest movement we recorded 296 km minimum distance through water. Importantly, 115 of the identified manta ray sites (90.5%) are distributed within 13 of the 20 BHS MPAs, and 95.9% of sightings (3,887 of 4,052), 89.5% of individuals (1,231 of 1,375) and all 4 identified nursery areas were from within MPAs in Raja Ampat, indicating the Raja Ampat MPA network, and the broader BHS MPA network within which it is nested, are critical for the conservation of manta rays in West Papua.
... The metapopulation concept has been a very powerful one in ecology, raising interest in research areas like dispersal ecology [1][2][3] or population genetics [4][5][6][7][8]. The concept also had a strong impact on the development of conservation concepts [9][10][11][12]. Yet there are also issues about its generality as the occurrence of metapopulations may be restricted to a rather small region in parameter space with many spatially structured populations basically behaving like patchy populations, mainland-island systems, non-equilibrium metapopulations or panmictic populations [13,14]. ...
Full-text available
Background: The idea that populations are spatially structured has become a very powerful concept in ecology, raising interest in many research areas. However, despite dispersal being a core component of the concept, it typically does not consider the movement behavior underlying any dispersal. Methods: Using individual-based simulations in continuous space, we investigate the emergence of a spatially structured population in landscapes with spatially heterogeneous resource distribution and with organisms following a simple area-concentrated search (ACS). Results: We found that foraging success increased with increasing resource density and decreasing number of resource clusters. In a wide parameter space, the system exhibited attributes of a spatially structured populations with individuals concentrated in areas of high resource density, searching within areas of resources, and 'dispersing' in straight line between resource patches. 'Emigration' was more likely from patches that were small or of low quality (low resource density). 'Looping' from patches was more likely if patches were large and of high quality. With the ACS implemented, individuals tended to move deeper into a resource cluster in scenarios with moderate resource density than in scenarios with high resource density because of more space between resource items. Conclusion: Our simulations demonstrate that spatial structure in populations may emerge if critical resources are heterogeneously distributed and if individuals follow simple movement rules (such as ACS). Neither the perception of habitat nor an explicit decision to emigrate on the side of acting individuals are necessary for the emergence of spatial structure. Mechanistic models like ours may help to close the gap between movement ecology and spatial ecology theory.
... Ensuring the persistence of such large carnivores in insular or isolated patches often requires management interventions to be implemented at the metapopulation level in addition to those that focus on securing individual populations (Dolrenry et al., 2014). Such conservation strategies can account for movement of individuals between patches, facilitate gene flow between sub-populations and allow for rescue effects, thereby prolonging the overall viability of populations (see Akçakaya et al., 2007). ...
Elevated rates of anthropogenic impacts on land‐use regimes have pushed terrestrial megafauna to the brink of extinction. Consequently, it is critical to adopt conservation approaches that safeguard individual populations, while retaining connectivity among these populations. Conserving spatially structured populations of imperiled species at large scales is often complex; the past decades have therefore seen a rise in spatial conservation prioritization exercises aimed at shaping landscape‐scale conservation programmes. We present a framework for informing nationwide connectivity conservation, linking ecological and administrative scales, to maximise relevance for management. We assessed connectivity of the endangered dhole Cuon alpinus among 155 potential source populations across India using a data‐driven approach combined with graph and circuit theory. We used clustering algorithms to identify ecologically meaningful conservation landscapes; within each landscape, we identified priority source populations based on their connectedness, and quantified pixel‐specific habitat accessibility. We superimposed administrative boundaries on our findings to provide conservation recommendations at this management‐relevant scale. We first mapped potential dhole movement across India. Dhole populations fell within three primary clusters—Western and Eastern Ghats (WEG), Central Indian Landscape (CIL), and North‐East India (NEI)—of which NEI had the highest forest cover, most diffuse connectivity, and lowest human density, while WEG had the highest protected area coverage, and overall connectedness. Within each conservation landscape we evaluated the relative importance of Protected Areas and accessibility to high‐quality patches. Parts of the Eastern Ghats had low habitat accessibility, yet high potential for dhole landscape connectivity. In 114 identified administrative units of priority for habitat restoration, we highlight those with low accessibility, i.e., areas where restoration needs to be spatially targeted for maximum benefits. Synthesis and applications. We make recommendations for spatially‐informed habitat restoration to enhance dhole connectivity in India, highlighting the importance of improving matrix permeability where dhole movement is currently restricted. More broadly, the framework we present is useful across species and management contexts, as it combines spatial and administrative scales to make ecologically‐informed assessments of high relevance to management. Synergistically integrating species ecology, threats, and administrative considerations in connectivity conservation plans can enhance success of species conservation programmes.
... Recently, a broader refugial concept has gained traction in the identification of areas that provide intermediate refuges and long-term refugia from biotic and abiotic conditions to create population holdouts in the face of rapid land-use and climatic change (Keppel et al. 2012, Monsarrat et al. 2019. Although the importance of refuges in ecology and conservation biology are widely recognized (Akçakaya et al. 2006), and would be ideal targets for translocations and species recovery, they are rarely identified prior to reintroductions (although see Struebig et al. [2015], Conner et al. [2018]). Indeed, information on those fundamental attributes of a potential refuge including resistance to environmental change (i.e., landscapes that resemble historical conditions and are protected from future perturbations), connectivity (i.e., the potential for dispersers to naturally recolonize surrounding areas), and demographic potential (i.e., a net exporter of individuals) are rarely documented, even though they would enhance both translocation success and the future persistence of recovering populations. ...
Rapid environmental change is reshaping ecosystems and driving species loss globally. Carnivore populations have declined and retracted rapidly and have been the target of numerous translocation projects. Success, however, is complicated when these efforts occur in novel ecosystems. Identifying refuges, locations that are resistant to environmental change, within a translocation framework should improve population recovery and persistence. American martens (Martes americana) are the most frequently translocated carnivore in North America. As elsewhere, martens were extirpated across much of the Great Lakes region by the 1930s and, despite multiple translocations beginning in the 1950s, martens remain of regional conservation concern. Surprisingly, martens were rediscovered in 2014 on the Apostle Islands of Lake Superior after a putative absence of >40 years. To identify the source of martens to the islands and understand connectivity of the reintroduction network, we collected genetic data on martens from the archipelago and from all regional reintroduction sites. In total, we genotyped 483 individual martens, 43 of which inhabited the Apostle Islands (densities 0.42‐1.46/km2). Coalescent analyses supported the contemporary recolonization of the Apostle Islands with progenitors likely originating from Michigan, which were sourced from Ontario. We also identified movements by a first‐order relative between the Apostle Islands and the recovery network. We detected some regional gene flow, but in an unexpected direction: individuals moving from the islands to the mainland. Our findings suggest that the Apostle Islands were naturally recolonized by progeny of translocated individuals and now act as a source back to the reintroduction sites on the mainland. We suggest that the Apostle Islands, given its protection from disturbance, complex forest structure, and reduced carnivore competition, will act as a potential refuge for marten along their trailing range boundary and a central node for regional recovery. Our work reveals that translocations, even those occurring along southern range boundaries, can create recovery networks that function like natural metapopulations. Identifying refuges, locations that are resistant to environmental change, within these recovery networks can further improve species recovery, even within novel environments. Future translocation planning should a priori identify potential refuges and sources to improve short‐term recovery and long‐term persistence.
... Protected area networks are usually designed for multiple species (Margules and Pressey, 2000), though making decisions for multiple species is challenging (Albert et al., 2017). To combat this, conservation managers often select 'umbrella' species, using one species with similar needs as a surrogate for others (Akcakaya et al., 2007), thereby assuming that the species community as a whole will benefit (Bennett et al., 2015). Previous studies have demonstrated the capacity of parameters in metapopulation models to predict dynamic patterns of other related species (e.g. ...
Full-text available
Protected area networks seek to ensure the persistence of multiple species, but their area and extent are limited by available land and conservation resources. Prioritising sites based on their quality, quantity, size, or connectivity is often proposed; potentially using the occupancy and metapopulation dynamics of individual threatened species as surrogates for network effectiveness. However, the extent to which the dynamics of species with overlapping habitat requirements differ, and the implications of this for the optimal network designs for multiple species, are rarely tested. We parameterise metapopulation models for 5 papyrus-specialist birds occupying a network of papyrus swamp in Uganda, each of which possess subtly different ecological characteristics and habitat preferences. We estimate how each responds to different strategies based on prioritising patch size, number, quality and connectivity. The optimal approach differed depending on the metapopulation structure and characteristics of each species. The rank order of strategies also varied with the overall wetland area available and the desired persistence threshold. For individual species, prioritising habitat quality achieved the highest levels of persistence and population size for an equivalent amount of land area conserved. However, connected patches showed greatest overlap across species, thus the most effective strategy to conserve multiple species in the same network prioritised habitat connectivity. This emphasises the importance of individual species' characteristics using the same habitat networks in conservation planning, and demonstrates the utility of prioritising protected sites based on the spatial connectivity of habitat patches, when aiming to conserve multiple species with differing or uncertain habitat requirements.
... The metapopulation concept provides an operational framework for both (evolutionary) ecologists and conservation managers [1]. Classically defined as a set of interacting populations for which frequent local extinctions are balanced by recolonization [2,3], a metapopulation can also broadly refer to patchy populations [4], that is, to any set of local populations potentially related by movements of individuals in a landscape [5]. ...
Full-text available
Understanding the functioning of natural metapopulations at relevant spatial and temporal scales is necessary to accurately feed both theoretical eco-evolutionary models and conservation plans. One key metric to describe the dynamics of metapopulations is dispersal rate. It can be estimated with either direct field estimates of individual movements or with indirect molecular methods, but the two approaches do not necessarily match. We present a field study in a large natural metapopulation of the butterfly Boloria eunomia in Belgium surveyed over three generations using synchronized demographic and genetic datasets with the aim to characterize its genetic structure, its dispersal dynamics, and its demographic stability. By comparing the census and effective population sizes, and the estimates of dispersal rates, we found evidence of stability at several levels: constant inter-generational ranking of population sizes without drastic historical changes, stable genetic structure and geographically-influenced dispersal movements. Interestingly, contemporary dispersal estimates matched between direct field and indirect genetic assessments. We discuss the eco-evolutionary mechanisms that could explain the described stability of the metapopulation, and suggest that destabilizing agents like inter-generational fluctuations in population sizes could be controlled by a long adaptive history of the species to its dynamic local environment. We finally propose methodological avenues to further improve the match between demographic and genetic estimates of dispersal.
... Minimum area requirements vary greatly among species; for example, some invertebrates may have minimum area requirements of 1 m 2 , while some large carnivores need more than 1000 km 2 (Ewers and Kapos, 2011). Larger patches also can provide animals and plants with high quality interior habitat and sustain large population sizes (Laurance, 2000;Akçakaya et al., 2007), and support more species (MacArthur and Wilson, 1967). However, this does not mean that smaller patches are useless and can be removed when 6. Relative (%) changes of habitat area (dA) and connectivity (dECA) given by the effect of removing forest patches smaller than 5, 10, 20 and 50 ha in the studied landscape. ...
Full-text available
Habitat fragmentation hinders the dispersal of species and reduces the range of suitable habitat, thereby threatening the conservation of biodiversity. Even in protected areas, an assessment of the landscape connectivity of suitable habitat for species is still essential. In this paper, we investigated the overall forest connectivity, and determined patch characteristics and their role in species dispersion for species with different dispersal abilities in the World Natural Heritage Site of Bogda, Xinjiang, China. In our study site, the overall landscape connectivity was low for species with short- and medium-distance (≤500 m) dispersal, but high for species with long-distance (>500 m) dispersal abilities. We ranked the importance of patches according to their role in maintaining overall connectivity. Two patches were identified as being the most important: one mainly provided habitat area and fluxes (i.e., the flow of species into and out of the area) for species, while the other acted as a ‘stepping-stone’ for dispersal and provided fluxes. All species could use smaller patches (≤50 ha) as stepping-stones, and some of these patches could provide special environmental conditions for endemic species with short- and medium-distance dispersal abilities. Our study offers a way to prioritize the conservation of patches in a forest network for biodiversity and ecosystem health.
... This would highlight aspects related to metapopulation processes and sourcesink dynamics among islands and/or subpopulations within each archipelago. Given that the viability of a population may depend on surrounding populations, this information would greatly benefit the delineation of conservation strategies and reserve spatial planning (Akçakaya et al. 2007). ...
Full-text available
The uptake of natural living resources for human consumption has triggered serious changes in the balance of ecosystems. In the archipelagos of Macaronesia (NE Atlantic), limpets have been extensively exploited probably since islands were first colonized. This has led to profound consequences in the dynamics of rocky shore communities. The specific objectives of this thesis were to: 1) develop and characterize species-specific microsatellite markers for the limpets Patella candei (d’Orbigny 1840) and Patella aspera (Röding 1798), endemic to the Macaronesia archipelagos; 2) assess their genetic diversity, population structure and contemporary levels of connectivity throughout Macaronesia; 3) conduct a morphometric analysis of the P. candei complex to complement molecular data; 4) evaluate the temporal and spatial variation in recruitment of P. candei and study its association with real-time environmental data; 5) assess the effect of temperature on larval development of P. candei; and 6) provide general recommendations to foster the sustainable exploitation of limpets in Macaronesia. A total of twelve and seventeen microsatellite markers were described for P. candei and P. aspera, respectively. These showed clean polymorphisms and speciesspecific markers were combined in three optimized multiplex reactions. For P. candei, a highly significant genetic break between archipelagos following isolation by distance was detected. Contrastingly, significant genetic differentiation among islands (i.e. Azores) was absent possibly indicating ongoing gene flow via larval exchange between populations. Significant shell shape differences among archipelagos were also detected using both distance-based and geometric morphometric analyses. Adaptive processes associated with niche differentiation and strong barriers to gene flow among archipelagos may be the mechanisms underlying P. candei diversification in Macaronesia. As for P. aspera, genetic analyses showed significant population structure between populations from Azores and populations from Madeira and Canaries, and absence of current or historic gene flow between these. Results also suggest that both population clusters have experienced demographic changes over time. Heterozygote deficits were common across populations, which can be better accounted for by inbreeding than by null alleles or Wahlund effect. Such levels of inbreeding are likely a consequence of a significant reduction of reproductive units due to decades of intense exploitation. The monitoring program applied to track P. candei recruitment showed that early recruits occurred throughout the entire duration of the program, but its intensity varied in space and time. In general, a marked peak in recruitment occurred during winter/spring months, the period of greatest reproductive activity, when sea surface temperatures are lower and wave turbulence higher. Significant wave height was probably the most important proximate cue triggering the recruitment of P. candei, which eventually depends on adequate ultimate drivers for spawning and reproduction (i.e. temperature). Indeed, as a winter-breeder, P. candei larvae seem to perform better and attain higher fitness at colder temperatures. In fact, experimental treatments on larval rearing showed that larval development was faster at increasing temperatures but cumulative survivorship decreased; about 25% of larvae at higher temperatures survived to the end of the experiment, a 2-fold decrease from the average survivorship of ~ 50% at lower temperatures. Overall, the outcomes of this thesis fill a gap in our knowledge about processes involved in determining the connectivity patterns between limpet populations and the environmental factors influencing such patterns across the Macaronesia region. The present study is an important first step in this direction of using multi-faceted approaches to understand complex processes operating at the marine environment, while providing a fundamental asset to define stocks and thus inform specific conservation strategies that foster the sustainable exploitation of limpets throughout Macaronesia archipelagos.
The southern coastal cliffs of Reunion Island (tropical island in the Western Indian Ocean) host a unique flora and fauna: the last populations of Manapany day gecko (Phelsuma inexpectata, an endemic reptile in critically endangered), relics of indigenous vegetation including endemic and/or threatened species (e.g.: Euphorbia viridula, Psiadia retusa, Latania lontaroides) and breeding colonies of three native seabirds (white-tailed tropicbirds, Phaethon lepturus ; brown noddies, Anous stolidus and wedge-tailed shearwaters, Ardenna pacifica). This biodiversity is threatened by habitat transformations due to invasive plants, human activities (urbanization and culture) and invasive mammals (especially cats, felis catus, and rodents). Moreover, little is known about biology and ecology of the native species, which does not allow the implementation of effective conservation strategy. Based on hand-in-hand collaboration between researchers (UMR ENTROPIE) and managers (CDL, NOI, AVE2M) working on different taxa, the aim of this thesis was to provide multispecies conservation prescriptions on cliffs study for the Manapany day gecko and the wedge-tailed shearwater. We undertook a progressive approach from describing of species conservation states through understanding threatening processes to the prescription and monitoring of management actions. Three research topics were targeted: (i) demography and reproductive biology, (ii) terrestrial habitat requirements, and (iii) impacts and management of invasive mammals (especially cats). Our results highlighted the critical conservation state of Manapany day geckos and wedge-tailed shearwaters populations. Invasive plants and mammals (especially cats) are threats to the conservation of native biodiversity. We provide several local and general conservation prescriptions, including management of invasive species, multispecies terrestrial habitat restoration and captive head-start program of Manapany day geckos. Several of these prescriptions were implemented during this thesis (invasive species management and captive breeding program) and monitored as part of active adaptive management approach. This multispecies study at the interface between research and management must be continued and supported by a strong federating regulatory tool as a National Nature Reserve (NNR). Keywords: Ardenna pacifica, biological invasions, captive head-start program, Capture-Mark-Recapture, cat control, cat tracking, conservation biology, Felis catus, habitat selection, multispecies management, Phelsuma inexpectata, population dynamics, Population Viability Analyses, Reunion Island, Spatial Mark-Resight, tropical island
Technical Report
Full-text available
1. Recent studies have unveiled drastic declines in the diversity and numbers of insects worldwide, unfolding over the last decades. These results have brought insects to the forefront of conservation attention, ranging from mitigation actions planned by dedicated conservation agencies to efforts undertaken by the general public. Further, the conservation focus has necessarily shifted from single (highly endangered) species to multi-species plans, but such a shift can become a daunting task when dealing with several dozens of species. In the last two decades powerful methods have emerged to describe and analyze populations inhabiting fragmented landscapes that are the dominating landscape type in most highly industrialized countries. Moreover, especially studies of butterfly species have enabled deep metapopulation insights, in theory and practice. 2. The quantitative framework I present consists of two connected components. The first aims at mapping a butterfly species list onto habitat patches in a landscape to be restored, uncovering the structure of the butterfly-specific metapopulations. This mapping relies mainly on larval host plant species and the vegetation types they inhabit, but it can accommodate additional resources that ultimately foster the species’ presence. Further, the mapping produces additional data that can conveniently be analyzed using methods from network theory, yielding ancillary insights for restoration planning. Subsequently, the second component aims at analyzing the metapopulations guided by metapopulation theory, to quantify the effect of available restoration actions on ecological and genetic aspects. The weighted, summarized results at the community level are now amenable to decision analysis that ultimately allows selecting the most effective and efficient strategies in a rational way, based on project objectives, strategies, and constraints. I hope the presented framework – as a whole or parts of it – may help inform and guide butterfly restoration projects.
Full-text available
Contemporary acceleration of biodiversity loss makes increasingly urgent the need to understand the controls of species coexistence(1,2). Tree diversity in particular plays a pivotal role in determining terrestrial biodiversity, through maintaining diversity of its dependent species(3,4) and with them, their predators and parasites. Most theories of coexistence based on the principle of limiting similarity suggest that coexistence of competing species is inherently unstable; coexistence of competitors must be maintained by external forces such as disturbance(5,6), immigration(7) or 'patchiness' of resources in space and time(8,9). In contrast, storage theory postulates stable coexistence of competing species through temporal alternation of conditions favouring recruitment of one species over the other(10,11). Here we use storage theory to develop explicit predictions for relative differences between competitors that allow us to discriminate between coexistence models. Data on tree species from a primary forest on the Mexican Pacific coast support a general dynamic of storage processes determining coexistence of similar tree species in this community, and allow us to reject all other theories of coexistence.
Full-text available
The habitat is the basic unit for developments in life history, population dynamics, landscape ecology and conservation of organisms. It is frequently treated as a particulate, invariant and homogeneous entity (a patch). Here we examine the implications of using this concept of habitat in butterfly biology. In doing so, we suggest the alternative approach of applying a functional resource-based concept of habitat. This recognises the fundamental requirements of organisms, consumables and utilities, the latter describing suitable environmental conditions as well as essential substrates. We argue that a resource-based concept is critical for butterfly conservation and call for the development of a resource database on butterfly biology.
Wild dogs have been eradicated from most of South Africa. However, a large number of smaller isolated reserves offer the potential for metapopulation conservation management of this species through continued translocations among reserves. Wild dogs were released into the Pilanesberg National Park (500 km2), South Africa, in June 1999 from a combination of wild captured and captive-bred individuals. The reserve has lions but no spotted hyaenas. We document post-release spatial use, prey selection and breeding biology. Dogs used a very small area (13.4 km2) for their first denning period, and then ranged more widely, but avoided the central parts of the park. Movement patters and den site locations suggested that dogs avoided the presence of lions. Major prey species were kudu (50 %), impala (32 %) and waterbuck (14 %). Large prey, including adult male kudu (250 kg) and waterbuck (260 kg) were taken regularly through use of the boundary fence as an aid to capture. Wild dogs have bred three times since introduction, indicating that reserves as small as 200 km2 may be suitable for introduction of wild dogs, and metapopulation management strategies may be a viable option as long as sympatric large predator populations are absent or managed appropriately.
Three mathematical models for assessing extinction risks are introduced. The first is a Monte Carlo model that describes the growth of an age-structured population under environmental and demographic stochasticity. The parameters are the mean value, variation, distribution and correlations of life history traits such as survival, fecundity and migration, and density dependence in recruitment. Probabilities of population decline and extinction are computed as a way to evaluate ecological risk. The other two models describe the dynamics of multiple population systems (metapopulations) given their spatial structure and the rate of migration among populations. One of these metapopulation models is based on the occupancy of habitat patches, and the other on the population dynamics within local populations. These models evaluate the risk of extinction of the whole metapopulation as well as the local populations. Three computer programs (RAMAS library) for building single and multiple population models are described. The programs are used to build age-, stage- and spatially-structured models. They provide summaries regarding the expected number of individuals as a function of time, and the probability that the population size will fall below a threshold.
This chapter discusses butterfly metapopulation dynamics. The metapopulation concept is swiftly spreading into common usage in population and conservation biology. The term “metapopulation” itself is not new, as it was coined by Richard Levins, 25 years ago. It took 20 years before biologists at large began to employ it. The model is based on a generalized incidence function and it provides a simple yet realistic and practical means of modeling real metapopulations. Melitaea cinxia is one of the many European butterflies that have declined during the past decades. More than half of the species not forming local populations failed also to satisfy the second condition, because they have large “mainland” populations. During the past 10 years the metapopulation approach has been applied in several butterfly studies in both North America and Europe. Migration between habitat patches is a key process in metapopulation dynamics. It is not generally agreed that metapopulation dynamics in the sense described in this chapter are often critical for long-term persistence of butterflies and other taxa. It is difficult to disagree about rampant environmental changes, and it is equally clear that many populations have been exterminated because the habitat became unsuitable. Melitaea cinxia may be exceptional, but until a score of other equally comprehensive studies have been completed, there is no reason to rush to a general conclusion.
1. It has been suggested that the habitats which a species occupies can be divided into sources and sinks, depending on whether or not local reproduction is sufficient to balance mortality. Source populations are those where reproduction exceeds mortality, surplus individuals dispersing to sink populations where mortality exceeds local reproduction. Sink populations would not be viable in the absence of immigration. 2. A difference equation model is constructed to show that sources and sinks cannot be identified from a simple comparison of the demographic rates between populations, as measured by the numbers of births and deaths. 3. Viable populations may appear to be non-viable simply because the dispersal of individuals into them depresses fecundity or increases mortality as a result of density-dependence. The consequence is that local recruitment appears insufficient to balance local mortality. 4. Viable populations that appear as sinks, as a result of the dispersal of individuals into them, are termed here as `pseudo-sinks'. They will clearly be difficult to distinguish from genuine sinks on the basis of a simple comparison of the numbers of births and deaths in different populations. 5. Examples of source and genuine sink populations and the data required to establish them are discussed.
1. The grasshopper Bryodema tuberculata requires open and dry habitats. In Central Europe, it survives only on gravel bars along braided rivers in the Northern Alps. Even there, many populations of B. tuberculata have became extinct in the last 50 years. 2. The dynamics of braided rivers are characterized by succession and floods. Catastrophic floods occur at irregular intervals. They are capable of washing away entire gravel bars and of building new, vegetation-free gravel bars. Succession eventually leads to an almost complete loss of habitat suitable for B. tuberculata on each single gravel bar. 3. Bryodema tuberculata can persist only as metapopulations, i.e. when local extinctions due to succession or flood events are compensated for by colonization of newly created gravel bars. 4. A simulation model was used to examine how the spatial and temporal dynamics of succession, flood regime and colonization determine the ability of B. tuberculata to survive in flood-plains. 5. The results show that small populations on relatively old gravel bars are important to the persistence of B. tuberculata, even though they usually only survive for a short time, due to demographic noise. 6. The effect of catastrophic floods is ambivalent: persistence is low if time intervals between floods are too short or too long. If floods are too frequent many subpopulations are extinguished at the same time and if hoods are to infrequent, local populations are eliminated by succession. 7. It is concluded that most extinctions of B. tuberculata populations in the Northern Alps are due to changes in the flood regime caused by humans. 8. Many other spatially dynamic animals and plants occupy successional habitats. We suggest that the form of model outlined in this paper, based on a dynamic habitat mosaic, be used for such organisms.